Optimization of direct conversion of wet algae to

Bioresource Technology 102 (2011) 118–122

Contents lists available at ScienceDirect

Bioresource Technology journal homepage: www.elsevier.com/locate/biortech

Optimization of direct conversion of wet algae to biodiesel under supercritical methanol conditions Prafulla D. Patil a, Veera Gnaneswar Gude a, Aravind Mannarswamy a, Shuguang Deng a,*, Peter Cooke b, Stuart Munson-McGee a, Isaac Rhodes c, Pete Lammers c,1, Nagamany Nirmalakhandan d a

Chemical Engineering Department, New Mexico State University, Las Cruces, NM 88003, USA Electron Microscopy Lab, New Mexico State University, Las Cruces, NM 88003, USA Chemistry and Biochemistry Department, New Mexico State University, Las Cruces, NM 88003, USA d Civil and Environmental Engineering Department, New Mexico State University, Las Cruces, NM 88003, USA b c

a r t i c l e

i n f o

Article history: Received 27 March 2010 Received in revised form 26 May 2010 Accepted 7 June 2010 Available online 29 June 2010 Keywords: Biodiesel Wet algae Supercritical methanol Response surface methodology

a b s t r a c t This study demonstrated a one-step process for direct liquefaction and conversion of wet algal biomass containing about 90% of water to biodiesel under supercritical methanol conditions. This one-step process enables simultaneous extraction and transesterification of wet algal biomass. The process conditions are milder than those required for pyrolysis and prevent the formation of by-products. In the proposed process, fatty acid methyl esters (FAMEs) can be produced from polar phospholipids, free fatty acids, and triglycerides. A response surface methodology (RSM) was used to analyze the influence of the three process variables, namely, the wet algae to methanol (wt./vol.) ratio, the reaction temperature, and the reaction time, on the FAMEs conversion. Algal biodiesel samples were analyzed by ATR-FTIR and GC– MS. Based on the experimental analysis and RSM study, optimal conditions for this process are reported as: wet algae to methanol (wt./vol.) ratio of around 1:9, reaction temperature and time of about 255 °C, and 25 min respectively. This single-step process can potentially be an energy efficient and economical route for algal biodiesel production. Ó 2010 Elsevier Ltd. All rights reserved.

1. Introduction Biodiesel can be produced from algal biomass and oils by extraction-transesterification, direct methanolysis and transesterification methods (Johnson and Wen, 2009; Belarbi et al., 2000). Traditionally, algal biodiesel is produced from wet algal biomass in a series of steps including preparation of dry algae powder, extraction of algal oils with chemical solvents, and conversion of the algal oil to biodiesel with a catalyst (Chisti, 2007). Drying the biomass and extraction of algal oils by conventional methods are both energy- and cost-intensive. An alternative to the conventional extraction and transesterification methods is supercritical process. Using water in wet algae as a tunable co-solvent in supercritical methanol process not only accelerates the conversion of fats and algal oils to fatty acid methyl esters (FAMEs), but also increases solubility and acidity. In this work, a single-step supercritical process for simultaneous extraction and transesterification of wet algal biomass is demonstrated. In the proposed process, FAMEs can be produced from polar phospholipids, free fatty acids, and triglycerides by * Corresponding author. Tel.: +1 575 646 4346; fax: +1 575 646 7706. E-mail address: [email protected] (S. Deng). 1 Present address: Solix Biofuels, 430-B North College Ave., Fort Collins, Co 80524, USA. 0960-8524/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2010.06.031

increasing fluidity and volatility while reducing the polarity of the high-energy molecules in algae at supercritical conditions. Other advantages with the supercritical process are that they operate at modest temperatures and have lower energy requirements compared to conventional extraction and transesterification methods due to single pot conversion of algal biomass to biodiesel. The cost of biodiesel production from vegetable oils through supercritical process is estimated to be $0.26/gal which is about half of that of the conventional transesterification methods, $0.51/gal (Anitescu et al., 2008). This may very well be applied to algal oils as the algal biomass contains much higher levels of unsaturated fatty acids, lipids and triglycerides. The objective of the present study is to demonstrate the direct liquefaction and conversion of wet algal biomass into biodiesel via a single-step supercritical methanol process in the presence of nitrogen, and to optimize the process parameters that influence the super critical transesterification reaction using response surface methodology (RSM).

2. Experimental 2.1. Materials and methods Algal paste from outdoor open raceways (Inoculum Nannochloropsis sp. (CCMP1776) provided by CEHMM Artesia, NM was used in

119

P.D. Patil et al. / Bioresource Technology 102 (2011) 118–122

this study. For GC–MS analysis, methyl heptadecanoate (C17), standard for GC, was purchased from Fulka, Milwaukee, WI. Extra pure (99%) methanol, hexane, acetic acid and sulfuric acid were purchased from Acros Organics, New Jersey. For the purification of crude algae FAME, SPE Silica columns were procured from Thermo Scientific, Waltham, MA. The supercritical methanol process was carried out in the PARR 4593 Micro-reactor with a 4843-controller (Parr Instrument Company, Illinois, USA). Transmission electron microscopy (TEM) of frozen and residue (after SCM) algal biomass was examined with a model H-7650 electron microscope (Hitachi High-Technologies, Pleasanton, CA) operated in the bright field mode.

2.2. Characteristics of Nannochloropsis algal species The ash-free dry weight of the algae sample and lipid yield on dry weight basis were found to be 69.8% and 50%, respectively. Lipid extraction report for Nannochloropsis sp. has the following composition: triglycerides: 37.74%; other non-polar hydrocarbons, isoprenoids: 8.72% and polars, glycolipids, phospholipids: 3.54%. The gross estimation of non-polar hydrocarbons and triglycerides was done using thin layer chromatography (TLC) and densitometry technique. Previous work by other researchers (Damiani et al., 2010; Pyle et al., 2008) showed that total fatty acids accounted for 30–50% of dry biomass, depending on different culture conditions. Qualitative elemental analysis of crude algal biomass was determined by scanning electron microscopy (SEM, HITACHI S3400 N) equipped with energy-dispersive X-ray spectroscopy (EDS). The major elements and their approximate composition (wt.%) were carbon (72%), oxygen (21%), sodium (1.5%), magnesium (0.41%), silicon (0.93%), phosphorous (0.47%), chlorine (1.52%), potassium (0.96%). The FTIR spectra of the Nannochloropsis algal species show the general features indicating: (i) the highly aliphatic character of the residues revealed by the strong absorption at 720 cm1, (ii) the presence of hydroxyl groups characterized by the absorption

centered at 3400 cm1, and (iii) the presence of carboxyl groups characterized by the absorption band at 1710 cm1, (iv) The presence of carbonyl groups indicated by the absorption band at about 1735 cm1. 2.3. Biosynthesis of triglycerides in microalgae and SCM transesterification mechanism The biosynthesis route of triglycerides in microalgae may consist of the following three steps: (a) the formation of acetyl coenzyme A (acetyl-coA) in the cytoplasm; (b) the elongation and desaturation of the carbon chain of fatty acids; and (c) the biosynthesis of triglycerides (Huang et al., 2010). Similar to other higher plants and animals, microalgae are able to biosynthesize triglycerides to store biomass and energy. In general, L-a-phosphoglycerol and acetyl-coA are two major elements required for the biosynthesis of triglycerides. In the supercritical state, depending on pressure and temperature, the intermolecular hydrogen bonding in the methanol molecule will be significantly decreased. As a result, the polarity and dielectric constant of methanol are reduced allowing it to act as a free monomer. Subsequently, methanol at supercritical conditions can solvate the non-polar triglycerides to form a single phase of lipid/methanol mixture and yield fatty acid methyl esters and diglycerides (Saka and Kusdiana, 2001; Kasim et al., 2009). In a similar way, diglyceride is transesterified to form methyl ester and monoglyceride, which is converted further to methyl ester and glycerol in the last step. 2.4. Experimental design The key variables in the proposed process affecting the FAME content of the product are the wet algae to methanol (wt./vol.) ratio, the reaction temperature, and the reaction time. A response surface methodology (RSM) was used (Myers and Montgomery, 2002) to analyze the influence of these three process variables on

Table 1 Experimental design based on RSM for direct transesterification of wet algal biomass. Run order

Std order

Temperature (oC)

1 5 6 26 11 20 25 24 13 16 3 12 22 19 7 14 9 15 17 27 8 18 21 4 23 10 2 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

240

Methanol (wt./vol.)

Reaction time (min)

Observed FAME%

Predicted FAME%

4

10

24.38

240

4

30

240 240

8 12

20 10

240

12

30

250 250 250

4 8 8

20 10 20

250 250 260

8 12 4

30 20 10

260

4

30

260 260

8 12

20 10

260

12

30

25.12 20.15 68.13 53.15 38.35 32.05 28.15 55.15 62.62 55.56 76.05 82.15 77.20 83.25 79.15 78.15 76.07 84.15 70.91 45.07 39.31 66.37 54.12 68.30 70.00 72.35 78.40 85.75

56.02 49.32 27.37 59.01 63.03 66.70 77.79

88.88 77.30 44.43 57.15 71.19 69.97 82.69

120

P.D. Patil et al. / Bioresource Technology 102 (2011) 118–122

the fatty acid methyl esters (FAMEs) content. Based on experience and economic feasibility, a three factorial subset design proposed by Gilmour (2006) was employed. The design contains three levels on three factors that could be represented by a cube with six replications at the center. The six replications at the center offer better approximation of the true error which statistically helps in determining significance of the variables. Another advantage of this method is its symmetry in design with regard to the center, which offers equal importance to all levels of all parameters. The total number of experimental runs was 28 with replications as shown in Table 1. The wet algae to methanol ratios (wt./vol.), reaction times, and reaction temperatures were varied in the ranges of 1:4–1:12, 10–30 min, and 240–260 °C, respectively. The lower temperature limit, 240 °C, was just above the critical temperature of methanol and the upper temperature limit, 260 °C, was determined by the decomposition temperature limit of algal biomass (based on several trial runs). A general second order linear model with the deconstructionist approach was employed for its flexibility and ease of parametric evaluation for the predicted response surface. Statistically insignificant terms were excluded from the proposed design based on design hierarchy for the construction of the response surface. Also, the interaction terms considered manifests a better estimation on the combination effect of any two parameters considered. Linear least square method was used to predict the values of parameters involved. The confidence level of the statistical analysis conducted was 95%. 2.5. Experimental procedure

structionist approach was followed which indicates the consideration of a complete quadratic model and eliminating terms which were not significant as the analysis continued. All further analysis was carried out using both coded and uncoded variables. Method of least squares was employed to ascertain the values of the model parameters and ANOVA to establish their statistical significance at a confidence level of 95% (in our case). 3. Results and discussion 3.1. Development of regression model The central composite design (CCD) matrix and the response obtained from the experimental runs are shown in Table 1. The estimated regression coefficients for response and analysis of variance (ANOVA) for response are shown in Tables 2.1 and 2.2, respectively. The regression analysis indicates that all the three parameters had significant influence on the fatty acid methyl ester content, which is confirmed by the P-values. The response surfaces were fitted using process variables that were found to be significant after the analysis. The P-value of the lack of fit analysis is 0.133, which is more than the 0.05 (confidence level is 95%). The regression model provides accurate description of the experimental data indicating successful correlation among the three transesterification process parameters that affect the yield of algal biodiesel. This is further supported by the correlation coefficient, R2 of 0.921. 3.2. Effects of process parameters and optimization

The experimental protocol for one-step supercritical methanol process is as follows: From the aliquots prepared previously (frozen biomass in 50 mL falcon tubes at 80 °C), 4 g of wet algae paste (10% solids) was subjected to a non-catalytic supercritical methanol (SCM) process in a 100 mL PARR micro-reactor under a matrix of conditions: constant pressure of 1200 psi; reaction times of 10, 20, and 30 min; reaction temperatures of 240, 250, and 260 °C; and wet algae to methanol (wt./vol.) ratios of 1:4, 1:8, and 1:12. After the reaction was completed, the reactor contents were transferred into a 50 mL round-bottom flask and freed of methanol and volatiles at a reduced pressure in a rotary evaporator. The remaining products were taken in hexane and then centrifuged (3200 rpm) for 5 min. The upper organic layer containing non-polar lipids was extracted and run through a short column of silica (Hyper SPE Silica). Neutral components were eluted with the solvent. For qualitative analysis, an internal standard, methyl heptadecanoate (C17:0) was added quantitatively to the eluted neutral component-solvent solution and analyzed by gas chromatography-mass spectroscopy (GC–MS). The content of the fatty acid methyl ester in the final product was calculated quantitatively by comparing the peak areas of fatty acid methyl esters to the peak area of the internal standard (methyl heptadecanoate, C17:0) obtained from GC–MS.

Fig. 1 shows the response contours of FAME yield against reaction temperature and wet algae to methanol (wt./vol.) ratio at three different reaction time intervals and fixed reaction pressure of 1200 psi. The values and signs on the regression coefficients suggest that the reaction time affects the response positively for temperatures up to 255 °C; however, reaction temperatures above 255 °C were not suitable for transesterification reaction of the algal biomass at fixed

Table 2.1 Estimated regression coefficients for response (the analysis was done using coded units). Term

Coef

SE Coef

T

P

Significant

Constant Temp Meth Time Temp  Temp Meth  Meth Time  Time Temp  Meth Temp  Time Meth  Time

77.012 10.933 7.133 11.088 20.213 10.303 6.562 5.638 4.729 2.047

1.938 1.337 1.337 1.337 3.38 3.38 3.38 1.418 1.418 1.418

39.73 8.179 5.336 8.295 5.981 3.048 1.942 3.977 3.336 1.444

0 0 0 0 0 0.007 0.068 0.001 0.004 0.166

Yes Yes Yes Yes Yes Yes No Yes Yes No

R2 = 94.46%, R-Sq (pred) = 84.35%, R-Sq (adj) = 91.69%.

2.6. Statistical analysis A general linear model which accounts for the single parameters’ linear and quadratic effects with their interaction effects was considered. The following is the general linear model for our analysis:

l ¼ b0 þ

3 X i¼1

bi xi þ

3 X i¼1

bii x2i þ

2 3 X X

bij xi xj

i¼1 j¼iþ1

where, x1, x2 and x3 are the levels of the factors and l is the predicted response if the process were to follow the model. A decon-

Table 2.2 Analysis of variance for response. Source

DF

Seq SS

Adj SS

Adj MS

F

P

Regression Linear Square Interaction Residual error Lack-of-fit Pure error Total

9 3 3 3 18 5 13 27

9870.7 5280.7 3656.5 933.5 578.9 257.6 321.4 10449.6

9870.7 5280.7 3656.5 933.5 578.9 257.6 321.4

1096.74 1760.23 1218.83 311.17 32.16 51.51 24.72

34.1 54.73 37.9 9.67

0 0 0 0.001

2.08

0.133

P.D. Patil et al. / Bioresource Technology 102 (2011) 118–122

121

As expected, a longer reaction time allows the transesterification reaction to proceed to completion and results in a higher yield of FAMEs from algal biomass. According to Vieitez et al. (2009), higher reaction time beyond particular limit in supercritical alcohol process for vegetable oil may lead to greater losses of unsaturated FAME due to degradation reactions. Nevertheless, Fig. 1 shows that the effect of reaction time is more prominent at wet algae to methanol (wt./vol.) ratio of 1:9 and reaction temperature around 255 °C at a fixed reaction pressure of 1200 psi. In water added supercritical methanol reaction, the water–methanol mixture has both strong hydrophilic and hydrophobic properties that help speeding up the reaction significantly (Kusdiana and Saka, 2004). Based on the experimental analysis and RSM study, the optimal conditions for this process are reported as: wet algae/methanol (wt./vol.) ratio of around 1:9, reaction temperature and time of about 255 °C, and 25 min respectively. 3.3. TEM of frozen algal biomass (raw material) and residual Nannochloropsis algal sample after SCM process

Fig. 1. FAME yield against reaction temperature and wet algae wt/methanol volume ratio at different reaction times using RSM.

For analysis of elemental composition, raw and residual samples were washed with distilled water and centrifuged pellets were excised from centrifuge tubes, air-dried and glued to carbon adhesive tabs (Electron Microscopy Sciences, Hatfield, PA) on aluminum sample stubs. Elemental spectra were collected at 15 kV using a model S-3400 N scanning electron microscope (Hitachi High-Technologies, Pleasanton, CA) equipped with a model Noran System Six 300 energy-dispersive spectrometer system (Thermo Electron Corp., Madison, WI). From TEM analysis report of frozen (raw) and residual algal biomass, it was found that at SCM condition, algal cell wall structure was totally disturbed and fragmented while EDS report showed the evidence for thermal degradation of algal biomass (wt.% of ‘C’ increased in residue) due to high content of unsaturated fatty acids in lipid. 3.4. Analysis of algal biodiesel conversion

pressure of 1200 psi. This may be due to the fact that the oil/lipid and the alkyl esters tend to decompose or become thermally unstable above the specified temperature owing to the high content of unsaturated fatty acids (Gui et al., 2009). It was observed that at high temperature and pressure, unsaturated fatty acids tend to decompose due to the isomerization of the double bond functional group from cis-type carbon bonding (C@C) into trans-type carbon bonding (C@C), which are naturally unstable fatty acids (Imahara et al., 2008). The FT-IR spectrum (reaction temperature of 270 °C) shows a trans-type carbon bonding (C@C) group at the wavelength of 960 cm1 which is absent at this wavelength for the reaction temperature of 250 °C. The greater the percentage of unsaturation in fatty acids or esters of algal biomass, the more it is susceptible to oxidation. The extent of unsaturation of in esters can be reduced easily by partial catalytic hydrogenation of the oil (Dijkstra, 2006). Wet algae to methanol (wt./vol.) ratios have a positive effect on the yield up to 1:9 but have a negative impact at higher levels. Higher ratio of biomass to methanol could shift the reversible reaction forward (as observed) perhaps due to increased contact area between methanol and lipid, resulting in higher yield of FAME and it also contributes to the lower critical temperature of the mixture. However, its interaction with reaction temperature can, on the other hand, cause a reduction in the yield of FAME due to either the decomposition of FAME or the critical temperature of the reactant/product mixture between methanol and FAMEs becomes highly dependent on the concentration of methanol and may decrease the critical temperature of the reactant/product. When reactant/product mixture is heated above critical temperature, it has the tendency to decompose (Hegel et al., 2008).

For the quantification of reaction product, the algal biodiesel samples were analyzed by a gas chromatography–mass spectrometry (GC–MS) system incorporated with an Agilent 5975 C MSD and an Agilent 7890 A GC. The content of the fatty acid methyl ester in the final product was calculated quantitatively by comparing the peak areas of fatty acid methyl esters to the peak area of the internal standard (methyl heptadecanoate, C17:0) obtained from GC–MS. It is noted from GC–MS results that algal biodiesel contains a major proportion of mono and poly unsaturated fatty acid methyl esters. The major fatty acids were palmitoleic acid (C16:1, 30–33%), oleic acid (C18:1, 35–38%), eicosapentanoic acid (EPA, C20:5n3, 5–8%), palmitic acid (C16:0, 5–10%,) and arachidonic acid (C20:4n6, 1–3%). From the GC–MS peak and total ion chromatography (TIC) data, it was observed that the algal biodiesel contains olefins, fatty alcohols and sterols in minor quantities along with saturated and unsaturated FAMEs. The relative weight compositions of organic compounds present in the algal biodiesel were analyzed using GC–MS, and summarized in Table 3. ATR-FTIR spectra of petro-diesel, camelina biodiesel, and algal biodiesel are shown in Fig. 2. The IR spectra were obtained using a PerkinElmer Spectrum 400 FT-IR/FT-NIR spectrometer. The main components of diesel are aliphatic hydrocarbons, whose chemical structures are similar to long carbon chain of the main components of biodiesel. The observation of an absorption Peaks around 1200 cm1 may be assigned to the antisymmetric axial stretching vibrations of CC(@O)–O bonds of the ester, while peaks around 1183 cm1 may be assigned to asymmetric axial stretching vibrations of O–C–C bonds (Silverstein and Webster, 1998). In addition,

122

P.D. Patil et al. / Bioresource Technology 102 (2011) 118–122

Table 3 GC–MS peak TIC data of crude biodiesel obtained from algal biomass. Peak

Retention time

Area (%)

Name

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

3.264 3.802 4.609 4.905 5.250 5.547 5.759 6.354 7.933 8.305 9.524 10.245 13.587 13.632 14.125 15.124 16.250 16.923 17.089 17.878

0.63 0.03 0.01 0.2 0.1 0.1 0.35 0.18 0.05 3.14 19.72 0.02 1.74 4.5 0.2 7.34 10.2 9.92 33.28 8.29

2H-pyran Octanoic acid methyl ester Nonanoic acid methyl ester Dodecane Benzenamine Indole Decanoic acid methyl ester Naphthalene,2,6-dimethyl cis-5-Dodecenoic acid methyl ester Undecanoic acid, 10-methyl ester Tridecanoic acid methyl ester 8-Heptadecene 7-Hexadecenoic acid methyl ester Cyclopropaneoctanal,2-octyl-9-Eicosyne Cyclohexaneethanol 9-Hexadecenoic acid methyl ester Tricyclo decane Pentadecanoic acid,13-methyl ester 9-Octadecenoic acid methyl ester Eicosanoic acid methyl ester

Fig. 2. FTIR results of algal biodiesel, petro-diesel, and camelina biodiesel.

since biodiesel is mainly mono-alkyl ester, the intense C@O stretching band of methyl ester appears at 1743 cm1 for algal and camelina biodiesel which is absent in petro-diesel spectra. 4. Conclusions The single-step process for wet algal biomass (Inoculum Nannochloropsis sp., CCMP1776) may offer the benefits of shorter reaction time, simple purification of products and maximum conversion of

triglycerides into their corresponding fatty acid methyl esters. The single-step process favors the energy requirements for biodiesel production by eliminating the needs for drying and extraction of algal biomass. Process optimization using response surface methodology (RSM) design proved to be a valuable tool for evaluating the effects of the process variables on the FAMEs yield. The single-step process has the potential to provide an energy efficient and economical route to algal biodiesel production. Acknowledgements This project was partially supported by New Mexico State University Office of Vice President for Research and State of New Mexico through a New Mexico Technology Research Collaborative grant. The authors are thankful to CEHMM Artesia, NM for providing the wet algal biomass for biodiesel testing. References Anitescu, G., Deshpande, A., Tavlarides, L., 2008. Integrated technology for supercritical biodiesel production and power cogeneration. Energy Fuels 22, 1391–1399. Belarbi, E.-H., Molina, G.E., Chisti, Y., 2000. A process for high yield and scalable recovery of high purity eicosapentaenoic acid esters from microalgae and fish oil. Enzyme Microb. Technol. 26, 516–529. Chisti, Y., 2007. Biodiesel from microalgae. Biotechnol. Adv. 25, 294–306. Damiani, C., Popovich, C., Constenla, D., Leonardi, P., 2010. Lipid analysis in Haematococcus pluvialis to assess its potential use as a biodiesel feedstock. Bioresour. Technol. 101, 3801–3807. Dijkstra, A.J., 2006. Revisiting the formation of trans isomers during partial hydrogenation of triacylglycerol oils. Eur. J. Lipid Sci. Technol. 108, 49–64. Gilmour, S.G., 2006. Response surface designs for experiments in bioprocessing. Biometrics 62, 323–331. Gui, M., Lee, K., Bhatia, S., 2009. Supercritical ethanol technology for the production of biodiesel: process optimization studies. J. Superc. Fluids 49, 286–292. Hegel, P., Andreatta, A., Pereda, S., Bottini, S., Brignole, E., 2008. High pressure phase equilibria of supercritical alcohols with triglycerides, fatty esters and cosolvents. Fluid Phase Equilib. 266, 31–37. Huang, G., Chen, F., Wei, D., Zhang, X., Chen, G., 2010. Biodiesel production by microalgal biotechnology. Appl. Energy 87, 38–46. Imahara, H., Minami, E., Hari, S., Saka, S., 2008. Thermal stability of biodiesel in supercritical methanol. Fuel 87, 1–6. Johnson, M., Wen, Z., 2009. Production of biodiesel fuel from the microalga Chizochytrium limacinum by direct transesterification of algal biomass. Energy fuels 23, 5179–5183. Kasim, N., Tsai, T.-H., Gunawan, S., Ju, Y.-H., 2009. Biodiesel production from rice bran oil and supercritical methanol. Bioresour. Technol. 100, 2399–2403. Kusdiana, D., Saka, S., 2004. Effects of water on biodiesel fuel production by supercritical methanol treatment. Bioresour. Technol. 91, 289–295. Myers, R.H., Montgomery, D.C., 2002. Response Surface Methodology: Process and Product Optimization Using Designed Experiments. Response Surface Methodology, second ed. Wiley, New York. Pyle, D.J., Garcia, R.A., Wen, Z.Y., 2008. Producing docosahexaenoic acid (DHA)-rich algae from biodiesel-derived crude glycerol: effects of impurities on DHA production and algal biomass composition. J. Agric. Food Chem. 56, 3933–3939. Saka, S., Kusdiana, D., 2001. Biodiesel fuel from rapeseed oil as prepared in supercritical methanol. Fuel 80, 225–231. Silverstein, R., Webster, F., 1998. Spectrometric identification of organic compounds, sixth ed. Wiley, New York. Vieitez, I., Silva, C., Alckmin, I., Borges, G.R., Corazza, F., Oliveira, J.V., Grompone, M.A., Jachmanián, I., 2009. Effect of temperature on the continuous synthesis of soybean esters under supercritical ethanol. Energy Fuels 23, 558–563.