DE-AC07-76ID01570 A Proceedings/Compendium of Papers HOY 2 SOLVENT SUBSTITUTION Based on The First Annual International Workshop on Solvent Substitution December 4-7, 1990 Phoenix, Arizona sponsored by The U.S. Department of Energy Office of Technology Development Environmental Restoration and Waste Management and U.S. Air Force Engineering & Services Center
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
DE-AC07-76ID01570
A Proceedings/Compendium of Papers
HOY 2
SOLVENT SUBSTITUTION
Based onThe First Annual International
Workshop on Solvent SubstitutionDecember 4-7, 1990
Phoenix, Arizona
sponsored by
The U.S. Department of EnergyOffice of Technology DevelopmentEnvironmental Restorationand Waste Management
andU.S. Air ForceEngineering & Services Center
CONF-901285-
DE92 003262
A Proceedings/Compendium of Papers
SOLVENT SUBSTITUTION
Based onThe First Annual International
Workshop on Solvent SubstitutionDecember 4-7, 1990
Phoenix, Arizona
MASTERsponsored by ^ UNUMITSD
OFTH1SDOCUM^..?
The U.S. Department of EnergyOffice of Technology DevelopmentEnvironmental Restorationand Waste Management
andU.S. Air ForceEngineering & Services Center
DE-ACO7-76IDO1570
Proceedings/Compendium of Papers
SOLVENT SUBSTITUTION
based onThe First Annual International
Workshop on Solvent SubstitutionDecember 4-7. 1990
Phoenix, Arizona
The DOE Environmental Restoration and Waste Management Office of Technology Development and the AirForce Engineering and Services Center convened the First Annual International Workshop on Solvent Substitutionon December 4-7, 1990, at rhe Executive Conference Center, The Pointe at Tapatio Cliffs, Phoenix, Arizona.The primary objectives of this joint effort were to:
•Share information and ideas among attendees in order toenhance the development and implementation of required newtechnologies for the elimination of pollutants associatedwith industrial use of hazardous and toxic solvents.
•Aid in accelerating collaborative efforts and technologytransfer between government and industry for solventsubstitution.
This highly successful 2-1/2 day event brought together over 300 leading national and international experts fromindustry, federal and state government agencies, various branches of the Armed Services, research laboratories,universities and public interest groups.
There were workshop sessions focusing on Alternative Technologies, Alternative Solvents, Recovery/Recycling,Low VOC Materials and Treatment for Environmentally Safe DisposaJ. TTK 35 invited papers presented covereda wide range of solvent substitution activities including: hardware and weapons production and maintenance,paint stripping, coating applications, printed circuit boards, metal cleaning, metal finishing, manufacturing,compliance monitoring and process control monitoring.
This publication includes the majority of these presentations. In addition, in ord.̂ r to further facilitate informationexchange and technology transfer, the U.S. Air Force and DOE solicited additional papers under a general "Callfor Papers." These papers, which underwent review and final selection by a p.̂ er review committee, are alsoincluded in this combined Proceedings/Compendium.
For those involved in handling, using or managing hazardous and toxic sdvents, this document should proveto be a valuable resource, providing the most up-to-date information on current technologies and practices insolvent substitution.
Prepared by theWeapons Complex Monitor Forums
Under DOE Contract No. DE-AC07-76ID01570
TABLE OF CONTENTS
SECTION I - ALTERNATIVE TECHNOLOGIES
Invited Papers Presented at Workshop
Surface Cleaning by Laser AblationH.C. Peebles, N.A. Creager and D.E. Peebles 1
CO2 Pellet Blasting for Paint Stripping/Coatings RemovalWayne N. Schmitz 11
The Evaluation of Alternatives to Ozone-Depleting ChlorofluorocarbonsRobin L. Sellers 15
Spray Forming as a New Processing TechniqueScott A. Ploger and Lloyd D. Watson 25
Reduction of Solvent Use Through Fluxiess SolderingF. Michael Hosking 31
Plasma Stripping of Magnetic ComponentsT.J. Gillespie and T. Mehrhoff 43
Sodium Bicarbonate Blasting for Paint StrippingN.E. Wasson, Jr. and Michael N. Haas 49
Low Toxicity Paint Stripping of Aluminum and Composite Substrates
Nona E. Larson 53
Papers Selected under the "Call for Papers"
Precision Parts Cleaning with Supercritical Carbon DioxidePaula M. Gallagher and Val J. Krukonis 79
Carbon Dioxide Pellet Blasting Paint Removal For PotentialApplication On Warner Robins Managed Air Force Aircraft
Randall B. Ivey 91
Alternative Technologies for Environmental ComplianceJ. Michael Locklin 95
High Pressure Supercritical Carbon Dioxide Efficiency in RemovingHydrocarbon Machine Coolants from Metal Coupons and ComponentsParts
Robert F. Salerno 101
Closed Loop Alternative to the Use of Hazardous Chemicals in IndustryDan F. Suciu I l l
SECTION II - ALTERNATIVE SOLVENTS
Invited Papers Presented at Workshop
Biodegradable Solvent SubstitutionAnne E. Copeland 115
DOE/DOD Solvent Utilization HandbookA.A. Chevei and M.D. Herd 119
The Elimination of Chlorinated, Chlorofluorocarbon and OtherRCRA Hazardous Solvents from the Y-12 Plant's EnrichedUranium Operations
D.H. Johnson, R.L. Patton and L.M. Thompson 121
Printed Circuit Board Defluxing: Alternatives to OzoneDepleting Substances
Dual Cure Photocatalyst SystemsSteven J. Keipert 245
in
Screening of VOC Control Technologies: Technology Optionsand Comparative Costs
Victor S. Engleman 251
SECTION V - TREATMENT FOR ENVIRONMENTALLY SAFE DISPOSAL OFTOXIC SOLVENTS
Invited Papers Presented at Workshop
General Overview of Hazardous Waste IncinerationPhilip C. Lin 257
Chemical Oxidation Treatment of Industrial Organic Waste
Penny M. Wikoff and Dan F. Suciu 275
Papers Selected under the "Call for Papers"
Mediated Electrochemical Oxidation of OrganicsLeonard W. Gray, Robert G. Hickman and Joseph C. Farmer 281
Towards a Protocol to Determine Waste Management Propertiesof Solvent Substitutes
Benerito S. Martinez, Jr., Ricardo B. Jacquez,Walter H. Zachritz II and Martha I. Beach 285
SECTION VI - ISSUES TO CONSIDER
Invited Papers Presented at Workshop
Alternatives to Chlorinated Solvents: Health and EnvironmentalTradeoffs
Katy Wolf 291
Formation of Specifications for New ProductsCaptain Daniel T. Witt 297
Appendix I: Attendees - International Workshopon Solvent Substitution 301
Appendix II: Contributing Authors to Proceedings/Compendium 337
IV
Section I
ALTERNATIVE TECHNOLOGIES
SURFACE CLEANING BY LASER ABLATION
H. C. Peebles, N. A. Creager, and D. E. PeeblesSandia National LaboratoriesAlbuquerque, New Mexico
ABSTRACT
Nd:YAG laser cleaning of metal oxide: from304L stainless steel surfaces has beencharacterized. Thin chromium oxide films canbe completely removed from the surface usinga single 10 nsec pulse of laser radiation withan average surface irradiance greater than 120MW/cm2. Laser etching of thicker iron oxidefilms exhibit a self-limiting effect that preventsoveretching into the stainless steel substrate.
INTRODUCTION
Recent international agreements will virtuallyeliminate the use of chlorinated hydrocarbonsin the cleaning of manufactured parts in thenear future. Effective alternative cleaningtechnologies must be developed to replacecleaning procedures which currently utilizethese solvents. At the present time, twoapproaches are being pursued. The mostcommon approach seeks to identify non-chlorinated substitute solvents which areenvironmentally acceptable. Possiblecandidates include aqueous and terpene basedcleaners as well as simple alcohols. A lesscommon approach involves total elimination ofthe solvent, substituting instead an alternativecleaning technology. Alternative technologiesinclude dry processing techniques such asplasma etching and laser ablation. This paperwill focus on the application of laser ablationas a cleaning technique for material surfaces.Data will be presented characterizing theremoval of metal oxides from 304L stainlesssteel surfaces.
TECHNIQUES AND PROCEDURES
Laser ablation involves the use of very shortpulses of high peat: power laser radiation torapidly heat and vaporize thin layers ofmaterial surfaces. Laser cleaning of a surfaceis accomplished by rastering the laser beamacross the surface of the material. When usedas a cleaning technique, this process must beperformed in a chamber containing an inertgas environment in order to preventrecontamination of the surface by reactive gas-phase species. The ablated surface materialforms a dense cloud of hot vapors which,upon cooling, condenses into submicrondiameter particles. This paniculate wastemust be removed from the near surface regionof the part by entrainment into a flowing gasstream or recontamination of the surface willoccur. One possible method for the removalof the paniculate waste is shown schematicallyin Figure 1. Inert gas from the near surfaceregion of the part is drawn into a nozzlecoaxial with the incident laser beam. The gasentering the nozzle exits through a side portand is then passed through a canister filterwhich removes the paniculate material fromthe waste stream. The filtered inert gas isreturned to the chamber.
Because the incident laser radiation can befocused to a small spot size, laser ablation canclean with high spacial selectivity, allowingapplication on partially or completelyassembled parts which could be damaged byother cleaning methods. The waste generatedby this process is limited to the volume ofmaterial vaporized from the surface. For atypical 1 mm thick iron oxide film with amaterial density of 5.2 g/cm3 (Fe2O3), thewaste generation rate will be approximately480 mg per square foot of surface areacleaned. Thus, as a cleaning process, laser
ablation exhibits high potential for wasteminimization.
The experimental apparatus used tocharacterize the etching of iron and chromiumoxides from 304L stainless steel is shownschematically in Figure 2. Target substrateswere placed in a small stainless steel chamberfitted with an 8 mm thick fused quartzwindow. The chamber was purged to removeoxygen and other reactive gas contaminants inthe ambient atmosphere by evacuation to 1micron pressure followed by backfilling thechamber to 1 atmosphere with 99.999% purehelium gas. This purge process was repeateda total of three times. All etching experimentsdescribed in this paper were performed at 1atmosphere pressure in helium gas. Laseretch craters were formed on the substratesurface by directing laser radiation onto thesurface at a near normal angle of incidencethrough the quartz window in the chamber.The laser used in these experiments was aSpectra Physics Model DCR-2A Q-switchedpulsed Nd:YAG laser emitting radiation at1 064 mm wavelength. The laser beam has adivergence of 0.5mradmrad and a full width athalf height pulse length of 10 nsec. The laserbeam intensity profile is gaussian in shapewith a diameter (measured at the points wherethe intensity is l/e2 of the maximum value) of5.9 mrn. A 2.5 mm diameter aperture wasplaced in the optical path of the laser beam atthe center of the beam to form a pseudo flat-top intensity profile. Ai. approximate relativeintensity profile for the DCR-2A output beamindicating the clipping points of the aperture isshown in Figure 3. All laser exposuresoccurred at a single spot on the substrate.The laser beam was not focused with a lens.
Prior to laser etching, substrate surfaces wereprepared by polishing to give a surfaceflatness of 200 nm and an arithmetic surfaceroughness less than 5 nm. The straw yellowoxide film was formed on polished 304Lstainless steel substrates by baking thesubstrate in air at 650RC for 6 minutes. Thisproduces a film which is composedpredominantly of chromium oxide and is less
than 5 nm in thickness. The iron oxide filmwas produced on the polished stainless steelsurfaces by baking the substrate in air at 800RC for 30 minutes. This produced an oxidefilm that was blue in color with a filmthickness of approximately 3 mm. Laser etchcraters were profiled with a Dektak 3030profilometer using a 2.5 mm radius diamondtip with a vertical applied force of 20 mg.The surface profiles have a vertical resolutionof 0.1 nm and a horizontal resolution of 5mm.
RESULTS AND DISCUSSION
Annealed 304L stainless steel substrates witha thin ( < 5 nm) overlayer of chromium oxidewere irradiated with multiple pulses ofNd:YAG laser radiation. Topographica!surface profiles along a line through the centerof the irradiated spot are shown in Figure 4for multiple pulse laser exposures of 1, 8, and20 laser pulses. The pulse energy for eachlaser pulse was 100 mJ, corresponding to atime averaged irradiance of 240 MW/cm2 atthe center of the laser beam and 170 MW/cm2at the edge of the beam. Note in the figurethat the horizontal scales are all identical, andare roughly a factor of 10,000 larger than thevertical scales. The vertical scale increasesfor each successive profile. The verticaldashed lines accompanying each profile markthe visible boundaries of the region where thechromium oxide film was totally etched awayby the incident laser radiation. Theseboundaries were determined by microscopicinspection of the light reflection properties atthe surface of the laser irradiated spot duringacquisition of the surface profile.
A single 100 mJ laser pulse was sufficient tocompletely remove the chromium oxide filmover the area directly irradiated by the incidentlaser beam. Single pulse experimentsperformed at lower pulse energies revealedthat a laser pulse energy of 50 mJ or greateris required to completely remove thechromium oxide film from the 304L stainlesssteel surface. This pulse energy corresponds
to a centra! beam irradiance of 120 MW/cm2.
The surface profile in Figure 4 correspondingto a single 100 mJ pulse of laser irradiationshows no significant etch crater. This isbecause the chromium oxide film is muchthinner than the apparent surface roughness ofthe 304L stainless steel substrates. As aresult, the etch crater corresponding toremoval of the chromium oxide film cannot bedetected in the profile. Note, however, thatthe apparent roughness of the irradiatedsurface has increased significantly whencompared to adjacent areas which did notreceive laser irradiation.
The surface profile corresponding to 8consecutive laser pulses in Figure 4 shows adefinite etch crater resulting from laser etchinginto the 304L stainless steel substrate. Notethat the profile of this etch crater exhibitsdefinite peaks and valleys. The apparentsurface roughness in the laser irradiated regionhas increased dramatically when compared tothat observed for a single pulse of laserirradiation. Many peaks are observed in theetch crater profile which extend above theplane of the original substrate surface. Theheight of the peaks generally increases towardthe center of the laser irradiated spot,suggesting that the extent of peak formationincreases with increasing surface irradiance.
The depth of the etch crater and the magnitudeof the surface roughness in the laser-irradiatedregion continually increase with increasinglaser pulse exposure as seen by theprogression in the etch profiles shown inFigure 4. A scanning electron micrograph ofthe etch crater resulting from a 10 pulseexposure to laser irradiation is shown inFigure 5a. The surface features responsiblefor the increased surface roughness appearingat the center of the laser irradiated region arereadily apparent in the figure. This regionexhibits a large number of exposed grainboundaries, each corresponding to theinterface between differently oriented singlecrystal grains. The centra) region of the laser-irradiated spot is shown at increased
magnification in Figure 5b. The linearboundaries bisecting crystal grains in themicrograph reveal the presence of crystaltwins. Preferential etching by the laserradiation along grain and twin boundaries isresponsible for the exposure of theseboundaries in the micrograph. The presenceof crystal grains extending above the originalplane of the highly polished substrate surfaceindicates that some vertical growth of thecrystal grains occurs during the laser etchingprocess. This growth could be the result ofthermally driven diffusion of metal atomsalong the substrate surface or the redepositionof laser vaporized material onto a more stablecrystal plane.
Chemical reagents are also known to etchpreferentially along grain and twin boundaries.An identical 304L stainless steel substrate waspolished to a 0.05 micron alumina grit andelectrolyticaliy etched in 10% oxalic acid for60 seconds at 5 volts DC. The resultingsurface microstructure is shown in the opticalmicrograph in Figure 6. The grain size isapproximately defined by ASTM GS 7.Comparison of Figure 6 with Figure 5b showsthat the chemically etched surface and thelaser-etched surface exhibit a similarmicrostructure. Laser etching does notsignificantly alter the microstructure in thenear surface region of 304L stainless steel.
A plot of laser etch depth in 304L stainlesssteel as a function of laser exposure for 100mJ laser pulses is shown in Figure 7. Etchdepth is defined in this case as the centralcrater depth averaged over the apparentsurface roughness. The data shown in Figure7 suggests a linear relationship between etchdepth and laser pulse exposure. The least-squares fit of a line to the data gives a laseretch rate of 0.0068 mm per laser pulse for304L stainless steel.
A plot of laser etch depth in an iron oxidefilm on a 304L stainless steel substrate versuslaser exposure for 100 mJ laser pulses isshown in Figure 8. For laser exposures lessthan 30 pulses, the data show a lineardependence between etch depth and laser
exposure. The least-squares fit of a line to thedata in this segment of the figure gives a laseretch rate of 0.10 mm per laser pulse. Notethat the etch rate in the iron oxide layer is 15times greater than that observed in 304Lstainless steel. As a result, one would predictthat as soon as the laser penetrated the ironoxide layer, etching would effectivelyterminate just past the metal-metal oxideinterface. This behavior is demonstrated inFigure 8. For laser exposures greater than 30pulses, the etch rate is effectively zero,indicating a 3 mm thick oxide layer. Thisself-limiting effect is highly advantageous inlaser cleaning operations because it provides asimple and ready method for endpoint control.If the opposite were true, i.e. laser etching inthe substrate proceeded much more rapidlythan laser etching of the contaminantoverlayer, it would be very difficult to removethe contaminant layer without seriouslyoveretching into the substrate. Mosicontamination layers on metal substrates willexhibit this self-limiting effect due to thecharacteristically high reflectivity of metals at1.064 mm wavelength.
SUMMARY AND CONCLUSIONS
Nd.YAG laser cleaning of metal oxides from304L stainless steel surfaces has beencharacterized. Thin chromium oxide films canbe completely removed from stainless steelsurfaces using a single 10 nsec laser pulsewith an average surface irradiance greater than120 MW/cm2. Laser etching proceedspreferentially along specific crystallographicplanes resulting in increased surface roughnessfor highly polished surfaces. No laser inducedchange in the near surface microstructure ofthe stainless steel is observed. Laser etchingof thicker iron oxide fiims from stainless steelsurfaces exhibits a self-limiting effect. Thelaser etch rate in the iron oxide is 15 timesgreater than in the 304L stainless steelsubstrate resulting in elective terminationof laser etching just past the metal-metal oxideinterface.
A CKMO WLEDGEMENTS
The authors gratefully acknowledge numerousdiscussions with C. V. Robino which proveduseful in the interpretation of themetallographic data presented in this paper.This work was performed at Sandia NationalLaboratories supported by the U.S.Department of Energy under contract numberDE-AC04-76DP00789.
CleaningChamber
Figur* 2.
VacuumPump
Experimental apparatus used to characterize laseretching of iron and chromium oxides on 3041.stainless steel surfaces.
figur* i. Schematic of particulate waste collection methodCor laser aBIacive cLeaning of surfaces.
03OC
o
• 1
1 1
• •
1 1
1 I
• 1
1 I'
l 1
1 1
1 |
1 1
IT
IT
I |||
1
||1
1 1
' '
' '
i I
1 I
t 1
1 i
I I
M i
I I
1 I
I i
I I
i a
i i
1 .
i i
i 4
i i
i •
- 6 - 4 - 2 0 2 4 6
Radial Position (mm)
Figure 3. Approximate laser beam relative intensityprofile. Dashed lines represent the laser bea»profile cut-off points for a 2.5 mm beamaperture.
0.1Z
0.06
0.00
-0.06
j
i1 1L Ik
1
1
! Pulse
Ih
:
r.UL ;
1'1 ' 11 liF :
W P ' :| S
1
0.2
§ 0.0
^ -0.2
r> 0.6
0.2
-0.2
-0.6
8Pul»e«1ll
11«•!
IJ
1 I I ! 'rlLl. .'lu
MjllIJIBlUUH
i l
f'l
20Pul8ea
1
1- ,i
!ills :
LuHTM
1 i0 1 2 3 4 5 6 7
Horizontal Dimension (mm)
Figure 4. Laser etch crater topographical profilesresulting from 1, 8, and 20 consecutive 100 nJlaser pulses on a chromium oxide covered 304Lstainless steel substrate.
(a)
25KU X30. 18.10.1000.-8U SNLi
Figure :. (a) Scanning electron micrograph of the etchcrater resulting from 10 consecutive laser pulseson a chromium oxide covered 304L stainless steelsurface. (b) View of the central region of theetch crater at high magnification.
Figure 6. Bulk microstructure of electrochemically etched3 04L stainless steel
BOO
LJ
S 10 15 20
Number of Laser Pulses25
Figure ?. Etch depth versus pulse number for laser etchingof chroaluB oxide covered 304L stainless steelsubstrates using 100 BJ laser pulses. The slopeof the line indicates an etch rate of 6.8 na perlaser pulse in the stainless steel substrate.
35000
20 40 60
Number of Lacer Pulses80
Figure 8• Etch depth versus pulse number for laser etchingof 3000 na thick iron oxide filas on 3O4Lstainless steel substrates using 100 BJ laserpulses. The slope of the line indicates an etchrate of loo ru> per laser pulse in the iron oxidefilm.
CO2 PELLET BLASTING FOR PAINT STRIPPING/COATINGS REMOVAL
Wayne N. SchmitzMcDonnell Aircraft Company (MCAIR)
St. Louis, Missouri
MCAIR's MOTIVATION
A major element of every MCAIR program isIntegrated Logistic Support (ILS) of thefielded weapon system. As more aircraft aremanufactured with composite materials toreduce weight while maintaining high strengthstructures, the requirement to effectively(completely) remove paint, primer and rainerosion coatings to facilitate bonded repairsbecomes critical.
While composite structures are not subject tocorrosion or fatigue cracking, the remainingmetal portions of the airframe must beinspected periodically to preclude catastrophicfailures from metal fatigue. Again, surfacecoatings must be completely removed tofacilitate inspection.
Current paint stripping activities employing theuse of phenol-based methylene chloridechemicals are not acceptable because:
•Composite materials are susceptible todamage, and
•The use of hazardous materials resultingin maintenance personnel injury and toxicwaste generation must be curtailed.
MCAIR, subsequently, began a search for newpaint stripping technologies to satisfy ILSrequirements.
THE SEARCH
Our search for a new process to strippaint/primer and any variety of surfacecoatings was conducted under the followingconstraints:
•The process must not compromise thestructural integrity of the aircraft-thisrequirement is far more stringent thanmerely not damaging the surface ofthe substrate as it includessubstructure.
•Toxic v/aste and the use of hazardousmaterials must be eliminated.
•Disposable materials, i.e., removedpaint, etc. plus any worn out,contaminated media must be reducedby 90%.
•The process must reduce maintenancemanhours, overall stripping cost, andaircraft cycle time by 50%.
The technologies investigated fell into threecategories:
•Dry Media Blasting-CO2 pellets-Plastic grit-Wheat starch-Walnut shells and the like
•Liquid Media Blasting-Medium pressure (7,000 Psi)Water jet
-High pressure (32,000 Psi)Water jet
-Water ice slurry-Sodium bicarbonate slurry
•Pulse Light Energy-Lasers-Xenon flashlamps
Each of the technologies exhibited one ormore of the following problems:
11
•Potential to damage substrates and/orsubstructure (as a secondary effect)
A comparative study of the technologies listedabove was conducted to determine which ofthe processes could effectively strip militaryspecification paint and primer within thespecified constraints. Two additionalcategories were added to those identifiedproblems in order to arrive at the "bottomline" effectiveness of each process: 1) Lifecycle cost benefits for a total weapon system,and 2) Aircraft thru-put rate.
MCAIR chose CO, pellet blasting as thetechnology offering the greatest benefits bothin terms of maintenance cost reduction andenvironmental issues compliance. Because theCO2 pellets are made from liquified CO2 gas,a natural atmospheric element, and sublimateinstantly on contact back to that gas, theyrepresent an operator/environmentally safeprocess. Since there is no media of which todispose, only removed paint chips remain andtheir volume, compared to toxic wastegenerated by chemical stripping processes,represents a 96 percent reduction. CO2 pelletblasting is generally benign to most substrates
though there are some concerns which will bediscussed later. Since there is no solid media,intrusion is not a factor, and the vastpercentage of aircraft masking is eliminated asis post-stripping clean-up, media disposal cost,and the requirement for a mediaseparation/recycling system. Elimination ofthese tasks reduces maintenance manhours byat least 50 percent. Another benefit of CO2
pellet blasting is its ability to remove a broadrange of aircraft surface coatings, sealants andadhesives. Best of all, there is no need to pre-clean the aircraft; the process instantlyremoves grease, oil, etc. while stripping paint.The economic bottom line to these benefits isan overall stripping cost of $5/ft2 compared to$19 plus/ft2 for current chemical processes.
CO2 PELLET BLASTING CONCERNS
Every paint stripping technology weinvestigated exhibited negative characteristicsto a greater or lesser extent and none of theprocesses has been thoroughly tested withrespect to all potential effects on aircraftstructures. For all of its excellent benefits,CO2 pellet blasting still requires further testingon fatigue life degradation, crack growthpotential and the possibility of inducing micro-cracking in composite materials.
Some of the effects of CO2 pellet blasting arevisual. At the blast pressures required toeffectively remove paint/primer, softaluminum skins less than 0.032 inch thickshow evidence of peening. Thermosetcomposite materials are easily damaged unlessvery close attention is paid to dwell time andstand-off distance. One other aspect of CO2
pellet blasting is a relatively slow strippingrate (0.5 ftVmin) on alclad-coated aluminumskins and thermoset composites. Clearly,further optimization of the process is requiredbefore CO2 pellet blasting is used to removepaint/primer from the wide range of aircraftsubstrates.
12
FUTURE PLANS
Because CO, pellet blasting offers outstandingenvironmental gains, MCAIR will continueresearch to enhance its performance.Preliminary results of combining CO2 pelletblasting with other paint stripping technologieslook promising, proving once again that thereare no simple solutions and no one process isa panacea for all problems.
THE EVALUATION OF ALTERNATIVES TO OZONE-DEPLETINGCHLOROFLUOROCARBONS
Robin L. SellersNaval Avionics CenterIndianapolis, Indiana
INTRODUCTION
In the spring of 1988, the electronics industrywas facing a serious set of problems. TheMontreal Protocol, an international agreementsigned on 16 September 1987, required a 50%reduction in production of chlorofluorocarbcn1,1 ,2- t r ich loro- l , 2 ,2 - t r i f l uo ro -e thane(CFC-113) by 1998.' In 1986, the electronicsindustry used an estimated 80 millionkilograms of CFC-113 to remove flux fromprinted circuit board assemblies, representing45 percent of worldwide CFC-113consumption.2 As the largest worldwide userof CFC-113, the electronics industry wasconfronted with a significant challenge.
In addition, the electronics industry was facingthe challenge of implementing surface mounttechnology (SMT). With SMT's smallerstand-off heights and higher density designs,the traditional cleaning processes and tests formeasuring cleanliness were being questioned.
Alternative cleaning materials and processeswere being used by some commercialelectronic manufacturers for bothplated-through-hole technology and SMT;however, the U.S. military specificationslimited the choices available to the militaryelectronics manufacturer. Further analysis ofthe situation showed that U.S. militaryspecifications also influenced a substantialportion of the commercial electronics industry.Due to their association with highly reliableproducts, military specifications had becomede facto world standards, driving an estimated50 percent of the world's use of CFC-113.2
Due to the unique cooperation and hard workof representatives from the EnvironmentalProtection Agency, the Department ofDefense, and industry, this set of problemswas transformed into an opportunity toimprove electronics manufacturing processesand military specifications.
AD HOC SOLVENTS WORKING GROUPAND THE CLEANING AND
CLEANLINESS TEST PROGRAMS
In the spring of 1988, the EnvironmentalProtection Agency (EPA) was preparing toissue the regulations that would implement theMontreal Protocol provisions within theUnited States. The pronouncement ofproposed reductions in the supply of CFC-113brought strong comments from the suppliers ofCFC-113 products, electronic manufacturers,and the military's technical community. Facedwith changing cleaning materials andprocesses, the U.S. military and electronicmanufacturers wanted assurance that the newmaterials/processes would perform as well asthe CFC-113 they had used for years.However, there was no widely acceptedmeasure of "how clean is clean" or "howclean is clean enough."
To answer these questions and expedite thetransition from CFC-113 to other materialsand processes, Dr. Stephen Andersen, EPA,formed the ad hoc Solvents Working Group.The ad hoc Solvents Working Group iscomprised of representatives from materialssuppliers, equipment manufacturers,commercial electronic manufacturers, militaryelectronic manufacturers, Department ofDefense, the Institute for Interconnecting andPackaging Electronic Circuits (IPC), the EPA,
15
and other interested organizations (see TableI). Utilizing this broad participation, the adhoc group has worked (and continues to work)at an extraordinary pace to provide a uniformand timely evaluation of materials/processeswhich reduce the level of CFC-113 used inelectronic cleaning processes.
One of the first and most critical tasks for thead hoc group was to define the scope ofeffort. When choosing an alternate materialand/or process for cleaning of electronicassemblies, there are a number of technicaland economic issues that an electronicsmanufacturer should consider. These includeenvironmental, worker safety and health,performance, and cost issues. As Figure Ishows, the list of issues that require definitionfor each alternative is extensive. In addition,the cleaning process, although critical, is justone of a number of interconnected processesthat affect the reliability of an electronicassembly.
In order to be effective in the shortest timepossible, the group decided to tackle the issuesthat could not be addressed by other interestedparties, and which required the broader reviewthat the ad hoc group could provide. Theseitems are denoted with the crosshatches inFigure I. The ad hoc group chose to focus onissues related directly to the ability of aprocess and its associated materials to producean assembly that was "as clean as" or "cleanerthan" an assembly produced using CFC-113.In addition, the group chose to tie thecleanliness levels to the effect on electricalperformance. In this way the test programaddressed the questions "how clean is clean"and "how clean is clean enough."
The ad hoc group was assured by the EPAand the suppliers of the alternate materials thatthey would answer questions concerningenvironmental issues, such as the ozonedepletion potential (ODP), the global warmingpotential (GWP), energy efficiency, regulationas a volatile organic compound (VOC), andwaste handling and disposal. In addition, theEPA assured the group that they would work
with other government agencies to expedite theevaluation of alternatives with regard to theireffect on worker safety and health. Thisinformation is supplied by materials vendorson the Materials Safety Data Sheets that arerequired by law.
The electronics manufacturer, with help frommaterial and equipment suppliers, was left thetask of deciding which materials and processesmake the most sense for their designs,processes, materials, equipment, corporateculture, specifications, facility, and budget.The ad hoc group did not attempt to answereach and every manufacturer's question of"what should I switch to?" The answer to thisquestion is as varied as the differencesbetween manufacturers, and even between anindividual company's plants and productionlines. Instead, the ad hoc group developed thestandard test that can be used by electronicsmanufacturers and by the military for decidingwhich alternatives are capable of cleaning "aswell as" or "better than" CFC-113.
Once the focus had been established, the adhoc group proceeded to develop the objectivetest protocols that are tied to actualmanufacturing conditions and answer thequestions "how clean is clean?" and "howclean is clean enough?" Beginning in thespring of 1988, the ad hoc Group developed athree phase test program. The goal of PhaseI/Benchmark is to produce the cleanlinessreference data for a typical CFC-113 process.The objective of Phase II/Alternate CleaningMaterials and Processes is to evaluate a widevariety of cleaning materials and processes,compare them to the cleanliness reference dataobtained in Phase I, and to makerecommendations with regard to theirsuitability for military use. The goal of PhaseIll/Alternate Manufacturing Media andProcesses is to evaluate alternate solderingmaterials and processes in combination withalternate cleaning materials and processes.Replacement of soldering materials/processescould eliminate the need to clean withCFC-113, and might even eliminate the needto clean at all. Current Phase III efforts
16
include the Water Soluble Flux Evaluation,No-Clean Fluxes, and Controlled AtmosphereSoldering. The ultimate goal for each of thesephases is to reduce the level ofozone-depleting chemicals used in themanufacture of electronic assemblies.
The link between all three phases of the testprogram is the standard printed wiring board(the 1PC-B-36) developed by the ad hocgroup. The IPC-B-36 incorporates boththrough-hole and surface mount features withten electrical circuits that can be used forevaluating cleanliness using surface insulationresistance (SIR). The assembly design alsoprovides a reasonably tough cleaning challengewith its 0.050" pitch leadless chip carriersmounted on 0.005" stand-offs and 0.020" viasthat permit liquid flux to flow up the via holesand under the carriers.
In addition, the ad hoc group developed asecond standard printed wiring board (theIPC-B-24), to be used in the evaluation ofalternate soldering processes and materials inPhase III. A relatively simple design, theIPC-B-24 has four SIR circuits that can beused for evaluating the interaction betweensoldering fluxes or pastes, metals, and thecleaning materials.
The Cleaning and Cleanliness Test Programfor Phases I and II3, and the Phase III WaterSoluble Fluxes Test Program4, define thematerials and processes to be used in eachphase. (Note: Test plans for the controlledatmosphere and no-clean flux portions ofPhase III are in the process of development.)Not only does the IPC-B-36 link the threephases, but the materials and processes arealso kept as uniform as possible. TheIPC-B-36 and IPC-B-24 test sequencessimulate surface mount assembly processesutilizing solder pastes, infrared or vapor phasereflow, liquid fluxes, and wavesoldering. Thetest results include data for boards whicl; iiaveNot Been Exposed to the typical electronicassembly residues and which represent thecleanest condition. The test results alsoinclude data for assemblies which have been
exposed to the typical electronic assemblyresidues and have Not Been Cleaned. Theseunclean assemblies represent the dirtiestconditions and indicate what would happen ifan electronic assembly were not properlycleaned. The final two sets of test resultsinclude the data for assemblies which havebeen exposed to typical electronic assemblyresidues and which have Been Cleaned.These cleaned assemblies represent the relativeperformance of a cleaning material/process,and indicate what effect the cleaningmaterial/process and any residues would haveon the electrical performance of an electronicassembly.
Standard cleanliness tests were developed anddefined for each of the three phases. ForPhases I and II, four standard cleanliness testswere defined. Ionic cleanliness levels aremeasured using the ionic conductivity test andthe commercially-available Omegameter600SMD. Non-ionic cleanliness levels aremeasured using two methods: the residualrosin test, which utilizes an ultraviolet/visiblespectrophotometer, and high performanceliquid chromatography (HPLC). The final testis surface insulation resistance (SIR), which isa measurement of electrical performance afterprolonged exposure to elevated temperatureand humidity. Phase III utilizes three of thefour standard tests used in Phases I and II:ionic conductivity, HPLC, and SIR. Inaddition, Phase III requires tests using ionchromatography. Using both ionic andnon-ionic tests, the ad hoc group ensured afull picture or "how clean is clean." Inclusionof SIR provides a good indication of "howclean is clean enough."
Although not included as pass/fail criteria,visual examination is also required. Each ofthe test plans requires photographs ofrepresentative assemblies and of all anomalies.
The only differences between Phase I andPhase II are the cleaning material and thecleaning process used. Differences betweenPhase II data and the Phase I/Benchmark dataare attributable to the alternate cleaning
17
material/process. The Phase II data areevaluated to determine if it is "worse than,""as good as," or "better than" the PhaseI/Benchmark data.
Phase III processes are significantly differentfrom both Phases I and II; however, thecommon test methods and the use of theIPC-B-36 provide a link to the Benchmark.
TEST MONITORING ANDVALIDATIONCOMMITTEE
Citing the need to expedite the evaluation ofalternatives, the ad hoc group decided thattesting should not be confined to a single site.In order to ensure that the test plans werestrictly followed and that tests werecomparable between sites, the ad hoc groupestablished the Test Monitoring and ValidationCommittee (TMVC). The TMVC is a subsetof the ad hoc group. Members of the TMVC(called the Test Monitoring and ValidationTeam) attend all official tests, monitor allofficial testing, review all reports, anddetermine if an alternate material/process is"worse than", "as good as," or "better than"the Benchmark. The sponsor of the alternatematerial/process is required to publish all testresu'ts, even in the case of "worse thanresults", and provide copies of the report uponrequest. The Phase II TMVC is chaired byDr. Leslie Guth, AT&T. The Phase IIITMVC is chaired by Dr. Laura Turbini,Georgia Tech.
PROGRESS/PHASE I
The Electronics Manufacturing ProductivityFacility (EMPF) and Naval Avionics Center(NAC) completed the Phase I tests in Januarya n d F e b r u a r y 1 9 8 9 u s i n g anitromethane-stabilized azeotrope of CFC-113and methanol. The tests were made possibleby significant contributions of labor, materials,and equipment by members of the ad hocSolvents Working Group. EMPF and NAC
estimate that they alone spent two and one halfman years of effort preparing for and actuallyperforming the tests.
Results of the Phase I/Benchmark tests werefully reviewed by the ad hoc group and arereported in IPC-TR-580, "Cleaning andCleanliness Test Program Phase I TestResults,"5 which is available from the IPC.The primary result of Phase I/Benchmark is abenchmark data set for comparison of alternatematerials and processes to the performance ofa typical CFC-based cleaning process. Inaddition, many of the processes used in PhasesII and III were defined during Phase I.
PROGRESS/PHASE II
David Bergman, IPC, has coordinated thePhase II tests. As of 25 January 1991, sevenmaterials and their associated processes havebeen tested and received Test Monitoring andValidation Team approval as part of Phase II.The results of these tests are summarized inTable II. Three of the alternate materials arehydrogenated chlorofluorocarbons (HCFCs);three of the alternate materials aresemi-aqueous; and the remaining material is aCFC-113 based solution which uses lessCFC-113 than is used in the Benchmarksolvent. Copies of the test reports areavailable from the sponsors of the tests (seeTable III). In addition, for those interested inperforming Phase II evaluations, DavidBergman's address and phone number arelisted. Additional Phase II tests areanticipated.
Results of the Phase II tests have beenforwarded with recommendations for inclusioninto military specifications to an extensive listof technical contacts within Department ofDefense and industry, including members ofthe Military Electronics Technical AdvisoryGroup (METAG) CFC Subcommittee. Inaddition, members of the ad hoc group havebeen involved in numerous briefings at avariety of levels within the Department ofDefense (DOD), including the DOD CFC
18
Advisory Committee which was established byCongress to provide feasibility and costestimates of CFC chemical substitutes andalternative technologies, and assist intechnology transfer.* The METAG CFCSubcommittee, a group established by theDeputy Assistant Secretary of Defense (TotalQuality Management) and the DeputyAssistant Secretary of Defense (Environment),and chaired by Harold Rife, Crane NavalWeapons Support Center, has recommendedthe following:
".. .that Phase I Benchmark become theacceptance criteria for all electronicassemblies"7
"...any environmentally safe andcompatible process proven capable of cleaningmaterials to the benchmark be allowed formilitary electronic assemblies, even if acontract modification is required."7
"...that U.S. military standardsofficially recognize and cite the benchmarktest procedure ... and the test results ... as thefull confirmation of the capability of cleaningas well or better than CFC-113"7
PROGRESS/PHASE HI
The Phase III Water Soluble Flux Test Planwas issued in August 1990 after months ofdevelopment. Naval Avionics Center waschosen as the primary test site for the WaterSoluble Flux (WSF) Evaluation, which wasdesigned to demonstrate the performance ofwater cleaning in combination with watersoluble fluxes and pastes. The WSFEvaluation consists of two parts: evaluation ofthe flux/substrate interaction using theIPC-B-24 assembly and the evaluation ofcleanliness using the IPC-B-36 assembly (sameassembly used in Phase I/Benchmark andPhase II testing). Process development hasbeen completed at Naval Avionics Center.The evaluation of three fluxes and three pastesusing the IPC-B-24 test is scheduled for 1-3February 1991. The "best" paste and the"best" flux from the IPC-B-24 test willundergo the IPC-B-36 test at NAC, tentatively
scheduled for April 1991.
The Phase III No-Clean Flux Test Plan isbeing developed by a subset of the ad hocgroup, primarily through the efforts of theIPC. Compared to the traditional rosin-basedfluxes and CFC-113 cleaning, no-clean fluxesand pastes (which require no cleaning at all)promise the greatest environmentalimprovements. However, they also raise themost questions with regard to performance andlong-term reliability. Dr. Laura Turbini,Georgia Tech, is leading this formidableeffort.
The Phase HI Controlled Atmosphere TestPlan is also being developed by a subset of thead hoc group and through the efforts of theIPC. Using a variety of inert or reactiveatmospheres, new soldering equipment andprocesses, which do not require traditionalfluxes, have been developed. LarryLichtenberg, Motorola, is leading this effort.
CONCLUSIONS
The EPA/DOD/lPC/Industry ad hoc SolventsWorking Group has worked effectively andefficiently to identify viable alternatives forozone-depleting CFC-113 for use in cleaningelectronic assemblies. Thus far, the ad hocgroup has:
(1) developed a set of standard tests forevaluating alternate materials and processes.
(2) established a group and method formonitoring and recognizing all official tests ofalternatives.
(3) produced a benchmark data set forcomparison of alternate materials andprocesses to the performance of existingCFC-based cleaning materials.
(4) monitored and validated the testing ofseven alternate cleaning materials andprocesses.
19
(5) worked with Department of Defenserepresentatives to modify militaryspecifications and existing contracts.
More than ever, the industry /governmentcooperation demonstrated by the ad hocSolvents Working Group needs to continue.The Clean Air Act of 1990 includes legislationconcerning CFCs, methyl chloroform, andvolatile organic compounds (VOCs). Thislegislation challenges ihe U.S. electronicsindustry to reduce the use and release ofozone-depleting and global warmingchemicals. In addition, the stringency, scope,and timetable for the Montreal Protocol wererecently reviewed by United NationsEnvironment Programme Committees and theprotocol was modified on 29 June 1990 inLondon, England (see Table IV).
The Protocol now calls for a 100% phase-outof CFC-113 by the year 2000. Methylch loroform (o therwise known as1,1,1-trichloroethane) was also added to thelist of regulated substances. 100% phase-outof methyl chloroform, which was at one timeconsidered a viable alternative for CFC-113 inelectronics cleaning, is required by 2005. Aresolution calling for the use of hydrogenatedchlorofluorocarbons (HCFCs) only whereother alternatives are not feasible was alsoincluded, with a notice that 100% phase-out ofHCFCs should be expected no later than theyear 2040 (Note that the Clean Air Act callsfor phase-out of HCFCs by 2030).
At the recent International Conference on CFC& Halon Alternatives held 27-29 November1990 in Baltimore, Maryland, EileenClaussen, Director of the Office ofAtmospheric and Indoor Air Programs at theU.S. EPA, announced that dependent onfurther scientific and technical evaluations, thephase-out of CFCs may be moved to 1997.Obviously, the challenge to the electronicsindustry remains just as imperative today asit did in the spring of 1988.
ACKNOWLEDGEMENTS
The author would like to recognize themembers of the ad hoc Solvents WorkingGroup who donated the necessary labor,materials, equipment, and technical support toexpedite the transition to non-ozone-depletingalternatives within the electronics industry.
REFERENCES
1. United Nations EnvironmentalProgram (UNEP), "Montreal Protocolon Substances that Deplete the OzoneLayer", 1987.
2. UNEP Solvents Technical OptionsCommittee, "Electronics Cleaning,Degreasing, and Dry CleaningSolvents Technical Options Report",30 June 1989, p. 13.
3. Ad hoc Solvents Working Group,"Cleaning and Cleanliness TestingP r o g r a m , a J o i n tIndustry/Military/EPA Program toE v a l u a t e A l t e r n a t i v e s toChlorofluorocarbons (CFCs) forPrinted Board Assembly Cleaning",published by the IPC, 1 September1990 revision.
4. Ad hoc Solvents Working Group,"Phase 3/Water Soluble Fluxes,Cleaning and Cleanliness TestingP r o g r a m , a J o i n tIndustry/Military/EPA Program toE v a l u a t e A l t e r n a t i v e s toChlorofluorocarbons (CFCs) forPrinted Board Assembly Cleaning",published by the IPC, 1 September1990.
5. Ad hoc Solvents Working Group,"Cleaning and Cleanliness TestProgram Phase 1 Test Results",IPC-TR-580, published by the IPC,October 1989.
20
6. "Charter for the Department ofDefense Chlorofluorocarbons (CFC)Advisory Committee" established bySection 356 of the National DefenseAuthorization Act for Fiscal Years1990 and 1991, Public Law 101-189,29 November 1989.
7. Harold Rife, "Department of DefenseChlorofluorocarbon SubcommitteeMeeting of 14 June 1990",Memorandum 5050/648, 603A, dated9 July 1990.
21
TABLE I
PARTICIPANTS / AD HOC SOLVENTSWORK GROUP
Alternative Producers
Allied SignalDuPontICI ChemicalsPetrofermDuBois ChemicalsBy Pas of ToledoAdvanced Chemical Tech.Advanced Chemical Tech.Alpha Metal*Cham-Tech internationalDow ChemicalGAF ChemicalKester SolderLondon ChemicalMartin Marietta LabsMirechemPemwalt Corp.Van Waters I Rogers
Institute for Intercon-necting and PackagingElectronic Circuits (1PC)Halogtr.ated Solvent Indus-try Association (HSIA)
Government Agencies/Other
EPA U.S. Air ForceU.S. Navy (EMPF) U.S. ArmyU.S. Navy (NAC) D.O.D.Sandia National Laboratories Georgia Tech.Underwriters Laboratory NASARobisan Laboratory DESCInternational Conservation Center Foundation (ICF)Naval Weapons Support Center (Crane)
22
TABLE II
TESTED PHASE II ALTERNATIVES
MATERIALAllied SignalGenesolv*2010
Martin MariettaMarclean™-R
PetrofermBioact" EC-7
OuPontAxarel" 38
AlliedGenesolv" 2004
DuPontFreon" SMT
DuPontKCO 9434
TYPE
HCFC *
S/A •*
S/A
S/A
HCFC
CFC
HCFC
MATERIALDATE
Sept. 89
Jan. 90
Feb. 90
Feb. 90June 90
Apr. 90
June 90
July 90
TESTRESULTS
Passed
Passed
Passed
Passed
Passed
Passed
Passed
Results reported are for specific material, equipment, and operating parameters.
* HCFC = hydrochlorofluorocarbon
** S/A = semi-aqueous material
TABLE III
CONTACTS FOR PHASE II
Institute for Interconnecting andPackaging Electronic Circuits (IPC)
David Bergman7380 North Lincoln AvenueLincolnwood, IL 60646708-677-2850
Allied-Signal, Genesolv/Baron-BlakesleeDr. Kirk Bonner2001 N. Janice AvenueMelrose Park, IL 60160708-450-3880
Martin MariettaDr. Maher TadrosSystems, Baltimore Division103 Chesapeake Park PlazaBaltimore, MO 21220301-247-0700
PetrofermDr. Mike Hayes5400 First Coast HighwayFernandia Beach, FL 32034904-261-8286
E.I. duPont de Nemours & Co., Inc.Carroll SmileyChestnut Run PlazaP.O. Box 80711Wilmington, DE 19880-0711302-999-2629
23
TABLE IV
SUMMARY OF LONDON AMMENDMENTS TOMONTREAL PROTOCOL
CHLORO-FLUOROCAR-BONSStudy isy1?92 to seeif esriierp*iase-outis possible
HALONS'Exemptionfor essentialuses
OTHER FULLYHAIOGE-MATED CFCs
CARBOMTETRA-CHLORIDE
WTHYLCHIORO-FORM
HCFCsTransit ionalSubstancesfor use onlyyhere otheralternativesare notfeasible
1993
20X
20X
freeze
1995
50X
50X
85%
30X
(Percent of
1997 2000
85% 100X
100X'
85X 100X
100X
70X
Reduction)
2005
100X
2020
100XTargeforPhaseout
2040
100XRequiredPhase-out
materials co
toxtctty•xposur«fttnunabtJrty
stabtlrTy
FIGURE 1
CRITERIAA CFC
mpatatxlrty
VOCi
disposal
FOR CHOOSINGALTERNATIVE
Performance.deaning eff ioency
> effect on electrical
\ yC°^«ubii«,\ ^f ^^K co«t per %Q. foot board
^^T need for drymg^ " ^ ^ energy efficiency
^ maintenance
dnposaie<}uipment cost
rhc lifettm*
•nergy •ftkicncy
24
SPRAY FORMING AS A NEW PROCESSING TECHNIQUE
Scott A. Ploger and Lloyd D. WatsonCustom Spray Technologies, Inc.
Rigby, Idaho
INTRODUCTION
A versatile process has been discovered forspraying solid materials in various forms fromliquid feedstock. As displayed in Figure 1,the basic principle involves aspirating liquidinto an inert gas nozzle, where incomingstreams are efficiently nebulized into adirected mist of fine droplets. The liquid istypically in a molten state, such that dropletscool rapidly in flight before collecting againsta substrate where solidification is completed.
Although seemingly similar to other "gas-atomization" techniques, the ControlledAspiration Process (CAP) offers uniqueadvantages for spray forming numerousmaterials. As discussed below, the CAPapproach emphasizes tailoring plumecharacteristics for individual applications.Critical aspects involve designing spraysystems for specific purposes, usingsophisticated control over componenttemperatures, and employing comprehensivemonitoring of performance parameters.Consequently, free-standing objects andadherent coatings have been sprayedsuccessfully. Whereas most activities thus farhave been aimed at rapidly solidified metals,this technology has also been extended topolymer-based materials.
Spray forming normally offers economic andenvironmental benefits to complement superiorproduct properties. By comparison totraditional processing techniques, spray-forming technology is highly efficient in useof energy and conversion of feedstock. It issafe for personnel, and it produces virtually nohazardous solid, liquid, and airborne wastes.Spray forming can be marketed successfully toindustries interested in the broad realms ofwaste minimization and environmental
compliance, where the attractive "bottom line"considerations on product properties andprocess economics can readily offset capitalcosts of retrofitting. In particular, sprayforming could conceivably eliminate solventemissions in several commercial situations ofimmediate concern.
BACKGROUND
The CAP version of spray forming wasoriginally developed at the Idaho NationalEngineering Laboratory (INEL). As outlinedin Reference 1, EG&G Idaho, Inc., devotedinternal seed funds toward elaborating uponpioneering research at the University Collegeof Swansea and the Massachusetts Institute ofTechnology (MIT). These two institutionsessentially adapted the standard nozzle used tofabricate metal powder2 into devices forspraying consolidated metal deposits. In bothcases the driving forces were reducing metalprocessing costs and improving mechanicalproperties.
Nozzles designed for powder metallurgy wereprimarily aimed at convenience. As such,they consist of an open crucible with aplugged hole in the base. When a stopper rodis pulled, molten metal drains down the hole,whereupon the metal stream is nebulized by aring of inert gas jets. Metal on the outside ofthe stream is nebulized more efficiently than atthe core, yielding droplets with spatiallyvarying size distributions. Consequently, mostof the mass remains near the plume centerline,producing a Gaussian-shaped deposit when thedroplets impact onto a substrate beforecompletely solidifying. Microsfuctures andmechanical properties change over the depositbecause of differing thermal fluxes andcooling rates.
25
The rate at which molten metal is nebulized inthe standard design is largely determined bythe liquid level in the crucible. Instead, theINEL chose to investigate an approach wherethe liquid feed rate is governed by aspiration.Here, over a certain range of gas pressure,liquid metal is drawn by suction into the throatof a converging/diverging nozzle, much asgasoline enters a venturi carburetor. Thenebulization process is very efficient in thephysically confined ineraction zone, resultingin unusually fine droplets. Furthermore,operating pressures are significantly lowerthan in conventional metal-spraying nozzles,producing a gentle, low-velocity plume thatresists entrapping gas bubbles in deposits.The U.S. Department of Energy obtained apatent that recognized the unique features ofthe INEL version of spray forming3.
EG&G Idaho employees also developed anozzle prototype for spraying wide, flatdeposits, as detailed in1. This rectangularnebulizer contained a slot-shaped gas throatfed by multiple liquid orifices from a heatedtrough (tundish). This design was critical toattracting a DOE program under the SteelIndustry/Federal Laboratories ResearchInitiative, in conjunction with MIT, Oak RidgeNational Laboratory, and eight industrialparticipants. The main objective is reducingthe energy consumed in manufacturing steelstrip by eliminating most rolling steps.
Another environmentally beneficial projectwas initiated at EG&G Idaho at the same time,where funding from the Engineering andServices Center at Tyndall Air Force Base wasdirected at spraying high-performancecoatings. The ultimate intention is eliminatinghazardous wastes from the chromiumelectroplating process, which demandscoatings of high-melting-point metals withestablished resistance to wear and corrosion.First, however, the feasibility had to beconfirmed at low temperature with tin4. Inthis low-budget study, a bench-scale spraysystem readily produced dense adherentcoatings, and rapid solidification alsostrengthened the coating layer. Over 99
percent of the metal sprayed was consolidatedinto each deposit and virtually no wastes weregenerated. The low heat flux delivered to thebase material shows that CAP systems areideal for depositing metal coatings on plastics,cellulose fibers, and heat-sensitive alloys.These novel coating attributes yielded apending DOE patent5.
ACTIVITIES OF CUSTOM SPRAYTECHNOLOGIES, INC.
The two principal investigators technicallyresponsible for spray-forming growth at theINEL (Lloyd Watson and Scott Ploger) electedto form their own research company underDOE's "technology transfer" auspices.Custom Spray Technologies (CST) is thus atwo-person private venture created to explorepossibilities for rapidly commercializingcertain spray-forming applications outside thedomain of direct INEL interests. To this end,CST constructed its own 4000 square-footlaboratory in 1990 to offer timely, economicaldesign and testing investigations. Consultingsupport will also be available during scale-upand pilot-plant stages to properly escortparticularly promising applications into fullproduction.
As noted, CSTs primary objective isconducting industrially oriented research anddevelopment. Nevertheless, CSTs first twoyears will be occupied honoring previouscommitments to federal agencies. SuccessfulPhase I efforts on the USAF coatings projectnaturally led to approved funding for a high-temperature Phase II demonstration withhardfacing alloys, where CST provided full-time consulting support to EG&G Idaho.Phase II was also successful, yieldingmicrocrystalline cobalt-chromium coatingswith hardness values among the highest evermeasured on metallic materials-in the realmof cemented carbide cutting tools6. More CSTsupport was thus requested by Mountain StatesEnergy, Inc., to assist on Phase HI pilot-plantexperiments.
26
Because CSTs two-person staff mightotherwise limit the commercial implementationof spray-forming technology, CST has enteredinto a teaming agreement with SCIENTECH,Inc.. of Idaho Falls. This establishedengineering firm will be responsible forscaling up prototype sytems developed byCST, as well as providing procurementservices, contract management, human factorsinput, approved drawings, system operatingp r o c e d u r e s , and other pedigreeddocumentation required for productionpurposes. The first example of this jointarrangement is a net-shape-forming project forMartin Marietta Energy Systems, Inc., whichis intended to minimize generation ofcontaminated wastes at Oak Ridge NationalLaboratory.
DEVELOPMENT METHODOLOGY
Successfully confirming the feasibility of sprayforming on diverse applications depends uponseveral inherent features of a CAP nebulizingsystem. Of particular importance is the abilityto control temperatures of all majorcomponents. Figure 1 reveals that heating isemployed at the melt furnace, tundishreservoir, nozzle body, and incoming inert gasplenum. Auxiliary heating and/or cooling isalso frequently implemented at the substrate toinfluence deposit wetting and adhesion, alongwith solidification rates. For any fixednebulizer geometry, these independent thermalcontrols offer considerable flexibility foroptimizing both plume characteristics andproperties in the consolidated droplets.
Performing meaningful experiments furtherrequires carefully monitoring and recordingcritical operating variables. Figure 1illustrates the typical positions ofthermocouples for temperature measurements,plus the pressure and flow transducersnecessary to control the nebulizing gas.Measurement signals are displayeddynamically on computer screens for on-linemonitoring through the use of versatile data-acquisition software. Menu-driven screens are
configured for all operating modes desired,including detailed characterizations ofindividual components prior to assembly of theintegrated system. Thoroughly understandingsubsystem performance has proven essential tolater data interpretation from sprayingexperiments. Dynamic computer displays ofkey system parameters aid manual adjustmentsof regulated gas pressure, along with visualobservations of droplet spraying anddeposition behavior.
Figure 1 also shows that spraying is normallyconducted in a sealed, inert environment toeliminate interactions with reactive gases.Additional benefits are protecting personnelfrom heated components, electrical hazards,and any ingestible unconsolidated particulates.As indicated, particle concentrations aremeasured on both sides of the filter todetermine filter efficiency and to absolutelyguarantee against significant atmosphericdischarges . Comparing chamberconcentrations to spraying rates further assistsin calculating melt-to-deposit consolidationefficiencies, as well as estimating needs forrespiratory protection on portable open spraysystems that would be designed with inertsheathing gas flows.
SOLVENT SUBSTITUTIONAPPLICATIONS
Ozone non-attainment and air toxic problemsin the United States are heavily influenced byemissions of volatile organic compounds fromstationary area sources7. Such small areasources include auto body shops and cabinetmakers, where applying polymer-based paintsand protective films invariably frees largeamounts of solvents to the atmosphere.Fortunately, the multi-component temperaturecontrol already discussed for spray formingmetals enables polymer films to be sprayed ina molten state with little or no solventemission and no later curing step. This newapproach is straightforward for linearpolymers, but the kinetics of crosslinkingreactions can also be accommodated.
27
In fact, feasibility of spray forming linear andcrossiinked polymers has already beendemonstrated at EG&G Idaho. Here theobjective was fabricating polyphosphazenemembranes for extracting sulfur compounds.Despite the narrow focus of this exploratoryexercise, no major impediments wereencountered that would prevent spraying otherpolymer-based materials. Furthermore,performance of the rectangular nebulizergeometry was verified with regard todepositing films of uniform thickness. Theseresults were embodied in another DOE patentidea record8. As co-inventors and a qualifiedsmall business, CST has applied to the DOEOffice of Patent Counsel for associatedtechnology transfer rights.
Another potential application for spray-formedpolymers relates to the semiconductorindustry, where protective films are used topackage printed circuit boards and otherelectronic modules against environmentaldegradation. Here spray forming would havetwo benefits beyond eliminating solventemissions. In current situations, difficultyoccurs frequently in obtaining a consistent filmthickness, because the dissolved polymer tendsto "run" prior to curing. Since sprayedmolten polymers could set up instantly bysolidification, deposited layers would resistlater relocation.
Protective packaging must also be kept fromcovering isolated areas such as electricalcontacts, which often requires masking duringthe application process. This would be thecase as well with spray forming, but a highlycollimated plume can be achieved byoptimizing the exit contour of aconverging/diverging nozzle. Consequently,spraying through a template would leave thepolymer film only where desired. Provisionsmust be made, however, for periodic templatecleaning, lest accumulation narrow theopenings.
Hazardous airborne emissions from thesemiconductor industry are not confined topolymer solvents, and attention must also be
paid to cleaning agents. The most notoriousof these agents are the ozone-depletingchlorinated fluorocarbons (CFCs), which areespecially important for removing flux fromsoldered connections on printed circuit boards.It is conceivable (albeit ambitious) thatconductors could be sprayed onto printedcircuit boards in a manner that altogethereliminates the soldering process.
The concept for solderless PC boards ispresented in assembly-line fashion in Figure 2.Beginning on the left side, a PC board entersthe process upside-down with electroniccomponents already mounted underneath. Asenvisioned, component connections would beflat tabs bent over both to secure thecomponents in place and to present aperpendicular surface to the spray plume. Thefirst step shown is placement of a templateover the PC board base, followed by acleaning and roughening operation such asgrit-blasting. Not only would this step removesurface contaminants from the exposedconnectors, but it would also roughen boththem and the board base for mechanicalbonding to conducting metal. Once remaininggrit has been blown free, a conductive coatingis sprayed through the template. At this point,the template is detached for removal ofaccumulated metal (probably by scraping),recycling of surplus conductor, and subsequentreuse of the template.
In evaluating this concept, it should berecognized that spray-formed mechanicalbonds can be quite strong. References 4 and6 describe coating bonds exceeding 3000pounds per square inch, with negligibleinterfacial porosity. Another attractive featureis that this approach opens the choice ofconductors to a wide range of alloys, ratherthan merely plated copper. And, of course,the photo-resist techniques now required toetch conducting paths on PC board backingswould no longer be necessary.
An additional method for reducing solventemissions by spraying metals concerns"coating-free" materials. The Environmental
28
Protection Agency is seeking new ways tocover exterior surfaces that do not demand aninitial layer of paint, as well as periodicreapplication of protective coatings7. Althoughbeyond the scope of this paper, a tandemnozzle concept exists whereby a corrosion-and wear-inhibiting layer could be co-deposited while spraying items such asaluminum and vinyl siding.
CONCLUSIONS
Spray forming has been introduced as aversatile process for fabricating materials,including both high-performance metals andpolymer coatings. With further research anddedicated development, this technology hasconsiderable potential for reducing solventemissions in several areas of immediatenational concern.
REFERENCES
and Article Produced Thereby," U.S.Patent Application No. 7.599.773.October 18, 1990.
Ploger, S.A. et al., Spray Coating ofMetals. Phase II: Proof of Concept.U.S. Air Force Engineering andServices Center, Tyndall AFB, FL, tobe published.
Kosusko, M., "Demonstration ofEmerging Area Source PreventionOptions for Volatile Organics," U.S.Environmental Protection Agency, Airand Energy Engineering ResearchLaboratory, Proceedings of the AIChE1990 Summer Meeting. San Diego,CA, August 19-22, 1990.
McHugh, K.M. et al., "Spray-Forming Process for Polymer Films,"DOE Case Number S-71.998. assignedAugust 20, 1990.
Watson, L.D. et al., "Nozzle-Aspirated Metal Forming," Paperpresented at the MetallurgicalSociety's International Symposium onCasting of Near-Net-Shape Products.Honolulu, HI, November 13-17,1988.
Hall, E.J., "Process for DisintegratingMetal," U.S. Patent 1.659.291.February 14, 1928.
Alvarez, J.L. and Watson, L.D.,"Apparatus and Method for SprayingLiquid Materials," U.S. Patent4.919.853. April 24, 1990.
Ploger, S.A. et al., Spray Coating ofMetals, Phase I: Feasibility ofConcept. ESL-TR-89-61, U.S. AirForce Engineering and ServicesCenter, Tyndall AFB, FL, May 1990.
Ploger, S.A. and Watson, L.D., "LowTemperature Process of ApplyingHigh Strength Coatings to a Substrate
29
isolctionChamoer
Housing
P — Pressure "ransducer F — Gas FlowmeterT - Thermocouoie S - Speea Sensor
REDUCTION OF SOLVENT USE THROUGH FLUXLESS SOLDERING*
F. Michael HoskingSandia National LaboratoriesAlbuquerque, New Mexico
ABSTRACT
Conventional soldering typically requiresfluxing to promote wetting. Halogenatedsolvents must then be used to remove the fluxresidues. While such practice has beenroutinely accepted throughout the DOEweapons complex, new environmental lawsand agreements will eventually phaseout theuse of these solvents. Solvent substitution oralternative technologies must oe developed tomeet these restrictions. SNL, Albuquerque ischaracterizing and developing alternativefluxless soldering technologies that will reducesolvent use and be compatible with prototypicpackaging materials. The program is focusingon controlled atmosphere (vacuum,inert/reducing gas, reactive plasma, andactivated acid vapor) soldering, metallizationand i n h i b i t o r t e c h n o l o g y , andthermomechanical surface activation (laser,infrared, solid state diffusion, and ultrasonic)soldering. Since there is no universal methodthat can be applied to every electronicapplication, the study is defining technologicaloptions and limitations. Fluxless solderingwould reduce the number of cleaning stepsand the subsequent volume of mixed solventwaste. This paper will present an overview ofthe effects of atmosphere, materials, andprocessing conditions on attaining a fluxlessoperation. Examples of applying thesetechnologies to electronic packaging will begiven.
•This work is supported by the U.S. Department ofEnergy under Contract DE-AC04-76DP00789.
INTRODUCTION
There has been increasing concern about theenvironmental effects of chlorotiuorocarbons(CFCs) by the scientific and politicalcommunity over the past decade. CFCs havebeen identified as a source of the depletion ofstratospheric ozone. Continued ozonedepletion would seriously affect both theenvironment and human health. The evidencefor this scenario is well documented (1-3).The most celebrated example of ozonedepletion is the ozone hole discovered over theAntarctic. It is being extensively studied andmonitored. The hole has been associated withthe emissions of fully halogenated CFCs andhalons. The Montreal Protocol, which is aninternational agreement that was originallydrafted and submitted for signature in 1987and has gone through several revisions,attempts to reverse this depletion problem.The current Protocol schedule requires acomplete phaseout of controlled CFCs by2000. Other fully halogenated CFCs, carbontetrachloride, and methyl chloroform will bealso affected by the Protocol controls (4).
The international restrictions on CFCs willsignificantly impact the electronics industry.Cleaning is a major element in electronicmanufacturing, especially as a part of solderprocessing. An electronic package is typicallypopulated with many devices (surface mountdevices, capacitors, resistors, chips carriers,leaded devices, etc.) that are attached to thehost board by one of several solderingmethods. Whether the operation is manuallyperformed with a soldering iron or batchprocessed with a wave soldering machine,each method has the common feature of usinga flux to help the molten solder alloy wet thebase material (5). The flux has threefunctions. The first is to chemically remove
31
surface oxides and provide a protective layerover the cleaned surface while solder wettingoccurs. The second is to assist heat transferto the joining surfaces. The third is to assistthe removal of the reaction products. Thereaction products and flux residue must beremoved after soldering. Although theresidues are generally nonconductive, they arecorrosive and could create a reliabilityproblem, especially for applications whereextended storage in uncontrolled environmentsis expected. These conditions make mandatorytheir complete removal from the assembly.
Most military electronic applications use rosinfluxes when soldering. The flux residues aretypically removed with halogenated solvents.This practice is changing because of theimpact of the Montreal Protocol. Newsolvents, fluxes, and cleaning methods arebeing consequently developed to satisfy theCFC phaseout. Terpene solvents, aqueousbased cleaning, water soluble fluxes, and lowsolids ("no-clean") fluxes are showingpromise. Alternative technologies, such asfluxless soldering, must also be developed tosupplement these other activities.
Fluxless soldering is not intended to eliminateall cleaning during electronic manufacturing,but it will reduce the total number of cleaningsteps and the subsequent quantities of mixedsolvent waste that must be handled. Anexample of this is step soldering where two ormore solder alloys with different meltingtemperatures are used to attach more than onecomponents in multiple processing sequences.If fluxing could be reduced or eliminatedduring these multiple steps, the need forcleaning could also be reduced and asignificant quantity of solvent saved. Thisquantity is dependent on the number and sizeof parts processed, but for a typical hybridmicrocircuit, up to 250 ml of mixed solventwaste can be generated from one cleaningcycle.
The purpose of this paper is to present anoverview of the Fluxless Soldering activitiesthat Sandia National Laboratories,
Albuquerque (SNL) has in progress. SNL ischaracterizing and developing severalalternative technologies that will be applied towaste minimization in the Department ofEnergy (DOE) weapons complex. The workis being funded by the DOE Office ofTechnology Development (DOE/OTD) whichis committed to developing faster, better,cheaper, and safer processes and materialswhich can achieve and sustain environmentalrestoration and waste management compliance.The project objective is to safely integratethese alternative technologies onto theproduction floor.
FLUXLESS SOLDERINGTECHNOLOGY OVERVIEW
SNL's Fluxless Soldering effort covers abroad range of technologies (6) that eitherreduce surface oxides or prevent surfaceoxidation prior to and during soldering. Mostof the technology currently exists but has notbeen fully applied to soldering. Fluxlesssoldering is consequently not well understoodand must be better characterized anddeveloped if it is to succeed in reducingsolvent use at the manufacturing level. Thereare four key elements to the SNL task. Theyinvolve the characterization and developmentof Controlled Atmosphere Soldering,Thermomechanical Surface ActivationSoldering, Metallization Technology, andInhibitor Technology, Figure 1. Theseactivities are being supported bythermodynamic and kinetic analyses andwetting experiments. Figure 2 lists theprincipal contacts in SNL's MetallurgyDepartment 1830 who are involved in theinvestigation.
Controlled atmosphere soldering utilizesvarious "clean" or reducing atmospheres tomaintain or produce a solderable base surface.These atmospheres typically depend on avacuum, inert or reducing gas, reactiveplasma, or dilute acid vapor-inert gas mixture(eg. formic acid and nitrogen). More will besaid on the use of these controlled
32
atmospheres for fluxless soldering in the nextsection.
Therrnomechanical surface activation solderingdepends on kinet ic or directedthermomechanical energy to spall or ablate thesurface oxide and facilitate wetting of theunderlying, pristine metal. Laser, solid statediffusion, and ultrasonic soldering are typicalways in which this can be accomplished.These processes can be done in air or in acontrolled atmosphere. An example offluxless laser soMering, coupled withmetallization and controlled atmospheretechnology, will be given in the next sectionon controlled atmospheres.
Ultrasonic soldering uses an ultrasonic probewhich is immersed in a solder bath andgenerates ultrasonic vibrations that reduce thinoxide layers through cavitation. It is difficult,however, to accurately direct these ultrasonicwaves. There is also a lack of fundamentallyunderstanding the interaction effects betweenthe process parameters and materials. SNL isconducting experiments to characterize thesefundamental properties. Cu and Al substratesare being fluxlessly and ultrasonically tinnedwith elemental Sn. The effects of tinningtemperature, probe separation, probe power,probe angle/position in reference to the basesurface, and vibration time on wetting arebeing studied. Preliminary results on flat Cusubstrates have shown that excellent wettingcan be achieved on both sides of the immersedsubstrate, but cavitation is very sensitive tosample thickness.
Protective coating technology also show-,promise as a compliment to fluxless soldering.Nonoxidizing surfaces, such as Au, have along history of being readily wettable withoutfluxing. Their deposition, however, must beclosely controlled. A thick layer of Au isgenerally required to guarantee completecoverage and wettability of the underlyingmetal. However, the metallization must notbe too thick or the extra Au will produce abrittle solder joint. Au metallizations shouldgenerally not exceed 3-5 wt. % in a
63Sn-37Pb solder joint. The resultingcompromise in thickness typically results in athinner, porous layer of Au that exposes theunderling metallic surface, usually Ni, anddegrades subsequent wettability underoxidizing conditions, Figure 3. These porousmetallizations can be protected by applyingorganic inhibitors, especially if stored in anuncontrolled environment before soldering.SNL is working with the University ofCalifornia at Berkeley to characterize themicrostructures and the fluxless wettability ofNi-Au platings and the State University ofNew York at Stony Brook to study thebonding behavior of organic inhibitors onmetallic surfaces and their effect onsubsequent solder wetting. Work is underwayto examine the effects of Au thickness andporosity on the degradation of wetting underfluxless soldering conditions. These protectivecoatings generally work in both air or acontrolled atmosphere, although a dry,nonoxidizing cover gas is more effective.
The above fluxless soldering technologies mustbe compatible with not only the base and fillermetals, but also with any neighboringmaterials that might be exposed to the sameprocess during soldering. Sensitivity to lasers,infrared heating, or reactive plasmas is ofspecial concern since they could effect thefunctional performance of an electroniccomponent. Materials such as alumina, glassfrits, epoxy, polyester, phenolic, polyimide,plastics, and conformal coatings could bedegraded by exposure to these processes.
CONTROLLED ATMOSPHERESOLDERING TECHNOLOGIES
Controlled atmosphere soldering (7) utilizesvacuum, inert or reducing gas, reactiveplasma, or acid vapor-inert gas mixtures thatfunction as either a protective or reducingcover during processing. Vacuum andinert/reducing atmospheres restrict the supplyof oxygen to the workpiece with oxygen levelsas low as 5 ppm. Although thermodynamicdata suggests that the reduction of metallic
33
oxides in hydrogen or vacuum is feasible, thekinetics for it to occur at typical solderingtemperatures, 200-300°C, is negligible and theoxide remains relatively stable requiring theuse of a flux. If fluxless vacuum or inert gassoldering is to succeed, therefore, the basemetal must be oxide-free throughout theheating and wetting cycle. Gas flow rates areimportant because volatile contaminants mustbe removed from the work area with adynamic flow of "clean" process gas.Metallizations, whether they are plated ortinned, provide an added margin for fluxlesssoldering in vacuum or inert atmospheresControlling the time at which the solder alloyis molten is also critical since extendedsoldering times could cause excessive solderalloy and base metal reaction and produce anew surface (eg. intermetallic) that coulddewet. Infrared heating helps to minimizesthermal gradients and heating times and isconsequently being applied to variouscontrolled atmosphere soldering systems.
A Controlled Atmosphere Solder WettabiiitySystem is being constructed at SNL to studythe effects of vacuum, inert gas, and dilutehydrogen reducing gas atmospheres onfluxless wetting. A schematic of the system isshown in Figure 4. The system uses aneiectrobalance to measure wetting force as afunction of time. An auxiliary video systemcan record the wetting event and analyze thewetting images to determine the effects ofprocessing conditions (pretreatment,atmosphere type, soldering temperature,immersion time, flow rates, etc.) and materials(base metal, metallization, solder alloy,inhibitor, etc.) on achieving fluxless wetting.A second system is available to performarea-of-spread (sessile drop) experiments inactivated acid vapor-inert gas atmospheres.
SNL has developed fluxless laser soldering tofabricate the closure joints on an electron";radar package. The process utilizes thecombined features of controlled atmosphere,metallization, and thermomechanical surfaceactivation soldering. The application joinsNi-Au plated Kovar pieces with Sn-Pb,
In-Pb-Ag, or In-Pb solder alloys and a 100watt CW Nd:YAG laser, Figure 5. The laserbeam is directed on the solder preformbecause of the reflection properties of the Auplating that would inhibit the absorption of thelaser energy. Satisfactory hermetic jointswere achieved with a 90 watt, 0.4 mm spotsize, and 5 mm/s travel speed laser setting ina forming gas cover of 5 vol. % hydrogen inargon. Figure 6 shows a cross-section of atypical laser soldered joint from the parametricstudy. Although the process is beingdeveloped for closure joints, it can be readilyapplied to attaching discrete leaded devices.
Reactive gases or plasmas are also beinginvestigated. Reducing plasmas can be used ina two step process that cleans the base metaland solder alloy during the first step and usesan auxiliary heat source, such as a heatedplaten, laser, or infrared heater, to make thesolder joint in the second step, Figure 7.Compatibility between the plasma and thepackaging materials is an importantconsideration. Reactive gases have thepotential for effectively reducing surfaceoxides. For example, thermodynamic datasuggests that atomic and ionic h\drogen havea higher copper oxide reduction potential thanmolecular hydrogen at 250°C, Figure 8. Ionichydrogen appears especially effective.Preliminary cathodic plasma cleaningexperiments on heavily oxidized Cu haveresulted in oxide-free solderable surfaces,Figure 9. Experiments are underway tocomprehensively characterize the effect ofionic hydrogen on Cu and Ni oxide reductionand fluxless wetting.
The final element of SNL's controlledatmosphere soldering effort is focused on acidvapor-inert gas mixtures. Although there arecommercial systems available that usevariations of this process, the fundamentals oftheir operation are not well characterizedThe process is readily applicable to wave orbatch furnace soldering. Figure 10. Diluteadditions of formic or acetic acid vapor areadded to argon or nitrogen to promote fluxlesswetting through the reduction of metallic
34
oxides:
MO + 2HCOOH — > M + 2CO2 + H2+ H2O
In actual practice, it is difficult to reduce mostsurface oxides and an adipic acid additive isgenerally required to achieve wetting. Adipicacid is a major constituent of low solids or"no-clean" fluxes. As with the low solidsfluxes, the adipic acid residues left on asoldered component must be completelyremoved to satisfy the long term reliabilityrequirements imposed on most militaryelectronic applications. SNL is characterizingthe effect of these acid vapor additions onoxide reduction and fluxless soldering. Theacid-gas mixture, gas flow rate, activatingadditives, soldering temperature, time, andmaterials are important parameters thatinfluence wetting and their interaction effectsare being determined. The objective is todevelop a scientific understanding of how theprocess works and what must be done to attaintrue fluxless wetting.
SUMMARY
Fluxless soldering is a viable and supplementaltechnology to solvent substitution for theelectronics industry. It has a high potentialfor reducing the storage and handling ofhazardous fluxes, environmentally harmfulsolvents, and the subsequent mixed solventwaste generated by flux residue removal.Since there is no universal method that can beapplied to every soldering application,technologies must be identified, characterized,and developed to satisfy the increasing numberof environmental, safety, and healthregulations. The objective is to quicklyintegrate these fluxless processes into fullscale manufacturing.
SNL, Albuquerque has an active program thatis evaluating various fluxless solderingtechnologies. It includes controlledatmosphere soldering, thermomechanicalsurface activation soldering, metallization
technology, and inhibitor technology. Theseprocesses offer a wide range of fluxlesssoldering options which can be combined toenhance the reliability of the final product.Laser and atmosphere soldering are excellentexamples of this dual technology concept.The key to fluxless soldering is to maintain a"clean" surface that the molten solder willdirectly wet or to reduce surface oxides thatthe solder will not wet. Materialscompatibility must also be considered.Otherwise, the functional performance of thefinal product could suffer if the selectedprocess degrades sensitive components nearthe solder joint.
ACKNOWLEDGEMENTS
The author would like to acknowledge thework of members from the SNL FluxlessSoldering Task group. Charlie Robino, PaulVianco, Darrel Frear, Dave Keicher, MarkSmith, and Rob Sorensen were especiallyhelpful in providing background andexperimental information. I would like to alsoacknowledge the work of Rusty Cinque,Choong-Un Kim, and Bill Morris ofUC-Berkeley and Clive Clayton ofSUNY-Stony Brook. I also appreciate theprogram support of Joan Woodard, SNL, PamSaxman, DOE/AL, and Clyde Frank,DOE/OTD.
REFERENCES
1. Molina, M. J. and Rowland, F. S.," S t r a t o s p h e r i c S i n k f o rChlorofiuoromethanes: Chlorine Atom- Catalyzed Destruction of Ozone,"Nature, 249, 810-812 (1974).
2. Derra. S.. "CFCs, No EasySolutions," R&D Magazine, 56-66(May 1990).
3. Anderson, S. O., "Progress by theElectronics Industry on Protection ofStratospheric Ozone," 40th Electronic
4. Schuessler, P., "CFC Alternatives:E x a m i n i n g the E m e r g i n gTechnologies," SUNY-BinghamtonSoldering Technology for ElectronicPackaging Symposium Program,November 12-13, 1990.
5. Wassink, R. J. K., "Soldering inElectronics," ElectrochemicalPublications, Scotland, 2ndEdition, 204-262 (1989).
6. Hosking, F. M. (PrincipalInvestigator), "Fluxless Soldering toReduce Solvent Use," DOE/OTDTechnical Task Plan, revised January30, 1991.
7. Hosking, F. M., "Fluxless Solderingwith Controlled Atmospheres,"SUNY-Binghamton SolderingTechnology for Electronic PackagingSymposium Program, November12-13, 1990.
• Inhibitor Technology - Rob Sorensen, 1834Mike Hosking, 1833(SUNY-Stony Brook)
Figure 2. Fluxless Soldering group responsibilities in SNL's Metallurgy Department 1830.
37
Oxide Free Surfaces Are Necessary If Vacuumor Inert Gas Soldering Is to Succeed
Oxidation and Corrosion TransportThrough Thin, Porous Au Suriaca
Submit*
• Au metallizations can provide anoxide free, solderable (fluxless)surface.
• Control of the Au thickness iscritical; too thick and Au inter-metallics will embrittle the joint;too thin and porous Au will allowoxidation of the underlying metal.
• Recommended Au thickness is50-75nin. (1.3-1.9 urn).
• Fraction of Au in a 63Sn-37Pbsolder joint should not exceed3-5 wt. %.
FMH-1833
Figure 3. Fluxless soldering can be achieved in a controlled atmosphere by overlaying the basemetal with Au.
Solder Wettability In Controlled AtmospheresCan Be Determined With A Wetting Balance
Controlled Atmosphere Solder Wettability System
Elactrobalanca
VldaoCamara
UanlseusHalght.h
vst
SampiaComputer
Solder Pot
Step Motor
I Fw
FUH-U33
Figure 4. Diagram of the Controlled Atmosphere Solder Wettability System which is underconstruction and will characterize the effect of controlled atmospheres on fluxless wetting.
Objective: Attach discreteelectronic components in aprotective inert or formingcover gas with laser heatingand no fluxing.
Electronic Device
X-Y Stage
Figure 5. Laser and controlled atmosphere soldering can be combined to produce a fluxlesssoldering operation.
250 50
Figure 6. Optical micrographs of a fluxlessly laser soldered joint (Ni-Au plated Kovar with60Sn-40Pb solder) showing excellent solder wetting and flow.
39
Two Step Plasma Cleaning and Soldering IsBest Suited for Batch Processing
EXAMPLE
SUD/HUCPlasma
Elodrode
Quartz Barrel Etcher
• Cleaning Variables -a) powerb) chamber pressurec) time
Soldering (flux/ess)assisted with auxiliaryheating (hot stage,laser, infrared)
Reducing plasma produced byilectric field.
anRF
FMH-1133
Figure 7. Fluxless soldering can be accomplished in a two step operation: plasma cleaningimmediately followed by soldering with an auxiliary heat source.
Thermodynamics of the Reduction of CU2O at 250°CSuggests That Ionic Hydrogen Has the Best Potential
Gas Reduction>pecies Factor
HiHH-H+
168
39
Sptcia
FMH-1833
Figure 8. Thermodynamic data demonstrating the oxide reduction potential of ionic hydrogen.
40
Figure 9. Heavily oxidized copper tube cathodically cleaned with a reducing plasma.
Typical Process Variables of ActivatedAcid Atmosphere Soldering
of formic acid)• Low Solids Additive (1 l/hr)• Preheat and Soldering Temperatures- Board Throughput
' Difficult to wet surfaces mayrequire a dicarboxylic acid additive(eg. adipic acid). This "no clean",low solids addition can be variedfrom 0.5 to 1.5 % and applied with anultrasonic atomizer in an alcoholcarrier.
FMH-1S33
Figure i(). Dilute additions of formic or acetic acid vapor to an argon or nitrogen atmospherehave the potential for fluxless soldering.
41
PLASMA STRIPPING OF MAGNETIC COMPONENTS
T.J. Gillespie and T. MehrhoffGE Neutron Devices*
Largo, Florida
ABSTRACT
A plasma-stripping process and its associatedproduct fixtures and equipment has beendeveloped and evaulated for stripping wireinsulation on both coils and magneticassemblies. The stripping process usestetrafluoromethane in oxygen as the activeplasma gas and a metal plasma containmentsystem. The system provides residue-freestripped leads when inspected at 200Xmagnification using Scanning ElectronMicroscopy (SEM). An evaluation of theplasma-stripping process was conducted on afly spec inductor and a converter assembly.The results showed that the parts met alldrawing requirements. In al! cases, thefixtures provided a sufficient shield to preventthe plasma from attacking areas of the productwhich were not to be stripped. Materialspecifications for the plasma gas mixture anda Manufacturing/Engineering EquipmentInstruction (MEEI) have been issued tosupport the plasma stripping activity inproduction.
INTRODUCTION
In 1985, the magnetics production assignmentwas assumed by GE Neutron Devices(GEND). One of the processes required bydrawing as a part of this transfer involved theuse of Iso Verre™ chemical stripper to stripwire insulation on the many types of products.Iso Verre is formulated in France and containshydrofluoric and formic acid, phenol and
*GE Neutron Devices operates the Pinellas Plant for theU.S. Department of Energy under Contract No. DE-AC04-DP00656.
methylene chloride. The material is providedas a kit with a bottle of thinner supplied witheach four bottles of active stripper. Inproduction, the basic stripper must be thinnedto a viscosity that can easily be applied to thewire. The stripper is a very active chemicalagent and causes frequent burning of operatorsand severe fume problems. One of the mainobjectives of this study was to eliminate theuse of the Iso Verre chemical stripper inmagnetics production.
In 1986, GEND and Sandia NationalLaboratories (SNL), Albuquerque, initiated ajoint study on the development of a process toplasma-strip wire insulation. The initial studywas designed around the fly spec inductor,which was the most difficult product to strip,since it required the leads to be cleaned suchthat they could be thermosonically bonded andrequired stripping within .010" of the core(see Figure 1).
In addition to the flyspec inductor, there weremany magnetic assemblies where up to 4 coilswith a total of 18 leads were mounted indiallyl phthalate (DAP) contact assemblies.Each lead on these assemblies requires that theinsulation be stripped so that the lead can besoldered to a specific contact on the contactassembly. Some examples of these assembliesare shown in Figure 2.
PLASMA PROCESS DEVELOPMENT
Initial plasma stripping studies were performedin plasma cleaners with quartz-barrel reactorsusing various types of plasma gases includingargon, oxygen, helium, pure air, sulfur
43
hexaflouride and combinations of these gases.The studies were performed on polyester-coated copper and gold wire. Results of thesestudies indicated that complete removal of allinsulation residue and oxicdation of the gold-coated wire would be major problems. Anadditional major problem that surfaced at thistime was development of a masking materialthat could survive the plasma treatment andprovide a tight seal around the leads, sinceany small opening in the seal would causeetching of the product. These problems weresolved by switching to a primary-type plasmacleaner in an all-metal enclosure (shown inFigure 3) using tetrafluoromethane in oxygenas the active plasma gas and using unfilledSylgard™ as the sealing material.
PRODUCT PROCESS DEVELOPMENT
Ferrite Core Inductor
The inductor is wound on a ferrite core withan outside diameter of .050" and an insidediameter of .020". The wire is AmericanWire Gauge (AWG) No. 40 (.003"). Theconductor is gold plated and is covered by apolyester insulation. A typical ferrite coreinductor is shown in Figure 4.
Winding Development
Early in development of the processes for thisinductor, the intent was to develop both newstripping and winding processes, since manualwinding of the inductor was very slow.GEND purchased two winders for this productthat were fabricated by the JovilManufacturing Company of Danbury,Connecticut. Initial evaluation of the windershowed that the polyester-coated wire was tooweak to withstand the winding process; itcracked at each core/wire interface.
The General Electric Research Laboratoryrecommended that GEND evaluate the winderusing other insulations. Following this lead,
GEND evaluated polyester amid imide,polyimide and polyester imide nylon.Evaluation of these insulations showed that allof the new insulations resisted cracking duringthe winding operation. However, a newproblem emerged; all wires were found tostretch during the winding process. Theresulting reduction in cross section of the wirecaused the product to violate the productspecification. No further winding activity hasbeen performed to date.
Having established that the combination of themetal p l a sma c l eane r and thetetrafluoromethane/oxygen gas would provideacceptable stripping of the wire insulationwithout oxidation of the gold-plated coating, itbecame a matter of designing the right fixtureto achieve an acceptable product. The fixtureshown in Figure 5 is designed to hold thewound inductors in a Sylgard pocket with thewires extending through the seam of thepocket. Twenty inductors can be processed ineach fixture, and up to three fixtures can beprocessed at one time in the plasma cleaner.
Two evaluations of the plasma strippingprocess have been performed in production todate, with 200 products being made in eachevaluation. All products met thespecifications. The stripping process wasperformed at .5 Torr, 400 watts for 45minutes.
PLASMA STRIPPING OF MAGNETICASSEMBLIES
When an operator is preparing to finish-soldera magnetics assembly, he or she routes thewire in the specified path to the contact towhich the wire is to be soldered. The wire isthen bent around the contact, and the bend isused as a stripping guide with all theinsulation being removed beyond the bend inthe wire. The stripped wire is then loopedaround the contact twice and soldered in place.
44
In designing a fixture to strip a magneticsassembly, the outside edge of the strippingmask defines where the stripping action starts.A spring-loaded retainer is positioned in thefixture to hold the wire rigid during stripping.After stripping has been completed, the wireshould be attached to the designated contactwith the wire following the normal wirerouting. Figure 6 shows a photograph of twoconverter assemblies in position in thestripping fixture and a diagram of theassembly after wiring has been completed.
CONCLUSIONS
A mixture of 8% tetrafluoromethane in oxygenis effective in stripping wire insulation suchthat no residue can be detected at 200X usingSEM. The plasma stripping process is veryfixture dependent, but it appears from theseevaluations that this process can be effective instripping wire insulation in production.
ALTERNATE METHOD OF STRIPPINGASSEMBLIES
Many of the magnetic assemblies, such as theone shown in Figure 6, consist of one or morecoils mounted into a DAP contact assembly.In these assemblies, the coils are secured inthe contact assembly with eposxy prior tobeing stripped. The alternate method ofstripping an assembly is to strip the coil todimension prior to securing the coil into thecontact assembly, the advantage being thatmany more products can be stripped at onetime and the number of fixtures required couldbe significantly reduced. A fixture has beendesigned to strip the coil on the magneticassembly previously mentioned and ispresently being fabricated.
45
Stripping Req
0.010 - 0.060"
Figure 1. Flyspec Inductor
Figure 2. Examples of Magnetic Assemblies RequiringStripping
46
Figure 3. Metal Containment Plasma Cleaner
Figure 4. Wound Flyspec Inductor
47
Figure 5. Stripping Fixture for Flyspec Inductor
a. Assemblies in Positionin Stripping Fixture
b. Assembly After Wiring
Figure 6. Converter Assemblies in Position in the Stripping Fixture and Assembly After WiringHas Been Completed
48
SODIUM BICARBONATE BLASTING FORPAINT STRIPPING
N.E. Wasson Jr. and Michael N. HaasU.S. Air Force
Kelly Air Force Base, Texas
ABSTRACT
The San Antonio Air Logistics Cente-(SA-ALC) is one of five industrial activities inthe United States that suppois themaintenance and modification requirements forthe United States Air Force. SA-ALC isespecially concerned about the corrosioncontrol requirements necessary to support theirmission, since they remove the coatings,inspect, and repaint over 1.6 million squarefeet of delicate aerospace structures each year.Recent legislation and public awareness hasencouraged the pursuit of alternatives to thecostly chemical stripping operation. PlasticMedia Blasting (PMB) has come forth as themost viable alternative in recent years. Theprimary weakness of the PMB process,however, is the generation of a substantialhazardous waste stream. SA-ALC decided toinvestigate the use of Sodium Bicarbonate asa coatings removal technique, because theprocess offered the potential to eliminate over90% of this waste stream. The SodiumBicarbonate, better known as the Bicarbonateof Soda Stripping (BOSS) process, is similarto the PMB process except that a smallvolume of water is injected into the blaststream at the nozzle to eliminate nuisancedust. SA-ALC decided to perform a thoroughevaluation of the process to determine whereit could be used in their industrialenvironment. This paper outlines the workinitiated for evaluation of the BOSS process;anothei report will summarize the results.
BACKGROUND
Four primary areas of study will be conductedto determine whether the BOSS process isviable for use on aerospace structures. Initial
work will be centered around aluminum alloyssuch as 2024-T3 and 7075-T6 clad and barematerials. First, corrosion tests will beconducted to determine what effects anyresidual sodium bicarbonate or its byproductsmay have on an aerospace structure whenexposed to the temperature and humidityconditions of a military operating environmentaircraft. Second, the waste stream will beexamined to determine the most suitable wastehandling and disposal methods. Third, anoptimization of the process, varying the manyparameters affecting material degradation andproduction rates, will be analyzed to developthe most suitable parameters for thin-skinned,aluminum aerospace structures. The finalwork will be a material characterization of theoptimized parameters to determine exactlywhat long term effects the process may haveon the life of the structure. Most of this workwill be performed by independent researchlaboratories under contract with thegovernment. The corrosion testing and thewaste stream evaluation are under contractalready and work has begun. Theoptimization and material characterization testshould begin by September, 1991.
CORROSION TESTING
Research has been performed over the last twoyears by J.H. Van Sciver Associates, whospecialize in materials and corrosionengineering. The corrosion work theyperformed has demonstrated there is noadverse corrosion effects on aluminum alloystested in sodium bicarbonate solutions,especially when compared to chemicalspresently used in aircraft paint stripping. TheAir Force Corrosion Program Office ispresently managing a Corrosion Study
49
contracted to Battelle Laboratories. Thepurpose of the study is to evaluate the BOSSprocess for military weapon systems. Theprogram consist of two phases. Phase I willaddress the thermal stability of potentiallyentrapped BOSS media when exposed toaircraft operating environments. Phase II willdetermine the relative corrosivity of the BOSSmedia and byproducts vs. existing paintstripping methods (Chemical and PMB) usedon aerospace materials. Phase I will becompleted in March, 1991, and Phase II willbe completed by October, 1991.
WASTE CHARACTERIZATION
A contract has been awarded to EG&G IdahoInc. to perform an evaluation of the wastestream of the BOSS process and how to treatit. The work to be performed by EG&G willinclude:
1) Development of a test plan forcharacterization studies.
2) Characterization of new and spent BOSSmedia.
3) Identification and evaluation of materialseparation technologies.
4) Prepartion of a final report on materialscharacterization and separationtechnologies.
This work should be completed by the end of1991. The primary objective of the study willbe to develop an efficient method of separatingthe paint particles from the waste stream.Then the bulk of the waste could be disposedof as a non-hazardous component in thesanitary sewer system; the hazardouscomponent consists primarily of paint chips,will be reduced to an absolute minimum. TheBOSS process will help the Department ofDefense work towards the goal of reducing itshazardous waste stream.
PROCESS OPTIMIZATION
Test specimens will be primed, painted andartificially aged. The specimens will beblasted to determine the optimum parametersto produce the best combination of strippingrate with the least amount of damage to thesubstrate. Parameters to be varied are:traverse speed, standoff distance, angle ofimpingement, nozzle pressure, media flowrate, and water pressure. Once theseparameters have been established, they will beused to perform a material characterizationstudy, quantifying the damage imparted by theabrasive effects of the sodium bicarbonate onclad and bare aluminum alloys. This workwill probably be contracted out to anindependent test laboratory. This will be(Phase I) with the follow-up materialcharacterization to be performed by the samelaboratory. The award of a contract isexpected to occur by September, 1991 withwork completed by January, 1992.
MATERIAL CHARACTERIZATION
Phase II of the contract will be to characterizethe effects of the optimized process parameterson 2024-T3 bare aluminum. We feel the datagenerated through the evaluation of the PMBprocess has adequately characterized theeffects of abrasive paint removal processes. Asubstantial amount of time and money will besaved by limiting the material characterizationstudies without affecting the accuracy of theassessment of die BOSS process. Stresssaturation curves will be generated by usingalmen strip data. This data can be correlatedto the existing PMB data without having to doextensive fatigue, tension, crack growth, andsurface flaw fatigue testing. Some X-RayDefraction testing with some minor fatiguedata will be generated for accurate correlationto the existing PMB data. This work will becompleted by March, 1992.
50
SUMMARY
Coatings removal, utilizing sodiumbicarbonate as the abrasive, appears to be aviable alternative. Concerns of long-termcorrosion effects of any media left in anaircraft or component will be resoived in thisstudy. The material characterization isexpected to mirror that of PMB and otherabrasive blasting processes. The wastecharacterization will determine the truepotential of this process. The promise of theelimination of one of the Department ofDefense's largest waste streams offers themost benefit. The results of this work will bepublished upon final completion of all testsdiscussed in this paper.
51
LOW TOXICITY PAINT STRIPPING OF ALUMINUM ANDCOMPOSITE SUBSTRATES
Nona LarsonBoeing Aerospace
Seattle, Washington
INTRODUCTIONPart 1
The effective chemical paint strippers foraerospace coatings are toxic and present ahazard to both personnel and the environment.They contain phenol, methylene chloride andhexavaJent chromium materials which havebeen targeted by governmental regulations forfuture elimination. These strippers are alsodetrimental to composites. This report focuseson impact damage inflicted on compositesubstrates by some of the mechanical methodsmeant to replace chemicals.
OBJECTIVE
The objective of this program is to find benignalternates to hazardous chemical paintstrippers.
APPROACH
The approach with the least impact onproduction was taken for this project; i.e.,identify and evaluate commercial chemicalstrippers which did not contain the targetedmaterials; if none were available, develop achemical stripper; and finally evaluate non-chemical methods. To accomplish this work,the following three tasks were identified:
Task 1: Evaluate Chemical Strippers
Identify coatings and substrates to beevaluated; contact commercial sovrces ofstrippers; evaluate strippers based upon theirability to strip the coating, length of time ittakes, and damage sustained by the substrate.If necessary, blend Boeing proprietaryformulas.
Task 2: Evaluate Non-ChemicalStripping Methods
Investigate the mechanical, radiation and otherstripping methods being tested throughoutindustry and evaluate only those which appearto offer solutions to aerospace problems.
Task 3: Specification Coverage
Change Boeing specifications for abrasiveblasting to include effective and non-damagingpaint removal methods. This will includeProcess Specification Departures for interimsolutions and specific applications.
CHEMICAL STRIPPER ALTERNATIVES
Conclusions from Boeing Document D180-30690-4, Comprehensive Chemical ReductionResearch Projects Final Report 1989,summarize the work from Tasks 1 and 2. Nolow toxicity chemical paint strippers werefound in this program which can berecommended to replace those presently used;however, the four strippers: Brulin EXP 2187mod., ManGill LP4566, and Turco 6088 maybe used effectively on selected coatings. Forexample, EXP 2187 mod. is extremelyeffective on removing melamine enamel andmilitary epoxy primer; Turco 6088 is effectivein removing phenolic resin varnish. Themechanical methods appear to be the bestoverall approach to removing the highperformance aerospace and commercialcoatings. Results are shown in Tables 1 and 2.
Recommendations made in Dl 80-30690-4 statethat we should monitor (rather than duplicate)data on alternate stripped methods, such asplastic media blasting and laser paint removal.
53
TECHNICAL ACCOMPLISHMENTS
For 1990 this program was to continue on amonitoring basis. Processes included were:plastic media, sodium bicarbonate, and carbondioxide blasting, high pressure water jet,xenon lamp and laser paint removal.Envirostrip wheat starch media was notavailable in 1989.
Wheat Starch Media
In April 1990, a new paint stripped media wasintroduced'. This media, made of 100%crystallized wheat starch, is non-toxic,biodegradable in the true sense of the word,and made out of a renewable resource (asopposed to petroleum based plastics). Sincethis product was new, independent researchfor us to monitor, we started our ownevaluation program. The first step in thestudy of this new process was to send samplesto the vendor (Ogilvie Mills). When theycame back, these samples looked good enoughto pursue this process more actively thanprogram planning had anticipated. Mediacharacteristics and results are summarized inTable 3.
The coatings we tested were chosen becausethey were the most difficult to remove of themilitary and commercial airplane specificationcoatings. Some of these were brought to uson scrapped or test parts as a challengebecause the owners did not believe that thewheat starch media would remove the coating.Table 4 lists the coatings removed, and Table5 lists the effects.
Composite Testing
Wheat starch blasting caused the least damageof any other blasting method we testeddirectly. Comparisons were made by blastingidentical panels (composites laid up at thesame time by the same person), or by blastingdifferent pans of a single large panel. Figures1 to 18 show these comparisons. These cross-
sections were photographed at 200x.
Aluminum Testing
No damage was detected on 0.020-inch thickaluminum, with the exception of somedeformation of the clad surface. Figures 19and 20 show results of stripping the thinaluminum panels. Sandwich corrosion testswere performed per D6-17487, CertificationTesting of Aircraft Maintenance Materials.As expected, the media passed both in thecrystalline state and dissolved in deionizedwater.
SUMMARY
Discussion of Envirostrip Wheat StarchBlasting
It is not difficult to explain the differences inresults obtained with this and other blastingmedia. Figures 21 & 22 show the fracturesurfaces of new and used media. A particlehitting the paint surface at a pressure aboveapproximately 30 psi will break. The glassyfractures shown in the figures explain why themedia remains effective cycle after cycle. Itbecomes more effective as the particles getsmaller simply because there are more sharpedges per pound of media in the blast stream.The limiting factor for size of this media isdrag. When the particles break down enoughthat dust hangs in the air, the operator cannotsee through it. A dust separation systemeasily removes this.
The fact that the particle breaks at around 30psi makes the media very forgiving. Turningup the pressure will increase the flow rate, butwill not make the media itself moreaggressive.
One common concern when using "wheat" ina blasting system is the so called "silo effect,"where a high concentration of wheat dust willspontaneously combust. This is not expectedto be a problem because crystallized starchmolecules are different in structure than thenaturally occurring polymer. Nevertheless,
54
we asked the vendor to supply data on theexplosive limits of the dust. The followingtable summarizes the finds.
Envirostrip Properties:
Media Size
Explosive Limits
Ignition Min. ExplosiveMesh (US Std) Temperature Concentration
12/3030/5050/250Dust Bin
°C
>850>850
530460
oz per ft3
Non-explosiveNon-explosiveNon-explosive0.060
There are many interesting and usefulproperties to this media which we discoveredduring our study.
a. Envirostrip does not remove alodine.The aloding remains after depaintingpasses 7-day salt spray exposure, butit is dehydrated so paint adhesionsuffers. This can be remedied by aquick dip in an alodine tank. Tenseconds should be sufficient.
b. New media is less aggressive than usedmedia. When stripping Kevlar, newmedia gives the operator morecontrol than the used media.
c. Large media works better on theelastomeric coatings, while smallmedia works better on the more brittlecoatings, i.e., epoxy primers andtopcoats.
d. The media works best by breakingthrough the paint surface and peelingthe paint up from the edges of thestripped area.
e. Some of the harder coatings (i.e.,BMS 10-11 type 1 epoxy primer) willbreak the media down faster thanothers. This is not true of Mil-P-23377.
Discussion/Results of Other Methods
There are a number of paint stripping methodscurrently commercially available. Each ofthese has good and bad points. Theinformation listed herein has been gatheredfrom a variety of sources in addition to thedata generated by this project.
Xenon Flash lamp Depainting
This method does not lend itself well toproduction use. Coated surfaces are exposedto high intensity pulses of light. A specialhead to focus the light must be used for eachdifferent part configuration. The threeprimary drawbacks are that it is not veryeffective on light colored coatings, it treatscomposite substrates the same as paints, andacutely toxic gases are released whenpolyurethane paints are broken down withoutadequate oxygen flow. An oil smut is leftbehind the flashlamp, so a cleaning step suchas carbon dioxide pellet blasting is required.Dark, low-gloss topcoats (i.e., MIL-C-83286)can be removed at rates up to one square footper minute. Light colors are removed muchmore slowly, and high gloss white is notefficiently removed. No damage to metalsubstrates has been found with this process;the metallic surface completely reflects thexenon flash. Composite surfaces do notreflect the flash, and are therefore removed inlayers analogous to paint.
Laser Depainting
This method is similar to the Xenon flashlampmethod, but is more easily controlled. Laserdepainting is geared toward roboticapplications where capital cost is high, butsome applications justify the cost.Optimization is difficult due to theuncontrollable variations in paint thickness.Light colored and high-gloss topcoats areremoved less efficiently than darker coatings,but they can be removed in a reasonableamount of time. Two companies are in theprocess of making this method commerciallyavailable.
55
Sodium Bicarbonate Blasting
This method, more than any other mentionedin this report, is a case of trading off pros andcons. The media is very effective withoutcausing much impact damage to the substrate.Since rinsing is very difficult, there is a highpotential for corrosion problems. The sodiumbicarbonate breaks down in water, formingsodium sesquicarbonate, which has a pH ofapproximately 10. Leaving this on aluminumparts will be very detrimental to the part. Theusual solution to this problem is to use a diluteacid rinse. Small parts are well suited to this,but large parts or structures are not. Thedilute acid rinse also creates more hazardouswaste, driving up the cost of the total paintremoval process.
The media itself is inexpensive at about $0.50per pound. Unfortunately, it is not recyclable.Enormous amounts of water are required todissolve the spent media before sewering. Tokeep the dust down, it is blasted with waterwhich contributes to the corrosion potential.The positive aspect of non-recyclable media isthat a dedicated facility is not required.
Carbon Dioxide Pellet Blasting
CO, pellet blasting has been the subject of anenormous amount of testing. This process isideal in the sense that no toxic substances aregenerated or released, a dedicated facility isnot required, and precleaning or surfacepreparation of the painted surface is notneeded. However, there are some veryserious areas of concern with this process.Some of which limit the use of CO2 blasting tosteel or very thick aluminum parts. Very highpressures are used to accelerate the particles.The particles impact the surface at extremelyhigh velocities, high enough to leave dents in0.020 inch thick aluminum.
There are two systems commercially availableto produce CO2 pellets for blasting. The morewidely used system extrudes the pellets,producing small cyclinders. They are notoptimally shaped for removing paint, and
unfortunately the optimization of the systemthat has been performed to date has consistedof changing pressure, impingement angle andstand-off distance. There has been someplanning to optimize the process from thefront end, beginning with producing pelletswith different shapes, sizes, and/or hardness.This new approach to optimization shouldincrease the strip rates. Previous rates havebeen slower than acceptable for productionoperations.
Plastic Media Blasting
Plastic media blasting has been studied ingreat detail. There are a number of completeprograms, both military and commercial,which report widely varying results (Ref. 2 to12). The process is partly accepted (one cycleonly) by the FAA, and it is widely used by themilitary (Ref. 13 & 14). There are seventypes of media, varying from soft to hard andmild to aggressive. The soft media requires alonger dwell time, so it does not necessarilycause less damage than the harder types.Plastic media is quite aggressive oncomposites, causing an unacceptable amountof erosion and/or fiber damage, with theexceptions of graphite or boron epoxy.
Ice Crystal Blasting
The Canadian Navy funded a program todevise a way to remove coatings from theinteriors of submarines. This method must besafe in a confined environment with minimalair flow. This method is very effective on theinterior coatings it was designed to remove.The company which invented the process isnow optimizing to remove aerospace and otherhigh performance exterior coatings. Progressis being made in this area. The coatings canbe removed with very little damage to anysubstrate tested, including composites and thinaluminum. This process will be watchedclosely during this program, In addition tobeing effective, it has the attractive propertiesof requiring no precleaning of the paintedsurface, minimizes the waste generated, doesnot require a dedicated facility, and is
56
completely non-hazardous to personnel.
General Concerns
Two primary concerns have been repeatedlyexpressed about any blasting method for paintremoval of structures: crack closure andmedia entrapment.
Crack closure has been a concern ever sincethe early days of plastic media blasting.Research has shown (15) that different plasticmedia types do not close cracks, but the mediacan become lodged in the crack and prevent itfrom showing up during dye penetrantinspection. This is more likely to happen withplastic media than with other types that arewater soluble.
Media entrapment is of serious concern tosome, while others do not consider it a majorproblem. Any accelerated particles will findways to enter openings. This can add weightto an aircraft and interfere with moving parts.Ideally, these areas would all be maskedbefore stripping, but in reality, this is notalways done properly. The other side of thisissue was expressed very well by a shopforeman (16) who said, "I can live with that,"comparing media entrapment to residualchemical paint stripper that was oozing out offaying surfaces on the repainted airplaneparked behind him.
CONCLUSIONS
During 1990 this program investigated anumber of mechanical paint stripping methods.Most of these methods were evaluated bymonitoring research done by others. Due tothe newness of the Ogilvie Mills Envirostripmedia, there was no outside work to monitor.Therefore, we did our own evaluation.Overall, this media proved to be the bestcurrently available technology. Clearly thereis no panacea, but this method is more widelyapplicable than any other. Recommendationsfor using this method on aircraft are awaitingfatigue testing, which the Boeing Commercial
Airplane Group plans to do in 1991.
On the other hand, chemical strippers willremain a problem, especially when pH is aconsideration. There is simply no lowtoxicity, low vapor pressure analogy formethylene chloride and phenol. When tankstripping is possible, the N-methyl-2-pyrrolidone based formulations can be used atelevated temperatures. However, these casescan usually be mechanically stripped, makingpreferential the wheat starch media. When pHis not a factor, concentrated, low molecularweight organic acids will work, but they arerelatively slow and unpleasant to use.Therefore, it can be concluded that a highlyinnovative approach will be required fordeveloping a good neutral pH, roomtemperature and chemical aerospace coatingstripper.
Part 2
FLUIDIZED BED ABSORBENTCLEANING
After demonstrating the feasibility of non-solvent substitutes during 1989, the objectivein 1990 was to develop methods for their usethat were suitable for industrial scale-up. Themost promising materials at that time wereabsorbents such as starch and uncalcineddiatomaceous earth. This could lead to theconclusion that very fine particulate matter isbest, but developments in 1989 dealt primarilywith substitution materials usable in the sameapproximate manner as a wipe solvent. Whileit can be generalized that a finer particleprovides more surface area per volume,density and other properties become majorconsiderations with other delivery systems.When fluidized bed technology wasinvestigated as a possible means to scale-upand automate absorbent cleaning, theabsorbent media had to be completely re-evaluated.
While it was theorized that the particles wouldaggregate with oil absorption and gain
57
sufficient bulk and density to drop off the testpanels, in reality the adhesion of the Zyglotest oil to the panel was the dominant force inthe system. Aggregated particle mass neverreached the point where it could exert thetensile and sheer stresses necessary toovercome the oil-panel bond. Thisprecipitated a search for denser, largerabsorbent materials which were alsosewerable. Flours, cornmeal, and finallygrain cereals were tried. Whole grains cerealsproved to possess the right balance of particlesize, absorbency and density. Specifically,Bear Mush, branch whole wheat cereal wasused as the active absorbent in our mostsuccessful trials.
The actual fluidized bed material used was"Envirostrip," a modified wheat starchresembling large sugar crystals. TheEnvirostrip crystals provided the abrasivenessand bulk necessary to remove the oil ladenwheat particles without manual assistance.This is a critical consideration for scale-up andprospective process automation.
The media selection thus dictated theprocedure which evolved to: 1) Immersion ofracked parts or sheet stock in the wheatcereal; 2) Removal of the wheat/oilconglomerate in the fluidized bed turbulentphase. Zyglo penetrant inspection oil wasused for artificial contamination because itcould be observed during bed operation;traditional cleaning tests indicated comparableproperties with the MIL-L-7870 protective oilcommonly used by our sheet stock suppliers.
The lab scale bed diameter was 14 cm with atotal height of 43 cm. 301 CRES screeningwas used as the air spreader (funnel).Oil/wheat removal occurred almostinstantaneously when the bed was turbulent.
Operating parameters were determinedexperimentally, and it was found that the mostefficient cleaning was in a region referred toas "rapid bubbling." Some spouting occurs inthis region. Combined with the rapid rollingof the bed, this spouting results in the
maximum contact of the media and soiledsurface. Figure 23 shows the characterizationof pressure as a function of flow rate for thesystem. Figure 24 shows pressure as afunction of velocity.
Scale-up of the fluidized bed can be calculatedfrom the best fit curves of the plots shown inFigures 25 through 29. These plots are basedon experimental values, with the exception ofA P. AP was calculated using Ergun'sequation. (4) The more commonly usedequation established by Baeyens and Geldart(5) was determined to be inappropriate for thissituation. Their equation does not hold truewhen the bed volume is small, or when thefluid and particle densities are orders ofmagnitude different, both of which were trueof our system.
Figure 25 shows the operating air velocities atvarying bed heights. Using the best fit curvefor this plot, a bed height of one meter willrequire an air velocity of approximately 0.28meters per second for efficient operation.
Scale-up to a bed height of one meter shouldbe achievable using available plant air.Multiple ports may be necessary to achieve theoperating flow rate. This size bed willapproximate the size of many vapordegreasers.
Another non-solvent substitute investigated in1990 was an oil absorbing cloth made by 3Mcalled "Oil Sorbent, Type T-151." This clothis made of 1/4 inch thick coarsely feltedpolypropylene. 3M markets this material forcleaning oil spills, and it is available to ourshops for that purpose.
Our interest in this material was precipitatedby an inquiry to Environmental Affairs fromthe Everett Production Drawing area. Thisarea uses large quantitities of Freon to removefinger prints from the mylar drawing film. T-151 samples were tested in a variety of oilremoving situations including the coated mylarfilm. It performed very well demonstrating ahigh degree of competitiveness for any oil
58
accretion. On the basis of the initial work, itis concluded that these cloths could replace thenaphtha/petroleum distillate preclean nowemployed in a two-step solvent cleaningprocess involving a non-polar precleaningsolvent followed by a polar final cleaningsolvent.
REFERENCES
1. Lenz, Ruben, "Envirostrip, A Non-Petroleum Based Natural Dry BlastMedia Engineered for the AerospaceIndustry," DOD/Industry AdvancedCoatings Removal Conference,Atlanta, GA, May, 1990
2. D204-14436-1, "Impact Study ofPlastic Bead Paint Removal onCorrosion Prevention and Control,"The Boeing Company, BoeingAerospace, Seattle, WA, March, 1987.
3. Panciera, H., "Surface Finish Removalfrom Advanced Composites Prior toRepair and Refinishing, Naval AirRework Facility, Alameda, CA.
4. "Coating Removal via Plastic MediaBlasting, "NAVAIR EngineeringSupport Office, Materials EngineeringDivision, Naval Air Rework Facility,Pensacola, FL.
5. NESO Code 34132, "PreliminaryReport on Plastic Media PaintStripping from GraphiteEpoxy Surfaces, "MaterialsEngineering Lab, March, 1984.
6. 86-E3B2-19, "Impact Study of PlasticBead Paint Removal on CorrosionPrevention and Control, "The BoeingCompany, Boeing Aerospace, Seattle,WA.
9. N00019-83-G-0049 IMP Contract No.,Project WBS-235, "AutomatedPainting and Stripping Project, SixthQuarterly Report," Department of theNavy, prepared by GrummanAerospace Corp. (Confidential).
10. 00-143 Stage 2 PRAM Project,Roberts, R.A., "Mechanical PaintRemoval Process," Interim report onstripping paint from the first F-4Eprototype at Hill AFB, UT, July,1984, and May, 1985.
11. AFWAL-TR-85-4138, Childers,Sidney, et al., "Evaluation of theEffects of a Plastic Bead PaintRemoval Process on Properties ofAircraft Structural Materials,"December, 1985.
12. Bullington, J.B., Williams, D.R.,"Organic Coating Removal viaMultiple Plastic Media Blast Cycles onClad Aluminum Airframe Skins,"Corpus Christi Army Depot ChemicalBranch, Engineering Branch, CorpusChristi, TX.
13. A.21 Process Standard, "Plastic MediaBlast Cleaning and Paint Removal,"Training Guidj for Corpus ChristiArmy Depot.
14. Manufacturing Operating Instruction 8-1951985937, "Plastic Media BlastCleaning for CH-47D ModificationProgram," July, 1985.
no effectno effectno effectno effectminimal or no effectwhen done properly
minima] or no effectwhen done properlyno effectminimal damage
minimal or no effectwhen done properly
Commentspending fatigue data,see figures
Possibly lessdamaging than currentmethod, see figures
fibers exposed, verydelicate operation
Corres.Figure19, 20
13 to 16
17, 18
1, 2
2 to 12
63
Figure 1. Graphite Surface Before Depainting with Envirostrip
Figure 2. Graphite Composite After Depainting with Envirostrip(No Noticeable Fiber Damage or Delamination)
64
Figure 3. Cyanate Ester Panel Before Depainting
Figure 4. Cyanate Ester Panel After Depainting with Plastic Media (Type V;
65
Figure 5. Cyanate Ester Panel After Depainting with Plastic Media (Type V)
Figure 6. Cyanate Ester Depainted Using High Pressure Water at 10-15K psi
66
Figure 7. Cyanate Ester Depainted Using High Pressure Water at 18K psi
Figure 8. Cyanate Ester Stripped to Primer Only with Envirostrip
67
Figure 9. Cyanate Ester Depainted with Envirostrip
Figure 10. Cyanate Ester Panel Top View After Depainting with Envirostrip
68
Figure 11. Cyanate Ester Panel After Depainting with High Pressure Water
Figure 12. Cyanate Ester Panel After Depainting with Type V Plastic Media(Note: All of the fibers are exposed. Plastic mediaremoves the resin.j
69
Figure 13. Fiberglass-Piece of Scrapped AWACS Rotodome Before Depainting
Figure 14. Fiberglass-Piece of Scrapped AWACS Rotodome Before Depainting
70
Figure 15. Fiberglass-AWACS Rotodome After Depainting with Envirostrip
Figure 16. Fiberglass-AWACS Rotodome After Depainting with Envirostrip(Impossible to determine if damage was caused by theEnvirostrip depainting or the previous rework.)
71
Figure 17. Epoxy E-Glass After Depainting with Wheat Starch
Figure 18. Epoxy E-Glass After Depainting with Wheat Starch
72
Figure 19. 2024 Aluminum Panel Before and After Depainting with Envirostrip(This photomicrograph is representative of all thealuminum photos. No damage is visible.)
- . . • « •
• / .
. • : * . . • • ' • • • ? : • • ' * -
~ "' v ' .'
Figure 70. .020 Inch Clad Aluminum Panel (150X) Upper Surface Depainted with Envirostrip(Very limited damage testing was performed on aluminum clad;therefore, quantitative data is not reported.)
73
Figure 21. New Envirostrip Media
Figure 22 Used Envirostrip Media(Note the particles have the same sharpangles as the new media.)
74
400,
0.000 0.001 0.002 0.003 0.004
Row Rate (cu m/sec)
1st BubbleRapid BubbleTurbulentEmpty Bed
Figure 23. Flow Rate vs Pressure of Empty Bed at Bed Height 20 cm
3.
400
300
200
100.
Pressure
Velocity (m/s)
Figure 24. Velocity vs Pressure at 20 cm Bed Height
PRECISION PARTS CLEANINGWITH SUPERCRITICAL CARBON DIOXIDE
Paula M. Gallagher and Val J. KrukonisPhasex Corporation
Lawrence, Massachusetts
INTRODUCTION
Over the past two decades, supercritical fluidshave been utilized as solvents foraccomplishing separations of materials asdiverse as foods, polymers, Pharmaceuticals,petrochemicals, natural products, andexplosives. More recently they have beenused for non-extractive applications, such asrecrystallization, deposition, impregnation, andsurface modification. Today, supercriticalfluid extraction is being practiced in the foodsand beverage industries; there are commercialplants for decaffeinating coffee and tea,extracting beer flavoring agents from hops,and separating essential oils and oleoresinsfrom spices. The use of supercritical fluids,especially carbon dioxide, for cleaning metal,ceramic, or composite parts is an almostnatural outgrowth of these previousdevelopments.
W i t h t h e r e c e n t s c r u t i n y ofchloroflaorocarbons (CFCs) and the newregulations concerning the phaseout of theseand other ozone-depleting chemicals, thesearch for alternative solvents and technologiesis becoming widespread. Phasex Corporationhas been developing a process calledDriClean" in which carbon dioxide is used asa CFC replacement for cleaning intricate partssuch as gyroscopes, laser optics components,accelerometers, nuclear valve seals, andthermal switches. Carbon dioxide has theadvantage of environmental acceptability, isnon-flammable, non-corrosive, and is "workerfriendly." Additionally, carbon dioxide has noozone-depletion potential, and while it doeshave some global warming potential, its use incleaning operations would contributeinsignificantly to global warming in
comparison with, for example, automobileemissions.
This paper will describe the application ofcarbon dioxide to the cleaning of precisionparts, specifically on the removal of organic-based contaminants from these pans; thecontaminants are actually dissolved by thecarbon dioxide and not simply dislodged byflow or abrasion. Included will bebackground information on supercritical fluidsand their behavior as solvents, as well aspreliminary data which demonstrates theeffectiveness of the DriClean" process.
BACKGROUND
The "simple" supercritical fluids, such ascarbon dioxide and the light hydrocarbons, aretypically gases at room temperature andpressure. Above their respective criticalpoints (carbon dioxide, for example, has acritical temperature of 31°C and a criticalpressure of 1072 psi), these fluids can be usedas solvents; even liquid carbon dioxide has theability to dissolve some materials. SCFs havehigh density properties and attractive transportproperties such as gas-like viscosities anddiffusivities, which render these fluids capableof penetrating very small pores and intersticesof complex parts. Perhaps the most uniqueproperty of an SCF is that its "dissolvingpower" is pressure-dependent, such that,simply put, at higher pressures more materialwill dissolve in the SCF than at lowerpressures.
The phenomenon of supercritical fluidsolubility was first reported over 100 yearsago by Hannay and Hogarth. As an exampleof the pressure-dependent dissolving power of
79
a supercritical fluid solvent. Figure 1 showsthe solubility of a simple solid, naphthalene, incarbon dioxide and ethylene. Above thecritical pressure of each gas, it is clear thatrelatively small increases in pressure result inlarge increases in the solubility ofnaphthalene. The directed line between Points1 and 2 shows the very large change insolubility that results when the pressure of asaturated solution at 200 atm is lowered to, forexample, 100 atm. Because of such dissolvingcharacteristics, it is possible to design aprocess to extract, purify, or fractionatematerials based on changes in pressure of asupercritical fluid solvent. At high pressurean extraction (i.e., a dissolution) can becarried out, and by lowering the pressure, aseparation of the dissolved material can bemade to occur. The solute-free solvent canthen be recycled to the extractor. The processcan be carried out in either a batch or acontinuous mode, depending upon the natureof the feed and the nature of the extraction,i.e., whether it be a purification (or topping),fractionation, or extraction from a reactionmass. Most of the work that is being carriedout industrially usually involves an extractionof one material from a mixture. Theoperation of an extraction process will bedescribed.
A schematic diagram of a process which usesa supercritical fluid as a pressure-dependentsolvent to extract an organic substance isgiven in Figure 2a. Four basic elements ofthe process are shown, viz., an extractionvessel, a pressure reduction valve, a separatorfor collecting the material dissolved in theextractor, and a compressor for recompressingand recycling fluid. (Ancillary pumps,valving, facilities for fluid makeup, heatexchange equipment are omitted from thefigure for clarity and ease of presentation.)Figure 2b shows extensive data on thesolubility of naphthalene in carbon dioxide asa function of temperature and pressure.Reference to Figures 2a and 2b is made inexplaining how a supercritical fluid processoperates. Some process operating parametersare indicated on two solubility isobars in
Figure 2b: Point 1 represents conditions in theextractor, e.g., 300 atm, 55°C, and Point 2the conditions which exist in the separator, 90atm, 32CC. The extractor vessel is assumed tobe filled with naphthalene in admixture withsome other material, which for the purpose ofthis discussion is assumed to be insoluble incarbon dioxide. Gas at condition 1 is passedthrough the extraction vessel, wherein itdissolves (and extracts) the naphthalene fromthe insoluble material. Leaving the extractor,the carbon dioxide-naphthalene solution isexpanded to 90 atm through the pressurereduction valve as indicated by the directedpath in Figure 2b. During the pressurereduction step, naphthalene precipitates fromthe solution, because as Figures 1 and 2bshow, the dissolving power of carbon dioxideis low at low pressure. The precipitatednaphthalene is collected in the separator, andthe carbon dioxide leaving the separator isrecompressed and returned to the extractor.This recycle process continues until all thenaphthalene is dissolved and extracted, thedirected line segment 1-2 in Figure 2b and itsreverse on the solubility diagram representingapproximately the cyclic process. (The otherdirected lines, e.g., 1-3, 4-5, etc., designateother extraction/separation paths; they showthat isobaric conditions can also be used toseparate a material and a case-by-caseevaluation will dictate the appropriateoperation of any specific process.)
Many materials including other SCF solublesolids, polymers, and oils will exhibit thegeneral behavior shown in Figures 1 and 2b,and thus the principles involved in theextraction process described by these figuresare directly applicable to precision partscleaning. The next section describes theresults of prior solubility studies on which theeffectiveness of the DriClean" process isbased.
PREVIOUS RESULTS
There has been a considerable amount ofresear :h devoted to the study of the behavior
80
of oils and polymers in various supercriticalfluids (SCFs). Previous studies by PhasexCorporation have demonstrated that SC carbondioxide is an excellent solvent for oils such ash y d r o c a r b o n s , e s t e r s , s i l i c o n e s ,perfluoropolyethers, halocarbon-substitutedtriazines, and organosilicones with variousreactive functionalities; many of these oils areassociated with the manufacture of precisioncomponents such as gyroscopes andaccelerometeij. The ability to dissolve apanicular oil or polymer at any given pressurewill greatly depend on the molecular weightand structure of the material. Theimplications of these characteristics aredramatic, particularly for polymers which areinherently composed of a range of molecularweight chains that give rise to a widepolydispersity. For example, by conducting agradually increasing pressure profile, apolymer can be fractionated into increasinglyhigher molecular weight fractions, eachfraction having a much narrowerpolydispersity than the parent polymer.Likewise, for both synthetic and natural oils,which generally contain chains of varyingmolecular weight, separation based on chainlength has been demonstrated. The workdescribed subsequently was carried out at thePhasex laboratories; much of it was reportednearly six years ago. Since that time, therehas been a significant amount of investigationsinvolving solubility studies of various otheroils and polymers in SCF's. Much of thisprior research was the basis for subsequentgyroscope cleaning work done for DraperLaboratories, Naval Avionics Center, andHoneywell Avionics.
As an example of the separation properties ofsupercritical fluid solvents, a high molecularweight silicone oil (Mw = 90,000) wasfractionated with supercritical carbon dioxide.This oil was analyzed by size exclusion (gelpermeation) chromatography to determine themolecular weights of the fractions. Table Igives the molecular weight values (both Mn
and Mw) of the parent silicone oil and of thesupercritical fluid fractions. Carbon dioxideover a pressure range of 5500 to 1800 psi at a
temperature of 80°C was used to carry out thisfractionation. The data in the table show thatsupercritical fluid fractionation can producecuts with very narrow molecular weight rangeand of course, they show that the silicone wascompletely dissolved.
As is frequently found with synthetic oils, thisparticular silicone (which was a commercial8000 centistoke oil) contained a small fractionof very low molecular weight material whichcould be isolated easily; analogously, the datain the table show that a small fraction of veryhigh molecular weight material of about150,000 (Mw) was also present in the siliconeoil. Figure 3a shows the gel permeationchromatogram (GPC) of the parent siliconeoil; the small peak of low molecular weightsilicone is evident. Figure 3b reproduces theparent GPC and onto which are superposedthe GPC's of the highest and the lowestmolecular weight fractions. Thus, from anexamination of the GPC traces shown inFigure 3b, it is seen that the dissolving powerof carbon dioxide can be adjusted toselectively extract the "lows"; additionally, itis interesting to point out here thatsupercritical carbon dioxide can dissolve asilicone polymer of 150,000 molecular weight,as the data for Fraction 6 in Table I and thesuperposed trace in Figure 3b show.
As another especially pertinent example, afractionation was made on a commercially-available, high molecular weightperfluoroalkyl-polyether oil which is beingused in specialty lubricant and barrier fluidapplications; the oil had previously undergoneprocessing in three sequential moleculardistillation steps to obtain as narrow amolecular weight range and to remove asmany lower molecular weight species aspossible. Using supercritical carbon dioxideat a temperature of 80°C and over adecreasing pressure range of 4000 psi to 1500psi, five fractions ("of an approximately equalweight) were obtained from this parentperfluoroether oil. The vapor pressure of theperfluoroether oil was too low to permitcharacterization of the fractions to be made by
81
gas chromatography, and the viscosity of eachfraction was measured as a means ofindicating molecular weight.
The viscosity-temperature curve of eachfraction and of the parent oil are shown inFigure 4. The room temperature viscosities ofthe fractions range from 800 cps to 9000 cps;the viscosity of the parent oil is 2000 cps.Published data give the molecular weight (Mw)of the perfluoroether at about 5000; if aneight-tenths power relationship betweenviscosity and molecular weight is assumed forthe oil, the average molecular weight of thefractions is calculated to range from about1600 to 13,000. Thus, although the particularcommercial perfluoroether oil has beenmolecularly distilled in three steps, the data inFigure 5 show that supercritical fluidextraction can separate the oil into stillnarrower fractions and that there is still alarge component of low molecular weightfluoroether that molecular distillation cannot"extract."
One final example of the solvent capabilitiesof supercritical fluids is given here. Althoughthe use of supercritical ethylene is described,it has recently been demonstrated for NavalAvionics Center that supercritical CO2 also hasthe ability to dissolvs and fractionatechlorotr ifluoroethy lene.
Chlorotrifluoroethylene (CTpE) oligomer oilsare currently being evaluated as flotation fluidsin several applications. Narrow molecularweight ranges, or single oligomers, aredesirable for these applications. Currently,only preparative GC or HPLC is satisfactoryfor achieving the narrow distributions desired.Supercritical fluids were able to process thischemical species also; the results of afractionation test with CTFE are described.Supercritical ethylene at a temperature of 80°Cover a pressure range of 4800 psi to 1900 psiwas used to fractionate the oligomers ofCTFE. Eight fractions were (arbitrarily)obtained, and GC-MS analysis of the fractionsprovided the information on the separationsachieved. There were primarily four
cligomers in the oil, with molecular weightsranging from about 700 to 1048; based uponthe structure of the molecule and the measuredmolecular weight, it can be calculated that themixture consists of species with six, -seven, -eight and -nine-mers, respectively. Thecomposition of mers in each of the eightfractions obtained from the test and in theparent are given in Table II. (The fractionswere not of equal weight, and the right handcolumn gives the weight contribution of eachfraction.)
The composition data in Table II show thatsubstantial changes in the ratio of mers in afraction can be achieved by supercritical fluidfractionation; in Fractions 3 and 6, forexample, the predominant mer representsabout 90% of the fraction.
The examples detailed above are only a few ofthe many systems that have been studied.The ability of supercritical fluids to dissolvemany types of oils and organic materials,coupled with the ability to penetrate minusculepores and interstices of metal, ceramic, andcomposite parts, suggest that these fluids couldpartially replace CFC. Although SCF's aretailored for the cleaning of intricate parts, theyare by no means recommended as a generalpurpose CFC alternative.
While each potential cleaning application mustbe evaluated on a case-by-case basis, it is ofvalue to point out here that there are somesituations which are not amenable to cleaningwith SC carbon dioxide. In general,contaminants which do not dissolve in carbondioxide cannot be removed by this process;these contaminants include rust, scale, lint ordust, ionic species, metal salts, and m?iiy (butnot all) fluxes. Removal of rosin flux residue,for example, is one of the more commonlyencountered cleaning problems wherereplacements for CFCs are being sought.Rosin is mainly composed of isomers ofabietic acid which, in itself, is soluble incarbon dioxide but which tends to polymerizeduring solder reflowing when heat is applied;the solder process renders the flux residue
82
insoluble. Water soluble fluxes whichtypically contain polyglycols as carrier fluidsmay have the potential for removal by SCcarbon dioxide; parts contaminated with thistype of flux residue are being evaluated.Certain polymers, such as high densitypolyethylene and cross-linked polymers, (e.g.epoxies and phenolics), will not dissolve incarbon dioxide; this fact may be beneficial forthose situations where it is desirable to leavethe polymer untouched during cleaningprocedures, however.
There are many misconceptions (perhaps,more accurately stated, exaggerations) beingcirculated through the precision parts cleaningcommunity that, for example, "very highvelocity" streams of CO2 can shear off theparticulates; this is not true except in somefortuitously placed large adherent particle onthe outside surface of the gyroscope part.(Note CO2 pellet blasting, one trade namebeing "Cold Blast", can remove only line ofsight particulates; pellet blasting cannotremove the oils from interstices or pores.)This abrasive procedure may be suitable forstripping large surface area components,ranging from circuit boards to airplane wings,but it is important to distinguish this processfrom DriClean" in which the contaminants tobe removed are actually dissolved by the CO2,not simply dislodged by flow.
PRELIMINARY CLEANING DATA
Much of the information regarding theeffectiveness of the DriClean" process wasobtained through preliminary feasibility testingfor several private companies that provided theparts to be cleaned. All of the pre- and post-cleaning analyses was conducted by therespective companies. A brief summary of theoperation of the DriClean" process as well assome of the results of the research is presentedin this section.
Process Operation
In concept, precision parts cleaning is exactlyanalogous to the extraction procedure
described in the background section. Parts tobe cleaned are placed into a vessel; carbondioxide, at some pressure and temperature, ispassed over the parts to dissolve the oils; theparts, now free of contaminating oils, areremoved from the vessel. The contaminant-laden stream of CO2 leaving the vessel isdirected to a second chamber in which thepressure is lower'1 and the contaminant oilthat was dissolved in the CO2 drops out ofsolution. If the gas is to recycled, it may bepassed through a bed of activated carbon toremove residual contaminants and thenrecompressed. Typically, a batch-continuousmode of operation would be employed. Sincecarbon dioxide is a gas at one atmosphere,there is no danger of solvent residueremaining on the parts. The dissolving andtransporting properties of carbon dioxidedescribed earlier, plus the integrated industrialexperience in large scale (e.g., 60,000,000lbs/yr of decaffeinated coffee at one plantlocation) operations combine to point out thatprecision parts cleaning can be scaled toessentially any manufacturing level.
RESULTS
As an example of the capability of SC CO2 toreplace CFC-113, a perfluoropolyether(Krytox 143 AD) contaminated berylliumoxide part which typically requires 20 cyclesof CFC-113 rinsing, was cleaned withsupercritical fluid: one pass with CO2 resultedin a degree of cleanliness much better than 20rinses with CFC-113. Figure 6 shows theanalytical results using FTIR. This work wascarried out for Draper Laboratories. In late1990 Phasex carried out tests for the NavalAvionics Center, Indianapolis, Indiana on thedeclassification of two Pulsed IntegratingPendulums (Mod B - Poseidon). Carbondioxide extracted literally all the CTFE asdetermined by visual inspection. Theinstrumental analytical data was not availablefor this publication.
A series of cleaning tests was conducted for a
83
communications company; the samples wereceramic substrates contaminated withfingerprints. It was anticipated that the oilswould be removed but that the ionics wouldnot; the results confirmed these expectations.The samples were intentionally contaminatedwith a solution of fatty acids and salts thatwould approximate the composition offingerprint residue. Prior to cleaning, surfaceanalysis using XPS (x-ray photoelectronspectroscopy) was conducted. The low levelsof ionic species on the original sample asindicated in Table 3 are strictly an artifact ofthe analytical technique; XPS samples only thetop ~30A and the ionics were essentially"hidden" beneath the fatty acid layer. Table3 gives the results of the cleaning tests incomparison to using Freon TA (a mixture ofCFC-113 and methanol) and peroxide.Supercritical CO2 was able to clean the partsto a level comparable to that using Freon TA.
information necessary to determine thepotential for success. Issues such as ozonedepletion and global warming potentials,flammability and toxicity, safety and health,process compatibility and ease of installabilityneed to be addressed, as well as economicconsiderations, such as operating and capitalcosts, throughput, process flexibility, floorspace r e q u i r e m e n t s , and was tetreatment/disposal/recycle/reclamation issues.However, for the particular case ofgyroscopes, accelerometers, and other intricateparts that are contaminated with the oilsenumerated above, carbon dioxide presentsitself, at least at this period in time, as theonly potential solvent to replace CFCs for theproblem of rf moving every oil that isassociated with machining, assembly, filling,or declassification of precsion parts.
SUMMARY
It has been demonstrated that the DriClean"process is indeed an effective replacement forCFCs in the cleaning of precision parts. Theprocess can be tailored for many types ofcleaning operations by modifying processconditions, and it can be employed as acomplementary second-stage cleaningprocedure. Although a case-by-caseevaluation of each potential application mustbe conducted, there exists an extensiveknowledge base from which to derive the
84
Fraction
Parent123456
Table I. Molecular Weight of Silicone Uii fractions
Molecular Weight WT % of Parent
100%4.05.1
27.727.928.37.0
42,500428
3,31027,10043,00058,900
112.600
90,000789
11,50053,20057,50091,500
149,900
Table II. Composition of Chlorotrifluoroethylene Fractions
Fraction
Parent12345678
6
4.551.124.47.61.80.7
——
7
49.537.673.586.475.338.75.70.63.0
Composition,
8
39.512.32.06.0
22.958.194.166.455.5
9
6.5————2.50.1
33.041.5
Wt%of Parent
100%5.88.6
20.215.222.815.610.51.3
Table III. XPS Data
Sample
Contaminated
Freon TA
Peroxide
supercritical-CO2
C
88
45
67
43
10
30
21
35
Ai
0.2
14
9
16
Na
0.2
4
0.5
2
Cl
-
3
-
1
F
0.2
0.7
-
-
Ta
-
0.1
-
1.5
Cu
-
0.4
-
1
Si
-
0.3
-
-
S
1.3
I
-
N
-
0.2
1
0.5
Sn
-
.2
-
-
85
10 15 20
Pressure, MPa
25 30
Figure 1. Solubility of Naphthalene. A, in Ethylene at 35°C, and B, CO2 at 45°C. 1 and 2 areExplained in the Text (To Convert MPa to Bar, Multiply by 10).
Figure 3b. GPC of Parent Oil and High and Low Molecular Weight Fractions.
87
10,000
1025 50 75 95
TEMPERATURE AT WHICH VISCOSITYWAS MEASURED C O
Figure 4. Viscosity of Perfluoroether and of Fractions Extracted with Supercritical CC>2-
88
Figure 5a.
U1O 20O0TIME SCAN
GC-MS Trace of Parent Chlorotrifluoroethylene(Nurr1- ;rs by Peaks Indicate MERS).
I JLAU'SO 2000
TIMC SCANll>40
Figure 5b. GC-MS Trace of Fraction 3 of Chlorofluoroethylene(Numbers by Peaks Indicate MERS).
Figure 5c.
UJO iffOOTtMC SCAN
GC-MS Trace of Fraction 6 of Chlorofluoroethylene(Numbers by Peaks Indicate MERS).
89
0.02 i
0.01 -
0.00
0.02
0.01
0.00 -
1350.0 1300.0 1200.0
20 FREON U\S
0.02
0.01 -
0.00-
1350.0 1300.0 1200.0 1350.0 1300.0 1200.0
1 PASS 2 PASSES
Figure 6. Krytox Removal Using (a) Freon-113, (b and c) SC CO2.
90
CARBON DIOXIDE PELLET BLASTING PAINT REMOVALFOR POTENTIAL APPLICATION ON
WARNER ROBINS MANAGED AIR FORCE AIRCRAFT
Randall B. IveyAir Force Corrosion Program Office
Robins Air Force Base, Georgia
INTRODUCTION
WR-ALC has been striving toward thereduction or elimination of environmentallyunacceptable waste produced from aircraftrepair processes. One of the largest sourcesof hazardous waste comes from the aircraftpaint removal operations. In this operation,the aircraft will go through several processes,all of which produce some form of wasteproduct. This paper will discusi theinvestigation of the carbon dioxide pelletblasting (CDPB) process as a potentialenvironmentally compliant replacement forcurrent coatings removal technologies.
CHEMICAL PAINT STRIPPINGDEFINED
Prior to chemical paint stripping, the aircraftis washed to remove oil, dirt, and grimewhich may prevent the chemical strippersfrom working properly. The wash and rinsewater requires industrial sewage treatmentbecause of the soils and, in many cases, thecontents of the soap. The chemical strippersare then applied to the aircraft.
Chemical strippers have several ingredientswhich make them hazardous and dangerous touse and discard. Chemical strippers typicallycontain up to 50 percent methylene chloridesolvent, which the Occupational Safety andHealth Administration (OSHA) has listed as aknown carcinogen. OSHA has limited workerexposure to levels that are not achievable in anaircraft stripping environment. In addition,methylene chloride may soon be classified asa volatile organic compound (VOC), furtherrestricting its use. Chemical paint removers
may also contain phenolic compounds inconcentrations up to 20 percent, aconcentration which cannot be treatedeffectively in industrial sewage treatmentplants. Therefore, the phenolic sludge must becollected, dried, stored, and discarded as a"hazardous waste."
Chemical strippers are applied to the aircraftand scraped off. Because the paints on U.S.Air Force aircraft are chemically resistant,they are difficult to remove with this process.Therefore, the application and scraping of thestripper may be repeated up to nine times untilall of the paint is removed. The aircraft isthen washed and rinsed, and any paintremained is removed mechanically. On theF-15 aircraft, this process yieldsapproximately 6,800 pounds of hazardouswaste, plus rinse water.
PLASTIC MEDIA BLASTING DEFINED
Plastic media blasting (PMB) has beenauthorized as an alternative tc chemical paintremoval operations. PMB is similar tc sandblasting in operation. The aircraft is firstwashed, as with the chemical depaint process,causing a waste stream then particles of plastic(acrylic, melamine, or polyester) are propelledat the paint using compressed air. Theprocess generates from 1,500 to 3,000 poundsof dry hazardous waste per F-15 aircraft.This waste is made up of plastic media whichhas been abraded to particles too small to berecycled and the paint removed from theaircraft. It is the lead, cadmium, andchromium from the paint chips which causethe end product to be classified as a hazardouswaste.
91
Plastic media also tends to ingress the aircraftthrough seals and seams in the aircraft skins.Man-hour intensive masking operations areused to reduce this ingress to a minimum.After the PMB process, media removal mayalso be required.
CARBON DIOXIDE PELLETBLASTING - A POSSIBLE SOLUTION
As a possible solution to these problems,Warner Robins Air Logistics Center(WR-ALC) has been investigating the use ofCDPB as a potential new aircraft coatingsremoval process. Carbon dioxide pelletblasting is similar to PMB and sand blasting,in that small particles are accelerated towarda painted surface using compressed air.However, the first two processes relyprimarily on the abrasive action of theparticles to remove the old paint; the newpellet blasting process uses carbon dioxide inits solid form-namely, dry ice. Carbondioxide pellets can be blasted at the coating atsubsonic, sonic, or supersonic speeds,depending on the particular application. Instripping paint, these particles provide notonly an abrasive action, but also a thermalshock from the -109 degree Fahrenheittemperature of the particles. After the carbondioxide pellets strike the paint, they simplysublime (evaporate directly from a solid to avapor without a liquid phase), leaving only theremoved paint as residue. Because thestripping action is not degraded by surfacegrime and the media is not recycled, there isno requirement to wash the aircraft prior tousing this process.
A demonstration of this process in April, 1990illustrated that this process had matured to thepoint that it could safely be used to stripcoatings on certain parts of the aircraft. Itwas also established that this process couldreduce the hazardous waste produced fromeach F-15 aircraft to 240 pounds, as comparedto the other stripping methods listed above.This hazardous waste consists only of the paint
removed from the aircraft.
Because the carbon dioxide pellets sublime,there is no intrusion problem. Any mediawhich intrudes the aircraft simply sublimesand does not have to be removed. Masking ofthe aircraft is reduced to only what is requiredto protect certain delicate materials such asaircraft canopies.
CARBON DIOXIDE TEST PROGRAMSAND PRELIMINARY RESULTS
There are several WR-ALC sponsored testprograms under way to evaluate the CDPBprocess. The major aircraft stripping effort isbeing undertaken by the F-15 ProductDirectorate under an engineering servicescontract to Mercer Engineering ResearchCenter (MERC) and has, so far, prototypestripped one F-15 aircraft. This first aircraftdemonstrated that the process could beapplied, with limited success, to the F-15aircraft. However, there were several areas ofconcern identified that require resolution priorto production implementation.
The first concern is the slow strip rate of theprocess. Mercer Engineering Research Centerhas indicated, in its preliminary Phase I TestResults (11 Jan 91), that instantaneous striprates varied from 1.0 square-foot-per-minuteto 0.1 square-feet-per-minute, depending onthe substrate that was stripped. The netaverage strip rate on the F-15 aircraft wasreported to be approximately .189 square feetper minute of nozzle on-time (.13 square feetper minute with worker effectiveness factoredin). This net average strip rate is marginal atbest. The slowest strip rates were experiencedon Alclad surfaces. While only 20 percent ofthe F-15 has Alclad surfaces, other Air Forceaircraft have up to 80 percent Alclad surfaces.Strip rates of aircraft with a larger percentageof Alclad will be extremely slow.
The second problem concerns the surfacecondition of Alclad and other aircraft surfaces.The MERC Phase I Test Results Report
92
indicates that the current process does notremove all of the paint from the Alcladsurfaces. The surface that is left is not readilyrepaintable. The residue paint must beremoved by other processes in order toprovide an adequate surface for repainting.This is obviously a potential problem for anyaircraft with Alclad skins.
The third problem concerns the ergonomicsinvolved with using the process. The weightand thrust of the blast nozzles and hoses hasbeen identified from the very first test as ahuman factor fatigue problem. The hose andnozzle weigh approximately 20 pounds whenheld at chest level. If blasting underneath anaircraft surface, the blast process thrust addsapproximately 10 pounds to this weight.MERC has indicated in the preliminary PhaseI Test Results that the ergonomic difficulties"reduce stripping effectiveness (speed) by 150percent" for the F-15 aircraft. Workers in theF-15 demonstration manual CO2 booth (Bldg137) currently switch the blasting duty everyfifteen minutes. This requires two personnelfor each bias* nozzle. MERC recommends theuse of some type of assistive device to reduceoperator fatigue and thus increase strippingspeed and quality.
The fourth concern is that the process, ascurrently defined, is too aggressive for use onthin skin aluminum alloys which areunsupported. Aluminum .032 inches thick andthinner are susceptible to peening-type damageat the pressures required for effective paintstripping. This is only a minor problemfor the F-15 aircraft, but could prevent the useof CDPB on up to 20 percent of cargo aircraftskins.
In order to resolve the above-listed concernsusing CDPB, several projects are currentlyunderway. The problems with the ergonomicswill be resolved by a robotic system which hasbeen installed and should be operational byAug 91. This robotic system will, at least,double the current strip rate. The precision ofthe robot, may provide an acceptable surfacecondition (especially on Alclad) as compared
to what is seen with the manual CDPBprocess. Other assistive devices, such asmanipulator arms, could also be utilizedinstead of robotics.
Also under development is combined CDPBand flashlamp stripping head, which mayboost s t r ipping speed to threesquare-feet-per-minute. The combined systemwill also significantly reduce theaggressiveness of the process, allowing eventhin skin materials to be stripped withoutdamage. This project will be completed by 15May 92, with the demonstration of the systemon an F-15 aircraft at WR-ALC.
A WR-ALC and MERC test effort will look atthe positive and negative attributes of usingpaint softeners prior to using the CDPBsystem. Paint softeners will allow the use ofmuch less aggressive parameters allowing theuse of CDPB on even thin-skin materials.Stripping speed should also increase to wellover one square-foot-per-minute. The possibledrawbacks of using chemical softeners includean increase in disposable waste, additionalaircraft processing steps, and the potentialeffects of the softeners on aircraft materials.
In order for CDPB to be widely accepted asan aircraft coatings removal process, it mustbe capable of adequately removing the entirecoating system from aircraft materials withoutdamaging substrate materials. In order forCDPB to be cost effective as a coatingsremoval cnrl, as compared to other paintremoval processes, it must remove coatings arate of .25 to .70 square-feet-per-minute(depending on the aircraft).
CONCLUSIONS
The intent is for the F-15 aircraft to achievean acceptable strip rate and surface conditionIncreases in using the robotic system currently
93
undergoing installation at WR-ALC. Alcladsurfaces may still require some alternativepaint removal prior to repaint. Advances instrip rate and Alclad surface condition throughthe use of combined flashlamp and CDPBsystems, or the use of paint softeners, willmake this an even more cost-effective system,strip rate, the ability to strip thin skins withoutdamage, and complete paint removal are allrequired prior to the acceptance of this processon the WR-ALC managed cargo aircraft.
94
ALTERNATIVE TECHNOLOGIESFOR
ENVIRONMENTAL COMPLIANCE
J. Michael LocklinDouglas Aircraft Company
Long Beach, California
ABSTRACT
Many Americans have hailed the 1990's as thedecade of the environment. While someconsider this recognition long overdue, otherssee it as a narrow, self-serving scheme tocreate jobs for environmentalists that willultimateiy degrade the worldwidemanufacturing competitiveness of the UnitedStates. For many years now, increasingpublic awareness has led to more and morespecific environmental regulations throughoutthe country, but especially in SouthernCalifornia. These regulations have had, andwill have, a serious impact on productivity inAmerican industry. Whether this impact willbe advantageous will depend upon the effortsof environmental professionals who researchand implement alternative technologies. Thedirection we take with these new processeswill contribute greatly to the long-termmanufacturing capability of the United States.
Over the past couple of decades we have seenseveral foreign nations displace the UnitedStates as world leaders in a variety of productlines. For example, Japan's impact on theautomobile industry is well-documenied, asare their efforts in the computer andelectronics fields. Korea is rapidly becominga world leader in the manufacture of steel.And the soon to be declared EuropeanEconomic Community is already wellrepresented in the transport aircraft communityby the Airbus consortium; not to mentionother European aerospace leaders such asBritish Aerospace (Great Britain), Dassaultand Arianespace (French), DeutscheAerospace, MBB and Dornier (German),Aeritalia and Aerospatiale (Italy), CAS A
(Spain), and Fokker (Netherlands).
The competition is serious. However, all isnot lost. A prompt and thoughtful resolutionof today's environmental concerns mayprovide the opportunity needed to re-establishthe world manufacturing leadership that theUnited States has enjoyed in the past. Whileforeign nations have been directing effortstoward increased manufacturing capability,environmental concerns often have taken aback seat to the lure of immediate profits anda higher consumer standard of living. Butthere is a strong movement in the UnitedStates to clean up the environment. We nowhave the opportunity to establish the U.S. as aleader in both environmental andmanufacturing technologies.
DOUGLAS AIRCRAFT COMPANYPROGRAMS
At the Douglas Aircraft Company (DAC) inLong Beach, California, a division of theMcDonnell Douglas Corporation, we areworking to continuously improve ourmanufacturing technologies whilesimultaneously reducing the impact of thesetechnologies on the local and globalenvironment. The following alternativet e c h n o l o g i e s a r e b e i n ginvestigated/implemented at theAircraft Company:
Douglas
1. High solids topcoats - in lieu ofconventional topcoats and/or exempt solventtopcoats.2. Chromium elimination - in paints,primers, sealants, process chemicals, etc.
95
3. Alkaline/aqueous degreasing technologiesin lieu of solvent vapor degreasing.4. Alternative Handwipe Solvents/cleaners.5. CFC Elimination.6. Resource Recovery/Waste Minimization.
Each of these projects contributes to theultimate goal of eliminating the negativeimpact of manufacturing processes upon theenvironment. Perhaps the future will providemethods that are even beneficia'. to theenvironment. Meanwhile, achieving this goalwill require considerable training and"retooling" of american industry. At DAC weare working diligently with our suppliers andsubcontractors to develop, test and implementnew alternative technoiogies.
High solids topcoats is an alternativetechnology to painting aircraft withconventional topcoats which contain highlevels of solvents for sprayability and drying.While it was recently acceptable to simplysubstitute exempt solvents, such as 1,1,1trichloroethane, in order to comply withregulations, it is now apparent that exemptsolvents also have damaging environmentalproperties and will eventually be banned bythe governing agencies. The MontrealProtocol set guidelines for the elimination ofmany solvents including exempt solvents.Locally, the Southern California Air QualityManagement District Rule 1124 requirestopcoat paints to contain fewer than 420 g/1 ofVOC as of July 1, 1991.
The term "high solids" implies lesssolvent/more solids, i.e., resins and heavymetals. The application of the new high solidspaints will require some training andfamiliarity since they will have differentcharacteristics. Tests have indicated that theygenerally require approximately 10% more drytime than conventional topcoats, althoughexperimentation with accelerators continues.The pot life of the new paints is shorter byone half. High solids primers are displayingapproximately a 200% longer dry time thantheir conventional counterparts.
As of April, 1991 we have painted three (3)MD-11 customer aircraft with high-solidstopcoats. Since these coatings are typicallymore glossy than the old topcoats, the needfor a clear coat is eliminated. Both thecustomer and - The Air Quality ManagementDistrict (AQMD) are pleased.
Chromium elimination covers a variety ofprocesses, such as painting, sealing, plating,and chemical processing. Chromium has longbeen a main ingredient of many airframeprocesses because of it's excellent corrosionand wear resistant properties. But, sincehaving been identified as a human carcinogen,efforts are now underway to reduce its' usageto the point of eventual complete elimination.
In 1990 DAC expanded the use of a thin filmsulfuric acid anodize process to include somecommercial work, thus reducing the use of thepopular chromic acid anodize. Somejustification for this change was provided bymilitary contracts which, in fact, specify thesulfuric process. A dilute chromic acid seal iscurrently being used with this process. Weare testing other seals to eventually replaceeven this usage of chromium but, as yet, havenot approved an acceptable alternate. Ourbest results to date have been with organicseals.
Chrome-free aircraft sealants are also beinginvestigated by DAC personnel. Workingwith our sealant suppliers, who are also awareof evolving environmental regulations, we aretesting newly developed chrome-free and lead-free formulations for compliance andefficiency. Although none has yet beenapproved, several show promise, including amanganese cure system which still requiressome work.
Non-chromated deoxidizers are also beingresearched. Airframe manufacturerscommonly use deoxidizers for cleaning,brightening, and removing corrosion fromaluminum prior to subsequent processing. Forthis reason, chromium is a common ingredientof aluminum deoxidizers. Finding an
96
acceptable substitute that will provide the sameprotection and a safe, durable airframe butwill not negatively impact the worker or theenvironment is no easy task. Again, oursuppliers are working diligently to developalternate deoxidizers while our laboratories aretesting the new formulations to aircraftspecifications. We hope to have an acceptablenon-chromated deoxidizer approved and inplace by the end of this year (1991).
Alternative Plating Technologies are alsobeing investigated under the auspices ofchromium elimination since chrome platinghas always been one of the industry's favoritemetal plating treatments due to its'aforementioned wear, and corrosion resistantproperties. Three potential alternatives nowavailable are:
1. Detonation Gun Coatings. This thermalplating technology, while proprietary to UnionCarbide Corporation, provides excellent wearand corrosion resistant properties but can onlybe accomplished by Union Carbide. Since itadheres to titanium better than chrome plating(which tends to flake off), this process hadalready been approved for use on certaintitanium airframe components prior to theconcerns about chromium. It is beingconsidered for additional airframe applicationsas a replacement for chromium processes.
2. Plasma Spray. This non-proprietarythermal plating process is approved for use incertain applications. Since it is a chrome-freeprocess, it is also being considered foradditional applications.
3. Electroless Nickel. This process usesdifferent percentages of nickel and phosphorusto achieve the desired hardness and corrosionresistance. It has been approved at DAC foruse on certain areas of the landing gear. Weare still evaluating even harder formulationsthat include boron and thalium which mayprovide additional applications in the future.
Vapor Degreasing is a cleaning process thatuses solvent vapors alone to effectively
remove a variety of contaminants from theworkpiece. It is a relatively simple, one stepprocess that provides a clean, dry part readyfor subsequent processing. However, thispopular process has now been identified as amajor contributor to ozone depletion. DACpresently uses 1,1,1 trichloroethane (TCA) asthe solvent of choice for vapor degreasing.Some areas of the country usetrichloroethylene (TCE), but TCE has beenprohibited in California. Tests are presentlybeing conducted on various immersion typecleaners to replace solvent vapor degreasing.Some of the candidates are aqueous cleaners,terpene based cleaners, and the use ofultrasonic technology with immersion cleaners.
Aqueous cleaners are typically alkaline innature, their pH being in the range of 9-11.Many chemical suppliers already providealkaline cleaners on the open market. In fact,alkaline cleaning is presently approved forcertain applications at Douglas Aircraft.Alkaline/aqueous cleaning is presently theleading contender to replace vapor degreasing,but the implementation of this will requiresome change in process and equipment whichwill, in turn, require some operator trainingand/or familiarity.
As mentioned above, the vapor degreasingprocess provides a clean and dry workpiecedirectly from the vapor degreaser equipment.This will not be the case with the newimmersion cleaners. Since the new cleanerswill not evaporate as quickly as the solventswith which we are familiar, the workpiece willcome out of the cleaning tank wet. And, sincethese new cleaners may also leave a residuethat could impact subsequent processing or theworkpiece itself, a rinse cycle is likely tobecome necessary with the new process. Ofcourse, the part will come out of the rinsecycle in a wet condition also. Therefore, itmay be necessary to add a dry cycle to theprocess in order to prevent corrosive action.
Lab testing is presently in process at DAC ona variety of candidate cleaners, most of them
97
alkaline in nature. Several companies providea terpene-based metal cleaner, which containsthe active ingredient d-Limonene. Althoughthese cleaners have a low vapor pressure, theyare high in volatile organic compounds(VOCs). A major defense contractor has aisoreported that terpenes do not respond well topurification via ultrafiltration. The use of newimmersion cleaners will likely require somemethod of removing the contaminants gatheredin the cleaning solution or it will rapidlybecome less effective. There are severalmethods, and many commercial filtration unitsare available. The best one for a specificapplication will depend upon the choice ofwhich cleaner you choose and equipment.
Handwipe solvents are used extensively forclean up and repair during the manufactureand assembly of transport aircraft. Not unlikethe solvents used for vapor degreasing, thesesolvents may be ozone depletors and/orcarcinogens. Again, we are working with oursuppliers to develop cleaners that will workeffectively at ambient temperatures to removethe common aircraft industry contaminants.One of the most difficult to removecontaminants is a semi-cured polysulfide basedsealant used extensively throughout theaircraft. This work has typically requiredexhorbitant solvent usage, much to thedetriment of the ozone layer! Preliminarytesting at ambient temperatures has indicatedthat the aforementioned terpene based cleanersseem more effective against this contaminantthan the aqueous cleaners. Additional testingis required.
CFC elimination. Chlorofluorocarbons(CFCs) are used as cleaners of smallelectronic parts such as printed circuit boards(PCBs), in wire assembly areas, and in manymaintenance tasks. They are also used in airconditioners and machine tool chillers, and aspropellants in aerosol can applications. Oneof our immediate substitution efforts at DACis directed at replacing CFCs as aerosolpropel Iants. Again, our suppliers arecooperating as they are aware of the need forchange in their products. It is important to
evaluate the impact of any changes to productsthat are governed by Mil Specs.
There are several lubricants and mold releasecompounds at DAC thaf are comprised ofCFCs. These are not used in high volume,but do provide an opportunity for reduction ofCFC emissions. As yet, we have not beenable to concentrate on replacing thesecompounds, but have begun the process ofsubstituting the propellants used in some of theaerosol lubricants (see above).
Resource recovery/waste minimization isanother broad category that encompasses manytechnologies: chemical processing, wastedisposal, recycling and housekeeping are justa few. There are many opportunities forimprovement under this category asenvironmental technology continues toadvance.
Waste minimization was highlighted duringDAC's recent implementation of the TotalQuality Management philosophy. This effortenlightened and encouraged every employee toconsider his or her impact on the environmentand the workplace. The impact of ForeignObject Damage (FOD) was emphasized duringthis introduction since it is a familiar topic tothose in the aircraft industry. This type ofdamage can be attributed to NOT minimizingwaste.
As an example of resource recovery, DAC'srecently completed chemical processing facilityincludes a controlled room for applying thechemical milling maskant which containsperchloroethylene, a toxic air contaminant.Through state-of-the-art controls, we are ableto capture some of the perchloroethylene andsell it back to our maskant supplier. Anothercase of recycling technology is found in thesame new processing facility, where a CausticEtchant Regeneration (CER) unit purifiesspent chemical milling solution byhydromechanically removing aluminum insolution and collecting it as aluminumhydroxide. This, in turn, can be used as rawmaterial for various manufacturing
98
technologies. and computers when not in use.
An extremely cost effective example of wasteminimization was recently accomplished bysimply reducing the size of ourvendor-provided wipe rags. An on-site surveyconducted to evaluate the usage of wipe ragsdiscovered that the three foot square rags weretoo large for convenient wipe operations. Thesupplier agreed to provide smaller rags at noadditional contract cost to DAC, thus reducingboth the volume and weight of rag-generatedwastes. Because of the wide variety of usesfor wipe rags, they are liable to becomecontaminated with many products includinghazardous substances that require the disposalof these rags as hazardous waste. Weanticipate reducing rag waste volume by 50%,thus realizing a cost savings in the hundreds ofthousands of dollars.
DAC has contracted with an outside recyclingcompany to continuously recycle on-site suchthings as machine tool coolant and hydraulicfluids. The effective life of these solutions isthus extended, further reducing theprocurement and disposal costs of additional,potentially hazardous, fluids and chemicals.
The Douglas Aircraft Company is located insouthern California and is surrounded byresidential neighborhoods. For this reason,DAC has always tried to be careful tominimize its wastes. DAC has hadwaste-water treatment systems operational formany years. One system uses sulfur dioxideto first convert hexavalent chromium totrivalent chromium. It is then neutralized andprecipitated out as metal hydroxides. Afterfurther clarification and filtration, it iseventually discharged to the public sewersystem. Nevertheless, we continue to trainemployees to use proper operating proceduresand encourage measures to further reduce thegeneration of wastes. Some of these measureshave been mentioned above, such aselimination of chromated process solutions andrecycling. Other examples that everyone isfamiliar with are: improvement of generalhousekeep'r.g practices and turning off lights
SUMMARY
DAC continues to investigate and implementnew environmental technologies with thecoordinated efforts of the different programsand core groups. McDonnell Douglasemployees, like so many other Americanstoday, are becoming aware of the manyadvantages of an aggressive environmentalcompliance/technology program. A primeexample of the growth of environmentalconcerns is the fact that ten years ago disposalcosts were seldom considered when justifyinga new expenditure (whether for installing anew machine tool or purchasing processchemicals). But today, responsible planningfor the handling of tomorrow's wastes can notonly improve our environment, but can proveto be very cost effective in the long term.With all the attention that environmentalconcerns are getting today, the attending rulesand regulations can sometimes be veryconfusing. While the EnvironmentalProtection Agency oversees national policies,its major concerns may not address certaincritical concerns at the local level. Balancingthe requirements of both agencies may requirea change to a governing specification. As newtechnologies are developed, specifications arebeing updated. But the industry is growingrapidly and different agencies have authorityover different aspects of environmentalconcerns. Local, national and internationalagencies assume responsibility forenvironmental legislation. Therefore, anychange in process that may effect theenvironment requires a thorough investigationof the governing rules and regulations.
In conclusion, alternative technologies arebecoming increasingly necessary to meet theever-tightening demands of an aware publicwhen it comes to environmental legislation. Itis noteworthy that environmental professionalstend to share technological developments andbreakthroughs in spite of the fact that theymay work for competing companies. This
99
indicates the importance of making our planeta better place to live. Although it sometimesrequires a delicate balancing act to satisfy thevarious applicable agencies, regulations, andspecifications, we now have the opportunity tolead the countries of the world to a better,cleaner, healthier future.
100
HIGH PRESSURE SUPERCRITICAL CARBON DIOXIDEEFFICIENCY IN REMOVING HYDROCARBON MACHINE COOLANTS
FROM METAL COUPONS AND COMPONENTS PARTS
Robert F. SalernoOrganic Material/Surface ModificationEG&G Mound Applied Technologies
Miamisburg, Ohio
High pressure, supercritical carbon dioxideefficiency in removing hydrocarbon machinecoolants (production process contaminants)from metal coupons and component parts wasevaluated. Solubility experiments wereperformed on Cimperial 1011, Cimperial 15,Gulfcut 11D machine coolants. Extractionexperiments were conducted on machinecoolant contaminated aluminum and 303stainless steel coupons (1.5 in. x 0.25 in.), aswell as detonator production components.The solubility/fractionation experiments wereconducted in a screening supercritical carbondioxide system. The solubilities of Cimperial1011 and Cimperial 15 were measured at50°C, 13.8 Mpa. The solubilities ofCimperial 1011 and Gulfcut 11D were alsomeasured at 35°C, 13.8 Mpa. In addition,coolant fractions were collected for gaschromatography analysis. Extraction(cleaning) experiments were conducted in asupercritical carbon dioxide feasibilitysystem, utilizing a 300 ml process vessel.Cleaning trials were conducted at 35°C, 13.8Mpa with carbon dioxide contact times of15-30 minutes. Residual machine coolantconcentrations on coupons and componentscleaned with high pressure, supercriticalcarbon dioxide were determined by a hexanerinse/capillary gas chromatography analysisprocedure.
The studies revealed that the three machinecoolants are all soluble in supercritical carbondioxide with solubilities ranging from 1.06weight percent (wt %) to 4.69 wt %. Also,cleaning experiments conducted showedvariations in the amount of residual machinecoolants on coupons and componentsregardless of carbon dioxide contact time.
Residual contaminants ranged from 3.0 to 840mg.
Under the stated experimental conditions thisstudy has demonstrated that high pressuresupercritical carbon dioxide shows potential asa cleaning media for removing hydrocarbonmachine coolants from metal substrates. Ifoptimized in production cleaning applicationsthe use of such a process would reduce plantwaste streams significantly.
INTRODUCTION
Until recently, production cleaning processeshave relied on halogenated solvents ascleaning media for removal of productionprocess contaminants. However, governmentregulations concerning the use and disposal ofthese products are becoming more and morerestrictive. As a result, considerable interesthas been generated in developing cleaningprocesses that use environmentally acceptablecleaning agents and that reduce or eliminatehazardous waste.
As part of the Department of Energy (DOE)waste minimization efforts, EG&G MoundApplied Technologies (MAT) has beenworking to develop a final cleaning processfor production parts using supercritical carbondioxide as a substitute for halogenatedcleaning agents. The objective of this studywas to evaluate the efficacy of high pressure,supercritical carbon dioxide in removinghydrocarbon machine coolants (productionprocess contaminants) from stainless steel andaluminum coupons and component parts(Inconel-glass ceramic).
101
SCOPE OF EXPERIMENTATION
Four solubility trials and eleven extractionexperiments were conducted. Thesolubility/fractionation experiments werecarried out in a small screening system. Fourtrials on the three coolants were completed.Solubilities of Cimperial 1011 and Cimperial15 were measured at 50°C, 13.8 Mpa. Thesolubilities of Cimperial 1011 and Gulfcut11D were also measured at 35°C and 13.8Mpa. In addition to determining solubilities,various fractions of the coolants werecollected for subsequent analysis.
Cleaning experiments were conducted in aFeasibility System, utilizing its 300 ml processvessel. Cleaning trials were conducted at atemperature of 35°C, a pressure of 13.8Mpa, and contact times of 15-30 minutes.The objective of the program was todemonstrate that supercritical carbon dioxidecould remove contaminants to a residua' levelon the order of 1-10 mg/cm2. Determinationof the effects of operating pressure andtemperature on the solubility of the threemachine coolants was also sought. Residualoil concentrations on coupons andcomponents cleaned with supercritical carbondioxide were determined by a hexanerinse/capillary gas chromatography analysisprocedure.
PROCESS EQUIPMENT
Screening Unit
The solubility/fractionation experiments wereconducted in a Screening Unit, which consistsof a 300 mL extractor, one 70 mL separator,a solvent pump, heat exchangers, and a backpressure regulator. A diagram of thescreening unit is shown in Figure 1.
A screening unit operates as follows: Liquidcarbon dioxide from a dip-tube storagecylinder is subcooled using a glycol-cooledheat exchanger to prevent vaporization in the
solvent pump. For carbon dioxide, cylinderpressure is the saturation pressure at ambienttemperature, about 6.2 Mpa. The subcooledsolvent is compressed from cylinder pressureto the extraction pressure, 13.8 Mpa for thisseries, using a reciprocating, packed-plungerpump.
An electric heat exchanger raises the highpressure solvent to the extractiontemperature, which was maintained in therange of 35 °C - 50°C for this series. Thesolvent, now at supercritical conditions, flowscontinuously upward through a sample ofmachine coolant in the extractor vessel.Electric band heaters under temperaturecontrol maintain the extractor at the desiredtemperature.
Supercritical solvent, containing a solublefraction of the contaminants or test materialfrom the feed, leaves the extractor and flowsthrough a back pressure regulator, which iselectrically heated. Flow through this valvereduces the stream's pressure to atmosphericso that the fluid entering the separation vesselis now a gas and no longer has good solventproperties. Electric heaters on the separatorvessel are used to control the separationtemperature. The material previously dissolvedin the solvent precipitates in the separator asa solid or liquid that can be easily removedfrom the system.
For the solubility/fractionation experiments,the solvent flow was stopped, the separatorremoved, and the soluble fraction collectedafter each approximately 600 g of carbondioxide passed through the system. In thisway, multiple fractions (3-4), which could beaveraged to determine an approximatesolubility, were collected. In addition, thefractions could be physically examined andanalyzed to observe their different physicalcharacteristics.
Feasibility Unit
The cleaning experiments were conducted in afeasibility unit. A diagram of the unit is
102
shown in Figure 2. The feasibility unitconsists of a 300 mL extractor, one or two 70mL separators, and associated pumps, heatexchangers, and pressure control valves. It isused to conduct batch extractions of smallsamples of solids or liquids to determinefeasibility of desired separations. The unitincludes a co-solvent pump for adding a smallamount of liquid co-solvent to thesupercritical fluid to modify the solubilitycharacteristics of the solvent system. Inaddition, the unit has a recirculation pump,which moves supercritical fluid through theextraction vessel at a high velocity. Also, theunit has been designed to allow flammablesupercritical solvents, such a.« hydrocarbons,to be safe'y employed.
The feasibility unit operates as follows:Liquid carbon dioxide from a dip-tube storagecylinder is subcooied using a giycol-cooledheat exchanger to prevent vaporization in themain pump. For carbon dioxide, cylinderpressure is the saturation pressure at ambienttemperature, about 6.2 Mpa. The subcooiedsolvent is compressed from cylinder pressureto the extraction pressure, 13.8 Mpa for thisseries, using a reciprocating, packed-plungerpump. An electric heat exchanger raises thehigh pressure solvent to the extractiontemperature, which was maintained at 35°Cfor this series. The solvent, now atsupercritical conditions, flows continuouslyupward through a batch of materials (feed) inthe extractor vessel. Electric band heatersunder temperature control maintain theextractor at the desired temperature.
Supercritical solvent, containing solublecomponents from the feed, leaves theextractor and is split into two streams. Themajority of the solvent is recirculated backinto the extraction vessel via the recirculationpump. A small flow enters a separation areacontained in a temperature controlledoven. Within the oven, the solvent flowsthrough a pressure control valve, which iselectrically heated. Flow through this valvereduces the stream's pressure so that the fluidentering the separation vessel is more "gas-
like" and no longer has good solventproperties. Control of the oven temperatureand of the electric heaters on the separatorvessel itself determines the separationtemperature. The material previously dissolvedin the solvent precipitates in the separator asa solid or liquid that can be subsequentlyremoved from the system.
The solvent leaving the first separator canflow through an additional control valve intoa second separator, where material may becollected under different pressure andtemperature conditions. Solvent leaving thesecond separator can be directed through acold trap, cooled by dry ice, to collect veryvolatile components. From the cold trap, thesolvent stream passes through a dry test meterto measure flow rate and is then vented to theatmosphere.
TEST COUPONS AND COMPONENTS
To determine the efficiency of supercriticalcarbon dioxide in the removal of machinecoolants, thirty test coupon disks, 1.5 in. diamx 0.25 in. thick, were used. Twenty of thedisks were 303 stainless steel and ten werealuminum. An additional twenty small Inconelglass ceramic components, having an overallcylindrical shape and approximately 0.5 in.diam x 0.625 in. long, were tested.
The coupons were cleaned two at a time.They were positioned on edge and suspendedone above the other in the center of thecleaning vessel. Small pieces of stainless steelscreen were used to keep the coupons in avertical position and away from the vesselwalls. Carbon dioxide entered through a portand distributor at the bottom of the cleaningvessel, passed up over the surfaces of the twocoupons, and exited the cleaning vesselthrough the cover. This assembly providedhigh solvent velocity and good contact acrossthe coupon surfaces.
103
MATERIALS
Carbon Dioxide
Commercial grade liquid carbon dioxide fromLiquid Carbonic Corporation was used for allexperiments in this program. The standardsfor this grade of carbon dioxide are shown inTable 1.
Machine Coolants
The three hydrocarbon machine coolantsstudied were Cimperial 1011, Gulfcut 11D,and Cimperial 15.
EXPERIMENTAL PROCEDURES
Cleaning
The coupons were coated with oil in thefollowing fashion: First, a clean, dry couponwas weighed. Approximately 50 mg ofmachine coolant was then poured onto a non-linting, absorbent towel. The oil was wipedonto the surface of the coupon with care toevenly coat the entire surface. Finally, thecoupon with machine coolant was weighedagain; the weight gain was recorded. Thisprocedure resulted in an average of 5.5 mg ofoil being deposited onto the coupon.
After completing the coating procedure, thecoupons were loaded, two per experiment,into the 300 mL cleaning vessel. The vesselwas closed, and the system was purged withlow pressure carbon dioxide for about 5minutes. After purging, the vessel waspressurized via the solvent pump. Control ofsystem pressure was maintained by adjustingthe pressure control valve on the outlet of theextractor.
Solvent flow was established through thesystem and controlled by varying the strokerate of the solvent pump. Simultaneous withthe pressurization of the extraction vessel,temperatures were brought to the desiredlevels and maintained via electric heaters.Once operating conditions were reached, the
recirculation pump was turned on, and a highflow of solvent was maintained up throughthe cleaning vessel. A steady flow ofsupercritical solvent was maintained so thatthe so!vent-to-feed ratio increased with time.
When the desired cleaning time had beenreached, solvent flow was stopped by turningoff the solvent pump and the recirculationpump. The system was depressurized byslowly venting the solvent through theseparator and out the vent system. Once theextiaction vessel had returned to atmosphericpressure, the vessel was opened and thecoupons were carefully lifted out for analysis.
ANALYTICAL
Hexane Rinse
The cleaned coupons and components werecarefully removed from the extraction vesseland rinsed with two, 4.0 mL aiiquots of highpurity hexane. The two hexane aiiquots werecombined and evaporated down to about 1.0mL in volume. This sample was thentransferred to a reaction vial and furtherevaporated to 500 mL.
Capillary GC Analysis
All samples were analyzed on a Perkin Elmer#8320 Capillary Gas Chromatographyequipped with a flame ionization detector. A15-m, RTX-1 non-polar column (0.53 mmi.d., 0.5 micron film) was used. The details ofthe chromatographic method are provided inTable 2. Hexane concentrate injections of 1.0ml. were made with the GC in a split injectionmode. Figures 3 and 4 are chromatograms ofthe hexane rinses of the coupons andcomponents after cleaning with CO2.
RESULTS
The three machine coolants studied, Cimperial1011, Gulfcut 1 ID, and Cimperial 15, are allsoluble in supercritical carbon dioxide. Atable of the solubility measurements is as
Although the table of solubility measurementsis incomplete, it appears that Gulfcut 11D ismore soluble than Cimperial 1011, which ismore soluble than Cimperial 15.
Supercritical carbon dioxide can cleanCimperia! 1011, Gulfcut 11D, and Cimperial15 from coupons and components at theoperating conditions studied. Average residualcontamination on the order of 0.65% of theinitial loading was found at the conclusion ofnearly every cleaning experiment, independentof experimental conditions studied to date.
CONCLUSIONS
Supercritical carbon dioxide is veryeffective in removing Cimperial 1011,Cimperial 15, and Gulfcut 11D fromthe surface of aluminum and stainlesssteel coupons at operating conditions of35 °C and 2000 psig. Residual oilconcentrations equivalent to less thanor equal to 0.15 mg/cm2 (99.94%removal efficiency) were achieved.
Supercritical carbon dioxide is veryeffective in removing Gulfcut 11Dfrom the surface of Inconel-glassceramic components at operatingconditions of 35°C and 13.8 Mpa.Residual oil concentrations equivalentto less than or equal to 8.20 mg percomponent (99.97% removalefficiency) were achieved.
At the operating conditions used forcleaning (50°C, 13.8 Mpa), 15 min ofcontact with supercritical carbon
dioxide is sufficient to clean testcoupons and Inconel-glass ceramiccomponents. A longer contact time of30 min did not result in significantlycleaner parts.
4. The solubility of the three machinecoolants studied varies as a function oftheir composition. Although the tableof solubility measurements isincomplete, it appears that Gulfcut 1 IDis more soluble than Cimperial 1011,which is more soluble than Cimperial15.
5. For the one material measured at both35°C and 50°C, Cimperial 1011,solubility increased by more than 50%at the lower temperature.
DISCUSSION
Under the stated experimental conditions thisstudy has demonstrated that high pressure,supercritical carbon dioxide shows potential asa cleaning media for removing hydrocarbonmachine coolants from metal substrates. Itappears that the machine coolant's solubilityin supercritical carbon dioxide was greatlydependent on CO2 density. This was apparentwhen the solubility of one of the machinecoolants increased dramatically as thetemperature was dropped from 50 °C to 35 °Cwith the pressure maintained at 13.8 Mpa.This drop in temperature, while maintaininga constant flow rate, increased the density ofCO2 from 0.675 g/cc to 0.825 g/cc and thesolubility of Cimperial 1011 from 1.33 wt %to 2.17 wt %. This change in solubility as afunction of CO2 density under controlledconditions clearly indicates the potential utilityof supercritical CO2 as a versatile cleaningmedia for production cleaning operations.
ACKNOWLEDGEMENTS
EG&G Mound Applied Technologiesappreciates the efforts that Supercritical
105
Processing Inc. put forth in completing therequested experiments for this study.
Charpentier, B. A. and M. R. Sevenants, ACSSymposium Series 366 (1987).
Davidson, P., R. D. Gray, Jr., M. E.Paulaitis, and J. M. L. Penninger, Ann ArborScience Publishers (1983).
Johnston, K. P. and J. M. L. Penninger, ACSSymposium Series 406 (1988).
Krukonis, V. J., M. A. McHugh, J. M. L.Penninger, and M. Radosz, Volume 3 (1985).
Column - RTX - 1 (non-polar)- 15 m length- 0.53 mm i.d.- 0.5 micron film thickness
GC Parameters
Initial Oven Temperature - 120°CHold Time - 2 minRamp 1 - 30°C/min
Second Oven Temperature - 268°CHold Time - 5 minRamp 2 - 20°C/min
Final Oven Temperature - 310°CHold Time - 1 min
Injector Temperature - 350°CDetector Temperature - 350°CPressure - 19.0 psig
GC ran in split injection mode with a vent flow of about 6 mL/min.
Sample Preparation:
- The sample was rinsed two times with 4 mL of high purity hexane.- The two rinses were combined in a sample jar.- The hexane rinse was evaporated to about 1 mL volume and transferred to a 1 mL reaction vial where it was evaporated
to a volume of 500 mL.- A 1 mL sample was injected onto GC column with syringe.
108
SOLVENT PRESSURE
OLVENT
COMPRESSED GASLIQUID
PRESSURE REDUCTION
EXTRACT
-VENT
I IELEVATED TEMPERATURE
PRESSURE: TEMPERATURE SEPARATOR: FLOW:
• PUMP • HEATING TAPES • &P - • ROTAMETER• COMPRESSOR • TEMPERATURE BATH GLASS VESSEL • DRY TEST METER
• OVEN LOW PRESSURE • MASS FLOW METERVESSEL
• A T -HIGH PRESSUREVESSEL
Figure 1
CARBONDIOXIDE
FLOWTOTALIZER
Figure 2
109
METHOD CURRENT
A t4 l . I I 86Ui040
u 3
U 6
Ilk3.33
Standard 0.475ug/ul Clmparial 1011o i l in Baxans
riGDRE 3
13.89-BEND
U 5
A 64 C 18 • »CH O.QIT
O.lBug/ul Cinperial 1011 Oil in Bexane AfterSupercritical Carbon Dioxide Cleaning
of Sample I At) 10
FIGURE 4
110
CLOSED LOOP ALTERNATIVE TO THE USE OFHAZARDOUS CHEMICALS IN INDUSTRY
Dan SuciuTEC/NIQUES International Ltd.
Benton Harbor, Michigan
ABSTRACT
With the advent of new technology inconnection with alternatives to CFCs and thelowering of VOCs worldwide, the search foralternatives that are the most practical,productive, and cost-effective is a monstrous,time consuming, and expensive undertaking.
This paper deals with the research anddevelopment of a "Closed Loop" systemwhich will assist industry in making thosedecisions. The system is a time-saving,money-saving, effective and efficient methodof replacing old technology, equipment, andchemical waste treatment with new, viable andstate-of-the-art alternatives.
This "Closed Loop" system has beendeveloped by a coalition of experts in thefollowing three fields: aqueous and semi-aqueous chemicals; suitable types ofequipment for application of those chemicals;and various methods of waste treatment foraqueous and semi-aqueous spent andcontaminated effluent. The end result is asystem that works cost effectively andefficiently. It eliminates the need of investingmassive amounts of time and money in manyindividual interviews in the search for properequipment and materials. Decision makingwill be greatly simplified with less confusionand with minimal disruption in the changeoverto a viable and totally functional system.
INTRODUCTION
The "Closed Loop" system is formed througha coalition of three companies representingaqueous and semi-aqueous chemicals, aqueous
and semi-aqueous chemical cleaningequipment, and effluent waste treatment andwaste minimization. The members of thecoalition, TEC/NIQUES International, are:E n v i r o n m e n t a l R e s e a r c h andDevelopment, Inc. (ERAD); The RansohoffCompany; and 3D, Inc.; representingtreatment, equipment, and chemicals forsolvent substitution. This coalition eliminatesthe need for plant environmental andoperations personnel to become experts in theavailable substitute chemicals, chemicalcleaning equipment, and waste treatmenttechnologies, while minimizing the timerequired in researching prospective vendorsand reviewing various literature and papersdescribing all the "best" technologies. At thesame time the substitution can be customdesigned to incorporate existing and expectedneeds of the company. In this custom fit, theequipment, technologies, or chemical cleanersmay not all be manufactured by members ofthe coalition, but the procurement andimplementation will bv coordinated throughthe coalition, thereby providing a completeoperating product. Each aspect of thecoalition is discussed separately below.
CLEANING CHEMICALS
Many indus t r i e s have rep lacedperchloroethylene with 1,1,1-trichloroethane.This replacement only involved removal of theperchloroethylene from the vapor degreaserand addition of the 1,1,1 -trichloroethane. Nosignificant change in operation occurred. Theoperators only observed a name change.Substitution of chlorinated hazardous solventswith non-hazardous cleaners cannot beaccomplished by this "easy" one-stepoperation.
I l l
Cleaning with the hazardous solvents offersgood oil removal and cleaning capabilities.The solvent evaporated quickly and providedspot-free parts normally requiring no auxiliarycorrosion protection. There is little treatmentof the solvent required. The sludge (dirt, oil,grit) is pumped from the tank (vapordegreaser) to be drummed for disposal. Thesolvent is replenished regularly by the additionof new solvent. The solvent is lost due todrag-out and evaporation and most oftenrequires replenishing at a significant rate.
Extremely efficient substitute aqueous andsemi-aqueous cleaners are available to replacethe hazardous solvents. The cleaner chosenfor a specific job is dependent on manyvariables, such as the type of soil to beremoved, the metal to be cleaned, and theprocess to follow the cleaning. A number ofvariations of these cleaners are availabledepending on the application equipment. Forexample, when used in a high pressure cabinetwasher or when high pressure spray is appliedto the surface being cleaned and foaming isundesirable, low foam or de-foamed versionsare effective alternatives: Passivators can beused to prevent flash rusting on certain ferrousmetals, and to prevent corrosion on non-ferrous metals during cleaning cycles or in therinse cycle. All variables must be reviewed inorder to choose the proper aqueous or semi-aqueoi-s cleaning chemical that will bringoptimal results.
Efficient use of the substitute aqueous or semi-aqueous cleaners cannot be achieved in theexisting vapor degreaser. In many casesrinsing and drying of the part is required.However, equipment is available which makesthe entire process a continuous, totallyintegrated operation.
Substitute aqueous and semi-aqueous cleanerswill require treatment within the cleaningprocess to remove the grits and oils and toprevent them from adhering to the part andtreatment of the spent substitute cleaners priorto disposal. Although the substitute cleanersare non-hazardous, the cleaning process may
produce a hazardous waste by adding soils,oils and greases, and metals to the spentcleaning solution. In order to m\nhr.:*je thehazardous waste disposal cost and the long-term liabilities associated with the waste, thespent cleaning solution should be treated priorto discharge.
CLEANING EQUIPMENT
In order to successfully and effectively replacean immersion, spray or vapor-type solventdegreasing system with an aqueous-basedprocess, the following considerations shouldbe examined.
PARTS HANDLING
Parts handling is an important feature of anaqueous cleaning system. It must be designedto address the often complex configurations ofparts being processed. If parts have intricateinternal passages, pockets or crevices that tendto cup solution or if there are areas difficult toreach with a cleaning solution spray, parthandling becomes critical.
The handling system must assure that allsurfaces of the parts are positioned formaximum spray impingement. Parts that arenot free-draining must be rotated or tilted toprevent cross contamination or carry-out ofprocess solutions. The part transfer designmust prevent damage to delicate, machinedparts and, most importantly, the handlingsystem must be designed to integrate withexisting methods of parts loading andunloading.
CLEANING PROCESS
Selecting the proper cleaning process willassure that part cleanliness and finalappearance will conform to applicationrequirements. Spray, agitating immersion,ultrasonic immersion or a combination of thesecan be selected to address specific
112
applications. Although single-stage processesare often adequate for many applications, it issometimes necessary to utiiize multiple-stageprocesses when part cleanliness andappearance are important.
An effective rinse is necessary to assure thatcleaner residue is completely and thoroughlyremoved. If the plant water is of poor qualityand water spotting is a major concern, it maybe necessary to include a final deionized waterrinse. A controlled part temperature and awell designed, high-velocity air knife dry-offwill also help prevent part spotting orstreaking.
If a corrosion-preventive treatment is calledfor, it can easily be applied during the rinsecycle, or just prior to dry-off.
The quality of part rinsing and part dry-off inthe aqueous cleaning/treating process isimportant if the process is to equal theperformance attained with the hazardoussolvent cleaning system.
BATH CONTAMINATION
The type and amount of contamination beingremoved from the parts must be consideredwhen designing an aqueous process.Historically, very oily parts would dictatesolvent cleaning because distillation couldseparate solvent from the accumulated oilfor reuse.Recent advances in oil-removal methods foraqueous cleaning solutions have made itpossible to deal with high volumes of non-soluble oil entering the cleaning bath on acontinuous basis. The oil removal systemshould be a decant style, designed as anintegral part of tne solution holding tank. Itmust prevent the oil that is removed frombeing recirculated through the process pump.This prevents the centrifugal pump fromblending or emulsifying the oil, makingseparation more difficult. This integral oiltrap should then skim off collected oil into adecanting chamber where a more complete
separation can be achieved. Also note that thecleaning chemical utilized should be one thatreadily releases or separates the collected oil.
In applications where a high volume of w?ter-soluble oils that do not easily separate fromthe cleaning solution are being removed it isrecommended that a continuous overflowapproach be used with direct cleaner injectioninto the overflow supply line. Non-solubleoils are much easier to deal with in anaqueous cleaning process.
Various methods of solution filtration also canbe applied to the cleaning system. Filtrationwill remove fine, suspended particulates thattend to build up in the cleaning bath and canultimately redeposit on the parts. Theselection of the proper filtration systemcoupled with an effective method of oilremoval will assure optimum life of thecleaning bath.
CONSTRUCTION MATERIALS
The aqueous cleaning system should beconstructed from materials that are compatiblewith the process cleaning solutions. Mild steelconstruction usually is acceptable for analkaline cleaner stage as the alkaline detergentoffers adequate protection against corrosion.With mildly acidic washes, fresh water ordeionized water rinses, it is recommended thata type-304 grade stainless steel be used. Mildsteel construction should also be acceptable forthe dry-off section.
There are various types and styles of aqueouscleaning systems available. They range fromsmall, single-station dip tanks to complexautomated-transfer, mi/iti-siage systems.Several aspects need to be considered indetermining the proper machine type for theapplication. These include the basic machineconsiderations relative to continuous part flowor batch part flow along with partconsiderations, such as: whether the parts aredelicate, susceptible to marring or scratching;part sizes; part segregation; part orientation
113
and volume of parts to be processed. Otherconsiderations are the area of the plant inwhich the equipment will be installed;adequate sizes of clearances; floor loadlimitations; pad and pad curbing; floor drainsand containments; and location of the wastetreatment facility.
WASTE TREATMENT
Determination of the applicable treatment ofthe cleaning waste effluent is dependent on thetype of cleaner used, the availability ofexisting treatment processes, the type of soilbeing removed, and the availability of spacefor installation of the treatment system. Pointsource treatment of the cleaner can, in somecases, be implemented into the design of thecleaning equipment. Point source treatmentcan be complete, treating the waste to a levelrequired for discharge into surface water or asewage treatment plant or it can treat thewaste to a level required for discharge into anexisting industrial waste treatment plant. Insome applications, the treated water can berecycled into rinse tanks or other processlines, such as electroplating, requiring make-up water. In some of the applications, thetreated water is a better quality than thepresent plant water.
Effluent waste treatment processes includechemical precipitation, chemical oxidation,biological oxidation, absorption on granularactivated carbon, ion exchange separation,filtration, distillation, and separation of theoils and greases. Filtration removes the dirtfrom the cleaning solution and permits reusewithout these dirt particles adhering to thecleaned part. However, in many cases, thecleaner is degraded with use through breakdown of the active ingredients or dilution. Insome cases, the cleaner life can be extendedby pulling off a side stream for treatment andadding new cleaner or active ingredients of thecleaner. In this manner, only the side streamof cleaning solution and any oils and greasesseparated in the cleaning need to be treated.Presently, the oils and greases are not treated
in the plant, but are shipped at a nominal costfor processing. They are either burned orprocessed for reuse.
Chemical precipitation processes are used toremove metals and organics from the wastestream. The organics are co-precipitatedthrough the addition of alumina or ferric ironand the appropriate flocculents. Metalprecipitation to discharge limits can beachieved with the sodium sulfide/ferroussulfate metal treatment process whileminimizing the production of sludge. Thesludge produced in these processes must bedisposed of as a hazardous waste.
Chemical precipitation processes require theselection of the proper clarification system toenhance the coagulation and settling of theparticulates. The clarifier design is dependenton the process flow, type of particulates. sizeof particulate, and the available space forprocess implementation. The alternativecleaners may impact the clarification process,requiring the use of additional iron or aluminato achieve clarification.
CONCLUSIONS
This technological approach for thereplacement of toxic or hazardous solventswith environmentally safe cleaners offerssignificant advantages for users. The foremostadvantage is that one organization, consistingof a coalition of selected experts in the fieldsof treatment, equipment and chemicals workstogether as a team on each problem ensurethat every aspect in the solvent substitutionprocess operates as a cohesive, fully integrated"CLOSED LOOP" system.
114
Section II
ALTERNATIVE SOLVENTS
BIODEGRADABLE SOLVENT SUBSTITUTION
Anne E. CopelandDirectorate of Environmental Management
Tinker Air Force Base, Oklahoma
INTRODUCTION
The Air Force Logistics Command (AFLC)overhauls and repairs jet engines, aircraft andvarious components. The cleaning processes,which must precede all inspection and repair,use a variety of solvents includingperch loroethylene, methyl chloroform,trichlorotrifluoroethane, and PD-680(stoddard solvent). Because of theenvironmental and health hazards associatedwith these solvents, AFLC has recognized thatthese solvents must be replaced with saferproducts.
In 1987 the Air Force Engineering ServicesCenter (AFESC), at the request of AFLC,began the research and development projecttitled "Biodegradable Solvent Substitution."The objective of the program was to find safesubstitutes for the solvents used for metalcleaning at A F L C installations. The work,contracted to EG&G Idaho, Inc., wasperformed at Tinker Air Force Base. Thispaper presents the user's perspective of theprogram.
APPROACH
The program was structured in three phases,each lasting one year. EG&G performed amajority of the work at a pilot plant facility,located at Tinker's industrial wastewatertreatment plant (IWTP). This pilot plant is asmall scale replica of the IWTP.
Phase I - Phase I consisted of datacollection, establishment of criteria forsubstitute cleaners, a market search ofavailable products and screening tests.
EG&G surveyed AFLC to determine theircurrent cleaning processes, what solventswere used and in what applications. Thesurvey revealed the primary solvent use wasin the metal cleaning using perch loroethylene,methyl chloroform, trichlorotrifluoroethane,and PD-680 (stoddard solvent) to removeoils, greases, carbon, and masking wax usedfor selective plating.
The criteria AFLC established for newsubstitute cleaners included: (I) efficiency -substitutes must be at least as efficient ascurrent solvents; (2) flashpoint - flash pointmust be greater than or equal to 200 degreesFahrenheit; (3) biodegradability - a productmust biologically degrade, as measured by itschemical oxygen demand (COD), in six hours(actual retention time) in Tinker's IWTP to theNPDES permit limit of 150 mg/1; (4)corrosiveness - products must not causecorrosion rates to exceed 0.3 mil/yr onspecified metals (see Table 1) as measured byASTM Methods F483-77, "Total ImmersionCorrosion for Aircraft MaintenanceChemicals," and F519-77 for hydrogenembrittlement.
After searching the market for availableproducts, EG&G contacted 215 companiesand selected 175 samples to screen. Theproducts were screened for biodegradability,soil solubility, cleaning efficiency, andcorrosiveness. Of the products that passed thescreening tests, six were chosen for continuedevaluation in the program. It should be notedthat all of the products corrode magnesium ata rate greater than 0.3 mil/yr. Two of the sixproducts were later dropped from theprogram, one for low flashpoint and the otherfor a toxic component. During Phase III anew product, already tested for
115
performance in one of Tinker's overhaul shops,was incorporated in the program. Of thesefive final products, two are aqueous and threeare organic, or not water dilutable.
Phase II - In Phase II the chosen productswere subjected to extended performance tests.These tests included process enhancements(temperature, agitation, ultrasonics), cleaningcapacity, rinsing requirements and the impacton Tinker's IWTP. Results showed that aprocess temperature of 140 degreesFahrenheit coupled with pressurized sprayingor vigorous agitation and rinsing gave the bestresults. Ultrasonic enhancement, due to itsnumerous variables, was not pursued furtherin this program. It was, instead, placedunder its own separate program.
To determine the effects of these products onTinker's IWTP, the candidate substitutes wereprocessed through the pilot plant. The usedproducts were sequentially fed through thepilot plant and their effects on each unitprocess were monitored. The results werethat while all were biodegradable in thelaboratory jar tests, only one product wassuccessfully treated in the pilot plant. Oneproblem encountered with the other productswas that they floated the metal sludge that hadbeen intentionally removed from solution andprecipitated. While with some of the cleanersthis effect could be counteracted by addingferric chloride, the IWTP personnel were notin favor of adding another chemical to theprocess or the associated costs of this addition.Another problem, which occurred with only
one of the products, was that when thecleaner was mixed with the rest of the wastestream, the bacteria would not acclimate toit. In other words, the bacteria preferred theother "food" available and would notconsume our cleaner. It therefore passedthrough the plant untreated.
Phase III - In Phase III the cleaners weretested in full scale production. The processconditions shown to be optimum in Phase IIwere demonstrated in an agitated immersion
tank and a cabinet spray washer, similar to alarge dishwasher. Agitation was achieved byrecirculating the cleaner through a pumplocated outside the tank and reinjecting in thetank through submerged jet spray nozzles.The pump rate turned over the volume of thetank once every two minutes. Only theaqueous cleaners were tested in the spraywasher due to the explosion hazards associatedwith heating and atomizing the organicproducts. The tests were conducted in twoproduction shops at Tinker using actualengine and aircraft parts. The parts weresoiled with plating wax, oil, grease, lightcarbon deposits, and heavy, baked-on carbonfrom the hot sections of the jet engines.These parts, normally scheduled for chemicalcleaning, vapor degreasing or cold solventcleaning, were rerouted to our test process.The acceptance criteria for productperformance levied by our process engineerswere that the parts had to be clean enough (1)to undergo fluorescent penetrant inspection (amethod of nondestructive inspection) and (2)to accept paint. Only four of the fiveproducts were tested in production since twoof the products are chemically very similar.
RESULTS AND CONCLUSIONS
Overall the aqueous cleaners, 3D Supreme andFremont 776, performed far better than theorganic products. They successfully removedoil and grease from 100% of the parts andlight carbon deposits from approximately80% of the parts subjected to the tests in 5-15minutes process time. The shop operatorsreported the aqueous products cleaned betterthan vapor degreasing. These products didnot, however, completely remove the maskingwax or the heavy baked-on carbon, even after90 minutes in the immersion tank.
Of the two organic products, Orange-Sol's De-Solv-It removed the masking wax moderatelywell, with a process time of 30-45 minutes. Itwas, however, restricted by partconfiguration. Tinker's "worst case" part wasvery intricate and so ail wax was never
116
removed. Other less intricate parts weresuccessfully cleaned. Exxon's Exxate 1000also removed the wax, though not as well asthe De-Solv-It. One disadvantage of theorganic cleaners is odor. De-Solv-It, aterpene based cleaner, has a heavy citrus odorand the Exxate 1000, an acetate ester, is verypungent. Exxate 1000 did cause headacheswhen not vented.
The baked-on carbon deposits from the hotsections (combustion, turbine, exhaust, andafterburner sections) was never successfullyremoved within acceptable process times.Though in some cases the aqueous cleanerssucceeded with much time and effort, they didnot remove heat scale or corrosion. Since thesolutions currently used to remove scale andcorrosion (acids, bases, and oxidizers) alsoremove the heavy carbon ve*-y quickly, theywill not be replaced with any of the cleanersfrom this program.
As a result of this program, Tinker's overhauland maintenance shops are beginning toreplace their vapor degreasers and cold solventtanks with bom of the aqueous products, 3DSupreme and Fremont 776. Even though the3D Supreme won't be treated through the baseIWTP, it cleans better than Fremont 776 insome applications and has numerousadvantages over halogenated solvents,lternate treatment and possible recyclingmethods for 3D Supreme will be pursuednder another program, an expansion of thisone, sponsored by the Department of Energy(DOE). All test data from both the AFLCand DOE programs will be available throughG&G's Solvent Handbook
A. A. Chavez and M. D. HerdIdaho National Engineering Laboratory
Idaho Falls, Idaho
Chlorinated hydrocarbon solvents andchlorofluorocarbons are used extensively incleaning operations throughout the Departmentof Energy (DOE) defense program, thenuclear weapons complex, the Department ofDefense (DOD) weapons refurbishmentfacilities, and in industry. The objective ofthe Solvent Utilization Handbook Task Forceis to provide guidelines for the selection ofnontoxic environmentally safe substitutesolvents for these operations. The informationcontained in this handbook will includecleaning performance, corrosion testing,treatability operations, recycle/recoverytechniques, volatile organic compoundemissions and control techniques as well asother information. The handbook will beupdated on an annual basis with informationon new solvent substitutes that appear in themarketplace. Toxicological information,handling and disposal, and economics ofsolvent usage will also be included in thehandbook when available.
The near term objectives of the handbookprogram will be to screen replacementsolvents for cleaning/degreasing operations.The operations to be conducted in FY-91 willbe cleaning performance testing, corrosiontesting and handbook data base revisions.Longer range objectives will be to performextended performance testing which willinclude extended corrosion testing, hydrogenembrittlement testing, continued screening ofnew solvents, and full scale demonstrations.These longer range objectives will beperformed in FY-92 and beyond.
All samples received for testing in solventutilization undergo an initial screening processto determine their suitability for testing.MSDS information is examined to determineif the solvent is low toxicity. Currently the
flash point must be j>200 °F in order for thesolvent to meet combustibility criteria. If thesolvent meets the criteria of low toxicity andflash point, the solvent is entered into thecleaning performance testing program. Thecriteria for cleaning performance is determinedby weight loss of the soiled test sample. Thesolvent must remove 95 % of the soil on thetest sample, currently a SS pipe nipple. Uponpassing the cleaning performance test, thesolvent is then entered into corrosion testing.The corrosion tests are done in accordancewith ASTM F483-87, using 1,1,1-trichloroethane as a control. If the solventcauses corrosion J>. than found with 1,1,1-trichloroethane for any metallurgy, it is then isentered into further testing for VOCemissions, recycle/recovery and treatability.
Approximately 300 solvents have been testedwith soils specific to DOD and have beenevaluated for their cleaning performance andtheir corrosion characteristics. An additional100 solvents are currently being tested fortheir cleaning performance on soils specific toDOE and DOD facilities. The 16 soilscurrently being used to evaluate the solventsare: vaseline, WD-40, hydraulic fluid, Sani-Tuff handcream, dioctylphthalate, Trim Solmachining fluid, Krylon Acrylic Sprayovercoat, Versamid 140 epoxy curing agent,Epon 828 epoxy resin, Kester 197RMA solderflux, carnauba wax, Amber B2 wax, Ram 225mold release, MS 122 mold release, lanolin,and molybdenum sulfide grease. The alternatesolvents are also being tested for theircorrosion characteristics on metals and alloysthat are specific to DOE/DOD facilities. Thecorrosion testing will be according to ASTMp\thod F483-87 and will be conducted withthe following 25 metallurgies: Alloy 25,Alloy 52, Al 1100, Al 2024, Al 5456, Al6061, Al 7075, Al A356. carbon steel 1020,
119
carbon steel 4340, copper alloy CDA 110,copper alloy CDA443, copper oxygen-free,Inconel X-750, Inconel 625, Kovar, MgAZ31B, Monel K500, Ni 200, SS 304L, SS310, SS 4iO, SS 17-4 PH, Ti grade 2, andWaspaloy.
The handbook database is currently underrevision. Modifications to the databaseinclude utilizing a PC-based system with agraphical user interface operating under aUnix operating system and possibilities forfuture network capabilities. All data collectedfor the solvent utilization program will beincluded into the handbook database, as willdata obtained from collaborations with BoeingAerospace, Hughes, and others.
Phase II of the solvent utilization program willbe initiated in FY-92. Phase II will involvetoxicity testing (ATP), solvent loading andlifetime studies, rinsing requirements,extended corrosion testing, and hydrogenembrittlement studies.
Phase III will be full scale demonstrations ofthe alternative solvents selected and will beperformed at Tinker AFB. Phase HI will bedone in later years.
120
THE ELIMINATION OF CHLORINATED, CHLOROFLUOROCARBON, AND OTHERRCRA HAZARDOUS SOLVENTS FROM THE Y-12 PLANT'S ENRICHED URANIUM
OPERATIONS
D.H. Johnson, R.L. Patton and L.M. ThompsonOak Ridge Y-12 Plant
Martin Marietta Energy Sytems, Inc.Oak Ridge, Tennesse
INTRODUCTION
The Oak Ridge Y-12 Plant* is one of severalplants which make up the Department ofEnergy's manufacturing facilities forproduction of nuclear weapons componentsand subassemblies. The plant is acomprehensive manufacturing facility withoperations encompassing materialmanufacture, component fabrication andsubassembly generation. All associatedinspection and certification functions areperformed in the plant. The plant's twoprimary products are uranium and lithiummaterials.
A major driving force in waste minimizationwithin the plant is the reduction of mixedradioactive wastes associated with operationson highly enriched uranium. High enricheduranium has a high concentration oftheuranium-235 isotope (up to 97.5%enrichment) and is radioactive, giving offalpha and low level gamma radiation. Thematerial is fissionable with as little as twopounds dissolved in water being capable ofproducing a spontaneous chain reaction. Forthese reasons the material is processed insmall batches or small geometries.Additionally, the material is completelyrecycled because of its strategic and monetaryvalue.
Since the early eighties, the plant has had anactive waste minimization program which has
•Managed by Martin Marietta Energy Systems, Inc., forthe U.S. Department of Energy under contract DE-AC05-84OR21400
concentrated on substitution of less hazardoussolvents wherever possible. The followingpaper summarizes efforts in two areas-development of a water-based machiningcoolant to replace perchloroethylene andsubstituion of an aliphatic solvent to replacesolvents producing hazardous wastes asdefined by the Resource, Conservation, andRecovery Act (RCRA). A summary of theplant's overall solvent substitution andreduction program can be found elsewhere1.
A WATER-BASED MACHININGCOOLANT FOR USE WITH ENRICHED
URANIUM
A 50% mixture of perchloroethylene(tetrachloroethylene) and mineral oil had beenused in Y-12 for machining enriched uraniumfor nearly twenty years, but changingregulatory conditiions made its use verydifficult. Both the Clean Water Act and theToxic Substance Control Act listed perk as ahazardous substance and RCRA declaredwaste sludge of perchloroethylene to behazardous. For these reasons, a new coolantwas developed.
Perchloroethylene has several properties whichmake it ideal for use as a machining coolantfor uranium. It is non-reactive with uraniumand all known machine tool materials, it canextinguish small uranium chip fires, itenhances nuclear criticality safety due to thepresence of the chlorine-35 isotope which is aneutron poison, and it facilitates recycle ofchips due to the ease with which it evaporates.Any new coolant has to maintain thesecharacteristics. Additionally it has to be safe
121
for humans and generate no RCRA hazardouswastes with the over-riding issue being nuclearcriticality safety at the expense of any otherdesirable characteristics.
A secondary issue was an operatingphilosophy based on best managementpractices which encourages the use of genericrather than proprietary chemicals.Examination of available literature shows anumber of commercial, water-based coolantsare available which can be used with uraniumafter modification to insure nuclear criticalitysafety. However, formulations are notspecified, are subject to change, and varyfrom lot-to-lot; all of which conditions areunacceptable conditions in nuclear operationsand support a decision vo utilize a specifiedformulation.
The coolant formulation selected consists of a50/50, by volume, mixture of water andpropylene glycol to which is added 90 g/Lsodium borate, 1000 ppm sodium nitrate, anda few drops of Azure Blue dye. The sodiumborate is a neutron poison and provides thenecessary criticality safety margins for thecoolant. The sodium nitrate is a corrosioninhibitor and the dye is a coloring agent addedto facilitate quick visual verification that thecoolant in use is nuclear safe.
The new coolant was implemented in January,1985. Perch Ioroethylene usage in Y-12dropped from 1,200,000 pounds in 1984 toless than 130,000 pounds in 1986; howeverall chlorocarbons or chlorofluorocarbons werenot eliminated since degreasing agents andwater removal chemicals were still required.Any residue of sodium borate left on themachining chips must be removed prior tochip recycle in order to maintain the requirednuclear characteristics of the material stream.The chips are washed in distilled water toclean off the borate residue which leavesabsorbed water on the chips. This water isdisplaced by dipping the chips in Freon-113which is immiscible with water. Thedisplaced water then floats on the Freon-113and is skimmed off for recovery of any
residual uranium or disposal. Theseoperations led to an increase in Freon-113usage in-plant of about 60,000 pounds,yielding a net reduction in controlledsubstance usage of approximately 1.1 millionpounds.
ELIMINATION OF THE GENERATIONOF RCRA HAZARDOUS WASTES
FROMSHOP FLOOR OPERATIONS
Background
On December 20, 1989, Region IV of theUnited States Environmental ProtectionAgency (EPA) issued a regulatoryinterpretation memo2 concerning solventwipers which said that solvent wipes and rags"used in cleaning and degreasing operationswith any solvent or mixture of solventsidentified under the RCRA hazardous wastecodes, F001-F005, at 40 CFR $261.31" are"considered to be a listed hazardous waste(i.e., a spent solvent)."
On April 12, 1990, in the U.S. District Courtfor the District of Colorado, Judge Lewis T.Babcock issued a Memorandum Opinion andOrder3 in a civil suit between the Sierra Club,Plaintiff, versus the U.S. Department ofEnergy and Rockwell InternationalCorporation which stated that "Atomic EnergyAct" process residues are regulated underRCRA as mixed radioactive waste until theradioactive components are separated from theRCRA waste components.
These two rulings required an immediatechange in the way Y-12 was doing business.Firstly, all shop floor cleaning operationswhich used Freons, methyl chloroform, or anyvolatile organic compound (VOC) whichproduced wastes classified as RCRAcharacteristically hazardous had to be treatedas RCRA hazardous wastes and were nowsubject to manifesting and associated controlrequirements prior to disposal. Since thesewastes were incinerated to recover any
122
uranium residues, treatment facilities for suchwastes now had to be permitted as hazardouswaste treatment facilities, almost animpossibility for such "land-banned" wastes.Therefore, an immediate program wasundertaken to eliminate generation of allRCRA wastes from enriched uraniumoperations.
SELECTION OF A NEW SOLVENTSYSTEM CALLED WATERCHASER
140 FOR SHOP FLOOR USE
Several criteria were considered in selecting asubstitute solvent system. They includedrequirements that the new solvent should cleanas well as the solvent it was replacing, yieldnon-hazardous wastes as defined by RCRA,have low or minimal toxicity, be a non-airpollutant, be compatible with all weaponsmaterials, and require minimal changes in theplant's operational areas in order to complywith applicable safety and fire codes.Additional considerations were the universalityof the solvent and potential costs.
Examination of these requirements drives oneto the conclusion that the solvent should havea flash point greater than 139°F to meetOccupational Health and Safety Act (OSHA)requirements and be a Class III liquid asdefined by the National Fire ProtectionAssociation (NFPA) in order to allow openshop usage. A study of Y-12's productionoperations showed that the primary solvent inuse for part and component cleaning wasFreon-113 with lesser amounts of methylchloroform, naphtha and various low flashpoint alcohols. Comparison of thecharacteristics of these solvents with availablematerials using Hansen Solubility theory45 ledto the conclusion that a medium weightaliphatic hydrocarbon mixture with a smallamount of a polar co-solvent additive wouldmeet the cleaning requirements as well as coderequirements. Figure 1 shows a comparisonof cleaning characteristics of a number ofsolvents and methods using electronspectroscopy for chemical analysis (ESCA)
methodology to determine cleanliness. As canbe seen Water Chaser 140 fulfills the statedneeds.
A survey of available commercial solventsshows that a number of blends are availablewhich meet these general needs. Usually theyare called hydrocarbon blends or "varsols,"and are almost always refinery fractions. Assuch, they are subject to the variabilityinherent in these operations. To preventregulatory liability due to the presence ofuncontrolled or unknown (primarily aromatic)chemicals in the solvent due to lot-to-lotvariability, generic specifications requiringcertification of contents were generated andused for procurement.
The solvent selected for general shop usageconsists of a mixture of aliphatic hydrocarbonswith 5% dipropylene glycol monomethyl ether(DPM) which Y-12 calls Water Chaser 140.The DPM has alcoholic functional groups aspart of its structure which lends some polarcharacter to the solvent and causes water tobead up on uranium surfaces. Because of thestrong hydrophilic nature of the surface oxidefilm which forms when uranium is exposed toair, water forms a tightly bonded film onuranium. The DPM acts as a surfactant whichbreaks the bonds, causing the water to bead sothat it can be easily wiped from the part orcomponent. Table 1 gives a summary of theprocurement specifications for the solvent andTable 2 gives a summary of the pertinentcharacteristics of the material.
Based upon the data shown. Water Chaser 140meets the majority of the characteristicsdesired when the study started; however, someobvious problems arise. The flash pointcauses the material to be classed ascombustible, leading to increased potential fireloads in shops where either chlorocarbons orchlorofluorocarbons were previously used.Secondly, the material has a toxicity rating of2 based upon the TLV-TWA. For thesereasons, best management practices wouldindicate that the solvent wipes and rags shouldbe stored in closed, nuclear-safe, approved
123
containers. The low vapor pressure increasesdrying time, but also prevents exceeding theTLV if any ventilation at all is present. Theincreased drying time is not a problem inpractice if operators wipe parts to removeexcess liquid and, in fact, leads to cleanerparts because wiping is a much better cleaningmethod than air drying which is the generalpractice with highly volatile solvents.
Procurement was accomplished via standardbid processes with bids received from threechemical companies. In-house certification ofthe material indicated all specifications hadbeen met by the low bidder and the materialwas accepted for use as a weapons-approvedcleaning agent.
Implementation occurred during June, 1990.No major problems surfaced. To-date, no userelated health incidents have occurred; i.e., nodermatitis, allergic reactions, or skinirritations have been reported. Industrialhygiene monitoring has been on-going and to-date no incidence in which the TLV wasexceeded has occurred. On one occasion, aworst case scenario for exposure was mocked-up in which a two-gallon bucket of solventwas poured onto a machine; the maximummeasured air concentration was 25ppm. Withwiping, part cleaning times are comparable toprevious experience.
CONCLUSIONS
Two major generators of RCRA wastes in theY-12 Plant have been eliminated. A water-based machining coolant has beenimplemented to replace a perchloroethylenebased coolant and an aliphatic hydrocarbonbased solvent has been implemented to replaceprevious solvents which produced RCRAhazardous wastes when used in shop floordegreasing and cleaning operations.
4.
5.
REFERENCES
Thompson, L.M.; Simandl, R.F.; andRichards, H.L.; Y-xxxx, "ChlorinatedSolvent Substitution Program at the Y-12 Plant"; January, 1991.
Memo: James H. Scarbrough, Chief,RCRA Branch, United StatesEnvironmental Protection AgencyRegion IV Office, Atlanta, Georgia; toRCRA Branch Personnel; Subject:Regulatory Status of Solvent Wipers;December 20, 1989.
Memorandum Opinion and Order inthe United States District Court for theDistrict of Colorado; Lewis T.Babcock, Judge; Civil Action No. 89-B-181; Sierra Club, Plaintiff vs.United States Department of Energy,and Rockwell In ternat ionalCorporation, a Delaware corporation;Filed April 12, 1990.
Hansen, CM. ; "The Universality ofthe Solubility Parameter," IE & CProduct Research and Development;ACS Publication Volume 8, Number1; 1969.
Barton, Allan F.M.; CRC Handbookof Solubility Parameters and OtherCohesion Parameters; CRC Press,Inc.; Boca Raton, Florida; 1983; 594pp.
124
TABLE 1SUMMARY OF PROCUREMENT SPECIFICATIONS FOR WATER CHASER 140
Flash Point
Specific Gravity
Evaporation Residue
Acidity
Volume % Aromatics
Doctor Test (for Sulfur)
DPM
Aliphatic Hydrocarbons
141 °F Minimum per ASTM-D-56 TCC Method
0.777-0.827® 60°F
< 200 micrograms per gram
Neutral per ASTM-D-1093
5 % maximum per NMR Methodology
Negative per ASTM-D-235
5 +/- 1 % per G C Mass Spec or NMR
> 90% per G C Mass Spec or NMR
TABLE 2APPLICATION CHARACTERISTICS OF IMPORTANCE FOR WATER CHASER 140
Flashpoint
NFPA Class III Liquid
TLV-TWA
142°C
Allows open usage without necessityto store in a flammable storagecabinet during off shift
100 ppm
* Can cause irritation to eyes* Prolonged exposure to skin can cause dermatitis* Excessive inhalation can cause irritation, headaches,
or asphyxiation
Vapor Pressure 0.50 mm Hg
125
FIGURE 1 - ABIUTY OF SOLVENTS TO REMOVE UGHT OILS FROM 304L SS
70-,
o 00-
DC
[3Q
30-I
20-
g 10z
Jp
to
3
d
as
Itz
SAMPLES WERE CLEANED ULTRASONICALUT, CONTAMINATED.• FLUSHED WITH 10 ML SOLVENT, AND WIPED DOT
Katy WolfInstitute for Research and Technical Assistance
Los Angeles, California
INTRODUCTION
The two solvents most widely used forremoving the flux from printed circuit (PC)boards after components have been soldered tothem are 1,1,1-trichloroethane (TCA) and1,1,2-trichloro-1,2,2-trifluoroethane (CFC-113). These solvents will be bannedworldwide over the next decade or so becausethey contribute to stratospheric ozonedepletion. Over the short term, users mustbegin to adopt methods of reducing the use ofthese substances and over the long term, usersmust identify and implement alternatives.This paper discusses the methods for reducingor eliminating the use of the two solvents inPC board defluxing.
OZONE DEPLETION AND THEREGULATORY REGIME
In 1974, two professors first put forth thetheory of ozone depletion. There weresubstances called chlorofluorocarbons or CFCsthat were very stable. They were so stablethat they survived in the atmosphere withoutdecomposing for some 100 years. During thattime, they made their way from thetroposphere or lower atmosphere to thestratosphere or upper atmosphere. Oncethere, ultraviolet light decomposed them andthe chlorine they contained was released. Thischlorine was then available to catalyticallyreact with ozone, depleting the protectiveozone layer that shields us from harmfulradiation in the so-called B range. Each
chlorine atom was capable of reacting with100,000 times its own mass in ozone.
In 1978, the U.S. unilaterally banned the useof CFCs as propellants in nonessential aerosolapplications. For some years, this action wasassumed to be adequate because suchapplications represented about one-third ofworld CFC use. In 1985, the Antarctic ozonehole was discovered and the world becameaware that ozone depletion was a seriousproblem. In September of 1987, most majorworld nations signed the Montreal Protocol,the agreement to limit the use of ozonedepleting substances. The Protocol called fora phasedown to half the 1986 production levelof the CFCs by 1998. In 1988, the OzoneTrends Panel reported that ozone depletionwas more serious than had been previouslythought. In June of 1990, at a meeting inLondon, the Montreal Protocol wasstrengthened. The London Amendmentscalled for a complete phase out of the CFCsby the year 2000; TCA was added to theprotocol and would be phased out somewhatlater-in the year 2005. The Clean Air Act,recently reauthorized in the U.S., calls for adomestic CFC ban in 2000 and a ban on TCAin 2002. As an additional incentive todiscourage the use of zone depletingsubstances, Congress placed a tax of $1.10 perpound on CFC-113 in January of 1990 and atax of $0,137 per pound on TCA in Januaryof 1991.
127
PRINTED CIRCUIT BOARDDEFLUXING
In the PC board assembly process, flux is firstadded to the boards to facilitate solder flow, toeffect heat transfer and to prevent oxideformation. Flux is generally composed of avehicle, usually alcohol, and activators whichare solids. The components are then solderedto the boards. In the case of through holeboards, the components are soldered throughholes that have been drilled in the boards. Inthe case of surface mount boards, thecomponents are soldered to the surface of theboards. There is a trend toward surfacemount technology because a higher density ofcomponents can be obtained. CFC-113 orTCA combined with alcohol are the solventsmost commonly used for removing the fluxand other contaminants from the boards.
Military specifications (mil specs) currentlygovern the use of solvents in the cleaningprocess for those with military contracts. Themil specs have become the worldwide de factostandard. Mil-Std-2000 calls out the solventsthat can be employed in the cleaning process.It allows the use of CFC-113, TCA, variousalcohols or combinations of the solvents withalcohols. Water can be used only with specialpermission. The standard also requires theuse of a particular flux-rosin flux. Althoughthe use of recycled solvent is not expresslyforbidden, the standard refers to solvent purityspecs that cannot be met with recycledsolvent. The moisture level requirement isvery low and it is doubtful that most virginsolvent could meet the standard either.
In recognition of the fact that CFC-113 andTCA will be banned, the Department ofDefense (DOD) is aware that Mil-Std-2000must be changed to allow the use of othercleaning methods. Although two differentgroups are involved in the changes, there is noindication that the standard will be changed inthe near future.
METHODS OF REDUCING ORELIMINATING THE USE OF ZONE
DEPLETING SUBSTANCES
In the light of the ban on CFC-113 and TCA,users must employ measures for reducing theuse of the substances over the short term andeliminating their use altogether over the longterm. Short term measures include conversionfrom CFC-113 to TCA; adopting improvedequipment; use of recycled solvent; and vaporrecovery. Long term alternatives includesubstitution of flammable solvents,combustible solvents, HCFC or HFC solvents,aqueous based cleaning, no clean flux, andcontrolled atmosphere soldering.
TCA is currently four times less costly thanCFC-113. Its ozone depletion potential is alsolower-at only one-eighth that of CFC-113.Users can reduce their costs and improve theenvironment by converting to TCAimmediately. TCA is a more aggressivesolvent than CFC-113 and compatibilitytesting must be conducted to assess itssuitability for a particular board.Polycarbonate materials, for instance, areincompatible with TCA and some PC boardscontain it.
Many defluxing units allow high solventemissions. Adding refrigerated freeboardchillers and extending the freeboard can bettercontain the solvent vapors and reduceemissions. Automated hoists which removeparts at a standard 11 feet per minute canreduce dragout losses on boards. Hot vaporrecycle which relies on superheating thevapors can also reduce dragout.
Better solvent equipment with close to zeroemissions is being offered by European andJapanese manufacturers. Tiyoda, themanufacturer of one new unit, guarantees thatlosses will be no grater than 10 kilograms permonth. Such tight equipment is generally veryexpensive-between $125,000 and $200,000.
128
For users who will convert to alternative non-solvent processes, the capital investmentwould be significant for a short period oftime. For some users who plan to continueusing CFC-113 or TCA for several moreyears, this equipment could be cost-justified.
The Congressional tax applies only to virgin,not recycled solvent. Users can purchase anduse stills on-site and use the reclaimed solventin place of virgin solvent. Recyclers alsooffer recycled solvent that can be used inplace of virgin solvent at a reduced cost. Atthis stage, the demand for recycled solvent isvery high and it is not always easily available.Users subject to the mil spec probably cannotemploy recycled solvent until Mil-Std-2000 ischanged.
Vapor recovery methods can be used to reducesolvent vapor losses. Traditional carbonadsorption with steam desorption has beenused for many years with CFC-113 basedsolvents. TCA is extremely unstable tohydrolysis and steam is not a good choice fordesorption. Carbon can be used with inert gasregeneration, however. Another method, theBrayton Cycle Heat Pump, is beingdemonstrated in various solvent applications.It employs carbon and inert gas forregeneration and is said to reduce energy use.
Flammable solvents like alcohols and mineralspirits were used more frequently in the pastand they are an option to replace CFC-113and TCA. These low molecular weighthydrocarbons pose a workplace dangerbecause they are flammable and, because theyare photochemically reactive, many local airdistricts will not grant a permit for their use.EPA has issued a final test rule requiringtoxicity testing for isopropyl alcohol, one ofthe potential alternatives.
So-called combustible solvents are highermolecular weight hydrocarbons with flashpoints in the combustible range. Theseinclude solvents like terpenes, Dibasic esters(DBE), N-methyl pyrollidone (NMP) and alkylacetates. These solvents are not volatile and
they require a water rinse. The semi-aqueousprocess involves spraying with solvent, rinsingin an emulsion bath, rinsing in a water bathand drying. Equipment has been designed touse these solvents. Although they arephotochemically reactive and require a permit,emissions are likely to be relatively low. D-limonene, a major ingredient of terpeneformulations, has given a positivecarcinogenicity test in male rates. Most of theother solvents have not been tested for chronictoxicity or adequately scrutinized for theirenvironmental effects
DuPont and Allied are marketing mixtures ofHCFC-123 and HCFC-141b with alcohol forPC board defluxing. The HCFCs have lowerozone depletion potential then CFCs becausethey contain hydrogen and therefore breakdown more readily in the lower atmosphere.The two HCFCs are more aggressive thanCFC-113 but less aggressive than TCA. Theyare not compatible with polycarbonate.Depending on the specific blend, the boilingpoints of the solvents are in the 80 to 90degree F range. Solvent losses will be veryhigh unless the chemicals are used inextremely conservative equipment that hasbeen designed to minimize losses. HCFC-123and HCFC-141b are currently being tested bya consortium of international CFC producersfor chronic toxicity. The results of the testsshould be available in the 1993/1994 timeframe.
Asahi Glass is examining another HCFC,HCFC-225. It is an extremely gentle solventwith properties that are very close to those ofCFC-113. It has just gone into animal testingand the results are likely to be available after1995. The solvent is composed of twoisomers, one of which appears to be verytoxic. It is not clear whether a manufacturingprocess that selectively produces the non-toxicisomer can be developed.
One HFC is being marketed for PC boarddefluxing. Pentafluoropropanol has not beentested for chronic toxicity and is not likely tobe so tested in the future. It has a very low
129
workplace exposure level, reflecting its highacute toxicity.
Aqueous cleaning formulations can be used intwo ways. First, water can be employed withrosin flux and a surfactant. Second, water canbe used with organic acid or water dispersibleflux; no surfactant is necessary because theflux is water soluble. Closed ioop waterrecycling cannot be used with surfactantbecause it builds up in the system; on-goingresearch is attempting to solve this problem.Water is a suitable cleaner for through holeboards. Some users claim that wster cannotbe used for surface mount boards where thespacing between the board and the componentsis less than 10 mis. They argue that water,because of its contact angle, cannot penetrateunder the components and carry away thecontaminants. The effluent from boards thathave been cleaned is likely to requiretreatment to remove the lead from the solder.This can be done using an ion exchange.Water based cleaning requires more energy fordrying and generally more floor space.
Low solids flux contains only about 3 to 5percent solids compared to traditional fluxwhich contains 25 to 30 percent solids.Because of the low solids content, the fluxdoes not have to be removed from the boardafter soldering. A problem with low solidsflux is that much more stringent processcontrol is required to deposit the solids on theboard in a uniform manner. Not all users willbe sophisticated enough to employ thetechnique.
Controlled atmosphere soldering using areducing atmosphere like hydrogen or za\ inertatmosphere like nitrogen is another option.The equipment is expensive and morestringent process control is again required.
CONCLUSIONS
CFC-113 and TCA the major solvents usedtoday for defluxing PC boards will be phasedout over the next decade. Over the shortterm, users can adopt measure that willminimize the use of the solvents. Over thelong term, users can test and implementalternative processes that will eliminate the useof the ozone depleting substances. Virtuallyall of the alternatives have limitations andmany of them have unknown and unexploredhealth and environmental effects. The mostsuitable alternative will have to be selected ona case-by-case basis.
130
ELECTRONIC ASSEMBLY SOLVENT SUBSTITUTES
Alex Sapre, Ph.D.Environmental TechnologyHughes Aircraft CompanyLos Angeles, California
SUMMARY
In the worldwide movement to protect theenvironment, perhaps the most significantevent so far has been the signing of theMontreal Protocol in 1987, where over 40countries agreed to phase out the production ofchlorofluorocarbons and other ozone-depletingchemicals (ODCs). In June 1990, the LondonAmendment to the Montreal Protocolsubstantially accelerated the phase-outschedule and also covered additional ODCs,methyl chloroform and carbon tetrachloride.
The defense electronics industry issignificantly affected by these decisionsbecause ODCs are used in the manufactureand assembly of high quality electronicshardware. Moreover, in most instances suchuse is dictated by military specifications(milspecs) or industry-wide standards.
Many technology alternatives are beingdeveloped to replace ODCs and other solvents.However, no technology has completed all thenecessary requirements for wide-spreadimplementation at this time, particularly indefense electronics applications.
The cheapest and the quickest way to reducethe consumption of ODCs is to improveoperating and, wherever practical, engineeringpractices. There is a long list of recommendedoperating practices such as preventing draftsaround degreasers, covering degreasers,optimizing production schedules and solventreplenishment, repairing leaks, increasingoperator awareness, and training employees.
Engineering controls can also help reduceconsumption of ODCs. Some examples are
increasing freeboard height, retrofittingautomatic hoists and programming them forproper entry and exit speeds, installingautomatic covers, installing extra cooling coilsetc.
Organic solvents such as ketones, aromaticsand alcohols can remove solder fluxes andmany polar contaminants. However, they aretypically volatile, flammable, and theiremissions are considered VOCs, which incombination with nitrogen oxides cause smog.Accordingly, they are not desirablealternatives.
Chlor inated compounds such astrichloroethylene, perchloroethylene, andmethylene chloride are effective cleaners, butthey all are considered carcinogenic. Hence,they also will not be viable alternatives.
Water is an excellent solvent for removingionic contaminates and water soluble fluxes.Water in combination with a saponifier canremove water insoluble substances such as oiland rosin fluxes. The saponifier chemicaldoes not dissolve rosin but reacts with it toform an alkaline/amine salt which is watersoluble. In most instances, aqueous saponifiercleaning is claimed to be as effective as CFCblends in removing ionic and nonionic rosinflux residues. Aqueous saponifier cleaningmay also be used as an interim step to longterm conversion to organic water solublefluxes.
Saponifier cleaning does have severaldrawbacks. Aqueous saponifier cleaningrequires more process control than CFCsolvent cleaning to achieve consistent cleaningefficiency. The active concentration ofsaponifier in the wash tank is reduced over
131
time as a result of reaction with rosin and bythe addition of water to the tank to replacelosses from drag-out. Saponifier is an alkalinechemical which is corrosive to solder joints ifnot totally removed. This can put limitationson geometry of designs. Also, sapomfiers canbe incompatible with certain materials.
Regulations placing limitations on waste watercontaminants which can be discharged into apublic sewer without pretreatment are also animportant consideration when converting toaqueous cleaning.
A newer technology for aqueous removal ofrosin flux residue is the use of terpene solventcleaning. The terpene is mixed with a largeamount of a surfactant which acts as anemulsifier so that the terpene and rosin can beflushed off the circuit boards with water. Theterpene actually dissolves the rosin rather thanreacting with it as in the case with saponifiercleaners. The solvent is non-corrosive and amuch lower amount of lead is dissolved intothe rinse water.
There are some drawbacks to this cleaningprocess. Conventional aqueous spray cleaningequipment cannot be used because theemulsifier generates excessive foam. Themixture represents a potential safety hazardbecause it is combustible, with a flash point ofapproximately 120°F and is used at elevatedtemperatures. Terpenes are considered VOCs.In addition, water discharge regulations needto be taken into account even in this case.
In the arena of new solvents, Hughes isdeveloping two unique technologies. One,known as the reacting aqueous defluxingsystems (RADS), involves aqueous cleaning,while the other takes advantage of uniquesolvent properties of some substances undersupercritical conditions.
RADS is a four-part process consisting of areaction phase, neutralization rinse, final rinseand drying. In addition, a separate loop forregenerating the RADS solvent is added tominimize the effluent from the process. The
reaction phase operates at about 140°F with apH of 10.5-11. The neutralization rinse usesdeionized water at 120°F, while the final rinseconsists of deionized water at 100°F. Thedrying sequence can consist of any viablewater removal method.
Unlike other aqueous systems, RADS is anexcellent defluxing technology with no VOCemissions and with desirable environmentalcharacteristics. Its material compatibilitybehavior also appears very acceptable.Substantial work in optimizing the process andin further minimizing its environmental impactis currently underway for this technology.
Some chemicals develop good solventproperties under supercritical conditions.Currently, carbon dioxide is being evaluatedfor cleaning flex cables, cryogenic detectorcomponents, bearings, hybrid micro-circuitsand navigational gyros. Initial results on solderflux removal from a printed circuit board,designed to function as a power control unit,are very promising.
The use of organic water soluble fluxes andlow solids "leave on" ("no clean") fluxesoffers another way tc eliminate the use ofsolvents in defluxing.
Organic, water soluble fluxes have been usedfor wave soldering of circuit boards for manyyears. Nearly half of electronics companies inthe U.S. use water soluble flux. It has beenused in the computer telecommunications,industrial, automotive and consumerelectronics market segments. Water solublefluxes generally possess higher activity thanrosin fluxes so their soldering performance isbetter than rosin fluxes, particularly whenboards or component leads are excessivelyoxidized. Improvements in the technologyover the last several years have resulted influxes which result in better ionic cleanlinessand surface insulation resistance characteristicsthan previously attained.
Most water soluble fluxes possess a chloridecontent of 2-3 percent. Most are acidic,
132
highly ionic and potentially corrosive socircuit board assemblies must be designed sothat flux entrapment is avoided.
Surface insulation resistance after cleaning isusually reduced when using an organic, watersoluble flux. Reliability problems with highimpedance circuitry, are also experienced withthese fluxes.
Very few halide-free water soluble fluxes arecurrently available for wave soldering ofcircuit board assemblies. Their thermalstability is not as good as chloride-containingwater soluble fluxes in that they occasionallytend to form insoluble residues.
Water soluble fluxes offer the same waterdischarge considerations as aqueous cleaners.Because of their acidic nature, water solublefluxes can leach lead from the circuit boards.The large amount of organic material in thewash tank discharge can make total recyclingof the wash water for reuse on-siteimpractical.
One way to eliminate both the use of ozonedepleting solvents and also the potential waterdischarge problems is to use low solids "noclean" fluxes. Also, they offer additionalbenefits. For example, manufacturing floorspace required for a given line can be reducedby elimination of cleaning equipment.Components, which were hand soldered ontoan assembly after wave soldering because theycould not withstand the cleaning operation,can be inserted onto the circuit board prior towave soldering.
The "no-clean" fluxes are halogen-free with asolids content of 2-5 percent. Carboxylic-typeacids are the main activation system for theflux. Most assemblies will appearcosmetically clean after soldering. A smallamount of residue is left behind after solderingso the flux must not be capable of degradingreliability of the assembly to be suitable foruse.
The range of process parameters at which
optimum soldering performance can beobtained is narrower for low solid fluxes thanthat for the traditional flux types. Therefore,it would be more difficult to implement thesefluxes. Significant effort is necessary todevelop "no-clean" fluxes for high reliabilityapplications.
So far, the Department of Defense has notviewed favorably the use of water soluble and,particularly, low solid, "no-clean" fluxes.
Fluxless-soldering, inert atmosphere solderingand organic solders (conductive adhesives)offer another set of alternatives toreduce/eliminate solvent usage in cleaning ofelectronics hardware.
A project is underway at Hughes to develop aseries of lead/tin base solder alloys that willnot require fluxes for joining of aerospaceelectronics. Small amounts of a highlyreactive material such as lithium are added tochemically reduce the oxides of lead and tinwhen the solder is fused. The products of thesolder oxide reduction reactions are inertoxides of the additives. Several eutecticlead/tin solder alloys containing lithium orindium sulfide particles in the range of 0.1 to0.5 atomic percent additions have beenprepared using powder mixing techniques andmelting characteristics of alloys. Improvedtechniques for incorporating lithium particlesinto the solder are under evaluation.
In the area of inert atmosphere soldering,nitrogen inerted wave soldering machines areattracting considerable attention. In additionto nitrogen, formic acid is sometime injectednear the solder wave in order to scavengeresidual oxygen in the nitrogen, as well as theoxides on the board. In order to promote holefilling and to promote wetting of the clippedleads, substances such as adipic acid areapplied to the circuit assembly using a sprayfluxer.
Potential organic solder substitutes arepresently marketed as conductive adhesives,typically metal-filled epoxies, and are used for
133
many hermetic microelectronics applications.Currently a project is underway at Hughes todevelop a detailed specification describing therequired physical, mechanical and electricalproperties of the potential substitute materials.Of particular importance is preservation ofthese properties under effects of temperatureand/or humidity cycling and vibration.Approximately 30 organic solder sampleswould be evaluated in the project.
Table 1
Estimates of Ozone Depletion Potentials (ODP) and Global Warming potentials (GWP) ofSubstances Included in the Montreal Protocol (Normalized with respect to CFC-11)
•Global warming potential of CFC-11 is estimated to be 4,000 times the potentia] of carbon dioxide (CCy ona mass basis.
134
CHLORINATED SOLVENT SUBSTITUTION PROGRAM AT THE OAK RIDGE Y-12PLANT*
L. M. Thompson, R. F. Simandl and H.L. RichardsMartin Marietta Energy Systems, Inc.
Oak Ridge, Tennessee
INTRODUCTION
In recent years, several regulations regardingchlorinated solvents have been established.The Montreal Protocol, which has beenratified by the United States, calls for a banon production of chemicals such as methylchloroform and trichlorotrifluoroethane due totheir association with the depletion of theozone layer. Other chlorinated solvents suchas perchloroethylene and methylene chloridehave been identified as suspect carcinogens.All of these solvents mentioned above arelisted wastes under the Resource Conservationand Recovery Act (RCRA) which strictlyrestrains handling and disposal. Theseregulations make substitution of these solventsvery appealing. The Oak Ridge Y-12 Planthas been actively seeking .substitutions forthese solvents for the past 7 years.
CONSIDERATIONS FOR SUBSTITUTION
The first step in our substitution program wasto determine the uses of the chlorinatedsolvents. This step was done by conductingusage surveys in the plant and bv compilinginformation of purchases from the Y-12 storesand other purchasing systems. The main usesof these solvents were determined to be forcleaning purposes. The uses included cleaningparts after machining and prior to inspection,cleaning chips in the chip cleaning facility,cleaning urethane foam guns, cleaning meter
•Managed by Martin Marietta Energy Systems, Inc., forthe U.S. Department of Energy under Contract DE-AC05-84OR21400.
mix machines, and dissolving adhesives. Inlooking for substitutes for chlorinated solvents,there are several pitfalls one must avoid.
Large problems can result from overlookingsmall details. Factors which we take intoconsideration include compatibility, toxicity,flammability, means of disposal, effects onproduction, and ability.
Compatibility issues concerning the material tobe cleaned are usually addressed byconducting submersion tests and looking forsigns of corrosion. When addressingcompatibility issues, one must look not only atthe material to be cleaned but at the handlingmaterials such as gloves, wipes, anddispensers. Degradation of these materialscan transfer unwanted contamination to thepart as well as risk personnel exposure. Thisissue is usually addressed by conductingsubmersion tests with the materials in thepotential solvent substitute or conductingsurface analysis studies on a sample which hasbeen cleaned while using these materials.
Toxicity issues are addressed by searching forhealth properties of the solvent in sources suchas Sax's Dangerous Properties of IndustrialChemicals. Registry of Toxic Effects ofChemicals, and the Hazardous Substance DataBank. Our Industrial hygiene Department alsoevaluates the solvent and/or cleaning operationfor health concerns and determines ifmonitoring is needed or the personalprotective equipment to be used.
When it is necessary to replace a halogenatedsolvent with another organic solvent, one mustbegin looking at flammable so'vents. Due tostringent requirements by the OccupationalSafety and Health Administration (OSHA)
135
regarding highly flammable solvents, ourapproach has been to use solvents with a flashpoint of 140°F or higher. This also enableyou to be above the 140°F limit given as acharacteristic RCRA waste. Other concernswhich must be addressed concerningflammable solvents deal with the proper use ofthese solvents which includes storage, use, anddispensing. The ability of the solvent to formperoxides is examined so that the solventcould be handled safetly.
The means of disposal must be examined toprevent generation of a waste which cannot behandled. Generally, a biodegradable solventis desirable although incineration is acceptable.Bacteria from our biodegradation pond areplaced in a given amount of solvent. Thismixture is then tested to determine if andwhen the solvent is degraded.
Effects on production include several itemsdepending upon the next operation involved.When using a solvent with a high flash point,the evaporation rate is much slower thana chlorinated solvent. This results in changesin production whether it be simply waitinglonger before doing the next step or dryingwith a paper towel. Other effects includeeffects on bond strength, effects on welding,or effects on handling.
In testing the ability of a substitute solvent orcleaning method, comparative studies betweenpossible substitutes and the current cleaningmethod are first conducted on small samples.Surface analysis is conducted using X-rayPhotoelectron Spectroscopy (XPS/ESCA).Using this technique, a surface is bombardedwith X rays and the energy of the electronsemitted from the surface is measured. Theelectrons from different elements or elementsin specific bonding states have differentbinding energies. Thus, one can determinethe specific elements or combination ofelements on a surface from the measure ofthese different energy levels. XPS/ESCA iscapable of examining microlayers of a surface.Data are recorded for a contaminated surfaceto get a feel for what elements are present due
to the contamination. Peak height ratios of themain element associated with thecontamination to the base metal are calculatedfrom the XPS/ESCA data. These ratios arecompared to determine the effectiveness of thesolvent and/or cleaning operation. The lowerthis ratio, the cleaner the surface. The abilityto remove other elements associated with thecontamination is also examined. Afterconducting these studies, the solvent and/orcleaning operation is tested on a larger scale todetermine if there are any problems associatedwith its use.
ULTRASONIC CLEANING
Substitution efforts at Y-12 have been dividedinto two main efforts. The first effort was thereplacement of large vapor degreasers utilizingchlorinated solvents with ultrasonic cleanersusing aqueous detergent and water. Ultrasoniccleaning works by cavitating a liquid andforming small micro bubbles which burst onthe surface to be cleaned. This providesmechanical as well as chemical cleaningaction. Three variables can influence theeffectiveness of ultrasonic cleaning: 1) thefrequency of the ultrasonic cleaner, 2) theliquid medium, and 3) the coupling betweenthe cleaner and the liquid. In order to cavitatethe liquid, a frequency of at least 18 kHz isrequired.
Ultrasonic cleaning has been shown toperform as well as if not better than cleaningwith vapor degreasers as shown in Fig. 1. Inthis study, samples of uranium/6% niobiumwere initially cleaned ultrasonically for 8 h ina detergent, isopropanol, and demineralizedwater solution at 50°C, rinsed indemineralized water, fingerprinted, andhandled thoroughly. One sample was retainedas a control specimen while the remainingspecimens were dipped in a rust preventativeoil and allowed to dry. The specimens werecleaned using the. techniques described. All ofthe specimens with the exception of thespecimens degreased in isopropanol hadcomparable cleanliness levels. The sample
136
which was soaked in detergent had slightlyhigher levels of contamination. However, thislevel may be acceptable for some operations.In the majority of our operations, the power ofultrasonics was required to enable the liquid toclean in small cracks and crevices whichwould otherwise not be cleaned.
There are some drawbacks with ultrasoniccleaning. The equipment requires an initialcapita] investment, a rinse step must beincluded in the process and a drying step isalso necessary.
SOLVENT SUBSTITUTES
The second phase of the substitution programhas been the replacement of squirt bottle orspecialty type operations with other organicsolvents. The main application was thecleaning of parts after machining or prior toinspection by wiping. The contaminants beingcleaned from the surface were the usualmachine shop contaminants such as machiningcoolant, rust preventative oil, lapping oil,lubricants, and fingerprints. Initially, possiblesolvent substitutes were chosen using HansenSolubility Parameter Theory.2 Using thistheory, solvents which have similar parametershave similar solvent properties. A wide rangeof solvent types have been tested for thedifferent contaminants which are present. Theexperimental procedure used was to initiallyclean samples of type 304L stainless steelultrasonically in aqueous detergent and waterin order to establish a baseline level ofcleanliness. A sample was retained as acontrol sample. The remaining specimenswere coated with the contaminant and allowedto dry overnight. One contaminated samplewas also retained to determine what elementsare present due to contamination. Eachsample was then squirted with a given amountof solvent being tested and wiped dry. Thesamples were submitted to XPS/ESCA foranalysis. The main element present due to thecontamination was carbon. Therefore, a peakheight ratio of carbon to chromium (whichrepresents the base metal) was calculated and
comparisons were made of this ratio.
Figure 2 shows the results of the studycomparing solvents for the cleaning of rustpreventative oil. Solvents such as diprcpyleneglycol methyl ether (DPM), ethyl lactale,anisole, propylene glycol methyl ether acetate(pm acetate), ethanol denatured with acetone(EtOH/acetone), and isopropanol did notremove the rust preventative oil sufficientlyenough to enable the ESCA to see the metalsurface. A terpene-based cleaner and N-methyl pyrrolidone (NMP) worked as well asCFC-113. Solvent 140, which is a high flashmineral spirit, worked as well as the methylchloroform and better than CFC-113.Ultrasonic cleaning with aqueous detergentyielded the cleanest surfaces.
Figure 3 shows the results of a studycomparing the ability of solvents to removelapping oil. DPM, ethyl lactate, and aterpene-based solvent yielded the dirtiestsurfaces followed by anisole, pm acetate,NMP, and Solvent 140. A solvent blend,which consists of 95% Solvent 140 with 5%DPM, yielded the best results of the possiblesolvent substitutes compared to methylchloroform and CFC-113. Ultrasonic cleaningwith aqueous detergent again yielded thecleanest surfaces overall.
Figure 4 shows the results of a studycomparing the ability of solvents to remove awater-based machining coolant known as TrimSol. Solvents such as anisole, pm acetate,isopropanol, Water Chaser 140, and Solvent140 yielded surfaces comparable to thosecleaned with CFC-113. Surfaces cleaned withDPM, ethyl lactate, and a terpene-basedcleaner were somewhat dirtier and theethanol/acetone blend yielded the dirtiestsurface overall. The cleanest surface wasfound by cleaning with ultrasonic cleaningwith aqueous detergent.
Figure 5 shows the results of a studycomparing the ability of solvents to removefingerprints. The ethanol/acetone solventmixture and the ultrasonic cleaning yielded the
137
best results overall. Methyl chloroform,isopropanol, and Solvent 140 gave the nextbest results followed by Water Chaser 140.CFC-113 yielded the worst results and did notappear to remove the fingerprint oils. All ofthe organic solvents left behind inorganiccontamination from the fingerprints such assodium, nitrogen, sulfur, potassium, chlorine,and calcium.
Due to these results, the Y-12 Plant iscurrently in the process of changing to astrategy using Solvent 140 and Water Chaser140. Since Solvent 140 is totally immisciblewith water, a blend of Solvent 140 with DPMwas developed. Adding the DPM enables thesolvent to be slightly miscible with waterwhich aids the solvent in its ability to "chase"water-based machining coolants. This alsoadds the power of a polar solvent to that of anonpolar solvent. These solvents have lowtoxicity, are non-RCRA, are easily handledunder fire code considerations, and arecompatible with materials used at Y-12.Solvent 140 will be used in moisture-sensitiveareas of the plant while Water Chaser 140 willbe used in the remainder of the plant. Thedrawbacks of these solvents are tnat they areflammable and are slow evaporators whichrequire a change in production operations.
REMOVAL OF URETHANES ANDEPOXIES
Other uses of chlorinated solvents, namelymethylene chloride, at Y-12 have been for thecleaning of urethane foam spray guns,dissolving of urethane adhesives, and removalof epoxies.
Three solvents have been found to be effectivein the cleaning of urethane foam spray gunsand dissolving of urethane adhesives. Thesesolvents are anisole, dibasic esters, and N-methylpyrrolidone. The anisole works wellbut is a characteristic RCRA waste because itsflash point is below 140°F. Dibasic esterswork but are slow evaporators and one mustbe careful regarding compatibility with O-
rings using this solvent. The N-methylpyrrolidone also works well but is alsoa slow evaporator.
Several experiments have been conductedregarding the swelling of epoxy bonds.Methylene chloride is a good solvent for thisapplication because it is a small molecule andcan penetrate into the epoxy structure easilyand swell the epoxy. Substitutes such asanisole and N-methylpyrrolidone are largemolecules and cannot penetrate the structure aseasily. Thus, they act more slowly at swellingthe epoxy. However, adding a solvent with asmaller molecular structure to these solventsappears to penetrate the structure so that thelarger molecules can penetrate and work onthe epoxy. A 10% solution of acetone in N-rnethylpyrrolidone or a 10% solution ofmethanol in anisole appears to cut back on thetime required to swell epoxy compared to N-methylpyrollidone or anisole by themselves.
CONCLUSIONS
Chlorinated solvents are widely usedthroughout industry for cleaning purposes.However, because of health and environmentalproblems associated with their use,substitution of these materials has becomedesirable. The Y-12 Plant has successfullysubstituted ultrasonic cleaningwith aqueous detergent as a substitute for largevapor degreasers and Solvent 140 and Water
Chaser 140 for chlorinated solvents used insquirt bottle type operations. There aredrawbacks associated with the use of thesesubstitutes but these drawbacks can beovercome.
REFERENCES
1. Sax, N. Irving, Dangerous Properties ofIndustrial Materials, Van Nostrand ReinholdCompany, New York, New York.
138
2. Hansen, C. M., The Universality of theSolubility Parameter, IE&C Product Researchand Development, ACS Publication, Vol. 8,1969.
139
FIG. 1 - COMPARISON OF ULTRASONIC CLEANING WITH OTHER CLEANING TECHNIQUES
O.i-
IOA-
*.*
CLEANING TREATMENT
FIG. 2 - ABILITY OF SOLVENTS TO REMOVE UGHT OILS FROM 304L SS
701
g-ia:
90-
40-
•
JO-I
20-
W- sI 6
f
SAMPLES WERE CLEANED ULnUSOHKAUT, CONTAMINATED.A A A FLUSHED WTTH 10 UL SOUrtMT. AND WIPED DRY
s+1
I
2s'
•4 J"
•oiymruKO
140
FIG. 3 - ABILITY OF SOLVENTS TO REMOVE LAPRNG OIL FROM 304L SS
X WO
ii3.
lg
I«)
40-
s
I3
*
|
|
SAMPLES CLEANED UU1US0MCAUX C0KTAMIKA1ED.RUSHED WITH W UL SOIYEHT. AND WIPED D«Y
X 2• a
L IFIG. 4 - ABIUTY OF SOLVENTS TO REMOVE TRIM SOL FROM 304L SS
I
g »5
o
I
o
1
41
2
3
|
lAiina CLEANED ULTKASONICAUy. COHTAMINAItD.A fUJSHED W.TM 10 ML SOiyEKT. AND WIPED DRY
3 £
SOiyENTUSED
141
FIG. 5 - ABILITY OF SOLVENTS TO REMOVE FINGERPRINTS FROM TYPE 304L SS
8I
s •
2
I5
om
O
IN0K6AKIC SKCCS SUCH AS HA. M, S, K. Cl , AND CA LETT KHIND
s
iS
CLEANING 1KEATMENT
142
SOLVENT SUBSTITUTION FOR ELECTRONIC ASSEMBLY CLEANING
M. C. Oborny, E. P. Lopez, D. E. Peebles and N. R. SorensenSandia National LaboratoriesAlbuquerque, New Mexico
ABSTRACT
The Department of Energy (DOE) is strivingto eliminate the use of chlorofluorocarbon andchlorinated hydrocarbon solvents fromweapons production. A major use of thesematerials is in the cleaning of electronicassemblies during production. In seeking toeliminate these materials, a screening studyhas been completed to identify alternatematerials/processes. This screening studyinvolved parallel investigations into: 1) theuse of alternate cleaners for removing therosin-based flux and mold release materialcurrently being used, and 2) the use of watersoluble fluxes in place of the rosin-based flux.The evaluation criteria used in this screeningstudy were: the environmental, safety andhealth impact upon production operations,cleaning efficacy, corrosion potential, and, thebondability and high voltage breakdownresistance of cleaned surfaces. Uponcompletion of the screening evaluation,oxo-decyl acetate and a terpene cleaner havebeen selected for further study.
INTRODUCTION
Like industry in general, the DoE weaponscomplex utilizes chlorofluorocarbon andchlorinated hydrocarbon solvents for mostcleaning and degreasing operations. However,environmental, safety, and health (ES&H)concerns and regulations, now dictate that theuse of these materials must be minimized andeventually eliminated. Since electronic
assembly cleaning processes account for alarge percentage of total halogenated solventusage within the weapons production complex.
Sandia National Laboratories (SNL) and AlliedSignal/Kansas City Division (AS/KCD) haveestablished a joint program to identify, qualifyand implement alternative materials andprocesses that would eliminate halogenatedsolvents from electronic assembly cleaningprocesses at AS/KCD. This program ispresently focussed on a major electronicsystem. The fabrication of this system, whichcontains eight major electronics assemblymodules and numerous secondary components,requires 49 separate solvent cleaning stepsusing trichloroethylene (TCE). These cleaningsteps are primarily for the removal of rosinsolder flux residues after soldering, cleaningof completed subassembly and assemblymodules prior to foam encapsulation, andpost-encapsulation cleaning of silicone moldrelease from encapsulated units.
HALOGENATED SOLVENTELIMINATION OPTIONS
In seeking to eliminate halogenated solventsfrom electronic assembly cleaning processes,two separate options were investigated. Thefirst of these options is the use of alternatecleaning materials in lieu of the halogenatedsolvents now being used for the removal ofthe rosin mildly activated (RMA) solder fluxand silicone mold release residues. Thesecond option involves alternative processingusing water soluble soldering fluxes ratherthan the RMA solder flux now being used.Water soluble fluxes have the advantage thatthey can be cleaned with water, thuseliminating the need to use halogenatedsolvents for flux residue removal. Since moldrelease cleaning studies had indicated that thesilicone mold release agent can be removedwith isopropyl alcohol (IPA), it wasHypothesized that a two step cleaning process.
143
hot deionized (DI) water followed by IPA,would effectively remove residues from boththe water soluble solder flux and the siliconemold release agent. Additionally, using IPAas the second cleaning step would remove anywater remaining from the first step and rapidlyevaporate to leave dry assemblies.
ELECTRONIC ASSEMBLY CLEANINGPROGRAM
A three phase program was designed to carryout the studies necessary to identify, qualify,and implement alternative cleaning processesfor electronic assembly cleaning. Phase 1 isan initial screening study, involving a limitednumber of alternative cleaning agents andwater soluble solder fluxes, to identify a singlecleaner or flux for further evaluation. Phase2 involves more extensive testing andevaluation of the alternative cleaner or watersoluble flux that was identified in Phase 1.Finally, given a successful outcome of thePhase 2 testing and evaluation, Phase 3 will bethe implementation of the alternative cleaneror water soluble flux into productionoperations.
At this time, the Phase 1 screening studieshave been completed. In these studies fivesubstitute cleaning materials were evaluated aspotential replacements for halogenatedsolvents: a terpene, oxo-decyl acetate, twoproprietary aqueous cleaners, and isopropylalcohol. Additionally, two water solublesoldering fluxes, Kester 2120 and Kester2224, were evaluated as possible substitutesfor the Kester 197 RMA solder flux now inuse. A two step cleaning process, hot DIwater followed by IPA, was used in the watersoluble flux evaluation studies. Evaluationcriteria that were used in the Phase 1screening studies were: ES&H impact atAS/KCD, cleaning efficacy, cleanercorrosivity, bondability of cleaned surfaces,and high voltage breakdown resistance of highvoltage assemblies after cleaning. Thebondability and high voltage breakdownresistance testing were necessary due to
specific issues related to the production andfunction of this particular system.
ES&H IMPACT ASSESSMENT
An ES&H impact assessment was done byenvironmental, fire safety, and industrialhygiene personnel at AS/KCD. As a result ofthis assessment it was determined that,properly used, none of the alternativecleaners or water soluble fluxes would presenta significant ES&H problem with eithercurrent or anticipated future regulations.
CLEANING EFFICACY STUDIES
Cleaning studies were carried out at both SNLand AS/KCD. The SNL studies wereprimarily focussed on the removal of solderflux and silicone mold release agent from foursubstrate materials common to the electronicssystem under study. These materials werebare copper, bare copper which had beenfluxed and Sn/Pb solder dipped, 17-4 PHstainless steel, and E-glass/polyimide printedwiring board material. The bare copper andsolder dipped copper substrates were used tosimulate electronics materials before and aftersoldering. The 17-4 PH stainless steel is usedas the structural housing material for thee l ec t ron i c s a s sembl ie s and theE-glass/polyimide printed wiring boardmaterial is used in all of the electronicsassemblies in the system.
For the alternate cleaner studies, coupons ofthe four substrate materials were contaminatedwith each of the two contaminants: the Kester197 RMA solder flux and the silicone moldrelease agent. These coupons were thencleaned using TCE or one of the fivesubstitute cleaning materials. The TCE-cleaned samples were necessary to provide abasis for comparison between the currentcleaner and the alternative cleaners. Aftercleaning, the coupons were analyzed todetermine cleaner efficacy for eachcombination of substrate, contaminant, and
144
cleaner.
In the water soluble flux studies, coupons ofthe four substrate materials were contaminatedwith each of the two water soluble fluxes andthen cleaned using the two-step cleaningprocess. As in the alternate cleaner studies,the cleaned coupons were them analyzed todetermine cleanliness levels for eachcombination of substrate material and watersoluble flux.
The solder-flux contaminated coupons fromboth the alternate cleaner and water solubleflux studies were analyzed visually and witheither Auger electron spectroscopy (AES) orx-ray photoelectron spectroscopy (XPS) toidentify and quantify surface contaminants.An additional set of solder dipped coppercoupons was analyzed using an Omegameterto measure residual ionic contamination levels.
All silicone mold release-contaminatedcoupons from the alternate cleaner studieswere analyzed visually and also with eitherAES or XPS. Silicone mold release cleaningefficacy for the alternate cleaners was ahodetermined using contact angle goniometermeasurements. For these measurements,copper coupons were contaminated withsilicone mold release agent and then cleanedwith TCE or one of the five alternate cleaners.After cleaning, water drop contact anglemeasurements were made to determine relativesurface cleanliness levels.
In addition to the above SNL cleaning study,cleaning studies were also carried out at bothAS/KCD and SNL to determine theeffectiveness of TCE, the five alternativecleaners, and the two step DI water/IPAcleaning process in removing a number ofgeneral contaminants that are present in theproduction area. These studies were necessarybecause experience has shown that thesematerials occasionally end up on productionunits and must be removed. These studieswere performed on bare copper, barealuminum, and solder dipped copper substratesusing various oils, greases, mold releases and
body oils. Cleaning efficacy in these studieswas determined visually, by weight lossmeasurements, MESERAN (Measurement andEvaluation of Surfaces by Evaporative RateANalysis) values, water drop contact anglemeasurements, and Grazing AngleReflectance-Fourier Transform InfraredSpectroscopy.
Analysis of data from all the SNL andAS/KCD cleaning studies indicated that, of thefive substitute cleaners, the oxo-decyl acetateand terpene cleaner were the most effective inremoving the RMA solder flux, silicone moldrelease agent and general contaminants thatwere studied. These data also showed that theoxo-decyl acetate and terpene cleaner bothcleaned as well as TCE in these studies. Datafrom the water soluble flux and silicone moldrelease cleaning studies indicated that the twowater soluble fluxes and silicone mold releaseagent can be effectively removed using a twostep DI water/IPA cleaning process.However, the general contaminant studiesfound that this cleaning process would be lesseffective for removing some of the productionarea contaminants.
CLEANER CORROSIVITY STUDIES
Two types of tests were conducted at SNL toevaluate the relative corrosivity of thesubstitute cleaners and TCE. The first ofthese tests was an immersion test to determinedissolution rates in each of the cleaningsolutions. In this test, preweighed coupons ofcopper and Sn/Pb solder were immersed ineach of the cleaners for one week, at ambienttemperature. At the end of this time, thecoupons were removed, rinsed, dried, andreweighed. These weight loss measurementsprovided an indication of the relativecorrosiveness of each cleaner. As expected,the iwo aqueous cleaners exhibited greaterdissolution rates than the organic cleaners(TCE, IPA, oxo-decyl acetate and theterpene).
The second part of the corrosion evaluation
145
testing was designed to assess potentiallong-term corrosion problems due to cleanerresidues left from incomplete rinsing aftercleaning. In this test, bare copper and Sn/Pbsolder coupons were prepared under no rinse,partial rinse, and full rinse conditions for eachof the cleaners. These coupons were thenplaced in an environmental chamber and agedfor 30 days at 40°C, 70% relative humidity.At the end of this time, the coupons wereremoved and a visual assessment was made ofthe relative amounts of corrosion present.Results of the immersion and cleaner residuetests indicated that all five substitute materialswere acceptable.
BONDABILITY TESTING
Post-cleaning bondability of materials wasinvestigated at SNL. These tests werenecessary due to the large number of bondingand encapsulation processes which occurduring production of the system. In thesestudies coupons of 17-4 PH stainless steel, andE-glass/polyimide wiring board werecontaminated with the RMA flux and the twowater soluble fluxes. An additional set ofSn/Pb solder-dipped copper coupons wasprepared by dip soldering bare copper couponsthat had been fluxed with the RMA flux andthe two water soluble fluxes. Couponsprepared using the RMA flux were cleanedusing TCE or one of the five alternate cleanerswhile coupons prepared with the water solublefluxes were cleaned using the two step DIwater/IPA cleaning process. After cleaning,thin-wall steel cylinders were bonded to thecleaned substrates with an encapsulating resinand torqued to failure to determine theshort-term adhesive shear strength between theresin and coupon. Within experimentalscatter, all coupons yielded the sameshort-term adhesive shear strength.
HIGH VOLTAGE BREAKDOWNTESTING
The final evaluation testing in Phase 1
involved high voltage breakdown testing atAS/KCD. High voltages are present inseveral modules of the electronic system understudy and previous experience had shown thathigh voltage breakdown problems are often thefirst indication of incomplete cleaning and/orcontamination of these modules. High voltagebreakdown testing was done using a speciallydesigned high voltage test assembly. Threeseparate groups of test assemblies werefabricated using the RMA flux and the twowater soluble fluxes. After fabrication, u.:RMA fluxed assemblies were cleaned withTCE or one of the five alternate cleanerswhile the water soluble fluxed assemblies werecleaned with the two step DI water/IPAcleaning process. After cleaning, each testassembly underwent high voltage stress testingto determine breakdown behavior. Results ofthis testing indicated no significant highvoltage breakdown failures due to the use ofany of the cleaners or water soluble fluxes.
CONCLUSIONS
At the completion of Phase 1 testing, both theoxo-decyl acetate and terpene cleaner wereselected for further evaluation in Phase 2 ofthis program. This selection was based uponcleaner performance and systems-levelconcerns associated with the use of aqueous orsemi-aqueous cleaning processes.
As has been previously discussed, theoxo-decyl acetate and terpene cleaner werefound to be as effective as TCE in removingthe RMA flux, silicone mold release agent,and general production area contaminants.Additionally, in contrast to the proprietaryaqueous cleaners and water soluble solderfluxes, these cleaners neither contain norrequire water for their use. Although the useof aqueous processing was not a concern whenthese investigations were started, during thesestudies a systems level decision was madethat, if possible, no moisture should beintroduced into the weapon system as a resultof aqueous processing methods. This decisionhindered any possible use of the proprietary
146
aqueous cleaners and water soluble solderfluxes.
At this time, Phase 2 studies are underway.Among the issues being addressed in thesestudies are: materials compatibility, electronicfunctionality of cleaned assemblies, and thelong-term reliability of cleaned assemblies.
ACKNOWLEDGEMENTS
The authors wish to acknowledge the effortsof numerous individuals at Sandia NationalLaboratories and Allied Signal/Kansas CityDivision who contributed to this program.
The general contaminant cleaning studies atAS/KCD and SNL were done by M. G.Benkovich and P. J. Nigrey, respectively. J.E. Reich performed the contact anglegoniometer studies. M. E. Smith providedtechnical assistance for the XPS and AESmeasurements, and G. A. Poulter assisted inthe corrosion studies. Bondability testing wasdone by T. R. Guess, M. E. Stavig and D. L.Zamora. B. N. Harnden was responsible forthe high voltage breakdown testing atAS/KCD. This work was supported by theU.S. Department of Energy.
147
ALTERNATIVE SOLVENTS/TECHNOLOGIES FOR PAINT STRIPPING
M. N. Tsang and M. D. HerdIdaho National Engineering Laboratory
Idaho Falls. Idaho
Paint stripping is a necessary part ofmaintenance at U. S. Air Force Air LogisticsCenters (ALC). The waste from Air Forcepaint stripping operations contains toxicchemicals that require special handling andmust be disposed of as hazardous waste atconsiderable cost. Emi^ions from thesesolvents into the atmosphere as volatileorganic compounds (VOC) are another sourceof pollution. These wastes are hazardous tothe environment and to operating personnel.The paint stripping wastes are regulated by theU. S. Environmental Protection Agency(USEPA), which can impose fines onoperations whose wastes exceed theestablished limits.
The purpose of this program is to identify andtest alternative solvents and/or technologies tostrip paint from aircraft parts and equipmentefficiently with the overall objectives ofminimizing hazardous wastes and volatileorganic compound (VOC) emissions.Commercially available chemical paintstrippers will be tested for paint strippingefficiency, biodegradability, and corrosioncharacteristics. An extensive literature searchhas been completed to obtain backgroundinformation, abstracts, patents, militaryspecifications, and ASTM testing standards forpaint stripping. Several mechanical paintstripping methods have been discovered duringthe literature search and are being monitoredfor their applicability in Department ofDefense operations. A joint program has beenestablished between Boeing Aerospace, PacificNorthwest Laboratory, and the INEL with thegoal to expand the collaboration effort withother industries.
Phase I of this program has been completed.Phase I involved gathering baselineinformation, conducting laboratory screening
of potential alternative paint strippers for theirability to remove paint, determining thebiodegradability of the solvent, anddetermining the corrosion characteristics of thesolvent. Phase II is currently beingimplemented in FY-91 and will involveextended performance testing of the alternativepaint strippers surviving Phase I testing.Phase III will be conducted in later years andwill involve full scale demonstration of thepaint strippers selected.
Paint stripping efficiency was evaluated bydetermining the ability of the stripper toremove various types of paint systems frommetal coupons. The test methods weredeveloped from rr.il'tary and federalspecifications for paint stripping. Apreliminary test was conducted on all samplesto eliminate those that cannot remove paintunder moderate conditions. The effects ofeach stripper on the paint system weredetermined by visual inspection of the couponafter paint stripping, since this is the standardprocedure at the ALCs. For the preliminarytest, Al 2024 and an epoxy paint system wereselected as the representative paint system.Paint strippers passing the screeningrequirements of _>.50% paint removal weresubjected to a more stringent test to provideaccurate performance data. The second testused Al 2024 and carbon steel 1010 as thesubstrates and utilized six paint systems forthe test.
The paint systems used were:
1. Epoxy polyamide primer withepoxy polyamide topcoat.
2. EJastomeric polysulfideprimer and urethane topcoat.
5. Epoxy polyamide primer,polysulfide sealant, epoxypolyamide primer, and aurethane topcoat.
6. Epoxy polyamide primer withepoxy polyamide topcoat thatdiffered in formulation from#1.
The painted coupons were subjected toaccelerated aging by immersion in 2%hydrogen peroxide for 18 hrs. Thisaccelerates oxidation, which normally occurswith ultraviolet (UV) light and time. Couponsfor the preliminary test were not aged beforetesting.
Corrosion testing will be performed inaccordance with ASTM F483-87, ImmersionCorrosion Testing. Those paint removalsolvents passing the extended performancetesting will be subjected to corrosion testing,hydrogen embrittlement testing by ASTMF519-77, and biodegradation studies.
Currently 60 paint removal formulations havebeen screened. Out of this screening test, 10immersion paint strippers passed into extendedtesting. These ten paint removal solvents are:Chemical Methods CM-3707, ChemicalSolvents SP-800, Fine Organics FO 606,Frederick Gumm Clepo Envirostrip 222, GAFM-Pyrol, McGean-Rohco Cee Bee A245,McGean-Rohco Cee Bee A477, Patclin 126Hot Stripper, Rochester Midland PSS 600, andTurco T-5668. These solvents are currentlyundergoing extended testing, VOC studies,and recycle/recovery studies.
Several new process technologies for paintremoval have also been identified and arebeing monitored by the INEL. These newtechnologies include: wheat starch blasting,CO2 ice blasting, ice blasting, water jetblasting, flash lamp stripping, laser stripping,and sodium bicarbonate blasting.
150
A PROPOSED "MORE DEMANDING" PWB DESIGN AND TEST PLAN TO EVALUATEAQUEOUS AND SEMI-AQUEOUS CLEANING TECHNOLOGIES
K.K. Asada. K.S. Hill and M.D. WalleyRadar Systems Group
Hughes Aircraft CompanyLos Angeles, California
ABSTRACT
A "more demanding" PWB design and a testplan were developed through a project jointlyfunded by Hughes Aircraft Company andCalifornia's South Coast Air QualityManagement District (SCAQMD). This testplan follows the guidelines developed by theIPC/DOD/EPA committee for removal of fluxresidues using alternatives to CFC solvents.Hughes Aircraft Company also proposed twofollow-on phases which will establish theperformance of fluxes and solvents throughscreen testing and involve full qualificationtesting necessary to gain customer approvals.
The design and test plan were derived from anextensive literature and industry review ofexisting solvents, fluxes, and alternativetesting programs. Based on the need for morecomplex test vehicles, two printed wiringboards were designed for more demandingcleaning requirement configurations foradvanced complex electronics: one for surfacemount and one for plated through-hole. Thesurface mount design consists of 15 differentcomponent types and sizes ranging from 2 to408 leads with lead pitch from 0.050 to 0.020inch, and stand-offs from 0.001 to 0.013 inch.The plated through-hole design consists of 7component types and sizes ranging from 2 *o68 leads.
The comprehensive test plan involved themodification of the IPC process flows for eachboard type, as well as the creation of thenecessary travelers and process instructions.The test plan is divided into three segmentswhich describe in detail cleaning mediascreening, assembly screen testing andcustomer qualification testing. This is the first
comprehensive test program that addresses theevaluation of alternative cleaning technologiesfor defluxing electronic assemblies ofaerospace complexities.
INTRODUCTION
In 1987 the Montreal Protocol was developedand adopted to establish a production controlplan for chlorofluorcarbon (CFC) emissionsand other ozone-depleting chemicals to resultin a total phaseout by the year 2000. Giventhe potential for environmental damage, alongwith the pending regulations and new taxes onCFC usage, it has become imperative for theaerospace electronics industry to eliminateCFC usage wherever possible. CFC-basedsolvents are extensively used by the aerospaceindustry for defluxing electronic assembliesbecause of their compatibility with othermaterials and hardware, as well as their abilityto meet the stringent cleanliness requirementsmandated by the aerospace customers, inparticular, the Department of Defense (DOD).Because of the need for industry to complywith the new laws and regulations, a jointcommittee consisting of the Institute forInterconnecting and Packaging ElectronicCircuits (IPC). the DOD, and theEnvironmental Protection Agency (EPA) wasorganized to evaluate Alternatives to CFCsforPrinted Board Assembly Cleaning. Thiscommittee has established a three-phase CFCreplacement test plan for defluxing electronicassemblies.
The first phase establishes a benchmark of"how clean was clean" using CFC113/'methanol/nitromethane stabilized solvent.The second phase evaluates alternative
151
cleaning media for rosin flux. The third phaseevaluates alternative technologies, such aswater soluble and no-clean fluxes.
One major concern was identified whenHughes Aircraft Company reviewed theIPC/DOD/EPA Phase 1 Test Program. Theresults from this test program would notvalidate the reliability of many aerospaceelectronics industry designs. One key quotefrom the purpose section of the test plan was,"in addition, this testing program does notaddress where printed wiring board assemblieshave more demanding cleaning requirements;for example, where the spacing betweenconductors is significantly less than providedon the test board." Hughes, as well as othermembers of the aerospace electronics industry,decided that more testing would be required.
A three-phase program was designed and toevaluate alternative cleaning technologies togain customer acceptance for defluxingelectronic assemblies. Phase I consists of thedevelopment of a comprehensive test plan andthe design of two printed wiring boards typicalof complex aerospace configurations. Thisphase was co-fiinded by Hughes and theCalifornia South Coast Air QualityManagement District, the results of which arerepresented in this report. Phases II and IIIare proposed as follow-up efforts to performthe evaluations of alternative fluxes andsolvents. Phase II will establish theperformance of fluxes and solvents throughscreening tests established in Phase I. PhaseIII will involve full qualification testingnecessary to gain customer approvals based onthe results of Phase II.
LITERATURE AND INDUSTRY REVIEW
In order to ensure a complete review of theissues related to existing and developingelectronic defluxing technologies, amultifaceted approach was taken. First, anadvisory commit tee of Hughespersonnelconsisting of members from the sixdifferent Hughes Aircraft Company Groups
was established. Second, a group of industryadvisors was also formed. In both cases, thegroup charter was to focus on defluxingefforts within Hughes and throughout industry.A second facet was to contact several militarycustomers in order to determine their positionon defluxing issues, as well as theirrequirements. A third facet was to conduct anextensive literature review, including over 200articles. The three topics reviewed weresolvents and solvent development, fluxes, andindustry testing. The last facet was to contactother companies in the industry in order todetermine their progress in CFC eliminationfor defluxing applications.
Design Review of Aerospace Electronics
The objective was to review designrequirements in the aerospace industry (e.g.PWB configuration, materials andperformance) and compile a single set ofrequirements that would best representaerospace electronics. Internal and externalreviews were conducted to ensure that theneeds of aerospace companies would be met.
A review was conducted at Hughes todetermine the various design, material andperformance requirements. This informationwas vital to the understanding of therequirements and for determining theconfiguration of the PWBs that needed to bedesigned for future cleaning testing. It hasalso reaffirmed the need for a test vehicle withmore demanding cleaning configuration.
Eleven aerospace companies or affiliatesexternal to Hughes in Southern Californiawere invited to participate in the review c f thedesign, materials and performancerequirements identified by Hughes. Only fivecompanies participated: AeroJetElectrosy stems, Aerospace Corporation,Northrop - Hawthorne, Rockwell Rocketdyneand TRW.
Customer and Technical Requirements
The philosophy of this program was discussed
152
with the major military customers: the Navy,the Air Force and the Army technical researchfacilities. All three agencies voiced similaropinions on the use of alternative fluxes andcleaning media and were actively involved inthe joint IPC/DOD/EPA Ad Hoc WorkingGroup. All three agencies were actingindividually in regard to testing reviewsbecause there was no centrally identifiedagency that would speak for all branches.They all stated that as a minimum, theIPC/DOD/EPA test plans would have to beperformed in order to obtain material approvalfor alternative fluxes and solvents.Additionally, testing would have to beperformed on hardware configurationsrepresentative of the products to be cleaned.Finally, the consensus was that each individualcontract would have to be modified to allowalternative fluxes and solvents because blanketapprovals through changes to the existingmilitary specifications would not beforthcoming.
Solvents Literature Review
A literature search was conducted on cleaningsolvents and their associated equipment. Thealternatives to CFC cleaning were identifiedand divided into four categories - drop-in,near term, semi-aqueous, and aqueous. Eachsystem is defined as follows:
Drop in - A solvent that can be used inexisting equipment with only minor or nomodification to the equipment. Suchmodification might entail adjusting thethermostatic controls on the refrigeration coils(i.e., dilute CFC 113 and stabilized 1,1,1-trichloroethane (TCA), and TCA alcoholblends).
Near Term - A solvent that will be availablein the near future. It may be used in thecurrent process but will require newequipment or extensive modification toexisting equipment, such as the addition ofcolder condensat ion coils ( i . e . ,hydrochlorofluorocarbons (HCFCs).
Semi-Aqueous - In general, organic solventsthat are used to remove flux and then rinsedfrom the printed wiring assemblies (PWAs)with water (i.e., hydrocarbons, surfactantswith water rinse and isopropyl alcohol (IPA)and water).
Aqueous - Water with saponifiers and/orsurfactants, or just water (i.e., saponifier andwater, and water and surfactant).
There is not one specific material category thatstands out as "the" likely CFC replacement;however, a few solvents have passed theIPC/DOD/EPA Phase 2 testing, and severalmaterials are undergoing IPC/DOD/EPAPhase 3 testing.
Flux Literature Review
A literature review was performed to evaluatethe current research use and of water solubleand no-clean flux formulations by electronicmanufacturers and vendors. Both types offluxes are used; however, specific industriesconsistently use each type. The no-clean fluxformulations have been limited to commercialindustries, while water soluble fluxformulations are being used in thetelecommunications, commercial, industrial,military and aerospace industries. Theliterature reviewed suggested that the use ofno-clean flux formulations is limited becauseof its recent appearance into the flux market.The articles reviewed also suggested that bothtypes of fluxes needed to be furtherunderstood to predict long-term reliability.
During the literature review, it was found thata single data base was not available forcomparing the fluxes. In order to evaluate thevarious fluxes and match them to theircorresponding assembly application, a database would be beneficial. Numerous vendorswere contacted and requested to submittechnical data sheets and material safety datasheets (MSDS) for their available products. Acompilation of this data resulted in anextensive, detailed centralized flux data basecontaining over 100 materials available for
153
soldering electronic hardware.
Alternative Testing Review
An industry literature search was performed togather current and historical information onalternative cleaning methods and testing forrosin and non-rosin fluxes. This search wasaccomplished through the review of publishedarticles, attending technical conferences, andreviewing past technical reports and technicalconference proceedings. Very little data wasavailable and that most people felt thatimplementation of alternative cleaningtechniques or non-rosin fluxes would requirespecific testing depending on an individualcompany's hardware configuration, geographiclocation and financial ability.
In the past 30 years, there have been manytesting programs that have evaluated eitheralternative cleaning media or alternative fluxes(non-rosin based) with their associatedcleaning media. However, only three majormilitary funded programs were identified thataddress alternative cleaning media oralternative fluxes. The first was in 1966,sponsored by the Army and performed byRadio Corporation of America (RCA). Thesecond was in 1983, sponsored by the Navyand performed by the Navy's ElectronicManufacturing Productivity Facility (EMPF)at China Lake. The third program is theIPC/DOD/EPA Ad Hoc Working Group'sPhase 1, 2 and 3 testing, currently beingperformed by various government facilitiesand manufacturing companies.
The Army and Navy evaluations arrived atconflicting conclusions. The Army evaluationstated that water soluble fluxes (WSF's) couldbe used for military applications if specificprecautions were taken, whereas the Navyevaluation stated that WSF's could not be usedbecause of their devastating effects on surfaceinsulation resistance (SIR) and ioniccleanliness.
Hughes conducted a survey of aerospacecompanies to determine the industry usage of
non-rosin fluxes. Table 1 summarizes theresults of the survey by identifying thecompany, application, materials and processesused.
TEST PROGRAM - ALTERNATIVESOLVENT AND ALTERNATIVE FLUX
TESTING
The alternative solvent testing program(IPC/DoD/EPA Phase 2) for more demandingcleaning requirement configurations is theevaluation of replacements for CFC solventsand associated equipment. This is limited toelectronic manufacturing cleaning processesfor the removal of MIL-F-14256 rosin-basedfluxes. The testing includes:
a. Cleaning media screen testing -performed on glass slides.
b. Assembly screen testing cleaningmedia - performed on the IPC-B-36PWB and on more demanding cleaningconfigurations (both surface mount andthrough-hole) using modifiedIPC/DOD/EPA Ad Hoc WorkingGroup Phase 2 process flows,which utilized soldering processesother than the vapor phase and wavesoldering process. The quantity ofhardware required is not as extensiveas for customer qualification testing.
c. Customer qualification - the testingsequence is the same as that for theassembly screen testing with theexception of the statistically accept-able quantity of hardware.
The alternative flux testing (IPC/DOD/EPAPhase 3) is the evaluation of non-rosin basedfluxes and their corresponding cleaningsystems as replacements for CFC solvents.This is limited to electronic manufacturingcleaning processes. The testing includes:
a. Flux screen testing - characterizationof the flux based on material and
154
performance.
b. Flux/equipment screen testing - glassslide testing of the combinations ofalternative fluxes and their associatedcleaning systems.
c. Assembly screen testing of fluxes andequipment - performed on the 1PC-B-36 PWB and on more demandingcleaning configurations (both surfacemount and through-hole) usingmodified IPC/DOD/EPA Ad HocWorking Group Phase 3 process flowswhich utilize soldering processes otherthan the vapor phase and wavesoldering process. This testing willprovide preliminary test data.
d. Customer qualification - the testingsequence is the same as that for theassembly screen testing with theexception of the statistically acceptablequantity of hardware.
All pertinent travelers and process instructionswere developed and written for the differenceprocess flows and will be available throughthe SCAQMD. Upon the completion of thistesting ail IPC/DOD EPA Ad Hoc WorkingGroup Test requirements should be satisfied.This testing will also provide additional datafor the more demanding cleaning requirementconfigurations typically found on aerospacehardware.
PRINTED WIRING BOARD DESIGNS
Surface Mounted PWB Configuration
The HAC-SMT-615 PWB (see Figure 1) wasdesigned to be used as the test vehicle formore demanding cleaning difficulties typical ofaerospace Surface Mount Technology (SMT)component selection and placements. Thereare two unique features of the componentselection and placements. First, the same 68I/O LCC component that is used on the IPC-B-36 PWB has been designed into this board
to provide correlation to the IPC boards.Second, axial and chip capacitor componentsare placed around the perimeter of the 408 and224 leaded components tc simulate denselypopulated assemblies. Because of the tightcomponent spacings on many designs and theirpotential for inhibiting the cleaning of a PWB,these components were necessary. A list ofthe components along with their correspondingdescriptions can be found in Table 2.
There are a total of 36 SIR patterns; threesimulate the IPC-B-36 SIR patterns (0.006-inch line widths and 0.006-inch spacesbetween lines) and the remaining have 0.008-inch widths and 0.010-inch line spaces. Thereare also eight daisy chain patterns, utilizingcircuit board traces, which are for referenceonly. The boards shall be fabricated to therequirements of MIL-P-5511OD and madefrom standard epoxy material conforming toMIL-P-13949 (Prepreg, MIL-P-13949/12, andPlastic Sheet, MIL-P-13949/4) Type GF. Theboard thickness is 0.070-inch nominal withfour layers, utilizing one ounce copper. Toensure the quality of the PWB's, chemicalcharacterization and Group A testing per MIL-P-55110D shall be performed.
Through-Hole Mounted PWB Conflguration
The HAC-PTH-616 PWB (see Figure 2) isdesigned to be used as the test vehicle for themore demanding cleaning requirement platedthrough-hole (PTH) configurations andconsists of many different componentconfigurations. A list of the componentsalong with their descriptions can be found inTable 3. The selection of this design andthese components is intended to present amore demanding cleaning requirementconfiguration typical of aerospace platedthrough-hole designs. The SIR patterns,fabrication requirements, and materials are thesame as that for the surface mountconfiguration. The same chemicalcharacterization and Group A testing per MIL-P-55110D shall also be performed.
155
CONCLUSIONS
This program has provided a comprehensivetool to assist aerospace companies in theirquest to eliminate CFCs in electronicdefluxing processes. Conclusions that can bemade from this program are:
•There is a need for more complexPWB design test vehicles forcleanliness validation.
•The IPC/DOD/EPA testing programsprovide the prerequisite data requiredby the military customers. TheHughes/SCAQMD program, whichcovers more demanding cleaningrequirement configurations,parallels the IPC/DOD/EPA testing.
•Blanket approvals of materials inexisting military specifications seemunlikely.
•No one material (technology) standsout as "the" likely CFC replacement.Selection of the materials for acompany to pursue as a CFCreplacement is a very complexprocess. Most replacement materialsare ecologically better than CFC-113but still create environmental damage.In some cases, damage results fromsmog and, in others, from watershortages and waste water products.Each company will need to makeeducated decisions on what to pursuebased upon their product complexity,geographic location, and financialability.
•The detailed test program - includingthe artwork, travelers, and processinstructions-will be available throughthe SCAQMD.
DEVELOPMENT OF A SOLVENT DATABASE SOFTWARE PROGRAM
Ralph D. HermansenHughes Aircraft Company
El Sequndo, California
ABSTRACT
A novel computer database software programhas been developed to assist materialsscientists/engineers with decisions aboutsolvent replacement or reduction in high VOCproducts, such as cleaners, coatings, primers,impregnants, etc., by presenting comparativedata quickly and easily. The softwaredevelopment effort was jointly funded by theSouth Coast Air Quality Management Districtand Hughes Aircraft Company. The solventdatabase is unique in that it providesinformation not only about solvent physicalproperties, but also about environmentalcompliance status and safety characteristics aswell. Altogether, 27 fields are filled for eachsolvent record. Over 150 solvent recordshave been entered. The solvent database is astand-alone program, which runs on IBM PCsand compatibles. The program is menu drivenand is designed to be user-friendly. Helpscreens explain functions of the database at akeystroke. Solvents can be added, modified,or deleted from the database. The searchfunctions allow for combinations ofrequirements. The solvents found meeting thescreening criteria can be visually examined orprinted out. A manual was written to explainthe functions of the database. Hughes AircraftCompany intends to make the solvent databaseavailable to interested parties in the nearfuture.
BACKGROUND
The public is highly concerned aboutenvironmental problems and as a result, therehave been stricter and more numerousregulations issued affecting solvents andsolvent-containing products. Environmental
compliance from a laboratory standpoint ofteninvolves reformulation of coatings, primers,impregnants, cleaning solvents, and othersolvent-containing materials to eliminate,replace or reduce the proportions of offensivesolvents. However, the selection ofreplacement solvents can be a complex taskfor the env i ronmen ta l mate r ia lscientist/engineer because many other factorsbesides the specific environmental regulationmust be addressed. One such group of factorsto be 'dressed consists of those physical andchemical properties which caused the solventto work properly in the original formulation.Examples of such properties are chemicalfamily, evaporation rate, or solubilityparameter. Another group of factors includesflammability, toxicity, and other areas ofenvironmental compliance.
When e n v i r o n m e n t a l m a t e r i a l sscientists/engineers attempt to reformulatenoncompliant materials, budget and/orschedule restraints often prevent them fromdoing an in-depth search for replacementsolvents or to evaluate the resultant newformulation for a wide spectrum of concerns.Thus, a software program is needed to displaythe necessary solvent information quickly andeasily. Tne program should be user-friendlyand should run on IBM-type microcomputers.The user should be able to add, modify andremove solvent data and should be able toconduct searches based upon individualizedsearch strategies.
OBJECTIVE
The objective of the effort was to develop asolvent database software package, whichwould be readily available to environmentalmaterials scientists/engineers to assist them intheir reformulation assignments.
161
METHODOLOGY
The methodology followed in this effortconsisted of the following phases:
Phase 1 - Survey of EnvironmentalEngineers. In order to build a useful softwaretool, it was first necessary to interview thescientists and engineers who were working inthe field to determine what problems needed tobe solved and what information wouiu beprovided by the solvent database program.
Phase 2 - Literature Search of SolventProperties. Mailings and telephonecanvassing of solvent manufacturers wasconducted in order to obtain technical data onthe solvents. In addition, the HughesTechnical Library was used to obtain booksand periodicals on the topic of solvents.
Phase 3 - Database Design andProgramming. Many different propertiescould be included in the database, but aneffort was made to prune the list down tothose properties most needed and most useful.Two main groupings of fields were decidedupon: physical and environmental. Thesolvent database software program was writtenin dBase III Plus using the programmingcapabilities of the utility to design a user-friendly, menu-driven program.
Phase 4 - Entry of Solvent Data. Solventproperty values were taken from severalsources including manufacturer's literature,Material Safety Data Sheets, technicalhandbooks, and databases. The source ofeach data item is included in the database andcan be examined by the user.
Phase 5 - Completion of Prototype SoftwarePackage. A menu-driven program wasdesigned. Detailed coding was written to havethe program function in a coordinated manner.The dBase III Plus software from Ashton-TateCompany was used. It was compiled into aself-contained program using Clipper(Nantucket Software). A manual was writtento describe the solvent database software.
Phase 6 - Field Testing of PrototypeSoftware Package. The first version wasdistributed within Hughes for evaluation. Thecomments and suggestions from the reviewerswere noted and plans made for upgrading thesolvent database program.
Phase 7 - Incorporation of Improvements.The obvious "bugs" in the program werecorrected. In addition, the program was mademore user-friendly by adding help functions toexplain the various functions as they areencountered. This version was shared withother agencies that are pursuing similarefforts. Specifically, two agencies werevisited: Idaho National EngineeringLaboratory (INEL) and New MexicoEngineering Research Institute (NMERI).
Phase 8 - Final Release of SoftwarePackage. The completed software packagewas assembled and fifty copies produced.Thirty copies were distributed to concernedindividuals within the Hughes groups.Comments and suggestions from the userswere recorded.
ACCOMPLISHMENTS
Self-Contained Software. The solventdatabase program was designed to be a self-contained unit. That is, the program does notrequire that another software package beresident in order for it to run. Although thesolvent database program was written in dBaseIII Plus, the source code was compiled usingClipper. Henceforth, the software will run onany IBM PC or clone independent of dBaseIII. The package was designed so that theuser could add new solvents to the database,modify existing ones and delete obsolete ones.The user can also run searches and print outinformation.
Description of Database Format. There area multitude of attributes used to describesolvents. So, it was decided to restrict the listof attributes in the database to thosecommonly used in solvent comparisons. The
162
format of the database consists of threegroupings of fields as shown in Figure 1.Overall, there are 27 fields per record. Thefirst grouping is a description of the solvent.Fields are: solvent name, alternate name,tradename, manufacturer, and type. All ofthese fields are string type fields (i.e., stringof characters) except for tradename, which isa logical field (i.e., true/false). Type refers tochemical family or category for the solvent.The different categories are shown in Figure2.
The second grouping of fields is called"physical characteristics." The fields ;;e:specific gravity, flash point, freezing pcint,solubility parameter, evaporation rate,molecular weight, vapor pressure, low boilingpcint, high boiling point, and VOC. Althoughall of these fields appear to be numeric, theyare actually string fields, which can beconverted to numeric fields by the program asrequired. The function of these fields is self-evident by their name.
The third grouping of fields is called"environmental characteristics." The fieldsare: Rule 1401, AB2588 AI, AB258S All,VOC exempt, Prop 65, Rule 442, acute oraltoxicity, acute dermal toxicity, CAS number,inhalation PEL, inhalation STEL, and ozonedepletion factor. The first five of theseenvironmental fields holds a Y or N (yes orno) character. The rest of the fields are stringtype.
User-Friendly Software Environment. Thesolvent database program is intended for useby hundreds of different engineers andscientists. Therefore, the software was writtenwith "the user" in mind. First of all, theentire program is menu-driven for ease of use.Figure 3 shows the appearance of the mainmenu. From the main menu, one may gethelp, add new solvents, search through thesolvents, print out a report of the solvents, orquit the so'vent database program.
During the process of adding new solvents orsearching through the solvents, one may forget
the meaning of the fields as shown in Figure1. Another user-friendly feature is availableto facilitate its use. One merely places thecursor over the field of interest and pressesthe Fl key. A window appears and helpfulmessages about that field appear. During thesearch mode, helpful messages appear inwindows to let the user know what theprogram is doing or has done. For example,a window appears for each numeric searchfield selected and asks the user whether thefield value will be a maximum, a minimum,or a range. If it is to be a range, then thewindow message asks for the lowest andhighest values of the range. During thesearch, a message announces that the programis searching. Upon completion of the search,a message is given telling how many solventswere found meeting the search criteria. Themessage also tells the user how to view thecomplying solvents or how to print out thedata on them.
Listing of Data Sources. Some users willwant to compare the data very rigorously.That process may involve a greater knowledgeof how the data were generated and by whom.For each field value in the database, there is asource listed showing from where the valuewas obtained. In order to view these datasources, one goes to the search screen fromthe main menu, locates the solvent of interest,and presses "R" for reference. A windowappears as shown in Figure 4. The windowdisplays a listing of the data sources by fieldfor that solvent.
Solvents in Database. The solvents in thedatabase were selected so that fairly equalrepresentation exists for the different solventcategories (see Figure 2). The philosophyused for selecting solvents was to include both"good" and "bad" solvents. Here, goodsolvents would be solvents which are currentlyregarded as safe and environmentallycompliant. Bad solvents would have someobvious problem. The immediate temptationwas to list only the"good" or "recommended"solvents. However, with furtherconsideration, it was realized that these "good
163
or bad" ratings change dynamically with time.Solvents which were favored not long ago(i.e., methylene chloride or 1,1,1-trichloroethane) are slated for eliminationtoday. Therefore, it was decided to includeany common solvent.
Manual in Software Package. The manual,which is part of the solvent database package,provides the user with several services. First,the manual explains how to install the softwareprograms on your computer. Secondly, themanual describes the solvent databaseprograms. For example, it explains themeaning of the fields in the database. Itexplains how to add new solvents, deletesolvents or search for solvents which complywith your search strategy. Thirdly, themanual includes a listing of the solventproperties as the database would display them.This compilation of solvent properties is quiteuseful in itself.
Individualized Usage of Package. Each userof the software will want to adapt the softwareto his/her particular interest. This can bedone primarily by adding those solvents ofinterest to the user and deleting solvents of nointerest. The solvent data can t\; updated andthe source of the data included in the referencedatabase. Many hundreds more solvents canbe added. The storage capacity of thecomputer is the limiting factor.
CONCLUSIONS
A useful software package has been developedto assist scientists and engineers with decisionsconcerning solvent replacements. The solventdatabase program is a user-friendly, menu-driven system to be used on IBM PCcompatibles.
FUTURE WORK
A follow-on effort is planned to test andimprr.v? the solvent database. A special groupwill be selected to evaluate the software,wherein a diversity of company interests arerepresented. Modifications to the existingsolvent database should be made based uponthe prioritized comments and suggestions. Todate, it has been suggested that the programbe modified so that solvent blends can behandled by the database and that a version ofthe solvent database be written for Macintoshusers.
ACKNOWLEDGEMENTS
At Hughes Aircraft Company, Larry Lippassisted in design of the database format.Patrick Shuss did the majority ofdBase/Clipper programming and wrote themanual. Phyllis Kelleghan and Liza Julienresearched and entered the solvent data intothe database. In addition, Angela Chavez atINEL and Dr. Jonathon Nimitz of NMERIwere very cooperative in sharing their findingson solvent properties and offering commentson the Hughes/SCAQMD solvent database.
164
FIGURE 1
SOLVENT NAME:
ALTERNATE NAME:TRADENAME?:
PHYSICAL CHARACTERISTICS
SPECIFIC GRAVITY :FLASH POINT (C) :FREEZING PT (C) :SOLUB. PARAMETER :
Miscellaneous compound. Not covered inany other type.
166
FIGURE 3
H
A
P
QVersion 1.0 s]990
SOLVENT DATABASE UTILITY
MAIN MENU
GET HELP OR INSTRUCTIONS
ADD NEW SOLVENTS
PRINT A REPORT OF ALL SOLVENTS
QUIT THIS PROGRAM
Hughes Aircraft Company
FIGURE 4
PHYSIC
SPECIFIFLASH PFREEZINSOLUB.
ENVIRO
RULE 14AB2588AB2588
VOC EXEMPT'
SOLVENT: ACETONE
SPECIFIC GRAVITYEVAPORATION RATEMOLECULAR WEIGHTFLASH POINT <"F)LOW BOILING POINT i°C)HIGH BOILING POINT <°C)
FREEZING POINT I ' C IVAPOR PRESSURESOLUBILITY PARAMETERVOCOZONE DEPLETIONINHALATION PELINIHALATION STELACUTE ORAL TOXJCJTYACUTE DERMAL TOX1CITY
1ACUTE DERM <LD 50) OZONE DEPLETION |j
167
EVALUATION OF ALTERNATIVE CHEMICAL PAINT STRIPPERS1
Keturah Reinbold and Timothy RaceU.S. Army Construction Engineering Research Laboratory,
Champaign, Illinois
Ronald JacksonU.S. Army Toxic and Hazardous Materials Agency,
Aberdeen Proving Ground, Maryland
Ronald StevensonSacramento Army Depot, Sacramento, California
INTRODUCTION
Background
In 1984 the U. S. Environmental ProtectionAgency (U. S. EPA) placed a limit of 2.13mg/L on the allowable concentration of TotalToxic Organics (TTO) which can bedischarged from metal finishing operations.Some of the metal finishing operations find itdifficult to comply with this regulation whenusing the popular cold chemical paintstrippers. These strippers contain methylenechloride. Besides contributing to TTO,methylene chloride is a suspected carcinogenand is affected by restrictions on emissions ofair toxics under amendments to the Clean AirAct. Many of the strippers are also classifiedas hazardous wastes after use. Disposal of thepaint stripper wastes will become moredifficult and costly as many of these wastesare banned from land disposal under the U. S.EPA schedule.
Objective
The objective of this study is to identify paintstrippers which are operationally effective andenvironmentally acceptable replacements formethylene-chloride-based paint removers. Thegoal is to alleviate
1 This research was supported by the U.S. Army Toxicand Hazardous Materials Agency (P347.OJ. 16 Evaluationof Alternatives to Toxic Organic Paint Stnppers).
TTO compliance problems and to minimizeenvironmental and health risks and disposalliabilities.
Approach
To test operational success, alternative stripperformulations were first evaluated in thelaboratory. Materials meeting the criteriaestablished for success in the laboratory weretested on a pilot scale. The environmental,health, and safety aspects of the selectedstrippers were also evaluated to ensure that thecandidates are acceptable replacements. Thefinal step in this process is a full-scale fieldtest.
EVALUATION CRITERIA FOR ANACCEPTABLE PAINT STRIPPER
Criteria for a successful paint stripper weredeveloped in collaboration with SacramentoArmy Depot (SAAD). The following criteriawere selected: (1) acceptable stripping speed(SAAD upper limit of 2 hours), (2) effectivefor a broad spectrum of coatings, (3) notrapidly evaporated or depleted and easilyreplenished when it does become depleted, (4)no TTO contributing chemicals, (5)environmentally acceptable, (6) safe to use,(7) relatively easy to dispose. (8)commercially available, (9) easy to procure.
In addition, more specific criteria forenvironmental, health, and safety acceptability
169
of an alternative paint stripper were selected incooperation with SAAD. These criteriaincluded characteristics relating to toxicity(acute and chronic, human andenvironmental), environmental fate, safety(corrosivity, reactivity, and ignitability), andregulatory restrictions.
SELECTION OF COATINGS ANDSTRIPPERS FOR TESTING
Test coupons were prepared using three paintsystems: zinc-chromate alkyd primer with analkyd topcoat applied to aluminum, water-thinnable epoxy primer with a CARC urethanetopcoat on aluminum, and epoxy polyamideprimer with an epoxy polyamide topcoat onboth aluminum and steel.
Candidate replacement strippers were solicitedfrom industry and were previewed forprobabl- success before inclusion in the testprogram. Organic strippers may contain anyor all the following materials: (1) primarysolvents, (2) cosolvents, (3) activators, (4)retarders, and (5) surfactants.
MS-111, which conforms to Mil-R-46116(now cancelled), is a stripper which containsmethylene chloride, phenol, and formic acid.It was included in this study as a controlagainst which alternatives would be measured.Only one of the strippers evaluated containedmethylene chloride or phenol both of whichare TTO contributors. Common solvents inthe other alternative strippers include 2-(2-butoxyethoxy) ethanol, n-methyI-2-pyrrolidone, monoethanolamine, and aromatichydrocarbon solvents.
PROCEDURES AND METHODS
Stripper Performance
The laboratory evaluation of strippingperformance was based on a laboratory-scalemockup of a typical stripping process. Stepsin the test were immersion in stripper, caustic
dip for acidic strippers, water rinse, and steamcleaning. Thirty-two strippers were evaluatedfor stripping efficiency under controlledconditions with the four coating/substratec o m b i n a t i o n s . M a n u f a c t u r e r ' srecommendations for optimal operationalconditions were followed. Strippingconditions are listed in Table 1. Stripperswere evaluated for percentage of coatingremoved for each paint system at specific timeintervals.
Alternative strippers which performedadequately in the laboratory tests wereevaluated in a 25-gallon pilot test using depotparts rather than test coupons. Strippingtemperatures and dilution ratios with waterwere the same as in the laboratory tests.
Alternative strippers performing successfullyin the pilot scale evaluation were consideredfor full scale production use. Full scaleevaluation was performed during normalproduction operations at SAAD in a 1500gallon tank. Qualitative results as well asperiodic quantitative coupon analysis werereported. Quantitative coupon analysis wasconducted employing prepared specimenssimilar to those used in laboratory evaluations.
Environmental, Health, and SafetyEvaluation
After criteria for evaluation of environmental,health, and safety acceptability were selected,we developed a procedure to assign numericalratings to permit a quantitative comparison ofthe hazards associated with each stripper.Each criterion was scored for each componentof the stripper. A total weighted averagescore for the total stripper mixture wascalculated by summing the result of the scorefor each component times the percent of thatcomponent in the stripper. A total score foreach stripper was determined by summingcharacteristic scores. In addition, in order toevaluate the potential effect of a particularlyhazardous component, a worst case hazardscore was calculated for each stripper bysumming the highest score for a component in
170
each hazard category.
Several existing scoring procedures forindividual criteria (toxicity, bioaccumulation,flammability, reactivity) were used(1,2,3,4,5). If a score were reported in theliterature for a criterion for a strippercomponent, that score was used. In othercases, the same procedure was used to assigna score. If data for assigning a score werelacking, the values were calculated if possible.
RESULTS
Five of the alternative strippers removed allfour coating systems in the laboratory tests.Eight other products removed three of thecoating systems and partially removed thefourth in the proscribed 2-hour limit. Thestrippers evaluated in the pilot scale tests atSAAD were McGean-Rohco Cee Bee A-477,Fine Organics FO-606, Oakite ALM, Patclin103B, Patclin 104C, and Turco 5668. Itshould be noted that laboratory, pilot, andproduction scale tests overlapped to a certaindegree. The materials evaluated in laterphases were judged relative to all the materialstested up to that time. Thus, in some cases.materials with good performance were notevaluated in pilot-scale and production scaletests while lesser products were.
AH the strippers evaluated in pilot scale testsat SAAD except for Patclin 103B performed ata level consistent with laboratory test results.A GC-MS analysis indicated the presence ofchloracetic acid in Patclin 104C. This Patclinproduct contains glycolic acid which is madeby reacting NaOH with chloracetic acid.Incomplete conversion may have been thecause of the unacceptable chloracetic acidfound in the stripper. Further consideration ofPatclin 104C was withdrawn because of thechloracetic acid content. The remaining fourstrippers were considered to be candidates forfull-scale production tests.
Environmental, Health, and SafetyEvaluations
The results of the ratings of the mostpromising candidate strippers compared toMS-111 for environmental, health, and safetyhazard are shown in Figure 1. The higher thetotal score, the greater the risks. The totalweighted average scores show that any of thesix candidate strippers are less hazardous thanMS-111. Table 2 lists the order of relativehazard for the alternative strippers based onweighted average and worst case scores.Turco 5668 appears to be slightly preferablefrom an environmental and safety hazard pointof view. Hazardous waste concerns andenvironmental regulations restrictingdischarges are also much more favorable forthe candidate strippers compared to MS-111(Figure 2).
FIELD TESTS
Oakite ALM was evaluated in a full-scale fieldtest. Results were disappointing. Depletionof the active components, either throughevaporation or absorption in the tank, led toan early decline in stripping power. Thestripping time became excessive both forproduction parts and for uniform test coupons.Because of the nature of the productionoperations at the test site, the evaporation ratewas above an acceptable level. Themanufacturer analyzed the depleted stripperand provided a replenishment solution. Thereplenishment proved to be of marginal help.The full-scale evaluation of Oakite ALM wasdiscontinued after approximately 9 months.
Besides the stripper performance evaluation,environmenta! monitoring was done during thefield test with Oakite ALM. Results ofanalyses of stripper and rinse water revealedthe presence of a complex mixture of aromatichydrocarbon compounds besides thecompounds reported by the manufacturer ascomponents of the stripper. It is notsurprising that the manufacturer did not knowof the presence of these compounds in their
171
stripper. A petroleum distillate such as wasused in ALM is a complex blend ofhydrocarbons which is characterized primarilyby its boiling range. The identifiedcompounds increase the hazard and decreasethe environmental acceptability of the strippercompared to the estimates based on thecomposition listed by the manufacturer.
FO-606 was also evaluated in production-scaletests. Performance was initially acceptablewith strip times of less than 2 hours noted formost production parts. The evaporation ratewas quite high, and frequent additions of freshstripper were necessary to maintain asufficient depth in the tank. Stripperperformance declined over the duration of the12-month test. Significantly longer contacttimes were needed to remove the samecoatings. Epoxy coatings were especiallydifficult to remove.
CONCLUSIONS ANDRECOMMENDATIONS
Four candidate strippers demonstratedperformance in laboratory tests consistent withthe goals of the study. A total of six stripperswere evaluated on a pilot scale. Four of thepilot-scale candidates tested were initiallyrecommended for full-scale production use onan experimental basis. Oakite ALM wasevaluated in full-scale production and did notmeet the minimum requirements r.cr did itperform at a level consistent with laboratoryand pilot-scale tests. In addition, chemicalanalyses of ALM stripper and rinse watersamples revealed the presence of a complexmixture of aromatic hydrocarbon compoundswhich pose considerable health andenvironmental risk. FO-606 was alsoevaluated in production-scale tests. The userindicated initially acceptable performancewhich declined to marginai as evaporation ofthe stripper occurred. Material cost wasunacceptably high, especially considering therelatively high evaporation rate. Cee Bee A-477 and Turco 5668 have also beenrecommended for use on a trial basis. Oneuser has reported dissatisfaction in a trial with
Turco 5668, but the authors have notconducted a comparable field test to comparethat product with the two which have beenfully tested.
ACKNOWLEDGEMENTS
The authors wish to acknowledge thecontributions of S. Glascock, the technicianwho performed the laboratory strippingperformance tests, and P. Hoglund and G.Barrett, research assistants who compiled theenvironmental, health, and safety data.Assistance from the Sacramento Army Depotduring the project's planning, pilot testing,and field testing stages is also appreciated.
REFERENCES
1. Sax, N. I. Dangerous Properties ofIndustrial Materials, Sixth Edition.Van Nostrand Reinhold Company,New York. 1984.
3. Comprehensive EnvironmentalResponse, Compensation, and LiabilityAct.
4. Vector Scoring System for thePrioritization of EnvironmentalContaminants. Prepared by CanTcxInc. and Senes Consultants Ltd. andPriority List Working Group, OntarioMinistry of the Environment. March1988.
5. Reinbold K. A., and G. Barrett.Environmental Hazard Assessment ofChemical Paint Strippers. DraftReport. March 1988.
172
Table 1. Laboratory Tvating Baaulta
stripper stripping 0.5 hr 1.0 or l.s hr 2.0 hrTaaparatura
MS-111
FO-606
Patclin 104C(1:1 with water)
Turco 5668
Cee Bee A-477
Enthone S-26
Enthone S-26(1:4 with water)
Enthone S-26(1:9 with water)
Patclin 126
Ambient
82 C
82 C
82 C
82 C
Ambient
Ambient
Anbient
88 c
ABCD
ABCD
ABC
A
AC
A
A
CD
Patclin 106Q 82 C
Intex 879S 82 C
Patclin 103B 82 C(1:3 with water)
Ardrox 2 302
Patclin 125
Ardrox 5300-w
Oakite AUI
71 C
82 C
82 C
82 C
AD
AD
AC
ABCD
AD
A
AC
AD
AC
CD
AD
ACD
AD
AC
AD
Eldorado HT-2230 82 CEZE 570-81 71 C
AAD
AAD
ABCD
ABD
AC
AD
ACO
ACD
ACD
AD
AC
AD
AD
A
AAD
ABCD
ABCD
ABCD
ACDB-95%
ACDB-50%
ACDB-20%
ACDB-85%
ACDB-65%
ACDB-50%
ACDB-65%
ACDB-50%
ABDC-35%
ACDADC-30%
Onega R-824 71 C
Pentone R-3936 82 C
Pavco Decoater 82 C3400
Enthone S-26 Ambient(1:19 with water)
SafeStrip-66 54 C
Cham-Lube X-177 93 C
Chemical Solventa 88 CSP-825
Cham-Lube XH-36 93 C
Chemical SolventsSP-HNP
Key Chemical 570
NonMeth 120
NonMeth 140
Pavco Decoater3321
Turco 5555-B
NonMeth 161
Envirosolv L
Brulin SafetyStrip 1000
1 82 C
79 C
49 C
49 C
82 C
71 C
Ambient
Ambient
Ambient
Brulin Non-chlorinated 82
A
A
A
A
A
A
C
AD
AD
AD
AD
A
ADB-90%C-80%
ADC-5%
AD
ADB-401C-20*
ADB-10%
O15%
AD
AD
DA-40%B-60%C-55%
AD-90%
A
A
A
A
A
A
A-50%
A-30%
•Ailino-ohromata/alkyd(aluminum); Bi epoxy polyamida/apoxypolyamida (aluminum); Ct vator thinnable epoxy/ CMtc uxathane(aluminum); Di epoxy polyamid/apoxy polyamida (ataal).
Table 2. Total Stripper Hazard Scores from Worst to Best
AQUEOUS DEGREASING: A VIABLE ALTERNATIVE TO VAPOR DECREASING
J.T. SnyderMartin Marietta Astronautics Group
Denver, Colorado
INTRODUCTION
The aerospace industry has used chlorinatedhydrocarbons in vapor degreasers for over 35years to remove dirt, greases and oils. TCEwas used almost universally for a number ofyears as a degreaser solvent because it is avery aggressive cleaner with a high boilingpoint and will remove most oils, greases andwaxes. Replacement of TCE by 1,1,1 TCAbegan in the late 60's and early 70's when thehealth hazards associated with the TCE usagewere more clearly understood and publicized.Martin Marietta Astronautics Group (MMAG),a division of Martin Marietta Corporation,changed the solvent used in productiondegreasers from TCE to TCA around 1972.A number of companies continue to use TCEbecause changes are usually met with somekind of resistance and, for most vapordegreasing applications, TCE still performsbest.
The Montreal Protocol of 1987 and recentupdates of the Protocol are accelerating thephasing out of several CFCs primarily: R l l ,R125 113, 114, 115. Two chlorinatedhydrocarbons, TCA (trichloroethane) andcarbon tetrachloride, have recently been addedto the list. Martin Marietta's EnvironmentalManagement Task Force has directed thecorporate divisions to develop a plan to meetthe deadlines of the Montreal Protocol. Asearch for an available and an environmentallyacceptable solvent substitute which could meetour requirements was unsuccessful. Theperformance, cost and possible environmentalrisks of substitute VOC's directed us towardthe growing family of aqueous degreasers,which exhibit low safety and environmentalrisk potentials.
OBJECTIVES
The primary objective of this effort was toeliminate the usage of 1,1,1 TCA in ourproduction vapor degreasers with strongconsideration to be given to cost andperformance of a substitute material. Vapordegreasers are utilized in the chemicalprocessing area as a means of removing grossdirt in preparation for other chemicalprocessing. With assistance from materialsengineering and manufacturing, it wasdetermined that vapor degreasing was not anindispensable processing step. Aqueousdegreasing was a reasonable substitute. Alldetail parts and some assemblies processed inthe factory require at least one of thefollowing:
•Deoxidizing•Pickling and passivation•Etching for dye penetrant inspection•Zinc phosphate coating•Chromate conversion coating•Acid etch for bonding prep•Chem milling
All of the above processes were preceded byvapor degreasing and alkaline cleaning. Thevapor degreasing step was used in everyprocess to remove gross dirt before proceedingto alkaline cleaning. Because water is used inall processes following vapor degreasing, theaqueous solution doesn't pose a water problemon hardware.
An evaluation of the construction of our vapordegreasers revealed that conversion to aqueousimmersion tanks was both practical and costeffective. Stainless steel was used exclusivelyin the fabrication of the degreasers.
177
TEST PROCEDURES •Aluminum mill stamps
Six materials were selected as candidates forevaluation.
•Bioact EC7•Simple Green•Daraclean 282•Quaker 624GD•Turco 3878•Coors Bio T
These materials were then screened using thefollowing parameters:
•health hazards•treatability in our industrialwastewater treatment plant
•corrosion potential•performance
Material safety data sheets for each of thecandidate materials were evaluated by ourindustrial hygiene department to determinehealth or safety concerns. Coors Bio T andBioact EC7 were eliminated in this screeningbecause of flammability concerns. Treatabilitystudies of the remaining solutions eliminatedSimple Green. Long term corrosionevaluation on 2014-T6 aluminum was initiatedwith the three remaining products. A twomonth immersion test in 10% solutions of thethree materials revealed that Turco 3878produced visible corrosion, while the Quakerand Daraclean products did not.
The remaining two candidate materials,Quaker 624GD and Daraclean 282. weretested in a 10% solution at 130°F with mildagitation for cleaning performance on 2014samples coated with the following materials:
One cleaning material, Daraclean 282,remained after performance tests. Thismaterial was targeted for further testingrequired for approval and usage in amanufacturing process on productionhardware. 2014 aluminum was selected for allcorrosion and performance testing because ofits extreme susceptibility to corrosion.
PROTOTYPE TESTING
A small vapor degreasing tank, measuring 4ft. wide x 6.5 ft. deep x 12.5 ft. long, wasselected for prototype testing in the productionchemical processing area. Because the insideof the tank, boiling sump, and chiller coilswere heavily coated with chloride scale, sandblasting was selected to remove thecontamination. Sand blasting and saddle tankremoval were the only requirements forconverting the small vapor degreaser to anaqueous degreaser. The tank, constructed of1/4-inch stainless steel, was then filled to adepth of 64 inches with a 10% solution ofDaraclean 282 in tap water. An air agitationline was installed to enhance cleaningperformance and the tank lid was removed toaccelerate evaporation at 130°F. Parts wererinsed over the modified degreaser toeliminate dragout. Evaporation enhancementprovided enough water loss to preventoverfilling the tank. Engineering requirementsfor inclusion of aqueous degreasing into amanufacturing process dictated that the newsolution meet cleaning performance tests andthe salt fog test for chromate conversioncoated aluminum specimens specified in MIL-C-5541. Using vapor degreasing as a baselinecleaning performance tests indicated that theaqueous degreased specimens were visuallycleaner and exhibited a more nearly waterbreak-free surface than the vapor degreasedspecimens. Specimens processed throughaqueous degreasing before chromateconversion coating exhibited superior salt fogcorrosion resistance to those which were vapordegreased prior to chromate conversion
178
coating. CONCLUSIONS
RESULTS
Evaluation of test data and performance resultsby materials engineering and qualityengineering found the new material to besuitable for use in production operations. Anew manufacturing process entitled "AqueousDegreasing" is now used as an alternative tovapor degreasing for processing productionhardware. Following conversion of the largeproduction vapor degreaser (7ft wide x 14 ftdeep x 28 ft long) to aqueous completed inearly 1991, 1,1,1 TCA usage at MartinMarietta Astronautics Group was reduced byover 95%. Conversion included adding shearspray rinse, filter, oil skimmer, and pumpagitation. Transfer pumps, degreaser still andmotorized cover were eliminated. The saddletanks were sand blasted and cleaned forconversion to surge tanks. One of the saddletanks also contains the oil skimmer. Theaqueous degreasing process was later certifiedfor usage on all titanium, stainless steel andcarbon steel alloys.
•Performance of aqueoas degreasing issuperior to 1,1,1 TCA vapor degreasing.
•Manufacturing personnel readily accepted thechange tc aqueous degreasing because ofimproved performance and minimal safetyconcerns.
•No known environmental impact.
•Cheaper to operate and maintain than withTCA.
•Recyclable with simple filtration and oilseparation.
•Cost of degreaser conversion to aqueousdegreasing about 1/3 that of a new state-of-the-art vapor degreaser.
1.4
1.2
1
° 0.8
S. 0.6
I 0.40.2
0
Figure 1
1.21
TOXIC CHEMICALRELEASES
1987 1988 1989I
1990' 1991*
•ESTIMATED
179
CHEMICAL SUBSTITUTION FOR 1,1,1 -TRICHLOROETHANE AND METHANOLIN AN INDUSTRIAL CLEANING OPERATION
Lisa M. Brown and Johnny SpringerRisk Reduction Engineering LaboratoryU.S. Environmental Protection Agency
Cincinnati, Ohio
Matthew BowerAPS Materials, Inc.
Dayton, Ohio
INTRODUCTION
Passage of the 1984 Hazardous and SolidWaste Amendments (HSWA) to the ResourceConservation and Recovery Act (RCRA) of1976 has redirected the U.S. environmentalpolicy towards waste minimization to improvethe quality of the environment. In its efforts topursue the objectives set forth by Congress inthe HSWAs to RCRA, the USEPA hasestablished a national comprehensive pollutionprevention program. Implementation ofprojects to achieve several of the pollutionprevention program objectives is accomplishedthrough research conducted by the PollutionPrevention Research Branch of the RiskReduction Engineering Laboratory. Thisresearch addresses the intent of theAmendments to reduce the release andtransport of hazardous, toxic, and non-hazardous materials through the air, water andsolid media. The principal goal of thePollution Prevention Research Branch is toencourage the identification, development, anddemonstration of processes and techniqueswhich result in less waste being generated, inorder to promote a more rapid introduction ofeffective pollution prevention techniques intobroad commercial piactice.
1,1,1-trichloroethane (TCA) is used as a coldsolvent degreasing agent in many industrialdegreasing processes. In 1986 TCA wasidentified as a hazardous waste (F001) thatmust be managed under Subtitle "C" of theResource Conservation and Recovery Act. Asa result of this action, industries began looking
for ways to avoid the use of TCA cleaningsolvents. The EPA decided to target the metalfinishing industry for participation in a jointresearch project to examine the possibility ofsubstituting a terpene-based cleaner for TCAin degreasing operations. APS Materials,Inc., a facility in Dayton, Ohio participated inthe research project. APS Materials, Inc. is ametal parts finishing company which generatesTCA and methanol (hazardous waste F003)waste from cold solvent degreasing operationsassociated with their plasma spray depositionprocess.
PLASMA SPRAY DEPOSITIONPROCESS
The plasma spray deposition process hasemerged as a major means to apply a widerange of materials on diverse substrates. Thedeposition process is accomplished with theuse of a plasma gun. The plasma, generatedinside the gun, exits as a high velocity flamethrough the nozzle of the gun. A powderedfeedstock is injected into the flame via acarrier gas (usually argon). The injectedpowder accelerates, melts, and is carried atsonic velocities to the substrate on which theparticles impact and solidify rapidly, buildinga well-adhered protective coating.(l)
While APS Materials, Inc. employs thefundamental plasma spray deposition process,a few changes were made to betteraccommodate the plasma spray workperformed by their company. First. APS
181
Materials performs its plasma spraying in aninert atmosphere chamber. This is done forcooling and to prevent the titanium powderused in many of its coating applications frombecoming oxidized thus forming brittlecoatings. APS Materials also uses helium, inthe spray gun, as well as to adjust the heatlevel and arc length.
PROCESS DESCRIPTION
Original Process
In the APS Biomedical Parts Division, thecompany primarily coats cobalt/molybdenumparts and titanium parts with a porous titaniumalloy. In order to achieve a strong andadhesive coating, the parts were cleaned withTCA or m e t h a n o l ( T C A forcobalt/molybdenum and methanol fortitanium). TCA is more economical thanmethanol but weakens titanium over time.The cleaning process consists of several steps.The parts undergo visual inspection, tapemasking, and grit blasting. After the grit blasthas been completed, the masking tape isremoved. The part is then immersed in asmall pail containing TCA or methanol. Thepail is placed in an ultrasonic bath containingwarm water for 15 minutes. Solids from gritblasting, oil and grease from themanufacturing and handling of the parts, andany adhesive residuals from the masking tapeare removed in this cleaning process. WasteTCA and methanol were being generated atthe rate of 1/2 barrel per month each. Afterthe ultrasonic bath, a graphite maskingsuspension is applied to the part on surfaceswhere the plasma spray coating is not wanted.The part is then plasma sprayed and cleanedagain to remove excess titanium and thegraphite mask.
As a check system, APS runs small one-inchdiameter disks of the same composition as thepart to be coated - called "test buttons" -through the same cleaning and coatingprocess. The test buttons are placed on atensile strength testing machine which
measures the tension required to separate thecoating from the substrate as a quality controlmeasure.
Description of Initial Bench ScaleExperiments
For this test, DuSQUEEZE (DuBoisChemicals, Inc.) was the product used todetermine substitution feasibility.DuSQUEEZE is a blend of surfactantscontaining 25% limonene. Limonene wasselected as a possible substitute for TCA andmethanol because of its disposal qualities.Disposal of dilute solutions of DuSQUEEZEcould be accomplished by flushing it to asanitary or industrial sewer, according to localsewer use permit requirements. Thefeasibility of substituting a dilute, terpene-based cleaner (DuSQUEEZE) for TCA andmethanol was determined by assessing thetensile strength of the plasma coating bondsmade after cleaning with dilute DuSQUEEZEsolutions. Five tests were performed, four onplasma-coated test buttons to assess the tensilestrength of bonds made after cleaning with theDuSQUEEZE solutions, as compared to thetensile strength of bonds made after cleaningwith methanol and TCA, and one test todetermine if any limonene remained on thebuttons after being cleaned. In the frst test,four titanium test buttons were placed in astainless steel beaker containing a 20:1 dilutesolution of DuSQUEEZE and water. Thesolution was agitated for 20 seconds. The testbuttons were then placed in a stainless steelbeaker containing deionized (DI) water whichwas agitated for 20 seconds. The test buttonswere then blow-dried and plasma- sprayed.The tensile strength of the bond between theplasma arc coating and the substrate wasmeasured using a Tinius Olsen tensile tester.
In the second test, four titanium buttons wereplaced in an ultrasound bath containing a 50:1dilute solution of DuSQUEEZE for 10minutes. Next the buttons were placed in astainless steel beaker containing DI water for30 seconds. The titanium buttons were blow-dried for 60 seconds and then plasma-
iS2
sprayed. Then tensile strength of the bondswere tested in the same manner as the firsttest. The third test followed the sameprocedure as test two, using a 100:1 dilutesolution of DuSQUEEZE. In the fourth test,the buttons were cleaned by the same processas the third test, but the buttons were analyzedfor residual limonene and were not plasma-sprayed and tensile-tested. In the fifth test.cobait/molybdenum buttons were used insteadof the titanium buttons with the test protocolidentical to the third test.
Modifications to Existing System
APS purchased a heated ultrasonic bath witha timer for the conversion. However, whenthis ultrasonic bath malfunctioned, a heaterwas added to the old ultrasonic bath. TheTCA/methano! cleaning system did not requirea DI water rinse, so a DJ water system waspurchased along with a stainless steel bath andimmersion heater. With the new cleaningsystem, the parts took longer to dry, so a heatgun was purchased to speed-up the dryingprocess.
SAMPLING AND ANALYSIS
The overall purpose of the sampling andanalysis project at APS Materials was tosupport a pureiy qualitative judgement of thecleaning capabilities of the substitute cleaningsolution (i.e.. limonene). The sampling andanalysis protocol for this project was set up inthree parts; sampling spent solutions ofmethanol ai.j TCA, sampling the terpene-based cleaning solution after modificationswere made to the cleaning system, anddeveloping data for a comparative analysis ofplasma-coating bond strengths between thecoatings of test buttons that were cleaned withmethanol/TCA prior to coating and thecoatings of test buttons that were cleaned withthe terpene-based solution prior to coating.
Pre-Modification Sampling
The first part of the sampling process was
performed prior to any modifications. Thissampling was performed in order to determinethe type and amounts of contaminants found inthe cleaning solvents. Samples of themethaiiol and TCA cleaning solutions weretaken and analyzed for oil and grease,dissolved solids, suspended solids, titaniummetal and cobalt metal. This sampling alsoestablished the baseline performance formethanol and TCA. The samples were takenby mixing the material in a plastic bucket andthen pouring a sample from the bucket througha glass funnel into a glass bottie. The dataderived from this sampling served as a benchmark for the ensuing substitution sampling.
Post-Modification Sampling
The second part of the sampling scheme wasperformed after the modifications were madeto the system in order to determine theeffectiveness of the terpene-based solvent incleaning the parts. Sampling of the cleaningsolution was performed throughout a typicaloperating cycle. Samples were recovered atthe beginning of a bath cycle (i.e.. when thetank contents were completely replaced withfresh cleaning solution) to establish baselineconcentrations. A second sample was takenmidway through the effective life of thecleaning solution, A final sample wasrecovered prior to removing the spent solutionfrom the dip tank.
In addition to taking samples of the cleaningsolution, wipe samples were taken of thecleaned parts. Wipe samples were taken toevaluate the cleaning efficiency of the solutionover time by analyzing for residualcontaminants (oil and grease) on the parts.One wipe sample was taken from the cleanedmetal parts during each sampling interval todetermine if there was a residual of oil andgrease. The wipe sample was performedusing sterile, uncontaminated clcth. Sterilegloves were worn to prevent contamination ofthe cloth with oil and grease. The wipingprocedure was consistent for each sample. Aglass container of sufficient volume was usedto hold the cloth after sampling. Three wipe
183
samples were taken over the life of thelimonene c>.aning solution, to coincide withthe three liquid samples described above.
Analysis for metals was performed usinginductively-coupled plasma atomic emissionspectroscopy (ICP). Oil. grease anddissolved/suspended solids were analyzedusing gravimetric analytical techniques.Spikes and replicate analyses were also doneto check for accuracy and precision and toidentify the presence of any matrix effectsassociated with sample preparation ormeasurement. Data were then combined andstatistically evaluated.
The analysis of plasma-coating bond strengthcompared current data collected by APSMaterials regarding the strength of coatingsapplied after parts were cleaned with dilutesolutions of DuSQUEEZE and historical dataof bond strength resulting from parts cleaningwith TCA and methanol. Data generated twomonths prior and two months following theconversion to the limonene solution were usedfor this comparison.
RESULTS AND DISCUSSION
Bench Scale Experiments
The before and after tensile strength resultswere comparable. Overall, the bondingstrengths were actually slightly better for thedilute limonene cleaner (see Table 1). Noresidual limonene was detected (detection limit1 ppm) for cleaner at 100:1 dilution.
Analyses For In-piant Operations
The initial tests for contaminants in methanoland TCA used for cleaning yielded the resultsshown in Table 2. The samples for thesearalyses were taken when the baths wereconsidered spent, just prior to being dumped.
The amounts of oil and grease found in thewipe samples, shown in Table 3, were verylow at about 1 mg or less. The increase in oil
and grease from the bath dump, as comparedto the fresh bath, was very small for onesample and was less than the fresh bath in thesecond bath dump sample. This latter resultcould have resulted from the wipingtechnique. The parts seemed to be cleanedjust as well at the time the bath is dumped aswhen the bath is fresh.
Table 4 shows results from GC/MS method8270 (SW-846) analyses for residual limoneneon the parts. Limonene was not detected inthe rinse samples, thus indicating that all ofthe limonene was removed during dragout andsubsequent drying of the parts.
The results in Table 5 indicate that dissolvedsolids and oil and grease were much higher inthe fresh bath and the bath used to clean partsonly prior to plasma spraying (Dump#l), thanin the bath used also for cleaning after plasmaspraying (Dump#2). The reverse was true forthe suspended solids. The graphite in the bathmay affect the DuSQUEEZE cleaning solutionto create these differences.
A comparison of the DuSQUEEZE cleaningsolution with the previous methanol and TCAsamples, reveals that the oil and grease levelsin the DuSQUEEZE are much lower than theother cleaning solvents. Suspended solids forthe DuSQUEEZE are lower than the previoussolvents, except for the sample containinggraphite, which is roughly equivalent.Dissolved solids for DuSQUEEZE are muchhigher than the other solvents. The higherdissolved solids may reflect the fact that theDuSQUEEZE is an emulsifying agent whichconverts the oil and grease to dissolved solids.This would explain the lower oil and greaselevels for DuSQUEEZE.
Although the data generated by the samplingand analysis program, shown in Table 6,indicates that the terpene-based cleaneradequately cleaned the parts for this process,since wipe samples were not taken for theoriginal process, a statement of comparisonbetween the former and present cleaningtechniques is not feasible.
184
Economic Analysis
Although the old ultiasonic bath was in use atthe time of the test, economic analysis isshown for the system that APS is nowoperating. A summary of the economicanalysis is found in Table 7.
CONCLUSIONS
In summary, it has been determined thar aterpene-based cleaner can adequately cleanmetal parts without adversely affecting theperformance of the plasma-arc coatingapplication. The use of a terpene-basedcleaner in place of methanol and TCA. hasproven to be an environmental and economicsuccess. Elimination of the disposal problemsassociated with methanol and TCA coupledwith the maintenance of plasma-arc coatingquality, make the use of terpene-basedcleaners more attractive than other plasmaspray coating processes, as well as other metalcleaning/coating operations. The annual costsavings, as well as the short payback period,also make the cleaner attractive from aneconomic standpoint.
REFERENCES
1. Herman, Herbert, "Plasma SprayDeposition Processes", MRSBulletin, p. 60 - 63, December 1988.
2. Test Methods for Evaluating SolidWaste. SW-846, Third Edition, U.S. Environmental Protection Agency,November 1986.
3. Environmental Monitoring andSupport Laboratory. Methods forChemical Analysis of Water andWastes. EPA-600/4-29-020, U.S.Environmental Protection Agency,Cincinnati, Ohio, March 1983.
185
TABLE 1. TENSILE STRENGTH TEST RESULTS FOR BENCH SCALEEXPERIMENTS
Test 3uttons Cleaning TensileAgent Strength (psi)
titanium methanol 6300+/-1260
titarr.um DuSQUEEZE* 7000 + /-570
cobalt/TT.olybdenum TCA 5150+/-1990
cobalt/molybdenum DuSQUEEZE* 5400+/-1290
* Tensile strengths neasured for test button cleaned withvarious dilutions of DuSQUEEZE showed no trends or statisticaldifferences, so values shown include all measurements.
TABLE 2. RESULTS OF ANALYSES OF SOLVENT SAMPLES FOR CONTAMINANTS
ALTERNATIVES TO CFCs IN PRECISION CLEANING:A NEW HCFC BASED SOLVENT BLEND
R.S.Basu and P.B.LondonEngineered Materials Sector
Allied-Signal Inc.Buffalo Research Laboratory
Buffalo, New York
and
E.M.Kenny-McDermottGuidance Systems Division
Allied-Signal AerospaceTeterboro, New Jersey
INTRODUCTION
Recent findings of ozone layer depletion hasprompted United Nations EnvironmentalProgram (UNEP) to amend the MontrealProtocol, for an accelerated phase-out ofchlorofluorocarbons (CFCs), carbontetrachloride and halons by the year 2000 andphase-out of methyl chloroform by the year2005. In addition, U.S. Congress has passedthe Clean Air Act which puts similar phase-out dates for CFCs and methyl chloroform.Because of these regulations, a search forreplacements for all the regulated CFCmolecules has been intensified. In addition,global warming is also emerging as anothermajor environmental problem making thesearch more and more complex.
In this article, we are going to discuss theperformance characteristics of a newstratospherically safe alternate totrichlorotriflucroethane (CFC-113) as aprecision cleaning solvent. The alternates arebased on hydrochlorofluorocarbons, theselected ones are 1,1-dichloro-l-fluoroethane(HCFC-141b) and 1,1-dichloro- 2,2,2-trifluoroethane (HCFC-123) with blendsbased on these compounds. These compoundsare formed by the addition of hydrogen toCFCs to make them shorter lived in thetroposphere and therefore less harmful to the
stratospheric ozone. The Montreal Protocolhas also put restrictions on the use of thesesubstances because of their low but non-zeroozone depletion potential. HCFCs areconsidered interim replacements with phaseoutdates starting from 2020 and completephaseout no later than 2040. The U.S. CleanAir Act assigns earlier phase-out dates forHCFCs, starting with a freeze in production inthe year 2015 with complete phase-out by2030.
SOLVENT SELECTION
In this section we are going to discuss how thesolvent for precision cleaning application isselected. Fre"ision cleaning encompassesmanufacturing aerospace components,gyroscopes in missile guidance systems,medical devices, computer disks, siliconwafers, etc. These parts to be cleaned containvarious metals and plastics and CFC-113 is theuniversal choice in these applications.CFC-113 is non-flammable, non-toxic, stable,has good solvency characteristics and iscompatible with these materials.
Solvent selection depends on a number ofenvironmental factors, along with the ozonedepletion and greenhouse warming potentialmentioned before. The potential ofcontributing to smog in the lower atmosphere
189
and polluvion to ground water by way of watereffluents are also important considerations.The objective of solvent selection is not totrade one environmental problem with another.There are federal and state laws prohibitingthe emission of volatile organic compounds(VOCs). These materials break down in thelower atmosphere and contribute to smogpollution. Over and above these, themolecules have to be relatively non-toxic sothat the workplace remains relatively safe.Effluents from plants may also pose a problemwhich is of a lesser concern for solvents thanfor aqueous systems. The energy requirementis in general lower for solvent cleaning ascompared to aqueous cleaning. However,energy consumption also has to be consideredin detail because of its contribution to thegreenhouse effect.
Among hydrochlorofluoroethanes HCFC-123(1,1-dichloro-1,2,2-tri-fluoroethane) andHCFC-141b (l,l-dichloro-l-fluoroethane),with boiling points between 80 and 90 F, arebeing considered as alternates to CFC-113 insolvent applications. The rest of thehydrochlorofluoroethanes are not found to besuitable because of their toxicity, flammabilityor other undesirable characteristics. Thesetwo HCFCs are far less ozone depleting thanthe current CFCs. Molecular structure of thesecompounds are shown in Figure 1. Thelifetimes, ozone depletion potentials (ODP) ofthese chemicals and also the greenhousewarming potentials (GWP) are given in Table1. The ODP and GWP are measured relativeto CFC-11 which is assigned a value of 1.0.The ODP and the GWP numbers for themolecules are still being revised by AFEASand values are expected to be finalized by theend of the year. These two HCFCs are notconsidered VOCs by US EnvironmentalProtection Agency and therefore, they do notcome under that regulation.
Our tests have shown that since HCFC-141bor HCFC-123 alone are not equivalent toCFC-113, a blend of the two is required.HCFC-123 lowers both the ODP and theflammability characteristics of the blend.
HCFC-123 is more aggressive towards theplastics and the presence of HCFC-141bmakes the blend more compatible to plastics.So blends are preferred to the purecomponents both from an environmental andperformance standpoint.
The preferred blend for precision cleaning isa new azeotrope-like non-segregating blendconsisting of HCFC-141b and HCFC-123. Thecomposition of the blend is 80 percent byweight HCFC-141b and 20 percent by weightof HCFC-123. Some of the physical propertiesof the blend as compared to CFC-113 areshown in Table 2. The blend is commerciallyavailable from Allied-Signal under thetradename Genesolv*2020.
As we compare the properties we find that themajor difference in physical propertiesbetween the CFC-113 and HCFC based blendis their boiling point. This would normallyindicate that Genesolv 2020 would evaporatefaster than CFC-113 resulting in highe: lossrates of the solvent but in reality theevaporation rates of the solvents at roomtemperature are equivalent. This is due to thefact that Genesolv 2020 has higher heat ofvaporization. This property makes it anacceptable substitute for CFC-113 despite itslower boiling point.
Another important use of CFC-113-basedblend is in printed circuit board cleaning. Theazeotrope- like blend containing HCFC-141b,HCFC-123 and methanol marketed byAllied-Signal under the tradename Genesolv2010 is the solvent of choice for thatapplication. We are not going to discuss thatapplication in this article. However, wewould like to refer the readers to several otherarticles detailing the use of this blend indefluxing [2,3,4].
SOLUBILITY AND CLEANLINESSSTUDIES
In selecting a new solvent, solvency andboiling point are chosen as major physical
190
properties. The solvency oi' various light oilsis determined by measuring their solubility inthese solvents. A boiling point range of 25° to75C has been chosen as the range for thesesolvents and various solubility models are usedas a tool to select solvents on the basis ofsolvency. Cleaning tests were performed tofinalize the selection.
In the cleaning tests, metal coupons are soiledby various types of oils and heated to 200°F.This is done to partially simulate thetemperature attained while machining andgrinding in the presence of these oils. Themetal coupons thus treated, are degreased in avapor phase degreaser machine. The couponsare held into the vapor and are vapor rinsedfor a period of 15 seconds to 2 minutesdepending upon the oils chosen. A short timeperiod is selected so that the solvents can becompared easily. The blend mentioned here iscompared to CFC-113 in cleaningperformance. Cleanliness of the coupons aftercleaning is determined by carbon coulometerwhich detects amounts of carbon physicallyadsorbed on the surface of the metals. Sincemost oils are hydrocarbon-based this methoddetects the amount of hydrocarbon left on thesurfaces.
The overall cleanliness test results withvarious types of oils are shown in Table 3.The results show that the performance ofazeotrope-like blend Genesolv 2020 is equal toor better than CFC-113 in cleaning variousoils. The cleaning results shown in Table 3are in percentages of oils removed fromvarious substrates.
PRECISION CLEANING STUDY OFGYROSCOPE COMPONENTS
High precision gyroscopes contain somemechanical assemblies with tolerances on theorder of 1-5 micrometers. As a result, theyare extremely sensitive to paniculatecontamination as well as to very low levels offoreign fluids and polymeric films. Accuracy,precision and reliability of gyroscopes used for
military and aviation applications depend on acleaning precis' being able to remove thesedeleterious contamiiirnts without degrading thematerials of construction. Gyroscopecomponents are fabricated using a variety ofmaterials, including light metals (i.e.beryllium and aluminum), plastics, adhesivesand elastomers. CFC solvents and blendshave been eminently suited for thisapplication. Finding a suitable alternativecleaning agent or process will be a difficulttask. Extensive material compatibility andcleaning efficacy studies must be conducted onany new solvent system to ensure productintegrity. Beryllium metal parts preclude theuse of aqueous based cleaning or 1,1,1-trichloroethane which hydrolyzes easily andmay cause corrosion. Also, any new solventsystem must not attack the polymeric materialsin the device.
Present cleaning processes include vapordegreasing, pressure cooking, soxhletextraction, cold spraying in enclosed booths,cold cleaning in ultrasonics and gyroscopeflushing with CFC solvents and blends. Partsand sub-assemblies are typically cleanedseveral times during assembly in order toremove handling and processing contaminants.Typical soils include particulates, silicones,hydrocarbons, finger oils, bromofluorocarbonbalancing fluids, flux and excess adhesive.
This work examines a few of the aspects oftesting the new HCFC solvents replacementfor CFC solvents for precision cleaning ofgyroscope parts and assemblies. It is only thebeginning of the extensive research programrequired to implement these new solvents intothe manufacturing mainstream. We are goingto talk about the cleaning studies in thissection. Cleaning efficacy is determined bycomparing the HCFC cleaned surface to onecleaned with CFC using Fourier TransformInfrared Micro and Reflectance Spectroscopy,water break test or weight change of thecoupon.
Fourier Transform Infrared reflectance spectraof the surfaces of the coupons cleaned in
191
CFC-113 and Genesolv 2020 are shown inFigures 2 and 3. The spectra show peaks ofvarious heights indicating materials left onstainless steel coupons after cleaning withCFC-113 and Genesolv 2020. The soils usedare phenyl methyl silicone and silicone grease.In case of both of these soils, couponsdegreased in Genesolv 2020 show fewer peaksindicating better cleaning compared toCFC-113.
TOXTCITY AND MATERIALCOMPATIBILITY
Introduction of new compounds in themarketplace requires extensive toxicity studies.Presently both HCFC-123 and HCFC-141b areundergoing thorough toxicological studies. AProgram for Alternative FluorocarbonToxicity (PAFT) has been set-up by aconsortium of a large number of current CFCmanufacturers all over the world. Repeateddose toxicity studies have been completed andno significanct adverse effects have beenfound. Allied-Signal has been givenpermission by the US EnvironmentalProtection Agency to manufacture and sellHCFC-141b for solvent use. Of the twoHCFCs, HCFC-141b has been found to beless toxic. At this point with availableinformation, an interim PEL of 500 ppm isassigned to HCFC-141b and 50-100 ppm isassigned to HCFC-123. These are interimvalues, final PELs are expected to be assignedfollowing completion of the chronic studies,which are underway.
Due to lower boiling points, loss of HCFCsolvent vapors from older degreasers may betoo great for reasons of health and safety andalso effect the processing costs. This requirestighter or low emission machines. Existingequipment should, therefore, be evaluated inthis respect and be either upgraded to newequipment, or perhaps, be retrofitted tomaintain safe and healthy work environment.
Hydrolytic stability of the solvent blend is
tested by refluxing the solvent in the presenceof various metals and water for a two (2)week period. The solvent even without anystabilizer, showed excellent stability inpresence of water. Its stability compares verywell with CFC-113 and it seems far superiorto 1,1,1-trichloroethane. It appears thepresence of a fluorine in the molecule andstronger carbon-fluorine bonding has increasedthe stability of HCFC-141b over1,1,1-trichloro-ethane. One important pointto note is that, HCFC-141b, HCFC-123 andGenesolv 2020 are all compatible withberyllium metal which makes the solvent blendextremely attractive to clean gyroscopes. Asmentioned before, various components in agyroscopes are fabricated using beryllium andits alloys which makes beryllium compatibilitya very important factor in solvent selection.
In commercialization of a new product forprecision cleaning another important propertyto study is its compatibility with variouselastomers. These elastomers are present invarious components cleaned. The elastomercompatibility is done where the elastomers arerefluxed in solvent for a two week period.The solvent blend seemed to have reasonablecompatibility with a large number ofelastomers. A detailed account of materialcompatibility appears on references [5,6] andis also available from Allied-Signal uponrequest.
CONCLUSIONS
In this paper we have shown that a substitutefor CFC-113 has been found with a muchlower ozone depletion potential for use as asolvent in precision cleaning applications.This is an azeotrope-Iike blend ofHCFC-141b, HCFC-123 (Genesolv 2020) andis presently in the process of commercializa-tion. Solubility measurements with variousoils have shown that this blend has a verygood potential to be used as a solvent for theseapplications. This is further confirmed bymetal cleaning studies where the blend showedequivalent or better performance compared to
192
CFC-113. Finally application results weredone to demonstrate that Genesolv 2020performs very well in field applications ofcleaning gyroscope components. In thisapplication FTIR reflectance spectroscopy hasbeen used and is found to be an excellentmethod for cleanliness measurement.
The solvent is currently being produced inpilot quantities in the pilot plant working atBuffalo and is marketed under the tradenameGenesolv 2020. The actual commercialproduction will start by the end of 1991 or byearly 1992.
REFERENCES
Montreal Protocol. Final Act 19S7,UNEP. 1987; London Amendments1990.
R . S . B a s u . E . L . S w a n andM J. Ruckriegel.' 'CleaningAlternativefor 1990s and Beyond: HCFC BasedSolvent Blend and Low EmissionEquipment". Proc. of NEPCON EAST90, pp. 161-178.
R .S .Basu and J . K . B o n n e r ,"Alternatives to CFCs: New Solventsfor the Electronics Industry", SurfaceMount Technology, Dec. 1989, pp.34-37.
J.K.Bonner, "Solvent Alternatives forElectronics for the 1990s", Proc. ofNEPCON EAST 90, pp. 189-202.
J.K.Bonner. "Solvent Alternatives forthe 1990s", Proc. of NEPCON WEST90. pp. 1601-1608.
R.S.Basu. K.D.Cook and E.L.Swan,"Stability and Compatibility of HCFCBased Solvent Blends for Defluxing",Proc of NEPCON WEST 91, pp.926-931.
193
Table 1: ODP and GWP Values for Selected Compounds
Solvents
CFC-11CFC-113
HCFC-123HCFC-141b
Formula
CCljFCXI3F3
CHCUFjC,H3C1,F
B.P.fF)
8289
75118
Lifetime
1.510
60102
ODP
0.020.12
10
.0
.8
GWP
1.01.3
0.020.12
Table 2: Physical Properties Comparison
Genesolv 2020
Ozone Depletion PotentialGreenhouse Warming PotentialFlash PointBoiling Point (=F)Liquid Density(g/cc)Kauri-Butanol ValueSolubility Parameter(caJ/cc)'-Evaporation Rate(ether = l)Surface Tension (dynes/cm)Heat of Vaporization (Btu/lb)
Table 3:
OIL SOLVENT
Petroleum Based CFC-113Genesolv'2020
Semi-synthetic CFC-1Genesolv'2020
Synthetic CFC-113Genesolv'2020
0.10.1
None87.81.3858
7.61.218.491.0
CFC-113
0.81.3
None1191.5631
7.31.2
17.863.1
Cleaning Results
Aluminum
100.088.8
13 94.599.2
19.540.1
% Oil Removed
SubstratesStainless Mild
Steel Steel
100.097.9
94.599.2
14.638.1
100.098.3
98.399.0
7.654.4
194
HCFC Solvents Comparison
CFC 113
Methyl Chloroform1,1,1-Trlchloroethane
HCFC 141b
HCFC 123
Fi
Ci-c-ii
FCl1
Cl-C-1
ClCl1
Cl-C-1F
F1
F-C-
Cli
1C-F1
Cl
HIC-H1H
H1
C-H1H
Cl1
C-CI
F H
Fig. 1
SS3O3 STRIPS AFTER CLEANING
FHENYLMETHYL SiLlCONE OIL (500 CS) RESIDUE
ABS = C,00232
'^4/^^^^^^^HCFC
AES = C 0
i !
1 \
4000 3500 3000 2500 2000 1 5C(
Fig. 2
195
SS303 STRIPS AFTER CLEANING
SILICONE GREASE RESIDUE
CFC
35G0 3000 2500 2000 1500
WAVENUMBERS (CM-1;
rig. 3
1 00
196 I
Section III
SOLVENT RECOVERY AND RECYCLING
THE SUCCESSFUL IMPLEMENTATION OF A SOLVENT RECOVERY PROGRAM
Marcanne Lynn BurrellWaste Minimization Company
Bellevue, Washington
ABSTRACT
This paper provides a step-by-step approachfor obtaining the technical background andprogram support necessary for the successfulimplementation of a solvent recovery programbased on an existing program at a BoeingAerospace and Electronics (BA&E) site in thePuget Sound Area.
INTRODUCTION
It has become apparent that many of thesolvents being used in our manufacturingprocesses today are not only harmful to theenvironment, but to humans as well.Presently, there are two options to choosefrom when addressing the problem ofenvironmentally harmful solvents. The first,and more preferred option, is to replace thesesolvents and paints with substitutes which areas effective, but are not as harmful (eg.,replace organic-based materials with aqueous-based materials). Unfortunately, substituteshave not been developed for all processes andmay not be for many years. In the interim,harmful chemicals are being released into theatmosphere and disposal costs for these spentsolvents are escalating. Until substitutes areavailable, recycling remains an option thatminimizes the generation of hazardous wasteand results in considerable cost savings
PROCEDURE
Step One - Research Background
The first step is to determine the solvent
which will best suit your company's needs andevaluate the technical and economic feasibilityof recycling this solvent.
Conduct a shop survey. Find out whichsolvents the shop(s) are currently using. Askthe following questions:
• Would the shop(s) prefer using asubstitute that works better, but maynot be economically feasible at thistime?
•Would it be feasible to change to themore desirable solvent if it was beingrecycled?
• If the shop(s) are generating differenttypes of solvents, are they willing andable to switch to an universal solvent?(This will avoid the generation ofsmall quantities of different types ofspent solvents).
Review historical records (i.e., purchasing anddisposal records). This data will help todetermine what quantities were consumed andto illustrate trends. Historical records can alsobe used to verify the information supplied bythe shops and other personnel.
After the type of solvent(s) to be recycled hasbeen determined, obtain information onrecycling equipment that will be used.Perform a literature search. Interview vendorsand survey other organizations and companiesthat have processes similar to yours to find outif they already have a program in operation.Capitalize upon their past experience. Thiscan save time and money by preventing futureproblems and duplication of work.
197
Investigate the regulations and restrictions thatmay effect this program. This should be doneat the federal, state, local and company levels.Agencies at the federal level irelude theEnvironmental Protection Agency (EPA),Occupational Safety and Health Administration(OSHA) and National Fire ProtectionAssociation (NFPA). The state regulationsmay be more stringent than regulations on thefederal level. Local restrictions would includecity and county codes enforced by the firedepartment, building inspectors, the local airand water regulatory agencies and localtransportation criteria. These codes mayrequire permits for the equipment, thelocation, transportation of the materials andbuilding additions or modifications. Largercompanies may have their own firedepartment, fire protection engineering,safety, industrial hygiene and other governingorganizations that will have some prerequisitesfor this program. Be sure to consult theseorganizations early and keep them informed.
A pilot scale study is recommended. This canbe accomplished by renting, leasing orborrowing equipment from a vendor forseveral weeks to months depending upon thecomplexity and magnitude of the program.The pilot study will test the logistics of theentire system, determine the division of labor,and identify and troubleshoot any unexpectedproblems prior to the purchase of equipment.Samples should be taken during the pilot studyto verify the quality of the recycled solvent.The pilot study will indicate the size ofequipment necessary for full scale operation.When sizing equipment, it is recommendedmat the equipment be oversized to allow foradditional streams in the future, errors in dataand downtime due to maintenance and repairs.
Step Two - Develop Support
The second step is to develop support for theproject to ensure its success. First, gainmanagement support by demonstrating theeconomic and other benefits gained from theprogram. Show management that the processfits within the company structure, policy and
philosophy. Management support is needed toobtain the necessary resources such as capital,labor, and space. Management can alsomandate shop participation.
It is also important to continue working withthe regulatory agencies. Work with the fireagencies to insure that equipment, procedures,installation and location are all acceptable. UseOSHA regulations or work with OSHArepresentatives within your company todevelop safe operating procedures for theequipment and handling of recycled and spentsolvent. Safety, hygiene and toxicologypersonnel can also be helpful in developing aMaterial Safety Data Sheet (MSDS).
Develop a good rapport with the shop(s) andits management. Make the project a teameffort b / keeping the shop(s) informed fromthe beginning, soliciting their suggestions andmaking them understand that it is ultimatelytheir program. Shop support is crucial for theprogram's success.
Step Three - Implementation
Implementation of a program consists ofcoordinating and troubleshooting. Thelocation of the equipment must now bedetermined. Ensure that the equipment will beaccessible. The location should be acceptableto all parties involved (eg., the FireDepartment, Fire Protection Engineering, thesite environmental group, plant engineering,transportation, layout, the using organizationand the organization(s) generating the spentsolvent).
Training of personnel will also have to becoordinated. This includes the operator of thenew equipment and the shop personnelgenerating the solvent. The shop managementneeds to concur on exactly what is expected oftheir personnel to support the program once ifis implemented.
Construct a manual describing the processprocedure and assigning responsibilities to
198
each group involved in the program. Identifykey personnel to contact in each organization.This will eliminate grey areas. In the futurethe manual will be a useful reference for newpersonnel.
Following the guidelines discussed above willminimize problems. However, complicationscan arise. The equipment may fail. Contactths manufacturer to discuss troubleshootingand repairs should equipment malfunctionsoccur. Beware of misinformation from thevendor. As a rule, research all equipmentassociated with the process (i.e.. chillers,pumps, heat exchangers, condensers, etc.) andif possible get guarantees on all purchases. Ifit is not practical to return malfunctioning orinapplicable equipment to the manufacturer, itmay be necessary to make changes to theprocess, location or installation. For example.upon startup of this program it was discoveredthe chiller should not be operated outdoors orat ambient temperatures below 50°F. Thissystem was located outside and was operatedwhen the temperature was below 50°F.Unfortunately, this information was not knownprior to purchase and installation of thisequipment. During winter, the chiller beganto malfunction and shut down. To solve thisproblem, the chiller was moved indoors.
Key personnel may become obstacles. Theequipment operator may lack enthusiasm abouthis/her new duties. It may be necessary tospend additional time with the operator andprovide information which will explain thepurpose and goals of the program. The shoppersonnel may also cause problems. Theymay not care to participate or they maymisunderstand the procedures. The shoppersonnel may also require additional trainingor monitoring during the initial months of theprogram until the procedures become a part oftheir daily routine. Changes in key personnelinvolved with the program may occur from theinitial to the fina' stages of the project.Forming alliances with these new personneland educating them on the program willrequire additional time and effort. Regulationschange daily and as a result may require
adjustments in the program and its procedures.Keep abreast of new or anticipated changesthat may effect this program. Productionmay also change. Flexibility should be builtinto the program to cushion these productionchanges. The characteristics of the waste maychange. The product quality can be verifiedby taking samples throughout the program.
By conducting the pilot study, other problemscan be avoided. At the Boeing Aerospace andElectronics sites in the Puget Sound area,spent solvent and unused paint were combinedand disposed of offsite. This combined wastestream was initially distilled during the pilotstudy. It was discovered that the concentrationof paint in the solvent was too high to obtaina good separation via distillation. Segregationof spent paint and spent MEK prior todistillation was prescribed to solve thisproblem.
Different types of stills are available on themarket. The conventional still is a cylindricaltank with a manhole at the bottom to shovelout the sludge. Removing this sludge is achore, since it is a combination of waste paintand solvent that has been heated but has notcompletely evaporated. There aremanufactures that have attempted to remedythis situation by adding polypropolene orTeflon bag liners to contain spent solvent andprotect the walls of the still pots. This methodis convenient as long as the pot is not too deep(bag liners in deeper stills tend to tear whenbeing removed, therefore a shallow wide bowlis recommended for this technology). Thewaste should not contain constituents that willwear or dissolve the liner. This can beverified with the vendor.
Some vendors may try to incorporate othermanufacturers' technology into theirequipment. The utilization of bag liners instills has been adopted by many manufacturersand vendors because it makes the still bottomsmuch easier to handle, causes less wear andtear on the still and reduces the chances forspills. Although manufacturers and vendorsmay be correct when stating that their
199
equipment is capable of utilizing anothercompany's technology (eg., bag liners,condensers, etc.), it is recommended toresearch if there are patents which wouldprohibit use of the technology or requireroyalties to the patenting company. It is lesscomplicated to buy from the originator of theidea or patent (check on patents pending).
During the pilot study at BA&E it wasdetermined that the recycled solvent can onlybe used for cleanup applications because of thestringent specifications for solvents used inproduction operations. Use of the recycledsolvent for production would require costlyand time consuming tests on each batch toverify its quality. Fortunately, the quantity ofsolvent needed for cleanup purposes exceededthe quantity of recycled solvent produced.
CONCLUSION
In organizations where waste solvent isgenerated, solvent recovery is critical toproactive waste minimization. Thoroughresearch and an organized step by stepapproach are key to the success of a solventrecovery program. Implementation of asolvent recovery program requires obtainingtechnical information while gaining positiveprogram support from management, shops andregulating agencies. Incorporating technicaland persuasive skills in an organized manneris vital for the success of this program.Remember, the idea won't reduce hazardouswaste or save money unless it is successfullyimplemented.
Case Study Boeing Aerospace andElectronics' Plant II site in Seattle,Washington was consuming 19,000 gallons ofMethyl Ethyl Ketone (MEK) per year (basedon a 1989 Sara 313 Report). The disposalcost of spent MEK and associated wastes wasapproximately $330,000 annually. A pilotstudy found that it was possible to recover upto 90% of the spent solvent. Based on thisdata, Boeing purchased a distillation unitwhich was installed at a central hazardouswaste accumulation area at Plant II. Theinstalled cost of this unit was approximately$55,000, including a closed-loop chiller.Boeing anticipates an annual savings of$200,000 with a payback period of less than 4months. Once the program began, minimalproblems were encountered. Some obstaclesincluded minor equipment malfunctions, lackof enthusiasm from shop and other associatedpersonnel, skepticism towards using recycledsolvent, health related questions and batches ofspent solvent that were not reclaimable.BA&E Plant II site is now recycling spentsolvent and using this recycled solvent forcleanup operations.
200
RECOVERY OF WASTE SOLVENTS BY RECTIFICATION, AZEOTROPIC AND/OREXTRACTIVE DISTILLATION
Lloyd BergChemical Engineering Department
Montana State UniversityBozeman. Montana
Solvents are utilized for cleaning, strippingand various other maintenance operations ofaircraft parts and equipment. After use, manyof the solvents can be recovered and reused.However, several of the solvents currently inuse are chlorinated and emit volatile organiccompounds, which are toxic to theenvironment and to operating personnel.Wastes generated from these solvents areregulated by the U.S. EnvironmentalProtection Agency; and soon use andmanufacture of these solvents may berestricted.
The objective of the Solvent Recycle/RecoveryTask of the DOE Chlorinated SolventSubstitution Program is to minimize hazardouswastes by identifying recycle/recoverytechniques for the proposed substitutesolvents.
The following new and unused solvents havebeen received from the followingmanufacturers.
Company
Chemical MethodsChemical SolventsFine OrganicsFrederick GummGAFMcGean-RohcoMcGean-RohcoPate 1 in
Precision rectifications were made on each ofthe solvents listed above and the compositionof the unused solvents determined. The nextstep in the program will be to obtain samplesof these solvents after being used. They willbe screened, filtered and rectified to determinewhat changes occur during use.
The goal of the program will be to return theused solvents to a purity and composition of aquality which will restore their market value.This will be accomplished by the employmentof precision rectification, azeotropic and/orextractive distillation. The principalinvestigator has been awarded 95 patents inthe separation of acids, alkalies, alcohols,esters, ketones, amines, glycols, sulfurcompounds, nitrogenous compounds,heterocyclics and hydrocarbons.
The principal investigator has been awardedU.S. Patents in azeotropic and extractivedistillation to separate the following mixtures."
Rochester Midland PSS-600
201
Ethylbenzene from XylenesEthyl acetate from Ethanol (2)Isopropyl ether from Acetone (2)m-Xylene from o-Xylene (7)n-Butyl acetate from n-Butanol (2)n-Propanol from Ally] alcohol (3)Benzene from Non-aromatic hydrocarbonsn-Propyl acetate from n-Propanol (2)Formic acid from Water (4)Isopropyl acetate from Isopropanol (7)n-Amyl acetate from n-Amyl alcoholt-Amyl alcohol from Isobutanol (2)Ethanol from Isopropanol (2)n-Propanol from 2-Butanol2-ButyI acetate from 2-ButanolEthanol from t-Butanol2-Pentanone from Formic acidFormic acid from Dioxane (3)
Formic acid from 3-Methyl-2-butanone (2)Acetic acid from 4-Methyl-2-pentanoneBenzene from Acetonen-Propanol from t-Amyl alcoholEthyl benzene from Styrene (2)Ethylene glycoi from ButanediolsGlycerine from Polyols
Toluene from Non-aromatic hydrocarbonsIsopropyl ether from Methyl ethyl ketone (2)Acetone from Methyl ethyl ketoneMethanol from Acetone (3)Methyl acetate from Methanol (5)Isopropy! ether from Isopropanol (2)Isopropyl ether from IsopropanolEthanol from Water (2)Isobutyl acetate from Isobutanol (2)Methyl t-butyl ether from HydrocarbonsFormic acid from Acetic acid (2)n-Hexyl acetate from n-Hexanol (2)Isopropanol from t-ButanolPropionic acid from WaterAcetic acid from Water2-Butanol from t-Amyl alcoholAcetic acid from Dioxane (2)m-Diisopropyibenzene fromp-Diisopropylbenzene (7)Formic acid from 2-PentanoneVinyl acetate from Ethyl acetate (2)2,3-Butanediol from Propylene glycolFormic acid from 4-Methyl-2-pentanoneAcetic acid from 4-Methy!-2-pentanone2-Methyl-l-butanoI from Pentanol-1Methyl, Ethyl, Propyl & Butyl Lactates
202
RECYCLING ALTERNATIVES
James L. SchreinerExxon Chemical Company
Bayton, Texas
With increased awareness of the impact ofhandling and disposing waste streams, manycompanies are looking into recycling andrecovering materials used in their operations.In areas related to cleaning applications, therecovery and recycle of modem hydrocarbonand oxygenated solvents is important. Thesefluids are proposed to replacechloroflourocarbons (CFCs) and otherchlorinated solvents for a variety ofapplications.
RECYCLE
To facilitate recycle and for purposes of wasteminimization, it is important to haveknowledge up-front of the product'scomposition. That is, the product must be ofa quality which will not degrade in use orduring standard recycle operations.
For our purposes we will focus on twoqualities: 1) a defined and narrow boilingrange and 2) a chemistry with limitedunsaturation and/or reactivity. The firstcriterion simplifies as well as enables effectiverecycling. The latter ensures product qualityafter recycle, or in other words, eliminatedegredation of the produa as a result of use orrecycle.
It is with these in mind that varioushydrocarbon and oxygenated fluids are findingniches in cleaning applications. That is, theseproducts have compositions such that they maybe recycled with the reused material havingthe same properties as the original product.
Product Contamination and Clean-up
Cleaning operations impart many contaminantsto the cleaning fluid. Contaminants entering
a cleaning fluid were previously characterizedin a paper entitled "Reclamation andReprocessing of Spent Solvents" by ArthurTarrer, et. al. These contaminants includehydrocarbon oils, such as lubricating oils,greases, transmission oil, fuel oil; asphalts;tars; waxes: paint and varnishes; or soilymaterials such as clay and silt, cement, soot orlampblack. In addition, some contaminationfrom metallic fines should also be expected inmetal cleaning operations. Reclamation ofmetals will not be addressed in this paper.Information on this subject is availablethrough publications such as Metal Finishingor seminars such as SUR/FIN.
Relating to oil and soil contaminations ofhydrocarbon and oxygenated fluids, ExxonChemical has, in other industries orapplications, had similar challenges of productcontamination. In these applications, aproduct has been used in a process becomingsufficiently loaded (typically) with a heavy oil,such as a hydraulic fluid, where the product isno longer effective in the application. At thispoint, the user has several alternatives: 1)Disposal of the product, incurring significantcharges for handling the spent material whichwill be classified as a hazardous waste underRCRA. 2) Sale of the used product to anindependent reclaimer, recovering someportion of the original solvent investment. 3)Establishment of an outlet for reclaiming theseproducts for recycle in its own operations.
As the third alternative has been found to bethe more economic and environmentally safealternative, we frequently consult withcompanies working to recycle our products.Typically, the reclamation procedure involvesdistillation, but in some cases carbonadsorption has been used. Most of theconversion companies we have dealt with are
203
members of the Association of PetroleumRefiners or the National Association ofSolvent Recyclers.
One technique an analytical chemist may usewhen viewing a contaminated or spent productis gas chromatography. This technique can beused to identify materials based on thedistribution of the components by boilingpoint. This criteria is needed to determine theeffectiveness of distillation as a means ofrecycle and recovery.
An uncontaminated hydrocarbon oroxygenated fluid will have a discrete narrowboiling range with no indication ofcomponents trailing past the final boilmg point(FBP). This is indicated by a flat baseline ona gas chromatogram.
One of the fluids that has been contaminatedwill show peaks or aberations in the gaschromatographic baseline after the finalboiling point of the fresh fluid. Thesebaseline movements are indicative of the typeof contamination in the fluid. One example isthat of a hydraulic fluid contaminating ahydrocarbon fluid. This hydraulic fluid has aboiling range from 600-1040°F and a viscosityof 32 cSt at 100°F. The contamination ratioof cleaning fluid to hydraulic fluid is 96:4.
The use of a controlled batch distillationprocedure removed all of the contaminationfrom the cleaning fluid. The quality of thisrecycled material, as determined by gaschromatography. was identical to the unusedfluid. Additional analytical testing techniquesalso confirmed that the recycled productquality was identical to that of the unusedfluid.
Processing conditions to effect this recyclewere stated by the recycler in Indiana as beingquite complicated. A simplistic andgeneralized interpretation of the process.however, is one that would usually requireonly atmospheric capabilities if using a hot oilsystem or a fired reboiler to heat the material.However, for more difficult distillations or if
only high pressure steam is available, mildvacuum and a few theoretical stages will effectan optimal separation. The use of thin filmevaporators may also be effective.
Another example of the recycle capabilitiesinvolves the removal of contaminants from anoxygenated fluid system. The contaminationconsisted of carbon, high melt point wax anda 46 cSt oil. The loading level was close to15 wt%. The carbon was effectively removedusing a simple filtration. This filtered materialwas then distilled to return the solvent to itsformer quality. It is important to rememberthat this may only be accomplished if theproduct's chemistry does not degrade in use orin recycle. The use of vacuum distillationand/or other temperature controls duringdistillation will help reduce the risk ofdegredation during reclamation.
Distillation results in a recycled product withinthe proper boiling range of the originalproduct. Other product quality inspections,such as flashpoint, oxidation level, color,particulates. etc.. are necessary to approve arecycled product for reuse. These examplesare only a demonstration of a single screeninganalysis for reviewing the effectiveness of arecycle operation.
CAPTURE
Other considerations for reclaiming productwill now be discussed. These are capturetechniques which facilitate controllingemissions of products. While some 10-20technologies are under review by theDepartment of Energy (DOE) group SITE(Superfund Innovative TechnologyEvaluation), the following will address onlythe more common and established methods ofcapture, especially as they may be directlyapplied to cleaning equipment modifications.Many of these technologies were recentlydiscussed at the DOE-sponsored Conferenceon Industrial Solvent Recycling held inOctober in Charlotte.
204
The first technology is direct condensation tocontrol loss of vapor laden air. The air isrouted over refrigerated coils condensing theliquid for capture and recycle.
Another common capture technology is carbonadsorption. Usually, for efficiency ofoperation, two adsorbers are installed whereone is active in adsorbing vapors from theoperation and the other adsorber is goingthrough a regeneration procedure. Thisregeneration can be accomplished using eithersteam or inert gas. Using this technology,recovered material is desorbed from thecarbon adsorption bed by the steam or inengas and collected in a receiving vessel.
If steam is being used, the immiscibleproducts can be recovered in a separator whilethe miscible products are removed bysubsequent distillation. The preferred methodis desorption using an inert gas if that isavailable. Inert gas avoids the issues of waterquality in the desorption and regenerationphases of the recover\r and does not requireadditional distillative processing. Skidmounted units and special container carbonmodules are available from multiple suppliersto assist regeneration of carbon for smallerusers or for facilities not having steam or inengas available.
An emerging technology for large industrialapplications is called reverse Brayton cycle.The vapor enters a dehumidifier to removemoisture and then goes to a chiller to reducethe temperature. From here, some fluid maybe recovered with the remaining vapor sent toa turbo compressor where pressure isincreased dropping the temperature stillfurther. From here, the product is cooledthrough an interchanger and pumped throughthe expander of a free-spindle turbo bringingthe temperatures to as low as -80 = F. Thevapor in the gas stream is then recovered as acondensed liquid with the inen gas passedthrough a pump to the exhaust. The energyresulting from the pressure change through theexpander helps drive the compressor forenergy savings.
"From and to adsorber" shown on the diagramillustrates the use of the reverse Brayton cycleto desorb a carbon adsorption bed that hasbeen previously saturated with vapors. Thehot gas exiting the recovery operation is usedfor desorbing the volatile organic compound(VOC) and regenerating one adsorber,therefore providing energy savings andoperation efficiencies in the process.
Recovery of solvents by membranes is atechnology which continues to evolve andwarrants close attention over the next fewyears. All of the other technologies discussed,however, are immediately available and areapplicable to recovering and recyclinghydrocarbon and oxygenated fluids. Ascreening model is being prepared by ScienceApplications International Corporation for theDOE Office of Industrial Technologies. Themodel will review these technologies toaddress the most favorable economics forrecovering VOC types - consideringconcentrations and gas flow rates of theprocess stream.
Many challenges face us with the latest cleanair legislation. There are many optionsavailable to meet these challenges. Alltechnology and chemical replacements shouldbe reviewed in the pursuit of safe alternatives.Safety should be coupled with energyconservation to minimize waste. As recoveryand recycle are steps toward minimization.proper product selection is important to ensurethat the product chosen for use in a specificcleaning operation is capable of beingrecovered and recycled. Previous industrialexperience indicates that some hydrocarbonand oxygenated fluids are available which maybe recycled effectively, returning product toformer quality for reuse by the customer.This recycling may be conducted throughoutside tolling facilities or in purchased on-sitefacilities.
Although recovery technologies such asmembranes. solidification, and in-situvitrofication are developing quickly, variousrecycle and recovery technologies exist today
205
to assist waste minimization efforts, at thesame time controlling costs.
Exxon Chemical would like to acknowledgethe assistance of Dr. Victor S. Engleman, ofScience Applications InternationalCorporation, in the preparation of informationand graphics for this presentation.
This talk was prepared by Janet S. Catanachand Kishore K. Chokshi of Exxon ChemicalCompany.
206
THIN FILM EVAPORATION FOR REUSE/RECYCLEOF WASTE ORGANIC SOLUTIONS
W. N. WhinneryPaducah Gaseous Diffusion Plant*
Paducah, Kentucky
ABSTRACT
A thin film evaporator TFE has been used atPGDP to evaluate the feasibility of recoveringwaste organic solutions as reusable products orto reduce the waste disposal volume. Thesetup of the TFE and the hot oil heatingsystem has allowed a wide range of solutionsto be tested. These solutions include 1,1,1trichloroethane, trichloroethylene, wastedeplating solution, lacquer thinner (paintwaste), and a poly chlorinated biphenyls (PCB)and uranium contaminated waste oil solution.The recovery rate or waste volume reductionare presented for each solution tested. Costsavings for specific compounds andrequirements for the use of recoveredtrichloroethylene and 1,1.1 trichloroethane areexplained. Actual plant reuse of the recoveredchlorinated solvents and the lacquer thinnerhas occurred.
EXECUTIVE SUMMARY
The thin film evaporator (TFE) is being usedto evaluate the feasibility of recovering wasteorganic solutions as reusable products or toreduce their waste disposal volume. TheTFE has successfully recoveredtrichloroethylene, 1.1,1 trichloroethane andlacquer thinner. These chemicals have beenreturned to a large vapor degreaser. smallparts cleaning bath in the pump shop, and thepaint shop for field use testing. No problems
•Operated by Martin Marietta Energy Systems, Inc forthe U.S. Department of Energy under Contract No DE-AC05-840R21400
have been found with the recovered productscompared to the virgin solvents.
Testing of PCB (polychlorinated biphenyls)and uranium co-contaminated waste yielded a28% decrease in disposal volume. 1>econdensate recovered from the evaporationwas below 50 ppm PCB, the federallyregulated guideline. These test results provedvolume reduction of PCB iaden waste can beachieved prior to final disposal 'incineration).
The technology is applicable to other sites forreduction of hazardous and contaminatedwaste feeds prior to incineration.
Material Tested To-Date
Materia ls tested to-date includetrichloroethylene, 1,1,1 trichloroethane,nickel stripper solution, co-contaminated oil.and lacquer thinner. Percent solvent recoveryis defined as weight of solvent recovereddivided by the weight of solvent processed.Maximum recoveries achieved were 96% fortrichloroethylene and 98% for 1,1.1trichloroethane. Recovery rates were foundto be a function of:
1. Solids loading2. Number of passes through the TFE system3. Percent oil in the waste
Uranium removal has averaged 99% withtypical 25 ppm uranium starting levels beingconcentrated to 350 ppm in the bottoms andthe product reduced to 0.02 ppm.
The readdition of stabilizers to the recoveredproduct suppresses the generation andaccumulation of HC1 during reuse of the
207
solvent. This prevents corrosive reactions inthe steel storage drums which would lead topinhole leaks. To prevent corrosion, eachspecific solvent is inhibited with an acidaccepting stabilizer. The level of inhibition ofa solvent is measured by the non-amine acidacceptance (NAA). The average operatingrange of the NAA in the degreaser is 0.10 to0.13% as NaOH. The maximum amount ofinhibitor is needed in waste solvent afterprocessing due to the interaction of the usedsolvent and die oil from dirty equipment.1,1,1 trichloroethane is stabilized by 1,2butylene oxide and secondary butyl alcohol.Trichloroethylene is stabilized by 1,2butylene oxide, secondary butyl alcohol,cyclohexane oxide and diisopropyl amine.
The nickel-stripping solution is composed oftwo parts. The "A" component is a drypowder composed of sodium meta nitro benzylsulfonic acid. The "B" component contains2% ammonia and 20% aliphatic amine. TFEtreatment of the nickel-stripper solution hasyielded two distinct fractions, an ammonia-water mixture in the distillate and the Metexsolid with high boiling liquids in the bottoms.The volume reduction achieved by separationof the ammonia-water mixture from the feedmaterial was 55 %.
The co-contaminated waste oil solution washigh in suspended solids (75 g/1) and wasfiltered to 100-micron particle size prior toTFE processing. The PCB level of the initialsample prior to filtration was 6700 ppm. TheTFE concentrated the waste in the bottomssolution to 55 g/l solids content. Afterprocessing, the bottoms concentration of 5200ppm PCB's was lower than expected due tothe solids removed in the filtration. Thesolids filtered out had PCB's sorbed to thesolid material and produced the lower thanexpected bottoms concentration. Thecondensate contained less than 50 ppm PCB,which is below the federally regulatedguideline for PCB's. The volume of theriginal solution contaminated with PCB's wasreduced by 28% on a s'ngle pass through theTFE. Increased volume reduction would have
been obtained if a lower initial suspendedsolids concentration had been present. A 50%recovery of the available volatile and waterfraction of the solution was obtained. Thisproves volume reduction of mixed volatilewastes containing PCB can be achieved priorto final disposal (incineration).
Test Materials Chemical Parameters
The specifications for chlorinated solventswere taken from the manufacturers datasheets and our material specification.Additional parameters of interest were added,such as uranium and total suspended solids.The list of test parameters for the chlorinatedwaste is shown in Table 1.
Testing of the waste nickel stripper solutionwas monitored for percent volume decrease.Earlier testing in the laboratory showed theliquid fraction of the two-part, nickel-stripping compound was the active ingredient.The "B" liquid fraction is partiallyfractionated in the evaporation process. Thehigh boiling liquid in the bottoms fractioncould serve as an activator for the strippingprocess; however, the bottoms does have highsuspended solids and separating the liquidwas difficult. Recovery of the bottoms liquidfor reuse was not feasible. The driedbottoms residue was ineffective as an additiveto rejuvenate the stripping solution. Thedeplating operation would be affected byincreased nickel concentrations from additionsof the dried bottoms and would reduce thelife of the bath rather than extend it. Thecondensate (product) yields an ammonia-water mixture that was tested in the laboratoryand would not deplate nickel plated steelsamples. The use of the ammonia-watermixture as a fertilizer is a reuse application ofthe condensate solution.
The waste oil solution contaminated withPCB's and U was analyzed for thecharacteristics listed in Table 2.
The lacquer thinner material was tested for thesame general properties as the chlorinated
208
solvents. The lacquer thinner waste solutionproduced a very clear condensate solution.The waste was identified by GC/MScomparison with stores current product oflacquer thinner. The test run recovered 57%of the total waste solution on the first pass.The solution was prefiltered and removed thepaint fines that would have coated the TFEjacket walls and reduced heat transfer.
Flow Diagram of the TFE
The flow diagram of the thin film evaporators>stem is shown in Figure 1. This systemincludes an oil heater unit to provide heatingof the jacketed evaporator section. The uppertemperature limit of the hot oil heating systemis 550cF. The evaporator is composed of arotating center section, producing a thin filmon the inside wall of the evaporator jacket, acondenser, feed pump, condensate pump, andbottoms pump.
Optimum Machine Conditions
The major parameters of operation that weremonitored on all testing runs are listed inTable 3. Amperage load on the rotatingcenter shaft ranged 62 - 67% of full loadduring all the testing. Increases in the vacuumon the TFE produced quicker product rates.The vacuum setting in processing thechlorinated solvents was five psia. At vacuumconditions above 13.5 psia. the rotatingcenter section will vibrate. The vacuum canbe used to draw feed material into the TFE toaid in priming the feed pumps provided theback pressure valve is set low enough toallow this to occur. Optimum conditions forprocessing the waste tested are listed on Table4.
The feed rate will vary depending on thesolids loading, percentage of oil in thesolvent, and type of pump used. Alltrichloroethylene waste was processed in asingle pass. The temperature differencebetween 1.1,1 trichloroethane and
trichloroethylene represents the added heatrequired to process the higher boilingtrichloroethylene. The flow rate will dictatehow many passes the solvent should makebefore all available material is recovered. Thetime required to process a 55-gallon drum ofwaste chlorinated solvent was 4 - 6 hoursdepending on the percent solvent in the feed.
The co-contaminated solution was filteredprior to treatment. The TFE removed excesswater and volatile organics from the waste oilmixture. High solids loading resulted inpumping out the bottoms area frequently.
The lacquer thinner solution processed throughthe TFE was filtered to 30 microns prior totreatment to remove paint panicles from thesolution. Solution fed very smoothly after thepump was primed and the correct pump feedpressure was established. The temperature inthe vapor phase was still too low for singlepass complete recovery.
The product cycle time has been plotted incontrol chart form for each of the materialstested and out of control points can be tracedto changes in the feed pump settings.
DISCLAIMER
This report was prepared as an account ofwork sponsored by an agency of the UnitedStates Government. Neither the United StatesGovernment nor an agency thereof, nor any oftheir employees, makes any warranty,expressed or implied, or assumes any legalliability or responsibility for the accuracy,completeness, or usefulness of anyinformation, apparatus, product, or processdisclosed or represents that its use would notinfringe privately owned rights. Referenceherein to any specific commercial product,process, or service by trade name, trademark,manufacturer, or otherwise, does notnecessarily constitute or imply itsendorsement, recommendation, or favoring bythe United States Government or any agencythereof. The views and opinions of authors
209
expressed herein do not necessarily state orreflect those of the UniteJ States Governmentor any agency thereof.
210
TABLE 1TEST PARAMETERS FOR CHLORINATED WASTE SOLUTIONS
UNDER SPECIFICATIONS FOR MATERIALS
Trichloroethvlene 1.1.1 Trichloroethane
AcidityAcidity after oxidation testAlkalinityResidue on evaporationMoistureFree halogenColorCopper corrosionBoiling rangeSpecific gravity
Gregory C. Frye and Stephen J. MartinSandia National LaboratoriesAlbuquerque. New Mexico
ABSTRACT
On-line chemical monitoring systems can helpensure safe, environmentally sound operationof industrial processes using hazardouschemicals. Using polymer-coated surfaceacoustic wave (SAW) sensors, we havedemonstrated monitors that are capable ofdetecting dilute concentrations of volatileorganic species. Using changes in both wavevelocity and wave attenuation, the identity andconcentration of an isolated chemical speciescan be determined. A polysiloxane coatinghas been found to provide unique propertiesfor monitoring chlorinated hydrocarbons(CHCs) such as trichloroethylene: gooddiscrimination of CHCs from most otherorganic species, rapid and reversible sensorresponse, and low detection limits. Using thistechnology, a portable acoustic wave sensor(PAWS) system has been constructed.
INTRODUCTION
A wide variety of volatile organic species arecommonly used in industry, especially incleaning operations such as vapor degreasing.Proper use of these chemicals requiresaddressing several environmental, safety andhealth related concerns. For example,chlorinated hydrocarbons such astrichloroethylene (TCE) are often used ascleaning solvents. Their ozone depletingpotential makes their emission to theenvironment a grave concern. This hasresulted in severe restrictions in their use, asdictated by the Montreal Protocol and theClean Air Act. In addition, many CHCs areknown or suspected human carcinogens andtheir concentration in the workplace isregulated by agencies such as OSHA. Jn
some DOE and industrial applications,acceptable alternatives to CHC-based cleaningare not currently available. In thesesituations, continued operation will requiredocumenting that CHC usage is in compliancewith all relevant regulations.
Real-time, on-line monitoring of exhauststacks and workplace environments can play acritical role in enabling safe, environmentallysound operation of processes using volatileorganic species. These monitors can be usedto document emissions and verify thatconcentrations in the workplace do not exceedsafety standards. They can be an effective toolin waste minimization and pollution preventionefforts, too. This latter application can be assimple as using the direct chemicalinformation to optimize worker protocols or toevaluate changes in processes or equipment.Alternatively, these monitors can be integratedwith on-line process control systems tooptimize process operation for improvedproduct quality and yield as well as tominimize waste and prevent pollution.
There are several requirements for usefulindustrial monitors. First, they noed to havesufficient sensitivity to detect the chemicalspecies of interest in dilute concentrations.For many applications, this requires detectionlimits on the order of one ppm or less.Second, they need to be insensitive totemperature fluctuations and potential chemicalinterferants, especially the omnipresent watervapor. Sensitivity to these effects can becircumvented by constructing a sensor systemwhich enables the periodic determination ofthe baseline sensor response by excluding thechemical species of interest from the sensor.Third, for many applications, the monitorsmust be able to identify the chemical species
215
providing the sensor response to preventunwarranted reactions, e.g., the emergencyevacuation of a workplace due to amisinterpretation of a sensor response. Inaddition to these requirements, there are manyother critical issues which must be consideredsuch as cost, size, reliability, durability, speedof response, ease of integration into theindustrial process, and simplicity of operation.
In this paper, we will describe monitoringsystems that meet these basic requirements.These systems utilize surface acoustic wave(SAW) devices and can be applied tomonitoring a wide variety of volatile organicspecies, including CHCs. The current systemis unable to determine the concentrations ofmultiple species in mixtures; however, it isable to discriminate between responses due todifferent isolated chemical species, allowingspecies identification and quantification usinga single SAW sensor.
SURFACE ACOUSTIC WAVE SENSORS
Surface acoustic wave devices consist of inputand output interdigitated transducers patternedon a piezoelectric substrate such as quartz (seeFigure 1). When an alternating voltage isapplied to the input transducer, an alternatingmechanical strain is generated in theunderlying substrate which launches the wave.This wave travels along the surface andinteracts with a thin film formed on the devicesurface before being converted back into anelectrical signal by the output transducer [1].This SAW/thin film interaction can resuit inperturbations in both wave velocity and waveattenuation (i.e., loss of acoustic power) inresponse to perturbations in film properties[2.3].
As shown in Figure 1, an effective method formonitoring changes in wave velocity is tooperate the SAW device as the feedbackelement of an oscillator circuit. In thisconfiguration, relative changes in oscillationfrequency (f) can be directly related to relativechanges in wave velocity (v): Af/f0 = Av/v0
where f0 and vo are the unperturbed frequencyand wave velocity, respectively. Waveattenuation can be measured with a SAWoscillator system by incorporating a vectorvoltmeter to monitor changes in the input andoutput signal levels to the SAW device [2,3].Alternatively, if the oscillator circuit isoperated with constant input power to thedevice (i.e., amplifier saturation), then waveattenuation can be obtained by monitoring theRF signal level at a single point in the circuitafter the device, as shown in Figure 1.
The SAW device used in this study waspatterned on an ST-cut quartz substrate andoperated at 97 MHz. To demonstrate theextreme sensitivity of this device, a 1 Hzfrequency change corresponds to a change insurface mass of only 80 pg/cm2. The filmused in this study was a polysiloxane polymerfilm formed by plasma-assisted chemical vapordeposition of hexamethyldisiloxane. The filmthickness, based on profilometry, wasapproximately 1 iim. Since no oxygen wasadded to the chamber during deposition, thepolymer contains a large number of methylgroups, making the film hydrophobic innature. The resulting affinity of this polymerfor many volatile organic species, withcomparatively low affinity for water, makesthis film ideal for monitoring organics such aschlorinated hydrocarbons.
With this film, there are two film propertieswhich are altered by absorption of species andwhich result in detectable sensor responses.First, the increase in the mass of the filmresults in a velocity decrease with no changein the attenuation. This "mass loading"response is the most common interaction usedin SAW sensor applications [4]. Second, theviscoelastic properties of the film are alteredby the plasticizing action that occurs upondispersal of species between polymer chains.In contrast to the mass loading interaction, thisviscoelastic interaction results in changes inboth the velocity and the attenuation.
216
DUAL OUTPUT SENSORS FORMOLECULAR IDENTIFICATION
Figure 2 shows the response of thepolysiloxane-coated SAW device to variousconcentrations of TCE in a dry nitrogenstream. The relative velocity changes wereobtained from frequency changes, while theattenuation changes were determined using anHP 8505 A vector voltmeter. Vaporconcentrations were varied using computer-operated mass flow controllers to vary therelative flow rates of a "carrier" stream(saturated with TCE by passage through abubbler) and a dry nitrogen "mix-down"stream. Using mis system, partial pressures(P) from 3 to -17% of the saturation vaporpressure (P^J can be obtained (see referencesJ or 5 for details). Due to the large responsesobserved with TCE, the concentration scanhad to be cut short at 16% of saturation;excessive attenuation makes device operationimpractical.
The advantage of monitoring both velocity andattenuation responses can be demonstrated byplotting the attenuation response vs. thefrequency response as shown in Figure 3 (thegas phase concentration determines theposition along the curve for each species). Itis clear that a unique curve is traced out inthis piot for each chemical species; thus, thesetwo sensor responses are independent. Thisdemonstrates the capability for distinguishingspecies on the basis of these two sensorresponses; In fact, the relative magnitudes ofthese two responses indicate the chemicalspecies providing the sensor response. Withthe chemical species identified, theconcentration of that species can bedetermined from calibration curves for thatspecies (see Fig. 2). Thus, a single SAWdevice can be used to both identify andquantify an isolated chemical species [5,6].
The results of studies with this polysiloxanecoating have demonstrated several otheradvantages of this coating: rapid response dueto fast diffusion, large response to volatileorganic species (1000 ppm frequency response
to TCE at 15% of saturation), reversibleresponse, and relatively small sensitivity towater vapor (50 ppm frequency response with90% relative humidity). Based on the testsshown in Figure 3, along with tests with avariety of other species (acetone, 3-methylpentane, dodecane, ethanol. isopropanoland water), a trend was observed in the datafor all species except those with significanthydrogen bonding capability (e.g. water andthe alcohols). The velocity shift at a givenvalue of attenuation was found to beproportional to the liquid density of theabsorbing species (density values are listed inFig. 3). Since chlorinated hydrocarbons havesignificantly higher densities (1.4 to 1.6g/cm3) than typical organic solvents (0.6 to1.0 g/cm3), this trend provides a distinct set ofresponses for CHCs. Thus, this polysiloxanefilm is ideally suited for discriminatingbetween chlorinated hydrocarbon species ascompared with most other volatile organicsolvents. Studies with polybutadiene andpolybutadiene/polystyrene films showed thissame dependence on liquid density [3]. Thistrend appears to be due to the manner inwhich these solvents plasticize these films,specifically that the amount of plasticizingaction correlates with the volume of speciesabsorbed into the film [3]. In contrast, apolyimide film has also been tested and, inthis case, the responses appear to correlatewith the molecular weight of the absorbingspecies [5J. This suggests a "massspectrometer on a chip"; however, the slowdiffusional properties of this film make itimpractical for sensing all but a few smallspecies.
Differences in response trends betweenpolymers can be used to choose a polymerwhich will optimize the sensitivity ?jid thediscriminating capability for a given sensingapplication (i.e, a given set of chemicals to besensed along with potentially interferingspecies). In addition, for analyzing mixtures,arrays of SAW devices with different coatingscould be used along with pattern recognitionschemes [7]. Since each sensor provides twoindependent responses, the number of sensors
217
required to quantify a mixture can be reducedby a factor of two as compared with singleoutput sensors. This simplifies the oftendifficult task of finding a sufficient number ofoatings which provide independent responsesto the chemical species of interest [7].
PORTABLE ACOUSTIC WAVE SENSORSYSTEMS
In order to demonstrate the practicalapplication of these dual output SAW sensors,development has begun on a prototypeportable acoustic wave sensor (PAWS) systemusing this technology. A schematic of thePAWS system is shown in Figure 1. Thesensor module contains a coated SAW sensorin a gas test fixture, RF electronics to operatethe device, and gas handling equipment. TheRF electronics consist of: (1) two modularamplifiers (Pasternack) having a combinedmaximum gain of 55 dB, (2) a coupler afterthe output transducer to split a portion of thesignal to send to an external frequency counter(HP 5384A), and (3) a second couplerbetween the amplifiers which splits power offto an RF detector (HP 8471 A: converting RFpower level to a DC voltage) which isconnected to an external DC voltmeter (HP3457A). The frequency counter and voltmeterare monitored using a computer (HP 9816).
The PAWS gas handling equipment consists ofa pump (Romega) to draw a gas sample acrossthe SAW sensor from the environment to betested, an activated carbon scrubber and athree-way teflon solenoid valve (ValcorScientific) to direct the gas flow through thescrubber upon request. Since the activatedcarbon effectively removes volatile organicspecies without removing water vapor,periodically passing the stream through thescrubber reestablishes sensor baseline. Theprototype PAWS module was assembled in ametal box with dimensions of 30 x 30 x 10cm.
Using the polysiloxane-coated SAW device,preliminary testing of the PAWS module has
been initiated. Baseline response with novapor present has been characterized in orderto evaluate noise levels and sensor drift.Significant drift in the signals is observedduring experiments performed over manyhours. By monitoring the temperature in thebox using a thermocouple, it was determinedthat this drift is due almost solely to changesin the device temperature. Since one reasonfor choosing the ST-cut of quartz for SAWdevices is due to the small temperaturecoefficient around room temperatures, thistemperature-induced SAW sensor drift isprobably dominated by changes in theviscoelastic properties of the polysiloxanefilm. Increasing temperature acts to soften thefilm much like what happens with solventplasticization [3j. Since these viscoelasticproperties form the basis for the independenceof the two sensor responses, this sensitivity totemperature cannot easily be removed.However, it can be circumvented by theperiodic use of the activated carbon scrubberto reestablish sensor baseline or bytemperature control of the SAW test case
The sensor drift makes it difficult to establishan estimate of the noise level. Since this driftwas approximately linear over short times, therms error between a linear least squares fitand the data was calculated to estimate theshort-term noise. For frequency (used toindicate velocity), the noise level for one andfive minute intervals was only 0.8 and 1.8 Hz,respectively. For attenuation (calculated bycalibration of the DC voltage from the RFdetector), the noise level is 0.00037 and0.00043 dB at one and five minute intervals,respectively. This value is smaller than thatobtained using the more expensive and bulkyvector voltmeter.
Estimating the lower limit of detection withthis film requires comparing device sensitivity(i.e., slope of the response vs. concentrationcurve) to these noise level;. Since thesaturation vapor pressure of TCE at 20 °C is57 Torr (a concentration of 75.000 ppm withan ambient pressure of 760 Torr). a 5.4 Hzfrequency shift (three times the five minute
218
noise level) represents a gas phase detectionlimit for TCE of 0.5 ppm. For attenuation, a0.0013 dB shift corresponds to 1.6 ppm.indicating that the frequency response isslightly more sensitive for detecting TCE.These low detection limits will only be validfor situations where the species concentrationchanges over a short period, e.g.. in theapplication of these devices as gaschromatographic detectors. For on-linemonitoring applications, the ability to accountfor sensor drift and to accurately establishsensor baseline determines the detection limits.
Studies to date on the response of theprototype module to vapors have focused onthe response of the polysiloxane-coated SAWdevice to trichloroethylene vapors. Tosimulate "real-world" conditions, thesechallenges were performed by probing thehead space in small vials containing thesolvent. As shown in Figure 4. rapid,reversible responses in both wave velocity andwave attenuation are observed uponchallenging with TCE. The response is large(e.g.. responses of over 70 kHz and 10 dB).as expected based on the previous work withthis polysiloxane coating. Due to the dynamicnature of this experiment, TCE concentrationis continuously changing with time. Thesensor follows these concentration changes, asindicated by the varying signal levels. Thefrequency response is delayed with respect toattenuation due to the one second gate timeused for the frequency counter.
The ability of the activated carbon scrubber toprovide sensor baseline has also beenevaluated. With no vapor challenge, a small(approximately 100 Hz frequency shift) sensorresponse is observed when the scrubber isactivated. This response is reversible and thesignal level corresponds to that expected for aTCE challenge at a concentration of about tenppm. indicating the practical detection limitwhen the scrubber is being used to providesensor baseline. As shown in Figure 4. v.hena vapor is being passed over the sensor, theactivation of the scrubber results in return ofthe sensor responses to near their
"unchallenged" values in about ten seconds.The ability of the scrubber to remove about99% of the TCE from the gas phase (based onthe level of return of the signals in Fig. 4)should be sufficient to reestablish sensorbaseline. In conclusion, these preliminarystudies demonstrate the basic capabilities ofthe PAWS module: real-time detection ofvolatile organic species such as TCE. twosensor responses for species identification, andinsensitivity to sensor drift due to changes intemperature and relative humidity based on theuse of the activated carbon scrubber toperiodically reestablish sensor baseline.
ACKNOWLEDGEMENTS
We gratefully acknowledge helpful discussionswith L. Gilliom and A. J. Ricco and thetechnical assistance of B. L. Wampler. A. K.Hays. G. C. Cordes. and T. V. Bohuszewicz.all of Sandia National Laboratories (SNL). Aspecial thanks to L. Casaus of SNL for hiswork in assembling the prototype PAWSmodule and to H. Wohltjen of MicrosensorSystems. Inc. for useful suggestion on theconfiguration of the PAWS module. Thiswork was performed at SNL, supported by theU.S. Department of Energy under contract no.DE-AC04-76DP00789.
REFERENCES
D. P. Morgan. Surface-Wave Devicesfor Signal Processing. Elsevier, New-York. 1985.
S. J. Martin and A. J. Ricco."Effective Utilization of AcousticWave Sensor Responses: SimultaneousMeasurement of Velocity andAttenuation." in Proc. 1989 IEEEUltrasonics Symp.. IEEE. New York.1989. p. 621-625.
S. J. Martin and G C. Frye. "SurfaceAcoustic Wave Response to Changesin Viscoelastic Film Properties." Appl.
219
Phys. Lett., 57 (1990) 1867. 6. G. C. Frye and S. J. Martin, "DualOutput Acoustic Wave Sensor for
4. M. S. Nieuwenhuizen and A. Venema, Molecular Identification," U. S. Patent"Surface Acoustic Wave Chemical Application, Serial No. 07/592,383,Sensors," Sensors and Materials, 1 Filed Oct. 3, 1990.(1989) 261.
7. S. L. Rose-Pehrsson, J. W. Grate, D.5. G. C. Frye and S. J. Martin, "Dual S. Ballantine, Jr. and P. C. Jurs,
Output Surface Acoustic Wave Sensors "Detection of Hazardous Vaporsfor Molecular Identification," Sensors Including Mixtures Using Patternand Materials, 2 (1990) 187. Recognition Analysis of Responses
from Surface Acoustic WaveDevices," Anal. Chem., 60 (1988)2801.
220
eno
O
o>on
From EnvironmentUnder Test
ActivatedCarbon
Scrubber
iTest Case
Pump
M M| I Coated M ISAW
I N Device I N
RFDetector
ToExhaust
Coupler
Coupler
Voltmeter
LI Computer
FrequencyCounter
Fig- 1: Schematic of chemical monitoring system consisting of a coated surface acoustic wavesensor, a gas handling system to draw in a gas sample, and electronics to operate thedevice and monitor changes in wave velocity and attenuation.
221
Frequency Shift (ppm)
g1
••••••
•••
4"-
«1 A L
t/1 •• •
• •a •
• •• •• •• •• •
• •• •
• •• •
• •• •
» ••
••
Attenuation Shift fdB)
Fig. 2: Changes in frequency and attenuation for a polysiloxane-coated SAW device vs.relative panial pressure of trichloroethylene vapors. Either of these curves can beused as a calibration curve to determine vapor concentration if it is known that thechemical providing the sensor response is TCE.
222
•v
s
-800 -600 -400
Frequency Shift (ppm)
-200
Fig. 3: Plot of attenuation response vs. frequency response for a polysiloxane coated SAWdevice showing discrimination of chemical species based on differences in the relativemagnitudes of these two sensor responses: ( • ) hexane (liquid density p = 0.660g/cm3), ( • ) toluene (p = 0.867), (A) trichloroethylene {p = 1.464), and (•)dibromomethane (p - 2.497).
223
Frequency Shift (ppm)
Power Signal (mV)
Fig. 4: Response of PAWS module to a trichloroethyiene challenge (180 to 240 sec) obtainedby probing the head space in a vial containing TCE. Scrubber activation with avapor challenge (200 to 220 sec) showi the ability to reestablish sensor baseline.Scrubber activation with no vapor challenge (120 to 140 sec) shows a small response(undetectable on this scale).
224
EVALUATION OF LOW VOC MATERIALS AT THE BOEING COMPANY
Linda H. Hsu and Judith A. WernerMetals and Finishes
Boeing Commercial Airplane GroupSeattle, Washington
INTRODUCTION
Over the past several years, the regulatoryemphasis has been directed toward control andreduction of smog-forming solvents,sometimes referred to as volatile organiccompounds or VOCs. The Boeing Companyhas devoted considerable effort and resourcesto the development of environmentallyacceptable materials and processes.
The projects described below are generallythose in which significant research orproduction scale up efforts have been made.Many of the Boeing-developed processes arecurrently of a proprietary nature. Informationgiven in these areas is general and has beenprovided to indicate where environmentallyacceptable alternates are probably going to beavailable in the near future in the area ofsolvent substitution.
LOW VOC CHEMICAL RESISTANTPRIMERS AND TOPCOATS
The development efforts of low VOC chemicalresistant primers and topcoats for metals andcomposites have been focused on materialssubject to regulation in the South Coast AirQuality Management District (SCAQMD).The regulatory maximum levels for the VOCcontents of chemical resistant primers andtopcoats are 350 and 420 grams/liter,respectively. VOC levels of conventionalchemical resistant primers and topcoats areapproximately 650 and 600 grams/liter,respectively. Four methods have been utilizedby coating manufacturers to achieve low VOCcontents:
1. Use of chlorinated solvents(Exempt, non-smog formingsolvents)
2. Use of water as carrier (WaterReducible)
3. High solids4. Electrodeposition
Overall Boeing has evaluated more than 200coating alternative candidates. Extensivetesting requirements are imposed on all thecandidates. The main tests these primers andtopcoats must pass include salt spray corrosionresistance, chemical resistance (includingSkydrol. fuel, lube oil, solvents), water,sealant adhesion, condensing humidity, andprimer/topcoat compatibility. In addition, theurethane compatible and corrosion resistantprimers must pass filiform corrosion and rainerosion tests.
Interior Structural Primers
The interior structural primers compriseapproximately 35 to 40% of the total paintused on Boeing commercial aircraft. They arethe primary corrosion preventive coating usedon airplanes today. Of the numerouscandidates evaluated, there are only threeprimers which have met Boeing specificationrequirements. Two acceptable primers arechlorinated solvents containing chemicalresistant primers. These two coatings wereimplemented in 1988 and 1989 in the LosAngelos basin areas. However, these recentlyqualified chlorinated solvent primers are noiused for any kind of wet installations, to avoidthe possibility of corrosion from trappedresidual chlorine. The third qualified primeris the water reducible version of the chemical
225
resistant primer. Currently a version of thewater reducible primer is being used formilitary applications by the Boeing Aerospaceand Electronics Company. BoeingCommercial Airplane Group is currentlyconducting a manufacturing scale-up study forcommercial aircraft applications.
Work is continuing on water reducible andhigh solids formulations to replace theconventional solvent based primers. Severalof the otherwise promising candidates arefaced with the problem of topcoatincompatibility. The alternative primers aresensitive to the combination of substratesurface film conditions and enamel overcoat.A major effort has been devoted todevelopment and manufacturing evaluations ofthese primers.
Composite Finishing Primer
Organic finishes for composite surfacescomprise approximately 5 to 15 % of paintused on Boeing commercial aircraft. Similarto all other primers, the regulatory target is350 grams/liter for the SCAQMD. Theadvantage of the composite finishes over metalfinishes is that the formulation does notcontain chromates for corrosion protection.Currently a water reducible, non-chromatedprimer is under manufacturing suitability studyfor commercial aircraft applications oncomposite parts.
Fuel Tank Primer
The development of low VOC fuel tankprimer alternates represents a significantengineering challenge. The fuel tank primercontains stringent engineering requirementsbecause of its dual function of corrosionprotection and sealant adhesion necessary forfuel containment. The combination ofrequired properties has resulted in a time-consuming material development program.The fuel tank primer comprises about 10% ofthe total paint volume on a commercial
aircraft. The regulatory target to be met in1993 is at 420 grams/liter. Currently Boeingis conducting preliminary screening tests onwater reducible and high solids alternativessubmitted by suppliers.
Interior Structural Topcoat
The interior structural topcoat usage is about15 to 20% on a commercial aircraft. Theregulatory VOC content is targeted at 420grams/liter or less. Of the numerouscandidates that Boeing has evaluated, two highsolids materials have passed the qualificationtests. They will be implemented by thissummer at subcontractors in the South CoastAir Quality Management District jurisdiction.
Abrasion Resistant Topcoat
The abrasion resistant topcoat is anothercoating containing difficult engineeringrequirements for material qualifications. Theabrasion resistant topcoat is used on trailingedge flaps and other rub areas. The usage isabout one percent of the total paint volume onan aircraft. Currently the Boeing CommercialAirplane Group is evaluating primer/topcoatcompatibility with two high solids, low VOCcandidates. Material qualification is inprogress for the two alternatives.
Interior Cabin Finish System
Testing is in progress on a one-coat candidateto replace the existing primer/topcoat interiordecorative enamel system. The major criteriafor interior decorative coatings are stainresistance and flammability. A high solids,low VOC candidate is being evaluated to meetthe 420 grams/liter requirement.
Concurrently to research and development ofalternative materials, the Boeing Company isalso implementing environmentally compliantapplication equipment to reduce the VOCemissions. High Transfer Efficiency (HTE)
226
spray equipment has been evaluated and is inuse with a variety of coatings and applicationswithin the Boeing facilities. Electrostaticequipment is currently being used at theBoeing Fabrication Division Auburn facility.In addition, the electrostatic equipment is alsoused wideiy in many paint shops and painthangars. High Volume Low Pressure (HVLP)spray guns have also been implemented formilitary applications in the Boeing Seattle sitesand at the Boeing Helicopter division inPhiladelphia. Several commercial facilities inthe Puget Sound area are currently evaluatingthe HVLP spray guns for a variety ofapplications.
ENVIRONMENTALLY ACCEPTABLESOLVENT CLEANERS
Regulations on the VOC content of cleaningsolvents generally control the vapor pressureof the VOCs. The usual requirement is thatVOC vapor pressure not exceed 45 mm Hg at20C. Engineering and manufacturingevaluation of cleaning solvents with low VOCvapor pressures is in progress.
Low VOC solvent research and developmentactivities primarily include the replacement ofMethyl Ethyl Ketone (MEK). naphtha,toluene, and other high vapor pressure volatilecommercial cleaners. Several low vaporpressure cleaning solvents and mixtures havebeen qualified for use on commercial andmilitary hardware. These materials areVOCs. but, emissions are reduced because thematerials do not have vapor pressures as MEKor toluene. Since these alternatives do notevaporate as quickly as the traditional cleaningsolvents, the residual cleaning materials arewiped off from the hardware and captured.
The material qualification for general-usesolvents is an on going effort for the BoeingCompany. In addition to general cleaningapplications, several low VOC materials arebeing qualified to meet the stringentengineering requirements of cleaning beforepainting, sealing, and adhesive bonding.
Because of the non-conventional nature ofsome of these candidates, each solventcandidate is tested for compatibility withaircraft metals, plastics, coatings, and sealant.This requirement includes corrosionresistance, degradation testing of plastics,elastomers and paint, and hydrogenembrittlement testing.
CORROSION INHIBITING COMPOUNDS
As a part of the corrosion control program forthe airplane fleet, the application of corrosioninhibiting compounds on most in-serviceairplanes has expanded in the internationalaviation industry in the recent year. The areasof application required by design on Boeingairplanes includes most of the interior metalaircraft structure. As a result, the increasedvolume of the corrosion inhibiting compoundsused in the factory has presentedenvironmental, personal health, andproduction-rate concerns. The corrosioninhibiting compounds are not regulated at thistime. However, as a part of the Boeing effortto reduce the total VOC emissions, low VOCcandidates are being evaluated.
The current corrosion inhibiting scheme usesa duplex system (one thin coat plus one thickcoat) for high corrosion-susceptible areas.Candidates with reduced VOC contents arebeing evaluated to replace the duplex system.In addition, Boeing is currently developing asingle-coat, heavy-duty system, which will beequivalent in the performance requirements ofthe duplex system.
DRY FILM LUBRICANTS
Dry film lubricants are used to facilitate theinstallation of fasteners. Currently thedevelopment activities on dry film lubricantshave been aimed at qualifying water reduciblelow VOC alternatives to solvent-based cetylalcohol fastener coatings. A manufacturingfeasibility study is scheduled and thealternative coatings will be implemented prior
227
to the regulatory effectivity date. Theregulatory target for the SCAQMD is at 250grams/liter.
ADHESIVE BONDING PRIMER
The structural adhesive bond primer is acritical component of bonded aircraftstructures. The material qualification containsstringent engineering requirements andtherefore exhaustive testing programs are setup for new candidates. Two batches ofmaterials have passed the qualification testing;the manufacturing evaluation is beingconducted for the commercial airplaneapplications. The regulatory target forSCAQMD is at 250 grams/liter, which is adramatic reduction from current 850grams/liter.
CONCLUSION
The regulatory emphasis has been onreduction of VOCs in coatings, i.e. primersand topcoats. A more recent shift in theregulatory focus has been towards loweringvapor pressures of the cleaning solvents.Local and U.S. government have set aggresivegoals in reducing environmentally hazardouschemicals. Material research and developmentis a long-term involved process. Boeing willcontinue to devote major efforts to developenvironmentally acceptable materials andprocesses.
228
WATER -REDUCIBLE POLYURETHANE ENAMELS:CANDIDATE LOW VOC AEROSPACE TOPCOAT FORMULATIONS
David J. SwanbergMechanical Systems TechnologyBoeing Defense and Space Group
Seattle, Washington
ABSTRACT
Recently, air quality regulations haveprompted significant efforts by coatings resinproducers and coatings manufacturers todevelop new low Volatile Organic Compounds(VOC) coatings. One approach has been todevelop water-borne polyurethanes havingperformance approaching that of solvent-bornecoatings.
Several water-borne resins were tested inwhite topcoat formulations versussolvent-borne controls (MIL-C-83286 and Deft1-COAT). Selected chemical resistance andflexibility tests were performed to determinewhether solvent-borne performance could beachieved in a water-borne formulation.Coatings were formulated with and withoutcrosslinkers to determine whether chemicalresistance could be improved withoutsacrificing flexibility.
In general, chemical resistance of thewater-bornes was not as good as that of thesolvent-borne controls," all formulations failed7 days' immersion in Skydro! 500B.Crosslinking tended to improve chemicalresistance but in some cases decreasedflexibility of the coatings. Anaziridine-crosslinked formulation performedbest but is undesirable because of potentialhealth risks associated with aziridines. Withadditional development, water-borne coatingsshould be able to satisfy performancerequirements of the military specification forhigh-solids polyurethane topcoats.MIL-C-85285.
INTRODUCTION
In recent years, regulations that restrictallowable emissions of Volatile OrganicCompounds (VOCs) from coating processeshave affected virtually every area of thecoatings industry. Regulatory agencies insome states are requiring significant reductionsin solvent content of organic coating materials,whereas others are limiting VOC emissionsfrom spray booth stacks. In either case,regulations have prompted significantdevelopment efforts by coatings resinmanufacturers and innovative new productsfrom coatings formulators. All efforts areintended to help reduce VOC emissions belowregulatory levels while still providingequivalent coating performance.
Reducing VOC while maintaining performancepresents a significant challenge to theaerospace industry. Aerospace coatings oftenprovide specialized functions in addition toprotection of the substrate and are routinelysubjected to extreme environments. Whileperformance is a primary consideration, otherfactors will also influence the success orfailure of low VOC replacement coatings.These include application characteristics,appearance, cost, and scale of coatingoperations (i.e., painting large assemblies).
The state of the art in performance aerospacetopcoats for many years has been thesolvent-borne, two-component polyurethanemeeting Military Specification MIL-C-83286.In terms of emissions, this material containstypically 600 grams/liter VOC. High-solidspolyurethane technology (MIL-C-85285) isavailable at about 420 grams/liter VOC withsimilar performance. The goal of our work isto provide equivalent performance in a liquid
229
coating with VOC emissions approachingzero. Water-soluble or 100% solids coatingscan achieve this. Our work to date hasconcentrated on water-borne polyurethanes andpolyesters (1,2) with potential for use intwo-component formulations that can beexternally crosslinked in the applied film.
The long-term goal of a solvent-free aerospacetopcoat is consistent with the concept ofminimizing hazardous waste by eliminatinghazardous components from manufacturingprocesses. Aside from the obvious benefits tothe health of workers and the environment,major incentives for eliminating organicsolvents from coatings include the following:1) avoided cost of purchase, operation, andmaintenance of emissions control equipment.2) reduced hazardous material inventory, and3) reduced cost and liability of hazardouswaste disposal.
The practical objective of the water-reduciblepolyurethane enamel project is to develop anultra-low VOC performance topcoat that canbe spray-applied with high transfer efficiencyequipment and cured at ambient conditions.Formulations must be applicable for originalfinish processes as well as for re-finishoperations and touch-up in the field. Thecoating should be available in a range ofcolors and reflectance ranging from high glossto very low gloss (camouflage applications).The resin backbone should be aliphatic forexterior durability and should containfunctionality that can be crosslinked forenhanced chemical resistance. Finally, onceapplied, the coating should be strippable byavailable means with minimal hazard toworkers and the environment.
EXPERIMENTAL APPROACH
The current phase of this work involvedtesting fully pigmented water-borne coatingsformulations for selected chemical resistanceand flexibility requirements of MIL-C-83286and MIL-C-85285. In coatings development,flexibility and chemical resistance are oftenconsidered mutually exclusive since the higher
crosslink density required to improve chemicalresistance tends to decrease flexibility of thecoating. Details of performance requirementsand test methods used in this work appear inTable I.
Materials
The water-borne resins used in this study weresupplied as single-component materialssuitable for formulation of lacquer-typecoatings. The water-borne resins tested aredescribed in Table II. All containedcarboxylate functionality to impart watersolubility and allow dispersion of the polymersin an aqueous phase. Water-compatiblecrosslinkers that were tested are shown inTable II!. Ail of the experimental coatingswere formulated both with and withoutcrosslinkers to determine whether increasedcrosslinking could significantly improvechemical resistance without sacrificingflexibility.
Two solvent-borne polyurethane coatings werealso used for comparison with water-bornecoating performance. They were: 1)MIL-C 83286 two-component polyurethaneand Deft 1-Coat (see Table IV). The 1-Coatmaterial was included since it represents arecent development in aerospace coatingstechnology whereby corrosion resistance isincorporated into the topcoat eliminating theneed for the traditional epoxy primer (3). Inaddition, this material is based on high-solidspolyurethane technology with a VOC of lessthan 420 grams/liter.
Formulation of VVater-Borne Coatings
The water-borne coatings used in this studywere formulated by first dispersing titaniumdioxide pigment in a portion of the resin.This dispersion was then let down withadditional resin to the desired ratio of pigmentto binder (usually 0.75 by weight). Additivessuch as dispersing aids, leveling agents, anddefoamers were rarely used since earlierwork showed that they tend to cause severecratering, crawling, and fisheye even at
230
low addition levels. Table V.
After the letdown, formulations were thinnedwith de-ionized water to a viscosity of about30 seconds on a #2 Zahn cup and the pH wasadjusted to between 8 an 9 with ammoniumhydroxide. Formulations were usually setaside overnight to de-aerate prior toapplication. When used, crosslinkers wereadded just prior to application and pH andviscosity were adjusted as reqired.
Sam pie Preparation
Test panels were typically 3 x 5 x 0.020 inch2024 aluminum alloy pre-treated with achromate conversion coating (Alodine 1200).Low temperature flexibility specimens werebare, deoxidized 2024 substrates only.
The test panels were primed (with theexception of the 1-Coat) with eitherMIL-P-23377 or MIL-P-85582 epoxy primersto dry film thicknesses of 0.008 to 0.0012inch. These were then topcoated with thevarious test formulations lo dry filmthicknesses of 0.0015 to 0.0027 inch. Aftertopcoatmg. the test panels were allowed tocure for 7 days at ambient conditions prior totesting.
RESULTS
General coating properties, flexibility, andchemical resistance results for all formulationstested appear in Table V. Water-borneformulas are listed by resin type andresin/crosslinker combination. Wherecrosslinkers were used, the addition ratio islisted in percent based on resin solids of thebase resin. VOCs for water-bornes werecalculated values, disregarding the volume ofwater in the coating (4). VOC valuesfollowed by an asterisk (*) indicate thatorganic coalescing solvent was added to theformulation, contributing additional VOC overthat present in the base resin. For comparisonpurposes, results of tests on solvent-bornepolyurethanes appear at the bottom of
The results of these evaluations indicate that ingeneral the water-bornes did not developgloss, hardness, and chemical resistance of thesolvent-borne coatings they hope to replace.Skydrol resistance was particularly low. evenwith crosslinkers added. Although Skydrolresistance is not a requirement for someaerospace applications, it is an excellentindicator for comparing resin systemperformance. The best Skydrol result for awater-borne in this study was the XW110 withCXI00 (6B after 7 days).
CYosslinking water-bornes with XL25carbodiimide showed interesting results whencompared to the uncrosslinked coatings madefrom the same base resins. Both XW110 andXW121 urethanes showed lower flexibility andgreater sensitivity to deionized waterimmersion with XL25 added. Skydrolresistance was very slightly improved but stillbelow specification requirements. R-9637 andR-9000 urethanes both showed slightlyincreased Skydrol resistance with XL25 added.R-9637 flexibility increased slightly while theresult for the R-9000 was mixed. Please notethat the XL25 addition levels of 17% werewell above the m a n u f a c t u r e r ' srecommendation of 5-10% on resin solids.The higher addition level was chosen based onequivalent weights of diimide functionality inthe XL25. to carboxyl functionality in the baseresins.
Lfncrosslinked 72-7230 polyester was notlisted in Table V since the material did notfully cure at ambient conditions. Bothcrosslinked versions, however, had excellentgloss and good flexibility. Chemicalresistance was lower than the water-borneurethanes. High crosslinker additions mayhave been partially responsible for the watersensitivity. This material has excellentpoential for good performance, high glossappearance, and very low VOC withoptimized crosslinking.
Three fairly low molecular weight diamines
231
were also tested as crosslinkers for the R-9000polyurethane. All showed slightly improvedSkydrol resistance without significantreduction in flexibility. An attractiveadvantage of the amine crosslinkers is thatthey do not contribute VOC to theformulation. The R-9000/BAPP formulashowed overall performance as good as any ofthe water-borne materials tested, with a VOCbelow 100 grams/liter.
The results for the solvent-borne coatings wereexcellent in all categories with the exceptionof Skydrol resistance. The MIL-C-83286material showed Skydrol resistance somewhatlower than expected while Skydrol resistanceof 1-Coat was slightly better than that ofwater-borne polyurethanes with crosslinkersadded.
Other results, related this work but not listedin Table V, indicate that baked cure conditionsfor water-bornes can enhance performanceover ambient cure. A white topcoatformulation using XW121 polyurethane with10% XL25 added was baked for one hour at300°F after solvent flash-off. This formed ahard coating (3H) that showed good Skydrolresistance (2B after 7 days) and also passedthe low temperature flexibility test. In anotherexample, two of tbz amine-crosslinked R-9000formulations showed improved lowtemperature flexibility after baking for onehour at 150°F. These were very low VOCformulations with little organic coalescingsolvent present. It is possible that improvedlow temperature flexibility was related toimproved polymer coalescence and filmformation at the elevated temperature.
high-solids, so!vent-borne polyurethane(1-Coat). All water-borne formulationsappeared to satisfy the MIL-C-85285 hydraulicfluid resistance requirement with no significantdegredation of the films. Full evaluation ofthe best water-borne formulations perMIL-C-85285 requirements should bepursued.
An aziridine crosslinked polyurethane showedthe best Skydrol resistance of all water-bornestested relative to MIL-C-83286 requirements.This formulation however, will not berecommended unless health risks associatedwith aziridines are demonstrated to be minor.One approach would be to find a relativelyinnocuous analog to the ethylene-imine ring.
A water-borne polyester was tested that, whencrosslinked. produced coatings withperformance similar to the water-bornepolyurethanes. The main advantage of thismaterial is the ability to formulate at very lowVOC and produce glossy, decorative coatingswith substantial chemical resistance. Suchcoatings may have aerospace applications fornon-flight hardware.
A very promising water-borne systemevaluated in this study was R-9000polyurethane crosslinked with amines. Thedistinct advantage of this system is the abilityto provide overall performance as good as anyof the water-bornes tested, but in very lowVOC formulations (white topcoat < 100grams/liter). It is likely that the chemicalresistance of this polyurethane could beimproved even further by increasedcrosslinking.
CONCLUSIONS
In general, low VOC water-borne coatingswith performance appraoching aerospacerequirements can be successfully formulatedusing available materials. Although Skydrolresistance was low for all water-borneformulas, overall performance of crosslinkedpolyurethanes was nearly as good as the
In general, crosslinked water-bornepolyurethanes and polyesters show potential toprovide durable, chemically resistant coatingsfor aerospace applications. In addition, theyhave potential to provide this performance atvery low VOC and should continue to bedevloped per aerospace requirements.
232
REFERENCES
1. R.E. Tirpak and P.H. Markush."Aqueous Dispersions of CrosslinkedPolyurethanes," Proceedings of the 12thWater-Borne Higher-Solids CoatingsSymposium. University of SouthernMississippi, 1985.
2. M. Glavas, M. Hage. and A. Heitkamp,"Waterborne Dispersion Resins for VeryLow VOC Coatings," Proceedings of the17th Water-Borne Higher-Solids CoatingsSymposium. University of SouthernMississippi. 1990.
3. C.R. Hegedus, "Development of aPrimer/Topcoat and Flexible Primer forA l u m i n u m , " R e p o r t N o .NADC-87016-60. Naval AirDevelopment Center. Warminster. PA.March 20. 1987.
4. ASTM D3960-89. "Practice forDetermining Volatile OrganicCompound Content of Paints andRelated Coatings," 1990 Annual Bookof ASTM Standards, Vol. 6.01.American Society for Testing andMaterials. Philadelphia. PA, 1990.
233
Table I. Selected Coating Performance Requirements
no blisteringhardness: n.c.no blisteringhardness: n.c.no blistering
hardness: 41 unitno blistering
' hardness: n.c.
no blisteringhardness: H-2H
no blisteringhardness: F-H
Skydrol 5(X)B7 days, room temp.
severe softening <6Bmars easily
severe softening <6B
softened 6B
severe softening <6Bmars easily
severe softening <6Bloss of adhesion
severe soficning <6Bmars easily
coating removal
severe softening <6Bloss of adhesion
severe softening <6B
severe softening <6Bloss of adhesion
severe softening <6B
severe softening <6B
severe softening<6B
severe softening<6B
softening 4UKioss: • 1 2 %
severe soficning <6Bgloss: -85%
Mil, H-83282
24 hours, I5O°F
light slainhardness: +1 unit
light stainhardness: -1 unit
no slainhardness: n.c.
very light slainhardness: n.c.
very light slainhardness: n.c.
It. slain, gloss: -20%hardness: -1 unitvery light slainhardness: n.c.
light stainhardness: +1 unitvery light stainhardness: n c.
very light stainhardness: n.c.
very light slainhardness: -1 unitvery light stainhardness: n.c.
very light slainhardness: n.c.
very light slainhardness: n.c.
very light stainhardness: no change
very lighi stainhardness: no change
LOW VOC COATING ALTERNATIVES
Mark D. SmithMaterials Engineering
Allied-Signal Inc.Kansas City, Missouri
INTRODUCTION
Allied-Signal Inc., Kansas City Division,(KCD) is a prime contractor for theDepartment of Energy. KCD is one of themanufacturing sites for the Department ofEnergy, (DOE) integrated weaponscomplex. KCD manufactures variouscomponents, many of which require surfacecoating in one way or another. Recentchanges in DOE directives and applicationof state and local air quality regulationscaused coating operations to become aconcern.
AREAS OF SURFACE COATINGCONCERN
The two main areas of surface coatingconcern at KCD are the painting operations,(both product related and painting done formaintenance or non-production applications)and the application of dry film lubricants.As is detailed later, both operations were,prior to this work, using materials that werehigh in volatile organic compounds,(VOCs). This wiil cover the two operationsseparately, mainly because the regulationstreated them differently.
PAINTING
Status Prior to July, 1989
Prior to July of 1989, we were vaguelyaware of state and local air qualityregulations. Some effort was being made todevelop a long range "site plan" to addressthe situation as it was understood at KCD.
We were totally unaware of the existence oflow VOC paints. A cursory investigationinto this area had been done a few monthsearlier regarding the currently used MilitarySpecification paints. At this time KCD usedapproximately 25 paints on a routine basiswith internal Material Specifications existingfor another 250-300 that could be used atany time that they might be needed. TheVOC content of none of these paints wasknown. Our paints in use were a mixture;some were Mil. Spec, paints and some werenot, depending on what our customerwanted. In addition, at this time all of thepainting VOCs were being released to theatmosphere through the paint bootn stacks.
REGULATION IMPOSED
The air pollution regulation imposed in July1989 was "Missouri Air Pollution Rule 10CSR 10-2.230, Control of Emissions fromIndustrial Surface Coating Operations."This regulation applies to locations emittingmore than 6.8 kilograms per day or 2.7 tonsper year of VOCs. KCD was regulatedunder the provisions for painting"Miscellaneous Metal Parts" since ourproducts were not considered to be coveredby any of the other classifications in theregulation. This classification of productshad an emission limit of 3.5 pounds ofVOCs per gallon of coating as applied.
Compliance Actions. July 1989
On July 7, 1989, KCD voluntarily halted allspray painting operations. The process ofdeveloping a "site plan" for obtaining state
237
and local EPA approval was accelerateddrastically. A survey was started of thecurrently used paints to determine theirVOC contents. Figure 1 shows the resultsof that survey with the individual paintsaveraged into groups by chemical type. Theresults clearly show that the paints beingused at that time were ail significantlyhigher in VOC content than the 3.5 poundsper gallon allowed by the regulations.
A comprehensive search for paints"compliant" to the regulations was started.To our pleasant surprise, a large percentageof the currently used paints had low VOCcompliant versions available or indevelopment. Some paints were beingreplaced by the paint industry with othertypes of paints which were equivalent orbetter in performance. Samples or smallbatches of these compliant and replacementpaints were obtained and VOC-testing begunon them. Figure 2 shows the VOC contentof selected replacement paints as claimed bythe vendors. Actual test data confirmedthese values within experimental error.Along a parallel path, while the new paintswere being tested for VOC content, sampleswere being test sprayed and their physicalproperties checked to determine suitabilityfor KCD production applications.
On July 25, 1989, a meeting was held withother DOE integrated weapons contractorsto appraise them of the situation anddetermine possible alternate sources forproduction painting within the weaponscomplex. The prime consideration in thismeeting was the ability of the othercontractors to meet their own state and localair quality standards. It was not consideredfeasible to send DOE contracted work to alocation that was also not in compliancewith air quality regulations. Productionschedules and other restrictions eliminatedthe possibility of sending parts to other DOEcontractors for painting.
An exhaustive survey was also conducted ofpainting industry contacts to determine
possible alternate sources of productionpainting outside the weapons complex.Over 500 companies were contacted. Theprime consideration in these contacts wasalso their ability to meet their state and localair quality standards. As with the DOEcontractors, it was not considered feasible tosend DOE contracted work outside thecomplex to a location that was also not incompliance with air quality regulations.This survey found less than five contactsable to meet local regulations, fewer thatwould take on KCD work and fewer stillthat could meet the paint quality standardsrequired for the product.
A small activated carbon filtering system forthe production painting area was obtainedbut local EPA approval was needed beforeit could be installed and tested. It was smallenough that it would not handle allproduction painting but could be used on alimited basis through a partitioned section ofan existing paint hood. Start-up of this unitwas included in the proposed site plan whichawaited EPA approval.
CONCERNS IN SWITCHING TOREPLACEMENT PAINTS
The mixture of Mil. Spec, and non-Mil.Spec, paints used previously was not anoptimum situation for the type of work doneat KCD considering the customer for ourproducts. Since changes had to be made inthe types of paints being used anyway, itwas considered preferable to strive for apaint inventory of all Mil. Spec, paints.
The VOC regulations were written, ingeneral, to reduce the amount of chemicalsexpelled into the air that cause the formationof ozone and smog at low altitudes. Forthis reason, several chlorinated andfluorinated solvents are considered by theregulations to be "exempt" or compliant foruse in paint formulations. Many of the newlow VOC compliant paints were formulatedusing 1,1,1-trichloroelhane as the principal
238
thinning solvent, (shown on Figure 2 withan asterisk beneath the paint type), in placeof the high VOC hydrocarbon thinners suchas methyl ethyl ketone or xylene. Thesolvents considered exempt for the VOCregulations are in fact considered ozonedepleters in the upper atmosphere by otherregulations. Significant effort is beingexpended at KCD, as at other DOEcontractor sites, to minimize or eliminate theuse of chlorinated and fluorinated solventsfor that reason. In this case, two majorefforts to reduce air pollution wereessentially incompatible. Therefore it wasalso considered preferable to strive for apaint inventory as free from chlorinated andfluorinated solvents as possible.
There were a number of paints for whichthere were no direct low VOC replacements,including lacquers, wash primers andTeflon'-based paints. It was decided that inmost cases this impediment to changingpaint systems could be worked around sincethe number of products requiring thesepaints was small. In addition some concernexisted over the need for new paintingtechniques and equipment to properly applythe new paint systems. This was alsoconsidered a small obstacle to overcomesince without the new low VOC paints,painting operations could not resume.
use has been significantly reduced from 25high VOC versions down to approximately5 low VOC versions. A larger percentageof the paints now being used are also basedon Mil. Spec's and Federal Color Standardnumbers. Both of these factors eases ourpurchasing and qualification testing of thepaints. In addition, the inventory is easierto maintain, both in size and complexity andfrom the standpoint of shelf-liferequalification of fewer paints. So, alongwith reducing our VOC emissions, theamount of waste from unused, out of datepaints will, in the future, be reduced.
ANTICIPATED FUTURE PAINTWORK
KCD will continue to emphasize the use ofhigh-solids, low VOC polyurethane paints.Those meeting Mil-C-85285 have workedwell in replacing high VOC urethanes aswell as other types of paints such as epoxiesand acrylics. The reduction or eliminationof chlorinated solvent based coatings will becontinue to be pursued. In one case, achlorinated solvent based, zinc chromateprimer was replaced with a Mil. Spec,waterborne epoxy primer with satisfactoryresults. Other chlorinated solvent basedpaints have been replaced with the Mil-C-85285 urethanes.
PRESENT KCD PAINTING STATUS
Most spray painting applications are nowusing low VOC paints. In rare instanceswhere substitute paints are not yet available,limited use of high VOC materials is madewithin the special spray booth, the stackemissions being routed thiough activatedcarbon filters as approved in the new EPAsite plan. This filtering system has provedto be greater than 95% efficient in removingVOCs from that air stream and has notrequired changing over the past year.
The number of paint systems now in routine
One major program remains to be changedover from a high VOC system of primer,aluminum pigmented epoxy and a clearepoxy top-coat to low VOC waterborneprimer and low VOC urethane in place ofthe epoxy. Problems with the clear urethanecontinue to be worked through.
In anticipation of tighter EPA regulationsand Allied-Signal corporate directives,consideration is being given to the activatedcarbon filtration of all the paint booths,including those using low VOC paints,instead of just the special booth now usinghigh VOC paints. This would allow more
239
flexibility in painting and reduce theemissions closer to the "as low asreasonably attainable" level desired by thecorporate policies.
Alternate coating methods are beingevaluated including powder painting andcoating by wet chemical electrophoresis. Ajoint study of powder coatings applied by anoutside contractor is underway with SandiaNational Laboratories. A laboratory-scale,batch-type powder facility is being set-up atKCD to serve the DOE weapons complex.
DRY FILM LUBRICANTS
Present Dry Film Lubricant Status
Most of the dry film materials presentlyused at KCD are high VOC, ranging from6.42 to 8.12 pounds per gallon as applied,(Figure 3). The current Missouri regulationregarding emissions from surface coatingoperations does not regulate the applicationof dry film lubricants. However,considering the dynamic nature of air qualityregulations and the corporate policies toreuuce emissions, considerable work isunderway to evaluate low VOC dry filmlubricants or alternative lubrication methodsthat would be suitable for our applications.
Concerns in Switching Dry Film
The application techniques and equipmentrequired to apply some of the dry filmsbeing considered are significantly differentthan what is now is use and will requiresubstantial changes in operations. Some ofthe technologies being investigated requirelicensing and some are performed only bythe vendors. In certain cases, such asclassified or sensitive parts, this would beprohibitive.
The performance requirements of the presentmaterials are not well defined, therefore itwill be hard to define what properties asuitable replacement should have. Many of
the dry films were put into use to solve aparticular problem on a particular part yearsago and their use has been continued withoutthe problem and solution being fullyunderstood.
Some of the new low VOC versions ofexisting dry films are relatively unproven inactual use, similar to some of the low VOCpaints, which causes some trepidation inadopting their use in an ongoing productionenvironment. Substantial testing will berequired in various modes of lubrication forseveral different types of assemblies.
KCD DRY FILM ACTIONS
A joint Sandia National Labs./KCD grouphas begun studying dry film lubricants. Thedry film market is being surveyed forpossible low or zero VOC materials andtechnologies. Possible alternatives identifiedat this point include: Dicronite*, sputteringof MoS2, sputtering MoS2 followed by iontreatment, electropiioretic application ofMoS2 rich coatings, increased hardnesscoatings, Microseal* and low VOC versionsof the presently used E/M materials. Adatabase of the dry films being considered isbeing developed for future referenceincluding vendor data and manufacturer'sMaterial Safety Data Sheets.
The study is also attempting to define dryfilm performance requirements for existingand planned assemblies. This covershundreds of individual parts and assembliesand various different modes of lubricationneeded, some within the same assembly. Inaddition, the study will need to determineobjective inspection techniques forqualifying and comparing the dry films thatwill be tested. Several short term tests arebeing made on individual parts or assembliesand are generating useful data. However,the large number of parts that use dry filmssuggests that a broad study that will yieldconsistency in the future use of dry films isin order.
240
CONCLUSIONS
Although a great deal of progress has beenmade at KCD in the effort to reduce VOCemissions from surface coating operationsand the state and local regulations are beingcomplied with, work has not and will notstop in trying to further this effort. FutureEPA regulations concerning air quality andAllied-Signal corporate policies of trying toreach emissions levels of "as low asreasonably attainable" assure that this workis not completed.
241
Figure 1. VOC's of Selected High VOC PaintSystems, (as applied)
3 6-
uO s->•Oe 4-oaI 3.u :S '-OA
o
"5
iI
III
I? -1
o-K Uretbane EpoxyEnamel
Acrylic AJkyd EpoxyPrimer
Lacquer ZincChrm. TeflonPrimer
WashPrimer
Paint Systems
Figure 2. VOC's of Selected ReplacementPaints, (vendor data)
AJkyd WaterborncEpoxy Primer
Zinc Chrm.Primtr
EpoxyPrimer
Epoxy Urtthane
Paint Systems
242
Figure 3. VOC's of Selected E/M Corp.Dry Films. (as applied)
Type-A EL-620 99-A 4396-S
Dry Film Lubricants
243
DUAL CURE PHOTOCATALYST SYSTEMS
Steven J. Keipert, Ph.D.Corporate Research Laboratories
3M CompanySt. Paul. Minnesota
The 3M Dual Cure process is a method forproducing radiation cuied polymercompositions via a solventless coating process.The process involves the simultaneous cure oftwo different monomer types. The first typemay be either an epoxy monomer, or thepolyol and polyisocyanate precursors to aurethane polymer. These are polymerized byuse of an iron based organometallicphotocatalyst which gives a Lewis acid uponirradiation. The second monomer type consistsof acrylate monomers, which are polymerizedvia a tree radical photoinitiator to giveacrylate polymers. The two polymerizationtypes procede separately to give substantiallyindependent, interpenetrating polymernetworks. The properties of the resulting DualCure compositions are often found to besuperior to either component alone.
The work described in this paper wasdeveloped under a 3M/U.S. Department ofEnergy (DOE) cost share contract titled"Industrial Gaseous Waste Reduction" fundedby the DoE Office of Industrial Technologies.The research was performed in the 3MCorporate Research Science ResearchLaboratory, in cooperation with other 3Mlaboratories. The contract consists of threephases. Phase I, which is complete, involveddemonstration of technical feasibility on alaboratory scale. Several catalyst combinationswere evaluated in a model curablecomposition, and the physical properties of theresulting cured polymers were analyzed andcompared. Phase II, which is nearingcompletion, consisted of a pilot scaledemonstration, primarily involving protectivemetal coating applications. Phase III will
involve the demonstration of the technology infull scale commercial applications incooperation with internal and externalindustrial partners.
The objective of the Dual Cure program is toeliminate solvents from coating compositions,while simultaneously maintaining or improvingthe physical properties of the coating and theprocessing conditions used to prepare them.This will lead to reductions in solvent use,with a resulting reduction in energyconsumption and solvent emissions.
Projections for the use of various coatingtechnologies in the future show a reducedmarket share for solvent coating techniques.Most o'r this will be replaced by high solids,waterborne, and powder coating methods.Although decreasing as a percentage of thetotal, market growth means that, in absoluteterms a substantial amount of solvent coatingwill remain in absolute terms. Radiationcoating techniques offer an alternativereplacement for solvent based systems.Projections show this sector growing, butnever gaining a substantial market share. Wefeel that a major reason for this trend is theinability of currently existing U.V. curingprocesses to deliver the materials performancerequired by many coatings applications. Webelieve that the 3M Dual Cure process, whichallows an expansion of the number ofradiation curable monomer types, will enhancethe production of high performance coatings.This should facilitate the increased use ofradiation curing, and a further reduction insolvent use. Preliminary projections of energy-savings resulting from the adoption of thistechnology show savings of 1.5 X 1013
245
btu/year by the year 2010. combination with a radical photoinitiator.
We are primarily interested in two types ofDual Cure compositions. The first results fromthe combination of epoxies with acrylates togive epoxy/acrylate compositions. The secondresults from the combination of a urethane andan acrylate to give a urethane / acrylate. Theseare not to be confused with the acrylatedurethanes, which are photocurable, acrylatefunctionalized urethane oligomers whichpolymerize by a free radical acryiate typepolymerization. In the Dual Cure process,both polymers are formed directly from theirrespective monomers, with limitedinterconnection of the two polymer networksformed.
Five different catalyst combinations wereinvestigated in phase I of the contract.Systems 1 through 3 use a cationic ironcatalyst. These compounds are relatively inertin the dark, and are air-and-moisture stable.Upon irradiation, they produce acoordinatively unsaturated, Lewis acidic ironspecies through the initial loss of the areneligand. This is an active thermal catalyst forthe polymerization of epoxies and urethanes.They are also able to initiate free radicalpolymerizations, although the mechanism forthis is not clear at present. These ironcatalysts were used alone as in catalyst system2, or in combination with other radicalphotoinitiators such as benzil dimethylketal incatalyst system 1, and oxidants such asdiphenyliodonium salts in catalyst system 3. Aneutral binuclear iron species was alsoinvestigated in catalyst system 4. Although itwas found to be an active photocatalyst, itsreactivity in the dark limited its usefulness andthis system was abandoned early in phase I.Catalyst system 5 consisted of the controls.For epoxy/acrylates this consisted of acombination of a triarylsulfonium salt, whichgives a protic acid upon photolysis, incombination with a radical photoinitiator. Forurethane/acrylates, the control consisted ofdibutyltin dilaurate, a commonly used thermalcatalyst for urethane polymerization, in
The urethane portion of the urethane/acrylateDual Cure formulations are formed from amixture of a pilyisocyanate, preferably analiphatic polyisocyanate, and a polyol. Prior tothe discovery of the Dual Cure catalysts, thesematerials could only be polymerized bythermal processes. The two components willpolymerize alone, but the rate ofpolymerization is enhanced greatly byincorporation of a catalyst such a dibutyltindilaurate. Thermal polymerizations of this typehave several disadvantages which will bediscussed later. The acrylate portion of theurethane/acrylate Dual Cure formulations isgenerally a mixture of mono andpoly functional acrylates. The optimum ratio ofurethane to acrylate for protective coatingapplications his been found to beapproximately "0:30. The high viscosity ofurethane formulations has often limited theirprocessability without the addition of thinningsolvents. Incorporation of the acrylate resuitsin reduced viscosity, and improvedprocessability.
Because of the reactivity of the urethaneprecursors even in the absence of catalyst,these materials require storage as twocomponents, which are mixed prior to use.The Dual Cure compositions do offerprocessing advantages in comparison withthermally catalyzed urethane polymerizations.Figure 2 shows the effect of the differentcatalyst compositions on the cure speed of aurethane/acrylate formulation. These are thecure times to give a tack-free film (set tocotton) at 100 degrees Celcius after U.V.irradiation. Catalyst concentrations wereadjusted to give comparable cure speeds for allsystems. The iron catalyst containingcompositions (1-3) contained 300 ppmcatalyst, and the control (5) contained 40 ppmof dibutyltin dilaurate. Free radical initiatorsand oxidants were present at 0.1 % in theappropriate systems. Figure 3 shows thepotlife of these same formulations at roomtemperature in the dark. The useful potlife is
246
defined as the time required to double thecoating viscosity of the formulation. Thesedata show that although all compositions gavecomparable cure speeds, the compositionsbased on the Dual Cure catalysts ( 1-3 ) hadpotlives in excess of 12 hours, comparable touncatalyzed compositions. In comparison, thecontrol composition containing dibutyltindilaurate had a potlife of less than 1/2 hour.This demonstrates an advantage in using thephotoactivated Dual Cure catalyst in urethanesystems. The necessary tradeoff bc'"een curespeed and potlife found when using thermalcatalysts is eliminated. With Dual Cure, curespeed may be adjusted as desired, with noeffect on the potlife.
Synergistic effects are often observed withrespect to the physical properties of DualCure compositions. Figure 4 shows a tensiles t reng th compar i son be tween aurethane/acrylate composition and theidentically cured urethane and acrylateportions alone. The acrylate has a highertensile strength than the urethane. The DualCure composition is seen to retain most of thetensile strength of the stronger acrylateportion, even though it consists of only 30%acrylate. Figure 5 shows a comparisonbetween the elongation to break behavior ofthe same three compositions. The urethane hasa much higher elongation to break than thestrong but brittle acrylate. The Dual Curecomposition retains most of the elongationproperties of the urethane portion. Figure 6shows a comparison of the toughness (energyto break) of the three materials as measuredby the area under the stress/strain curve. Sincetoughness is a function of both high tensilestrength and high elongation to break, it is notunexpected that the Dual Cure composition isfound to be substantially tougher than either ofits components separately.
ERL-4211. These are combined with amixture of mono and polyfunctional acrylates.The most useful monomer ratio for protectivemetal coating applications was found to be60:40. Since neither of these monomers willpolymerize without activated catalyst orinitiator, they are stable one-componentmaterials as long as they are stored in thedark. These compositions are generally of lowviscosity and easily coated.
The synergistic effects observed in theurethane/acrylate system are also seen with theepoxy/acrylates. Figure 7 shows a comparisonof the tensile strength of the Dual Curecomposition with the separate epoxy andacrylate portions. The epoxy/acrylate is seento maintain a significant portion of the tensilestrength of the stronger epoxy component.Figure 8 shows a similar comparison of theelongation to break data. In this case the DualCure composition outperforms both of theseparate components. The toughness data isshown in figure 9. Once again, the Dual Curematerial is found to be significantly tougherthan either component separately.
In summary, it has been shown that the DualCure process offers several advantages withrespect to materials, properties andprocessing. The urethane/acrylate system.when compared to conventional urethanes.gives a lower viscosity, more easily coatableformulation, and avoids the cure time/potlifetradeoff. In addition, it has been demonstratedthat both the urethane/acrylate andepoxy/acrylate Dual Cure systems exhibitbeneficial synergistic effects, retaining the bestphysical properties of each component. PhaseIII of the contract will extend this technologyinto full scale commercial demonstrations.
The second class of Dual Cure materialswhich were investigated are the epoxy/acrylates. The preferred epoxies are thedifunctional cydoaliphatic epoxies such as
247
Figure 1: Catalyst systems.
af>//W/.
; _ ^ ' ^
lii
P
Figure 2: Cure time as afunction of catalyst system.
Uretnane/ Aery lateEffect of cnotocataiys:
ctiwt» ta ncurs
78!-
51-
IP
Figure 3: Potlife as afunction of catalystsystem.
248
iensne Prcperties CL' retnane/Acylate C mpos'.crs
m
Figure 4: Tensile strengthcomparison with purecomponents.
Tensile Propernes C o m p a r e n
Uretnane/Aery late Compos .tens
Compositcn
Figure 6: Toughnesscomparison with purecomponents.
Tensile Propertiesjret^ane/ Aery lateelT Elongglian at Brest
so I-
«f4 0 (•
2 111
3l—
Figure 5: Percentelongation to breakcomparison with purecomponents.
Tensile Properties CoTps i -scEpoxy/ Acv'iste Compos 'IO'-S
lensie Slrsngtl.uOe
Figure 7: Tensile strengthcomparison with purecomponents.
Figure 8: Percentelongation to breakcomparison with purecomponents.
Tensile Properties ConpaEpoxy/Acryiate Compos
Compos it en
Figure 9: Toughnesscomparison with purecomponents.
249
SCREENING OF VOC CONTROL TECHNOLOGIES:TECHNOLOGY OPTIONS AND COMPARATIVE COSTS
Dr. Victor S. EnglemanScience Applications International Corporation
San Diego. California
Emissions of volatile organic compounds("VOCs) can be controlled by recoverymethods and destruction methods. Recoverymethods include adsorption, absorption,condensation and separation options.Destruction methods include incineration andbiodegradation. A screening model has beendeveloped to perform preliminary evaluationson feasibility and relative costs of VOCcontrol technologies. Tne combinationsevaluated by the model are as follows:
•carbon adsorption with off-siteregeneration destruction
•carbon adsorption with on-site steamregeneration
•carbon adsorption with on-site inert gasregeneration and reverse Rankine solventrecovery
•carbon adsorption with on-site inert gasregeneration and reverse Brayton solventrecovery
•carbon adsorption with decoupled inert gasregeneration and reverse Brayton solventrecovery
Recovery technologies represent anopportunity to save both energy and money,since solvent recovered may be reused in thesame or a similar application. Each specificcase requires separate evaluation because ofthe specific requirements of the individualapplication. In some cases, it may not bepossible to use recovery technologies becauseof technical impediments.
The screening model has been used for theinitial evaluation of control technologies forspecific cases. The results of calculationsfrom this screening model are presentedbelow, in graphic form, to illustrate theregions where control technologies are cost-competitive for the specific case of tolueneemissions. Figures 1-3 represent the regionsin which each of the control technologiesrepresent the least-cost control option, giventhe input used for this panicular test (the mostimportant of which is the value for therecovered solvent of $0.20 per pound).Annualized costs were used as the basis forthese figures.
Figure 1 indicates the regions in whichadsorption-based technologies are the least-cost control method. Adsorption with off-siteregeneration is most favorable at lowconcentrations and low flow rates. Adsorptionwith on-site steam regeneration is mostfavorable at high flow rates and low tomedium concentrations. Adsorption withmobile inert-gas regeneration fills the gapbetween the first two methods. Adsorptionwith on-site Brayton regeneration overlaps theregion for steam regeneration, but extends tohigher concentrations. The region of higherconcentrations is also covered by adsorptionwith on-site Rankine regeneration. Since theBrayton system can achieve lowertemperatures more readily than the Rankinesystems, it tends to be more favorable for
streams requiring very low condensationtemperatures.
Figure 2 indicates the regions for incinerationtechnologies. Because of the good recoveryvalue used for toluene and the fact that tolueneis readily recoverable, even from steamregeneration, incineration technologies wereamong the least-cost control option only in alimited range. In the vicinity of 10.000 cubicfeet per minute and 100 ppm, regenerativeincineration was among the least-cost controloptions. If technical factors in the system hadmade adsorption technically unfeasible,incineration would have covered with a widerrange. Incineration is technically feasibleacross a broad range.
Figure 3 indicates the regions for condensationtechnologies. Both direct Brayton (also calledopen-cycle Brayton) and indirect Rankine(closed-cycle Rankine) condensationtechnologies were the least-cost controloptions at high concentrations across the entirerange of flow rates. The direct Brayton doesbecome limited in applicability above 10 or12% because it becomes difficult to condensethe VOCs by direct expansion.
The graphs shown do not apply universally.Higher or lower boiling points, difference inreactivity, and differences in recovery costs,among other factors, influence the regions inwhich contro) technologies are most favorable.
The model outputs include capital costs(equipment and installation), operating costs(labor, energy, maintenance, supplies,taxes/insurance, overhead and solventrecovery credit) and total annualized costs.Since this is a screening model, results areexpressed as ranges rather than as absolutenumbers. The ranges represent theuncertainty of the calculations, based on site-specific factors not considered.
The model is still under development andinterested parties are encouraged to contact theauthor at the above address or by phoning(619) 587-9071. ext. 169. Input" for cost
252
information, suggestions for inclusion ofadditional features, case studies for calibrationand general interest are solicited.
253
EXAMPLES OF RANGES FOR CONTROL TECHNOLOGIESTOLUENE (1 OF 3)
OTHER FACTORS AFFECTING VOC CONTROL SELECTIONoVALUEo # COMPONENTSo CONDENSATION TEMP
oREACTMTYo WATER CONTENToMBCBIUTY
100,000 n=
10,000
i1,000
dCO
100
ADSORPTION/MOBILE WEFT REGENERATION
REOENERATIOM
ON-SITC8TEAMREQENERATUN
10 100 1,000 10,000 100,000
VOC CONCENTRATION (PPM)
FICURE 1
ir
LUQ
EXAMPLES OF RANGES FOR CONTROL TECHNOLOGIESTOLUENE (2 OF 3)
OTHER FACTORS AFFECTING VOC CONTROL SELECTION
100,000
10,000
LJL
ow 1,000
d
100
oVALUEo # COMPONENTSo CONDENSATION TEMP
o REACTMTYo WATER CONTENTo MISCIBIUTY
10 100 1,000 10,000 100,000
VOC CONCENTRATION (PPM)
FIGURE 2
254
EXAMPLES OF RANGES FOR CONTROL TECHNOLOGIESTOLUENE (3 OF 3)
OTHER FACTORS AFFECTING VOC CONTROL SELECTION
cc
LLJ e -Q 2
oVALUEo * COMPONENTSo CCT«S«AT1ON TEMP
oHEACTIVrTYo WATER CONTENT
100,000
10,000
1,000
100
/ DIRECT
/
N. RANKME
10 100 1,000 10,000 100,000
VOC CONCENTRATION (PPM)FICURE 3
255
Section V
TREATMENT FOR ENVIRONMENTALLY SAFEDISPOSAL OF TOXIC SOLVENTS
GENERAL OVERVIEW OF HAZARDOUS WASTE INCINERATION
Philip C. Lin, Ph.D.Risk Reduction Engineering Laboratory
Office of Research and DevelopmentU.S. Environmental Protection Agency
Cincinnati, Ohio
ABSTRACT
Improper disposal of hazardous materialsgenerated by industries in the U.S. in the pastseveral decades has prompted the Congress toenact a series of environmental laws to governand cleanup hazardous waste. Conventionalmethods of waste disposal, such as landfillingand deep-well injection, arebeing restricted.Waste minimization, recycling, and treatmentare being vigorously pursued. Although thereare many different treatment methods, such asp h y s i c a l / c h e m i c a l t r ea tmen t andstabilization/solidification methods,incineration is the most universally applicable.Incineration, which is capable of the highestdegree of waste destruction, is considered tobe able to destroy the broadest range ofhazardous waste. However, incineration mayproduce residuals in the ash and emitundesired trace amounts of unburnedhazardous waste, incomplete combustionby-products, metals, and particulates. Thispaper is to review the current technologyavailable and practice of incineration ofhalogenated hydrocarbons; regulations andstandards for incinerators, boilers, andfurnaces; and results from field evaluations ofthe incineration of various hazardous wastes inthe early 1980's. Recent incineration researchby the USEPA is also discussed.
INTRODUCTION
Improper disposal of hazardous materialsgenerated by industries in the U.S. in the pastseveral decades has prompted the Congress toenact a series of laws to control the hazardouswaste. The first federal standards for controlof incineration emissions were enacted under
provisions of the Clean Air Act (CAA) in1970. Only paniculate emissions fromincineration sources were regulated. In 1976,the Resource Conservation and Recovery Act(RCRA), which defined and identifiedhazardous waste and provided provisions forcontrolling the storage, transport, treatment,and disposal of hazardous waste, was enacted.Incineration of polychlorinated byphenyls(PCBs) was controlled in May 1979 underrules established under the Toxic SubstancesControl Act (TSCA) of 1976. The rulesprohibited further manufacture of PCBs afterJuly 2, 1979, set limits on PCB use incommerce, and established regulations forproper disposal. In 1980, a national fund toassist the clean-up of uncontrolled waste sitescreated by poor disposal practices wasestablished under the ComprehensiveEnvironmental Response, Compensation, andLiability Act (CERCLA). On June 24, 1982,the final incinerator standards of performancewere published in the Code of FederalRegulations (CFR) under 40 CFR 264.343.The Hazardous and Solid Waste Act (HSWA)of 1984 amended and reauthorized RCRA toestablish a strict timetable for restrictinguntreated hazardous waste from land disposal.By 1990, most wastes were restricted andrequired to be pretreated by BestDemonstrated Available Technology (BDAT)before disposal. The 1986 SuperfundAmendments and Reauthorization Act (SARA)reauthorized the Superfund programs andgreatly expanded the provisions and funding cfthe initial Act. SARA also emphasized theneed to select clean-up technologies that wouldresult in a permanent decrease in the toxicity,mobility, or volume of hazardous materials.The impact of these various statutes will be asignificant modification of waste managementpractices.1
257
REGULATIONS AND STANDARDS efficiency.
Environmental regulations and standardspromulgated under various federal statutes arechronologically listed as follows:
Regulations and Standards for Incinerators
1. Clean Air Act (CAA. 1970)
Paniculate concentration emitted fromall incinerators constructed afterAugust 1971 must not exceed 0.08grains/dscf corrected to 12% CO2 inthe stack gas.
2. Resource Conservation and RecoveryAct (RCRA. 1976)
The final incinerator performancestandards were published on June 24,1982, listed in the Code of FederalRegulations (CFR) under 40 CFR264.343. In order to receive a RCRApermit to run the facility, it must attainthe following performance levels:
a. Principal organic hazardousc o n s t i t u e n t s ( P O H C s )designated in each waste feedmust be destroyed and/orremoved to a destruction andremoval efficiency (DRE) of99.99% or better.
b. DRE is defined by thefollowing formula:
DRE (%) = 100 x(Win- Wout)AVin
where Win = POHC feedrate to theincineratorWout = POHCemission rate from the
incinerator
c. Gaseous hydrogen chloride(HCL) emissions must eitherbe controlled to 4 Ibs/hr orless, or be removed at 99%
d. Paniculate emissions must notexceed 180 mg/dscm correctedto 7% O2 in the stack gas.The measured paniculateconcentration is multiplied bythe following correction factor(CF):
CF = (21 - desired O2) /(21 -measured O2)
The POHCs are listed in RCRAAppendix VIII which consists of 390organic and inorganic compounds firstpublished in the May 19, 1980 FederalRegister and updated semiannually in40 CFR 261. The Appendix VIIIconstituents, which are in the highestconcentration in the waste feed and aremost difficult to incinerate, are to beselected as POHCs in the trial burn.The heat of combustion of the POHCswas initially suggested as a measure ofcompound incinerability, but this hasbeen replaced by an EPA-generatedIncinerability Ranking Index.
To obtain a RCRA permit, a trial burnto meet these requirements must beconducted. The standards also specifythe requirements for waste analysis,operation, monitoring, inspection, andprocedures by which permits will begranted.
Toxic Substances Control Act (TSCA,1976)
Whenever disposal of PCBs isundertaken, they must beincinerated, unless the PCBconcentration is less than 50 ppm. Ifthe PCB concentration is between 50and 500 ppm, the waste can be used asa fuel in a high-efficiency boiler. ForPCBs exceeding 500 ppm, they mustbe incinerated with a combustionefficiency no less than 99.9% and a
258
DRE of 99.9999% and meet a numberof specific incinerator operatingconditions, such as combustiontemperature, residence time, stackoxygen concentration.1 Thecombustion efficiency (CE) is the ratioof the following:
CE = CO/(CO + CO2)
Further manufacture of PCBs wasprohibited after July 2, 1979.
4. Dioxin Rule2 (promulgated onJanuary 14, 1985 under RCRA)The incinerator must be capable ofachieving 99.9999% C :E forchlorinated dioxins or similarcompounds.
a. Reauthorized the Superfund programand expanded the provisions andfunding.
b. Emphasized the need to selectclean-up technologies that will resulta significant decrease in thetoxicity, mobility, and/or volume ofhazardous materials.
On May 6, 1987, EPA proposed rules tocontrol the burning of hazardous waste inboilers and furnaces. They were promulgatedin December, 1990. The followingperformance standards must be met:
1. 99.99% DRE for each selected POHCin the waste feed.
2. 99.9999% DRE for dioxin and otherextremely toxic substances.
3. Risk-based emission limits forcarcinogenic metals (arsenic.beryllium, cadmium and hexavalent
Incineration
Incineration is a process that employsdecomposition via thermal oxidation at hightemperature to destroy the organic waste. In1981, EPA estimated that the annualhazardous waste generation was about 250million metric tons (MMT) which wasconfirmed in separate studies by theCongressional Office of TechnologyAssessment (OTA) in 19833 and theCongressional Budget Office (CBO) in 1985."Of this, approximately 47 MMT per yearcould have been incinerated. However, theCBO projected that only 2.7 MMT were
259
actually incinerated in 1983. A 1981 EPAstudy also estimated that approximately 3.8MMT of hazardous waste was disposed inover 1300 industrial boilers and furnaces.5
These numbers are illustrated in Figures 1 and2. This estimate does not include wastesgenerated from uncontrolled hazardous wastesites. The implementation of the HSWA(1984) land disposal restriction regulations andgenerator concern for long-term liability willresult in increased utilization of incinerationfor ultimate disposal.
Thermal Destruction Systems
Three types of thermal destruction systems(incinerators, boilers, and furnaces) arediscussed below.
1. Incinerators
Various types of incinerators areavailable for handling differentphysical forms of hazardous waste.An incinerator typically includes aprimary and a secondary combustionchamber. Pollution control devices forreducing particulate, hydrogenchloride, sulfur oxides, metals andother emissions may be added. Themost common types of incineratordesigns are as follows:
a. Liquid injection
Liquid wastes are blended andthen pumped into thecombustion chamber throughatomizing devices.
b. Rotary kiln
Rotary kilns generally consistof a rotating kiln and anafterburner. Afterburners areused to ensure completecombustion of flue gasesbefore their treatment for airpollutants. The rotary kilnincinerator can generally be
e.
used for the destruction of solidorganic hazardous waste.
Fixed hearth
FixeJ hearth is also called acontrolled air, a starved air, ora pyrolytic incinerator. Thestarved air condition causesmost of the volatile fraction tobe destroyed in the primarychamber with 50-80% ofstoichiometric air and chambertemperature at 1200-1800F.The resultant smoke andpyrolytic products pass to thesecondary chamber whereexcess air is injected tocomplete the combustionreactions. This type ofincinerator generally has asmaller capacity than liquidinjection or rotary kilnincinerators.
Fluidized bed
Fluidized beds are eithercirculating or bubbling bedd e s i g n s . O p e r a t i n gtemperatures are maintained inthe 1400-1600F range andexcess air requirements rangefrom 100 to 150 percent. It isgenerally used for sludges orshredded solid materials.Fluidized bed incineratorsoffer high gas-to-solid ratios,high heat transfer, and uniformtemperatures through out thebed.
Fume incinerationFume incinerators are used todestroy gases or fume wastes.Wastes are injected by pressureor atomization through theburner nozzles.
260
2. Boilers and Furnaces
Boilers are constructed to producesteam. The concept of disposing ofhazardous wastes in boilers hascentered around industrial boilers.Wastes are cofired with conventionalfuels in these boilers. Furnaces usedinclude ceme.it kilns, blast furnacesand smelters.
Air Pollution Control Devices (APCD)
Combustion gases generally need to be furthertreated in an air pollution control system. Thepresence of chlorine and other halogens in thewaste requires a scrubbing or absorption stepto remove HC1 and other halo-acids.Paniculate emissions require collection devicesof moderate efficiency to meet the RCRAemission standard (0.08 grains/dscf). Themost common system used for the airpollution control is a quench, followed by aventuri scrubber (particulate removal), apacked tower adsorber (acid gas removal) anda demister.
1. Wet Scrubber Systems
These may include spray towers,centrifugal scrubbers, and venturiscrubbers, etc.
a. Venturi scrubbers
Venturi scrubbers involve theinjection of a liquid, usuallywater or a water/causticsolution, into the exhaust gasstream as it passes through thethroat. The fine-atomizedliquid entrains fine particlesand a portion of the absorbablegases from the gas stream.They are reliable and simple tooperate but they often requiresignificant pressure drop acrossthe throat (60-120 in. ofwater)1 which represents a
significant percentage of totalcost of operation.
b. Packed bed scrubbers
These are vessels filled withpacking material such aspolyethylene saddles. Theliquid is fed to the top of thevessel, with the gas flowingcountercurrent to it. Theliquid wets the packedmaterial to remove the acidgas from the stack gases.
c. Plate tower scrubbers
They are similar to packedbed scrubbers, relying onabsorption for the removal ofcontaminants. They aremostly used with liquidinjection incinerators forabsorption of soluble gaseouspollutants such as HC1 andSOx. For rotary kiln or fixedhearth facilities with high ashfeeds, venturi scrubbers arealso used in series withpacked bed scrubbers.
2. Dry Sorbent Injection (DSI)
Dry alkali sorbents are injected intothe flue gas downstream of thecombustor outlet and upstream of theparticulate matter control device toform salts. The removal efficiencydepends on flue gas temperature,sorbent type and feed rate, and theextent of sorbent mixing with the fluegas.
3. Wet Electrostatic Precipitation (ESP)
4. Ionizing Wet Scrubbers (IWS)
5. Fabric Filters
ESP, IWS, and fabric filters are used
261
for small particle removal.
6. Spray Dryer (SD)
Lime or limestone slurry is injectedinto the SD, the water in the slurryevaporates to cool the flue gas andthe Iime/CaCO3 reacts with acidgases to form salts that can beremoved by a paniculate mattercontrol device. The key design andoperating parameters that affect SDperformance are SD outlettemperature and lime-to-acid gasratio.
Residuals and Ash Handling
Ash is frequently accumulated on-site prior tohazardous waste land disposal. It may bedewatered or may be chemically(fixation/stabilization) prior to disposal.Residuals generated by the APCD containparticulates, absorbed acid gases which havereacted to become salts and small amounts oforganic contaminants. The suspendedcontaminants can be concentrated beforedisposal. The waters may either be returnedto die process or treated and discharged tosewers as necessary.
PERFORMANCE TESTING OFTHERMAL DESTRUCTION
FACILITIES IN THE EARLY 1980's
EPA conducted performance testing at severalthermal destruction facilities in the early1980's. Complete test reports have beenpublished for 8 incinerators, 11 industrialboilers, and 8 furnaces. These data as well astrial burn results from 14 RCRA applicants forincinerators have been summarized in an EPAreport, "Permit Writer's Guide to Test BumData - Hazardous Waste Incineration."7
Results of these studies (on the DREs, etc)have been summarized, and compared with thehazardous waste incinerator standards andproposed boiler and furnace regulations, byDempsey and Oberacker of the US EPA in
1988.5 Their comparisons are reproduced andshown in Figure 3 through Figure 16. It wasconcluded that properly designed thermaldestruction systems equipped with suitable airpollution control devices can meet or performbetter than the requirements set by theregulations.
RECENT STUDIES
Recent studies (1988) on hazardous wastecofired in industrial boilers under nonsteady oroffset conditions by EPA have demonstratedthat industrial boilers can provide adequatethermal environments for hazardous wastedestruction, achieving an average DRE of99.998% for RCRA toxic organics.89 Resultsfrom a pilot-scale boiler cofiring test (1988) toinvestigate nonsteady effects on DRE havealso revealed that unburned POHCs and PICscould be adsorbed on soot deposited on boilersurfaces during cofiring and desorbed backinto the combustion gases after waste cofiringceases, an effect which has been termedhysteresis.10 The impact of this hysteresis onthe DRE for POHCs was further tested in afull-scale hysteresis study on a watertubepackage boiler in 1990 by EPA to determineif the hysteresis effect actually exists, and ifso, to evaluate its effect on DREmeasurements. Results from POHC cofiringtests - Trichloroethylene (TCE) andMonochlorobenzene (MCB)- under sootingand nonsooting conditions indicated that DREswere generally low during the sootingoperations, only three to four "nines". Thehysteresis effect does exist: however, it doesnot significantly affect the DREperformance." During 1988 and early 1989,other halogenated hydrocarbons wastes wereincinerated.12 Over 300.000 gallons of EDB(ethylene dibromide) pesticide wereincinerated at commercial hazardous wasteincineration facilities. Sulfuric acid was firedinto the kiln to prevent the release of bromineto the atmosphere. EDB (C,H4Br2) is a liquidhalogenated hydrocarbon which wasregistrated as a pesticide in 1948 and iscarcinogenic, muiagenic, and has adverse
262
reproductive effects. The level of bromine inthe stack gas was non-detectable andpaniculate emissions were 0.0081 to 0.0123grains/dscf corrected to 7% CO:. Figure 17shows that the DREs of the EDB were at least6 nines. It is obvious that properly operatedincinerators are able to destroy a wide rangeof hazardous wastes.
CONCLUSIONS
Incineration is capable of the highest degree ofwaste destruction and is able to destroy thebroadest range of hazardous waste. It cangreatly reduce the solid waste volume, recoverheat energy from the combustion process, andpermanently destroy the waste. If theincinerators are properly designed andoperated, incineration remains the mostefficient and available technique for disposalof most organic wastes. More research isneeded to study on how to reduce the traceamounts of unburned hazardous wastes,products of incomplete combustion, haloacids.halogen gases, metais and particuiates in stackgases and residuals in the ash.
REFERENCES
E. T. Oppelt. "Incineration ofHazardous Waste: A CriticalReview," the Journal of the AirPollution Control Association(JAPCA). Vol.. No. 5. 1987.
U.S. EPA. "Dioxin Rule." FederalRegister. January- 14. 1985.
U.S. EPA. "Standards Applicable toOwners and Operators of HazardousWaste Treatment Facilities; InterimFinal Rule and Proposed Rule."Federal Register 47(122, PartV):27516-27535. June 24. 1982.
U.S. Congress. Congressional BudgetOff ice . "Hazardous WasteManagement: Recent Change and
9.
10.
Pol icy A 11 e r n a t i v e s. " U . S .Governmental Printing Office. 1985.
C. Dempsey and D. Oberacker."Overview of Inc ine ra t ionPerformance." Presented at theEngineering Foundation's Conferenceon "Hazardous Waste ManagementTechnologies." Mercersberg.Per sylvania. August 7-12. 1988.
U.S. EPA, "Polychlorinated Biphenyls(PCBs) Manufacturing, Processing.Destruction in Commerce, and UseProhibuion." Federal Register. Vol.52. No. 87. 1987.
U.S. EPA. "Permit Writers Guide toTest Burn Data - Hazardous WasteIncineration." EPA-625'6-36 012.1986.
M. Wool. C. Castaldini. and H. Lips."Engineering Assessment Report:Hazardous Waste Cofiring in IndustrialBoilers Under Nonsteady OperatingConditions." Acurex Summary ReportTR-86-I03 ESD. U.S. EPA'RREL.Cincinnati. Ohio. July 1989.
H. B Mason, et al. "Pilot-ScaleBoiler Cofiring Tests to InvestigateNonsteady Effects," Proceedings ofthe 14th Annual EPA ResearchSymposium on Land Disposal.Remedial Action, Incineration andTreatment of Hazardous Waste.Cincinnati. Ohio. May 9-11. 1988.EPA 600'9-88 '021. 7/88. pp.332-345.
H. B. Mason. J. A. Nicholson. M.Chan. R J. Derosier, and R. Gale."Pilot-Scale Testing of Boiler WasteCofiring-Hysteresis Effects." MidwestResearch Institute report. U.S. EPA.ORD. EPA Contract No 68-03-3241.August 1988.
263
11. G. D. Hinshaw, S. W. Klamm. G. L.Huffman and P. C. L. Lin, "Sorptionand Desorption of POHCs and PICs ina Full-Scak Boiler Under SootingConditions", Presented at the 16thAnnual research Symposium onHazardous Waste, Cincinnati, OhioApril 3-5. 1990.
12. D. Oberacker and C. Stangel,"Incinerating Ethylene Dibromide andDinoseb Stocks." Presented at theU.S. EPA's 15th Annual ResearchSymposium on Remedial Action.Treatment and Disposal of HazardousWaste, Cincinnati, Ohio, April 10.1989.
264
NON-INCINERABLE202.03 MMT
INCINERABLE47.25 MMT
Fig. 1. Annual Hazardous Waste
NON-THERMAL40.75 MMT
BOILERS & FURNACES3.8 MMT
INCINERATION2.7 MMT
Fig. 2. Incinerable Hazardous Waste
265
DRE (No of Nines, ie. 99.994 = 4.4)7 -
1 2 3 4 5 8 7 8
Facility Sites
Fig. 3. Avg. DREs from 8 EPA Tested Incinerators
HCL Removed (%)
Fig.
1 2 3 4 5 6 7 8
Facility Sites4. Avg. HCL Removal from 8 EPA Tested Incinerators
266
1000 ~T
8 0 0 - •
600 -
Particulate Emissions (Mg/Cubic Meter)
400 -
200 -
i 2 3 4 !i 0 7 8
Facility SitesFig. 5. Avg. Particulates from 8 EPA Tested Incinerators
1000
100
CO Emissions (ppm)
10 j=r
1 2 3 4 5 6 7
Facility SitesFig. 6. Avg. CO Emissions from 8 EPA Tested Incinerators
267
DRE (No of N ines , i e . 9 9 . 9 9 4 = 4 .4)ti -
b -
4 -
3 -
1 -
0 -
n^ ^ ^ ^ 99.99%
111•
ML!
1JnJ1J
1J1J1
\M
11J
_. •
11J1\M
•
J1JI1J1JJ11
1 2 3 4 5 8 7 8 9 10 11 12 13 14
Facilit}? S i tes
Fig. 7. Avg. DREs from 14 RCRA Applicants
100HCL Removed (%)
99.5 -
98.5 -
1 2 3 4 5 3 7 8 9 10 11 12 13 14
Facility SitesFig. 8. Avg. HCL Removal from 14 RCRA Applicants
268
500 -fParticulate Emissions (Mg/Cubic Meter)
400 -; •
30C -
200 -
100 -I
1 2 3 4 5 8 7 8 9 10 1] 12 13 14
Facility SitesFig. 9. Avg. Particulates from 14 RCRA Applicants
1000CO Emissions (ppm)
100 t
1 2 3 4 5 6 7 8 B 10 11 12 13 14
Facility SitesFig. 10. ,ivg. CO Emissions from 14 RCRA Applicants
269
3 -
5 -
DRE (No of Nines, ie. 99.994 = 4.4)
A 5 C D E F G H I J K
Facility SitesFig. 11. Avg. DREs from 11 Boilers
Particulate Emissions (Mg/Cubic Meter)
800 - f
600 -
400 -J
200 -
A B C D E F G H I J K
Facility SitesFig. 12. Maximum Particulates from 11 Boilers
270
CO Emissions (ppm)1Q000 »L_- -
1000 =
100 -
A B C D E F G H I J K
Facility SitesFig. 13. Maximum CO Emissions from ll Boilers
8 -
5 -
4 -
3 -
2 -
1 -
0 -*
DRE
A4
1
(No
fI1J
of Nines, ie. 99.
, 99.99%
m
•1Ii1J
994=^t - 4 )
1JJP
4•11•1 —
A B C D E F C H
Facility SitesFig. 14. Avg. DREs from 8 Industrial Furnaces
271
Particulate Emissions (Mg/Cubic Meter)300 f
250 -
200 -t
150 -
100 -(
50 -
A B C D E F G H
Facility SitesFig. 15. Avg. Particulates from 8 Industrial Furnaces
CO Emissions (ppm)1000 rrrr
100
A B C D E F G H
Facility SitesFig. 16. Maximum CO Emissions from 8 Industrial Furnac
CHEMICAL OXIDATION TREATMENT OF INDUSTRIAL ORGANIC WASTE
Penny M. Wikoff and Dan F. SuciuEnvironmental Research and Development, Inc.
Idaho Fails, Idaho
ABSTRACT
This paper presents chemical oxidationmethods including oxidation with hydrogenperoxide, oxidation with potassiumpermanganate, Wet Air Oxidation, andozonization. The benefits and limits arepresented. Process demonstration and costsare given where available.
INTRODUCTION
With the increased emphasis for use ofnontoxic, nonchlorinated solvents for cleaning,degreasing, or depainting operations and theincreased cost for disposal of the spentsolvents, an increased emphasis has beenplaced on recovery and/or treatment of thespent solvents. Recovery becomes moredifficult because many of the solvents are notsingle component systems and in some casesone or more of the components may beconsumed or degraded during use. Methodsavailable to treat the spent solvents includechemical oxidation, biological oxidation andadsorption processes such as granular activatedcarbon or ion exchange. The optimummethod is dependent on the process, processdischarge requirements, volume to be treated,the concentration of the spent solvent, andwhether the facility has an existing industrialwaste treatment plant (for example one with anactivated sludge system for treating theorganics).
The focus of this paper will be chemicaloxidation treatment processes. In chemicaloxidation processes, the toxic organiccompounds are oxidized to less toxic or non-toxic organic compounds or carbon dioxide.The chemical oxidation methods include
oxidation with hydrogen peroxide in thepresence of a catalyst, oxidation withpotassium permanganate, incineration, WetAir Oxidation, and ozonization. Incinerationwill not be discussed in this paper since it isbeing presented by Dr. Philip C. L. Lin inthis proceeding.
OXIDATION WITH HYDROGENPEROXIDE
Oxidation of organic waste with hydrogenperoxide (H2O2) is usually a slow process.However in the presence of a catalyst, theoxidation process proceeds rapidly.1 Catalystsinclude ferrous iron, copper, aluminum, andchromium. Ferrous iron is a good selectionsince it is not listed as a heavy or toxic metaland its use and subsequent precipitation doesnot create further hazardous waste problems.Hydrogen peroxide in combination with aferrous iron salt is commonly referred to asFenton Reagent. This reagent acts as anoxidizing agent through the formation ofhydroxyl free radicals:1
Fe+2 + H2O2 -» Fe+3 + OH + -OH (1)
The Fenton Reagent can act as an oxidizingagent for compounds such as alcohols,ketones, benzene, and phenols. The reactionwith phenol proceeds as shown in Reaction 2.1
There are no reports in the literature indicatingthe oxidation of phenol proceeds beyondmuconic acid. Optimum oxidation of phenolis achieved when 1 mole of phenol is treatedwith 1 mole of ferrous salt and 3 moles of
275
hydrogen peroxide. The reaction is unaffectedby pH in the range 3 to 7. However, in thepresence of acetate or phosphate buffers thereaction proceeds more slowly. With agreater degree of substitution, the reaction alsoproceeds more slowly, particularly when thesubstitution is in the ortho and para position.Halogenated phenols are oxidized rapidly withthe reaction rate decreasing according to thefollowing order: Cl > Br > I.'
The formation of intermediate products(semiquinones) produces a dark brown color.At the completion of the reaction, the solutionis a yellow or orange color due to thepresence of the ferric iron. The ferric ironcan be precipitated by increasing the pH eitherwith lime or caustic. Alum can be added toaid in floe formation. There is some co-precipitation of the remaining organics.Depending on the effluent dischargerequirements, further treatment may berequired to reduce the chemical oxygendemand (COD) to discharge limits. Theeffluent, however, may be applicable fordischarge to a domestic sewage treatment plantor biological industrial waste treatment plant.
The Fenton Reagent has been used to treatstripped refinery effluent and steel planteffluent. The stripped refinery effluent had aninitial COD due to phenol of 280 mg/L with atotal initial COD of 970 mg/L. (Note: Eachmg/L of phenol (C6H6O) is equivalent to 6mg/L of COD.) Nine moles of hydrogenperoxide and 1 mole of ferrous ammoniumsulfate was added per 1 mole of phenol, theCOD remaining due to phenol after reaction atpH 4 and 50°C for 30 minutes was 2 mg/L.The total COD remaining was 246 mg/L.1
CHEMICAL OXIDATION WITHPOTASSIUM PERMANGANATE
Certain naturally occurring organicrefractories or residual organics can be readilyoxidized by potassium permanganate.2
Functional groups are critical in determiningwhether or not a compound can be oxidized
by potassium permanganate. The carboxylgroup, carbonyl group, and hydroxyl group ofalcohols resist oxidation, while the carbonylgroup of aldehydes, the amino group ofamines and the carbon-carbon double bond ofunsaturated compounds are readily oxidized bypotassium permanganate. However oxidationdoes not proceed beyond the acid.2
Aromatic compounds require greaterconcentration of potassium permanganate.Approximately 9.3 moles of potassiumpermanganate were required to oxidize onemole of phenol to malaic acid. The reactionproceeds as follows:2
The wastewater requires further treatment toprecipitate the manganese. Depending ondischarge requirements, further organictreatment may be required to further reducethe COD.
WET AIR OXIDATION
Wet Air Oxidation can effectively treataqueous waste which is too dilute for costeffective incineration but too concentrated (tootoxic) for biological treatment.39 Wet AirOxidation is the aqueous phase oxidation oforganic and inorganic materials at elevatedtemperatures (34? to 608°F) and pressures(300 to 3000 psig). There is enhancedsolubility of oxygen in aqueous solutions atthese temperatures and pressures. The processrelies on the heat of oxidation to raise thetemperature to the required operating level.The incoming waste stream is typicallypreheated using a counter flow heat exchangerand excess heat from the treated stream. Theenergy required is the difference in enthalpyof the two streams, the energy required to heatfor stan up, the energy required to operate theair compressor. For the reaction to be self-sustaining (no auxiliary reaction fuel) theinitial COD must be a* least 15,000 mg/L.(The conditions required for autogenous
276
incineration are 1500 to 2000°F and 300,000to 400,000 mg/L COD.)3
The required oxygen is provided bycompressed air. The oxidation products areprimarily carbon dioxide and water. Sulfur isconverted to suifate and nitrogen to ammonia.These with the halogens stay in the aqueousphase. Metals remain in the aqueous phaseand can be precipitated prior to discharge.The heat from the treated waste stream can beused to generate steam.
Kalogenated aromatic compounds withoutother non-halogenated functional groups arerelatively resistant to Wet Air Oxidation.Electron donating constituents such ashydroxyl, amino, or methyl groups makearomatic rings more susceptible to destructionby Wet Air Oxidation.
Wet Air Oxidation has been demonstrated byMODAR on the pilot scale at CECO'sInternationa], Inc., Niagara Falls, New York,Hazardous Waste Treatment and DisposalFacility.56 The process was demonstratedusing a dilute isopropyl alcohol streamcontaminated with priority pollutants including1,1,2-tricnloroethane, nitrobenzene and 2-chlorophenol and a dielectric fluid containingpoly chlorinated biphenyis (PCBs). The pilotunit had an organic material flow capacity of50 gal/day. Greater than 99.99 percentdestruction of the organics was demonstrated.
Wet Air Oxidafion has been demonstrated atthe pilot level by several other companies,including Zirnpro in Rothschild, Wisconsin.7
Zimpro treated sewage sludge and industrialwaste sludges from paper and textile mills.Approximately 3400 kw-hr/day power wasrequired to treat 10,000 lbs of sludgecontaining 70 grams/liter of COD to 50percent reduction in COD.
Full-scale operations were not found in theliterature. However. MODAR is workingwith ABB Lummus Crest to scale-up theprocess and expects full-scale demonstration inthe near future. An estimate of the capital
equipment cost for a 10,000 gal/day wastestream containing sufficient organics to beself-sustaining was obtained from ABBLummus Crest. The estimated capital costwas $15 million.8
A Wet Air Oxidation demonstration unit(10.000 to 12,000 gal/day) has been inoperation for 4 years in Mississauga, Ontario,Canada. The unit has been treating wastecontaining up to 250.000 mg/L COD,reducing the COD by 89 percent. Oils andgreases are reduced by 99 percent,naphthalene by 99 percent, cresol by 94percent, polyethylene glycol by 98 percent.and trimethylbenzene by 99 percent.9
OZONIZATION OF INDUSTRIALORGANIC WASTE
If sufficient ozone is added, essentially allorganic compounds can be oxidized to carbondioxide.10"15 The source of the ozone caneither be by electrical production with anozone generator or through use of ultravioletlight to convert the oxygen available in thewaste stream to ozone. The use of an ozonegenerator, although requiring more initialcapital cost, is a more effective method oftreatment. Generation with ultraviolet light islimited by the solubility of oxygen in thewater.1011 The ozone generator processeseither air, oxygen enriched air or pure oxygeninto ozone. The ozonization process can beenhanced by the addition of hydrogen peroxideto the waste stream.
Complete and efficient ozonization isdependent on the type of waste, the degree ofoxidation desired, the reactor or contactorconfiguration, the contact time. Contactorconfigurations may include bubble towers orpacked columns. The presence of carbonatesmay inhibit ozonization at high pH.
Approximately 2 to 6 mg/L ozone are requiredper 1 mg/L ozone. Solutions with higherinitial phenol concentration react more rapidly.The ozone demand to oxidize ph.-nol at a pH
277
of 12 is half that required at a pH of 7.1: REFERENCES
The degree of ozonization desired is dependenton what process may follow the ozonizationprocess. For example, if a biologicaltreatment process is available, it may bedesirable to only treat the waste stream withozone to degrade the organics to the levelwhere they are not toxic to the biologicalsystem or can be degraded to discharge limitsby the biological system. Ozonization of high-molecular weight biorefractory, organiccompounds (humic substances) produceslower-molecular weight, more biodegradablesubstances.13 At the same time more oxygenis added to the waste stream for biologicaltreatment.
Treatment of ground water contaminated witht r i c h l o r o e t h y l e n e ( T C E ) andtetrachloroethylene (PCE) has beendemonstrated on a 2000 gpm unit.1415 Thegroundwater was contaminated with 200 /xg/LTCE and 20 fig/L PCE. In order to achieve80 to 90 percent destruction of thecontaminants. 4 mg/L ozone was requiredwith a ratio of hydrogen peroxide to ozone of0.5 by weight. The capital cost for the ozonegenerator was $85,000. Hydrogen peroxidecost was approximately $1.00/lb.15
CONCLUSIONS
Complete destruction of the toxic organics tocarbon dioxide by chemical oxidation is onlyachieved with Wet Air Oxidation andozonization. Wet Air Oxidation is efficientfor those concentrated wastes having at least15.000 mg/L COD. The Wet Air Oxidationprocess has not been demonstration full-scale,although MODAR expects demonstration inthe next year. Ozonization is effective andeconomical. especially with lowerconcentrations of organics. Table 1 lists acost comparison of the more commontreatment processes for organics.15
1. Eisenhauer, Hugh R., "Oxidation ofPhenolic Wastes," Journal of WaterPollution Control Federation.40.11,November 1968, pp. 1887-1899.
2. Spricher, R. G. and Skrinde, R. T.,"Effects of Potassium Permanganateon Pure Organic Compounds.' Journalof American Water Work Association.April 1965, pp. 472-484.
3. Teletzke, G. H. et al., "Componentsof Sludge and its Wet Air OxidationProducts," Journal of Water PollutionControl Federation. 39. 6, June 1967,pp. 994-1004.
4. Dietrich, M. J., Randall, T. L., andCanny, P. J., "Wet Air Oxidation ofHazardous Organics in Wastewater,"Environmental Progress. 4, 3, August1985, pp. 171-177.
5. Swallow, Kathleen C. et al., "TheMODAR Process for the Destructionof Haza dous Organic Wastes-FieldTest of a Pilot-Scale Unit," WasteManagement. 9, 1989, pp. 19-26.
6. Staszak, Carl N. and Malinowski,Kenneth C , "The Pilot-scaleDemonstration of the MODAROxidation Process for the Destructionof Hazardous Organic WasteMaterials," Environmental Progress.6. 1, February 1987, pp. 39-43.
7. Teletzke, G. H., "Wet Air Oxidation,"Chemical Engineering Progress. 60. 1.January 1964, pp. 33-38.
8. Evans. Brian, ABB Lummus Crest.H o u s t o n , T X , T e l e p h o n eConversation, February 15, 1991.
9. "Wet Air Oxidation Process Operatesat Low Pressures, Company Says,
278
"HazTech News. 5. 21. October 18.1990. pp. 159-160.
10. Sargent. J. W. And Sanks. R. L.,"Light-Energized Oxidation of OrganicWastes.", Journal of Water PollutionControl Federation. 46. 11. November1974. pp. 2547-2554.
11. Otake. Tsutao et al.. "Photo-Oxidationof Phenols with Ozone," Journal ofChemical Engineering of Japan. 12. 4,1979, pp. 289-295.
12. Niegowski, S. J.. "Destruction ofPhenols by Oxidation." Industrial andEngineering Chemistry. 45, 3. 1953.pp. 632-634.
13. Jones, Bonnie M.. Sakaji. Richard H..and Daughten. Christian G.. "Effectsof Ozonization and UltravioletIrradiaton on Biodegraddbility of OilShale Wastewater Organic Solutes."Water Resource. 19. 11. 1985.pp. 1241-1428.
14. Glaze, William H. and Kang, Joon-Wun. "Advanced Oxidation Processesfor T r e a t i n g G r o u n d w a t e rContaminated with TCE and PCELaboratory Studies." Journal ofAmerican Water Work Association.Research Technology. May 1988. pp.57-63.
15. Aieta. E. Marco et al.. "AdvancedOxidation Processes for TreatingGroundwater Contaminated with TCEand PCE. Pilot-Scale Evaluations."Journal of American Water WorkA s s o c i a t i o n . Research andTechnology. May 1988, pp. 64-72.
279
TABLE 1. TREATMENT COST COMPARISON15
Cost TvpeCapital Cost - $(Annual-
ized Over 10 Years)Operating and
Maintenance Costs - $GAC Replacement
Cost - $Total Annual ized
Cost - $Cost Per 1000 GAL $
AirAir Stripping
$ 48.400
$ 30.200
$78,600(0.075)
StrippingGas PhaseGAC
$108,000
$ 78.300
$105,100
$291,400(0.277)
PeroxideLiquid Phase Ozone
GAC
$192,500
$ 13.900
$210,000
$416,800(0.397)
AOP
$ 35.000
$ 63,900
$ 98.900(0.094)
280
MEDIATED ELECTROCHEMICAL OXIDATION OF ORGANICS
Leonard W. Gray, Robert G. Hickman. and Joseph C. FarmerLawrence Livermore National Laboratory
Livermore. California
INTRODUCTION
Replacement of halogenated with non-halogenated and toxic with non-toxic solventsis a worthy goal being sought within theDepartment of Energy (DOE) and Departmentof Defense (DOD) complexes. Today,however, the substitutions that would besuitable are not all identified. Consequently,there is continuing production of bothhazardous waste and mixed (hazardousradioactive) waste. Furthermore, existingback-log inventories of halogenated solvents atmany sites must be addressed, probablythrough some kind of destruction technology.Commercial technology exists for destructionof very dilute solutions of halocarbons inwater. We are seeking non-incinerationtechnologies suitable for destroyingconcentrated streams of various toxic organics.
OPERATING PRINCIPLES
Some metal ions, when dissolved in nitric orsome other mineral acid, can have theirvalence state raised at the anode of anelectrochemical cell. One such metal ion isAg+ which can be raised to Ag++ withoutsignificant oxidation of water at the anode.This is particularly true if the anode is madeof a material that exhibits a reasonably highoxygen bubble overpotential, such as Au orPt. Other examples of metal ions that may beraised to higher valence states and used aschemical mediators are Ce+ + * and Fe+ + + .Of these, silver is the most thoroughlystudied.(l)
The Ag++ formed at the anode almostimmediately forms a complex with a nitrate
ion, AgNO3+ which is very strongly absorbing
spectroscopically through a broad range of thevisible spectrum. Formation of the complexdoes not significantly reduce the oxidizingpower of the Ag+ + . The complex readilyattacks water to form OH free radicals andH+, at which time the silver is reduced to itsoriginal state, Ag+. A constant replenishmentof Ag++ ions is accomplished by continuouslycirculating the electrolyte past the anode.Hydroxyl free radicals are themselvespowerful oxidizers, particularly with respect tohydrocarbon molecules. In the case ofethylene glycol. the major component ofantifreeze. (CH2OH)2 reacts with bothAgNO3
+ and OH (from water) to form CO2.H+, NO3. and Ag+.
On the cathode side of the electrochemicalcell, NO, and water are produced by reductionof nitrate ion. If the NO2 is regenerated withoxygen in a separate reactor to produce nitrateion (which is recycled to the cell), then theoverall reaction is for ethylene glycol to beconverted to CO2 and H2O. Since the onlychemical added is the O2 used to regeneratethe nitrate ion, the stoichiometry is identical tocombustion, i.e. 2(CH:OH)2 + 5O : = 4CO:
+ 6ri :0.
The differences between the process justdescribed and combustion are obvious. One isdone in an aqueous electrolyte and requires anelectrochemical cell that operates near ambientconditions. Normal combustion operates nearambient pressure but at very high tempera-tures. Residence times in combustion arequite short. That gives some people concernabout stack gas emissions, whereas theresidence time in this aqueous system is quitelong, and complete destruction of the startingmaterial can be verified before anything isdischarged to the environment.
281
LAB SCALE EXPERIMENTS
The principles described above have beenchecked in many small experiments. Theyhave been used to develop several differentparts of the demonstration bench-scale systemthat was eventually built. For example, at theanode of the electrolytic cell, it is easy toevolve molecular oxygen if the anode potentialis raised too high. Scanning voltammetryusing a computer controlled potent'ostatallowed us to determine the maximum anodepotential that could be used to produce Ag+ +
but avoid oxygen evolution, a processinefficiency. Using rotating disc electrodemethods, diffusion coefficients for Ag+ in theelectrolyte were determined so the demon-stration cell could be optimized with regard toAg++ production. The reaction betweenAgNO3
+ and water to form OH and eventuallyO2 was followed spectrophotometrically. Thishelped determine the way that OH attackedhydrocarbon molecules. It also providedimportant information that was incorporatedinto the numerical model of the system.Kinetics of destruction of ethylene glycol weremeasured and a proposed mechanism wasdeveloped.(2) It should be applicable toketones, aldehydes, alcohols, organic acids,and the starting parent hydrocarbon such asoctane. This is shown in Figure I. Variousother organic compounds also were destroyed,such as tetraphenylborate and benzene.
MODELING
Numerical models were developed for thedevelopmental scale electrochemical cell. It isan annular design with a porous ceramicseparator between the gold anode and thestainless steel cathode. Because we hadmeasured or otherwise obtained the physicalproperties of the electrolyte and had measuredthe reaction kinetics of Ag++ with both waterand ethylene glycol, the mode! was able topredict how fast a given amount of glycolcould be destroyed as functions of celldimensions, electrolyte temperature andconcentrations, electrolyte flow rate, and
potential. A schematic diagram is shown inFigure 2. Parametric runs we e done on asmall computer to calculate both the optimumcell design as well as predict the system'sability to destroy antifreeze.
DEMONSTRATION
The demonstration hardware, which includednot only the electrochemical cell but otheritems as well, was installed in a glovedenclosure ventilated with untreated room air.The other hardware items included a demisterfor each side of the electrochemical cell. Thecathode side allowed NO, to be removedwithout entraining liquid droplets as the gasvented to the stack. Similarly, the anode sidehad a comparable unit to vent the CO2
produced and any O2 that may have beenproduced. Temperature was held at about70°C using electricai heaters immersed in theelectrolyte. Each side of the cell, of course,had its own circulation pump. We chose amagnetically coupled centrifugal design, withwetted surfaces made entirely ofpolyvinylidene fluoride. Piping and valueswere of the same material or else either teflonor titanium.
The results of the demonstration were verygratifying. We had determined from ti.cnumerical model that we could destroy 0.5Lethylene glycol in 22 hours. The destructionrate would be limited by the rate at which theelectrochemical eel! could produce Ag+ + .During the run, CO2, CO, and O2 weremonitored on-line in the off-gas from theanode loop. The first two used infrareddetectors and O2 used an electrochemicalmethod. Only CO2 and O2 were detected.This was later confirmed by grab samples ofgas analyzed by mass spectroscopy. We alsomonitored AgNO5
+ on-line in the anoSyte.This was done by absorption spectroscopyusing fiber optics to connect the probe to thespectrophotometer. At 20 hours, based onnegligible CO2 evolution, increased O:
evolution, and a high concentration ofAgNO3
+ being maintained in the anolyte. the
282
run was terminated. Using off-line totalorganic carbon analysis, it was determined thatwe had destroyed more than 99.9% of theethylene glycol in the anolyte. The numericalmodel had predicted the time for destructionwithin 10% of the actual time, which isexcellent agreement for this kind ofdevelopmental work.
SUMMARY AND FUTUREDIRECTIONS
The demonstration of the technology on thisscale gives us confidence in both our ability topredict the systems performance, and alsogives us some confidence in the destructionmechanism at work in the electrolyte. Otherclasses of organic compounds are under study,such as halogenated species. Also, destroying0.5L in 20 hours is a scale far too small to beof practical value. A new system is beingdesigned and constructed that has a capacityexpandable to about 120 times larger than ourinitial demonstration system.
2. J. Farmer, et.al., "Initial study of thecomplete mediated electrochemicaloxidation of ethylene glycol", LawrenceLivermore National Laboratory, Li verm-ore, CA. Report No. UCRL-106479(March, 1991).
283
Ceramicdiffusion barrier
HNO3 & AgNO3
anolyte
AnodeAnode
*• AgNO*
CathodeHNO3 + 2H* • 2e" *• HNO2 + H2O
CathodeHNO3 catholyte
Fiqure 1
—c — cI \
Acetaidehyde
I IHO—C—C—OH
Ethylene glycol
-OH—C —C
1 OHAcetic acid
AgNO:1
AgNOj AgNOj
!CO2(g) + —C
I
Carbon dioxide
—C —OHI
Methanol
•OH
— C —\
Formaldehyde
•OH
-*- —c\OH
Formic acid
AgNOjCO2 (g)
Carbon dioxide
Figure 2
284
TOWARDS A PROTOCOL TO DETERMINE WASTE MANAGEMENT PROPERTIESOF
SOLVENT SUBSTITUTES
Benerito S. Martinez jr., Meei-Huey Li and Ricardo B. JacquezDepartment of Civil, Agricultural, and Geoiogical Engineering
andWalter H. Zachritz II
Southwest Technology Development InstituteNew Mexico State University
Las Cruces, New Mexico
Martha I. BeachN-CON Systems
Larchmont, New York
INTRODUCTION
Halogenated solvents such as trichlorethylene(TCE), freon, and methylene chloride, areroutinely used in a wide variety of degreasingand cleaning operations (Higgens, 1989).Spent halogenated solvents released fromprocess washwater or in concentrate canadversely impact industrial wastewatertreatment plants and industrial pretreatmentprograms (Breton, 1988). Non-halogenatedsolvent substitutes with positive environmentalattributes are being strongly considered formany industrial operations. However, solventsubstitutes may suffer from contaminantcarryover unknown impact on the industrialwastewater treatment plant such as resistanceto biodegradation and physical/chemicalinteraction with waste components. Thenumber of commercial products marketed assubstitutes for halogenated solvents inindustrial processes is rapidly increasing.Very little is known about the impact of thesespecific solvent formulations on potentialgeneration of volatile organic compounds(VOC) emissions or impact on existingwastewater treatment facilities. While somemethods for assessing waste managementcharacteristics of substitute solvents have beendescribed, a standard procedure thatadequately characterizes the majorenvironmental factors is needed. A standardmethod for determining waste management
properties would provide broad guidelines formanufacturers and allow for more detaileddisposal information in MSD sheets. At aminimum the proposed standard testingprocedure should include volatilization andbiodegradability as basic test procedures. Thepurpose of this paper is to present apreliminary protocol for the rapiddetermination of volatilization andbiodegradability attributes of solventsubstitutes.
PROTOCOL FOR VOLATILIZATION
Several methods are available to determinevolatilization potential. The majority of thesemethods are based on ASTM standards forcoatings and paints and are summarized underEnvironmental Protection Agsncy Method 24.The volatilization of a compound can have farranging affects on the treatability andtreatment control technologies. Assessment ofvolatilization should encompass attributes ofboth the solvent tested and the wastewatersreceiving the solvent. Changes in pH,temperature variations or suspended solids canail impact volatilization rates of a particularcompound in a given wastewater. Theprotocol presented here is designed tocharacterize the basic volatilization propertiesof a terpene-based solvent and evaluate theinfluence of pH. Additional modifications to
285
this procedure can be used to systematicallyevaluate the presence of suspended anddissolved organic material or other wasteparameters on volatilization.
To determine basic volatilization rates, thesubstitute solvent was diluted to the expecteduse rate concentration with distilled water.For this test, a dilution of 1:2000. based onthe COD of 1,500.000 mg/L for the solventconcentrate. A 500 ml portion of this solutionwas placed in a 600 ml beaker and was stirredcontinuously for 10 hours with a 6.0 cmmagnetic bar at 300 rpm. Duplicates at theleast should be analyzed for each test. Atintervals of 2 hrs. a 6 ml sample was takenfrom each beaker for COD analysis. The rateof volatilization was determined by plottingCOD versus time for the test period anddetermining a best fit curve. The COD wasmeasured in accordance with StandardMethods procedure 508C (APHA, 1985). Theeffect of pH on volatility can be rapidlyevaluated by using the same procedure, butwith samples adjusted over the expected pHrange (pH 2 to 12 for this test) using either2N H,SO4 or NaOH. Baseline volatilizationdata were determined usii:<? a control sampleat the natural pH (9.2) of the solvent andwater mixture. Using a multiple stirring plate,up to four different pH values could beevaluated at any one time.
To test this protocol, a terpene-based solventwith the major active component listed asd-limonene (Penetone Corporation, T.nafly,NJ) was used. Additional componentsincluded: ethanolamine, glycol ether, EDTA.surfactants, and water. The pH of theconcentrated form was reported to be alkaline(10.4) and the Biochemical Oxygen Demand(BOD) and COD concentrations were reportedaj 2 5 0 , 0 0 0 - 4 0 0 . 0 0 0 mg/L and1.500.000-3,700.000 mg/L, respectively(Camacho. 1989). The flash point andfreezing point were reported to be 165CF(concentrate). 175°F (1:3 dilution), and ^2°F,respectively.
The results, shown in Figure 1. indicated that
the terpene-based, solvent mixtures decreasedover time as measured by COD. This lossappeared to follow a first-order, exponentialdecay and after 6 to 9 hrs the COD reductionwas about 46 percent. The volatilization ratedetermined for this data was 0.062 hr1. Thissolvent formulation was also tested over a pHrange 2-12. The loss of COD in each mixturefollowed a similar, consistent pattern, but pHdid not appear to affect the volatilization rateof the solvent. These results indicate that theterpene-based solvent formulation is stableover a wide pH range. Thus, the effect of pHon volatilization would be predicted to beminimal for many industrial applications.However, as the COD data indicate, the lossof unknown product components(s) wassignificant. Where capture of volatile organiccompounds (VOCs) is a requirement, the needfor appropriate control technology must beconsidered.
PROTOCOL FORBIODEGRADABILITY
Recent developments in respirometric analysishave provided a new technique that caneffectively determine biodegradability ofvolatile compounds. Jacquez et al., (1990)have provided data that describes the rapidevaluation of aqueous-based solventsubstitutes. Brown et al., (1990) indicate theappropriate test culture is needed formeasuring biodegradation. In developing theprotocol for rapid determination of biokineticparameters for substitute solvents, a series ofbatch and respirometric analysis wereperformed. Due to the volatile nature of thetest compound, the batch studies producedinconsistent and unreproducible data. On theother hand, the respirometric techniqueprovided reproducible and consistent datawhich provided for a more completecharacterization of biodegradability for thesolvent-substitute. A procedure using thistechnique will be described.
The biodegradability of the terpene-basedsolvent was evaluated by making respirometric
286
measurements with microbial seeds acclimatedto a variety of substrates and by evaluation ofMonod kinetic coefficients. Two acclimatedheterogeneous microbial seeds were preparedfor the biodegradability tests. The first seedwas obtained from the local wastewatertreatment plant and was acclimated to peptone.This seed was maintained at a sludge age of10 days, in a 10 L reactor mixed with air andit received a synthetic feed which containedpeptone as the sole carbon source. Theculture acciimated to the solvent wasdeveloped from the peptone culture andreceived a daily feed in which theterpene-based solvent was the sole carbonsource. The two cultures received the sameamount of carbon source (725 mg as COD)each day.
The biodegradation of the terpene-basedsolvent was then measured by using anautomated respirometric technique todetermine BOD and oxygen uptake rates(OUR). The instrument used in this test wasan eight station COMPUT-OX respirometer(N-CON Systems Company, Inc.. Larchmont,NY). For eacr« test series, the terpene-basedsolvent was diluted to a concentration range1000 to 0 ppm. A synthetic feed (Jacquez ;tal., 1990) was used to dilute the solution to500 mi. In order to minimize the ioss ofsolvent in the head space during mixing, 500ml of solution was placed in the 690 mlrespirometer bottle. Each reactor was seededto a concentration of 10 mg/L with the testculture and the contents were stirred with a6.0 cm magnetic stir bar.
The results of the biodegradation studies forthe culture acclimated to the solvent are shownin Figure 2. For concentrations exceeding400 ppm, the respiration curves show thatoxygen uptake was greater than the maximumoxygen transfer capabilities of the instrument.This characteristic has been previouslyreported for this instrument (Jacquez et al.,1989). The final results indicate that initiationof respiration occurred at 3, 9. and 18 hoursfor the cultures acclimated to the solvent,peptone, and domestic wastewater.
respectively. This increased lag period is areflection of the specificity of the culturepopulation developed through the acclimationprocess. In each case, biodegradation wasapparent because the oxygen uptake wasbasically proportional to the solventconcentration.
Evaluation of Monod kinetic coefficients usingrespirometric measurements is a recent butwell established procedure (Gaudy et al..1987; Dang et al., 1989). In order to evaluatethe Monod kinetic coefficients, threeparameters must be measured: initial biomassconcentration (XJ. growth yield (Y), and thebiomass COD equivalent (Ox. mg COO/mgbiomass). Y was estimated to be 0.4. from abatch growxh study. Ox was determined to be1.45 for the culture acclimated to the solvent.Based on the final results. umax was determinedto range between 0.045 ro 0.050 hr1 and K,ranged between 45 and 70 mg/L as COD.The magnitude for these parameters iscomparable to estimates determined by Dang,et al. (1989) using respirometric technique andchlorobenzene as the substrate. The meanvalues for umax, Kg, and Y were reported to be0.060. 3.52. and 0.42, respectively.Compared to the results obtained from thebatch growth study, measuring kineticparameters for a volatile substrate throughrespirometric technique is more reliable.
The respiration studies have clearlydemonstrated that the solvent is readilydegraded by a variety of bacterial cultures. Atthe concentrations tested, the terpene-basedsolvent may not be expected to pose a threatto either an industrial treatment plant or aPublicly Owned Treatment Works (POTW)which utilize biological processes in thetreatment scheme. But from the data, thetreatment plant might experience a short livedacclimation period which may extend 12-24hours. The potential impact carryover ofunknown hazardous materials ("heavy metals,toxic organics. etc.) resulting from the use ofthe solvent substitute has not br:en evaluated.Currently the protocol is being refined inorder to evaluate the biodegradability of these
287
attributes of substitute solvents.
CONCLUSIONS
The use of COD for evaluating volatilizationgives a direct measurement of the rate atwhich the solvent substitute is volatilizing.This method provides more useful informationthan that which can be obtained from EPAMethod 24. Other waste parameters that mayaffect volatilization can be evaluated by usingthe COD method.
The protocol outlined in this paper indicatedthat the respirometer can be a useful tool forrapidly evaluating biodegradability of thevolatile terpene-based solvent. Since therespirometer uses closed reactor vessels thetest was not biased due to changes inatmospheric pressure or the loss of volatilecomponents. Further refinements are neededto assure consistent test results. For example,test biomass either acclimated or unacclimated,must be grown in a controlled method toassure consistent determinations of lag timeand kinetic parameters. Useful parameterssuch as growth rates, yield coefficients, orinhibition coefficients must be agreed uponand ranges of typical values determined.Additionally, methodologies addressingcontaminant carryover and mobilizationresulting from used solvents must bedetermined.
REFERENCES
1. American Public Health Association(1985) Standard Methods for theExamination of Water and Wastewater.16th edition., Washington. DC.
7.
Breton, M., Frillici, P., Palmer, S.,Spears, C , Arienti, M., Kravett, M.,Shayer, A., and Suprenant, N. (1988)Treatment Technologies for SolventContaining Wastes. Noyes DataCorporation, Park New Jersey.
Brown, S.C., Grady, C.P. Leslie,Tabak, H.H. (1990) "BiodegradationKinetics of Substituted Phenolics:Demonstration of a Protocol Based onElectrolytic Respirometry." Wat. Res.24, 853-861.
Comacho, R. (1989) PersonalCommunication, Industrial WasteSpecialists, City of Austin, Texas,August.
Dang, J. S., Harvey D. M., JobbagyA. and Grady C.P.L. Jr. (1989)"Evaluation of Biodegradation Kineticswith Respirometric Data." Res. J.Wat. Pollut. Control Fed.61,1711-1721.
Gaudy, A.F., Jr., Rozich, A.F.,Garniewske, S., Moran, N.R., andDkambaram. (1987) "Methodology forUtilizing Respirometric Data to AssessBiodegradation Kinetic." Proceedings.42nd Annual Industrial WasteConference. Lewis Publishers,573-584.
Jacquez. R.B., Cadena, F., Prabhakar,S.. and Beach, M.I. (1989) "GasTransfer limitations in EnvironmentalRespirometry." Proceedings. 44thAnnual Industrial Waste Conference,Lewis Publishers, 425-433.
288
9. Jacquez, R.B.. Zachritz II, WH., Li.Meei-Huey, Beach, M.I. (1990)"Minimization of Hazardous WasteGeneration: Preliminary Investigationof a Solvent Substitute for TCE."Proceedings of the 1991 NationalConference on Environmenta]Engineering. Reno, NV.
10. U.S. EPA, CFR 40. Parts 53 to 60,Method 24. 979-981, July 1. 1990.
289
700Mixing Speed: 300 rpm
f
400
30010
TUne.hr
Figure 1. Volatilization of the Terpene-Bised Solveat Mixture
160
140
120
100
80
60
40
20
400 ppm
10 15
Tlme.hr
Figure 2. BOD for in Acclimated Cultureusing the Respirometer
20 25
290
Section VI
ISSUES TO CONSIDER
ALTERNATIVES TO CHLORINATED SOLVENTS:HEALTH AND ENVIRONMENTAL TRADEOFFS
Katy WolfInstitute for Research and Technical Assistance
Los Angeles, California
INTRODUCTION
Governmental agencies have not traditionallyused a systems approach to developing policiesor regulations governing the use of chlorinatedsolvents and ozone depleting substances. Asa consequence, policies have pushed usersfrom one set of chemicals dangerous in aparticular way to another set, dangerous in adifferent way. Policies have also transferredthe problem from one medium to a lessregulated medium. This paper discusses threehistorical case studies and focuses on twocurrent case studies that demonstrate the lackof an integrated approach. The shift awayfrom chlorinated solvents and ozone depletingsubstances is well underway and, becausethere is no coordinated policy, the alternativesthat are adopted may eventually pose seriousproblems to human health and theenvironment.
HISTORICAL CASE STUDIES
There are numerous case studies thatdemonstrate the lack of integration ofenvironmental policies. Regulations generallyfocus on reducing or eliminating the use of aparticular chemical or set of chemicals. Usersrespond by adopting other substances that havenot yet been targeted by regulation. Theymay be as dangerous as the chemicaloriginally used, but in different ways.Frequently, different government agencieshave different agendas. Even different officeswithin a particular government agency have
different aims. This can result in inconsistentand conflicting regulations that are confusingand that may not better protect human health
and the environment. The three historicalcase studies discussed here illustrate thesepoints.
Ban on DBCP
In the 1970s, a chemical calleddibromochloropropane or DBCP was used asa fumigant for certain crops to preventdestruction by nematodes, small worm-likesoil insects. The DBCP was being producedin an Occidental Chemical plant in Lathrop,California. Several workers in the plant,through conversations, eventually realized thata large fraction of them where unable to havechildren. Over time, it became apparent thatit was the DBCP that had caused sterility inthe workers at the plant. DBCP wassubsequently banned and the chemical thatreplaced it in -nany pesticide applications wasethylene dibromide or EDB, a suspectcarcinogen. It, too, was eventually banned forpesticide uses. This is a situation where onechemical was banned because of a particularproblem and it was replaced by anotherchemical that simply posed a health problemof a different kind.
VOCs and Exempt Solvents
The South Coast Air Quality ManagementDistrict (AQMD) in Southern California is oneof the most stringent air districts in the nation.Since the early 1970s, the District hasdeveloped a series of stringent rules designedto reduce or eliminate the use ofphotochemically reactive volatile organiccompounds (VOCs) that contribute to smog.The rules exempt certain substances becausethey do not contribute to photochemical smog.These exempt chemicals include thechlorofluorocarbons (CFCs), 1,1,1-
291
trichloroethane (TCA) and methylene chloride(METH). Many users in the Los Angelesarea have moved out of the VOCs into theCFCs, TCA and METH in response to theregulations. Indeed, the proportional use ofthese substances is much higher in the regionthan in the rest of the nation. The CFCs andTCA contribute to stratospheric ozonedepletion and are scheduled to be banned overthe next decade or so. METH is a suspectcarcinogen and is a toxic air contaminant.Again, this is a situation where users movedfrom VOCs, which cause smog, to otherchemicals that had different problems.
Aerosol Propellants
In the 1970s, three of the CFCs-CFC-11,CFC-12 and CFC-114--were widely used aspropellants in aerosol packaging. In 1978, theU.S. unilaterally banned the use of CFCs aspropellants in nonessential aerosol applicationsbecause of their contribution to stratosphericozone depletion. Hydrocarbons likeisobutane, which are flammable andphotochemically reactive, replaced the CFCsas propellants in most instances. Manyaerosol packers added METH to thehydrocarbons; it was an excellent co-solventfor the propellants and active ingredients.Several years later, the Consumer ProductSafety Commission asked industry to labeltheir products that contained METH with thewarning that the chemical was an animalcarcinogen. Rather than add this label, mostaerosol formulators converted from METH toTCA. TCA will be banned because itcontribute to ozone depletion. In California,the air districts are considering a ban onphotochemically reactive propellants. This isa situation where there was a conversion fromozone depleters to smog producing, flammablechemicals; then a partial conversion to asuspect carcinogen; finally a partial conversionto another, less strong ozone depleter.
CURRENT CASE STUDIES
There are two very interesting case studies
that are evolving today which will stronglyaffect nationwide chemical use patterns. Thefirst, dealt with here in detail, involves thesubstitutes for ozone depleting substances.The second focuses on METH and itswidespread use in paint stripping. Regulationson use of the latter chemical will becomeincreasingly stringent over the next decade andalternatives that are dangerous in other wayswill be adopted. These case studies, like thehistorical cases described above, demonstratethat government policies are developed withoutan integrated assessment of the consequences.
Ozone Depleting Substances
As mentioned earlier, TCA and CFC-113 willbe banned over the next decade or so becausethey contribute to ozone depletion. In thelight of this ban, many substitutes are beingmarketed in electronics, precision and generalmetal cleaning applications. These substitutes,fall into seven categories which are listed inTable 1, together with some of theircharacteristics.
The low molecular weight hydrocarbonsinclude flammable solvents like acetone,isopropyl alcohol and methyl ethyl ketone(MEK). These substances were used widelybefore chlorinated and CFC solvents wereintroduced into the market. The lattersolvents replaced them in many applications;they were perceived to be safer to workersbecause they were not flammable. The lowmolecular weight hydrocarbons arephotochemically reactive and some air districtsmay not grant permits for their use. Theirapplicability as alternatives is limited becausemany users do not want to deal withflammable materials. EPA has i:sued a test
rule requiring toxicity testing of isopropylalcohol; many of the other flammable solventshave not been tested for chronic toxicity.
The high molecular weight solvents have flashpoints in the combustible range. They includechemicals like dibasic esters (DBE), terpenes,alkyl acetates and N-methyl pyrollidone
292
(NMP). These solvents are not volatile andthey do not evaporate readily. They will leavea residue if they are simply oven dried. Awater rinse is required to flush themcompletely from parts. Thus they can be usedin applications where a residue does not matterand where a water rinse is feasible. The highmolecular weight hydrocarbons are not exemptfrom photochemical smog regulations so apermit is required for their use. Because theyare not volatile, however, emissions are notlikely to be extremely high. D-limonene, amajor ingredient of terpene formulations, hasgiven a positive carcinogenicity test in malerats. EPA has issued a proposed test rule onNMP and few of the other combustiblesolvents have been scrutinized for their healthand environmental effects.
The other chlorinated solvents includetrichloroethylene (TCE), perchloroethylene(PERC) and METH. All three are suspectcarcinogens and have been labeled toxic aircontaminants under the new Clean Air Actamendments. TCE and PERC are regulatedas smog contributors. These solvents havebeen used for many years and they performextremely well. They are likely to be moreheavily regulated in the future. WilliamReilly, the Administrator of EPA, hasdeveloped a list of chemicals; he is askingindustry to voluntarily reduce the use of thesechemicals by half over the next several years.The three chlorinated solvents are on the list.
The HCFCs are CFCs containing hydrogen;this makes them less stable in the atmosphereand they contribute less significantly to ozonedepletion than do the CFCs. Those suitablefor solvent use include HCFC-123, HCFC-141b and HCFC-225. The former two haveextremely low boiling points; their losses willbe very high and they will be extremelyexpensive to use. HCFC-225 is composed oftwo isomers, one of which is toxic. It is notclear whether a manufacturing process thatselectively produces the nontoxic isomer canbe developed. All three HCFCs are currentlyin toxicity testing and they will be fullyscrutinized by the time they enter the market.
Because they deplete the ozone layer to someextent, however, they are considered onlyinterim alternatives and they will eventually bebanned.
The HFCs and FCs contain no chlorine sothey do not contribute to ozone depletion.They do, however, contribute to globalwarming, particularly the FCs which havelong atmospheric lifetimes. One HFC,pentafluoropropanol, is being offered forsolvent applications. It has a very low workerexposure level and is not being tested forchronic toxicity. The FCs alone are not verygood cleaners and other substances must beadded to them to increase their performance.They are not currently in toxicity testing andthey will be very expensive to use.
Aqueous based formulations are anotheralternative to TCA and CFC-113. These willprove technically suitable for manyapplications. They have four drawbacks,however. First, water based cleaners do notdry readily and more energy will have to beused to achieve dry parts; this will exacerbateglobal warming. Second, these formulationsvirtually always contain additives and theadditives are sometimes known to bedangerous or they have not been scrutinizedfor their health and environmental effects.Many of the additives are photochemicallyreactive. Third, water based cleaning willresult in much more sewer loading of organicsand metals and it is not clear what theconsequences will be. Fourth, these cleaningprocesses require increased use of a scarceresource and this can be a significant problemin areas like California.
Various other processes are being proposed asalternatives to TCA and CFC-113 in variousapplications. Two of these-supercriticalcarbon dioxide and carbon dioxide snow-release carbon dioxide to the atmosphere. Itis not clear whether these releases willeventually be regulated to prevent globalwarming.
In the years to come, as TCA and CFC-113
293
availability become increasingly restricted, thealternatives will be implemented widely. Theygenerally pose known problems of a differentkind or they are unscrutinized. Governmentagencies and offices are not takingresponsibility for evaluating the alternativesand there is no evidence that they will be usedproperly or that their disadvantages will evenbe known before they are used. For example,the office of EPA responsible for banningozone depleting substances encourages users toadopt photochemically reactive substances assubstitutes. Since their agenda is todiscourage the use of TCA and CFC-113, theyare responsible for pushing users intosubstances that are unscrutinized. The AQMDin Southern California, on the other hand,discourages the use VOCs and has, for manyyears, encouraged the use of TCA and CFC-113. The Office of Toxic Substances at EPAhas been charged with evaluating the safetyand environmental characteristics of thealternatives. They have performed an initialscreen of only some of the alternatives andthey have received pressure from the Airoffice to minimize the problems with thealternatives they are examining.
The EPA has set up a special working groupthat was to facilitate changing Mil-Std-2000,the military standard governing the use ofprinted circuit board cleaning solvents. Thegroup has tested and qualified severalalternatives. In spite of the high priorityassigned to changing the military specification,however, the Department of Defense (DoD)has not yet taken action.
In this case study, it is clear that differentgovernmental agencies have different aims thatmay not result in better protection of humanhealth and the environment. The result is aweb of regulations which are inconsistent andconflicting and which encourage the useunscrutinized alternatives or alternatives thatare known to be dangerous.
Paint Stripping
METH is an excellent paint stripper and it has
been used widely for that purpose for manyyears. The Occupational Safety and HealthAdministration will lower the workplaceexposure level from 500 to 25 ppm in 1991.METH is considered a suspect carcinogen andhas been designated a toxic air contaminantunder the Clean Air Act amendments. It willbe increasingly regulated in the years to comeand users will be forced to adopt alternatives.
In maintenance stripping-aircraft stripping,for instance—abrasive and fracturetechnologies are being investigated. Some ofthese are promising but they present otherproblems. The dry abrasive techniquesrequire dust control which can be expensiveand there remains a question of whether or notthese techniques do long term substratedamage. The wet abrasive method may leadto corrosion and probably will not be adoptedfor whole airframes. The fracture technologiesmay also cause substrate damage.
In the consumer sector-household andcontract stripping-several chemical strippingalternatives are being examined. They includethe flammable and combustible solventsdescribed in the last case study. EPA's Officeof Toxic Substances performed an initialanalysis of the alternatives and concluded thatall but one alternative posed as severe or moresevere health and environmental problems asMETH. Many of them remain relativelyunscrutinized.
CONCLUSIONS
The historical case studies discussed heredemonstrate that there has not traditionallybeen a systems approach to regulation.Regulations have pushed users from one set ofchemicals dangerous in a particular way toanother set dangerous in another way. Thecurrent case studies show that this has notchanged and that society will adoptunscrutinized chemicals over the next decadeand will transfer the problem from the air tothe sewer.
294
A continuing problem is that governmentalagencies and offices each have a particularagenda and there is no incentive to use anintegrated approach to the problems. Theregulators also have only limited knowledge ofthe field that corresponds to their specialty.They know very little of other areas. Anotherproblem is that regulators rarely understandthe processes where chemicals are used. Theygrant and deny permits without knowinganything about the process itself. If thissituation is allowed to continue, currentpolicies will continue to push users into new,unscrutinized chemicals and processes and totransfer the problem to the least regulatedmedium.
To improve regulatory policy, there areseveral steps that could be taken. First thereshould be an admission that all chemicals andprocesses present problems of one kind oranother and that, in fact, there is no panacea.Second, there should be an acknowledgementthat there are no easy generic solutions. Thealternatives will have to be chosen on a case-by-case basis with the characteristics andlocation of the particular operation in mind.Third, governmental policies must begin totake a systems approach to the problem. Newchemicals should be tested for their health andenvironmental effects before they are marketedand a clear analysis of the tradeoffs should beundertaken before chemicals are heavilyregulated and substitutes adopted.
295
TABLE 1
CHARACTERISTICS OF CHLORINATED SOLVENT ALTERNATIVES
Chemical Class Examples Characteristics
Low MWHydrocarbons
High MWHydrocarbons
Acetone: isopro-panol; petroleumsolvents; MEK
Terpenes; DBE;NMP; alkylacetates
ChlorinatedSolvents
HCFCs
TCE; PERC; METH
HCFC-123; HCFC-225; HCFC-141b
HFCs and FCs pentafluoropro-panol
Flammable; photo-chemically reac-tive; evaporatereadily; many notscrutinized fortoxicity
Combustible; photo-chemically reac-tive; leave residueor require waterrinse; do not dryreadily; many notscrutinized fortoxicity
No flash point;evaporate readily;health effectss c r u t i n i z e d ;heavily regulated
No flash point;evaporate readily;low boiling point;health effects willbe scrutinized;will be banned
Health effectsunscrutinized;contribute toglobal warming
Water Higher energy use;increased sewerloading; additivesunscrutinized; useof scarce resource
296
FORMATION OF SPECIFICATIONS FOR NEW PRODUCTS
Captain Daniel T. WittTechnology and Industrial Support Directorate
San Antonio Air Logistics CenterKelly Air Force Base, Texas
BACKGROUND
The Environmental Program Office acts as aconsulting organization to the Aircraft, Engineand Support Equipment Directorates whichspecify the materials and processes used onSan Antonio Air Logistics Center (SA-ALC)managed equipment. This includes 1/2 of theAF Engines, 60% of AF Support Equipment,All the AF Electronic Test Equipment. C-5,T-38, T-37, and OV-10 Aircraft. Inconjunction with the equipment, SA-ALCcurrently manages over 45,000 TechnicalOrders. The Materials Engineering Section,which includes the TI Environmental Program,SA-ALC Corrosion Program, NondestructiveInspection and the High Technology MetalsInsertion Program Offices, recommends thematerials and processes specified in all SA-ALC Technical Orders.
TECHNICAL ORDER CHANGES
The cost to change one page in a TechnicalOrder (TO) is between 500 and 1000 dollarsregardless if the change involves one word orthe entire contents of the page. The referenceof a specification number in a TO eliminatesthe arbitrary changing of manufacturers dataand thus eliminates changing TO pages. Tofurther explain, if a Commercial Item isinserted into a TO, enough information mustbe provided in the TO so the user canpurchase the item. This includesmanufacturer's name, address, phone number,and product designation. This results inseveral problems:
1. If the manufacturer changes any of thedata, it costs the Air Force between $500 and$1000 to change the TO.
2. Sole sourcing is frowned upon, therefore,at least two products must meet the samerequirements before they can be inserted in theTO.
3. When users need to purchase a non-specified product, they go out on localpurchase. The installation procurement officegoes out on bid based on the description of theproduct. He gets a LOW BID product thatmay not meet the requirements of the productlisted in the TO.
4. The quality of the product can not becontrolled. The manufacturer can change theformulation of the product for any number ofreasons and not inform the organizationspecifying it.
By developing a specification and referencingit in the TO, the manufacturer's informationcan change numerous times and never affect asingle TO. The manufacturer's information ischanged on the Qualified Products List (QPL)the specification. Thus the cost formaintaining TOs is reduced and quality controlof the product affirmed. This insures thatfield maintenance personnel receive a productthat works.
FORMULATING A SPECIFICATION
When formulating a specification, the existingrequirements for current cleaners must beincorporated along with special requirementsdealing with environmental, safety, health andnew technologies. Some of the existing
297
requirements that have proven their worth are: CURRENT PROJECTS
Hydrogen Embrittlement of High StrengthSteels
Total Immersion Corrosion Tests ofAerospace Metals
Low Embrittling of Cadmium Plated SteelEffects on Unpainted MetalsSandwich Corrosion Tests of Aerospace
MetalsStress Crazing of Stretched Acrylics/PlasticsEffects on Painted SurfacesLong Term Storage StabilityEffects on Aerospace Sealants and RubbersHard Water StabilityHeat and Cold StabilitypH Value
As environmental and heaith concernsheighten, some special requirements arereviewed, keeping in mind new legislation andpublic concerns. These special requirementsare:
* BIODEGRADABILITY* TOXICITY* DISPOSAL CHARACTERISTICS* FLASH POINT* VOLATILITY* NEW TECHNOLOGIES
The largest stumbling block encountered is thelack of standard definitions for each of thespecial requirements. It becomes moredifficult when writing a performancespecification (versus a material specification)which does not limit the materials used in themanufacturing of the products. Added to thefact that specifications are used worldwide andregulations governing hazardous wastedisposal and hazardous materials change fromcountry to country, state to state and district todistrict, the task of defining common groundbecomes elusive to say the least. Establishedguidelines concerning the specialrequirements must be standardized before aspecification can satisfy the needs of everyinstallation. But before that happens sometough questions need to be answered abouteach of the special requirements.
The TI Environmental Program Office, inconjunction with the Energy ManagementDirectorate, (which manages the Air Forcecleaner specifications) SA-ALC/SF, iscompleting research and development onspecifications for cleaners passing the phase IItesting conducted at Tinker AFB, hoping toquickly implement these cleaners into the AF.The following specifications have been revisedor initiated for the purpose of hazardousmaterial substitutions.
SPECIFICATION: P-D-680A Amendment3, (added Type III)ISSUE DATE: 13 July 90NATIONAL STOCK NUMBER:
5 gal 6850-01-331-334955 gal 6850-01-331-3350
TITLE: DRY CLEANING ANDDEGREASING SOLVENT
This amendment was initiated by SA-ALC/TIESM due to the numerous requestsfrom field organizations to add a type ofmaterial with an increased flash point to over200 degrees. The amendment added a TypeIII petroleum distillate solvent. Theamendment was issued by SA-ALC/SF. Stillin development, it is an effort to addrequirements for recycled solvent into thespecification.
Advantages'.
•Min. 200 Degrees F flash point
•Low to no odor
•Non-chlorinated
•200 PPM Threshold Limit Value (TLV)low evaporation rate (3-4 times slower thanthe Type II material)
•Recyclable
298
Disadvantages:
•Higher cost ($8.00 per gallonvs $2.00 for Type II)
•Slower evaporation rate
SUBSTITUTION POTENTIAL: Expected toreplace P-D-680 Types I and II in numerousapplications. References of P-D-680 inTechnical Orders at SA-ALC estimated over300,000.
SPECIFICATION: MIL-C-87937ISSUE DATE: 29 Oct 90NATIONAL STOCK NUMBER: stillpending
TITLE: CLEANING COMPOUND,AEROSPACE EQUIPMENT
This specification was initiated by SA-ALC/TIESM to replace solvent cleaning ofaircraft, aircraft engines and supportequipment with a biodegradable, waterdilutable, less hazardous cleaner. Thespecification has two types of cleaningcompounds. Type I is Terpene based (citrusbased) cleaners and Type II is for generalcleaners with little or no solvents.
Advantages:
•Biodegradable
•Less toxic
•Diverse uses: spot cleaning, exterior A/Cwashing, dip tank cleaning, carbon removaland general cleaning
•Type I recyclable
Disadvantages:
•Type I, low flash point (minimum 125Degrees F, dilution with water raises theflash point)
•Requires water rinse except as a spot cleaner
•Higher cost ($9 a gl vs approx. $2 forcurrent
•Cleaners but dilution is 20:1 for exteriorcleaning which reduces cost to $.45 a gallon)
SUBSTITUTION POTENTIAL: Replacehazardous solvent based cleaners in exterioraircraft and engine cleaning, parts cleaning invapor degreasing, cleaning Aerospace GroundEquipment, hazardous carbon removers andgeneral cleaning where water rinsing isacceptable.
SPECIFICATION: MIL-C-83873ISSUE DATE: est. 1 Mar 91NATIONAL STOCK NUMBER: Pending
This specification was initiated by SA-ALC/TIESM to replace the hazardous solventsused prior to coating aircraft and supportequipment. This specification takes advantageof the unique properties ot enzyme basedcleaners to remove contaminants.
Advantages:
•Biodegradable
•Nontoxic
•Nonhazardous
•Easily disposed of
Disadvantages:
•Cost ($65.00 a gallon, dilution ratio is 15parts water to one part cleaner, $4.00 agallon diluted) No handling or disposal costs.
SUBSTITUTION POTENTIAL: Expected toreplace methyl ethyl ketone for exterior wipedown prior to aircraft and AGE painting.
299
Replacement for scuff sand and solvent wipeprior to recoating or touch up of aircraft andAGE.
SPECIFICATION: MJL-C-XXXXXISSUE DATE: 1 Mar 91NATIONAL STOCK NUMBER: Pending
TITLE: ACETATE ESTER BASEDSOLVENTS
This specification was initiated by SA-ALC/TIESM to replace low flash point, highlytoxic solvents used to clean aircraft partswhere rinsing of the parts is not feasible andfor electronics cleaning. There are six typesavailable with varying viscosities, flash pointsand evaporation rates.
Advantages:
•Not a Ozone Layer Depleting Substance
•Flash points range from 134 to 261degrees F
•Low toxicity
•Low residue (27 ppm and up)
Disadvantages:
•Alcohol based
SUBSTITUTION POTENTIAL: Furthertesting is required to determine the fullpotential of the cleaning products. Expectedto replace CFCs in cleaning electronicsequipment and general applications on aircraftsystems. Provides replacement potential forcleaning solvents used in painting equipmentclean up, coating thinness, spot cleaning andsolvent wiping.
300
IParticipants
The First Annual InternationalWorkshop on Solvent Substitution
December 4-7, 1990Phoenix, Arizona
David Albright
Gopal Annaznraju
Bob Anthony
Iris Artaki
Kelly K. Asada
R. W. Aubert
Environmental Protection SpecialistU.S. Environmental Protection AgencyOTS/GCD401 M Street SWWashington, DC 2 04 60202-245-4028
Program ManagerWright-Patterson Air Force BaseHQ-AFLC/DEVRDayton, OH 45433513-257-7053
Maintenance SupervisorCoors Brewing CompanyBC035Golden, CO 80401303-277-2092
AT&T Bell LaboratoriesP. O. Box 900Princeton, NJ 08540609-639-2585
Member of the Technical StaffHughes Aircraft CompanyRadar Systems GroupP. 0. Box 92426Los Angeles, CA 90009213-334-5235
Mgr., Environmental Resources Mgmt.General Dynamics, Air Defense SystemsP. 0. Box 2507Pomona, CA 917 69714-868-3472
301
APPENDIX I
Daniel Axelrad
R. Page Ayres
E. G. Baker
Gary E. Baker
James J. Baremore
Donald Bauer
Fred Bauer
John Beller
Economist, Office of Toxic SubstancesU.S. Environmental Protection Agency401 M Street, SW (TS-779)Washington, DC 20460202-382-3713
Environmental EngineerNewport News Shipbuilding4101 Washington AvenueNewport News, VA 2 3607
Battelle Pacific Northwest LaboratoriesMail Stop K2-12Richland, WA 99352509-375-2026
Senior EngineerSAIC635 West 7th, Suite 403Cincinnati, OH 45203513-723-2611
Mgr., Manufacturing ProgramsSandia National LaboratoriesDept. 6610P. O. Box 5800Albuquerque, NM 87185505-844-3690
Senior Manufacturing EngineerMartin Marietta Energy SystemsMail Point 150P O. Box 58 3 7Orlando, FL 32855407-356-5354
U.S. Department of EnergyTechnology Development Division785 DOE PlaceIdaho Falls, ID 83402208-526-0142
Senior Project EngineerEG&G Idaho, Inc.Technology IntegrationP. 0. Box 1625Idaho Falls, ID 83415208-526-1205
302
Mark G. Benkovich
Jim Bennett
Richard Benson
Lloyd Berg
James Berger
Kieran D. Bergin
Elliott Berkihiser
James Blackburn
Jack Blakeslee
Senior EngineerAllied-Signal Aerospace Co.P.O. Box 419159Kansas City, MO 64141816-997-5020
Business ManagerUnited Technologies USBI6000 Technology DriveHuntsville, AL 35806205-721-5509
Program ManagerLos Alamos National LaboratoryMail Stop F643Los Alamos, NM 87545505-667-4960
Montana State UniversityChemical Engineering DepartmentBozeman, MT 59715406-944-2222
Westinghouse HanfordP. 0. Box 1970Richland, WA 99352509-376-9942
Manager, Environmental ControlRockwell International CorporationP. 0. Box 2515Seal Beach, CA 90740213-797-2412
Manager, Chemical ReductionThe Boeing CompanyP. O. Box 2707Seattle, WA 98124206-393-4784
Associate Director, EERCUniversity of Tennessee423 S. StadiumKnoxville, TN 37996615-974-1779
Technical Program ManagerEG&G Rocky Flats, Inc.P. 0. Box 464Golden, CO 80005303-966-4642
303
Brian Blakkolb
Jim Block
Elliot W. Bloom
Phoebe Boelter
J. K. Bonner
Robert M. Bottom
Tom Boyce
Karla M. Boyle
Raymond J. Bradley
Materials EngineerTRWOne Space ParkRedondo Beach, CA 90278'13-813-8960
Creare Inc.Etna RoadP. O. Box 71Hanover, NH 037 55
Project ManagerUnited TechnologiesChemical Systems DivisionP. O. Box 49028San Jose, CA 95161408-365-5535
Weapons Complex Monitor Forums1715 North Wells, #34Chicago, IL 60614312-988-7667
Manager, Applications/Proc. Devel.Allied Signal Inc.2001 N. Janice AvenueMelrose Park, IL 60160708-450-3880
Plant Layout EngineerGeneral Motors CorporationAllison Gas Turbine DivisionP. O. Box 420, Code P-21Indianapolis, IN 46206317-230-3618
Research Leader, Larkin LaboratoryDow Chemical Company1691 North Swede RoadMidland, MI 48640517-636-1484
Hughes Aircraft CompanyBldg. R7/MS 406Los Angeles, CA 90009213-334-4430
Staff Engineer, Manufacturing Eng.Martin Marietta Manned Space SystemsP. O. Box 29304New Orleans, LA 70189504-257-2918
304 I
James L. Breece
Owen M. Briles
Walter Brodtman
Louis J. Brothers, Jr.
Lonnie Brown
Phillip G. Brown
Robert Brown
Mark E. Brynildson
Marcanne Burrell
Vice PresidentSafety-Kleen CorporationP. 0. Box 92050Elk Grove Village, IL 60009312-694-2700
Engineering ManagerSundstrand Corporation4747 Harrison AvenueRockford, IL 61125815-226-2930
Environmental EngineerU.S. Environmental Protection Agency401 M Street RD-681Washington, DC 20460202-382-2615
Products ManagerQuaker Chemical CorporationElm and Lee StreetsConshohocken, PA 19428215-828-4250
PresidentUniclean Products Inc.3700 Osuna, Suite 703Albuquerque, NM 87109505-344-9673
Col., USAFHeadquarters, Air Force Logistics CommandWright-Patterson Air Force BaseDayton, OH 4 5433513-277-6728
General ManagerUniclean Products Inc.3700 Osuna, Suite 703Albuquerque, NM 87109505-344-9673
Environmental ChemistSandia National LaboratoriesDivision 8543P. 0. Box 969Livermore, CA 94 551415-294-3150
Boeing Aerospace and ElectronicsMS 9E-06P. 0. Box 3999Seattle, WA 98124
305
Diane Bursey
William Cain
James D. Caldwell
Joe Camahort
Christopher J. Campbell
Charlie Carpenter
Robert Carrington
Robert A. Carter
Chemical EngineerPratt & Whitney Canada1000 Marie VictorianLongueuil, Quebec J4G1A1Canada514-647-3676
Chemical Engineer, Air Logistics Ctr.Tinker Air Force BaseOC-ALC/LAPEPTinker Air Force Base, OK 7 314 5405-736-5986
Industrial EngineerHill Air Force Base00-ALC/LAOPCROgden, UT 84056801-707-2050
Senior Staff EngineerLockheed Missiles & Space Company0/47-01 B/101P. O. BOX 3504Sunnyvale, CA 94088408-742-6936
Executive DirectorSouthern California CoalitionHazardous Materials Management355 South Grand AvenueLos Angeles, CA 90071213-683-8717
Engineering & Services LaboratoryTyndall Air Force BaseHQ AFESC/RDVS, Building 1117Tyndall AFB, FL 32403904-283-6015
Program ManagerNew Mexico State UniversitySW Technology Development InstituteP. 0. Box 30001, Dept. 3 SOLLas Cruces, NM 88003
334
John Zavodjancik Senior Manufacturing EngineerPratt & WhitneyMS 104-03400 Main StreetEast Hartford, CT 06108203-565-5030
Bill Zidar Manufacturing EngineerTalley Defense Systems4551 East McKellups RoadMesa, AZ 85205602-898-2568
335
APPENDIX II
CONTRIBUTING AUTHORS*
K. K. Asada
R. S. Basu
Martha I. Beach
Lisa Brown
Joseph C. Farmer
P. M. Gallagher
Leonard W. Gray
Ralph D. Hermansen
Robert G. Hickman
K.S. Hill
Randall B. Ivey
Hughes Aircraft CompanyP. O. Box 92426Los Angeles, CA 90009
Allied-Signal, Inc.Buffalo Research Laboratory20 Peabody StreetBuffalo, NY 14210
N-Con Systems2410 Boston Post RoadLarchmont, NY 10538
U.S. Environmental Protection AgencyRisk Reduction Engineering Laboratory2 6 W. Martin Luther King DriveCincinnati, OH 45268
Lawrence Livermore National Lab.L-370Livermore, CA 94551
Phasex Corporation3 60 Merrimack StreetLawrence, MA 01843
Lawrence Livermore National Lab.L-439Livermore, CA 94 551
Hughes Aircraft CompanyBldg. E-l, MS F1572 000 E. Segundo Blvd.El Segundo, CA 90245
Lawrence Livermore National Lab.L-4 3 9Livermore, CA 94551
Hughes AircraftP. O. Box 92426Los Angeles, CA 90009
Air Force Corrosion Program OfficeWR-ALC/CNCRobins Air Force Base, GA 31098
Note: Authors in this list are those whose papers were selected under the general"Call for Papers" but whose papers were not presented at the Workshop.
337
Ronald Jackson
Ricardo B. Jacquez
E. M. Kenny-McDermott
V. J. Krukonis
J. M. Locklin
P. B. Logsdon
Benerito S. Martinez, Jr.
Keturah Reinbold
Robert F. Salerno
J. T. Snyder
Johnny Springer
U.S. Army Toxic & Hazardous MaterialsAgencyAberdeen Proving Ground, MD 21010
New Mexico State UniversityDept. of Civil, Agricultural, andGeological EngineeringLas Cruces, NM 88001
Allied-Signal AerospaceGuidance Systems DivisionTeterboro, NJ 07608
Phasex Corporation3 60 Merrimack StreetLawrence, MA 018 4 3
Douglas Aircraft CompanyMC 36-413 855 Lakewood Blvd.Long Beach, CA 90846
Allied-Signal, Inc.Buffalo Research Laboratory2 Peabody StreetBuffalo, NY 14210
New Mexico State UniversityDept. of Civil, Agricultural andGeological EngineeringLas Cruces, New Mexico 88001
U.S. Army Corps of EngineersConstruction Engr. Research Lab.CECER-ENRP. 0. BOX 4005Champaign, IL 61824