Special Report 89-35USAmCop Noveber 989of Engineers Cold Regions Research & Engineering Laboratory Analytical methods for determining nitroguanidine in soil and water Marianne E. WaSh'- ;' I) A IW-7 A,"."' .'7 .- ~c "~-. O TIC T~ :AY.V, ~.ELECT JAN 02 1990 I Prepared for U.S. ARMY -1OXIC AND HAZARDOUS MATERIALS AGENCY CETHA-TE-CR-891 79 u.Approved for public release; distribution is unlimited. 90 01 02 066
29
Embed
Noveber 989of Engineers Cold Regions Research … Report 89-35USAmCop Noveber 989of Engineers Cold Regions Research & Engineering Laboratory Analytical methods for determining nitroguanidine
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
Special Report 89-35USAmCopNoveber 989of Engineers
Cold Regions Research &Engineering Laboratory
Analytical methods for determiningnitroguanidine in soil and water
Marianne E. WaSh'-
;' I)
A
IW-7A,"."'
.'7
n!
.- ~c "~-. O TIC T~:AY.V, ~.ELECT
JAN 02 1990
I Prepared for
U.S. ARMY -1OXIC AND HAZARDOUS MATERIALS AGENCYCETHA-TE-CR-891 79
u.Approved for public release; distribution is unlimited. 90 01 02 066
UNCLASSIFIEDSECURITY CLASSIFICATION OF THIS PAGE Form Approved
8a. NAME OF FUNDING/SPONSORING 8b. OFFICE SYMBOL 9. PROCUREMENT INSTRUMENT IDENTIFICATION NUMBERORGANIZATION (if applicable)
8c. ADDRESS (City. State, and ZIP Code) 10. SOURCE OF FUNDING NUMBERS
PROGRAM PROJECT TASK WORK UNITELEMENT NO. NO. NO ACCESSION NO
11. TITLE (Include Security Classification)
Analytical Methods for Determining Nitroguanidine in Soil and Water
12. PERSONAL AUTHOR(S)
Walsh, Marianne E.13a. TYPE OF REPORT 13b. TIME COVERED 14. DATE OF REPORT (Year. Month. Day) 15. PAGE COUNT
FROM _ TO , November 1989 29
16 SUPPLEMENTARY NOTATION
17, COSATI CODES 18. SUBJECT TERMS (Continue on reverse if necessary and identity by block number)
FIELD GROUP ROUP Analytical method, Munitions;, Explosives . Nitroguanidine ,')
Groundwater pollution, Pollution. (" -19. ABSTRACT (Continue on reverie if necessary and identity by block number)" I
i t
- Methods were develoed for determining nitroguanidine in soil and water. The soil method involves extracting a2-g sample with watr usiifgan ultrasonic bath. Soil extracts and water samples are filtered through a 45 nunembraneprior to determi -PLC.eparations are achieved on a mixed-mode RPI8/cation column eluted with 1.5mL/min of water; detection is bU ( = 263 nm). Certified reporting limits were estimated at 5.0 pig/L for water and0.5 ipg/L for soil.
20 DISTRIBUTION/AVAILABILITY OF ABSTRACT 21 ABSTRACT SECUR!TN' CLASSIFICATION
] UNCLASSIFIED/UNLIMITED [ SAME AS RPT [1 DITC USERS Unclassified22a NAME OF RESPONSIBLE INDIVIDUAL 22b TELEPHONE (Include Are Code) 22c OFFICE SYMBOL
Marianne E. Walsh 603-646-4 100 CECRL-EA
DD FORM 1473, 84 MAR 83 APR edit:on mriy be used until exnausted SLCU CLASSIFICATiON OF TH. PAGE
All other edcIons are obsoleteUNCLASSIFIED
PREFACE
This report was prepared by Marianne E. Walsh, Research Physical Scientist,Applied Research Branch, Experimental Engineering Division, U.S. Army ColdRegions Research and Engineering Laboratory. Funding for this research was pro-vided by the U.S. Army Toxic and Hazardous Materials Agency, Aberdeen ProvingGround, Maryland (R-90 Multi-Analytical Services), Martin H. Stutz, Project Monitor.
The author gratefully acknowledges Dr. Thomas F. Jenkins for technical guid-
ance and review of the manuscript, and Kathleen Koehler and Patricia Schumacherfor laboratory support during this project. Paul Miyares is also acknowledged forreviewing the manuscript.
The contents of this report are not to be used for advertising or promotionalpurposes. Citation of brand names does not constitute an official endorsement or
Results and discussion.................................................................... 4Column and eluent selection ........................................................ 4Optimum wavelength determination.............................................. 4Instrument calibration................................................................. 4Kinetic study ............................................................................ 5Extraction of various soils to test for interference .............................. 5Spike recovery studies ................................................................ 5
1:onclusions .............................................................................. 7Literature cited............................................................................. 7Appendix A: Computer output for USATHAMA IRPQAP software.......... 9Appendix B: Reversed-phase HPLC method for the determination of
nitroguanidine in water and soil in USATHAMA format................ 15
ILLUSTRATIONS
Figure1. Chemical structure of nitroguanidine........................................... 12. Chromiatogram of aqueous nitroguanidine on a RP18/cation exchange
column eluted with 1.5 mL/min of water................................ 23. Optimal wavelength determination for nitroguanidine.................... 44. Kinetic study for extraction of nitroguanidine from soil ................... 55. Reporting limit determination for nitroguanidine .......................... 7
TABLES
Table1. Physical properties of nitroguanidine .............................................. I12. Summary of high-performance liquid chiromatographic methods
for nitroguanidine ............................................................. 23. Calibration standards for nitrogUanidine ....................................... 34. Solutions for spike recovery test for water method.......................... 35. Spiking solutions for spike recovery test for soil method .................. 36. Retention times for nitroguanidine using three Columns tested.......... 47. Lack-of-fit and zero-intercept tests .............................................. 58. Recovery of nitroguanidine during 4-day spike recovery study for
water method .................................................................. 69. Recovery of nitroguailidin( during 4-day spike recovery study for
For many years the making of munitions for the The physical properties of nitroguanidine (NQ)Army resulted in contamination of the environ- are listed in Table 1. As a pure substance, nitrogua-ment surrounding production sites. In the 1970s, nidine exists in two crystal forms, alpha and beta,the Army sought to correct this situation by identi- which have the same melting point (U.S. Armyfying and cleaning up affected areas. As part of this 1984). The alpha form, whose crystals develop intoeffort, the U.S. Army Toxic and Hazardous Materi- long, thin, flat needles, is most commonly used asals Agency (USATHAMA), under the Installation an explosive. The two forms differ slightly in waterRestoration Program, has been actively developing solubility, but their solubility curves cross at 25 andanalytical methods for detecting unique military 1000C and have values of 4.4 and 82.5 g/L at thesecompounds, such as explosives and propellants, in two temperatures (U.S. Army 1984). Thus, nitrogua-environmental samples. Under the auspices of USA nidine has a water solubility an order of magnitudeTHAMA, the U.S. Army Cold Regions Research greater than most other explosives, such as TNTand Engineering Laboratory has been charged with and RDX. NQ is sparingly soluble in ethanol, metha-developing methods fornitramines, nitroaromatics, nol and acetone. It is essentially nonvolatile.tetrazene, and, most recently, nitroguanidine in Nitroguanidine in aqueous solutions exists inwater and soil. the two tautomeric forms shown in Figure 1 (Kemula
Nitroguanidine Nf'N= C(NH2)NHN& is a et al. 1970, Kaplan et al. 1982). The nitroimine existscomponent, along with nitroglycerine and nitro- in acidic, neutral or slightly basic solutions.cellulose, of triple base propellant. Its relativelyhigh solubility in water"A-gTl- --increases the likelihood of groundwater contami- NH* NH2
nation when water used to clean cutting blades and C -N . C - NH- NO?
Nitroguanidine's physical and chemical char-Crystal density (g/cm') 1.72 acteristics preclude analysis by gas chromatogra-
Melting point ( 0C) 232 phy, but several liquid chromatographic methods(decomposes) (Table 2) have been developed using both UV and
Solubility (g/L) electrochemical detection. M-ost of these methodswater 25'C 4.4 use a reversed-phase C8 or '18 column (Kaplan etwater 100°C 82.5 al. 1982, Burrows et al. 1984, Maskarinec et al. 1986,
CAS reg no. 1556-88-71 Ogle and Westerdahl 1986, Manning and Maskar-inec 1987) eluted with a mobile phase that is pre-
Table 2. Summary of high-performance liquid chromatographic methods fornitroguanidine.
Mobile Retention tine ConcentrationColumn phase (rain) range Reference
Dupont 90/10 water- 2.8 100 pg/L (water) Kaplan et a]. (1982)Zorbax ODS methanol 1 Pg/g (soil)(25 cm x 4.6 mm)
Dupont water 6.0 0.5-10 mg/L Burrows et al. (1984)Zorbax C8 0.8 mL/min(25 cm x 4.6 mm)
Dupont 20/80 5.0 - Maskarinec et al. (1986)Zorbax C18 propanol-aqueous Manning and Maskarinec (1987)(25 cm x 4.6 mm) sodium acetate
C8 water 1.78 0.5-26 mg/L Ogle and Westerdahl (1986)3 mL/min
dominantly water. Nitroguanidine is not well re- saved using a Hewlett Packard 9114B disk drive.tained on these columns and elutes early, making Data were also recorded on a Linear Model 500interferences likely when environmental samples strip chart recorder.are analyzed. This report outlines the development Separations were achieved on a mixed modeof a High Performance Liquid Chromatographic RP18/cation exchange column (Alltech Associates)(HPLC) method for the analysis of nitroguanidine that was eluted with 1.5 mL/min of degassed water.in soil and water samples. Separation is achieved Retention time for nitroguanidine was 4.4 minutes.using a mixed mode RP18/cation exchange col- Figure 2 shows a typical chromatogram.umn.
ChemicalsNitroguanidine was obtained from Aldrich and
EXPERIMENTAL METHODS was recrystallized from water. The crystals of ni-troguanidine were dried to constant weight in the
Instrumentation dark in a vacuum desiccator. Water used for theThe RP-HPLC determinations were made using spike recovery, preparation of standards and for
two systems. One system was composed of a Perkin the mobile phase was purified by a Milli-Q Type 1Elmer Series 3 pump and a Perkin Elmer LC-65T Reagent Grade Water System (Millipore). Watervariable wavelength UV detector. The other sys- for the mobile phase was degassed by first boiling,tem was composed of a Spectra Physics SP8810 and then, after cooling, vacuum-filtering through apump and a Spectra Physics SP8490 variable wave- Whatman CF-F microfiber filter. The water re-length UV detector. Both systems were interfaced mained under vacuum for at least I hour prior towith a Dynatech Precision Sampler Model LC-241 use.autosampler containing a Rheodyne Model 701 OAsample loop injector with a 100-iL sampling loop. SoilsData were recorded on a Hewlett Packard 3393A The soil for the spike recovery study wasdigital integrator set in peak height mode and USATHAMA standard soil (blank). Several other
soils from present and former Army sites weretested for interferences.
Figure2. Cliromato- Optimum wavelength determinationgran of aqueous iti- The optimum wavelength setting was deter-troguanidine (190 mined using the stopflow scan capability of thepag/L) on a RP18/ Spectra Physics variable wavelength detector. An
_____ cation exchange col- aqueous sample of nitroguanidine was injectedntin eluted with 1.5 onto the analytical column, and when the analyte
r,,, .. nL/iMin of water. was detected, the eluent flow was stopped and the
2
UV spectrum in the range 210-300 nm was re- Calibration standardscorded at a scan rate of 1 nm/s. An aqueous nitroguanidine stock solution was
prepared by dissolving 95.1 mg of recrystallizedKinetic study nitroguanidine in I L of water. Then two independ-
To determine the length of time required to ent series of standards were prepared by first dilut-extract nitroguanidine from soil, a kinetic study ing the stock standard 5 to 500 mL with water towas conducted. Actual contaminated soils were make duplicate 951-pg/L standards. Subsequentnot available for these experiments; therefore, four dilutions were madefrom thesestandards as shownsamples of USATHAMA standard soil were con- in Table 3.taminated in the laboratory by adding 1.0 mL of2000-gg/L aqueous NQ to 2-g subsamples of soil. Spike recovery studiesNQ concentration in the soil was thus I pig/g. Then,to thoroughly dry the soil and to hasten any inter- Wateraction between the nitroguanidine and the soil Reporting limits were obtained using the Hu-constituents, the soil samples were baked at 50'C baux and Vos (1970) method outlined in thefor 30 hours and then air dried for 56 hours. An USATHAMA Installation Restoration Quality Assur-unspiked soil was treated in the same manner. anice Program (USATHAMA 1987) for Class 1 certi-
The soil samples were extracted with 50.0 mL of fication. Samples were spiked with known quanti-water in a 2.5- x 20-cm glass test tube by vortex ties of nitroguanidine and analyzed on each of fourmixing for 30 seconds and sonicating in an ultra- days. A spiking stock solution was prepared bysonic bath (Cole Palmer Model 8855-10). During dissolving 100 mg of recrystallized nitroguanidinesonication, 5.0-mL aliquots were removed at 5,30, in I L of water. Then, a series of spiked water60, 120, 240 and 480 minutes. Soil extracts were samples corresponding to 0.5, 1,2,5 and 10 times aallowed to settle for 30 minutes and then filtered Target Reporting Limit (TRL) of 10 pg/L was pre-through Millex-HV (0.45-htm) filter units prior to pared. The 10-TRL sample was made by placinganalysis. 1.00 mL of the 100 mg/L stock solution in a 1-L
volumetric flask and bringing the flask to volumewith water. The concentration of the 10 TRL solu-
Table.Ciaio station was further diluted as shown in Table 4. Priornitroguanidine, to analysis, each water sample was filtered through
a 0.45-jm Millex-HV filter unit using a 20-mL
951igL voluaetric flask concentration disposable BD syringe. The first 10 mL of filtratestandard (was discarded and the remainintg 10 ,nL retainedfor analysis.
0.50 100 4.751.00 100 9.51 Soil2.00 100 19.0 A series of spiking solutions was prepared from5.00 100 47.5
10.0 100 95.1 the 100-mg/L spiking stock. The 10-TRL spike so-20.0 100 19050.0 100 475
Table 5. Spiking solutions for spike recovery testfor soil method.
Table 4. Solutions for spike recoverytest for water method. CapacityI of Equialent*
Aliquot of volumetric Solution concentration
Aliquot of Capacity of Solution lO.O-ing/L flask conceittrat iot il Soil
No dilution 10051NAssuming 1.(X mL of spike solution added to 2 g it soil.
3
lution was made by placing 50.0 mL of stock in a500-mL volumetric flask and diluting to volumewith water. The the 10-TRL spike solution (10.0mg/L) was diluted to make a series of spike solu-tions as shown in Table 5.
Several 2-g subsamples of USATHAMA stan-dard soil were weighed out to the nearest 0.1 g in2.0- x 25-cm test tubes equipped with Teflon-lined ,screw caps. Each soil subsample was spiked with gone of the spiking solutions listed in Table 5 and 2allowed to stand 1 hour uncapped. Then 50.0 mL of Fwater was added and each sample vortex mixed for30 seconds and sonicated for 2 hours. Samples wereallowed to settle for 30 minutes prior to filtrationthrough Millex-HV filter units.
RESULTS AND DISCUSSION
Column and eluent selection I I I I Figure3. Optinaluwav'-
Three analytical columns were tested during 2 10 230 250 270 290 h'igth deternationiforthe method development. These included a re- wovelenqth Inm) ntrgitanidie.versed-phase LC-18 column (Supelco, Inc.), a cat-ion exchange LC-SCX column (Supelco, Inc.), anda mixed mode RP18/cation column (Alltech Asso-ciates). Retention times for nitroguanidine usingthe various columns and eluents are shown in Optimum wavelength determinationTable 6. The flow rate was 1.5 mL/min for each The range 210-300 nm was scanned to deter-column-eluent combination tested. mine the optimum monitoring wavelength for
Since the Mixed Mode column resulted in the nitroguanidine (Fig. 3). The absorption maximumlong st retention time for nitroguanidine, it was was about 263 nm.chosen as the analytical column for this study. Alonger retention time is desirable to avoid interfer- Instrument calibrationencesfrom the earlyelutingcompounds often found To determine if the detector response for ni-in environmental samples and to improve the abil- troguanidine was a linear function of analyte con-ity to resolve NQ from other solutes. centration, the calibration data were subjected to a
regression analysis for a non-zero-intercept model(y = a + bx) and a zero-intercept model (i = I'x). The
Table 6. Retention times for nitroguanidine regression coefficients a, 11 and 1' were estimatedusing three columns tested. Flow rate was 1.5 using the method of least squares (Table 7).mL/min. The fitted equations for both models were sub-
jected to the Lack-Of-Fit (LOF) test (USATHAMARetoitio'j time 1987). A linear model was found to be acceptable at
Colunin Eluent 001,) the 95', confidence level. The intercept was thentested to determine if it was significant!y different
LC8 Water. 3.1 from zero. The F-ratio was calculated by dividingLC18 1:9 methanol-water. 2.5 the differences between the residual sum of squaresLC18 1.4 methanol-water for the non-zero-intercept and zero-intercept mod-
with ion-pairing reagent,pH=3. 2.4 els by the residual mean square for the model with
LCIS Water with ion-pairing the non-zero-intercept. Since the calculated F-ratioreagent, pH=3. 2.7 was less than the critical value at the 95% confi-
LCSCX 0.05 M KH2,I04. 2.8 dence level, the zero-intercept linear model wasLCSCX I:1) methanol:KHP0 4. 2.7 accepted. Thus, daily calibration may be obtainedRPIS/Cation 0.05 M KH211'O4 , 4.1RMS/Cation Water. 4.2 using a zero-intercept model and a single high-con-
centration replicated standard.
4
Table 7. Lack-of-fit and zero-intercept tests.
Analysis of Residual Variations
Modcl witl iotercept Mode)lo t t o::ih i originY = (611.2479980)+(199.9723840)X Y = (200.9022340)X
Zero intercept accepted Calculated F: 1.226919863 Critical 95'. F: 4.,6)
1.1 was repeated with 50.0 mL of solvent and filtrationwas much easier.
1.0- 0
- - Spike recovery studiesC 0.9 - Spike recovery studies were conducted to allow
estimation of reporting limits. Water and soil0.8 -e samples were spiked over the concentration ranges
0 o Rep2 of 5.0-100 pg/L and 0.25-5.0 i'g/g respectively.
0.7 * Rep 3 Blank water and soil samples were also analyzed.€I Rep 4 -
06 _e 4Certified reporting limits (CRLs) were calcu-
0.6 i I I I lated using the method of Hubaux and Vos (1970).0 40 80 120 160 200 240 Bartlett's test was used to compare the variances at
Time (min.) each target level (Tables 8 and 9). For water, the
Figuire 4. Kinetic studl for extractio1 of ,itroglalidin range of homogeneous variance included 0.5-5from Sol. fTRL levels. For soil, the range of homogeneousfrom soil. variance included the entire data set. The data for
each of the four days were pooled and tested forKinetic study lack of fit. For the water samples, data for the entire
Results of a kinetic study, using water and a range were linear; however, the intercept was sig-
sonic bath to extract soil, are shown in Figure 4. nificantly different from zero. When the highestEquilibrium is reached between 120 and 240 min- target value was dropped, the intercept's differ-utes. Aliquots removed at 480 minutes were a ence from zero became nonsignificant. For the databrownish-orange color, and the chromatograms for the soil samples, the model with intercept wasfor these samples contained several large peaks linear, but the model through the origin was not.that prevented the baseline from returning to zero The intercept (-0.089) was statistically significantlyprior to the start of the NQ peak. different from zero, but from a practical stand-
point, this is a small difference.Extraction of various soils to CRLs were obtained from the X values corre-test for interference sponding to the point on the lower confidence limit
Twelve soils from various Army sites were ex- curve where the Y values natched the values of Ytracted with 10.0 mL of water in a sonic bath for 480 on the upper confidence limit curve at X = 0 (Fig. 5).minutes. No interfering peaks were observed in the Method reporting limits were 5.0 pg/L and 0.5 pggchromatograms, but some of the extracts were forwaterandsoil, respectively. Method accuracies,extremely difficult to filter, even after centrifuga- based on percent recovery, were 106 and 989;tion at 3500 rpm for 30 minutes. The experiment
5
Table 8. Recovery of nitroguanidine during 4-day spike recovery studyfor water method.
Target Bartlett'sSpike Concentration Fomund cocetration (pg/L) testle'el (pg/L) Day I Day 2 Day 3 Day 4 Mean Vaijaee (X
6 -T -T I T T I I I I I I T 1 1 7 1 1 1 1 - L IT E R A T U R E C IT E D
Burrows, E.P., E.E. Brueggemann, S.H. Hoke, E.H.McNamee and L.J. Baxter (1984) Nitroguanidinewastewater pollution control technology: Phase II.Wastewater characterization and analytical meth-ods development for organics. USA Medical Bio-engineering Research and Development Labora-tory, Technical Report 8311.
3Hubaux, A. and G. Vox (1970) Decision and detec-0 0 tion limits for linear calibration curves. Analytical
o Chemishy, 42:840-855.o 2 Kaplan, D.L.,J.H. Cornell and A.M. Kaplan (1982)
Decomposition of nitroguanidine. EnvironmentalScience and Teclnology, 16: 488-492.
1 Kemula, W., M.K. Kalinowski, T.M. Krygowski,0 J.A. Lewandowski and A.J. Walasek (1970) Inves-
0 tigation of N-nitroderivatives. Equilibria of nitro-
010 o2 3 4 ureaand nitroguanidine in aqueous solutions. Bulc'-
CRL Target Concentration (pJg/g) tin tie L'Academie Polonaisedes Sciences, 18(8): 455-461.Manning, D.L. and M.P. Maskarinec (1987) Analy-
Figure 5. The reporting limit determination for sisofnitroguanidineinaqueoussolutionsbyHPLCnitroguanidine, with electrochemical detection and voltammetry.
Oak Ridge National Laboratory, Contract DE-AC05-84 OR 21400.Maskarinec, M.P., D.L. Manning and R.W. Harvey(1986) Application of solid sorbent collection tech-
CONCLUSIONS niques and high performance liquid chromatogra-phy with electrochemical detection analysis of ex-
Methods were developed for determining ni- plosives in water samples. Oak Ridge Nationaltroguanidine in soil and water. Soils are extracted Laboratory, Final Report, ORNL/TM 10190.with water for 2 hours in a sonic bath, then the ex- Ogle, E.E. and R.P. Westerdahl (1986) On-linetracts are filtered through 0.45-pm membranes. monitors for water pollutants. USA Armament Re-Water samples are simply filtered. Nitroguanidine search and Development Center, Contractor Re-is determined for both by RP-HPLC on a mixed- port ARAED-CR-86015.mode RP1 8/Cation column eluted with 1.5 mL/min U.S. Army (1984) Military explosives. Technicalof water and a UV detector set at 263 nm. Calibra- Manual TM 9-1300-214, September.tion data were found to be linear over the range USATHAMA (1987) U.S. Army Toxic and Hazard-4.75-951 ag/L. Spike recovery tests were carried ous Materials Agency Installation Restoration Qual-out. Certified reporting limits were estimated at 5.0 ity Assurance Program. Maryland: Aberdeen Prov-pg/L and 0.5 pg/g for water and soil respectively. ing Ground.
7
APPENDIX A: COMPUTER OUTPUT FOR USATHAMA IRPQAP SOFTWARE
Part 1: CRL for water
CERTIFICATION ANALYSIS Report Date: 10'17 R
Method Name: NQWATCERTTRUNI Laboratory: CRCompound: NQ Analysis Date: 04;12 83Units of Measure: UGL Matrix: WA
ANALYSIS OF RESIDUAL VARTATTONS
--- Model with Intercept . -.. Model through the Cr77n -Y = (0.787082774) + (1.062316560)X Y = (1.084567-00)X