The Phvsical & Chemical Characterization of Biodiesel Low w Sulfur Diesel Fuel Blends Sponsored by the National Biodiesel Board University of Missouri December 30,199s Final Report Principal Investigator Leon Schumacher and Research Associates Anand Chellappa, Will Wetherell, & Mark D. Russell
85
Embed
The Physical & Chemical Characterization of Biodiesel Low Sulfur ...
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
The Phvsical & Chemical Characterization of Biodiesel Low w Sulfur Diesel Fuel Blends
Sponsored by the National Biodiesel Board University of Missouri
December 30,199s Final Report
Principal Investigator Leon Schumacher
and
Research Associates Anand Chellappa, Will Wetherell, & Mark D. Russell
4. CONCLUSIONS OF TIXIS PROJECT. ....................................... 10 Table 5. Summary of properties that were analyzed ........................... 12
The Physical Characterization of BiodieseU Low Sulfur Diesel Fuel Blends
Sponsored by the National Biodiesel Board University of Missouri
December 30,1995 Final Report
Executive Summary
The University of ?+fissouri Department of Agricultural En-tiesring was contracted to examine the physical characteristics of biodiesel when mixed with low 4fi-x diesel tieI. Engine manufacturers remain concerned about the physical and chemical makeup of the fuel used to power the eqgines that they manufacture and sell C ummins Engine Company is particularly interested in the oxidative stability, thermal stability, cold flow propekes, and water separability of each respective blend.
Cummins, the EPA and the National Renewable Energ Laboratory requested that speciiic characteristics about the biodiesel and biodiesel blends be well documented. Their suggestions enabled the Agricultural Engineering Department to compile the list of physical and chemical characteristics found in Table 5 of this final report.
Anand Chellappa prepared three identical samples of 0, 20, 30, 50, 70, and 100 percent by volume blends using certikation diesel fuel purchased from Phillips 66. Two of these samples were analyzed by Cleveland Technical Center. The third sample remains on campus for backup purposes. Will Wetherell purchased petroleum diesel fuel local(v from MFA Oil Company. Will prepared three identical samples using the same percent by volume blend concentrations. TWO of these samples were analyzed by Cleveland Technical Center. The third sample remains on campus for backup purposes.
Anand Chellappa determined the pour point, cloud point, fatty acid content, refractive tide% viscosity, density and specdic gravity for the certification diesel-biodiesel blends. Cleveland Technical Center analyzed the balance of the certikation diesel-biodiesel chemical blend
properties found in Table 5. Cleveland Technical Center analyzed &I the diesel-biodiesel chemical blend properties found in Table 5 for the fuel blends that were prepared using the diesel fuel purchased from MFA Oil Company.
The values associated with the chemical character&&s were systematically identified. Specific objectives of the project involved assessing these properties with 0, 20, 30, 50, 70, & 100 percent biodiesel blends. Regression analysis was conducted’with these data and it was determined that simple mathematical interpolation methods could be used to predict some of the characteristics. For example, the cloud and pour points that were measured varied in a linear fashion with the amount of biodiesel blended with the certification diesel fkeL Therefore it appears that simple mathematical interpolation could be used to estimate or predict the cloud and pour point characteristics for a _tien blend of diesel fuel and biodiesel, provided the researcher is able to determine the cioud and pour points of the parent fuels.
Respectfully submitted,
Leon Schumacher Principal Investigator
2
1. INTRODUCTION. . .
Research funded by the National Biodiesel Board has shown that an engine tided with a 20
percent biodieseK30 percent low sulfix diesel blend exhibited exhaust emission levels that are much lower than that obtained when the engine was fueled with 100 percent low sulfirr diesel fuel. The potential exists that a 70 percent biodieselG0 percent low sulfkr diesel fuel blend could quaiiry as an alternative fire1 for federal fleets that have been mandated to use alternate fuels. Such an increasing interest in blended fuels calls for detailed information concerning the physical and chemical properties of fuel blends that range f?om 100 percent biodiesel to 100 percent low sulfur diesel me!. This knowledge would facilitate a more effective commercialization of biodiesel as a me1 for modem diesel en-ties.
For the most part, properties have been estimated on a proportional basis (proportional to content of biodiesel in the blend) between the properties of 100 percent biodiesel and 100 percent low sulfur diesel fiteL However, properties do not always change in a linear fashion with blend composition. As an example, Van Gerpen (1994) found that flashpoint changed in a non-linear fashion when increasing the concentration of biodiesel in the fuel. In this project, various blends of biodiesel and low sulfur diesel me1 were systematically characterized, so that correlations between the fuel properties could be derived.
2. OBJECTIVES.. .
The overall goal for this research activity was to characterize various biodieseUlow sulfur diesel fuel biends. Specific activities that were conducted to attain this goal are listed below:
1. Experimental measurement of fuel related variables, as outlined by the fuels division of the EPA, for blends of 20, 30, 50, and 70 percent biodiesel/diesel fuel blends.
2. Analysis of the data obtained to determine whether the variables vary in a linear fashion with blend composition Corn 100 percent biodiesel fuel to 100 percent low sulfix diesel fuel.
3. To determine if one can accurately predict a range of values that can be expected for the various fire1 related variables, in blends ran-&g from 100 percent biodiesel fuel to 100 percent low sulfk diesel fuel.
4. The data were analyzed using statistical methods to provide a measure of accuracy of the experimental procedures that were followed.
3. RESULTS/FINDINGS.. .
3.1 ASTM Test Methods Selection The physical and chemical properties were identified in the work proposed by MU. The ASTM
3
test method that would be used to determine these qualities are outlined in Table 5. Rich Heiden of Heiden and Associates and James Gardner, NBB subcontractors, reviewed these test procedures and reponed that these were adequate to assess the physical and chemical characteristics of the biodieseVdiese1 fuel blends.
3.2 SecurinF Fuel for Testing The me1 used to prepare the blends for our Batch 1 analysis was low sulfur Xi! diesel certification fuel. This was procured ii-om Phillips 66 Company in Bagor, Texas. The low suifk #2 diesel used to prepare Batch 2 for analysis was purchased locally from MZFA Oil in Columbia, MO. The biodiesel fuel was manufhcrured by Proctor & Gamble and procured f?om Midwest Biofiels.
3.3 Freuarstion of Blends When preparing the blends for this analysis, the specihc gravity (gramsIliter) of the parent fkel was used to insure that each blend was a 20, 30, 50 or 70 volume percent blend ofbiodiesel and petroleum diesel fuel. These blends were prepared using a Sartorius digital balance that is accurate to 0.1 mg. This method of preparing a blend is commonly used when preparing small blends (volume) of a _&en liquid. \
3.1 Analvsis and Oualitv Assurance Anand Chehappa determined the pour point, cIoud point, fatty acid content, refractive index, viscosity, density and specific gravity for the certification diesel-biodiesel blends. Cleveland Technical Center analyzed the balance of the chemical properties found in Table 5 for the biodiesel/certitication diesel fuel blends. Cleveland Technical Center analyzed & the chemical properties found in Table 5 for the biodiesel/diesel fuel blends that were purchased f?om MFA Oil company.
Cleveland Technical Center performed quality assurance tests to determine the repeatability of the tidings the reported for each blend. A summary of this report can be found in Appendix F. As noted in this report, ASTM DS6, D93, D130, D3338, D613, D13 19, D4294, D4S2, D664, D2075, and DIS9 were re-analyzed and the results of each second arklysis can be found in this report.
3.5 ComDarison of Physical\Chemical Prooerties Linear regression analysis was used to determine the relationship between the blend and the chemicaUphysica1 characteristics that were evaluated as a part of this investigation. The ‘R’ ” of each analysis was examined to determine the strength of each relationship. Variables that had an R’ that was less than .950, although strongly correlated, were deemed to have curvilinear tendencies. This tendency is clearly evident as one reviews the graphs that are found in Appendix C and D of this report. As such, variables that had an R* that was less than .950 were deemed curvilinear. Simple interpolation of the chemical/physical properties of a biodieseYdiese1 fuel blend was not an appropriate when the R* was less than .950. a’ . . .._
‘.’ __
’
4
3.5.1 Refractive Index
The refractive index of the mei samples was determined using an Abbe Digital Refractometer and a Haake Constant Temperature Bathe (precision 0.2”C). Ref?actie index was found to vary in a linear fashion with the amount of biodiesel bIended with the fuel The correlation coefficient R2 of the linear fit was found to be greater than 0.999.
An accurate reading concerning the concentration of the blend, however, can only be accomplished when the refractive index of the parent fuels are predetermined.
3.5.2 Fam Acid Methvl Ester (FAME) content
Analysis of the fuel blends for Fatty Acid Methyi Ester (FAME) content was done using gas chromatographic (G-C) techniques. A 6’ X l/S” stainless stee1 column packed with GP 3% SP-23 10 / 2% SP-2300 on 100/120 Chromosorb was procured from Supelco Inc. This column was mounted in a Varian 3700 GC. The GC is equipped with a Flame Ionization Detector (FID) and is connected to a Hewlett Packard integrator. Calibration was accomplished using standard ester samples. The column oven temperature was maintained at 190’ C for 2 min and then ramped to 220’ C at 2’ C/m& with nitrogen as the carrier gas at 20 ml/min.
The FAME content was found to vary in a linear fashion with the amount of biodiesel blended with the ftreL The GC chain ler@ distribution (Oh by weight) averaged over two analysis for the biodiesel was:
The data associated with the distillation points of the biodiesel blends are reported in Tables 1 and 3. The data varied in a curvilinear fashion as the concentration of biodiesel increased in relation to the petroleum diesel f$eL The linear regression coefficient R’ for each distillation point varied from 0.544 to .982 for the ZocafIy purchased low-sulfur diesel fuel (Table 4). The linear correlation coefficient for each distillation point of the certification diesel fuel ranged from .8 11 to .95 1 (Table 2) . Based on this analysis a simple mathematical interpolation is not an adequate method to determine the distillation curve of a biddiesel blend. Chemical analysis procedures or curvilinear regression analysis procedures must be used to determine the distillation-curve of a biodieseVdiese1 fuel blend (see Figures 1, la, 22, & 22a.).
5
3.5.4 Densirv
Anand Chellappa used a quartz thermometer (Hewlett Packard) which has an accuracy of 0.00 1” C, and a Fisher Scientific thermostat for accurate temperature controL A Canon-Fenske viscometer was used to determine these characteristics.
Density was found to vary in a curvilinear fashion as the concentration of biodiesel increased in reiation to the petroleum diesel fuel The correlation coefficient R’ of the curve fit was found to be greater than 0.999. In other words, simple mathematical interpolation is not an adequate method to determine blend density. Chemical analysis procedures or curvilinear regession analysis procedures must be used to determine the density of a biodiesel/diesel fuel blend (see [email protected] 2 & 23).
3.5.5 API Gravitv
The data associated with the API Gravity of the biodiesel blends is reported in Tables 1 and 3. The data varied in a linear fashion as the concentration of biodiesel increased in relation to the petroleum diesel fuel. The linear correlation coefficient R2 for API Gravity was .994 for the locally purchased low sulfur diesei me1 (Table 4). The linear correlation coefficient R2 for the certiiication diesel fuel was .999 (Table 2). A simple mathematical interpolation of the data, providing one knows the API Gravity of the blendstocks, can provide a good estimation of the calculated API Gravity for a biodiesel blend ( see Figures 4 & 25).
3.5.6 Cloud and Pour Points
Cloud point and pour point measurements for 0, 20, 30, 50, 70, and 100 percent biodieseudiesel fuel blends were carried out following the ASTM D2500 and D97 procedure. Three trials were carried out for each blend. Both cloud point and pour point varied in a linear fashion with the amount of biodiesel blended with the petroleum diesel fuel. The correlation coefficient R2 was found to be greater than 0.99 (see Figures 5, 6, 26, & 27). A simple mathematical interpolation of the data, providing one knows the cloud and pour points of the blendstocks, can provide a good estimation of the cloud and pour points for a biodiesel blend.
3.5.7 Cold Filter Plug Point
The data associated with the cold hlter plug point of the biodiesel blends are reported in Tables 1 and 3. The data varied in a curvilinear fashion as the concentration of biodiesel increased in relation to the petroleum diesel fuel. The linear regression correlation coefficient R* for the cold filter plug point of the certi&ation.diesel fuel was .964 (Table 2). The linear regression correlation coefficient R’, for the locally purchased low sulfur diesel fuel was .92 1 (Table 4). Simple mathematical interpolation is not an adequate method to determine the cold flter plug point of a biodiesel blend (see P&res 7 & 28). - ,,’ ej
.1 . . . -,
_
6
3.5.8 Flash Point
The data associated with the flash point of the biodiesel blends are reported in Tables 1 and 3. The data varied in a cur+bnear fashion as the concentration ofbiodiesel increased in relation to the petroleum diesel me1 The linear regression correlation coefficient R’ for the flash point of the certification diesel fuel was .759 (Table 2). The linear regession correlation coefficient R’ for the locally purchased low sulfur diesel fuel was .809 (Table 4). Simple mathematical interpolation is nor an adequate method to determine the flash point of a biodiesel blend (see Fi,g.nes 8 & 29). Chemical analysis procedures or curvilinear regression analysis procedures must be used to determine the flash point of a biodiesel/diesel fuel blend.
3.5.9 Corrosion
The data associated with corrosion of the biodiesel blends are reported in Tables 1 and 3. The data reported concerning corrosion was 1A regardless of the blend or the diesel fuel that was used to prepare the blend. In this case, mathematical interpolation is a mute effort.
3.5.10 Viscositv
The data associated with the viscosity of the biodiesel blends are reported in Tables 1 and 3. The data varied in a linear fashion as the concentration of biodiesel increased in relation to the petroleum diesel meL The linear regression correlation coefficient R2 for the viscosity of the certiilcation diesel fuel was -962 (Table 2). The linear regression correlation coefficient Rz for the locally purchased low s&k diesel meI was ,99 1 (Table 4). Simple mathematica.I interpolation can be used to estimate the viscosity of a biodiesel blend (see Figures 9 & 30).
3.5.11 Heat of Combustion
The data associated with heat of combustion of the biodiesel blends are reported in Tables 1 and 3. The data varied in a linear fashion as the concentration ofbiodiesel increased in relation to the petroleum diesel fuel. The linear regression correlation coefficient R2 for heat of combustion of the certihcation diesel fire1 was .990 (Table 2). The linear regession correlation coefficient R’ for the locally purchased low suIfix diesel fuel was .999 (Table 4). Simple mathematical interpolation can be used to estimate the heat of combustion of a biodiesel blend (see Figures 10 & 31).
3.5.12 Cetane Number
The data associated with the cetane number of the biodiesel blends are reported in Tables 1 and 3. The data varied in a curvilinear fashion as the concentration of biodiesel increased in relation to the petroleum diesel fuel. The linear regression correlation coefficient R’ for the cetane number of the certification diesel fuel was .595 (Table 2). The linear regression correlation coefficient R2 for the locally purchased low sulfur diesel fuel was .795- (Table 4). Simple mathematical
’ interpolation is not an adequate method to determine the cetane number of a biocliesel blend (see
7
Figures 11 & 32). Cetane engine test procedures (D6 13) must be used to determine the cetane number for a given biodiesevdiesel mei blend.
3.5.13 Aromatics Content
The data associated with the aromatics content of the biodiesel blends are reported in Tables 1 and 3. The data varied in a linear fashion as the concentration ofbiodiesel increased in relaticn to the petroleum diesel fkeL The linear regression correlation coefficient R’ for the aromatics content of the cetication diesel fuel was .998 (Table 2). The Linear regression correlation coefficient R2 for the locally purchased low sulfur diesel fuel was .997 (Table 4).
This test procedure (D13 19) does not measure the “real” aromatics associated with a biodiesel blend. D 13 19 mistakenly identiiies the double bonds commonly found in biodiesel as the double bonds (normally associated with benzene, or aromatics) that are attached to the diesel fuel molecule. Tn short, simple mathematical interpolation is not appropriate when D13 19 is used to determine the aromatics of a biodiesel/diesel fuel blend GC analysis procedures should be considered when determirk, - the aromatics of a biodiesel blend (see Figures 12 & 33).
3.5.14 Oleiinic Content
The data associated with the ole&ic content of the biodiesel blends are reported in Tables 1 and 3. The data varied in a curvilinear f&&ion as the concentration of biodiesel increased in relation to the petroleum diesel fkeL The linear regression correlation coefficient R* for the olefinic content of the certification diesel fuel was .I336 (Table 2). The linear regression correlation coefficient R2 for the locally purchased low sulfur diesel fuel was -947 (Table 4). Simple mathematical interpolation is not an adequate method to estimate the oleSnic content of a biodiesel blend (see Figures 13 & 34).
3.5.15 ParafEns Content
The data associated with the paraflins content of the biodiesel blends are reported in Tables 1 and 3. The data varied in a linear-fashion as the concentration ofbiodiesel increased in relation to the petroleum diesel meL The linear regression correlation coefficient R* for the parafFins content of the certikation diesel fuel was .998 (Table 2). The Linear regression correlation coefficient R* for the locally purchased low suLfurdiesel fuel was .997 (Table 4). Simple mathematical interpolation is an adequate method to estimate the parafEns content of a biodiesel blend (see Figures 14 & 35).
3.5.16 Carbon Content _. : ._ L.
The data associated with the carbon content of the biodiesel blends are reported in Tables 1 and 3. The data varied in a linear fashion as the concentration ofbiodiesel increased in relation to the petroleum diesel fueL The linear regression correlation coefficient R* for the carbon content of the certification diesel fuel was .996 (Table 2). .The L&ear regression correlation coefficient R* for
8
the locally purchased low .suKr diesei fbel was .985 (Table 4). Simple mathematical interpolation should provide a good estimate of the carbon content of a biodiesel blend (see Fi-gures 15 & 36).
3.5.17 Hvdrogen Content
The data associated with the hydrogen content of the biodiesel blends are reported in Tables 1 and 3. The data varied in a linear fashion as the concentration of biodiesel increased in relation to the petroleum diesel fuel. The linear regression correlation coefficient Rz for the hydrogen content of the certification diesel fuel was .990 (Table 2). The linear regression correlation coefficient R?- for the locally purchased low sulfkr diesel fuel was .946 (Table 4). Simple mathematical interpolation should provide a good estimate of the hydrogen content of a biodiesel blend (see Fi-mes 16 & 37).
3.5.18 Oxvgen Content
The data associated with the o,xygen content of the biodiesel blends are reported in Tables 1 and 3. The data varied in a linear fashion as the concentration of biodiesel increased in relation to the petroleum diesel fuel. The linear regression correlation coefficient RZ for the oxygen content of the certification diesel mei was .996 (Table 2). The linear regession correlation coefficient R for the locally purchased low sulfkr diesel fuel was .976 (Table 4). Simple mathematical interpolation should provide a good estimate of the oxygen content of a biodiesel blend (see Figures 17 & 38).
3.5.19 Sufir Content
The data associated with the m&r content of the biodiesel blends are reported in Tables 1 and 3. The data varied in a linear fashion as the concentration ofbiodiesel increased in relation to the petroleum diesel fuel. The linear regression correlation coefficient R* for the su.lfk content of the certikation diesel fuel was .972 (Table 2). The Linear regression correlation coefficient R* for the locally purchased low sulfix diesel fuel was .968 (Table 4). Simple mathematical interpolation should provide a good estimate of the sulfur content of a biodiesel blend (see Figures 18 & 39).
3.520 Neutralization Number
The data associated with the neutralization number of the biodiesel blends are reported in Tables 1 and 3. The data varied in a linear fashion as the concentration ofbiodiese1 increased in relation to the petroleum diesel fkeL The linear regression correlation coefficient R* for the neutralization number of the certification diesel he1 was .983 (Table 2). The linear regression correlation coefficient R’ for the locally purchased low sulfiu diesel fuel was -983 (Table 4). Simple mathematical interpolation should provide a good estimate of the neutralization number of a biodiesei blend (see Figures 19 & 40).
9
3.5.21 Iodine Value
The data associated with the iodine value of the biodiesel blends are reported in Tables 1 and 3. The data varied in a linear fashion as the concentration of biodiesel increased in relation to the petroleum diesel fkeL The linear regression correlation coefficient Rz for the iodine value of the certitication diesel fuel was -968 (Table 2). The linear regression correlation coefficient R’ for the locally purchased low sulfur diesel fuel was .999 (Table 4). Simple mathematical interpolation should provide a good estimate of the iodine value of a biodiesel blend (see Figures 20 & 4 1).
Z 3.22 Conradson Carbon Residue
The data associated with the carbon residue ofthe biodiesel blends are reported in Tables 1 and 3. The linear regession zmlysis of the carbon residue for the certification diesel fuel and for the locally purchased low sulfur diesel Fidel was not statistically si@ticant. Simpie mathematical interpolation is not an adequate method to determine the Conradson carbon residue of a biodiesel blend (see Fi_gure 2 1).
3.5.23 Ash Content
The data associated with the ash content of the biodiesel blends are reported in Tables 1 and 3. No differences were noted, regardless of the blend. As such, mathematical interpolation is a mute e%ort.
4. CONCLUSIONS OF.THIS PROJECT.. . .
b. -
Industry (OEMs) and the Environmental Protection Agency requested information concerning the physical and chemical properties ofblends that ranged Corn 100 percent biodiesel to 100 percent low sulfur diesel fireL It was their perception that this knowledge would facilitate the commercialization of biodiesel as a fuel for modem diesel en_gines.
In this project, various blends of biodiesel and low sulfur diesel fuel were systematically characterized, so that correlations between the fuel properties could be derived. As was noted in the literature, the chemical properties of biodieseVdiese1 fuel blends do not always change in a linear fashion (Van Gerpen, 1994). -. ,.. ‘.._.’
E.xperimental blends of 0,20, 30, 50, 70 and 100 percent b&die&l were prepared by the researchers. Ceitification diesel meI arid locally purchased diesel fuel (MFA Oil Company) were blended with the biodiesel on a volume basis. Fuel related variables, as &tIined by the fuels division of the EPA, were analyzed by the University of Missouri Researchers and by Cleveland Technical Center. A total of seven chemical characteristics were analyzed by MU researcher Anand ChelIappa for the biodiesel/certifkation diesel Fidel bIends. The balance of the analysis for the biodieseYcertification diesel fire1 blends were completed by Cleveland Technical Center.
10
Cleveland Technical Center performed all the analysis for the biodieseflocally purchased diesel he1 blends.
Several variables were found to vary in a linear fashion. The linear regression coefficients as noted in Tables 2 and 4 were quite high. However, one must interpret the linear regression coefficients (‘X2 “) with care. The ‘%I?’ for this study was small Only two different diesel fuels were used (certification LSD +2 and a locally purchased LSD g2). Further, only number two diesel fuel was tested. In short, additional analysis will be needed with a variety of diesel fuels (each with slightly different chemical properties) that are commonly available to the user. This would allow the researchers to develop a model that would accurately predict the chemical property in question.
Based on the aforementioned limitations of this study, several chemical characteristics of the bioclieseYdiese1 fuel blend will require additional data analysis due to the curvilinear tendencies noted in the data. The additional data will facilitate the statistical analysis of the data and eventually allow researchers to develop a curvilinear model that can be used to accurately predict these chemical characteristics for a given blend. Chemical characteristics that may require additional data to fhcilitate the development of a prediction mode1 include: distillation, cold filter plug-tig point, flashpoint, cecane number, olefinic content, and carbon residue.
It appears that simple mathematical interpolation (rather than linear regression) can be used to estimate some of the chemical characteristics for a given biodieseVdiese1 fuel blend. If the researcher lmows the chemical characteristic in question of the “parent” or ‘“olendstock” fuels, the following chemical characteristics can be mathematically interpolated: density, specsc gravity, MI gravity, viscosity, heat of combustion, pa&ins content, carbon, hydrogen, oxygen, neutralization number, sulfur, cloud point, pour point, and iodine value.
11
Table 5. Summary of properties that were anaiyzed for each respective blend.
Phvsical Prouerties ASTM No.
1.
2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.
Dkxiition Initial BP, degrees C TlO, degrees C T50, degrees C T90, degrees C End BP, degrees C Density Specific Gravity API Gravity Cloud Point Pour Point Cold Filter Plug Point Flash Point corrosion Viscosity @ 40C Heat of Combustion Cetane Number
Chemical Composition 1. Aromatics Content (volume percent) 2. Olefinic content (volume percent) 3. Pa&ins content (volume percent) 4. Carbon and Hydrogen 5. Hydrogen 6. Oxygen(weight percent) 7. Sulfbr (iu % wt levels) 8. Neutralization Number 9. Iodine Value of Drying oils and Fatty Acids 10. Carbon residue 11. Ash
D 86
D 1298 D 1298 D 257 D 2500 D 97 D 4539 D 93 D 130 D 445 D 3338 D 613
D 13i9 D1319 D 1319 D 5291 D 5291 D 5291 D 4294 D 664 D 2075 D 189 D 482
Note. Physical properties 3 and 6 are used only for petroleum diesel fuels.
12 c
‘I’nble 1. CllcniicnVl’llysic:11 Properties of Reference l)iesel/l~iotlicscl IWmls
80.5 99.6 1.2 I 18.3 1 80.05 76.4 I 12.29 12.05 9.29 12.12 0.02 0.014 0.39 0.62 81.3 133.9 0.02 0.01 0.02 co.0 I 1 I. Ash D482
Squ;lretl, Conshnt, Stnntl:lrtl Error of Constant, S Coefficients, St;lntlwtl Error of Coefficients for Reference ~iesel/l~iorliescl Blends.
Variables R Squared Conslant Standard Error of Constant
X Coefficients Standard Error of Coefficienls
I. Initial Boiling Point ‘. TIO (Distillation) /. T50 (Distillation) I. T90 (Distillation) ;. End Boiling Point 1. Density ‘. Specific Gravity :. API Gravity ). Cloud Point IO. Pow Point I I. Cold Filter Plug Poiul 12. Flash Point 13. Viscosity @ 40 “C 14. Heat of Combustion I 5. Cetnne Number IO. Aromatics Conlenl 17. Oletinic Content 18. Parallins C0nlc11l 19. Carbon !O. Hydrogen ! I. Oxygen !2. Sulfur !3. Neutralization Number .!4. Iodine Value !5. Carbon Residue
Dl298 0.8473 0.855 0.8581 0.8665 0.8708 0.8855 ‘. S eci IC Gravity
i I)287 35.5 34 33.4 32 31 28.3
I. A I Gravity D2500 IX 20 I8 26 ;. Cloud Point “C D97 ‘to -15 5 IO 1; 1. Pour Point “C
%:‘” 6 4 14 f:
1, Cold Filler Plug Point 130 2
;:5 I60 E5 310 !. Flash Point Dl30 IA IA I. Corrosioil D445 2.55 2.77 it7 3. I8 :ps5 z3 10. Viscosity @ 40 “C D3338 19682 I9156 I8957 18466 I7999 I7272 1 I. Heat of Coiiibuslioii D613 43.5 47.3 48.9 48.3 49.7 51.1 1 2. Cetarre Number
IJllemical Coin ositioii 6
Dl319 25.9 37.8 44.1 61.6 75.8 99.4 I. Aronlalics onteut (volume pelcenl) Dl319 4.2 3.1 3.2 2.2 ) Olefmic Coiiteilt (vohie percenl) Dl319 69.9 59. I 52.7 36.2 z;’ t ii ParaUins Coiileiil (volume perceiil) D529l 86.21 84.04 83.69 81.24 80.35 76.48 I. Carboll (weight erceiil)
P D529 I 13.4 12.91 12.66 12.45 12.41 Il.78
i. t lydrogeil (weig it percent) D529l 0.15 * 3.41 2.69 5.74 8.25 12.83 4. Oxygeli (weight percent) D4294 0.036
3’ 0.03 I 0.025 0.016 0.012
I. Sullilr (iri % wei 1111 levels) A
II664 0.01 0. I I 0.2 03 0.48 3. Nerltralizatioll umber D2075 20.5 43.9 53.6 7'). I 102.9 1~10.9 ‘). Iodine Value Dl89 <O.Ol co.01 co.0 I W.01 <O.Ol co.0 I IO. Cal boil Residue D482 co.0 I <O.Ol <O.Ol <O.Ol <O.Ol co.0 I I I. ASII
IlC 4.
cl~tnretl, Constant, Stnntlnrtl Et-tw of Constant, S Coefficients, Stantl:wtl Error of Coefficients for the LociIlly Pli~cllilsetl Low Sulfur
,;cl FoeVIIiodiesel INends.
I iables R Squared Corlslalll Slantlard Error of Conslalll
X Coefficienls Standard Error of Coeffkienls
lnitinl Boiling Point TIO (Distillntiou) T50 (Distillation) T90 (Distillation) htl Boilirlg Poillt Density Specific Gravity API Gravity Cloud Poht
Pow Poinl Cold Filter Plug Point Flash Point Viscosity @ 40 “C Heat of Combustion Cetane Nunlber Aromatics Content Olefinic Content Paraffins Content Carbon f-lydrogen Oxygen Sulfur Neutralization Number loditie Value
/j i ’ i / .i’i 1 m 1 j : i :-,;,‘- -- ,. -. ..+ --. ----.. L ----1.
.’ I
-- I
CY
0 co
1 / ;
/ : I
\ ,
.-
--
I . !
I
I
0 0.l
i-1. . . 1 .
0
I--- (00 ; I” !Eb F; L- .
> I- I L- co 4
I-
i
Y I .
.‘L..
/ ~
0 0 \
0 co
3 0 .a>
-E .- a
i= al
s* iii
CL
0 N
0
~--- - -- I
0 .- m .i - . .-- - -4
n > I- O
-!-
a, - _-_- -I-.---.--
j
I
I
I ! ..I.-----
- I 1 1 .~. I- O
c-9 N 7 0 d d
m d
.
I L
I
I ! :
: ;
,
i ..-.- 1 :
L i -- --
. I
1 ; i 1 I :
! .~- .I -- :.--i.. . !
0000 d-c\1000 \ \ ?.
.- ..--
Appendix E
Data Forwarded by Clevehnd Technical Center
17
.
CLEVELAND TECHNICAL CENTER A DIVISION OF CONAM INSPECTION INC. 935 Sunshine Road Kansas City, KS 661151122 (913) 281-9881 (800) 332-8055 FAX: (9 13) 28 l-9885
January 05, 1996
Dr. Leon Shumacher University of Missouri Department of Agriculture Engineering 235 Agriculture Engineering Building Columbia, MO 65211
RI!: Purchase Order Number W 1693 8
Enclosed are the revised test results for the biodiesel fuel samples and used oil samples that your Department submitted to us for testing.
We have corrected the Cloud Point test method &om ASTM D-2300 to ASTM D-2500. Also, the distillation temperatures for samples 8-12 (Batch 2) have been converted from degree Fahrenheit to degree Centigrade.
If you have any questions after reviewing your test results please do not hesitate contacting me at (9 13) 28 l-988 1. Thank you for considering Cleveland Technicai Center for this project.
Respectfilly,
Laboratory Supervisor
Enclosures:
CLEVELAND, OH - KANSAS CITY, KS. PHOENIX. Ai! - LINDEN. NJ. GOLDEN VALLEY, MN - SPOKANE, WA - ATLANTA. GA l PORTLAND. OR Expert Lubrication and Fuel Analysis Services
CLEVELAND TECHNiCAL CENTER A DIVLSION OF CONAM /NSPECTlON INC. 9% Sunshine Road
CLEVELAND, OH. KANSAS CITY, KS’. PHOENIX, AZ l LINDEN. NJ. GOLDEN VALLEY. MN. SPOKANE, WA l ATLANTA, GA l PORTLAND, OR . Expert Lubrication and Fuel Analysis Services
CLEVELAND TECHNICAL CENTER A DIVISION OF CONAM INSPECTiON INC. 92.5 Sunshine Road Kansas CI’~. KS 66115-1 m (913) 281-9881 (800) 332-8055 FAX: (913) 28 I -9885
BATCH #l SAMPLE 3 (CAN#302-303) 30% BIODIESEL
PHYSICAL PROPERTIES
DISTILLATION IBP 10 % 50 % 90 0,/o EP
COLD FILTER PLUG POINT FLASH POINT CORROSION HEAT of COMBUSTION CETANE NUMBER
CLEVELAND, OH l KANSAS CITY, KS - PHOENIX. AZ l LINDEN, NJ. GOLDEN VALLEY, MN * SPOKANE. WA l ATLANTA. GA. PORTLAND. OR - : _ _ : Expert Lubrication and Fuel Analysis Services
CLEVELAND TECHNICAL CENTER A DIVISION OF CONAM INSPECTION INC. 936 Sunshine Road Kansas City, KS 66115-1 IZ (913) 281-9861 (800) 332-8055 FAX: (913) 281-9885
CLEVELAND, OH l KANSAS CITY, KS. PHOENIX. AZ. LINDEN, NJ. GOLDEN VALLEY, MN. SPOKANE, WA. ATLANTA. GA. PORTLAND, OR
. , _; Expert Lubrication and Fuel Analysis Services
CLEVELAND TECHNICAL CENTER A DIVISION OF CONAM INSPECTfON INC. 935 Sunshine Road Kansas City, KS 66115-1122 (913) 281-9881 (800) 3324055 FAX: (913) 281-9885
CLEVE,+ND. OH * KANSAS CITY. KS. PHOENIX. Ai! l LINDEN. NJ l GOLDEN VALLEY. MN l SPOKANE, WA l ATLANTA, GA. PORTLAND. OR 2 L
Expert Lubrication and Fuel Analysis Services
CLEVELAND TECHNiCAL CENTER A lYlVlSiON OF CONAM INSPECTION INC. 935 Sunshine Road Kansas City, KS 661151iz.z (913) 281-9881 (800) 33243055 FAX: (913) 291-9885
DENSITY @ 15 DEG C. ASTM D-1298 SPECIFIC GRAVITY AS”l-bf D-1298 CLOUD POrNT ASTM D-2500 POUR POINT ASTM D-97 COLD FILTER PLUG POINT ASTM D-4539 FLASH POINT ASTM D-93 CORROSION ASTM D-130 VISCOSITY @ 40 DEG C. ASTM D-445 HEAT of COMBUSTION ASTM D-333 8 CETANE NUMBER ASTM D-613
CLEVWNO. OH l KiUQM CTTy, KS - PHOENIX, AZ. LINDEN, NJ l QCLOEN VALLEY, YN l S?OKANE. WA l ATLANTA. CIA l PORTLAMI, Of?
Expert Lubricatbn and Fuel Analysis Services -
PHYSICAL PROPERTIES
IXSIKLATION
10 % SO% 90 % EP
DENSITY@ 15 DEG C. SPECIFTC GRAvxTY cmuD POINT POUR POINT COLD FILTER PLUG POINT FLASH POINT CORROSION VTSCOSI’IY @ 40 DEG C. HEAT of C0MEWSTKIN CETANE NUMBER
CLEVELANO. OH - KWSAS CITY, KS - WCENIX, AZ . -INDEN, NJ l QOLDEN VALLEY. h!N l PCKANE. WA. ATMA, GA- PORTLMJD, OR Expert Lubrication and Fuel Analysle Swvlces
Appendix F
Cleveland Technical Center Quality Assurance Report
-- F - CLEVELAND TECHNICAL CENTER
A DMSION OF CONAM 1NSPECTlON /NC , 935 Sunshine Road
Department of Agricultural Engineering 2 3 5 Agriculture Eqjneering Buildi%
Wnbia, MO 6 52 i 1
Reference Purchase Order No. WI6938
QUALrrY ASSURANCE REPORT
Biodiesel Fuel Project
Cleveland Technical Center 935 Sunshine Road
Kansas City, KS 66 115
.
CLEVELAND, OH * KANSAS CITY, KS. PHOENIX, AZ l LINDEN. NJ l GOLDEN VALLEY. MN l SPOKANE. WA l ATLANTA, GA l PORTLAND, OR Expert Lubrication and Fuel Analysis Services
CLEVELAND TECHNICAL CENTER A DlVlSlON OF CONAM 1NSPECTlON INC. 935 Sunshine Road Kansas City, KS 661151122 (913) 281-9881 (800) 332~8GS5 FAX: (913) 281-9885
ASTM D-86 Distillation
Sample No. 1 Sample No. 1 Origjrtd Value, Deg C. Replicate Value. De:? C.
Sample No. 7 Sample No. 7 Ori.ginal Value, Dee C. Replicate Value. Deg C.
IBP 311 IBP 305 10% 329 10% 327 50% 334 50% 334 90% 351 90% 335 EP 356 EP 344
ASTM D-93 Flash Point
p-Xylene Standard True Value = 81 Deg C Actual = 60 Dleg C
ASTM Round Robin LU-9509 Robust Mean = 184 Deg C Actual = 182 Deg C
ASTM D-130 Copper Strip Corrosion
Laboratory Control Sample - Mobil DTE 797 True Value = IA Actual = 1A (l.b/19/95) True Value = IA Actual = IA (1.0/25/95)
. CLEVELAND, OH. KANSAS CITY, KS l PHOENIX. AZ l LINDEN, NJ l GOLDEN VALLEY. MN l SPOKANE. WA l ATLANTA, GA. PORTLAND, OR _ . Expert Lubrication and Fuel Analysis Services
--
- CLEVELAND TECHNlCAL CENTER - A DIVISION OF CONAM INSPECTlON /NC
935 Sunshine Road -- Kansas City, KS 651151122
(913) 281-988~ (800) 332-8055 FAX: (913) 287-9885
ASTM D-3338 Heat of Combustion
Standard - Benzoic Acid True Value = 11373 BTU/lb Actual = 11433 13TU/lb
CLEVELAND, OH l KANSAS CITY, KS l PHOENIX. AZ. LINDEN, NJ l GOLDEN VALLEY. MN l SPOKANE, WA l ATLANTA. GA l PORTLAND. OR - , -. -, - ..:: .>_ ~.. Expert Lubrication and Fuel Analysis Services
--
- CLEVELAND TECHNICAL CENTER - A DIVISION OF CONAM INSPECTION INC.
936 Sunshine Road -- Kansas City, KS 661151122
(913) 281-9881 (800) 3324055 FAX: (913) 281-9885
Sample No. 9 Origird Value
Sample No. 9 Replicate Value
48.9 46.4
ASTM D-613 Cetane Number
Sample No. 10 Original Value
48.3
Sample No. 10 Replicate Value
48.3
Sample No. 11 Original Value
49.7
Sample No. 11 Replicate Value
47.2
ASTM D-1319 FIA
Sample No. 10 Sample No. 10 Original Value Replicate Value
Standard = 0.000 Wt % Actual = 0.005 ‘% Standard = 0.052 Wt % Actual = 0.055 ‘%
Sample No. 1 Sample No. 1 Ori,ginal Value Replicate Value
0.014 Wt % 0.011 Wt %
CLEVELAIJD, OH l KANSAS CITY, KS - PHOENIX, AZ. LINDEN, NJ l GOLDEN VALLEY, MN. SPOKANE, WA ’ ATLANTA, GA. PORTLAND, OR .,- Expert Lubrication and Fuel Analysis Services.
CLEVELAND TECHNICAL CENTER A D/V/S/ON OF CONAM INSPECT/ON INC. 935 Sunshine Road Kansas City, KS 66115-l 122 (913) 281-9881 (800) 3324055 FAX: (913) 281-9685
Sample No. 7 Sample No. 7 Original Value Replicate Value
0.012 Wt % 0.007 Wt %
ASTM D-482 Ash
Sample No. 2 Original Value
Sample No. 2 Replicate Value
0.02 Wt % 0.03 Wt %
Sample No. 3 0ricr;inal Value
Sample No. 3 Replicate Value
0.01 Wt % < 0.01 Wt %
ASTM D-664 Total Acid Number
Standard = Mobil DTE 797 + Oleic Acid True Value = 0.51 mg KOH /g Actual = 0.49 mg KOH / g
Sample No. 7 Sample No. 7 Original Value Replicate Value
0.48 mg KOH / g 0.48 mg KOH / g
ASTM D-2075 Iodine Value
Sample No. 7 Original Value
Sample No. 7 Replicate Value
140.9 140.8 -- - . . . IL-.
CLEVELAI0. OH l KANSAS CITY, KS l PHOENIX. AZ l LINDEN, NJ l GOLDEN VALLEY. MN l SPOKANE:WA * ATLANTA, GA l PORTLAND. OR
. .; : I ;, : .: ..1 Expert Lubrication and Fuel Analysis Services _. ._
--
; CLEVELAND TECHNICAL CENTER - A DlVlSlON OF CONAM INSPECTION INC.