Journal of Undergraduate Chemistry Research, 2009, 8(1), 22

STUDY OF HETEROGENEOUS BASE CATALYSTS FORBIODIESEL PRODUCTION

Albert J. Gotcht, Aaron J. Reeder* 1,and Aleesha McCormick2

Chemistry Department, Mount Union College, Alliance, OR 44601, gotchaj@muc.edu1. current address Chemistry Department, Carnegie-Mellon University, Pittsburgh, PA 152132. current address Biomedical Engineering Department, University of Akron, Akron, OR 44325

Abstract

Solid metal oxides were tested as heterogeneous base catalyst for the production of biodiesel fuel derived fromvirgin and waste vegetable oil feedstocks. Compounds tested included MgO,CaO, SrO, and BaO. The most effectivecatalyst was 1.0 mol% BaO which showed >95% conversion of canola oil to biodiesel in 1 hour at 50GC.The relativeorder of effectiveness was BaO -SrO > NaOH » CaO - MgO.The SrO catalyst was studied further due to toxicityconcerns with BaO. Catalyst recovery was difficult and conversion efficiency dropped to 38% on re-use of the SrO.Further testing of silica and alumina as solid supports for SrO did not improve recoverability or conversion efficiency.The experimental results suggest a different paradigm such as solid acid catalysts may provide a more fruitfulapproach for the heterogeneous transesterification of vegetable oils to biodiesel.

iil:

Introduction

Keywords: biodiese1, heterogeneous, catalyst, 8rO, transesterification

Transportationfuels likegasoline, diesel, andjet fuelare currently derived nearly entirely fromnonrenewable petroleum with 2007 worldwideproduction at nearly 350 billion gallons (1). Incontrast,just under 18billion gallons ofbiofuels wereproduced in the same year (2).The Energy InformationAdministration projects a steady increase in world oilconsumption at ~ 1.3% per year through 2009 and oilprices of $120-140 per barrel (3). This has led toincreased efforts to promote renewable sources ofcarbon for the production of transportation fuels. Forexample, Europe has promoted a fuel standard of20%biofuels by 2020; Brazil has developed E85 vehiclesand a sugar cane based ethanol supply chain; and theU.S. has provided tax incentives for both ethanol andbiodiesel production. The diesel engine was originallydesigned as a high efficiency combustion engine (4)and has been the engine of choice in many industrialapplications like construction, transportation, andgeneration of electricity. Thus, the production ofbiodiesel is one important way to begin to addressissues of sustainability.0

IICH2O-C-R

I ?,HC-Q-C-R + R'-OH

I 'ifCH2O-C-R1riglyceride

catalyst ,.. Cj(

3 R'O-G-R

CH2OH

IHC-OH

ICH.oH

glycerol

+

alcohol esters

Figure 1. Transesterification of Vegetable Oils-. '..'~:>

The current method of producing biodiesel is thebase catalyzed transesterification of vegetable oil asnoted in Figure 1. Here an alcohol, usually methanolor ethanol, is added to a triglyceride using a basecatalyst such as NaOH, KOH, NaOCH3' or KOCH3(5). The process is fast and efficient at relative lowtemperatures. For example,using NaOH and methanolat 50GC a >95% conversion may be obtained inapproximately 1hour (6). However, cost competitiveproduction of biodiesel remains an important issue.Biodiesel production costs may be reduced by choiceof feedstock, production plant location, and increasedprocess efficiencies (7). For example, recent work hasfocused on the testing of waste cooking oil (8-10) andnon-food based oils like Jatropha as alternatefeedstocks for biodiesel production (11-14). Anotherarea where process ,efficienciesmay be realized is inthe use of heterogeneous rather than homogeneouscatalysts (15-17).

While the homogeneous process has manyadvantages there are a number of efficiency losses.The biodiesel product contains spent catalyst, crudeglycerol, and excess alcohol which requires additionalwashing and drying 'Stepsto gene_ratepure biodieselas noted in Figure 2. Also, the base catalyst is lost inthe wash step further decreasing effic.iency andincreasing costs. One way to increase>processefficiency is to use heterogeneous catalysts. Inprinciple the catalyst may be recovered and reusedleading to simpler biodiesel processing. A number of

SeparatorOil

Free Fatty Acids

Glycerol (85%)

Figure 2. Production of Biodiesel

MethanolRemoval

Drying

Biodiesel

recent studies (16-21) highlight the current interest inheterogeneous catalysis of the transesterificationreaction. Here we report the exploration of alkalineearthmetal oxidesas heterogeneousbase catalystswithvirgin and waste cooking oil for biodiesel production.

Experimental

Materials

Samples of 100% pure canola, com, and soybeanoil were purchased from Aldi Inc. under the Carlinibrand name. USP grade MgO was purchased fromMerck (Whitehouse Station, NJ), CaO was purchasedfrom Baker (Phillipsburg, NJ), 99.9% srO and BaOwere purchased from Aldrich (St. Louis, MO). HPLCgrade methanolwas purchased fromAldrich and driedover 4A molecular sieves. CertifiedACS grade NaOHwas purchased from Fisher (Pittsburg, PA). Sodiummethoxide was prepared by dissolving sodium metalinto methanol in a 0.1g:1OmLratio.All materials wereused as obtained unless otherwise specified.

Component characterizationCanolaandwaste cooking oil sampleswere analyzed

for their fatty acid content on a Clarus 500 GC/MS(Perkin Elmer, Waltham, MA) by first converting thetriglycerides into their respective fatty acid methylesters (FAME's). For this analysis 200 mL of oil wereadded to 5 mL of previously prepared 1%NaOCH3 inMeOH (m/v) in a 150 mm x 13 mm screw top vial.The closed vial was heated for 5 minutes in a 50°C

water bath. The resulting FAME's were extracted intohexane, washed with saturated NaCI water to removethe glycerol byproduct, and dried by passing througha columnofMgSO4'Then I mL of samplewas injectedonto an AT-Silar0.25 mm capillary column (Alltech,Deerfleld, IL) using a T-programmed run (22). Fattyacid content was verifi'ed using pure components,

Journal of Undergraduate Chemistry Research, 2009, 8(1), 23

standard mixes from Alltech, or a 117,000 componentNIST spectral library.

The pre- and post- sample acidity was determinedvia titration. First, 250 mg of the feedstock oil orFAME was dissolved in 50 mL of 50:50 ethylether:ethanol with 3 drops of phenolphthaleinindicator. Then 0.1 M standardized KOH in ethanol

was added until the endpoint was reached (23).

SrO catalysts supportedonAlz03 or SiOzwere madefrom 3 M Sr(N°3)Z solutions and the desired supportmaterial according to standard procedures (21). Therequired amount of solution to give the final % loadingwas determined and stirred with the solid supportmaterial with heating at ~ 250°C until dryness. Thenthe sample was calcined in air at 800°C in a tubefurnace for 6 hours to decompose the nitrate salt tothe oxide. The resulting catalyst was crushed with amortar and pestle to ~ 100 /-lm.Decomposition to SrOwas verified using a XE800 powder X-raydiffractometer (Philips Analytical, The Netherlands).

Biodiesel synthesis was tested using the basecatalyzed scheme shown in Figure 1. Each reactionmixture consisted of7.5 mL of methanol and 30 mLof oil added to a 125 mL round bottom flask. This

corresponded to a 6: I mol ratio of alcohol:oil whichis an actual mol excess of 2: 1 since the reactionrequires 3 mol of alcohol for every 1mol of oil. Theamount of base catalyst was varied between 0.1 to 1.0mol%.The reaction flaskwith vegetable oilwas placedin a water bath over a magnetic stirrer whose speedwas kept constant at ~ 1200 rpm. The bath temperaturewas set at 20, 35, or 50°C. After approximately 15minutes thermal equilibrium was reached and thefreshly made alcohpl plus catalyst mixture was added.Reaction progress was followed using the NMRmethod of Knothe (6). Approximately 1 mL of thereaction mixture was sampled at regular intervals,washed with 1 mL of water, and centrifuged. The toplayer was transferred to an NMR tube, tetramethyl-silane was added as an internal standard, and the NMRspectrum was obtained using a 60 MHz EM360spectrometer (Varian, Palo Alto, CA) with a digitalupgrade (Anasazi, Indianapolis, IN).

Results and Discussion

Figure 3 shows the results of testing the group 2basic metal oxides as transesterification catalysts for

Journal of Undergraduate Chemistry Research, 2009, 8(1), 24

100%

90%

80%

c 70%0'21 60%~§ 50%

(.)

w 40%

~ 30%

20%

10%

0%0 0.5 1.5

time (hr)2 2.5

Figure 3. FAMEConversion Using Basic Metal Oxides.Reaction conditions: 1200 rpm, 50°C, 6:1 MeOH:canolaoil. . SrO 1.0 mol%, . BaO 1.0 mol%, . NaOH 1.0 mol%,X CaO 1.0 mol%, + MgO 1.0 mol%.

biodiesel formation. MgO and CaO show very littleconversion at I mol % catalyst and 50°C over a 3 hourtime frame. Others have reported significantly higherconversions with CaO catalysts (15,24,25). Thisdifference is due to preparation conditions that affectthe surface area and surface states (25). SrO and BaOshow essentially complete conversion in approxi-mately 1 hour under the same reaction conditions.Results for a conventional NaOH catalyst are shownfor comparison. Since barium compounds generallyhave high toxicities BaO was dropped from furtherstudies and additional work focused on SrO.

Figure 4 shows a concentration of 0.5 mol% SrOwas nearly as effective as 1.0% at 50°C. If the time isextended to three hours, then 0.2 mol% SrO alsoqUilntitatively converted canola oil to FAME's.However, 0.1 mol% SrO is limited to approximately50% conversion. This is likely due to known catalyst"poisoning" effects such as losses from saponification

100%

80%

i60%

!~40%

20%

0%

0 0.5 1.5

time (h,)

2.5

Figure 4. Mole%of SrO. Reaction conditions: 1200 rpm,50°C, 6:1 MeOH:canola oil, SrO catalyst. X 1.0 mol%, 00.5 mol%, 00.2 mol%, OO.1I'1}ol%

0.00300-1.7 0.00310 0.00320 0.00330 0.00340 0.00350

-1.9

-2.1 y = -2356.6x + 5.2903

R' = 0.9898-2.3

:5 -2.5

-2.7

-2.9

.3.1

-3.3

-3.5

3 .NaOH .SralIT (1(1)

Figure 5. Arrhenius Plots for NaOH and SrO BaseCatalysts. Reaction conditions: 1200 rpm, 6:1 MeOH:canola oil, catalyst at 1 mol%. . SrO, . NaOH

side reactions (26) adsorption of free fatty acids (24)or the formation of carbonates via dissolved CO2fromthe atmosphere (19).

As expected, decreasing the reaction temperaturefrom 50 to 20°C led to decreased rates of reaction.Figure 5 shows an Arrhenius plot for SrO and NaOHcatalysts over this T range. Both gave good straightline fits and nearly identical activation energiesof 19.6kllmol for SrO and 21.4 kllmol for NaOH. This is ingood agreement with the value of 20 kllmol reportedby Dossin et al. for the mechanism of methoxyformation via methanol adsorption on MgO (18).

Additional experiments with SrO catalyst at 1mol%and 50°C indicated the catalyst could be generally usedwith other oils including corn and soybean. Alsosubstituting methanol with ethanol led to highconversions to FAME but at a slower rate. With 1-

propanol conversion was negligible even after 3 hours.

Waste cooking oil (WCO) was also collected froma local Kentucky Fried Chicken fast food restaurant.Figure 6 shows the results of the transesterification ofthis sample using SrO catalyst at 1 mol% andmethanol. The conversion went essentially tocompletion but in this case a 1-2 hr induction period

'was observed. A number of studies have reportedmixing to be an important factor in achieving goodFAME yields (8,24,25,27). The stirring rate used inthis study was 1200 rprll or two times higher than the600 rpm rate shown to be the lowest acceptable formaximum yields (27). S~nceWCO is known to have ahigher free fatty acid (FFA) content compared to virginoil, titrations were performed on raw and trans-esterified WCO. The acidity dropped from 4.01 mg

. . . ... . ... .

.X

x x Ix a a

xx a

D-0 0 0 0 0 0 4

..,0

100

80

fO'":I<

40

20

0

0 0.5 1.5 2

time (hr)

2.5 3.5

Figure 6. FAME Conversion Using SrO with WasteCooking Oil. Reaction conditions: 1200 rpm, 50°C, 6:1MeOH: waste cooking oil, SrO 1.0 mol%. Each point isthe average of three trials. Error bars are :1:10'.

KOH/mg oil to 0.35 mg KOH/mg oil, respectively.Previous work has shown basic catalysts are noteffective at transesterifying FFAs where the basecatalyst can be neutralized leading to a loss of catalystas noted above (8,9). In Figure 6 the amount of SrOpresent in the reaction mixture was 0.0022 mol. Theamount of acid neutralized calculated from the

difference in the KOH titration results correspondedto 0.0019 mol. Thus, it appears the

SrO + 2 RCOOH ~ RCOO-Sr-OOCR + H2O E-l

induction period is due to a competition betweenFAME formation and catalyst neutralization. Atest ofa canolaoil sample spikedto 44.7 mg KOH/g oilunderthe same reaction conditions showed only a 3%conversion to FAME after 63 hours.

Recovery of catalyst from the reaction mixture wasan important operational parameter. In this studyrecovery was plagued by loss of catalyst due tosecondary reactions as noted above and poor FAMEconversion efficiencies for recovered catalysts. Forexample, recovered SrO under the same reactionconditions in Figure 5 gave a FAME conversion of38% after 3 hours. To help determine where the Srwas going a flame test was performed on three productcomponents; excess methanol recovered by rotaryevaporation, the FAME layer, and the crude glycerollayer. The detection limit of Sr was determined to beapproximately 150ppm by dilutionof a stock solutionof Sr(NO3h- Both the recovered MeOH and theglycerol layer tested negative for Sr,while the FAMElayer tested positive. This behavior is in contrast tothat reported for NaOH catalyst where most of the

-

Journal of Undergraduate Chemistry Research, 2009,8(1),25

catalyst ends up in the glycerol layer (28). Further,the conductivity of each layer was determined to be0.0 /l-Susing an Orion l50A+ conductivity meterwhich has a detection limit of approximately 40 ppbNaCl. This is consistent with the loss ofSrO accordingto equation 1 and indicates ion exchange methods ofbiodiesel clean-up (15) may perform poorly if usedwith SrO catalysts.

Finally, SrO was loaded onto SiO2 or A1203inertsupport as noted in the ~xperimental section. All ofthese catalysts gave low FAME conversions similarto MgO and CaO as noted in Figure 3. Work byGranados et al. (19) indicates evenwith heterogeneousFAME catalysts the process is dominated by ahomogeneous mechanism. A number of authors havesited the importance of the solubility of the activespecies (15,19,24) even for heterogenous catalysts forvegetable oil transesterification. Our study confirmsthis view and extends it specifically to SrO. Thissuggests the alternative heterogeneous acid catalysts(17) for the transesterification of vegetable oils maybe a more fruitful approach for future pursuits.

Acknowledgements

AJG and AM gratefully acknowledge support fromthe National Science Foundation, Stark County Mathand Science Partnership, Grant EHR0226986. AJRgratefully acknowledges receipt of a PappenhagenSummer Research Fellowship from the Mount UnionCollege ChemistryDepartment in support of this work.AJG Dan.7:l4. .

References

(1). Based on worldwide total gasoline, kerosene, anddiesel oilproduction numbers fromthe lEA. InternationalEnergy Agency. Monthlv Oil Survev, April 2008.

(2). Food and Agriculture Organization of the UnitedNations. Bioenerf/Y.Food Security and Sustainability-Towardsand International Framework, Rome, Italy June3-5,2008.

(3). Energy Information Association. Short-Term Enerf/YOutlook, released July 8.

(4). W.Addy Majewski and Magdi K. Khair. DieselEmissions and Their Control, SAE International,Warrendale, PA2006.

(5). G. Vicente, M. Martinez and 1.Araeil. BioresourceTech., 2004, 92,297-305.

(6). G. Knothe. J. Am. Oil Chem. Soc., 2000, 77, 489-493.(7). B.A. Saville and S. Villeneuve. Feasibility Studv for a

Biodiesel Refining Facility in the Regional Municivalityof Durham, BEl Biofuels C~nada, Kitchener, ON,

i

T

. I III

-

~

Ii

Journal of Undergraduate Chemistry Research, 2009, 8(1), 26

February 2006.(8). M.G. Kulkarni andA.K. Dalai. Ind. Eng. Chern.Res.,

2006,45,2901-2913.(9). Y Wang, S. Ou, P.Liu, F. Xue and S. Tang.J. Malec.

Cat. A, 2006, 252, 107-112.(10). K.-T. Lee, T.A. Foglia and K.-S. Chang. J. Am. Oil

Chern.Soc., 2002, 79,191-195.(11). S. Shah, S. Shanna and M.N. Gupta. Energy & Fuels,

2004,18, 154-159.(12). M.M. Azam, A. Waris and N.M. Nahar. Biomass

Bioenergy, 2005, 29, 293-302.(13). R.A. Holser and R. Harry-O'Kuru. Fuel, 2006, 85,

2106-2110.(14). M.P. Dorado, E. Ballesteros, F.J.Lopez and M.

Mittelbach. Energy & Fuels, 2004, 18, 77-83.(15). P. Bondioli. Topics in Catalysis, 2004, 27, 77-82.(16). T.F.Dossin, M.-F. Reyniers and G.B. Marin. Appl.

Cat. B, 2006, 61, 35-45.(17). A.A.Kiss, F.Omota,A.C. Dimian and G. Rothenberg.

Topics in Catalysis, 2006, 40,141-150.(18). T.F.Dossin, M.-F.Reyneirs, R.J. Berger, G.B. Marin.

Appl. Cat. B, 2006, 67, 136-148.(19). M.L. Granados, M.D.Z. Poves, D.M. Alonso, R.

Mariscal, F.C. Galisteo, R. Moreno-Tost, J. Santamariaand J.L.G. Fierro. Appl. Cat. B, 2007, 73, 317-326.

(20). W. Xie and H. Li. J. Mol. Cat. A, 2006, 255, 1-9.(21). Z. Yangand W.Xie. Fuel Proc. Tech.,2007,88,631-

638.(22). T-program: first ramp 50°C to 150°Cat 10°C/min., 5

min. hold; second ramp 150°C to 200°C at 10°C/min.,10min. hold. Total run time = 30 min.

(23). P.G.Nouros, c.A. Georgiou and M.G. Polissiou.Anal.Chern.Acta, 1997, 351,291-297.

(24). M. Kouzu, T. Kasuno, M. Tajika,Y. Sugimoto, S.Yamanaka and J. Hidaka. Fuel, 2008, 87, 2798-2806.

(25). C. Reddy,V. Reddy, R. Oshel and J.G. Verkade.Energy Fuels, 2006, 20, 1310-1314.

(26). U. Schuchardt, R. Sercheli and R.M Vargas.J. Braz.Chern.Soc., 1998, 9, 199-210.

(27). R. Alcantara, J. Amores, L. Canoira, E. Fidalgo, M.J.Franco, A. Navarro. Biomass Bioenergy, 2000, 18, 515-527.

(28). J. Van Gerpen. Fuel Process. Tech., 2005,86, 1097-1107.

. .-