Page 1 of 23 Energy & Fuels

1
2
3
4
5 Biodiesel Production from Green Microalgae
6
7
8 Schizochytrium limacinum via in situ
9
10
11
Transesterification
12
13
14 Zheting Bi, † B. Brian He,*, † and Armando G. McDonald‡
15
16 †
17
Department of Biological and Agricultural Engineering, College of Engineering, University of
18 Idaho, 875 Perimeter Drive MS 2060, Moscow, Idaho 83844, United States
19
20
21 ‡Department of Forest, Rangeland, and Fire Sciences, College of Natural Resources, University of
22
23 Idaho, 875 Perimeter Drive MS 1132, Moscow, Idaho 83844, United States
24
25 KEYWORDS: microalgae, in situ transesterification, biodiesel, subcritical methanol, supercritical
26
27 methanol
28
29
30 ABSTRACT:
31
32 Microalgae are considered as one of the most promising feedstocks for biofuel production because
33 of their environmental and social benefits. However, challenges exist in converting microalgal
34
35 lipids into algal biofuels due to the unique characteristics of microalgae and the technologies for
36
37 processing them. This study aims at exploring an alternative sub-/super-critical methanol
38 (subCM/SCM) process technology that combines lipid extraction from whole microalgae and lipid
39
40 esterification/transesterification in a single step or in situ transesterification. A high lipid content
41
42 microalgal strain, Schizochytrium limacinum, was used for in situ transesterification in a batch
43 reactor. Temperature (170, 210, 250, and 290˚C), reaction time (30, 60, 90 and 120 min), and
44
45 lipid-to-methanol molar ratio (sRatio; 1:50, 1:75, and 1:100) were investigated for their effects on
46
47
the conversion efficiency. Temperature appeared as a most positive influential factor. Additionally,
48 operating temperature over 250˚C caused degradation of the lipids and/or algal biomass thus led
49
50 to the decline of the ester yield. The combination of reaction time and temperature had a
51
52
significant impact on the in situ transesterification reaction. The sRatio had a statistically
53 significant impact on the product yield and purity, and both these two responds factors reached
54
55 the maximum levels after sRatio reached 1:75. It was observed that the highest product purity
56
(37.5 wt%) occurred at sRatio 1:75, 211.6˚C, 120 min, with a product yield of 58.4 mol%. This
57
58
59
60 1
ACS Paragon Plus Environment
 Energy & Fuels Page 2 of 23

1
2
3 study shows that the in situ transesterification of microalgae bears some advantages over the
4
5 traditional two-step processes and has the potential to be applied to large-scale processing for
6
7 biodiesel production.
8
9
10
11
12 INTRODUCTION:
13
14 Microalgae are considered as one of the most promising feedstocks for biofuel production
15
16
because of their environmental and social benefits. However, challenges exist in converting
17 microalgal lipids into algal biofuels due to the unique characteristics of microalgae and the
18
19 technologies for processing them. In the past decades, extensive research has been conducted
20
on the selection of microalgal strains, cultivation of various microalgal strains for high lipid
21
22 production, and bioreactor technologies for microalgae production.1–5 Studies had also focused
23
24 on producing biodiesel from microalgae via various approaches. Most of these studies, however,
25
have used the traditional two-step procedure, i.e., solvent extraction to remove the algal lipids
26
27 from the microalgae biomass first, and lipid transesterification of the extracted lipids in a
28
29 separate unit.6–9 Two-step processes have many limitations including the increased processing
30
cost due to the additional purification process. Besides, strong acid or alkaline catalysts are
31
32 commonly used in two-step transesterification that catalyst disposal will harm the environment
33
34 in some extant.10–12 Therefore, to avert the drawbacks of two-step catalytic transesterification
35
process, in situ transesterification was studied.
36
37
38 Traditionally, in situ transesterification in literature refers to a biodiesel production from oil
39
40 seeds, with or without a catalyst, without conducting the oil extraction beforehand.4,13,14 In an in
41 situ transesterification, oil extraction and oil transesterification are performed simultaneously
42
43 in a single step, which simplifies the processing. So far, biodiesel production from microalgae
44
45 requires either co-solvent (i.e., chloroform, hexane, water) and/or catalyst assistance.5,15–20 In
46 situ transesterification of microalgae for biodiesel production has been investigated by several
47
48 researchers.17,21–23 For example, in situ transesterification of microalgae, Chlorella vulgaris, for
49
50
biodiesel production using NaOH as the catalyst achieved a maximum biodiesel yield of 77.6 wt%
51 using a catalyst to lipid ratio of 0.15:1 and a methanol to lipid molar ratio of 600:1 at 60°C for
52
53 75 min.9 In contrast, an acid-catalyzed in situ transesterification achieved up to 96.8 wt% of
54
55
yield with a catalyst (H2SO4) to lipid ratio of 0.35:1 and methanol to lipid molar ratio of 800:1 at
56 60°C for 20 h.9 Wahlen et al.8 studied the catalytic in situ transesterification of five microalgal
57
58 species, i.e., Chaetoceros gracilis (diatom), Tetraselmis suecica, Chlorella sorokiniana (green
59
60 2
ACS Paragon Plus Environment
Page 3 of 23 Energy & Fuels

1
2
3 algae), Synechocystis sp. PCC 6803 and Synechococcus elongates (cyanobactrium). They believed
4
5 the key parameters, which are essential to the successful conversion of algal lipids to biodiesel
6
7 by in situ transesterification, were alcohol type, alcohol to biomass ratio, temperature and
8 catalyst concentration. They achieved the product yields (percent FAME of extractable lipid) of
9
10 82% (C. gracilis), 78% (T. suecica), 77% (C. sorokiniana), 39% (Synechocystis sp. PCC 6803), and
11
12 40% (S. elongates)), with lyophilized microalgae (100 mg reacting with 2 mL of methanol
13 containing 1.8% (v/v) H2SO4 for 20 min at 80˚C). They also experimented on producing FAME
14
15 from C. gracilis with a water content of 50% (w/w) using the same procedure and obtained a
16
17 FAME yield of 84%. More recently, Tran, et al. demonstrated an innovative approach of using
18 wet microalgae with up to 90 %wt water as the feedstock to produce biodiesel with a co-
19
20 solvent.2,24 In their studies an excessive amount of methanol and immobilized lipase were
21
22 employed and achieved a high biodiesel conversion of 97.3% when using a wet microalga, C.
23 vugaris (lipid content 63 wt%), in the presence of 71% water.2,24 A combination of acid and base
24
25 catalysts was found most effective on converting microalgal lipids with a high content of free
26
27 fatty acids (FFA) via in situ transesterification. Work by Dong et al. using a two-step in situ
28 process was investigated to obtain a high FAME yield (94.9%) from microalgae (C. sorkiniana
29
30 UTEX 1602) that had high FFA content of 46.9% over the total lipids.17 This was accomplished
31
32 with a pre-esterification process using heterogeneous lipase catalyst (Amberlyst-15) to reduce
33 the FFA content prior to the base-catalyzed (KOH) transesterification.
34
35
36 Several researchers have also focused on converting the whole microalgae biomass (i.e., the
37
lipid and the cellulosic components in microalgae) into a liquid fuel through liquefaction
38
39 processes, but such a liquid fuel has no direct applications in the real life due to the complex
40
41 composition of the products before cost effective downstream separation/purification are
42 developed.25–28 Therefore, to take advantages of the microalgal lipids, the high quality energy
43
44 components, a study of targeting at producing FAME without converting the lignocellulosic and
45
46 other matters is desired. The transesterification of triglycerides via supercritical fluid (SCF) has
47 a greater strength in comparing to catalytic transesterification in biodiesel production. In this
48
49 non-catalytic process, the reactants contained in the microalgal biomass react with supercritical
50
51 alcohol condition to produce biodiesel, which simplifies the process relatively and make it
52 potentially cost-effective by avoiding the lipid extraction. The primary obstacle of triglycerides
53
54 transesterification is that they are immiscible with alcohols. During supercritical fluid
55
56 treatment, the solvent ability of alcohol is increased substantially so that a homogeneous
57 mixture is formed, which leads to higher reaction rate.29 Under the SCF conditions, various
58
59
60 3
ACS Paragon Plus Environment
 Energy & Fuels Page 4 of 23

1
2
3 properties of the fluid are placed between those of a gas and of a liquid. Under SCF condition,
4
5 the density of a supercritical fluid is similar to a liquid, its viscosity is similar to a gas, and its
6
7 diffusivity is between the two states. Because of its different physicochemical properties, SCF
8 has been used in several industrial applications, such as supercritical extraction. Due to their
9
10 low viscosity and relatively high diffusivity, supercritical fluids can penetrate easily through
11
12 solid materials therefore lead to a better extraction performance. One of the main
13 characteristics of a supercritical fluid is the possibility of modifying the density of the fluid by
14
15 changing its pressure and/ or its temperature. Since density is directly related to solubility, by
16
17 altering the pressure, the solvent strength of the fluid can be manipulated and its extraction
18 efficiency, in terms of higher extract yields and less extraction times, is much higher than
19
20 general solvent extraction.30 Tan et al. combined non-catalytic transesterification with
21
22 supercritical methanol (SCM) and investigated the impact of several parameters. A methyl
23 esters yield of 80% was achieved at 350˚C, 20 min reaction time, and alcohol to oil ratio of 40.31
24
25 In another study by Warabi et al., a catalyst-free transesterification was used on converting
26
27 rapeseed oil to biodiesel via supercritical alcohols. They concluded that the reaction rate of
28 triglycerides transesterification was slower than that of alkyl esterification of FFAs for any
29
30 alcohols employed.32 Kusdiana et al. also concluded that the reaction rate of triglycerides
31
32 transesterification was dramatically increased in the SCM state as compared to sub-critical
33 methanol (subCM) state.33
34
35
36 Keeping in mind the objective of simplifying the microalgae to biodiesel process under non-
37
catalytic condition, this study was conducted to develop and validate a process for obtaining
38
39 algal biodiesel directly from microalgae in a single step, i.e., in situ transesterification. Methanol
40
41 under subCM or SCM conditions were used as the solvent to simultaneously extract lipids and
42 esterify/transesterify FFA/triglycerides in microalgae to biodiesel. The effects of process
43
44 parameters on the process efficiency were explored. A statistical analysis of selected
45
46 parameters was conducted and used to predict the optimum conditions which can predict the
47 highest product yield and productivity for use in scaled-up operations. Furthermore, Fourier
48
49 transform infrared (FTIR) spectroscopic analysis was also performed on the chemical
50
51 composition of the products from various operation conditions and the degree of biomass
52 liquefaction.
53
54
55
56
57
58
59
60 4
ACS Paragon Plus Environment
Page 5 of 23 Energy & Fuels

1
2
3
4 METHODOLOGY:
5
6
Microalgae and chemical reagents
7
8 A lipid-rich green microalga Schizochytrium limacinum was obtained from ENN Energy Service
9
10 Co., Ltd (Langfang, China) as a dry powder biomass with 55 wt% lipid content, and used in all in
11
situ transesterification experiments in this study. The microalgae biomass was manually ground
12
13 in mortar to a relative uniformity in particle size (10-20 μm) for subsequent use.
14
15 Characterization was performed on the S. limacinum biomass, including proximate analysis,
16
ultimate analysis and fatty acid (FA) composition. Fatty acid profile of the lipids was analyzed
17
18 by gas chromatography-mass chromatography (GC-MS) as described by Hammond.34 It was
19
20 determined that the microalgal biomass contains 1.6 wt% of moisture, 78.5 wt% of volatile
21
matter, 8.9 wt% of fixed carbon and 11.0 wt% of ash content on dry basis. Elementally, the S.
22
23 lmacinum biomass consists of 61.4 wt% C, 8.9 wt% of H, 19.5 wt% of O, 4.0 wt% of N, and 0.6
24
25 wt% of S on dry basis. The gross heating value was calculated as 29.7 MJ/kg, based on the
26
formula proposed by Channiwala and Parikh.35
27
28
29 The samples for FA profile determination was obtained by extracting the S. limacinum lipids
30
31 using chloroform/methanol/water mixture at 1:1:0.9 volume ratio by following the procedure
32 in literature.14 The lipid samples were washed using warm water before analyzed. The FA profile
33
34 of the S. limacinum lipids, as determined via GC-MS (ThermoQuest Polaris Q instrument; ZB1
35
36 capillary column (30 m x 0.25 mm ID) by Phenomenex; 40°C (1 min) to 320°C at 5°C/min)36 ,
37 was 2.8 wt% of myristic acid (C14:0), 43.6 wt% of palmitic acid (C16:0), 1.3 wt% of stearic acid
38
39 (C18:0), 5.2 wt% of oleic acid (C18:1), 12.7 wt% of linoleic acid (C18:2), 1.8 wt% of linolenic
40
41
acid (C18:3), 25.4 wt% of docosahexaenoic acid (C22:6), and 6 wt% of unknown acid (possibly
42 clupanodonic acid, C22:5s). Furthermore, the analysis, according to AOCS official method Ca 5a-
43
44 40 37, indicated that the extracted microalgal lipids contained 16.6 wt% of FFA.
45
46 All chemicals used in this research were HPLC grade (n-hexane from Fisher Scientific
47
48 (Pittsburgh, Pa.); chloroform and methanol from EMD Millipore (Darmstadt, Germany) and
49
50 Macron Fine Chemicals (Center Valley, Pa.).
51
52 Experimental design
53
54
55 A response surface methodology (RSM) was used to determine the optimal operation condition
56
57
for maximizing the yield of FAME. The experimental design was based on a 23 full factorial
58
59
60 5
ACS Paragon Plus Environment
 Energy & Fuels Page 6 of 23

1
2
3 design and was constructed using Design-Expert v. 8.0 (Stat-Ease, Inc., Minneapolis, Minn.). The
4
5 process of in situ transesterification was investigated as a function of three factors, i.e.,
6
7 operating temperature, reaction time, and lipid-to-methanol molar ratio (sRatio). The
8 significance of the regression coefficients was determined at a p-value of 0.05. Microsoft Excel
9
10 was used for data processing.
11
12
13 Experiments and Procedures
14 All experiments were carried out in a 300-mL stirred batch reactor (Pressure Reactor 4560,
15
16 Parr Instrument Co., Moline, Ill.). The working volume of the reactant is approximately 150 mL
17
18 and the void space is approximately 150 mL controlled by a PC-based 4857 Reactor Controller
19 on temperature and agitation speed. A computer connected to the controller displays the
20
21 operating temperature, pressure, and agitation speed (500 rpm). In the preliminary
22
23 experiments, the reaction system was operated under SCM (250°C, 10.3 MPa) and subCM
24 (210°C, 7.6 MPa) conditions. Two levels of reaction time, i.e., 30 and 60 min, were employed.
25
26 The reactor was immediately immersed in to a cold water bath in order to stop the reaction
27
28 after the set reaction period. The final liquid slurry product was separated by centrifugation
29 (2500 rpm for 15 min) to obtain a liquid and solid biomass fractions. Liquid product was
30
31 concentrated to dryness to remove methanol. At last, condensed liquid product was collected
32
33 for GC-MS and FTIR analysis. Six sRatio of 1:50, 1:75, 1:100, 1:250, 1:300, and 1:350 were
34 investigated which are approximately equivalent to mass ratios of 1:1, 1:2, 1:3, 1:5, 1:6 and 1:7,
35
36 respectively. At the beginning of each experiment, CO2 was used to purge the system. After the
37
38 reaction, the oily products were collected and analyzed by GC-MS for the FAME content.31
39
40 A range of operation parameters was obtained from preliminary experiments, a systematic
41
42 investigation on the effects of selected operation factors was conducted. Experiments were
43 performed based on a factorial experimental design that included three numeric factors and
44
45 one categorical factor. The selected numeric factors are three levels of sRatio (i.e., 1:50, 1:75
46
47 and 1:100), four levels of operating temperature (i.e., 170, 210, 250, and 290°C), four levels of
48 reaction time (i.e., 30, 60, 90, and 120 min) with a constant 1.4 MPa gauge CO2 input to clean-up
49
50 the oxygen residue. According to this design, 48 sets of experiments were conducted and each
51
52 experiment was implemented in triplicates for a total of 144 experiments.
53
54
55
56
57
58
59
60 6
ACS Paragon Plus Environment
Page 7 of 23 Energy & Fuels

1
2
3
4
Response Factors
5
6 Two responds factors were used to reflect the effects of various operation condition on the in
7
8 situ transesterification, i.e., product yield (mol%) and product purity (wt%), based on the
9 following definitions:
10
11

   
  
12 Product yield (mol%) = [×
    
      ] × 100%
13
14 (Eq. 1)
15
16
   
  
Product purity (wt%) = [ ]× 100% (Eq. 2)
17
  
18
19
20
21
Spectroscopies on Microalgal Biomass
22
23
24 Spectroscopic analysis of Fourier transform infrared (FTIR) was performed on an Avatar 370
25
26 spectrophotometer (Thermo Nicolet) with an attenuated total reflection (ATR) probe (ZnSe
27 crystal). The spectra were ATR and baseline corrected using Omnic v9.0 software.
28
29
30
31
32
33 RESULTS AND DISCUSSION:
34
35
Biodiesel properties due to algal lipid composition
36
37 Fatty acid profiles of microalgal lipids are different among different algal species. In general,
38
39 microalgal lipids contain fatty acids that possess two unique signatures of relatively longer
40
chain length and high unsaturation. Microalgae have a common FA chain length from C12 to
41
42 C22, which is a crossover with typical vegetable oil FA ranges of C14 to C20, and a
43
44 polyunsaturated FA range of C20 to C22.38–40 Lipid composition of S. limacinum presents most of
45
the fatty acids (i.e., C14, C16 and C18) in typical vegetable oils but with a higher portion of
46
47 polyunsaturated fatty acids. C16:0 is the most abundant saturated FA in S. limacinum lipids, and
48
49 C22:6 is the highest among the polyunsaturated fatty acids. FA profile can directly affect
50 derived fuel properties. The longer chain length and higher saturation of a FA leads to a higher
51
52 viscosity of the corresponding FAME. Therefore, the combination of the long chain and the level
53
54 of unsaturation in FAs would decide the viscosity of a microalgal biodiesel. However, the
55 unsaturated FA is a concern due to their susceptibility to oxidation.41 Besides, microalgal lipids
56
57 used in this study contain even more polyunsaturated fatty acids (approximately 50%) than
58
59
60 7
ACS Paragon Plus Environment
 Energy & Fuels Page 8 of 23

1
2
3 those found in typical vegetable oils (typically in the range of 15%-35%), thus the algal
4
5 biodiesel made from such lipids tends to be oxidatively unstable. Fortunately, cold flow and
6
7 oxidative stability are two interacting factors controlled by the chain length and the level of
8 unsaturation of FAs. Long-chain saturated FAs cause poor clod flow property of corresponding
9
10 biodiesel but promise a decent oxidative stability property. Polyunsaturated FAs are more
11
12 reactive than saturated FAs due to the carbon-carbon double bonds that are easily to open up
13 and react with alcohols. In contrast, a good portion of microalgal lipids is polyunsaturated
14
15 which counter balances the poor cold flow property caused by the longer chain FAs.
16
17
18 Preliminary experiments of in situ transesterification
19
20
21 Microalgal lipids are mainly composed of triglycerides and some FFAs and phospholipids. In
22
23 this study, the microalgal lipids were extracted and converted to FAME (or biodiesel) by
24 transesterification with subCM/SCM without the addition of catalyst. This process is referred as
25
26 the in situ transesterification in this study. The ideal operation of such an in situ
27
28 transesterification is to carefully control the operating conditions so that only the microalgal
29 lipids are extracted and converted to FAME while keeping the microalgal cellular structure
30
31 intact. If severe operating conditions are employed then the whole microalga material is
32
33 liquefied and the process is then referred as thermal liquefaction, which is beyond the scope of
34 this study.7,17,42
35
36
37 Preliminary experiments used two temperatures (210˚C and 250˚C) close to the supercritical
38
point of methanol (240˚C and 8.1 MPa), 5 levels of sRatio (50, 75, 100, 250 and 300), and two
39
40 reaction times (30 and 60 min). The results showed that the product yield depends on the
41
42 combined effect of operating temperature and the reaction time at high sRatio. A spike of
43 product yield occurred at 210˚C with a reaction time of 60 min. Within a constant reaction time
44
45 of 60 min and >210˚C gave a negative result in the FAME yield (Figure 1).
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60 8
ACS Paragon Plus Environment
Page 9 of 23 Energy & Fuels

1
2
3
4 100%
5
90% 84
6
7 80% 76
8
9 70% 67
Product yield (mol%)

10
11 60% 170˚C, 6.1 Mpa
53
12
48 210˚C, 7.6 Mpa
13 50%
14 250˚C, 10.3 Mpa
15 40% 34
16 290˚C, 13.8 Mpa
17 30%
18 18
20%
19
10
20 10%
21
22 0%
23 sRatio 1:250 sRatio 1:300
24
25
26
27 Figure 1. Product yields of microalgal FAME at two levels of sRatio and 60 min reaction time.
28
29 There was a noticeable difference in the Figure 1 while sRatio rose from 1:250 to 1:300
30
31 (equivalent to algae to methanol mass ratio of 1:5 and 1:6, respectively), thus the higher sRatio
32
33
(1:350) was selected to further test the effect of sRatio levels. Besides, reaction time has been
34 pointed as a core parameter of transesterification.43,44 The results (Figure 2) showed no
35
36 significant impact of sRatio beyond 1:300 regarding to the product yield. Although, for S.
37
38
limacinum a sRatio of 1:300 is satisfactory in driving the in situ esterification/transesterification
39 to completion, this reaction condition. However, at this high sRatio a high amount of methanol
40
41 is required and can make the process uneconomic and therefore, a lower sRatio is required.
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60 9
ACS Paragon Plus Environment
 Energy & Fuels Page 10 of 23

1
2
3
4 100%
5
90% 84 85
6
7 79 77
80% 76
8 74
9 68 66
70%
Product yield (mol%)

10
11 60% 210˚C, 7.6 MPa, 30min
53 53 53 52
12
13 50% 210˚C, 7.6 MPa, 60min
14 250˚C, 10.3 MPa, 30min
15 40%
16 250˚C, 10.3 MPa, 60min
17 30%
18
20%
19
20 10%
21
22 0%
23 sRatio 1:250 sRatio 1:300 sRatio 1:350
24
25
26
27
Figure 2. Product yields of microalgal FAME at three levels of sRatio.
28
29
30
31 Supercritical CO2 (ScCO2) is an excellent solvent for lipid extraction and has been used in many
32
33 applications.45–47 It has been shown that ScCO2 and have similar behavior to the conventional
34 co-solvents, such as hexane or ethyl acetate.48 Furthermore, glycerol is almost insoluble in
35
36 ScCO2 and will precipitate out of methanol upon formation.45,49 Triglycerides are moderately
37
38 soluble in methanol while FAME product is highly soluble. Therefore, the property of FAME
39 tends to dissolve in ScCO2 and drive the reaction equilibrium forward to FAME formation. The
40
41 supercritical conditions of CO2 are 304K (31°C) and 7.4 MPa. It is often used with methanol or
42
43 ethanol as the co-solvent in supercritical processing.45,50 In this study, CO2 was applied to
44 enhance lipid extraction and transesterification reaction.
45
46
47 In situ transesterification of microalgae requires a higher alcohol to oil ratio mainly because
48
methanol serves as a solvent first then a reactant. Oil to alcohol molar ratio range of 1:100 to
49
50 1:800 with acid/base catalysts was commonly selected in microalgae in situ transesterification
51
52 studies.9,51 Meanwhile, other researchers studied in situ transesterification with physical
53
method aid, such as microwave. One study worked on base catalyzed transesterification of
54
55 microalgae lipids with microwave heating assistant. They operated this process with
56
57 Nannochloropsis sp. and Tetraselmis sp. (biomass and oil to methanol molar ratio of 1:12, 50˚C
58
59
60 10
ACS Paragon Plus Environment
Page 11 of 23 Energy & Fuels

1
2
3 and 16 h) afforded a maximum yield of 83.3% and 77.1%, respectively.52 Because of all the
4
5 evidence from previous studies show that in situ transesterification of microalgae can be done
6
7 at oil to alcohol molar ratio close to the stoichiometric level. Therefore, some lower levels of
8 sRatio, i.e., 1:50, 1:75, 1:100 and 1:250, were selected to investigate the effects of sRatio.
9
10
11 Experimental results indicated that the sRatio < 1:250 does not show a clear effect on product
12
13
yield (Figure 3). When the sRatio was increased to 1:250, the product yield reached its highest
14 value at 210˚C. In contrast, product yields showed a decreasing trend with a lower sRatio of
15
16 1:75 at 250˚C. As shown in Figure 2, the temperature is a more critical parameter than sRatio.
17
18
From lipid extraction and physical operation point of view, sRatio of 1:50 to 1:100 are
19 preferable because low methanol application than 1:50 would not make a good microalgae-
20
21 methanol suspension and limit the mass transfer of microalgal lipids into the liquid phase for
22
reaction. Additionally, the interaction of temperature and sRatio affected the results of product
23
24 yields due to a turning point at sRatio 1:75 to 1:100. Between sRatio of 1:100 and 1:250,
25
26 reaction temperature seems to be a more important factor since the product yield at sRatio
27
1:250 has dropped at 250˚C. Furthermore, literature suggests that a sRatio > 1:42 is suitable for
28
29 microalgae biodiesel transesterification.53 Therefore, in the second stage of this study, sRatio of
30
31 1:50-1:100 were chosen for further investigation.
32
33
34 100%
35
36 78
37
38 75% 69
Product yield (mol%)

39 63
40 56 58
53 sRatio 1:50
41 49
42 50% sRatio 1:75
43 37 sRatio 1:100
44
45 sRatio 1:250
46 25%
47
48
49
50 0%
51 210˚C 250˚C
52
53
54 Figure 3. sRatio effect on product yields at 60 min reaction time.
55
56
57
58
59
60 11
ACS Paragon Plus Environment
 Energy & Fuels Page 12 of 23

1
2
3
4
Systematic investigation of process parameters
5
6 Preliminary experiments revealed that in situ transesterification of the microalga is achievable
7
8 in methanol around its critical point. However, the complexity of the microalgal properties,
9 namely the physical and chemical properties of microalgal lipids versus the microalgal
10
11 cellulosic biomass, implies that the process operating parameters need to be systematically
12
13 investigated in order to achieve satisfactory product yield and purity which are the valuable
14 outcomes in evaluating the efficiency of the process from the angle of engineering and
15
16 technological development.
17
18
19 The objective of this systematic investigation was to find the suitable operating conditions at
20 which the highest product yield and/or purity can be achieved. Experiments were performed
21
22 based on a 4×4×3 factorial experimental design that includes four levels of operating
23
24 temperatures (i.e., 170, 210, 250, and 290°C), four levels of reaction time (i.e., 30, 60, 90 and
25 120 min), and three levels of sRatio (i.e., 1:50, 1:75 and 1:100). All experiments were conducted
26
27 in triplicate to ensure the repeatability and reliability. Tables 3 and 4 summarize the
28
29 experimental results.
30
31 Close examination of the experimental results leads to some preliminary conclusions. First, the
32
33 product yield ranged widely from as low as 12% to as high as 67%, highly depending on the
34
35 operating conditions. It was observed that the operating temperature affects the product yield
36 significantly. Generally, the product yield increases as the temperature increases in the range of
37
38 170 to 250°C. Once the operating temperature reaches 290°C, the some product yields were
39
40 even lower than those at temperatures below 290°C. This might be due to the faster rate of
41 decomposition of the formed product FAME, which exceeds the rate of FAME formation from
42
43 esterification and transesterification. It was observed that the quantity of residual after
44
45 reaction varied widely depending on operating conditions. For example, at 250°C and 1:75
46 sRatio, the residual varies from 7.02 g to 4.25 g as reaction time changed from 30 min to 120
47
48 min; if the operating temperature was 210°C, the range was 11.4 g to 23.4 g. The biomass
49
50 residual indicates, in addition to the lipid transesterified, the decomposition of algal biomass
51 and affects the FAME purity in the product mixture. Therefore, FAME purity is used as an
52
53 indicator indirectly for biomass conversion/ decomposition.
54
55
56
57
58
59
60 12
ACS Paragon Plus Environment
Page 13 of 23 Energy & Fuels

1
2
3 Table 1. Product yield (mol%) and product purity (wt%) of FAME via in situ transesterification with 1.4 MPa initial CO2.
4
5
6 sRatio 1:50 sRatio 1:75 sRatio 1:100
Time
7
(min)
Product yield (mol%)

8 170°C 210°C 250°C 290°C 170°C 210°C 250°C 290°C 170°C 210°C 250°C 290°C
9
10
30 16.5±5.7 24.7±12.7 48.6±10.4 27.6±13.1 10.6±3.1 27.3±11.4 62.7±11.1 59.2±11.4 9.3±11.1 28.6±7.6 60.6±13.1 53.9±9.1
11
12
13 60 16.7±3.2 39.6±13.8 39.4±12.1 26.8±12.2 20.7±3.4 41.9±9.9 60.5±9.2 38.5±9.7 25.9±14.7 48.3±9.1 58.7±15.2 46.6±8.7
14
15 90 26.4±4.7 44.4±12.3 36.4±11.1 26.7±13.1 34.9±2.1 46.2±12.4 52.9±11.4 29.0±10.1 28.8±9.4 52.5±7.7 52.5±11.0 26.6±8.4
16
17 120 36.1±4.8 47.6±13.4 27.4±10.3 17.2±14.3 38.2±1.6 64.5±11.8 47.2±10.0 31.7±11.1 36.7±10.1 60.2±7.6 40.9±12.3 20.7±9.7
18
19
sRatio 1:50 sRatio 1:75 sRatio 1:100
20 Time
21 (min)
Product purity (wt%)

22 170°C 210°C 250°C 290°C 170°C 210°C 250°C 290°C 170°C 210°C 250°C 290°C
23
24 30 10.3±7.1 18.2±13.7 23.6±12.4 22.4±10.0 15.1±9.4 21.7±11.3 25.4±3.7 40.6±4.7 18.6±4.1 26.6±10.1 46.8±10.7 28.5±10.1
25
26 60 12.7±8.6 25.9±20.4 22.2±15.1 18.1±5.6 20.4±10.7 30.6±10.9 35.2±4.6 24.1±4.3 23.8±9.3 28.3±12.2 20.6±13.5 19.4±9.6
27
28
90 21.0±11.2 26.9±14.3 20.5±11.6 13.4±6.3 28.7±9.4 29.9±14.2 36.7±3.9 20.5±4.4 37.2±13.1 29.6±10.7 18.4±11.2 22.9±11.5
29
30
31 120 22.3±11.4 27.1±10.1 15.8±1.7 10.6±5.1 29.3±8.2 33.1±13.7 24.9±4.8 24.0±4.7 34.2±10.2 34.7±9.3 20.8±13.1 20.6±7.6
32
33
34
35
36
37
38
39
40
41
42
43
44 13
45
46 ACS Paragon Plus Environment
47
48
 Energy & Fuels Page 14 of 23

1
2
3 Meanwhile, reaction time and operating temperature showed a correlated effect on the product yield, with
4
5 a few exceptions. At lower temperatures (170°C and 210°C), the product yield showed an increasing trend
6
7 as the reaction time is in the range of 30 to 120 min. Once the operating temperature was >210°C,
8 especially at 290°C, the product yield decreased as the reaction time increased. Similar to the reason
9
10 discussed above, this decrease in product yield might be due to the decomposition of FAME at high
11
12 temperatures and extended reaction time. It seems that there is an optimum range of operating conditions
13 at 210˚C (120 min) and 250˚C (30 min), at which both the product yield and the product purity reached
14
15 their peak values, and then both rapidly declined beyond temperature of 250˚C. Hence, the optimal reaction
16
17 condition point ought to exist between these two points regardless of the effect of sRatio level.
18
19 The sRatio statistically exhibits a significant effect on the product yield. Under the same operating
20
21 temperature and reaction time, the product yields are approximately the same in sRatio range of 1:75 to
22
1:100. The highest product yield and purity happened at sRatio 1:75. This indicates that the quantity of
23
24 methanol at sRatio 1:75 and 1:100 are adequate for the purpose of lipid extraction and transesterification,
25
26 which are higher than that of 1:42 as reported in the literature.33,53 However, the results of product yield
27
and purity are also affected by the interaction of sRatio and other parameters. When sRatio beyond 1:75,
28
29 the effect of methanol quantity reached a plateau and did not benefit the transesterification reaction
30
31 significantly. Therefore, another conclusion is that methanol amount does not have an effect on FAME
32
purity beyond sRatio of 1:75.
33
34
35 Response surface methodology for producing microalgal biodiesel
36
37
38 A RSM model was used to statistically analyze and determining the optimum operation condition. Table 5
39 shows the ANOVA results which indicates that sRatio (C) in the 1st order, temperature (A) in the 2nd order
40
41 and an interaction (AB) of temperature and time (B) were shown to be the most significant variables.
42
43 Secondly, variable A is another variable that presents a significant effect on the results. The model also
44 included temperature in the 2nd order and time based on their significance (α = 0.05). The R-square of the
45
46 model was 0.85, which is acceptable to provide a decent prediction on FAME yield under optimal operation
47
48 conditions. Thus the modeling results could be concluded by the equations 3 and 4:
49
50 Yield = 51.42 + 4.44 × A + 1.31 × B - 8.60 × C + 5.41 × C2 - 13.10 × AB - 1.31 × AC + 0.71 × AC2 – 0.76 × BC +
51
52 2.48 × BC2 – 21.40 × A2 – 2.56 × B2 (Eq. 3)
53
54
55
56 14
57
58
59
60
ACS Paragon Plus Environment
Page 15 of 23 Energy & Fuels

1
2
3 Purity = 28.92 – 1.83 × A + 1.21 × B – 5.24 × C + 2.99 × C2 – 6.74 × AB + 1.02 × AC + 0.14 × AC2 – 0.95 × BC +
4
5 2.71 × BC2 – 7.28 × A2 – 0.34 × B2 (Eq.4)
6
7
8 With product yield and product purity as the related responds factors, RSM provides a direct visual
9 analysis on the overall results. Since the optimum operation condition was given at the sRatio of 1:75,
10
11 sRatio at 1:75 was fixed in the 3D plot (Figure 5A). It is seen that the highest product yield lay at the point
12
13
of 210˚C and 120 min. The 2D contour plot (Figure 5B) indicates the optimization result based on the
14 quadratic model at sRatio of 1:75. The maximum predicted product yield is 58.4 mol% and product purity
15
16 is 37.5 wt% at 210˚C, 120min, and sRatio 1:75.
17
18 Table 2. ANOVA table of product yield analysis including all four parameters.
19
20 Sum of p-value
21 Source df Mean of Square F Value
Squares Prob > F
22
23 Model 9444.55 11 858.60 1862 <0.0001
24 A -Temperature 525.70 1 525.70 10.68 0.0018
25 B -Time 46.11 1 46.11 1.48 0.3234
26 C - sRatio 1815.92 2 907.96 11.10 <0.0001
27 AB 2542.18 1 2542.18 65.40 <0.0001
28
AC 23.02 2 11.51 2.41 0.7800
29
30 BC 86.10 2 43.05 1.11 0.4016
31 A2 4343.41 1 4343.41 97.97 <0.0001
32 B2 62.11 1 62.11 0.080 0.2529
33 Residual 1656.27 36 46.01
34 Corrected Total 11100.82 47
35
R2 = 0.85, adjusted R2 = 0.81, df = degree of freedom, F-critical = 3.02
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56 15
57
58
59
60
ACS Paragon Plus Environment
 Energy & Fuels Page 16 of 23

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21 (A) (B)
22
23
24
Figure 4. (A): 3D plot of RSM for in situ transesterification product yield based on a 4×4×3 factorial design
25 (fixed sRatio at 1:75); (B): Contour plot of RSM of in situ transesterification product yield at fixed sRatio of
26 1:75.
27
28 Spectroscopic analysis
29
30 FTIR spectroscopy was used to further analyze chemical functional group changes of the in situ
31
32 transesterification material (Figure 5). The main characteristics are attributed to the presence of esters and
33
34 hydrocarbon content and protein. A strong hydrogen bond O-H stretching band appeared at 3348 cm-1 due
35 to the presence of protein and polysaccharides This band was also observed in the FTIR spectra of
36
37 microalgae hydrothermal liquefaction biocrude oil.26,36 The distinct peak at 1435 cm-1 is detected in the
38
39 spectra of product mixture from a reaction of 210˚C and 120 min. Similar to hydrothermal liquefaction
40 biocrude oil from Spirulina algae, greater intensity and higher resolution of bands of C-H stretching at
41
42 3000-2840 cm-1 appeared in all spectra which suggested a more highly aliphatic in character.26,54
43
44 Absorption ester bands of C=O stretching at 1741 cm-1 from FA, hydroxy FA, and diacids in lipids were also
45 detected in all product mixtures.25,26,28 Moreover, the characteristics of the FA, esters and carbohydrate
46
47 derivatives, C-O stretch at 1360 cm-1 and C-O alcohol stretch at 1196 cm-1 and 1168 cm-1 are also presented
48
49 in all spectra. Next to C=O stretching at 1741 cm-1, a low level of N-H bending presents at 1680-1600 cm-1
50 associated with amide and amine compounds shown in all spectra.
51
52
53
54
55
56 16
57
58
59
60
ACS Paragon Plus Environment
Page 17 of 23 Energy & Fuels

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
29
30
31 Figure 5. FTIR spectra of the FAME product from in situ transesterification at 210˚C and 120 min.
32
33
34
35 However, some distinct differences could also be observed when comparing the spectra of in situ
36
37 transesterification oil products from different operation conditions. Figure 6 compares the spectra
38
differences of biocrude oils from operation conditions of 210˚C and 290˚C at 120 min. The spectra clearly
39
40 indicate the esters yield loss at a higher reaction temperature. The strong ester band (C=O stretching) at
41
42 1739 cm-1 is significantly reduced. In contrast, O-H stretching at 3373 cm-1, C-O stretching at 1290 cm-1, C-O
43
alcohol vibration at 1196 cm-1 and 1170 cm-1, and N-H bending at 1661 cm-1 are amply increased. This
44
45 phenomenon reveals that more protein was reacted at 290˚C than at 210˚C and less esters formed which
46
47 could be explained by the possibility of degradation reactions occurred after transesterification. The esters
48
were further reacted with the components of the microalgal biomass and formed other compounds (i.e.,
49
50 amide, amine, hydroxyl group and aromatic compounds). In liquefaction reaction, carbonyl groups reduced
51
52 to the corresponding alcohol with formate ion and water that would reduce FAME yield. The degradation
53
mentioned in the paper to explain the product yield lost when temperature beyond 250°C referred to the
54
55
56 17
57
58
59
60
ACS Paragon Plus Environment
 Energy & Fuels Page 18 of 23

1
2
3
reduction of newly formed carbonyl group to the corresponding alcohol with formate ion and water. This
4
5 statement was also claimed in a work by Demirbas.55
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34 Figure 6. FTIR spectra of in situ transesterification product mixture at 210˚C and 290˚C for 120 min.
35
36
37
38
39 The effect of reaction time was examined by comparing the FTIR spectra of product mixture from the same
40 operation temperature but different reaction time (Figures 7 and 8). When the in situ transesterification
41
42 was operated at 210˚C for 30 min and 120 min, the compositions of product mixture was not significantly
43
44 changed, except the O-H stretching and C=O stretching had a slightly increase. This finding supports the
45 fact that at 210˚C, biomass liquefaction was not promoted with an extended time period. Figure 8 presents
46
47 the comparison of FTIR spectra of product mixture from reaction conditions of 30 min and 120 min at
48
49 250˚C. A dramatic change of the product composition was detected for different period of time at 250˚C. At
50 this temperature, biomass liquefaction became the dominate reaction at 120 min. Therefore, an interaction
51
52 of operation temperature and reaction time affected the process greatly, which is consistent with the
53
54 results of statistical analysis.
55
56 18
57
58
59
60
ACS Paragon Plus Environment
Page 19 of 23 Energy & Fuels

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
29
30
31
32
33
34
Figure 7. FTIR spectra of the FAME product from in situ transesterification 30 min and 120 min at 210˚C.
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56 19
57
58
59
60
ACS Paragon Plus Environment
 Energy & Fuels Page 20 of 23

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
29
30
31
32
33
34 Figure 8. FTIR spectra of the FAME product from in situ transesterification for 30 min and 120 min at
35 250˚C.
36
37
38
39
40 CONCLUSION:
41
42
43 The experimental results showed that the direct conversion of algal lipids of microalga S. limacinum can be
44
achieved in one step. The subCM/SCM in situ transesterification without catalyst application has significant
45
46 results with respecting to product yield and purity. The actual effects on product yield are collectively
47
48 contributed by the operating temperature, which determines the fluid status of methanol, and sRatio. The
49
effect of reaction time was determined not significant but affected the system by its interaction with
50
51 operating temperature. Another important response factor is the purity of the targeted microalgal methyl
52
53 esters or FAME. Experimental data on product selectivities lead to similar conclusions as those observed on
54 product yield. Generally, the product purity increases as the operating temperature increases in the range
55
56 20
57
58
59
60
ACS Paragon Plus Environment
Page 21 of 23 Energy & Fuels

1
2
3 of 170-250˚C. A clear trend was noticed that when the operating temperature was at 290˚C, the selectivities
4
5 were generally lower than those at lower temperatures. The FTIR spectroscopy supported the statistical
6
7 analysis results in that formed FAME products may degraded and react with other algae components at
8 high temperature (> 250˚C) and long reaction time (>90 min). Overall, an optimum operation condition for
9
10 in situ transesterification was determined at 210˚C, 120 min with sRatio 1:75 and the highest product yield
11
12 was 58.4 mo% and product purity of 37.5 wt%. The advantage of such a proposed one-step, in situ
13 transesterification without addition of catalysts is to simplify the microalgae processing process and
14
15 overcome the technological challenges in traditional processes thus to enhance the processing efficiency
16
17 for viable algal biodiesel production.
18
19
20
21
22 ACKNOWLEDGEMENTS:
23
24
25 The authors gratefully acknowledge the technical and financial support from the National Institute for
26 Advanced Transportation Technology (NIATT) via University Transportation Centers, US Department of
27
28 Transportation (award No. DTRT12GUTC17), National Institute of Food and Agriculture (NIFA), US
29
30 Department of Agriculture, via hatch grant, and the Department of Biological and Agricultural Engineering at
31 the University of Idaho. The authors also express their sincere gratitude to Mr. Keegan Duff, engineering
32
33 support scientist, for his assistance in this study.
34
35
36
37
38 REFERENCE:
39
40
41 (1) Xu, R.; Mi, Y. J. Am. Oil Chem. Soc. 2010, 88 (1), 91–99.
42 (2) Tran, D.-T.; Yeh, K.-L.; Chen, C.-L.; Chang, J.-S. Bioresour. Technol. 2012, 108, 119–127.
43
44 (3) Sani, Y. M.; Daud, W. M. A. W.; Abdul Aziz, A. R. J. Environ. Chem. Eng. 2013, 1 (3), 113–121.
45
46 (4) Miao, X.; Wu, Q. Bioresour. Technol. 2006, 97 (6), 841–846.
47 (5) Huang, G.; Chen, F.; Wei, D.; Zhang, X.; Chen, G. Appl. Energy 2010, 87 (1), 38–46.
48
49 (6) Yen, H.-W.; Hu, I.-C.; Chen, C.-Y.; Ho, S.-H.; Lee, D.-J.; Chang, J.-S. Bioresour. Technol. 2013, 135, 166–174.
50
51 (7) Williams, P. J. le B.; Laurens, L. M. L. Energy Environ. Sci. 2010, 3 (5), 554.
52 (8) Wahlen, B. D.; Willis, R. M.; Seefeldt, L. C. Bioresour. Technol. 2011, 102 (3), 2724–2730.
53
54 (9) Velasquez-Orta, S. B.; Lee, J. G. M.; Harvey, A. Fuel 2012, 94, 544–550.
55
56 21
57
58
59
60
ACS Paragon Plus Environment
 Energy & Fuels Page 22 of 23

1
2
3 (10) Chew, T. L.; Bhatia, S. Bioresour Technol 2008, 99 (17), 7911–7922.
4
5 (11) Lee, J. S.; Saka, S. Bioresour Technol 2010, 101 (19), 7191–7200.
6
7 (12) Lam, M. K.; Lee, K. T.; Mohamed, A. R. Biotechnol. Adv. 2010, 28 (4), 500–518.
8 (13) Halim, R.; Gladman, B.; Danquah, M. K.; Webley, P. A. Bioresour. Technol. 2011, 102 (1), 178–185.
9
10 (14) Bi, Z. T.; He, B. B. Trans. ASABE 2013, 56 (4), 1529–1539.
11
12 (15) Amaro, H. M.; Guedes, A. C.; Malcata, F. X. Appl. Energy 2011, 88 (10), 3402–3410.
13 (16) Amalia Kartika, I.; Yani, M.; Ariono, D.; Evon, P.; Rigal, L. Fuel 2013, 106, 111–117.
14
15 (17) Dong, T.; Wang, J.; Miao, C.; Zheng, Y.; Chen, S. Bioresour. Technol. 2013, 136, 8–15.
16
17 (18) Haas, M. J.; Scott, K. M.; Foglia, T. A.; Marmer, W. N. J. Am. Oil Chem. Soc. 2007, 84 (10), 963–970.
18 (19) Chisti, Y. Biotechnol. Adv. 2007, 25 (3), 294–306.
19
20 (20) Gouveia, L.; Oliveira, A. C. J Ind Microbiol Biotechnol 2009, 36 (2), 269–274.
21
22 (21) Griffiths, M. J.; van Hille, R. P.; Harrison, S. T. Lipids 2010, 45 (11), 1053–1060.
23 (22) Patil, P. D.; Reddy, H.; Muppaneni, T.; Schaub, T.; Holguin, F. O.; Cooke, P.; Lammers, P.;
24
25 Nirmalakhandan, N.; Li, Y.; Lu, X.; Deng, S. Bioresour Technol 2013, 139, 308–315.
26
27 (23) Velasquez-Orta, S. B.; Lee, J. G. M.; Harvey, A. P. Biochem. Eng. J. 2013, 76, 83–89.
28 (24) Tran, D.-T.; Chen, C.-L.; Chang, J.-S. Bioresour. Technol. 2013, 135, 213–221.
29
30 (25) Yuan, X.; Wang, J.; Zeng, G.; Huang, H.; Pei, X.; Li, H.; Liu, Z.; Cong, M. Energy 2011, 36 (11), 6406–6412.
31
32 (26) Huang, H.; Yuan, X.; Zeng, G.; Wang, J.; Li, H.; Zhou, C.; Pei, X.; You, Q.; Chen, L. Fuel Process. Technol.
33 2011, 92 (1), 147–153.
34
35 (27) Vardon, D. R.; Sharma, B. K.; Scott, J.; Yu, G.; Wang, Z.; Schideman, L.; Zhang, Y.; Strathmann, T. J.
36
37 Bioresour. Technol. 2011, 102 (17), 8295–8303.
38 (28) Minowa, T.; Yokoyama, S.; Kishimoto, M.; Okakurat, T. Fuel 1995, 74 (12), 1735–1738.
39
40 (29) Tan, K.; Lee, K. T. Renew. Sustain. Energy Rev. 2011, 15, 2452–2456.
41
42
(30) Herrero, M.; Cifuentes, A.; Ibañez, E. Food Chem. 2006, 98 (1), 136–148.
43 (31) Tan, K. T.; Lee, K. T.; Mohamed, A. R. Energy 2011, 36 (4), 2085–2088.
44
45 (32) Warabi, Y.; Kusdiana, D.; Saka, S. Bioresour. Technol. 2004, 91 (3), 283–287.
46
47
(33) Kusdiana, D.; Saka, S. Fuel 2001, 80 (5), 693–698.
48 (34) Hammond, E. G. In Essential Oils and Waxes; Linskens, P. D. H. F., Jackson, P. D. J. F., Eds.; Modern
49
50 Methods of Plant Analysis; Springer Berlin Heidelberg, 1991; pp 321–330.
51
(35) Channiwala, S. A.; Parikh, P. P. Fuel 2002, 81 (8), 1051–1063.
52
53 (36) Liang, S.; McDonald, A. G. J. Agric. Food Chem. 2014, 62 (33), 8421–8429.
54
55
56 22
57
58
59
60
ACS Paragon Plus Environment
Page 23 of 23 Energy & Fuels

1
2
3 (37) AOCS. AOCS Offical Method Ca 5a-40: Free Fatty Acids.
4
5 (38) Mata, T. M.; Martins, A. A.; Caetano, N. S. Renew. Sustain. Energy Rev. 2010, 14 (1), 217–232.
6
7 (39) Verma, M. K.; Anand, R.; Chisti, A. M.; Kitchlu, S.; Chandra, S.; Shawl, A. S.; Khajuria, R. K. J. Essent. Oil
8 Bear. Plants 2010, 13 (3), 331–335.
9
10 (40) Harun, R.; Singh, M.; Forde, G. M.; Danquah, M. K. Renew. Sustain. Energy Rev. 2010, 14 (3), 1037–1047.
11
12 (41) Razeghifard, R. Photosynth Res 2013, 117 (1-3), 207–219.
13 (42) Yang, Y. F.; Feng, C. P.; Inamori, Y.; Maekawa, T. Resour. Conserv. Recycl. 2004, 43 (1), 21–33.
14
15 (43) Issariyakul, T.; Dalai, A. K. Renew. Sustain. Energy Rev. 2014, 31, 446–471.
16
17 (44) Valizadeh Derakhshan, M.; Nasernejad, B.; Dadvar, M.; Hamidi, M. Asia-Pac. J. Chem. Eng. 2014, 9 (5),
18 629–637.
19
20 (45) Soh, L.; Zimmerman, J. Green Chem. 2011, 13 (6), 1422.
21
22 (46) Breet, E.; Nortjé, Y.; van Greuning, C. J. Supercrit. Fluids 2011, 60, 38–44.
23 (47) Cheng, C. H.; Du, T. B.; Pi, H. C.; Jang, S. M.; Lin, Y. H.; Lee, H. T. Bioresour Technol 2011, 102 (21),
24
25 10151–10153.
26
27 (48) Soh, L.; Zimmerman, J. AP-133 - One-Pot Algal Biodiesel Production in Supercritical Carbon Dioxide.
28 Supercritical Fluid Technologies.
29
30 (49) Varma, M. N.; Deshpande, P. A.; Madras, G. Fuel 2010, 89 (7), 1641–1646.
31
32 (50) Wyatt, V. T.; Haas, M. J. J. Am. Oil Chem. Soc. 2009, 86 (10), 1009–1016.
33 (51) Ehimen, E. A.; Sun, Z. F.; Carrington, C. G. Fuel 2010, 89 (3), 677–684.
34
35 (52) Chee Loong, T.; Idris, A. Bioresour Technol 2014, 174, 311–315.
36
37 (53) Saka, S.; Kusdiana, D. Fuel 2001, 80 (2), 225–231.
38 (54) Vardon, D. Hydrothermal liquefaction for energy recovery from high-moisture waste biomass,
39
40 University of Illinois at Urbana-Champaign: Urbana, Illinois, 2012.
41
42
(55) Demirbaş, A. Energy Convers. Manag. 2000, 41 (6), 633–646.
43
44
45
46
47
48
49
50
51
52
53
54
55
56 23
57
58
59
60
ACS Paragon Plus Environment