Effect of minor components in soy biodiesel on cloud point and fi lterability

Size: px

Start display at page:

Download "Effect of minor components in soy biodiesel on cloud point and fi lterability"

Alexander Stanley
5 years ago
Views:

1 Effect of minor components in soy biodiesel on cloud point nd fi lterbility Lis Pflzgrf, Inmok Lee, Jmes Foster, nd George Poppe Biodiesel fuels hve been in commercil use in mny countries for up to 20 yers. Tody, in lmost every country in the world biodiesel production nd utiliztion re growing. High crude oil prices nd the desire to reduce dependence on foreign oil, improve the environment, nd provide opportunities for griculturl products re stimulting biodiesel use in the United Sttes. Biodiesel is renewble fuel derived from nturl fts nd oils, such s soyben oil, tht meets the specifi ctions of ASTM D 6751, Biodiesel Fuel Blend Stock (B100) for Middle Distillte Fuels (1). ASTM D 6751 defi nes biodiesel s fuel comprised of mono-lkyl esters of long-chin ftty cids derived from vegetble oils or niml fts. Under D 6751, biodiesel, or B100, tht meets this specifi ction my be used in blends with petroleum diesel t up to 20% (B20) by volume. The use of B20 nd lower blends mde with D 6751-grde B100 hs been lrgely trouble-free since the dvent of the biodiesel industry in the United Sttes over 10 yr go. However, unexpected filter plugging took plce in vehicles using B20 or lower biodiesel blends in the stte of Minnesot during the winter (2), even though the cold flow properties (cloud point [CP], cold filter-plugging point) of the biodiesel did not indicte filter-plugging potentil (2). Our initil testing of mteril recovered from plugged filters identified prepondernce of sturted monoglycerides (MG) in the orgnic component. In response to the filter problems encountered in Minnesot, cold incubtion ( cold sok ) filter test ws instituted for B100 used in mking biodiesel blends in the stte. This cold sok filter test consists of chilling the smple for specified time nd then llowing it to wrm to room temperture prior to filtering in modified ASTM D 6217 Prticulte Contmintion in Middle Distillte Fuels by Lbortory Filtrtion procedure (3). Although not prt of the officil ASTM D 6751 requirements for B100, this cold sok filter test hs become commonly used in the industry s mesure of filterbility. In ddition, some compnies hve lso implemented other filter tests, such s ASTM D 2068, Filter Plugging Tendency of Distillte Fuel Oils (4). Sttisticlly designed experiments re powerful tools for the chrcteriztion of complex systems. Through proper ppliction of estblished sttisticl design methods, they llow the experimenter to obtin the mximum informtion from the smllest possible number of experimentl runs. The gol of sttisticl design is to develop polynomil model tht dequtely describes the effect of the vribles on result of interest. This model cn be used to provide n estimtion of the effects of components within the rnge tested without hving to test every possible combintion of the components. In ddition, the sttisticl model cn help explin the effects of interctions tht re not esily ccessible by direct observtion only, leding to deeper understnding of fundmentl mechnisms of the observed effects nd thus better process control nd consistent product qulity. Modern commercilly vilble sttisticl design pckges re robust nd provide guidnce in proper experimentl design nd interprettion of results. The effects of vrious components nd processing residuls on the properties of finished soy biodiesel re lrgely unknown. In this lbortory, sttisticlly designed experiment ws developed nd conducted to determine the influence of MG, sterol glucosides (SG), sops, nd wter on the CP nd filterbility fter cold sok tretment. The rnge of concentrtion for ech component ws chosen to reflect possible expected vlues found in commercil soy biodiesel meeting D 6751 specifictions. The influence of these components reported for soy biodiesel my not describe their influence in biodiesel mde from other feedstock owing to differences in composition. MATERIALS AND METHODS Distilled soy methyl ester (SME) biodiesel (Steposol SB-D, Stepn, Northfield, Illinois, USA) ws obtined nd nlyzed to confirm the bsence of SG, sodium/sop, diglycerides (DG), triglycerides, nd free glycerol. Only trce of MG (0.02%, 200 ppm) ws found. The distilled SME ws dried under vcuum to remove ny trces of wter nd spiked with mesured quntities of MG, SG, sops, nd wter (Tble 1). MG were prepred from soy oil. Refined, bleched, deodorized soyben oil (Archer Dniels Midlnd, Dectur, Illinois) ws subjected to glycerolysis with sodium methoxide, followed by wiped-film distilltion to remove methyl esters, specil supplement to inform / 17

2 Biorenewble Resources No. 4 / September 2007 TABLE 1 Combintions of components dded to distilled soy biodiesel vs. cloud points nd filter times obtined Components dded, ppm Responses MG SG Sop Wter CP, ºC Filter time, s CP = cloud point; MG = monoglycerides; SG = sterol glucosides. free ftty cids (FA), methnol, nd wter. The purified product contined 95 wt% MG nd smll mount of DG, with typicl soy FA profile. SG were obtined from filter used in commercil production of soy biodiesel. Residue from the filter contined 21.5% SG, with the reminder consisting of methyl esters (77.2 wt%), MG nd DG (1.0 wt%), nd wter (0.3 wt%). The filter residue ws wshed with solvent, filtered, nd dried in 70 C oven overnight to obtin 99.8 wt% pure SG. The soy SG were dissolved in pyridine (Sigm, St. Louis, Missouri, USA) s 1% (wt/vol) solution. Sops were mixed from commercilly obtined mterils (Nu-Chek-Prep, Elysin, Minnesot, USA). Sodium plmitte, sodium sterte, sodium olete, nd sodium linolete were combined in the rtio of 0.105:0.041:0.232:0.621, respectively, to simulte the FA profile of soy. Sodium linolete ws used to represent ll the polyunsturted FA sops due to the oxidtive instbility of sodium linolente. The sop mixture ws dissolved in dry methnol (Acros, Morris Plins, New Jersey, USA; <50 ppm moisture) s 0.5% (wt/vol) solution. Neutrl ph wter ws obtined from the fcility deioniztion system. Four common components of soy biodiesel were chosen for this study. These components were dded to distilled soy biodiesel t levels relevnt to commercil soy biodiesel: 0 10,000 ppm (0 1.0%) MG, 0 40 ppm SG, 0 40 ppm sops, nd ppm wter. These levels were chosen bsed on previous filter test experience nd the expecttion tht filter test responses would be meningful compred to time limit of 900 s. A constrined mixture sttisticl experimentl design consisting of set of 30 experiments ws constructed nd nlyzed using commercil softwre (Stt-Ese Design Expert, Minnepolis, Minnesot). The system ws modeled s qudrtic eqution with first-order interctions. The trils were run in rndomized order. Tril btches were prepred by dding SG nd sop solutions s required to the weighed distilled SME. The solvents from the SG nd sop solutions s well s ny moisture were removed by heting the mixture under vcuum to 90 C nd stirring for 20 min. The mixture ws cooled to 70 C, vcuum ws broken, nd MG were dded s required nd stirred for 5 min. Finlly, wter ws dded s required by mens of volumetric pipet nd stirred for 5 min. After stirring, ech tril btch ws visully inspected to verify tht the btch ws cler nd ll components were completely dissolved t 70 C. Ech tril btch ws llowed to cool to room temperture nd stored under mbient conditions prior to testing. The CP of ech tril btch ws mesured per ASTM D 5773 Cloud Point of Petroleum Products (Constnt Cooling Rte Method) (5) using PCA-70X Automted Pour nd Cloud Point Anlyzer (Phse Technology, Richmond, British Columbi, Cnd). Smples (0.15 ml) t room temperture / 18

3 Effect of minor components in soy biodiesel on cloud point nd fi lterbility were cooled t 1.5 C/min to determine the CP opticlly, s indicted by the scttering of light cused by the formtion of crystls. The cold sok tretment nd filtrtion were crried out on 300-mL smples. After mixing ech tril btch, 300-mL smple ws trnsferred to 16-oz (450 ml) glss jr with Teflon-lined lid nd llowed to cool to room temperture. The smple ws then plced in 4.4 C (40 F) bth for 16-h (±10 min) cold sok. Cold sok filter times were mesured using modified ASTM D 6217 filtrtion procedure. After the 16-h cold sok, the smple ws removed from the bth nd llowed to wrm to C for 2.5 h (±15 min). The smple ws swirled briefly to disperse ny precipittes tht hd settled to the bottom of the jr nd immeditely filtered using modified ASTM D 6217 filtrtion procedure. This consisted of pouring the entire 300-mL smple into 650-mL stinless steel funnel with glss fiber filter (Whtmn GF/F, 47-mm dimeter, 0.7-µm pore size) supported by stinless steel filter support (Fig. 1). A vcuum of 23 in of mercury (585 torr) ws pplied, nd the time required for the entire 300-mL smple to pss through the filter ws mesured. RESULTS AND DISCUSSION The combintions of MG, SG, sop, nd wter dded to distilled soy biodiesel nd the CP nd filter times obtined re shown in Tble 1. Ech response ws the result of single mesurement. The distilled soy biodiesel with no components dded hd CP of 0.3 C nd cold sok filter time of 88 s. Sttisticl nlysis of vrince (ANOVA) ws performed using commercil softwre (Stt-Ese Design Expert). Terms tht were not sttisticlly significnt were eliminted from the models, nd selected point predictions of the CP nd filter time t vrious combintions of components were mde (Tble 2). As typicl soy biodiesel contins 6000 ppm MG, predictions were clculted t this level nd t two brcketing levels (3000 nd 9000 ppm). Note tht predictions cnnot be mde for levels of components tht re outside the rnges studied. The CP model developed in this study shows tht ddition of MG lone, especilly t 6000 ppm (0.6%) or higher, is predicted to produce the lrgest increse in CP (Tble 2). In ddition, sops or wter in combintion with MG t ppm nd higher lso would increse the CP. SG t the levels tested would not ffect the CP. The filter test model indictes tht the filter test response ws prticulrly sensitive to the presence of SG nd sops. Approximtely 10 ppm SG or sops would cuse similr undesirble increse in filter time s ppm MG using the modified ASTM D 6217 filtrtion method. Per the model, the presence of MG or SG cused the filter time to increse s the mount of dded wter incresed, but the presence of wter decresed the filter time when sop ws present t levels of 20 ppm or more. The interction of wter with other components bers further exmintion. FIG. 1. Modified ASTM D 6217 filtrtion method. specil supplement to inform / 19

4 Biorenewble Resources No. 4 / September 2007 Tble 2 Predicted CP nd filter times t set levels of MG nd vrible levels of SG, sop, nd wter SG, ppm Sop, ppm Wter, ppm MG = 3000 ppm Predicted CP, ºC Predicted filter time, s MG = 6000 ppm Predicted CP, ºC Predicted filter time, s MG = 9000 ppm Predicted CP, ºC Predicted filter time, s For bbrevitions see Tble 1. Although the presence of MG, SG, nd sops cn be controlled t the production stge, the moisture content in biodiesel my increse during trnsporttion, hndling, nd storge. These dt emphsize the importnce of creful hndling of biodiesel to prevent wter contmintion, even fter biodiesel leves the mnufcturing fcility. MG hve generlly been viewed s the single most hrmful component ffecting filter plugging. Although we observed tht MG negtively ffected filter times under the conditions of the cold sok filter test, filter test results were lso strongly ffected by the presence of other components nd by interctions between the components. The negtive effect of SG nd sops is not limited to cold conditions; they cn cuse filter-plugging problems even in wrm wether becuse these components cn be present s solid prticles t wrm tempertures fter prolonged storge. The cold sok is utilized to ccelerte the precipittion of sops nd SG. The MG effects in this study reflect the MG from soy oil feedstock. Soy oil hs bout 15% sturted FA (6), nd therefore soy biodiesel nd the MG present in soy biodiesel lso hve bout 15% sturtes. Soy biodiesel with 6000 ppm (0.6%) MG contins 900 ppm (0.09%) sturted MG. The CP of soy biodiesel increses significntly s the content of sturted MG increses (7). Exposure of soy biodiesel to cold tempertures during storge is known to cuse crystlliztion of sturted MG, which will concentrte in the bottom of storge vessels. Biodiesel drwn from the bottom of storge tnk my contin much higher levels of sturted MG thn those typiclly found in soy biodiesel. Some fts or oils, such s tllow or plm, hve sturted FA contents in excess of 40% (6), nd the impct of this level of sturted FA on CP nd filterbility of biodiesel mde from these feedstocks or soy biodiesel contining typicl levels of sturted FA cnnot be predicted from this study. For CP nlysis, the model hd n djusted R 2 vlue of 0.92 nd the precision, or signl-to-noise rtio, hd reltively high vlue of 20. A good model ws developed for CP, nd, overll, the CP model described the experimentl results nd cn be used to mke resonble predictions within the experimentl constrints. CP tempertures rnged from 0.3 to 3.9 C, nd the repetbility of the test instrument ws 1.3 C. Filter time response dt were trnsformed to log 10 (filter time) to fcilitte nlysis. This yielded model with high djusted R 2 vlue of 0.86 nd precision of 17. However, the model did not pss the lck of fit test. The reson for this is tht the qudrtic model used in modeling the system is indequte to explin the vrition in the experimentl design spce completely. Modeling with cubic eqution would ddress this issue but would require significntly more experimentl runs. However, the qudrtic model does n dequte job of explining the reltionships between the components in the center of the design spce nd only develops problems when multiple components re present t their mximum or minimum vlues. Therefore, the filter time model is considered suitble for predictions nd estblishing trends in the effects for the components studied, provided we re wre of the incresed possible error t high nd low levels of more thn one component. Finlly, severl trils were repeted to compre with the predicted results nd initil test results. Good repetbility ws obtined (Tble 3). The predicted vlues mtched the trends seen in the ctul results nd were lso in resonble greement with the ctul vlues (within the 95% confidence intervl). / 20

5 Effect of minor components in soy biodiesel on cloud point nd fi lterbility Tble 3 Repetbility of test results nd comprison of predictions with test results MG, ppm SG, ppm Sop, ppm Wter, ppm CP, ºC Predicted Initil tests , 0.6, Repet tests Filter time, s Predicted Initil tests , 362, Repet tests For bbrevitions see Tble 1. CONCLUSIONS Overll, MG, especilly t 6000 ppm (0.6%) or higher, hd the most significnt negtive impct on CP. In ddition, sops or wter in combintion with MG t ppm nd higher hd negtive impct on CP. SG did not hve negtive effect on CP. A good sttisticl model ws developed for CP. The filter test response ws prticulrly sensitive to SG nd sops. Filter times lso incresed with MG concentrtion, but the effect ws much less drmtic. The filter test response ws negtively ffected overll by wter, but interesting interctions between wter nd other components were observed. The negtive effect of wter emphsizes the need for proper hndling of biodiesel fter the production stge. The model for filter time fter cold sok hd some limittions but ws useful in estblishing trends in the effects of the components studied. Although this study clerly showed the interction between the four components studied, further exmintion of the interctions between the components is necessry to wholly understnd the effects of these components on biodiesel filterbility. Proposed future work includes further exmintion of component interction using n experimentl design for cubic eqution modeling, nd testing of ctul engine performnce using biodiesel contining the components exmined in this study. Lis Pflzgrf is mnger of oil process reserch, Archer Dniels Midlnd Co. (ADM), Dectur, Illinois, USA. Inmok Lee is ADM mnger of vegetble oil reserch. Jmes Foster is ADM director of corn processing reserch. George Poppe is ADM mnger of industril oils reserch. Contct him vi e-mil t: poppe@admworld.com. REFERENCES 1. Annul Book of ASTM Stndrds, Vol , ASTM Interntionl, West Conshohocken, Pennsylvni, D Stndrd Specifiction for Biodiesel Fuel Blend Stock (B100) for Middle Distillte Fuels. 2. Zemn, N., Remember December, Biodiesel Mgzine 3:52 56 (2006). 3. Annul Book of ASTM Stndrds, Vol , ASTM Interntionl, D Stndrd Test Method for Filter Plugging Tendency of Distillte Fuel Oils. 4. Annul Book of ASTM Stndrds, Vol , ASTM Interntionl, D Stndrd Test Method of Cloud Point of Petroleum Products (Constnt Cooling Rte Method). 5. Annul Book of ASTM Stndrds, Vol , ASTM Interntionl, D Stndrd Test Method for Prticulte Contmintion of Middle Distillte Fuels by Lbortory Filtrtion. 6. Hmmond, E.G., Sources of Fts nd Oils, in Introduction to Fts nd Oils (O Brien, R.D., W.E. Frr, nd P.J. Wn, eds.), 2nd edn., AOCS Press, Chmpign, Illinois, USA, 2000, pp Yu, L., I. Lee, L. Johnson, nd J. Vn Gerpen, The Influence of Trce Components on the Melting Point of Methyl Soyte, J. Am. Oil Chem. Soc. 75: (1998). specil supplement to inform / 21

631 ALLOWANES AND TOLERANCES

631 ALLOWANES AND TOLERANCES ANSI Fits. Tbles 3 through 9 inclusive show series of stndrd types nd clsses of fits on unilterl hole bsis, such tht the fit produced by mting prts in ny one clss will produce