Methods to Assess Biodiesel Quality

8/13/2019 Methods to Assess Biodiesel Quality.

1/8

ANALYTICAL METHODS USED IN THE PRODUCTIONAND FUEL QUALITY ASSESSMENT OF BIODIESEL

G Knothe

ABSTRACT. Biodiesel, an alternative diesel fuel derived from vegetable oil, animalfats, or ivaste vegetable oils, is obtainedby reacting the oil orfat with an alcohol transesterzfication in the presence of a basic catalyst to give the correspondingmono-alkyl esters. Two major categories ofmethods besides other miscellaneous ones have been reported in the literaturefor assessing biodiesel fuel quality andlor monitoring the transesterZfication reaction by ,vhich biodiesel is produced. Thet,I O major categories comprise chromatographic and spectroscopic methods. This article considers the various methods ineach categor.v, including advantages and draWbacks, and offers suggestions on selection of appropriate methods.Keywords. Biodiesel, Chromatographic methods, Fiber-optic probe, Fuel qualit} Gas chromatography, Gel permeationchromatography, High-performance liquid chromatograph.v, Mass spectrometry, N e a r i / ~ f i a r e d spectroscopy, Nuclearmagnetic resonance spectroscopy, Physical properties, Spectroscopic methods, TransesterZfication, l/iscometry.

B iodiesel has significant potential for use as analternative fuel in compres.sion-ignition dieselengines Knothe et aI., 1997: Dunn et al., 1997 . Itis technically competitive with conventional,petroleum-derived diesel fuel and requires no changes in thefuel distribution infrastructure. While some technicalimprovements regarding cold-flow properties, reduction ofNOx exhaust emissions, and oxidative stabili ty remain, amajor hurdle towards widespread commercialization is thehigh price of biodiesei. For this reason, in the United Statesthe commercialization of biodiesel is targeted towardsregulated fleets and mining and marine markets. thesemarkets. environn1ental and energy security concerns, whichare subject to legislation. can override economic aspects.Vegetable oils, such as soybean oil. rapeseed oil canolaoil , and in countries with more tropical climates, tropicaloils palm oil and coconut oil are the major sources ofbiodiesei. However, in recent years, animal fats andespecially recycled greases and used vegetable oils havefound increasing attention as sources of biodiesel, the latterprimarily as inexpensive feedstocks Mittelbach andTritthart, 1988 . Regardless of the feedstock, transesterification reactions are carried out to produce biodiesel. Early onduring the renewed interest in vegetable oils as alternativediesel fuels, it was observed that the resulting vegetable oilor animal fat esters did not exhibit the operationalproblems, such as engine deposits, coking of injector nozzles,etc., associated with neat oils Bruwer et al., 1980 .The transesterification reaction of triacylglycerols invegetable oils, usually with a monohydric alcohol such as

Article was submitted for review in March 2000: approved forpublication by the Power Machinery Division of S in December 2 XXlParts of this article were presented at the 1999 ASAE Annual Meeting asPaper No 99-6111.Product names are necessary to report factually on avai lable data;however. the USDA neither guarantees nor warrants the standard of theproduct. and the use of the name by the USDA implies no approval of theproduct to the exclusion of others that may also be suitable.The author is Gerhard Knothe. Research Chemist . USDA-ARS.National Center for Agricultural Utilization Research. 1815 N University St..Peoria. IL 61604; phone: 309-681-6112; fax: 3 9 681 634 ; e-mail:[email protected].

methanol. in the presence of a basic catalyst such as NaOHor KOB, yields the mono-alkyl esters of the fatty acids,mainly comprising vegetable oils besides glycerol as sideproduct see figure 1 . Methanol, being the least expensivealcohol. is rt1e most commonly used alcohol fortransesterification and accordingly yields methyl esters suchas methyl soyate and rapeseed methyl ester. Recently,modifications of the transesterification reaction, such asthose based on enzymatic catalysis. have been exploredNelson et aL 1996 .During the transesterif ication process, intermediateglycerols, mono and diacylglycerols, are formed, smallamounts of which can remain in the final biodiesel methylester product. Besides these partial glycerols, unreactedtriacylglycerols, unseparated glycerol, free fatty acids,residual alcohol, and catalyst can contaminate the finalproduct. The contaminants can lead to severe operationalproblems when using biodiesel, such as engine deposits, filterclogging, or fuel deterioration. Therefore, in the UnitedStates an ASTM American Society for Testing andMaterials standard is under development Howell, 1997 . Insome European countries, such as Austria, the CzechRepublic, France, Germany, and Italy, standards have beendeveloped that limit the amount of contaminants in biodieselfuel. In these standards, restrictions are placed on theindividual contaminants by inclusion of items such as freeand total glycerol for limiting glycerol and acylglycerols,flash point for limiting residual alcohol. acid value forlimiting free fatty acids, and ash value for limiting residualcatalyst. A more detailed discussion of the rationale for eachquality parameter in biodiesel fuel standards is given in theliterature Mittelbach, 1994; Mittelbach, 1996 . Anotherpaper Komers et al., 1998 briefly describes some methodsused in the analysis of biodiesel, which include proceduresfor determining contaminants such as water and phosphorusthat will not be dealt with here. The determination ofbiodiesel fuel quality is therefore an issue of great importanceto the successful commercialization of this fuel.Continuously high fuel quality with no operational problemsis a prerequisite for market acceptance of biodiesel.Accordingly, the analysis of biodiesel and the monitoring ofthe transesterif ication reaction have been the subject ofnumerous recent publications.

Vol. 44 21: 193 2Transactions of the ASAE

200I American Society of Agricultural Engineers 193


2/8

Figure I. The transesterification reaction of triacylglycerols withmethanol, yielding methyl esters and glycerol. The catalyst isusually a base, such as sodium or potassium hydroxide.

Generally, the major categories of analytical proceduresfor biodiesel comprise chromatographic and spectroscopicmethods, although papers dealing with other methods,including physical property-based ones. have also appeared.The different categories and procedures will be evaluatedhere. As a rule, primarily research with direct reference tobiodiesel analysis will be considered. Additional literature isavailable through the papers cited in the References.

GENERAL ASPECTSThe ideal analytical method for a product such as biodieselwould be able to reliably and inexpensively quantify allcontaminants even at trace levels with experimental ease ina matter of. at most, seconds or even faster for on-linereaction monitoring. No current analytical method meetsthese extreme demands. Therefore, compromises arenecessary when selecting methods for analyzing biodiesel ormonitoring the transesterification reaction.The categories mentioned above often overlap in organicanalytical chemistry due to the advent of hyphenatedtechniques such as gas chromatography - mass spectrometryGC-MS), gas chromatography - infrared spectrometry C-IR), and liquid chromatography - mass spectrometryLC-MS). Few reports exist in the literature on the use ofhyphenated techniques in biodiesel analysis. The mainreasons are likely the higher equipment costs and the higherinvestment in technical skills of personnel needed to interpretthe data. This is the case despite the fact that hyphenatedtechniques could aid in resolving ambiguities remaining afteranalysis by stand-alone chromatographic methods.It is important to note that, in order to satisfy therequirements of biodiesel standards, the quantitation ofindividual compounds in biodiesel is not necessary but thequantitation of classes compounds is. For example, for thedetermination of total glycerol, it does not matter whichmonoacylglycerol for example, monolein or monostearin)the glycerol stems from. The same observation, of course,holds for di and triacylglycerols. It does not even matter forthe determination of total glycerol which type of acylglycerolmono-, di or tri the glycerol stems from. Thatacylglycerols are quantifiable as classes of compounds byGC is a result of the method.Furthermore, virtually all methods used in the analysis ofbiodiesel are suitable if necessary, with appropriatemodifications) for all biodiesel feedstocks, even if theauthors report their methods on one specific feedstock.

a O HW I_ O HW I O H2

TriacylglycerolVegetable oil

catalyst+ 3CHsOH Methanola3 - _ O Hs +

Methyl ester

HO HIHO HIHO HGlycerol

CHROMATOGRAPHIC METHODSBoth GC and HPLC analyses and combinations thereofhave been reported for biodiesel. Gel permeation chromatography GPC) as an analytical tool for analysis oftransesterification products has also been reported. To date,most chromatographic analyses have been applied to methylesters and not higher esters such as ethyl. iso-propyl. etc. Itis likely that most methods discussed here would have to bemodified to properly analyze the higher esters. For example,when conducting gas chromatographic analyses, changes inthe temperature programs or other paran1eters may benecessary. The original work Freedman et aI., 1986) on GCanalysis reported the investigation of methyl and butyl estersof soybean oil. Apparently, not all individual componentscould be separated there in the analysis of butyl soyate, butclasses of compounds could be analyzed. HPLC analysis wasapplied to some ethyl, iso-propyl. 2-butyl, and iso-butylesters of soybean oil and tallow Foglia and Jones, 1997). an analytical method has been applied to esters higher thanmethyl, it is noted here accordingly.The first report on chromatographic analysis of thetransesterification used thin layer chromatography withflame ionization detection TLC / FlD; Iatroscan instrument)reedman et aL 1984). In another report Cvengros andCvengrosova, 1994), TLC / FlD was used to correlate boundglycerol content to acyl conversion determined by Gc It wasfound in this work that if conversion to methyl esters is >96 , then the amount of bound glycerol is < 0.25 wt.- .Although the TLC / FlD method is easy to learn and usereedman et., 1984), it has been largely abandoned becauseof lower accuracy and material inconsistencies, as well assensitivity to humidhy Freedman et aI., 1984) and therelatively high cost of the instrument Cvengros andCvengrosova, 1994).GAS CHROMATOGRAPHYGas chromatography has to date been the most widelyused method for the analysis of biodiesel due to its generallyhigher accuracy in quantifying minor components. Note,however, that accuracy of GC analyses can be influenced byfactors such qS baseline drift, overlapping signals, etc. It isnot always clear that such factors are compensated for in suchreports on biodiesel analysis. The first report on the use ofcapillary gas chromatography discussed the quantitation ofesters as well as mono-, di-, and triacylglycerols Freedmanet aI., 1986). The samples were treated with N O-bis trimethylsilyl)trifluoracetamide BSTFA) to give the corresponding trimethylsilyl TMS) derivatives of the hydroxy groups.This kind of derivatization has been carried out in subsequentresearch on GC quantitation of biodiesel. Derivatization toTMS derivatives is important because it improves thechromatographic properties of the hydroxylated materials,and in case of coupling to a mass spectrometer, facilitatesinterpretation of their mass spectra. While the originalresearch Freedman et aL 1986) used a short 1.8 m) fusedsilica 100 dimethylpolysiloxane) capillary column, otherresearch typically used fused-silica capillary columns coatedwith a O.l-llm film of 5 -phenyl)-methylpolysiloxane of1 to 15 m length. The analysis ofrapeseed ethyl esters wasalso carried out on a GC instrument equipped an FlD and witha 1 8 m x 4 mm i.d. packed column Cvengrosova et aI.,1997).

194 TRANSAC110NS OF 11iE ASAE


3/8

Most reports on the use of GC for biodiesel analysisemploy t1ame-ionization detectors FID), although the use ofmass spectrometric detectors MSD) would eliminate anyambiguities about the nature of the eluting materials sincemass spectra unique to individual compounds would beobtained. Two papers exist in the literature in which the useof mass spectrometric detection is described Mittelbach.1993: Mittelbach et aL 19961. It can be surmised that theadditional cost of MSDs and mass spectral interpretation)plays a role in deterring the commercial adoption of thisdetection method, although the benefits of mass spectrometry would likely more than compensate for the costs.The majority of GC-related papers discuss thedetermination of a specific contaminant or class ofcontaminants in the methyl esters. The original report onbiodiesel GC analysis Freedman et aI., 1986) quantifiedmona-. di-, and triacylglycerols in methyl soyate on a short100 dimethylpolysiloxane column 1.8 m 0.32 mm i.d.).Similar reports on the quantitation of acylglycerols exist ariani et aL 1991: Plank and Lorbeer, 1992). The maindifferences are in the specifications of the columns - both5 -phenyl1-methylpolysiloxane, with differences in parameters such as column length - and the temperatureprograms, as well as standards.Other papers discuss the individual or combineddetermination of other potential contaminants, such as freeglycerol or methanol. A paper describing the use of a massselective detector deals with the determination of glycerol ittelbach, 1993) and, in an extension thereof, a secondpaper describes the simultaneous quantitation of glycerol andmethanol Mittelbach et aI., 19961. Other authors have alsoreported the determination of glycerol Bondioli et aI.,1992a) or methanol Bondioli et aI., 1992b). Using the samegas chromatographic equipment as in the previousdetermination of glycerol Bondioli et aI., 1992a) with onlya modification of the oven temperature program, methanolcould be determined. Ethanol was used as a standard forresponse factor determination. The t1ash point of biodiesel

from palm oil and methanol content were correlated.Underivatized glycerol was detected with 1,4-butanediol asa standard on a short 2 m glass column 4 mm i.d.) loadedwith Chromosorb 10 1 Bondioli et aI., 1992a), while theother method used derivatization and a 60 m 0.25 mm i.d.column with 0.25-J.l.m film of 5 -phenyl)-methylpolysiloxane and was reported to be more sensitive Mittelbach,1993). The temperature program varied starting lower whendetermining methanol) Mittelbach, 1993: Mittelbach et aI.,1996), otherwise the column was the same.Two papers Mittelbach, 1993: Mittelbach et aI., 1996)discuss the use of mass spectrometry as a detection methodbesides t1ame ionization. In the determination of freeglycerol in biodiesel by GC-MS. selected ion monitoringIM) mode was used to track the ions z116 and 117 ofbis-Q-trimethylsilyl-l,4-butanediol from silylation of the1,4-butanediol standard) and z 147 and 205 oftris-Q-trimethylsilyl-l,2,3-propanetriol from silylation ofglycerol). The detection limit was also improved for r p e ~ e emethyl ester RME) when using MS in SIM mode 10-) compared to the FID detector 10-4 Mittelbach, 1993).In straightforward extension of this work, the simultaneousdetection of methanol and glycerol by MS in SIM mode wasreported Mittelbach et aL 1996). For detection of silylated)methanol trimethylmethoxysilane), peaks at z59 and 89Vol. 44 2): 193-200

were monitored, as were peaks at m 75 and 103 of theadditional silylated) standard ethanol trimethylethoxysilane). Mass spectrometry in SIM mode has the additionaladvantage that interfering signals can be avoided, and thusthe use of shorter columns is possible Mittelbach et aI.,1996).A further extension of the aforementioned research is thesimultaneous detemlination of glycerol as well as mona-,di- and triacylglycerols simultaneously by GC Plank andLorbeer. 1995). Here and in previous work Plank andLorbeer. 1992; Plank and Lorbeer, 1995) 10 m 5 -phenYl)-methylpolysiloxane columns with O.l-J.l.m film 0.25 mm i.d. in Plank and Lorbeer 1992) and 0.32 mm i.d.in Plank and Lorbeer 1995) - were used. Major differenceswere the lower starting temperature of the temperatureprogranl Plank and Lorbeer, 1995) and the addition of astandard L2,4-butanetriol) for the glycerol analysis. In thiswork Plank and Lorbeer, 1992; Plank and Lorbeer. 1995), acool on-eolumn injector was used instead of the morecommon split/splitless injector.Non-glyceridic materials that can be present in biodieselalso have been analyzed by Gc. Thus, the determination ofsterols and sterol esters in biodiesel Plank and Lorbeer,1993) has been reported. The stated reason is that theint1uence of these compounds, which remain in vegetableoils after processing and thus in biodiesel after transesterification because they are soluble in methyl esters), onbiodiesel fuel quality has not been determined Plank andLorbeer, 1993). Detection was carried out with at1ame-ionization detector, as in other GC-related research,although in this case MS detection would appear especiallydesirable. The method for detection of sterols Plank andLorbeer, 1993) is virtually identical to the other GC methodreported by the same authors Plank and Lorbeer, 1992). Theonly differences are the use of sterol standards and a slightmodification of the GC temperature program to spread thesterol peaks under condensation or even overlapping of thepeaks of the other classes of compounds. Derivatization wascarried out again with BSTFA with 1 trimethylchlorosilane), and the column again was a fused-si lica capillarycolumn coated with a O.l-J.l.m film of 5 -phenyl)-methylpolysiloxane. The total concentration of sterols in rapeseedmethyl ester was reported to be 0.339--0.500 , and sterolesters were 0.588--0.722 . In another paper on analysis ofsterol content in rapeseed methyl ester Plank and Lorbeer,1994a), the same authors reported a sterol content of0.70--0.81 . Other authors Mariani et al., 1991) also pointedout the presence of sterols and sterol esters in biodiesel, butno analytical data were given.HIGH-PERFORMANCE LIQUID CHROMATOGRAPHYA general advantage of HPLC compared to GC is thattime- and reagent-eonsuming derivatizations are notnecessary, which reduces analyses times. Nevertheless, thereare fewer reports of HPLC applied to biodiesel than GCanalysis. The first report on the use of HPLC Trathnigg andMittelbach, 1990) described the use of an isocratic solventsystem chloroform with an ethanol content of 0.6 ) on acyana-modified silica column coupled to two gel permeationchromatography GPC) columns with density detection. Thissystem allowed for the detection of mona- di-, andtriacylglycerols as well as methyl esters as classes of

195


4/8

compounds. The system was useful for quantitating variousdegrees of conversion of the transesterification reaction.HPLC with pulsed amperometric detection (the detectionlimit is usually 10 to 100 times lower than for amperometricdetection, and the detection limit is 1 lg/g) was used todetermine the amount of free glycerol in vegetable oil esters(Lozano et aL 1996). The major advantage of this detectionmethod was its high sensitivity. The simultaneous detectionof residual alcohol is also possible with this technique.Reaction mixtures obtained from lipase-catalyzedtransesterification were analyzed by HPLC using anevaporative light scattering detector (ELSD) (Foglia andJones, 1997). This method is able to quantitate product esters,free fatty acids, and the various forms of acylglycerols. Asolvent system consisting of hexane and methyl tert butylether (each with 0.4 acetic acid) with a gradient elutionprofile was used. It can be applied to esters higher thanmethyl, as discussed above.In an extensive study (Holcapek et aL 1999),reversed-phase HPLC was used with different detectionmethods: UV detection at 205 nm, evaporative lightscattering detection, and atmospheric pressure chemicalionization mass spectrometry (APCI-MS) in positive-ionmode. Two gradient solvent systems were used: oneconsisting of mixing methanol (A) with 5:4 2-propanol /hexane B from 100 A to 50:50 A:B - a non-aqueousreversed phase (NARP) solvent system - and the otherconsisting of mixing water (A), acetonitrile (B), and 5:42-propanol / hexane (C) in two linear gradient steps (30:70A:B at 0 min. 100 B in 10 min, 50:50 B:C in 20 min, andlast isocratic 50:50 B:C for 5 min). The fIrst solvent systemwas developed for rapid quantitation of the transesterifIcation of rapeseed oil with methanol by comparing the peakareas of methyl esters and triglycerols. The contents ofindividual acids (using normalized peak areas) were subjectto error, and the results differed for the various detectionmethods. The sensitivity and linearity of each detectionmethod varied with the individual triacylglycerols. PCI-MS and ELSD had decreased sensitivity with increasingnumber of double bonds in the fatty acid methyl esters, whileUV will not quantify the saturates. APCI-MS was stated tobe the most suitable detection method for the analysis ofrapeseed oil and biodiesel.GEL PERME TION HROM TOGR PHYOne report exists which describes the use of GPC (whichis very similar to HPLC in instrumentation except for thenature of the column and the underlying separation principle,namely molecular weight of the analytes for GPC) for theanalysis of transesterification products (Damoko et aI2000). Using a refractive index detector and tetrahydrofuranas mobile phase, mono-, di-, and triacylglycerols as well asthe methyl esters and glycerol could be analyzed. Themethod was tailored for palm oil, and standards were selectedaccordingly. Reproducibility was good, with the standarddeviation at different rates of conversion being 0.27-3.87 .LIQUID HROM TOGR PHY WITH GAS HROM TOGR PHYThe combination of liquid chromatography (LC) with gaschromatography (GC) has also been reported. The purpose ofthe combination of the two separation methods is to reduce

196

the complexity of the gas chromatograms and to obtain morereliable peak assignments (Lechner et aI 1997).In a study on analytical methods for detemlining biodieselin mixtures with conventional diesel fuel, silica cartridge(Sep--Pak) chromatography with hexane / diethyl ether assolvents was employed to separate biodiesel fromconventional diesel fuel (Bondioli et aI 1994).A fully automated LC GC instrument was employed inthe determination of acylglycerols in vegetable oil methylesters (Lechner et aI., 1997). Hydroxy groups wereacetylated, and then the methyl esters (sterols and esterifiedsterols elute with methyl esters) and acylglycerols werepre-separated by LC (variable wavelength detector). Thesolvent system for LC was hexane / methylene chloride /acetonitrile 79.97:20:0.05 and GC (flame-ionizationdetector) was performed on a 10 m (5 -phenyl)-methylpolysiloxane column. One LC GC run required 52 minutes.

LC GC was also applied to the analysis of sterols inbiodiesel derived from rapeseed oil (Plank and Lorbeer,1994a: Plank and Lorbeer, 1994b). On-line LC-GC wasapplied to the analysis of sterols from fIve different types ofmethyl esters (Plank and Lorbeer, 1994b). The vegetable oilmethyl esters were those of rapeseed oil, soybean oil,sunt10wer oil, high-oleic sunt10wer oil, and used frying oil.The sterols were silylated prior to analysis with methyl-N-trimethylsilyltrifluoracetamide (MSTFA). No saponifIcation and off-line pre-separation was required. The methylesters were separated from the sterols by LC with a hexane/ methylene chloride / acetonitrile 79.9:20:0.1 solventsystem. Gas chromatography was again carried out with a12 m (5 -phenyl)-methylpolysiloxane column and FIDdetection. Total concentrations of free sterols were 0.20-0.35weight- for the fIve sanlples, while sterol esters displayeda range of 0.15-0.73 weight- . Soybean oil methyl ester wasat the lower end (0.20 and 0.15 , respectively), whilerapeseed oil methyl ester was the higher end (0.33 and0.73 , respectively). In a comparison of two methods,saponifIcation and isolation of the sterol fraction withsubsequent GC analysis and LC GC analysis of sterols inrapeseed oil methyl ester (Plank and Lorbeer, 1994b), despitethe sophisticated instrumentation required, LC-GC wasrecommended because of additional information, shortanalysis time, and reproducibility. The total sterol content inrapeseed methyl ester was found to be 0.70-0.81 weight- .

SP TROS OPI M THO SSpectroscopic methods also have been reported for theanalysis of biodiesel and/or monitoring of the transestermsation reaction. These methods are 1H nuclear as well as bCmagnetic resonance (NMR) spectroscopy and near-infrared(NIR) spectroscopy.

NUCLEAR MAGNETIC RESON N EThe fIrst report on spectroscopic determination of theyield of a transesterification reaction utilized 1H-NMR(Gelbard et aI., 1995). Figure 2 depicts the IH-NMRspectrum of a progressing transesterification reaction. Theseauthors used the protons of the methylene group adjacent tothe ester moiety in triglycerols and the protons in the alcoholmoiety of the product methyl esters to monitor the yield. A

TRANSACTIONS OF THE ASAE


5/8

I I4.2 4.0 3. 6 3.6

i j I ii i i I I ii i i 5. 5 5.0 4.5 4.0 3.5 i I i i I I i i i 3.0 2.5 2.0 1.5 \. 0

Figure 2. 1H-NMR spectrum of a progressing transesterification reaction.The signals at 4.1-4.3 ppm are caused by the protons attached to theglycerol moiety of mono-, di-, or triacylglycerols. The strong singlet at 3.6 ppm indicates methyl ester -COlCH3 formation. The signals

at 2.3 ppm result from the protons on the CHl groups adjacent to the methyl or glyceryl ester moieties-CHlCOlCH3 for methyl esters . These signals can be used for quantitation.

simple equation.is given by the authors their terminology isslightly modified here :C lOOx 2A ] 1

CHwhereC = conversion of triacylglycerol feedstock vegeta-ble oil to the corresponding methyl ester.

A1\ E = integration value of the protons of the methylesters the strong singlet peak . CH = integration value of the methylene protons.The factors 2 and 3 derive from the fact that the methylene

carbon possesses two protons and the alcohol methanol-derived carbon has three attached protons.Turnover and reaction kinetics of the transesterification ofrapeseed oil with methanol were studied by 13C-NMRDimmig et aI., 1999 with benzene-d6 as solvent. Thesignals at approximately 4 5 ppm of the terminal methylgroups unaffected by the transesterification were used asinternal quantitation standard. The methyl signal of theproduct methyl esters registered at around 5 ppm, and theglyceridic carbons of the mona-, di-, and triacylglycerolsregistered at 62-71 ppm. Analysis of the latter peak rangeVol 44 2 : 193-200

allowed the determination of transesterification kinetics,which showed that the formation of partial acylglycerolsfrom the trig1ycerols is the slower, rate-determining step.NEAR-INFRARED SPECTROSCOPYMore recently, NIR spectroscopy has been used to monitorthe transesterification reaction Knothe, 1999 . The basis forquantitation of the turnover from triacylglycerol feedstock tomethyl ester product are differences in the NlR spectra ofthese classes of compounds. At 6005 cm- I and at4425-4430 em-I, the methyl esters display peaks, whiletriacylglycerols only exhibit shoulders see fig. 3 . Ethylesters could be distinguished in a similar fashion. Knothe,1999 . Using quantitation software, it is possible to developa method using partial least squares regression forquantifying the turnover of triacylglycerols to methyl esters.The absorption at 6005 cm-1gave better results than the oneat 4425 em-I. Note that the mid-range IR spectra oftriacylglycero1s and methyl esters of fatty acids are almostidentical and offer no possibility for distinguishing. Itappears that ethyl esters, and perhaps even higher ester, maybe distinguished similarly by NlR from triacylglycerols, butno results have been reported yet.

197


6/8

Figure 3. NIR spectra of soybean oil BO), methyl soyate SMEj, andSME containing significant amounts of methanol. The inscribed wavenumbers highlight the possibilities for distinguishing the spectra

and thus quantifying the components.NIR spectra were obtained with the aid of a fiber-opticprobe coupled to the spectrometer, which renders theiracquisition particularly easy and time-efficient.Contaminants of biodiesel cannot be fully quantified byNIR at the low levels called for in biodiesel standards. Theaccuracy of the NIR method in distinguishing triacylglycerols and methyl esters is in the range of 1-1.5 , although inmost cases better results are achieved. To circumvent thisdifficulty, an inductive method can be applied. The inductivemethod consists of verifying - by, say, gas chromatography

- that a biodiesel sample meets prescribed biodieselstandards. The NIR spectrum of this sample would then berecorded. The NIR spectrum of the feedstock would also berecorded, as well as the spectra of intermediate samples atconversions of, say, 25 , 50 , and 75 . With these samples.a quantitative NIR evaluation method could be established.When another transesterification reaction is subsequentlyconducted. the NIR spectrum would indicate that the reactionwithin the prescribed parameters such as time andtemperature) has attained conversion to a product that withinexperimental error of NIR) conforms to standards. It can besafely assumed that this result is correct, even if not allpotential contaminants have been fully analyzed. Only if asignificant deviation is indicated by NIR beyondexperimental error) would a detailed investigation by a morecomplex method such as GC be necessary. The NIRprocedure is considerably less labor-intensive, faster. andeasier to perform than GcWhile the first NIR paper used a model system to describemonitoring of transesterification and for developingquantitation methods. a second paper applied the method toa transesterification reaction in progress on a 6 L scale. Here.spectroscopic results were obtained not only by NIR but alsoby IH NMR and NIR Knothe, 2000). The results of both

2)

OTH R M THO SVISCOMETRYThe viscosity difference between component triacylglycerols of vegetable oils and their corresponding methyl estersresulting from transesterification is approximately one orderof magnitude Knothe et aI., 1997; Dunn et aL 1997). Forexample. the viscosity of soybean oil is 32.6 js 38C)and that ofrnetl1yl soyate is 4.41 js 0C) Knothe et al 1997, and references therein). The high viscosity of thevegetable oils was the cause of severe operational problems.such as engine deposits Bruwer et aI.. 1980; Knothe et aI..1997; Dunn et aI.. 1997). This is a major reason why neatvegetable oils largely have been abandoned as alternativediesel fuels in favor of mono-alkyl esters such as methylesters.The viscosity difference forms the basis of an analyticalmethod. viscometry, applied to determining the conversionof vegetable oil to methyl ester De Filippis et aI.. 1995).Viscosities determined at 20C and 37.8C were in goodagreement with GC analyses conducted for verification

spectroscopic methods. which can be correlated by simpleequations, were in good agreement. Two NMR approacheswere used. one being the use of the methyl ester protons peakat 3.6 ppm in figure 2) and the protons on the carbons next tothe glyceryl moiety , -CH2; peaks at 2.3 ppm in figure 2)Knothe, 2000), for which the equation for determiningconversion in ) is:xIM = lOOx_ I I M x T

where refers to integration values. and the subscripts MEand T G refer to methyl esters and triacylglycerol. Thesecond approach was the use of the methyl ester protons andthe protons of the glyceryl moiety peaks at 4.1-4.3 ppm infigure 2) in the triacylglycerols Knotlle. 2000).In connection with the spectroscopic assessment ofbiodiesel fuel quality and monitoring of the transesterification reaction. a paper that discusses determining the amountof biodiesel in lubricating oil Sadeghi-Jorabchi et al.. 1994)is noteworthy. The investigated problem is significantbecause biodiesel can cause dilution of the lubricant. whichcan ultimately result in engine failure. The dilution wasattributed to the higher boiling range of biodieseladeghi-Jorabchi et aI.. 1994; Siekmann et aI.. 1982)compared to conventional diesel fuel. whose more volatilecomponents have less chance to dilute the lubricant. Theseauthors used mid-range IR spectroscopy with a fiber-opticprobe to determine the anlOunt of biodiesel in lubricating oil.The range used was 1820-1680 em-I, which is typical forcarbonyl absorption and which is not observed inconventional diesel fuel nor in the lubricating oil note thatthis- range is not suitable for distinguishing vegetable oils andtheir methyl esters because they have nearly identicalcarbonyl absorptions in the mid-IR range). Previous to thiswork, other authors had used IR spectroscopy without theaid of a fiber-optic probe) in the range 1850-1700 cm- l toanalyze biodiesel in lubricating oil Siekmann et aI.. 1982).The carbonyl absorption at 1750 cm-I was not disturbed bythe absorption of oxidation products at 1710 em-I.

4000500000500cm 1

6000500000

TRANSAC11 NS OF THE ASAE


7/8

purposes. The viscometric method, especially resultsobtained at 20e. is reported to be suitable forprocess-control purposes due to its rapidity (De Filippis etaI., 1995). Similar results were obtained from densitymeasurements (De Filippis et aL 1995).TITRATION FOR DETERMINING FREE FATTY ACIDSTitration methods for determining the neutralizationnumber (NN) of biodiesel were described (Komers et aL1997). Two methods for determining strong acids and freefatty acids in one measurement were developed. One method,of particular interest, used potentiometry, while the otherused two acid-base indicators (neutral red, phenolphthalein).The potentiometric method is more reliable, and even withthe use of two indicators, the NN values derived from thetitration method are 10-20 rel.- greater than the real acidityof the sample.ENZYMATIC METHODSAn enzymatic method for analyzing glycerol in biodieselwas described to test for completeness of the transesterification reaction (Bailer and de Hueber, 1991). Solid-phaseextraction of the reaction mixture with subsequent enzymaticanalysis was applied. This method was originally intended asa simple method for glycerol determination, but reproducibility and complexity concerns exist (Bondioli et aI., 1992a;Lozano et aI. 1996).

IODIESEL LENDSBlends of biodiesel with conventional diesel fuelrepresent a common utilization of biodieseI. In the UnitedStates, B20 (a blend of 20 biodiesel with 80conventional diesel fuel) is recognized as an alternativediesel fuel under criteria of the Energy Policy Act (EPACT).In France, biodiesel is utilized as a lower-level blend withconventional diesel fuel. In this connection, a problem thathas found little attention to date in the literature isdetermining or verifying the blend level of biodiesel inconventional diesel fuel in such blends. NlR spectroscopyusing the same peaks for quantitation as those for monitoringtransesterification and fuel quality appears to be suitable forsuch purposes (Knothe, manuscript submitted for publication). Mid-IR spectroscopy as used for the determination ofbiodiesel in lubricating oil (Sadeghi-Jorabchi et aI., 1994)should also be suitable. The use of the NIR range, however,would permit using a spectrometer without any changes ininstrument settings for reaction and fuel quality monitoringas well as for determining blend levels. Generally, it appearsthat spectroscopic methods may be more suitable to addressthe problem of determining / verifying blend levels.Especially GC would appear to yield very complexchromatograms due to the numerous components ofconventional diesel fuel. In HPLC, as done previously, theclasses of compounds may elute and be analyzable, thusreducing complexity. Methods based on physical propertiesmay also be suitable for determining biodiesel blend levels.

SUMM RY ND OUTLOOKSeveral analytical methods have been investigated for fuelquality assessment and production monitoring of biodiesel.The most intensively studied method is GC, while HPLC,NMR, and NlR have also been studied. GC is also the methodused for verification that biodiesel meets prescribedstandards due to its ability to detect low-level contaminants,although improvements to this method are possible. Physicalproperties-based methods have been explored less, and itappears that this may be an area for further study. However,

no method can simultaneously satisfy all criteria ofsimultaneously determining all trace contaminants withminimal investment of time, cost, and labor. The followingapproach therefore appears to be a reasonable compromise.A fast and easy- ta-use method that may be adaptable toproduction monitoring, such as NIR (or viscometry), can beused for routine analyses. For example, if measurements byNIR (or viscometry) at several turnover ratios indicate thatthe transesterification reaction is progressing as desired andthat the NIR spectrum (viscosity) of the biodiesel productagrees with that of one known to meet biodiesel standards,then further, more complex, analyses would be unnecessary.Only if NIR (or viscometry) analyses indicate that there is apotential problem with the product would more complex andtime-consumirig analyses, for example by GC, be warrantedto determine the exact cause of the problem.Determining the blend levels of biodiesel in blends withconventional diesel fuel is also an area that can be expanded.

REFERENCESBailer. J.. and K. de Hueber. 1991. Determination of saponifiable

glycerol in bio-diesel. Fresenius Anal Chem. 340(3): 186.Bondioli. P.. Mariani. A. Lanzani. and E. Fedeli. 1992a.

Vegetable oil derivatives as diesel fuel substitutes: Analyticalaspects. Note 2: Determination of free glycerol. Riv. Iral.Sosran e Gra:, se 69(1): 7-9.

Bondioli. P. C. Mariani . E. Fedel i. A. M. Gomez. and S. Veronese.1992b. Vegetable oil derivatives as diesel fuel substitutes:Analyticataspects. Note 3: Determination of methanol. Rh: Iral.Sosran e Grasse 69(9): 467-469.

BondiolL P. A. Lanzani . E. Fedel i. M. Sala. and S. Veronese. 1994.Vegetable oil derivatives as diesel fuel substitutes: Analyticalaspects. Note 4: Determination of biodiesel and diesel fuel inmixture. Ri\ . Ital Sostan e Grasse 71(6): 287-289.

Brower, J. J .. B. van D. Boshoff. F. 1. Hugo. J. Fuls, Hawkins. N. van del' Walt. A. Engelbrecht. and L M. du Plessis. 1980.The utilization of s u n o w ~ r seed oil as a renewable fuel fordiesel engines. National Energy Symposium the ASAE 29September - 1 October. Kansas City. Missouri.

Cvengros. J . and Z. Cvengrosovu. 1994. Quali ty control ofrapeseed oil methyl esters by determination of acyl conversion. Am. Oil Chem Soc 71(12): 1349-1352.

Cvengrosovu, Z . J. Cvengros, and M. Bronec. 1997. Rapeseed oilethyl esters as alternative fuels and their quality control. Petrol.Coal 39(3): 36-40.

Damoko, D., M. Cheryan, and E. G. Perkins. 2000. Analysis ofvegetable oil transesterification products by gel permeationchromatography. Liq Chromo Rei Technol. 23 15):2327-2335.

Vol. 44(2): 193-200 199


8/8

De Filippis, P. C. Giavarini, M. Scarsella, and M. Sorrentino. 1995.Transesteritlcation processes for vegetable oils: A simple controlmethod of methyl ester content. Am. OilChem. Soc. 72 11 :1399-1404.Dimmig, T. W Radig. C. Knoll, and T. Dittmar. 1999.I3C-NMR-Spektroskopie zur Bestimmung von Umsatz undReaktionskinetik der Umesterung von Triglyceriden zuMethylestem [13C-l\i MR spectroscopic determination of theconversion and reaction kinetics of transesteritlcation oftriacylglycerols to methyl esters]. Chem. Tech. Leipzig) 51 6):326-329.Dunn, R. a G. Knothe. M. Bagby. 1997. Recent advances inthe development of alternative diesel fuel from vegetable oil andanimal fats. Recelll Research Developmellls in Oil Chemisf/} 1:

31 56.Foglia, T. A., and K. C. Jones. 1997. Quantitation of neutral lipidmixtures using high performance liquid chromatography withlight scattering detection. Liq. ChromaOgl: Relat. Tee/mol.20 12): 1829-1838.Freedman, B.. E. H. Pryde, and W F. Kwolek. 1984. Thin layerchromatography / name ionization analysis of transesteritledvegetable oils. Am.Oil Chem. Soc. 61 7): 1215-1220.Freedman, B., W F. Kwolek, and E. H. Pryde. 1986. Quantitationin the analysis of transesteritled soybean oil by capillary gaschromatography. Am. OilChem. Soc. 63 10): 1370-1375.Gelbard, G .. Bres, R. M. Vargas, Vielfaure. and U. Schuchardt. 1995. IH nuclear magnetic resonance determination

of the yield of the transesesteritlcation of rapeseed oil withmethanol. Am. Oil Chem. Soc. 72 10): 1239-1241.HoI: apek, M., P. Jandera, J. Fischer, and B. Prokes. 1999.Analytical monitoring of the production of biodiesel byhigh-performance liquid chromatography with various detectionmethods. ChromaOgl: A. 858 1): 13-31.Howell. S. 1997. Biodiesel fuel standard making progress. STMStandardi:ation News 25 4): 16-19.Knothe, G., R. a. Dunn, and M. a . Bagby. 1997. Biodiesel: Theuse of vegetable oils and their derivatives as alternative dieselfuels. Am. Chem. Soc. Symp. Series 666: 172-208.Knothe, G. 1999. Rapid monitoring of transesteritlcation andassessing biodiesel fuel quality by NIR spectroscopy using atlber-optic probe. Am. Oil Chem. Soc. 76 7): 795-800. . 2000. Monitoring the turnover of a progressing

transesteritlcation reaction by tlber-optic NIR spectroscopy withcorrelation to IH NMR spectroscopy. Am. OilChem. Soc.77 5): 489 493. . Determining the blend level of mixtures of biodiesel withconventional diesel fuel by NIR spectroscopy using a tlber-opticprobe. [Manuscript submitted for publication].Komers, K., F. Skopal, and R. Stloukal. 1997. Determination of theneutralization number for biodiesel fuel production. FeU/Lipid99 2): 52 54.Komers, K., R. Stloukal. J. Machek, F. Skopal, and A. Komersov, ,, P ~ - : ; ~ : : > \ t h P ~ N ~ lHin0is

TRANSACTIONS OF TH ASAE

Documents

Methods to Assess Biodiesel Quality