• Open Access

Optimizing a continuous flow lipid extraction system (CFLES) used for extracting microalgal lipids

Authors

  • Javed Iqbal,

    1. WA Callegari Environmental Center, Louisiana State University AgCenter, Baton Rouge, LA, USA
    Search for more papers by this author
  • Chandra Theegala

    Corresponding author
    1. Biological and Agriculture Engineering, Louisiana State University AgCenter, Baton Rouge, LA, USA
    • WA Callegari Environmental Center, Louisiana State University AgCenter, Baton Rouge, LA, USA
    Search for more papers by this author

Correspondence: Chandra Theegala, tel. + 1 225 578 1060, fax + 1 225 578 3492, e-mail: ctheegala@agcenter.lsu.edu

Abstract

A laboratory-made continuous flow lipid extraction system (CFLES) was devised to extract lipids from microalgae Nannochloropsis sp., a potential feedstock for biodiesel fuel, with a focus to assess the workable temperatures and pressures for future industrial applications. Using conventional solvents, the CFLES recovered 100% of the lipids extracted with conventional Soxhlet extraction. The optimum temperature and pressure were found to be 100 °C and 50 psi, respectively; conditions significantly lower than those normally used in pressurized liquid extractions requiring specialized equipment. Approximately 87% of the extracted oil was successfully transesterified into biodiesel fuel (fatty acid methyl esters). Preliminary calculations based on the tested lab-scale system indicated savings in energy, solvent consumption, and extraction time as 96%, 80%, and more than 90%, respectively, as compared to Soxhlet extraction. However, the true cost savings can only be assessed at scaled up level. Energy efficiency of CFLES was calculated as 48.9%. Residual water (~70%) in the biomass had no effect on the extraction performance of CFLES, which is expected to help the process economics at scaled up application. The effect of temperature and pressure on the fatty acids profile of Nannochloropsis sp. is also discussed. Based on the existing literature, the authors believe that a pressurized liquid extraction system with continuous solvent flow has not been reported for lipid extraction from Nannochloropsis sp.

Introduction

One of the primary hurdles hampering biodiesel from emerging as a renewable fuel is the availability of feedstock, the vegetable oil. It is estimated that feedstock accounts for 45–58% of total production cost for second generation biofuels (Hamelinck & Faaij, 2006; Al-Zuhair, 2007). Therefore, cheaper feedstock, which does not interfere with the food market, is the key to the future of renewable biodiesel fuel. Numerous non-crop feedstock have been explored as possible substitutes for vegetable oil e.g. Chinese tallow tree (Boldor et al., 2010), and Jatropha (Kaushik et al., 2007). Microalgae is reported as the most promising substitute for the current vegetable oil obtained from non-food crop along with its nutraceutical and pharmaceutical significance (Roessler et al., 1994; Sawayama et al., 1995; Sheehan et al., 1998; Banerjee et al., 2002; Miao & Wu, 2006; Chisti, 2007). Microalgae has the potential to yield 15–300 times more oil for biodiesel production than traditional crops (Chisti, 2007). However, there are issues to resolve before reaping the full advantages of microalgae as an alternative diesel fuel. Harvesting of biomass is the most energy intensive aspect of algal biofuel production. Recovery of minute microalgal cells suspended in more than 99.9% water (typically between 0.02% and 0.07%) requires economical harvesting to a concentration between 5% and 25% depending on the extraction process employed (Benemann & Oswald, 1996). Extraction is an important and costly step, which often involves the use of toxic solvents. The use of solvent extraction requires extra energy input to recover the solvents, and it has the potential to contaminate the algal solids, thereby, restricting options for their end use.

The use of high pressure and temperature with forced flow of solvent is a recently adopted method for extraction of chemicals from solid and semi-solid matrices for environmental analysis (Richter et al., 1996; Herrero et al., 2004; Schantz, 2006). Pressurized liquid extraction (PLE) or pressurized fluid extraction (PFE), which is also know by the trade name “Accelerated Solvent Extraction” (ASE) (Dionex, Sunnyvale, CA, USA), was introduced in 1996 (Richter et al., 1996). The advantages of PLE over Soxhlet extraction include lower solvent consumption and faster extraction. Pressurized accelerated hot solvent extraction further improves the speed and extraction efficiency of lipids. Higher temperature increases the extraction kinetics, whereas high pressure keeps solvents below their boiling point, thereby enabling safe rapid extraction (Richter et al., 1996; Macnaughtona et al., 1997). However, the limitations of PLE are its requirement for specialized instrumentation to achieve relatively high pressures and temperatures. Maintaining excessive temperature and pressure is a cost deriving factor in terms of maintenance and operation. Therefore, a technology that utilizes an optimum temperature and pressure is expected to help lower this cost. Moreover, selection of a suitable extraction solvent is probably the most important step in optimizing PLE for micro-algal extractions as the traditional solvents generally have environmental and health implications.

The main disadvantage of a PLE-based extraction process is the high cost associated with its infrastructure and operation. This study focused on a laboratory-made continuous flow lipid extraction system (CFLES), which was designed to improve the process economics of microalgae oil extraction while simplifying the overall extraction process. If proven successful, the CFLES system is anticipated to offer unprecedented advantages to the biodiesel, nutraceutical, pharmaceutical, and other industries that can utilize microalgal lipids. Traditionally, pressurized liquid extraction (PLE) uses extreme temperatures up to 200 °C or higher and extreme pressures up to 3000 psi or higher. The current study examined moderate temperatures (80–120 °C) and moderate pressures (ambient to 500 psi) in the custom-developed CFLES. These are normally workable operational temperatures and pressures for large scale operation systems. PLE systems use a specialized sample cartridge filled with an extraction fluid, which statically extract the oils under elevated temperature and pressure. The current CFLES system on the other hand uses a continuous flow of solvent through the extraction cell containing biomass.

Microalgal strain selection has a significant impact on the efficiency of lipid extraction due to inherent variability associated with cell wall strength, structure, and chemical composition (Cooney et al., 2009). Nannochloropsis sp., a marine microalgae, was selected for this study due to its tougher cell wall and its ability to tolerate a wide range of temperatures. Insoluble and non-hydrolyzable biopolymers called algaenans are reported to form a chemically resistant part of the outer cell wall in Nannochloropsis sp. (Tyson, 1995; Gelin et al., 1999; Sukenik, 1999; Rodolfi et al., 2003). Due to high growth rates and lipid contents, Nannochloropsis sp. is recognized as a potent renewable resource for production of biofuel (Rodolfi et al., 2009; Brown et al., 2010). The species was also well studied from an analytical point of view. The effect of temperature and pressure on the fatty acids profile of Nannochloropsis sp. is reported in more detail than previously reported (Sukenik & Carmeli, 1990; Rebolloso-Fuentes et al., 2001; Rodolfi et al., 2009; Brown et al., 2010). Furthermore, to our knowledge, a pressurized liquid extraction system with continuous solvent flow has not been reported for lipid extraction from microalgae.

Material and methods

Microalgae strain and culture condition

Nannochloropsis sp., microalgal strain, along with modified Guillards f/2 formula, Micro Algae Grow, and Crystal Sea® marine mix at 33 g L−1 salinity were acquired from Aquatic Ecosystems Inc. (Apopka, FL, USA, cat# LAC1Q, F2A6 and CM2, respectively). Cultures were grown in four 10-gal aquarium tanks, each filled with 3 gal of growth media. A 16/8 h light/dark illumination was provided by a 400 W high pressure sodium light bulb. Air was continuously bubbled through the media to prevent settling and to supply CO2. Ambient temperature was recorded in the range of 22–28 °C. Approximately 40 g of wet paste was produced in 2 weeks (equivalent to 64 mg L−1 d−1 of wet culture density). Water was separated through centrifugation in Thermo IEC, K centrifuge (Needham Heights, MA, USA) at 3500 rpm for 5 min, bringing down the moisture contents to 80%. The paste was then dried at 38 °C overnight to contain 30 wt% solids. The biomass was homogenously mixed before extraction.

Conventional soxhlet extraction

Soxhlet extraction was performed according to the method mentioned by Luque de Castro & Garcia-Ayuso (1998). In short, 3.3 g of algal paste (equivalent to 1 g dry wt.) was placed in a cellulose extraction thimble (Whatman, Clifton, NJ, USA, Cat. # 2800–338). A co-solvent system based on a well-known Bligh & Dyer (1959) method, in which 50 mL of chloroform, 100 mL of ethanol, and 40 mL of DI water. The extraction was performed for 8 h to achieve complete extraction. The extracts were transferred to a stoppered graduated cylinder and 40 mL of DI water was added. The cylinder was inverted 30 times and allowed to settle for 1 h to recover the bottom layer containing lipids and chlorophyll dissolved in chloroform. The chloroform layer was transferred to a 250 mL flask to evaporate the excess chloroform using rotary evaporator (Rotavapor R-210, Buchi Inc., Zurich, Switzerland). The final extraction volume was adjusted to 10 mL.

Continuous flow lipid extraction system (CFLES)

As depicted in Fig. 1, the in-house designed and built CFLES was driven by a HPLC solvent injection pump (Model 510, Millipore, Milford, MA, USA) capable of delivering liquid in the range of 0.1–10 mL min−1 at pressure up to 3000 psi. The pump was connected to a stainless steel check-valve, a ball valve (V1), a 10 feet long, ¼ inch ID copper tubing, coiled inside a temperature controlled oven (Isotemp vacuum oven, Model 285A, Fisher Scientific, USA), and a pressure gauge. The long copper tube facilitates pre-heating the solvent before it enters the sample extraction cell. The sample extraction cell was a 3-inch long and 3/8 inch ID stainless steel tubing (Fig. 2). The other end of the sample extraction cell was connected to a ¼ inch copper tubing, which exits the oven and was attached to a stainless steel ball valve (V3). After V3, the tubing ended in a 250 mL clear glass bottle to hold the liquid extracts exiting CFLES. Flow control through V3 controlled the system pressure. The entire system was tested for leaks at the maximum operating pressures and temperatures. The entire system was flushed with clean ethanol before the test runs and in between the sample runs to overcome any carry over.

Figure 1.

Schematic of continuous flow lipid extraction system (CFLES).

Figure 2.

Sample extraction cell used in continuous flow lipid extraction system (CFLES), along with a steel wool plug/filter.

Approximately 3.3 g of microalgal wet paste (70% moisture) was fed into the sample extraction cell. Both ends of the sampling cell were plugged with a mass of steel wool to filter and contain the microalgal cells inside the extraction cell (Fig. 2). The extraction cell was connected tightly with the copper tubing using stainless steel nuts, front and back ferules. Valve V2 was left fully open all the time, except during replacement of the extraction cell. Pressure inside the copper tubing and extraction cell was regulated using V3, for a given pump flow.

Solvents used for CFLES

A co-solvent extraction system consisted of ethanol and chloroform in 1 : 2 proportions. The flow rate of co-solvent was adjusted to 2 mL min−1. All the extractions were performed in triplicate. Starting with ambient temperature and pressure, the extractions were performed at 80 °C, 100 °C, and 120 °C temperatures, and at 50 psi and 500 psi pressures (Table 1). The sample extraction was terminated with a clear solvent draining into the extracts-collection bottle.

Table 1. Temperature and pressure parameters used in continuous flow lipid extraction system (CFLES)
Test nameTemperature, °CPressure, psiTest nameTemperature, °CPressure, psi
AmbT, PAmbientAmbient100T, AmbP100Ambient
AmbT, 50 psiAmbient50100T, 50 psi10050
AmbT, 500 psiAmbient500100T, 500 psi100500
80T, AmbP80Ambient120T, AmbP120Ambient
80T, 50 psi8050120T, 50 psi12050
80T, 500 psi80500120T, 500 psi120500

Post-extraction process

The extracts were transferred to a graduated cylinder. DI water, approximately equal to the amount of ethanol, was added forming a biphasic system. The top layer, containing water and ethanol, was removed. The bottom layer contained lipids and chlorophyll contents dissolved in chloroform. Final extraction volume was set to 10 mL.

Glycerides analysis

Mono-, di-, and triglycerides were analyzed using 1 mL of the extracts according to the method mentioned in Balasubramanian et al. (2011). Briefly, 1 mL of the extracts was silyated with 20 μL of N-methyl-N-trimethylsilyltrifluoracetamide (MSTFA) (ThermoScientific, Waltham, MA, USA, catalog# TS-48913) in a 5 mL vial (ASTM D6584). The solution was mixed thoroughly and reacted for 10 min at 70 °C in an oven. One μL of the final diluted aliquot was injected into a gas chromatograph (SRI, 8610C, Torrance, CA, USA) equipped with flame ionization detector. A siltek-treated stainless steel capillary column (14 m, 0.53 mm id, 0.16 μm df) with 2 meter Integra-Gap® built-in retention gap (Restek Corporation, Bellefonte, PA, USA, MXT-Biodiesel TG w/inl Gap, catalog# 70289) was used for mono-, di-, and tri-glycerides. The initial column oven temperature was 50 °C held for 2 min; raised to 380 °C at 15 °C min−1. Helium was used as a carrier gas at 4 mL min−1. Peaksimple software (SRI instruments, Menlo Park, CA, USA) was used to quantify peak areas and converted to concentrations using appropriate response factors.

Fatty acid methyl esters (FAMEs) analysis using GC/MS

FAMEs were determined using GC/MS by transesterifying 1 mL of the extract with 15 μL of 2N methoxide solution (11.2 g KOH in 100 mL methanol). The sample was centrifuged at 3000 rpm for 2 min to precipitate the free glycerol and chlorophyll. Hexadecanoic acid, 2-hydroxy-, methyl ester (CAS No. 16742-51-1, Indofine Chemicals, NJ, USA, Cat.# 24-1602) was used as an internal standard.

Fatty acid methyl esters were analyzed using a Varian 450-GC gas chromatograph (Walnut, CA, USA) with 1179 injector, equipped with Varian 250-MS ion trap mass spectrometer, and Varian CP-8400 autosampler. FAMEs were separated with a Varian FactorFour WAXms column (30 m, 0.25 mm, and 0.25 μm df; Varian Inc., Walnut Creek, CA, USA, catalog# CP9205). The MS electron multiplier voltage was set to 1400 V, ionization time of 25 000 μs in electron impact (EI) mode, with transfer line, ion trap and manifold temperatures set to 250 °C, 200 °C, and 50 °C. The MS was set to scan 50–1000 m/z with an ionizing voltage of 70 eV. One μL sample was injected with a split ratio of 20 and injector temperature set to 240 °C. The initial temperature of the column oven was set to 100 °C, held for 2 min, then raised to 255 °C at 12 °C min−1 held for 7 min; Helium was used as a carrier gas at 1 mL min−1. Data acquisition and analysis was performed using Varian MS Workstation version 6.5. The instrument was calibrated using a 37-component standards mix (Supelco No. 18919, CA, USA) containing C4 – C24 FAMEs (2–4% relative concentrations).

Chlorophyll a analysis

Chlorophyll a determination was done according to US EPA method 446.0. The method employs Jeffrey & Humphrey's (1975) Trichromatic Equations. The UV-VIS Spectrophotometer (Helios Aquamate, ThermoSpectronic, UK) was calibrated using a chlorophyll standard (MP Biomedicals, OH, USA; Catalog# 210221). Absorbance was measured at 750, 664, 647, and 630 nm. Chlorophyll a (mg L−1 of extracts) was calculated according to the following equation:

display math(1)

where, Chla is the concentration (mg L−1) of chlorophyll a in the extracts, converted to mg g−1 dw.

Gravimetric analysis

Total lipids were determined gravimetrically according to Bligh-Dyer modified method (Burja et al., 2007) using the following equation:

display math(2)

where, Wdry is the weight (mg) of aluminum dish and residues dried at 60 °C, Wdish is the weight (mg) of empty aluminum dish, Vext is the volume (mL) of final extracts, Vdry is the volume (mL) of extracts transferred to the aluminum dish, WS is the weight (g) of the sample extracted.

Elemental composition of algae feed and extracted oil

Total nitrogen was quantified by CHN analysis. Dry samples were combusted in a CHN elemental analyzer (Elementar, Vario EL III). Helium was used as a carrier gas. Acetanilide (C = 71.09%; N = 10.36%; H = 6.71%) was used to calibrate the instrument.

Statistical analysis

Analyses of variance (anova) of different treatments and Fisher's protected least significant difference (PLSD) test for pair wise comparison was performed using STATISTICA version 9 software (StatSoft Inc., Tulsa, OK, USA).

Results

Prolonged 8 h lipid extraction from Nannochloropsis sp. with conventional Soxhlet apparatus yielded 498.9 mg g−1 of oil (49.9% as total bound glycerides), 443 mg g−1 (44.3%) of fatty acid methyl esters (FAMEs), and 4.19 mg g−1 of chlorophyll a. The Soxhlet extraction consumed approximately 150 mL of solvent, which is one of the drawbacks associated with conventional Soxhlet extractions (Luque de Castro & Garcia-Ayuso, 1998; Wang & Weller, 2006). Complete extraction is indicated by clarity of solvent filtered through the biomass as well as discoloration of the biomass (Rao et al., 2007). Although conventional Soxhlet extraction is widely reported for its more efficient extraction, its longer time required for complete extraction, large volume of solvents wasted, and the high energy requirement for continuous distillation restrict its use for scaled up applications (Halim et al., 2011).

Oil contents (as total bound glycerides) extracted from Nannochloropsis sp. was 627.5 mg g−1 dw (62.8% dry wt.) using CFLES, which was approximately 13% higher (P < 0.05) than that of the conventional Soxhlet extraction (49.9%) (Fig. 3). Although these results agree with those previously reported for Nannochloropsis sp. to contain 31–68% oil on dry weight (dry wt.) basis (Rebolloso-Fuentes et al., 2001; Chisti, 2007; Gouveia & Oliveira, 2009; Rodolfi et al., 2009), only 87% of the CFLES extracted oil (i.e. 546 mg g−1 or 54.6% dry wt.) was saponifiable, which could be converted into FAMEs. High concentrations may also incorporate analytical errors associated with standards recovery (112.3 ± 5.6%) during glyceride analysis using GC-FID instrument. The final results for all analysis including those of the Soxhlet extraction were not corrected for standard's recovery. Many microalgae strains naturally have high lipid content (20–50% dry weight) (Hu et al., 2008; Brennan & Owende, 2010), Soxhlet extraction showed 491.2 mg g−1 of triglycerides. CFLES extracted 100% of the triglycerides (494.3 mg g−1) along with additional diglycerides (116.2 mg g−1), which was not extracted with conventional Soxhlet extraction (Table 2). The maximum total fatty acid methyl esters (FAME) produced from the total bound glycerides extracted using CFLES was approximately 87% of FAME produced from total glycerides using Soxhlet extraction (Fig. 4). This maximum yield was achieved at 100 °C temperature, 50 psi pressure, and 15 min time compared to Soxhlet's 8 h extraction (Fig. 5). The total FAMEs extracted under the aforementioned conditions was 385 mg g−1 (38.5% dry wt.) compared to 443 mg g−1 (44.3% dry wt.) of the Soxhlet extraction (i.e. 87% efficiency).

Figure 3.

Total bound glycerides (mono-, di-, and triglycerides) concentrations (mg g−1 dry weight) extracted from Nannochloropsis sp. under different temperature and pressure combinations in continuous flow lipid extraction system (CFLES) and Soxhlet extraction (n = 3).

Figure 4.

Total fatty acid methyl esters (FAMEs) concentrations (mg g−1 dry weight) extracted from Nannochloropsis sp. under different temperature and pressure combinations in continuous flow lipid extraction system (CFLES) and Soxhlet extraction (n = 3). Percent values are relative to that of Soxhlet extraction.

Figure 5.

Effect of temperature and pressure on: (a) lipids as total glycerides extracted from Nannochloropsis sp. in continuous flow lipid extraction system (CFLES), and (b) lipids transesterified into fatty acid methyl esters (FAMEs).

Table 2. Total free and bound glycerides (% of dry wt.) extracted from Nannochloropsis sp. under different temperature and pressure conditions in CFLES and Soxhlet extraction (n = 3)
Test nameFree glycerinMono-glyceridesDi-glyceridesTri-glyceridesTotal bound glycerides
  1. Amb, Ambient; T, Temperature, °C; P, pressure, psi; “–”, Non detect.

AmbT, P0.29 ± 0.011.21 ± 0.23.32 ± 0.513.72 ± 1.618.26 ± 2.2
AmbT, 50 psi1.32 ± 0.51.09 ± 0.52.09 ± 0.115.02 ± 2.418.22 ± 2.8
AmbT, 500 psi0.25 ± 0.011.64 ± 0.61.95 ± 1.111.22 ± 0.814.82 ± 1.3
80T, AmbP0.38 ± 0.14.76 ± 0.21.71 ± 0.415.28 ± 121.76 ± 0.8
100T, AmbP2.68 ± 0.33.64 ± 1.21.39 ± 0.512.9 ± 4.117.95 ± 2.6
120T, AmbP0.13 ± 0.011.27 ± 01.78 ± 0.211.83 ± 0.714.89 ± 0.8
80T, 50 psi0.34 ± 0.011.72 ± 0.22.57 ± 0.822.43 ± 3.826.74 ± 3.4
100T, 50 psi0.97 ± 0.11.69 ± 0.211.62 ± 2.149.43 ± 2.562.75 ± 0.8
120T, 50 psi0.3 ± 0.013.74 ± 0.33.23 ± 1.420.95 ± 1.727.93 ± 2.2
80T, 500 psi0.06 ± 0.010.9 ± 00.62 ± 0.25.38 ± 2.56.91 ± 2.7
100T, 500 psi0.08 ± 0.012.42 ± 0.21.95 ± 0.311.54 ± 0.915.92 ± 0.9
120T, 500 psi0.68 ± 0.10.3 ± 0.12.87 ± 1.43.88 ± 1.6
Soxhlet0.19 ± 00.57 ± 0.149.12 ± 3.449.89 ± 3.4

Data showed that pressure alone had no significant effect on extraction performance unless augmented with temperature (Figs 3 and 4). Only up to 21% of oil, of which 17% was esterifiable to FAMEs, was extracted at lowest or highest pressures. Once combined with temperature, the extraction of oil, and consequent FAMEs, at lowest pressure (ambient) promoted to 38% and 30%, respectively. An optimum pressure of 50 psi further maximized the extraction of oil and FAMEs yield significantly (P < 0.05) to 97% and 87%, respectively. Pressure higher than 50 psi did not show any beneficial effect. Temperature around 80 °C was not enough to attain maximum yield. Temperature around 120 °C had deleterious effect on physical and chemical properties of the biomass, including the mass transfer and reduced extraction of lower molecular weight fatty acids (Table 3). Temperature above 100 °C seems to deteriorate the lipids. For instance, C12:0 and C14:0 fatty acids reduced by 50% at 120 °C temperature and 500 psi pressure, compared to that of 100 °C and 500 psi. Deterioration of lipids may also be attributed to the cleavage of carbon-oxygen bond in fatty acids due to thermal decomposition (Frankel, 1980) or sensitivity of fatty acids to oxygen and temperature (Fournier et al., 2006). The effect of solvent temperature was complex. As the temperature increased, the biomass clumped into a hard cake inside the extraction cell, reducing the solvent diffusion and mass transfer (Fig. 6). The temperature increase also reduced the solvent's density considerably, thereby reducing the solvent-lipids contact, which in turn offset the lipids volatility, resulting in a net lower lipid mass transfer rates (Halim et al., 2011).

Figure 6.

Microalgal biomass extracted at 100 °C temperature and 50 psi pressure (R), and 120°C/500 psi (L) in the continuous flow lipid extraction system (CFLES).

Table 3. Elemental composition of dry Nannochloropsis sp. feedstock and extracted oil (bio-oil)
 C%H%N%O%aHigh heating value MJ kg−1
  1. a

    Oxygen contents were determination by subtracting the sum of C, H, and N from 100.

Dry feed48.74 ± 0.377.34 ± 0.076.9 ± 0.0837.03 ± 0.421.06 ± 0.1
Bio-oil67.95 ± 0.469.85 ± 0.051.23 ± 0.320.97 ± 0.433.61 ± 0.3

Fatty acids profile of CFLES extracted lipids showed C18:1 as the major component (~38%) (Fig. 7) followed by C16:0 (~23%), C12:0 (~12%), and C14:0 (~8.5%) (Table S1). The C12:0 has not been reported in Nannochloropsis sp. before. Total saturated fatty acids were slightly higher (~54%) than the unsaturated ones (~46%). No significant concentrations of polyunsaturated fatty acids (PUFA), eicosapentaenoic acid (EPA) [C20:5(n-3)], or docosahexaenoic acid (DHA) [C22:6(n-3)] were found in the algal culture, though the strain has been reported for significant production of EPA (Sukenik, 1999; Brown et al., 2010). Fatty acid profile of microalgal species, however, has been reported to be a function of its culturing conditions (Andrich et al., 2005; Rao et al., 2007). The Soxhlet extracted lipids were dominated by C16:1 (~21%), C18:ω3 fatty acid (~16%), C12:0 (~15%), C18:1 (12%), and C14:0 (~8%). Fatty acid profile, in agreement with previously reported results (Chisti, 2007; Umdu et al., 2009) indicated the potential of Nannochloropsis sp. for biodiesel production.

Figure 7.

Concentrations of saturated and unsaturated fatty acid components of C18 extracted from Nannochloropsis sp. at 100 °C temperature and 50 psi pressure in continuous flow lipid extraction system (CFLES) compared to that of the Soxhlet extraction (n = 3).

Gravimetric lipids concentration indicated similar trends as observed for bound glycerol and FAMEs (Fig. 8). Lipids concentration recovered at 100 °C/50 psi (662 ± 14 mg g−1) were not significantly different (P > 0.05) than that in the Soxhlet extracts (682 ± 22 mg g−1). None of the other temperature and pressure combinations had yields comparable to these two extractions (P < 0.05).

Figure 8.

Gravimetric lipid concentrations extracted from Nannochloropsis sp. using continuous flow lipid extraction system (CFLES) and Soxhlet extraction (n = 3).

Significantly higher concentrations of chlorophyll a were extracted at 120 °C temperature and ambient pressure than the 100 °C and 50 psi or the Soxhlet (P < 0.05) in Nannochloropsis sp. (Fig. 9). Results showed that high Chlorophyll a contents extracted was not associated with high yields of oil or FAMEs, or even the gravimetric lipid contents. Gravimetric lipids at 120 °C temperature and ambient pressure were 157 ± 6 mg g−1, which was significantly less than that of the 100 °C/50 psi or Soxhlet extractions (P < 0.05).

Figure 9.

Chlorophyll a contents extracted from Nannochloropsis sp. using continuous flow lipid extraction system (CFLES) and Soxhlet extraction (n = 3).

Discussion

CFLES extracted glyceride molecules, particularly diglycerides, which were not extracted with conventional Soxhlet extraction. This shows significantly higher lipid contents extracted with CFLES (Fig. 8), although the excess lipids failed successful esterification into FAMEs with the methylation techniques selected. This is likely due to the fact that Nannochloropsis sp. has a variety of polar and non-saponifiable lipids, including complex phospholipids, glycolipids, and phosphatidylglycerol (Schneider & Roessler, 1994; Halim et al., 2011). Phospholipids are suggested as a source of catalyst destruction during transesterification (Schneider & Roessler, 1994) and the phosphorus compounds in the oil do not easily carry over into the methyl esters (Gerpen & Dvorak, 2002).

Non-polar lipids (e.g. sterol esters, glycerides, hydrocarbons and carotenoids) are bound weakly by van der Waals forces and are relatively easy to extract if in contact with suitable solvents (Enssani, 1990). However, the impermeability of the microalgal cell wall is a physical barrier for solvents to reach the lipids (Bligh & Dyer, 1959). Therefore, combination of chemical and moderate physical processes (100 °C/50 psi, and the solvents) used in this study was sufficient enough to break the barrier between the solvents and the lipids, unlike pressurized solvent extraction processes, which require extreme pressure and temperature to achieve similar extraction efficiencies. Morrison & Conventry (1989) reported that fatty acids were more extractable at 100 °C as compared to ambient temperature, particularly saturated acids e.g. C16:0 palmitic and C18:0 stearic acid, while using hot propanol–water (3 : 1, v/v), water saturated butanol, methanol and methanol–water (85 : 15, v/v).

A co-solvent system has a limited carrying capacity to dissolve a lipid (or the lipids have a limited solubility in co-solvent) (Cooney et al., 2009). At a certain point the carrying capacity of the co-solvent was exhausted unless a fresh solvent was either added or re-circulated similar to Soxhlet extraction. In similar fashion, a fresh solvent was flushed continuously through the biomass in the CFLES extraction cell so that the carrying capacity of the solvent was never exhausted, resulting in high yield. Conversely, lower yields reported for PLE (Jaime et al., 2005; Rodríguez-Meizoso et al., 2008) were expected due to exhaustion of the solvent's carrying capacity. In PLE, the extraction cell is filled with biomass and solvent up to the desired pressure. Extraction is performed statically for a certain time. The same solvent along with solute is then removed for further processing and analysis.

This study has demonstrated that residual water (~70%) in the biomass did not affect the extraction performance. Water was reported to aid extraction through its swelling of the cellular matrix and its natural role as a polar co-solvent (Schwartzberg, 1997; Pourmortazavi & Hajimirsadeghi, 2007). This was important since (1) drying biomass before extraction has economic implications for microalgae as biodiesel feedstock (2) industrial-scale implementation of the extraction system is more economical with wet paste compared to dried biomass.

The chlorophyll a contents did not show significant relationship with lipid contents extracted at a specific temperature or pressure (Fig. 9). Degradation of chlorophyll may occur at a high temperature. For instance, pheophytins as degradation products of chlorophylls have been reported at 200 °C (Rodríguez-Meizoso et al., 2008). The microalgal lipid extracts had a dark green color due to chlorophyll presence. To make it a viable feedstock for biodiesel, further refining to remove these unusable constituents would be necessary. The transesterification process in the current study precipitated significant concentrations of these constituents. Various studies have been conducted to effectively remove chlorophyll from oil (Bahmaei et al., 2005).

Effect of temperature and pressure

As shown in Fig. 5, the extraction performance increased from ambient to 100 °C temperature and decreased with further increase. Similarly, the extraction performance increased from ambient to 50 psi pressure and decreased with further increase. Similar trends were observed for FAMEs, total bound glycerol, and gravimetric lipids contents.

Extraction at 120 °C temperature resulted in a hardened biomass which stayed green at the end of the extraction process (Fig. 6). A part of the biomass was also seen burnt. Similar results were observed at 150 °C temperature and 1000 psi pressure (data not shown). On the other hand, microalgal cells turned white at the end of extraction at 100 °C, indicating an efficient or complete extraction (Eroglua & Melis, 2010), as was observed in Soxhlet extraction. Extracted biomass at the end of 80 °C was light green indicating incomplete extraction.

CFLES reduced the solvent consumption, as well as comparable solute recoveries. Liquids under pressure act as solvents, allowing extraction of analytes at a temperature above the boiling point of the solvent. The analyte solubility is therefore enhanced and the desorption kinetics is accelerated. Pressure also facilitates increased transport of solvent to hard-to-reach corners, pores, surfaces and matrices (Richter et al., 1996; Cooney et al., 2009). Elevated pressure is reported to reduce the dielectric constant of immiscible solvents to values that better match the polarity of the lipids (Richter et al., 1996; Herrero et al., 2006; Cooney et al., 2009). Similarly, temperature increases the solvent potential of a solvent (Richter et al., 1996) by accelerating diffusion rates (Denery et al., 2004). The thermal energy helps overcome the cohesive (solute-solute i.e. lipids-lipids) interactions and adhesive (solute-matrix, i.e. lipids-cell matrix) interactions (Richter et al., 1996; Cooney et al., 2009). Increase in thermal energy increases molecular motion of the molecules and thereby decreases the molecular interactions of hydrogen bond, van der Waals forces, and dipole interactions (Cooney et al., 2009). Diffusion rates have been reported to increase roughly from twofold to tenfold upon increasing the temperature from 25 to 150 °C (Perry et al., 1984; Richter et al., 1996). Pressurized solvents at elevated temperature hence improve the efficiency of traditional extraction systems resulting in shorter extraction time and lower solvent consumption (Cooney et al., 2009).

Effect of higher temperature on fatty acids profile

Changes in fatty acid composition at high temperature have been reported previously (Tyagi & Vasishtha, 1996). They reported a significant decrease in unsaturated fatty acids with temperature during frying oil. Tri-unsaturated fatty acids (trienes) deteriorated faster than the di-unsaturated (dienes), which in turn deteriorated faster than mono-unsaturated (monoenes). The concentration of saturated fatty acids was reported increasing with temperature simultaneously. We found similar results while heating soybean oil at 100 °C for 1 h in a Rancimat oxidation stability test (data not shown), wherein linolenic acid methyl ester (C18:3ω-3) reduced by 98%, C18:2 reduced by 63%, and C18:1 reduced by 36%. The γ-linolenic acid methyl esters (C18:3ω-6) reduced by 27%. Almost similar results were observed here in the CFLES (Fig. 7). High concentration was seen for C18:1 compared to C18:2 and C18:3 contrary to that of Soxhlet extraction. Generally metals like nickel are used as a catalyst in hydrogenation of oil (Fernandez et al., 2007) and can also initiate fatty acid oxidation by reaction with oxygen (Frankel, 1980). Copper metal works as a catalyst for hydrogenation of oil, which has a much higher preference for linolenic acid (C18:3) and is further accelerated with increasing temperature (Coenen, 1976). Since copper tubing and high temperatures were used in this study, hydrogenation of the double bonds may have occurred during the extraction process.

Energy efficiency and energy savings

The cost of extraction (both energy and solvent) is anticipated to be significantly lower for CFLES. Although the true cost savings can only be assessed at pilot or large scale production system, preliminary calculations on the tested lab-scale systems indicated an energy consumption of 0.3 kWh for the CFLES (0.25 h, 1200W of solvent pumping and heating, continuous power consumption) and 8.64kWh for Soxhlet extraction (8 h of 1200 W heating and condensing, assuming a heater on time of 90%), which corresponds to approximately 96.5% energy savings. Since the current work focused on lipid extraction only, energy requirement for harvesting of biomass to produce algae paste is not considered in energy calculations.

Dulong formula Eqn (3) was used to estimate the heating value of the dry microalgae and extracted oil.

display math(3)

where C, H, and O are the weight percentages of carbon, hydrogen, and oxygen, respectively. The elemental composition and the heating value for microalgae feedstock and extracted oil (bio-oil) are given in Table 3. The heating value for bio-oils (~33.6 MJ kg−1) was significantly lower than that of petroleum crude oil (43 MJ kg−1) and biodiesel (~41 MJ kg−1) (Brown et al., 2010) but significantly higher than that of the dry microalgal feedstock (∼21.06 MJ kg−1). Energy efficiency calculated as:

display math(4)

was found to be 48.9% for the CFLES. Based on the heating values per 100 g of dry feedstock, energy recovered from algae was approximately 100% using CFLES compared to 79% of the Soxhlet extraction (Fig. 10). The calculations were based on 62.8% and 49.9% dry wt. oil contents extracted with CFLES and Soxhlet, respectively.

Figure 10.

High heating values (HHVs) for dry microalgae feedstock and bio-oil extracted with continuous flow lipid extraction system (CFLES) and Soxhlet. Percent values in the bars shows energy recovered compared to that of the dry feed (n = 3).

The solvent usage for the CFLES was also 80% less when compared to Soxhlet extraction without solvent recycling. Significant time and labor savings are also anticipated. Despite these promising figures, it has to be kept in mind that these savings may or may not reflect the true economics of continuous flow, industrial-scale lipid extraction systems. Process economics are expected to be more favorable with continuous process compared to a batch process.

The optimum temperature and pressure of 100 °C and 50 psi, respectively, used in the CFLES readily extracted the microalgal oil than is possible at extreme (low or high) temperatures and pressures. The extraction efficiency is higher than most of the extraction methods reported previously for microalgae including supercritical CO2 extraction (36%) (Valderrama et al., 2003; Krichnavaruk et al., 2008), thermochemical and hydrothermal liquefaction (37–64%) (Minowa et al., 1995; Sawayama et al., 1995; Brown et al., 2010), pressurized liquid extraction (20–40%) (Jaime et al., 2005; Rodríguez-Meizoso et al., 2008), and microwave assisted extraction (28–77%) (Lee et al., 2010; Balasubramanian et al., 2011). Although the optimum operating temperature and pressure (100 °C, 50 psi) are workable for scaled up continuous CFLES, the main outcome of the current study is an efficient extraction technique for bio-oil from microalgae at lower temperature and pressure. The authors suggest additional research to investigate the performance of CFLES designed with continuous flow of biomass into the extraction cell at a pilot scale, extraction of other significant non-fuel products, and the use of solvents least toxic to environment and health.

Acknowledgements

Special thanks are extended to Dr. Ralph J. Portier, Department of Environmental Studies, School of the Coast and Environment, Louisiana State University, Baton Rouge, LA USA for providing accessories to the CFLES system. Thanks to W.A. Callegari Environmental Center, LSU AgCenter, for using the facility to perform this research work.

Ancillary