Introduction

Algal oil continues to attract attention as a potential feedstock for biofuels (Borowitzka and Moheimani 2013). For oil recovery from an alga, the culture broth is first typically dewatered to a paste, and the paste is then dried and finally extracted with solvents. The recovery of biomass paste from a dilute algal slurry with a typical dry biomass concentration of 0.5–4.0 g L−1 is expensive. Except for extremely high-value products, the algal slurry cannot be directly centrifuged or filtered to a paste in a commercial operation in view of the capacity requirements of the separation equipment and the associated operational costs. Therefore, the slurry must be partly dewatered by using an initial flocculation step, typically involving inexpensive and safe mineral salt flocculants (e.g., aluminum sulfate, ferric chloride). This initial dewatering concentrates the slurry sufficiently (Chatsungnoen and Chisti 2016a, b) so that further dewatering by centrifugation or filtration becomes feasible. However, the cationic metal salt flocculants adsorb to the algal biomass irreversibly and have the potential to interfere with solvent extraction of the paste. This notwithstanding, flocculation–sedimentation is among the least expensive methods for initial dewatering of the algal slurries (Molina Grima et al. 2003; Vandamme et al. 2013). Other dewatering methods have been discussed in the literature (Chen et al. 2015).

Although the solvent extraction of algal oils from dry biomass is generally straightforward, drying is not feasible in a commercial production process if the intention is to use the oils for making the various possible biofuels (Chisti 2010, 2013). This is because drying is expensive and energy intensive (Chisti 2013). Therefore, the oil must be extracted from a moist biomass paste that typically contains about 85 % water by weight. Solvent extraction of algal oils has been further discussed in the literature (Cooney et al. 2009; Halim et al. 2012; Mercer and Armenta 2011; Molina Grima et al. 2013; Chatsungnoen and Chisti 2016c).

This work uses several freshwater and marine algae, including a diatom, to demonstrate the following: (1) extractability of the oils from the biomass paste in comparison with extraction from the freeze-dried biomass, (2) a lack of an effect of the adsorbed metal salt flocculants on solvent extraction of the oil from the biomass paste, and (3) an equivalence of the freeze-drying and the high-temperature oven drying methods of determination of the moisture content of the biomass paste. The latter aspect is important because the biomass concentration in a growth process is generally determined in terms of the oven-dried biomass or a method calibrated against it. However, the biomass used in oil extraction is typically freeze-dried because high-temperature oven drying can potentially damage the oils.

An equivalence of oil extraction from biomass paste and dry biomass using an appropriate solvent extraction protocol has not been previously demonstrated as a general principle applicable to diverse freshwater and marine microalgae. Similarly, no generalization has been made previously concerning the effects of adsorbed metal ion flocculants on oil recovery from the biomass of diverse microalgae.

Materials and methods

Five microalgae were used: the freshwater microalgae Chlorella vulgaris, Choricystis minor (Mazzuca Sobczuk et al. 2008; Mazzuca Sobczuk and Chisti 2010) and Neochloris sp., and the marine microalgae Nannochloropsis salina (CCAP849/3) and Cylindrotheca fusiformis (CCAP1017/2). The freshwater species were purchased from Landcare Research, Lincoln, New Zealand, and the marine species were from the Culture Collection of Algae and Protozoa (CCAP), Argyll, UK (Chatsungnoen 2015; Chatsungnoen and Chisti 2016a).

BG11 medium or its modifications were used to maintain and grow the microalgae (Andersen et al. 2005). The freshwater algae were maintained in BG11, while the marine algae were maintained in the BG11 formulated in seawater. Chlorella vulgaris was maintained separately in both BG11 and BG11 seawater media. The medium for the diatom (C. fusiformis) always contained silicate (sodium metasilicate nonahydrate, Na2SiO3·9H2O, at a concentration of 30 mg L−1 medium) (Chatsungnoen 2015). BG11 seawater medium was made using artificial seawater. This was prepared by dissolving 40 g of sea salt (natural unrefined Southern Pacific Ocean salt; Pacific Natural Fine Salt; Dominion Salt Ltd, Marlborough, New Zealand) in 1 L of distilled water and filtering with Whatman GF-C (0.45 μm) 90 mm microfiber filters before use (Chatsungnoen 2015). The seawater had a salinity of approximately 38.5 ppt (EcoSense EC300 conductivity/salinity meter; YSI Inc., USA).

Inocula were prepared by aseptically transferring the microalgal stock culture from liquid or solid agar media to 40 mL of BG11 in 250-mL Erlenmeyer flasks. The flasks were held in an incubator shaker at 130–140 rpm, 25 °C, under fluorescent light (~30 μmol photons m−2 s−1), for around 20–30 days. This culture (40 mL) was used to inoculate 360 mL of BG11 in a 1-L Duran bottle (borosilicate glass 3.3, LabServ, Biolab, New Zealand). These bottles were incubated at room temperature (~25 °C) for approximately 7–14 days. This culture (400 mL) was used to seed 1.6 L of fresh medium in a 2-L Duran bottle and grown for a further 7–14 days. This culture was split into 400-mL lots and transferred to five 2-L Duran bottles each with 1.6 L of BG11 or its modification. All Duran bottle cultures were maintained at room temperature (24–26 °C) under continuous light (~219 μmol photons m−2 s−1) from a bank of six tubes of fluorescent lamps (Philips TLD 58w/840, cool white, Thailand). All Duran bottle cultures were continuously bubbled (0.375 L min−1) with humidified air mixed with 5 % (v/v) carbon dioxide. The inlet and exhaust gas streams were sterile filtered by passing through 0.2-μm Teflon membrane filter (Midisart 2000; Sartorius AG, Germany). All cultures were grown as three replicate runs.

The cultures were harvested in the stationary phase (30–55 days post-inoculation) and kept at 4 °C in the dark. This broth was used within 7 days for biomass recovery by flocculation or centrifugation.

Biomass concentration, productivity, and growth rate

A 20-mL sample of the algal broth was vacuum filtered using a pre-weighed Whatman GF-C (0.45 μm, 90 mm) microfiber disc filter. The filter disc was washed with 2 × 20 mL of distilled water (for the cultures grown in BG11 freshwater medium) or with 2 × 20 mL of 0.5 M ammonium formate (for cultures grown in the BG11 seawater medium). The filtered biomass samples were dried at 80 °C in an oven overnight, cooled in a desiccator, and weighed to calculate the dry biomass in 20 mL of the algal broth (Lee and Shen 2004).

A sample of the broth used in the above dry weight measurements was serially diluted with the fresh medium, and the optical density was measured at 680 nm. The blank was the fresh medium. The dilutions were such that the maximum measured optical density did not exceed 0.6. The measured optical density was plotted against the precisely known dry weight concentration to obtain a linear calibration plot (Chatsungnoen 2015; Chatsungnoen and Chisti 2016a). Separate calibration plots were made for the different microalgae. Subsequently, the biomass concentration of an unknown sample was determined by comparing the measured absorbance of an appropriately diluted sample with the appropriate calibration curve. The biomass concentration C b (g L−1) of the sample was calculated using the equation (Chatsungnoen and Chisti 2016a):

$$ {C}_b=\frac{A_{680}\times \mathrm{D}\ }{\mathrm{S}} $$
(1)

where A 680 is the measured absorbance, D is the dilution factor, and S is the slope of the calibration plot. The S values (and the regression coefficients, r 2) were 5.1949 (r 2 = 0.997) for C. vulgaris grown in freshwater BG11, 7.5673 (r 2 = 0.999) for C. minor, 3.3285 (r 2 = 0.997) for Neochloris sp., 2.1368 (r 2 = 0.999) for C. vulgaris grown in BG11 seawater, 5.6773 (r 2 = 0.999) for N. salina, and 1.9134 (r 2 = 0.998) for C. fusiformis (Chatsungnoen 2015; Chatsungnoen and Chisti 2016a).

The measured biomass concentrations were used to calculate the biomass productivity (P b , g L−1 day−1) as follows:

$$ {P}_b=\frac{X_f-{X}_i}{t} $$
(2)

where X f (g L−1) is the final biomass concentration, X i (g L−1) is the initial biomass concentration, and t (day) is the time required for growth.

The specific growth rate (μ, day−1) was calculated from the slope of the linear regression line of the semilog plot of the biomass concentration versus time during the exponential growth phase; thus,

$$ \mu =\frac{ln\left({X}_2/{X}_1\right)}{t_2-{t}_1} $$
(3)

where X 1 and X 2 are the biomass concentrations (g L−1) at times t 1 and t 2, respectively, within the exponential growth phase.

Biomass recovery

A given batch of an algal broth was divided into two or more parts. The biomass from one part was recovered by centrifugation. The other parts were used to recover the biomass by flocculation–centrifugation. The centrifugal recovery was always at 8370×g (4 °C, 10 min; Hitachi bimac CR22GII centrifuge with rotor number R9A; Hitachi Kok Co., Ltd., Japan). In all cases, the recovered biomass was washed three times with deionized water (for algae grown in freshwater media) or 0.5 M ammonium formate (for algae grown in seawater media). Each wash volume was the same as the volume of the broth sample used in recovering the biomass. The recovered biomass paste was used for various extraction experiments. A part of the biomass was freeze-dried and then pulverized in a grinder (Breville Model CG2B, China) for other parallel extraction experiments.

For biomass recovery by flocculation, two different flocculants were used in separate experiments for a given batch of algal broth. The flocculants were aluminum sulfate (Al2(SO4)3·18H2O; Riedel-de Haen, Germany) and ferric chloride (FeCl3·6H2O; Acros Organics, Belgium). Stock solutions of the inorganic flocculants were made at 20 g L−1 concentration. The 20 g of salt (calculated as FeCl3 or Al2(SO4)3) was dissolved in deionized water, and the volume was made up to 1 L. The stock solutions were kept at room temperature (24–26 °C) and used within 14 days. All flocculant concentrations given here are for the nonhydrated forms of the salts, but the hydrated salts were used in the experiments (Chatsungnoen 2015; Chatsungnoen and Chisti 2016a).

For flocculation, 200 mL of the algal broth was mixed rapidly (80 rpm; 250-mL beaker) for 2-min with the required volume of the flocculant solution. The mixing was then continued at 20 rpm for 30-min to form and grow the flocs. The flocs were then allowed to settle for 30-min. The supernatant was discarded. The flocculated biomass slurry was further concentrated to a paste by centrifugation as specified above. The recovered paste was washed three times with either deionized water or 0.5 M ammonium formate, as specified above. The final recovered paste was used for various extraction studies. A portion of the paste was freeze-dried as explained above, for comparative extraction studies. The detailed flocculation methodology has been previously described (Chatsungnoen and Chisti 2016a).

For a given concentration of the biomass in the original broth, a previously identified optimal flocculant dosage was used for flocculation (Chatsungnoen 2015; Chatsungnoen and Chisti 2016a). The optimal dosages per gram of biomass (dry basis) are shown in Table 1.

Table 1 The dosages of the flocculants
Full size table

The flocculant mass that adsorbed irreversibly to the algal biomass was determined gravimetrically. A sample of the algal broth was divided into three parts. The biomass from the first part (the control) was recovered by centrifugation without the use of any flocculant, suitably washed, dried, and weighed as explained above. The second and the third parts of the broth were separately flocculated with the specified dose of aluminum sulfate and ferric chloride. The biomass was recovered by centrifugation, washed, and dried as explained above. The difference in weights of the dry biomass from the control sample and the flocculated sample was the weight of the flocculant adsorbed to the biomass.

Moisture content of biomass paste

The water content (w/w%) of the biomass paste was determined by drying it to a constant weight. Two methods of drying were compared for this determination. These were oven drying (105 °C, 3 h) and freeze-drying.

The moisture content of the oven dried biomass was calculated using the following equation:

$$ m=\frac{\left({W}_1-{W}_2\right)}{W_1}\times 100 $$
(4)

where m (w/w%) is the moisture content in the biomass paste, W 1 (g) is the weight of the sample before drying, and W 2 (g) is weight of the sample after drying. The dry biomass content (p, w/w%) of the paste was then,

$$ p = 100 - \mathrm{m} $$
(5)

In freeze-drying, a given quantity of paste biomass was freeze dried. The weight of the biomass before and after drying was used to calculate the moisture content as explained above for the oven drying method.

Total lipids

The total lipids in the biomass were measured using a modification of the Bligh and Dyer (1959) method. Depending on the experiment, the washed biomass used in extraction had been recovered by (1) centrifugation and freeze-dried, (2) centrifugation and left as a paste, and (3) flocculation/centrifugation and left as a paste. Total lipids were extracted from 1 g of freeze-dried biomass or 1 g dry equivalent of the paste biomass, using a solvent mixture of 5 mL of chloroform, 10 mL of methanol, and 4 mL of water (Chatsungnoen 2015; Chatsungnoen and Chisti 2016a). For the paste biomass samples recovered by flocculation, the 1 g (dry basis) sample used in extraction included only the mass of the alga and not the mass of the adsorbed flocculant.

The slurry of the solvents and the cells was homogenized for 3 min and stirred for a further 4 h at 760 rpm with a magnetic stirrer at room temperature. Chloroform (5 mL) was then added, and the mixture was blended for 30 s. Distilled water (5 mL) was added, and blending was continued for a further 30 s. The suspension was centrifuged (4000×g, 4 °C, 10 min) and allowed to separate into three layers. The top methanol/water layer was discarded. The chloroform layer (the bottom layer) was collected. The middle layer was the residual biomass. This was extracted again, as specified above for the fresh biomass and the chloroform layer was recovered. The residual biomass was then extracted a third time by contacting with 5 mL of chloroform for 1 h. The three chloroform extracts were combined. The volume of the combined extract was measured in a graduated cylinder. The total lipids in the extract were determined gravimetrically by evaporating an aliquot of the extract in a weighed aluminum dish for 12 h (room temperature in the fume hood) and further drying in a desiccator (12 h, room temperature). Using the measured volume of the pooled chloroform, the total lipid concentration in the extract, and the amount of dry biomass used in extraction, the total lipid content could be calculated as a weight percent of the dry biomass (Chatsungnoen 2015; Chatsungnoen and Chisti 2016a).

Depending on the size of the biomass sample, the above method was scaled in such a way that the volume ratio of chloroform, methanol, and water remained at 1:2:0.8 (monophasic step or the extraction step) and 2:2:1.8 (biphasic step or the separation step) (Chatsungnoen 2015; Chatsungnoen and Chisti 2016a).

In experiments involving the microalgal paste, the paste obtained after the centrifugation and washing steps was used for lipid extraction. The dry solids content of the paste was around 28 % (g per 100 g). This was determined exactly by drying a portion of the paste and comparing its pre-dried mass to that of the corresponding dried biomass. The lipids were extracted from an amount of the paste biomass that was equivalent to 1 g of the dry mass. In comparing lipid extraction from dry and paste biomass, the total lipids recovered from the 1 g dry-equivalent of the biomass paste were compared to the total lipids recovered from 1 g of the freeze-dried biomass. All lipid extractions were in triplicate.

Statistical analyses

Three replicate runs were used to grow the biomass of each alga. The data were used to calculate the growth parameters and their standard deviations. The average values of the lipid content of three replicates and their standard deviations were calculated. The mean values of the control and treatments were compared using the t test. The data analysis used the SAS software (version 9.1, SAS Institute Inc., USA).

Results and discussion

Biomass growth and lipid contents

The biomass growth curves of the algae are shown in Fig. 1. The replicate growth curves were highly reproducible (Fig. 1). The growth pattern generally comprised a short lag phase, a phase of exponential growth, a phase of progressively slowing growth due to light limitation caused by self-shading of cells, and a final stationary phase (Fig. 1). The biomass for extraction was harvested 30–55 days post-inoculation. This was consistent with the norms for commercial processes for producing algal biomass. For example, Pulz (2001) noted a batch culture period of 6–8 weeks post-inoculation in large raceway ponds and 2–4 weeks in photobioreactors.

Fig. 1
figure 1

Growth curves of the microalgae. Data are presented as mean values ± standard deviation of triplicate runs

Full size image

The batch culture kinetic parameters based on the growth curves (Fig. 1) are shown in Table 2. The freshwater microalga Neochloris sp. exhibited the highest biomass productivity of 0.08 ± 0.02 g L−1 day−1. C. vulgaris grown both in freshwater and seawater had the second highest biomass productivity (0.07 ± 0.00 g L−1 day−1) of the algae tested. Significantly, the biomass production of C. vulgaris, normally a freshwater alga, was not affected by growth in full-strength seawater. For the freshwater microalgae, the specific growth rates (i.e., during exponential growth phase) were generally higher than for the marine microalgae (Table 2). The specific growth rate of C. vulgaris grown in freshwater was ~50 % greater than the growth rate in full-strength seawater.

Table 2 The productivity of biomass and lipids
Full size table

From the perspective of oil production, the lipid productivity is of course more important than the biomass productivity (Griffiths and Harrison 2009; Rodolfi et al. 2009; Huerlimann et al. 2010). The lipid productivity of the algae harvested at the end of the batch culture (Fig. 1) is shown in Table 2. The highest lipid productivity (31.4 ± 0.6 mg L−1 d−1) was found for the marine alga N. salina (Table 2). The lipid productivity of C. vulgaris was relatively high (Table 2) and was not significantly affected by whether the alga was grown in seawater or freshwater (Table 2).

In addition to lipid productivity, the lipid weight fraction in the dry biomass is important. This is because less biomass needs to be extracted for a given amount of lipid if the biomass has a high lipid content. The lipid content of the microalgae at harvest ranged from 17.5 to 50.9 % by dry weight, depending on the alga (Table 2). The marine alga N. salina had the highest lipid content of 50.9 ± 0.1 % by dry weight. The microalga C. vulgaris had a relatively high lipid content of >30 % by dry weight (Table 2). Although the specific growth rate of C. vulgaris, a freshwater alga, was reduced in the seawater medium (Table 2), the biomass grown in seawater had a significantly higher lipid content (36.1 ± 1.1 %) compared to the biomass grown in freshwater (lipid content of 33.2 ± 1.7 %; Table 2). As a consequence, the lipid productivity in seawater was essentially the same as in freshwater growth (Table 2). No nutritional limitations were imposed during these cultures (Fig. 1), and therefore, the lipid content of the biomass was not unusually high (Table 2) as would commonly occur in nutritionally stressed biomass.

Equivalence of oven drying and freeze-drying of the biomass

Most literature measurements of the biomass concentration during growth have relied on the oven dry weight method or an optical density method calibrated against the oven dry weight. In contrast, the biomass used in extracting the oil is almost always freeze-dried, as oven drying has the potential to damage the lipids. As calculations of oil productivity interchangeably use the oven-dried biomass concentration and the freeze-dried biomass concentration, an assurance is needed that these two measurements are actually identical or reasonably close. Therefore, the moisture content of identical samples of the biomass paste of the various algae was determined both by oven drying and freeze-drying.

As shown in Fig. 2, the oven drying and freeze-drying gave statistically identical values. All data shown in Fig. 2 are for biomass recovered by centrifugation, without the use of flocculants. The moisture content of the washed paste depended on the alga (Fig. 2). The moisture content ranged from 67 to 88 % on a weight basis. These values are consistent with literature data. For example, a moisture content of ~80 % (w/w) has been reported for the biomass of Nannochloropsis oculata and Dunaliella salina recovered using centrifugation (Tanzi et al. 2013). In view of the results (Fig. 2), oven drying and freeze drying are equally effective in drying the algal biomass for the purpose of determining the moisture content.

Fig. 2
figure 2

A comparison of the moisture content (w/w%) of the biomass paste measured by oven drying and freeze-drying. Identical paste samples were used for the two determinations. The biomass had been recovered by centrifugation without the use of flocculants. Mean values of triplicate measurements ± standard deviations are shown

Full size image

Lipid extraction: paste versus dry biomass

The biomass samples used for extraction in this part of the work were recovered by centrifugation. No flocculants were used. The same biomass samples were extracted as paste or after freeze-drying.

A modified Bligh and Dyer (1959) solvent extraction was used, as it is a widely accepted standard method for complete extraction of microalgal lipids (Mercer and Armenta 2011). This method was originally developed to extract lipids from fish tissue and can be used to extract materials containing nearly 80 % water (Iverson et al. 2001). This notwithstanding, most oil extraction of microalgae reported in the literature has used freeze-dried biomass. Freeze-drying, or any form of drying, is too expensive and energy intensive for use in any process for producing fuel oils from algae (Chisti 2010; 2013). Therefore, an ability to quantitatively extract the oil directly from the biomass paste is important (Suganya and Renganathan 2012; Liu et al. 2013; Tanzi et al. 2013). As shown in Fig. 3, in nearly all cases, the apparent lipid content of the biomass was identical irrespective of whether the freeze-dried biomass or the paste biomass samples were used in the determination. Thus, the modified Bligh and Dyer (1959) method is confirmed as effective for the biomass paste as it is for the freeze-dried biomass. This is so for a variety of microalgae including freshwater species, marine species, and the diatoms.

Fig. 3
figure 3

A comparison of the total lipid content of the biomass determined by standard solvent extraction of the freeze-dried biomass and the wet paste biomass. Mean values of triplicate measurements ± standard deviations are shown

Full size image

Effect of flocculants on lipid extraction

The inorganic metal salt flocculants adhere irreversibly to the microalgal biomass (Henderson et al. 2010) and have the potential to interfere with lipid extraction. Therefore, experiments were performed to quantify the possible effects, if any, of the adsorbed flocculants on the extractability of the oil.

As shown in Fig. 4, the amount of the flocculant adsorbed to the biomass depended on the microalgal species. The highest retention of flocculant was found for C. minor where 68.3 ± 5.9 % of the aluminum sulfate dosage used was retained. Similarly, C. minor retained 76.3 ± 3.4 % of the ferric chloride used. C. vulgaris (seawater) generally retained the least flocculant (Fig. 4). In general, the algae that had a higher value of the optimal flocculant dose also retained a higher fraction of that dose (Chatsungnoen 2015; Chatsungnoen and Chisti 2016a). The differences shown in Fig. 4 are due mainly to two factors: (1) The differences in the cell sizes of the microalgae mean that for a given quantity of the biomass, the total cell surface area is different; and (2) differences in the areal density of the negatively charged functional groups (i.e., the groups responsible for adsorbing the cationic flocculant) on different algae.

Fig. 4
figure 4

Flocculant quantity (% of applied) irreversibly retained on the washed microalgal biomass. Mean values of duplicate measurements ± standard deviations are shown

Full size image

Irrespective of the algal species and the quantity of the flocculant retained on the biomass, the flocculants had no significant effect on recovery of the total lipids by extraction (Table 3). All biomass samples were extracted as paste (Table 3). The biomass samples for any given alga were from the same batch of the broth. The control paste was recovered by centrifugation without the involvement of any flocculants (Table 3). Within statistical limits, the lipid content determined by extraction of the control paste was essentially the same as the lipid content determined for the paste recovered by flocculation–centrifugation (Table 3).

Table 3 Measured lipid contents of the biomass recovered by flocculation and the same biomass recovered by centrifugation (i.e., control, no flocculants)
Full size table

A similar lack of any effect of flocculants on lipids recovery from the biomass was reported by Borges et al. (2011) for the microalgae Nannochloropsis oculata and Thalassiosira weissflogii. The biomass had been flocculated using polyacrylamide flocculants, unlike the inorganic flocculant salts used in the present study. The solvent extraction protocol used by Borges et al. (2011) was also different, and the extraction solvent system was chloroform/methanol in the volume ratio of 2:1.

Large-scale flocculation is practiced in certain industrial processes. By allowing most of the water to be easily removed, flocculation reduces the cost of further downstream biomass recovery operations. Its large-scale use for algal biomass recovery is technically feasible as discussed by Chatsungnoen and Chisti (2016b).

Utility of chloroform/methanol for large-scale lipid extraction

There is no fundamental constraint to large-scale use of chloroform and methanol. Both are widely used industrial chemicals. They can be potentially safely used in algal biomass extraction processes. For example, the annual global use of chloroform in other industries is around 200,000 t, and nearly 100 million t of methanol is used each year world-wide. Toxicity considerations notwithstanding, both methanol and chloroform are natural solvents. Some 660,000 t of chloroform is released into the environment annually. Nearly 90 % of this release is from natural sources. For example, certain seaweeds produce chloroform. Natural release from oceans appears to be the main source of biogenic chloroform.

In a solvent extraction process using chloroform/methanol, both the solvents can be readily recycled as they both have low boiling points (<65 °C). Furthermore, extraction processes can be designed for total containment of solvents.

Concluding remarks

For all algae, solvent extraction of oils from the biomass paste (~77 % moisture, w/w) was as effective as extraction from freeze-dried biomass. Therefore, there was no need of a prior freeze-drying step. For the biomass recovered by flocculation, a certain fraction of the cationic flocculants used in the process attached irreversibly to the negatively charged algal cells. This adhering flocculant did not interfere with the solvent extraction of the oil from the biomass. The Bligh and Dyer (1959) solvent mixture of chloroform, methanol, and water was as effective in extracting the oils from the flocculated biomass paste as from the biomass paste recovered without the use of the flocculants.