Extraction of oil from microalgae for biodiesel production: A review

Extraction of oil from microalgae for biodiesel production: A review

Biotechnology Advances 30 (2012) 709–732 Contents lists available at SciVerse ScienceDirect Biotechnology Advances journal homepage: www.elsevier.co...

3MB Sizes 0 Downloads 144 Views

Biotechnology Advances 30 (2012) 709–732

Contents lists available at SciVerse ScienceDirect

Biotechnology Advances journal homepage: www.elsevier.com/locate/biotechadv

Research review paper

Extraction of oil from microalgae for biodiesel production: A review Ronald Halim ⁎, Michael K. Danquah, Paul A. Webley Bio Engineering Laboratory (BEL), Department of Chemical Engineering, Monash University, Victoria 3800, Australia

a r t i c l e

i n f o

Article history: Received 8 March 2011 Received in revised form 14 November 2011 Accepted 4 January 2012 Available online 11 January 2012 Keywords: Microalgae Biodiesel Oil extraction Lipid extraction Organic solvent Supercritical carbon dioxide Downstream process Pre-treatment Direct transesterification

a b s t r a c t The rapid increase of CO2 concentration in the atmosphere combined with depleted supplies of fossil fuels has led to an increased commercial interest in renewable fuels. Due to their high biomass productivity, rapid lipid accumulation, and ability to survive in saline water, microalgae have been identified as promising feedstocks for industrial-scale production of carbon-neutral biodiesel. This study examines the principles involved in lipid extraction from microalgal cells, a crucial downstream processing step in the production of microalgal biodiesel. We analyze the different technological options currently available for laboratoryscale microalgal lipid extraction, with a primary focus on the prospect of organic solvent and supercritical fluid extraction. The study also provides an assessment of recent breakthroughs in this rapidly developing field and reports on the suitability of microalgal lipid compositions for biodiesel conversion. © 2012 Elsevier Inc. All rights reserved.

Contents 1. 2. 3. 4.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Microalgal lipid composition . . . . . . . . . . . . . . . . . . . . . . . Overview of downstream processes . . . . . . . . . . . . . . . . . . . . Lipid extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Organic solvent extraction . . . . . . . . . . . . . . . . . . . . . 4.1.1. Basic principles . . . . . . . . . . . . . . . . . . . . . . 4.1.2. Solubility parameters . . . . . . . . . . . . . . . . . . . 4.1.3. Selection of organic solvents . . . . . . . . . . . . . . . . 4.1.4. Operating variables . . . . . . . . . . . . . . . . . . . . 4.1.5. Modifications to organic solvent extraction . . . . . . . . . 4.2. Supercritical fluid extraction . . . . . . . . . . . . . . . . . . . . 4.2.1. Basic principles . . . . . . . . . . . . . . . . . . . . . . 4.2.2. Operating variables . . . . . . . . . . . . . . . . . . . . 4.3. Comparison between organic solvent extraction and SCCO2 extraction 5. Effect of cellular pre-treatment on lipid extraction . . . . . . . . . . . . . 6. Simultaneous extraction and transesterification of microalgal lipids . . . . . 7. Microalgal biorefinery . . . . . . . . . . . . . . . . . . . . . . . . . . 8. OriginOil Single-Step Extraction of microalgal lipids . . . . . . . . . . . . 9. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

⁎ Corresponding author. Tel.: + 61 3 990 53420. E-mail address: [email protected] (R. Halim). 0734-9750/$ – see front matter © 2012 Elsevier Inc. All rights reserved. doi:10.1016/j.biotechadv.2012.01.001

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

710 710 712 715 716 716 718 718 720 721 722 722 723 724 726 729 730 730 731 731 731

710

R. Halim et al. / Biotechnology Advances 30 (2012) 709–732

1. Introduction The search for sustainable and renewable fuels is becoming increasingly important as a direct result of climate change and rising fossil-fuel prices. Current commercial production of biodiesel or fatty acid methyl ester (FAME) involves alkaline-catalyzed transesterification of triglycerides found in oleaginous food crops with methanol. However, cultivation of these food crops for biodiesel (mainly rapeseed in Europe and soybean in the US) is no longer sustainable as it requires substantial arable land and consumes large amounts of freshwater (Chisti, 2007). Microalgae are currently considered to be one of the most romising alternative sources for biodiesel (Sheehan et al., 1998). Since many microalgal strains can be cultivated on non-arable land in a saline water medium, their mass farming does not place additional strains on food production (Widjaja et al., 2009). Their high photosynthetic rates, often ascribed to their simplistic unicellular structures, enable microalgae not only to serve as an effective carbon sequestration platform but also to rapidly accumulate lipids in their biomass (up to 77% of dry cell mass). Even using a conservative scenario, microalgae are still predicted to produce about 10 times more biodiesel per unit area of land than a typical terrestrial oleaginous crop (Chisti, 2007; Rosenberg et al., 2008; Sheehan et al., 1998; Shenk et al., 2008). There are, however, various technological and economic obstacles which have to be overcome before industrial-scale production of microalgal biodiesel can take place. The selection and successful outdoor large-scale cultivation of a robust microalgal strain, which has optimum neutral lipid content, possesses an elevated growth rate, and is immune towards invasion by local microbes, remain a major upstream challenge (Sheehan et al., 1998). On the other

hand, the development of an effective and energetically efficient lipid extraction process from the microalgal cells is critical for the successful upscaling of the downstream processes. Despite the routine use of laboratory-scale extraction protocols to determine microalgal lipid contents, the variables affecting lipid extraction from microalgal cells are not well understood and no method for industrial-scale extraction is currently established (Halim et al., 2011). This paper attempts to address the knowledge gap surrounding microalgal lipid extraction by summarizing and critiquing recent studies in the field. We report on the suitability of microalgal lipid compositions for biodiesel conversion and review the different conventional downstream bioprocessing steps required for microalgal biodiesel production. We then examine the technologies currently available for laboratory-scale microalgal lipid extraction, paying special attention to the use of organic solvent extraction and supercritical fluid extraction. We conclude with an assessment on how different cellular pre-treatment processes can effect microalgal lipid extraction as well as with an update on the recent advances in the field, such as the development of a simultaneous microalgal lipid extractionmethylation method and the establishment of a novel ‘single-step’ microalgal lipid extraction method by OriginOil, Inc. 2. Microalgal lipid composition A fatty acid (FA) molecule consists of a hydrophilic carboxylate group attached to one end of a hydrophobic hydrocarbon chain (Fig. 1). Fatty acids are constituents of lipid molecules (both neutral and polar) and designated based on their two most important features ‘the total number of carbon atoms in the hydrocarbon chain: the number of double bonds along the hydrocarbon chain’. Saturated fatty acids have no double bond, while unsaturated fatty

Fig. 1. (a) Fatty acid chains. Saturated fatty acid (C18:0 or stearic acid) on the left. Unsaturated fatty acid (C18:1 or oleic acid) on the right. Oleic acid is of cis-isomerism. (b) Lipid molecules. Triacylglycerol (neutral lipid) on the left. Phospholipid (polar lipid) on the right. R′, R″, R‴ in the triacylglycerol molecule represent fatty acid chains. Phospholipid molecule is negatively charged. Modified from Nelson and Cox (2000).

R. Halim et al. / Biotechnology Advances 30 (2012) 709–732

acids consist of at least one double bond (Nelson and Cox, 2000). When the carboxylate end of the fatty acid molecule is bonded to an uncharged head group (e.g. glycerol), a neutral lipid molecule is formed (e.g. triacylglycerol). On the other hand, the association of a fatty acid molecule to a charged head group (e.g. glycerol and phosphate complex) forms a polar lipid molecule (e.g. phospholipid). Lipids can be defined as any biological molecule which is soluble in an organic solvent. As mentioned above, most lipids contain fatty acids (Fig. 1) and can generally be classified into two categories based on the polarity of the molecular head group (Kates, 1986a): (1) neutral lipids which comprise acylglycerols and free fatty acids (FFA) and (2) polar lipids which can be further sub-categorized into phospholipids (PL) and glycolipids (GL). Neutral lipids are used primarily in the microalgal cells as energy storage, while polar lipids pack in parallel to form bilayer cell membranes. Acylglycerol consists of fatty acids ester-bonded to a glycerol backbone and is categorized according to its number of fatty acids: triacylglycerols (TG), diacylglycerols (DG), monoacylglycerols (MG). FFA is a fatty acid bonded to a hydrogen atom. It is noted that there are also some types of neutral lipids that do not contain fatty acids, such as hydrocarbons (HC), sterols (ST), ketones (K), pigments (carotenes and chlorophylls). Even though these lipid fractions are soluble in organic solvents (hence fitting the definition of lipids), they are not convertible to biodiesel. Readers are referred to other sources for a more comprehensive description regarding lipids, their varieties, and their molecular structures (Kates, 1986a; Mathews and van Holde, 1996; Volkman et al., 1989). The term ‘oil’ is often used to refer to any lipid fraction that exists as a liquid at ambient conditions. Since lipids, especially those obtained from microalgae, are extracted as composite mixtures consisting of various fractions, they do not always exist as liquids. As such, the term ‘oil’ is not used in any part of this study. Microalgal lipid content varies considerably from one species to another and could range, in terms of dry biomass, from 5 to 77 wt.% (Brown et al., 1997; Chisti, 2007). Brown et al. (1997) studied the nutritional properties of 40 different Australian microalgal species and concluded that they comprise, as a weight fraction of dry cell mass, between 5 and 20% lipids. Microalgal lipid composition also varies considerably from one species to another (Brown et al., 1997). During their study investigating the lipid compositions of various microalgal species, Lv et al. (2010) demonstrated that some microalgal species are richer in neutral lipids than other species. The composition and fatty acid profile of lipids extracted from a particular species is further affected by the microalgal life cycle as well as the cultivation conditions, such as medium composition, temperature, illumination intensity, ratio of light/dark cycle, and aeration rate (Guzman et al., 2010; Ota et al., 2009; Ramadan et al., 2008; Rao et al., 2007). Microalgal cells harvested during the stationary phase have lower polar lipid contents than the same species obtained during the logarithmic phase (Dunstan et al., 1993). Some microalgal species have been known to increase their lipid contents from ~ 10 wt.% to almost 20 wt.% during oxygen deprivation (Dunstan et al., 1993). Microalgal cells generally respond to nutrient starvation by intensifying the metabolic pathway which synthesizes neutral lipids. However, this increase in cellular lipid production usually does not result in an overall increase in oil productivity per unit mass as it is often performed by sacrificing growth and through the cessation of cell division. Due to the aforementioned inter- and intraspecific variations, the suitability of microalgal lipids for biodiesel production is difficult to assess and often needs to be examined on a case-by-case basis. Acylglycerols are desirable for commercial-scale biodiesel production for two main reasons. Firstly, industrial-scale alkaline-catalyzed transesterification is designed to process acylglycerols (TG, DG, and MG) and has limited efficacies on other lipid fractions, such as polar lipids and free fatty acids (Christie, 2007; Lang et al., 2001). Secondly,

711

since acylglycerols generally have a lower degree of unsaturation than other lipid fractions (i.e. polar lipids), they produce FAME with higher oxidation stability. As shown by the lipid profiles of three microalgal species (Nannochloropsis oculata, Pavlova lutheri, Isochrysis sp.) in Fig. 2, microalgal lipids usually comprise high levels of polar lipids and non-acylglycerol neutral lipids (HC, ST, K, FFA). As such, they often need to be purified before they can be transesterified. When comparing lipids obtained from logarithmic and stationary growth phase, stationary-phased lipids, despite having an abundance of polar lipids at 51–57 wt.%, contain higher levels of TG (20–41 wt.% of total lipid) and appear more attractive for biodiesel processing than logarithmic-phased lipids (Dunstan et al., 1993). Microalgal fatty acids range from 12 to 22 carbons in length and can be either saturated or unsaturated. The number of double bonds in the fatty acid chains, however, never exceeds 6 and almost all of the unsaturated fatty acids are cis isomers (Medina et al., 1998). The fatty acid profile of lipid extracted from Tetraselmis suecica during early stationary phase is shown in Fig. 3 (Volkman et al., 1989). T. suecica is a common green microalga and its fatty acid profile is used here as an example to illustrate the suitability of microalgal lipids for biodiesel synthesis. Having C16:0, C18:1, and C18:2 as its the principal fatty acids, Tetraselmis lipid appears to have the required fatty acid profile for conversion to high-quality biodiesel. Saturated

neutral lipids breakdown

neutral lipids breakdown

neutral lipids breakdown

Fig. 2. Compositions of crude lipids extracted from three microalgal species during logarithmic phase and stationary phase. Top: Nannochloropsis oculata, middle: Pavlova lutheri, bottom: Isochrysis sp. For neutral lipids, TG: triacylglycerols, HC: hydrocarbons, ST: sterols, K: ketones, FFA: free fatty acids. Modified from Dunstan et al. (1993).

712

R. Halim et al. / Biotechnology Advances 30 (2012) 709–732

30 25 20 15 10 5

C

14

:0 16 :0 C 18 :0 C 16 : C 1 16 :1 C t 18 :1 C 20 :1 C 16 :2 C 18 :2 C 16 :3 C 18 :3 C 16 :4 C 18 : C 4 20 :4 C 20 :5

0 C

wt% of total fatty acids

a

fatty acid chain

b

No. of double bonds in the fatty acid chain wt% of total fatty acids ≥4 20.8%

0 27.6%

3 14.2% 2 16.2%

1-trans 0.8%

1 20.3%

Fig. 3. Fatty acid composition of crude lipid extracted from the species Tetraselmis suecica at the end of logarithmic phase (the beginning of stationary phase): (a) in terms of fatty acid chain; (b) in terms of number of double bonds in the fatty acid chain. In (a), the letter t after the fatty acid name (C16:1t) denotes trans-isomerism. When no letter t appears, fatty acid is of cis-isomerism. In (b), the word -trans after the number of double bonds denotes that fatty acids are of trans-isomerism. When no isomerism is mentioned, fatty acid is of cis-configuration. Modified from Volkman et al. (1989).

fatty acid content (27.6 wt.%) is relatively low when compared to the total cis-unsaturated content (71.6 wt.%). This is desirable as FAME derived from cis-unsaturated fatty acids often has advantageous cold flow properties (a low cloud point and a low pour point). In contrast to saturated chains which pack rapidly upon temperature decrease to form tight semicrystalline structures, cis-unsaturated fatty acids are prevented from forming regular molecular packing due to the bends imposed by the cis double bonds and consequentially freeze at a much lower temperature (Lang et al., 2001; Mathews and van Holde, 1996). The extracted lipid contains a relatively modest amount of polyunsaturated fatty acids (PUFA) with 4 or more double bonds (C16:4 being the most abundant at 7.9 wt.%). This is desirable as highly unsaturated PUFAs are known to be responsible for the poor volatility, the low oxidation stability, and the tendency for gum formation observed in some oilseed-derived biodiesel (Lang et al., 2001). In terms of lipid classes, it is noted that acylglycerols generally have a lower degree of unsaturation than polar lipids and, as such, are more suited for biodiesel conversion. 3. Overview of downstream processes Fig. 4 shows the downstream processing steps required to produce biodiesel from microalgal biomass (Halim et al., 2011). Table 1 lists the different laboratory-scale technological options currently available for each step. The table also examines the scale-up potential of each technology. After the microalgal culture is harvested from the bioreactor, it is concentrated in a dewatering step. The concentrated microalgal culture is then processed in a pre-treatment step to prepare it for lipid extraction. During lipid extraction, lipids are extracted out of the cellular matrices with an extraction solvent. The lipids are then separated from the cellular debris, isolated from the extraction solvent and any residual water, and finally converted to biodiesel in

the transesterification step. Each of the downstream processes is explained in more detail below. Cultivation of microalgae is performed in either an indoor or an outdoor system (Chisti, 2007). Indoor cultivation systems normally use photobioreactors, while outdoor systems employ either raceway ponds or photobioreactors. In an outdoor system, the microalgae are grown in the open environment where cultivation parameters (temperature and light intensity) are dependant on day-to-day weather conditions. The microalgae grown in such a system often suffer from inconsistent growth rates and are more susceptible to local species invasion. On the other hand, the microalgae grown in an indoor system are placed in a greenhouse-type structure where cultivation conditions can be tightly controlled. Despite providing better protection against local species invasion, the indoor system is not preferred due to its high operating cost (Chisti, 2007). Throughout its cultivation, microalgal culture needs to be aerated with CO2 supply and replenished with growth medium consisting of essential elements, such as nitrogen, phosphorous, and iron (Chisti, 2007). Harvested microalgal culture exists as a dilute aqueous suspension (from 0.1 to 2 g dried microalgal biomass/L culture, depending on cultivation method) and needs to be concentrated in order to reduce the cost of downstream processing (Danquah et al., 2009; Molina Grima et al., 2003). Solid–liquid separation methods, such as centrifugation, filtration, and flocculation, are used to dewater the microalgal culture to a concentration between 10 and 450 g dried microalgal biomass/L culture. Concentrated microalgal culture is referred to as concentrate. When dewatered beyond 200 g dried microalgal biomass/L culture, the concentrate is transformed to a sludge suspension and is often referred to as paste or pellet. Developing a cost-viable and an energy-efficient dewatering technology is currently an active field of research. Among the myriads of dewatering technologies (Table 1), flocculation appears to be the most advantageous due to its low energy requirement (Uduman et al., 2010; Wijffels and Barbosa, 2010). During flocculation, microalgal cells adhere to one another to form heavy aggregates which then settle to become concentrate. Cationic, anionic, and non-ionic polyelectrolytes (or polymer) are typically used to flocculate microalgal cells. More details on microalgal dewatering and flocculation can be found elsewhere (Uduman et al., 2010). Post-dewatering, the concentrate undergoes pre-treatment process aimed at enhancing the efficiency of subsequent lipid extraction (Lee et al., 1998, 2010). As shown in Fig. 4, the pre-treatment process can take different pathways depending on the desired biomass alterations. The pre-treatment can be performed in either a single step or multiple steps. Sometimes, no pre-treatment is performed and the concentrate is directly processed for lipid extraction. In one option of the pre-treatment pathways, the concentrate is exposed to a cell disruption method (such as high-pressure homogenization) which ruptures the cellular structures. This pre-treatment forces the release of intracellular lipids to the surrounding medium, thus assisting the lipid extraction process. Microalgal concentrate that has been subjected to cell disruption process is referred to as disrupted concentrate. In a typical 2-step laboratory-scale pre-treatment pathway, the concentrate is completely dried and then milled into fine powders. This pre-treatment eliminates residual water known to prohibit effective mass transfer during the lipid extraction step and results in the formation of dried powder. Detailed discussion on the types and effects of pre-treatment are deferred to Section 5. During lipid extraction, the pre-treated microalgal biomass (existing as one of the following physical states: concentrate or disrupted concentrate or dried powder) is exposed to an eluting extraction solvent which extracts the lipids out of the cellular matrices. Concentrate or disrupted concentrate still contains a certain level of residual water, while dried powder will be completely devoid of residual water. The principles and processes involved in lipid extraction are discussed in detail in the following section (Section 4).

R. Halim et al. / Biotechnology Advances 30 (2012) 709–732

713

Fig. 4. Process flow diagram showing the downstream processing steps needed to produce biodiesel from microalgal biomass.

Microalgal lipid extraction generally uses either organic solvent or supercritical fluid (such as supercritical carbon dioxide) as an extraction solvent. After lipid extraction, the resulting mixture, consisting of extraction solvent, residual water (only when extraction is performed on

concentrate or disrupted concentrate), lipids, and cell debris, is submitted to a solid–liquid separation method, such as filtration, to remove the cell debris. For organic solvent extraction, a liquid–liquid separation method, such as distillation, vacuum evaporation, or solidphase solvent adsorption, is then employed to remove the extraction

714

R. Halim et al. / Biotechnology Advances 30 (2012) 709–732

Table 1 Different laboratory-scale technologies available for each downstream processing step required to produce biodiesel from microalgae. The scale-up potential of each technology is examined. . * * *: highly scalable. *: lack scalability. Process step Cultivation

Dewatering

Pre-Treatment: cell disruption

Pre-Treatment: drying

Pre-Treatment: particulate size reduction

Lipid extraction

Removal of cell debris from extraction solventresidual water-lipids mixture Removal of extraction solvent and residual water from lipids. This step is not applicable to supercritical fluid extraction.

Lipid fractionation

Transesterification

Technologies

Scale-up potential

raceway ponds

***

photobioreactors

**

agglomeration

***

centrifugation

**

filtration

**

flocculation

***

pressure dewatering

*

ultrasonication

**

high-pressure homogenization

***

french pressing

**

bead miling

** *

microwave

**

chemical lysis (acids & enzymes)

**

osmotic shock

**

oven drying

*

freeze drying

*

spray drying

*

milling with specific sieve

**

crushing with pestle and mortar

*

organic solvent extraction

**

supercritical fluid extraction

**

organic solvent extraction with Soxhlet apparatus

*

ultrasound-assisted organic solvent extraction

**

microwave-assisted organic solvent extraction

**

accelerated organic solvent extraction

**

sub-critical organic solvent extraction

**

filtration

**

centrifugation (not applicable to supercritical fluid extraction)

**

distillation

**

vacuum evaporation

*

solid-phase solvent adsorption

**

liquid chromatography

*

silicic acid column chromatography

*

acid precipitation

**

urea crystallization

**

acid catalyst

***

alkali catalyst

***

solvent and the residual water from the lipids. When non-polar/polar organic solvent mixture is used (refer to Section 4.1.1), residual water is removed from the extraction solvent and the lipids via biphasic separation and decantation. The liquid–liquid separation then removes the extraction solvent from the lipids. On the other hand, for supercritical fluid extraction, pressure decompression returns the extraction solvent as well as the residual water to their gaseous states and results in forced precipitation of the lipids (refer to Section 4.2.1). As such, no extra step is needed for the removal of extraction solvent and residual water. The isolated lipids, referred to as crude lipids or total lipids, can now be gravimetrically quantified. The term ‘total lipids’ is primarily used for analytical purposes. As previously mentioned in Section 2, in addition to acylglycerols, crude lipids obtained from microalgal biomass frequently contain polar lipids and non-acylglycerol neutral lipids (such as free fatty acids, hydrocarbons, sterols, ketones, carotenes, and chlorophylls). From the perspective of biodiesel production, any non-acylglycerol biochemical fraction is a contaminant and will have to be removed from the crude lipids. As such, crude lipids arising from microalgal biomass are often subjected to a fractionation step before they are transesterified. Different purification methods, such as liquid chromatography, acid precipitation, and urea crystallizations, are used for lipid fractionation (Medina et al., 1998). During transesterification, the fatty-acid-containing lipid fractions in the crude lipids are reacted with alcohol (methanol, ethanol, isopropanol, butanol) and converted to fatty acid alkyl esters. When methanol is used, the reaction produces fatty acid methyl ester (FAME) or biodiesel. Either an acid (such as H2SO4) or an alkali (such as NaOH or KOH) can be used as a catalyst for transesterification (Christie, 2007; Volkman et al., 1989). Since alkali catalysts have faster reaction rates (estimated at 4000× faster) and higher conversions than acid catalysts for the transesterification of acylglycerols, they are commercially used in the chemical industry for conversion of plant and animal oils to biodiesel (Huang et al., 2010). As previously noted in Section 2, alkaline-catalyzed transesterification has limited efficacies when applied to non-acylglycerol fatty-acid-containing lipid fractions, such as polar lipids and free fatty acids. During alkaline transesterification of acylglycerols, the catalyst cleaves the ester bonds holding the fatty acids to the glycerol backbone (Fig. 5) (Chisti, 2007). The liberated fatty acids are then reacted with methanol to form FAME. In lab-scale experiments where only small amounts of crude microalgal lipids are available, a large amount of methanol (substantially in excess of stoichiometric requirement) is often added to ensure quantitative transesterification. Once transesterification is completed, the reaction mixture, containing biodiesel, glycerol, reformed alkali catalyst, excess methanol, and un-transesterified lipids, then undergoes posttransesterification purification to remove by-product contaminants (glycerol, alkali catalyst, and excess methanol). A laboratory-scale post-transesterification purification method typically consists of 2 steps. The reaction mixture is left to settle under gravity to induce biphasic partitioning (top biodiesel/un-transesterified lipids phase and bottom glycerol phase). Once the biodiesel/un-transesterified lipids phase is decanted off, it is washed repeatedly with water to eliminate any alkali catalyst and excess methanol (Chisti, 2007; Demirbas, 2008; Demirbas and Karslioglu, 2007). More details on alkaline transesterification of acylglycerols and post-transesterification purification can be found elsewhere (Lang et al., 2001; Wahlen et al., 2011). Analyses of the FAME composition of the purified biodiesel/ untransesterified lipids phase are carried out using a gas chromatography (GC) system. Variables that affect FAME conversion during alkaline transesterification include the molar ratio of acylglycerol to methanol, the molar ratio of acylglycerol to catalyst, the reaction temperature, the reaction time, the FFA content of the crude lipids, and the water content of the

R. Halim et al. / Biotechnology Advances 30 (2012) 709–732

715

Triacylglycerol

Methanol

FAME (biodiesel)

Glycerol

Diacylglycerol

Methanol

FAME (biodiesel)

Glycerol

Monoacylglycerol

Methanol

Free Fatty Acid

Potassium Hydroxide

Monoacylglycerol

Water

FAME (biodiesel)

Soap

Free Fatty Acid

Glycerol

Water

Glycerol

Fig. 5. Various reactions involving the alkali catalyst, KOH or potassium hydroxide. [1], [2], and [3] illustrate the alkaline transesterification of acylglycerol with methanol to produce biodiesel as a main product and glycerol as a by-product. [4] illustrates the undesirable reaction between free fatty acid and KOH to form soap and water (saponification). [5] illustrates the undesirable reaction between acylglycerol (monacylglycerol is used as a representation) and water under an alkaline condition to from free fatty acid and glycerol. Modified from Chisti (2007) and Huang et al. (2010).

crude lipids. As illustrated in Fig. 5, FFA reacts with the alkali catalyst to form soap and water (saponification). As such, if the crude lipids to be reacted have a high FFA content, excess alkali catalyst must always be added in order to compensate for the saponification loss (Huang et al., 2010). Past works on alkaline transesterification of vegetable oil have shown that abundant presence of water in the crude lipids had an adverse effect on the reaction kinetics (Christie, 2007; Lang et al., 2001). As depicted in Fig. 5, water under alkaline conditions irreversibly reacts with acylglycerol to from free fatty acid, the formation of which, as mentioned above, consumes the alkali catalyst for its elimination. In a study by Lepage and Roy (1984), FAME recoveries during the transesterification of TG standards were found to substantially decrease once water content of the standards exceeded 20 wt.%. In order for microalgal biodiesel to be environmentally sustainable, the total CO2 emitted in the downstream processing steps must be

lower than or at least equal to the total CO2 originally captured by the microalgal cells during their cultivation. Therefore, processes selected in each step should aim at minimizing energy consumption. 4. Lipid extraction Depending on its pre-treatment pathway, microalgal biomass to be submitted to lipid extraction can assume one of the following physical states: concentrate or disrupted concentrate or dried powder. During lipid extraction, the microalgal biomass is exposed to an eluting extraction solvent which extracts the lipids out of the cellular matrices. Once the crude lipids are separated from the cell debris, the extraction solvent, and water (only when extraction is performed on concentrate or disrupted concentrate), their mass can be measured gravimetrically. Ideally, a lipid extraction technology for microalgal biodiesel production needs to display a high level of specificity

716

R. Halim et al. / Biotechnology Advances 30 (2012) 709–732

towards lipids in order to minimize the co-extraction of non-lipid contaminants, such as protein and carbohydrates. To reduce downstream fractionation/purification, the lipid extraction technology should also be more selective towards acylglycerols than other lipid fractions that are not as readily convertible to biodiesel, i.e. polar lipids and non-acylglycerol neutral lipids (free fatty acids, hydrocarbons, sterols, ketones, carotenes, and chlorophylls) (Medina et al., 1998). Additionally, the selected technology should be efficient (both in terms of time and energy), non-reactive with the lipids, relatively cheap (both in terms of capital cost and operating cost), and safe (Kates, 1986b). Since dewatering the microalgal biomass beyond a paste consistency (200 g dried microalgal biomass/L culture) is energy intensive, it will be economically beneficial if the selected lipid extraction technology is effective when directly applied to wet feedstock, i.e. concentrate or disrupted concentrate with concentrations between 10 and 200 g dried microalgal biomass/L culture (Halim et al., 2011). In this section, we will review the use of organic solvent extraction for routine lab-scale determination of microalgal lipid contents. We will also review the application of supercritical fluid extraction, an emerging green technology that is currently gaining considerable research attention, for microalgal lipid quantification. 4.1. Organic solvent extraction 4.1.1. Basic principles The principles underlying organic solvent extraction of microalgal lipids are anchored on the basic chemistry concept of ‘like dissolving like’. Due to the interactions between their long hydrophobic fatty acid chains, neutral lipids participate in weak van der Waals attractions between one another and form globules in the cytoplasm (Kates, 1986b; Medina et al., 1998). The proposed mechanism for organic solvent extraction is shown in Fig. 6 and can be divided into 5 steps. When a microalgal cell is exposed to a non-polar organic

solvent, such as hexane or chloroform, the organic solvent penetrates through the cell membrane into the cytoplasm (step 1) and interacts with the neutral lipids using similar van der Waals forces (step 2) to form an organic solvent-lipids complex (step 3). This organic solvent–lipids complex, driven by a concentration gradient, diffuses across the cell membrane (step 4) and the static organic solvent film surrounding the cell (step 5) into the bulk organic solvent. As a result, the neutral lipids are extracted out of the cells and remain dissolved in the non-polar organic solvent. A static organic solvent film is formed due to the interaction between organic solvent and cell wall. This film surrounds the microalgal cell and remains undisturbed by any solvent flow or agitation. Some neutral lipids are, however, found in the cytoplasm as a complex with polar lipids. This complex is strongly linked via hydrogen bonds to proteins in the cell membrane. The van der Waals interactions formed between non-polar organic solvent and neutral lipids in the complex are inadequate to disrupt these membrane-based lipid–protein associations. On the other hand, polar organic solvent (such as methanol or isopropanol) is able to disrupt the lipid–protein associations by forming hydrogen bonds with the polar lipids in the complex (Kates, 1986b; Medina et al., 1998). The mechanism in which the non-polar/polar organic solvent mixture extracts membrane-associated lipid complexes is also proposed in lower half of Fig. 6 and can be divided into 5 steps. The organic solvent (both non-polar and polar) penetrates through the cell membrane into the cytoplasm (step 1) and interacts with the lipid complex (step 2). During this interaction, the non-polar organic solvent surrounds the lipid complex and forms van der Waals associations with the neutral lipids in the complex, while the polar organic solvent also surrounds the lipid complex and forms hydrogen bonds with the polar lipids in the complex. The hydrogen bonds are strong enough to displace the lipid–protein associations binding the lipid complex to the cell membrane. An organic solvent–lipids complex is formed and dissociates away from the cell membrane (step 3). The

5

static organic solvent film

bulk organic solvent cell membrane and cell wall

4

1 3

2 cytoplasm

2

1

nucleus

4

3

5 Fig. 6. Schematic diagram of the proposed organic solvent extraction mechanisms. Pathway shown at the top of the cell: mechanism for non-polar organic solvent. Pathway shown at the bottom of the cell: mechanism for non-polar/polar organic solvent mixture. lipids, ○ non-polar organic solvent,◊ polar organic solvent. Both mechanisms can be described in 5 steps. Step 1: penetration of organic solvent through the cell membrane. Step 2: interaction of organic solvent with the lipids. Step 3: formation of organic solvent–lipids complex. Step 4: diffusion of organic solvent–lipids complex across the cell membrane. Step 5: diffusion of organic solvent–lipids complex across the static organic solvent film into the bulk organic solvent.

R. Halim et al. / Biotechnology Advances 30 (2012) 709–732

717

hexane system = 0.015 g lipid/g dried microalgal biomass, final total lipid yield of hexane/isopropanol system (3/2 v/v) = 0.068 g lipid/g dried microalgal biomass). When a non-polar/polar organic solvent mixture (such as hexane/ isopropanol or chloroform/methanol) is used, both solvents are added simultaneously to the microalgal biomass (existing as one of the following physical states: concentrate or disrupted concentrate or dried powder) in the desired volumetric ratio. Once cell debris is removed using a solid–liquid separation method (such as filtration), biphasic separation of the initially single-phase organic solvent mixture is induced by roughly equivolume addition of the non-polar organic solvent (hexane for hexane/isopropanol mixture and chloroform for chloroform/methanol mixture) and water. Upon complete biphasic separation, neutral and polar lipids will mainly partition in

organic solvent–lipids complex then diffuses across the cell membrane (step 4) and the static organic solvent film surrounding the cell (step 5) into the bulk organic solvent. As such, the addition of a polar organic solvent to a non-polar organic solvent facilitates the extraction of membrane-associated neutral lipid complexes. However, the process also inevitably leads to the co-extraction of polar lipids. In most laboratory practices, both non-polar organic solvent and polar organic solvent are added to the microalgal cells to ensure the complete extraction of all neutral lipids, both in the form of freestanding globules and in the form of membrane-associated complexes. During our previous study investigating lipid extraction from Chlorococcum sp. (Halim et al., 2011), the inclusion of isopropanol as a co-solvent was shown to improve the total lipid yield of pure hexane system by more than 300% (final total lipid yield of pure

a

b

c

concentrate microalgal culture dried

e

hexane/IPA (3/2 v/v)

powder

f

d

cell debris

hexane/ IPA methanol

h

g

KOH

crude lipids biodiesel,

hexane/IPA

glycerol, un-transesterified

hot plate stirrer

lipids Fig. 7. Schematic diagram showing the experimental steps typically undertaken for laboratory-scale production of microalgal biodiesel using an organic solvent mixture as a lipid extraction technology. Hexane/isopropanol (IPA) (3/2 v/v) mixture is used for lipid extraction. Cell disruption, lipid fractionation, and post-transesterification purification are not performed. (a) Cultivation with photobioreactors. (b) Dewatering with a centrifuge. (c) Drying (pre-treatment) with an oven. (d) Particulate size reduction (pre-treatment) with a pestle and a mortar. (e) Lipid extraction with hexane/isopropanol mixture. (f) Removal of cell debris with a filter. (g) Removal of extraction solvent with a distillation unit. (h) Transesterification with an alkali catalyst.

718

R. Halim et al. / Biotechnology Advances 30 (2012) 709–732

the organic phase (a mixture of non-polar organic solvent and polar organic solvent), while the aqueous phase (a mixture of water and polar organic solvent) will contain primarily non-lipid contaminants (proteins and carbohydrates) (Kates, 1986b; Medina et al., 1998). As such, biphasic separation removes not only residual water but also non-lipid contaminants from the mixture of organic solvents and lipids. The organic phase is decanted and evaporated to yield dry crude lipids, which are then fractionated and transesterified. Fig. 7 shows the experimental steps typically undertaken for laboratory-scale production of microalgal biodiesel using an organic solvent mixture as a lipid extraction technology. Table 3 summarizes the methods and the findings of recent studies which investigated organic solvent extraction of microalgal lipids. 4.1.2. Solubility parameters Among the myriad of thermodynamic parameters (polarity index, Kauri-butanol value, and Hildebrand solubility parameter) attempting to predict the solubility of a solute in a solvent, Hansen solubility parameters (HSP) appears to be one of the more widely accepted and promising systems (Gupta et al., 1997; Hansen, 2008; Snyder, 1974). HSP characterization predicts that a solute will dissolve in a solvent if the molecules of either substance have similar force of interaction. With HSP system, total energy of cohesion (E) can be quantitatively divided into three components (Hansen, 2008). These components account for the atomic dispersion (or Van der Waals) interactions (ED), the molecular dipolar interactions (Ep), and the molecular hydrogen bonding (electron exchange) interactions (EH). E ¼ ED þ Ep þ EH

ð1Þ

E is experimentally determined by measuring the energy required to evaporate the liquid (solute or solvent), thus breaking all of its cohesive bonds. Dividing Eq. (1) by the molar volume, V, yields the three components of Hansen solubility parameters. E=V ¼ ðED =VÞ þ ðEp=VÞ þ ðEH =VÞ

ð2Þ

where E/V = δ 2, ED/V = (δD) 2, Ep/V = (δP) 2, and EH/V = (δH) 2 2

2

2

2

δ ¼ ðδD Þ þ ðδP Þ þ ðδH Þ

ð3Þ

where δ is the Hildebrand total solubility parameter, δD is the Hansen dispersion parameter, δP is the Hansen dipolar parameter, and δH is the Hansen hydrogen bond parameter. The SI unit for all of the parameters is MPa 1/2. HSP characterization can be conveniently visualized with a spherical representation. Hansen solubility parameters of the solute are at the center of the solubility sphere and the radius of the solubility sphere (Ro) indicates the extent of interaction for solubilization to occur. Good solvents lie within the solubility sphere

and poor ones lie outside. Ra is the distance of the solvent from the center of the solubility sphere. 2

2

2

2

Ra ¼ 4ðδDS −δDP Þ þ ðδPS –δPP Þ þ ðδHS –δHP Þ

ð4Þ

Subscript s is for the solvent and subscript p is for the solute. If Ra b Ro, the solvent lies within the solubility sphere and the solute is soluble in the solvent. Relative Energy Difference (RED) is numerical representation of this graphical visualization. The smaller its RED value, the better the solvent is at dissolving the solute. RED ¼ Ra=Ro

ð5Þ

For a solvent mixture, composite Hansen solubility parameter, δi, mix, is calculated with: δi;mix ¼ φ1 δi;1 þ φ2 δi;2 þ …

ð6Þ

Subscript i signifies one of the three components of Hansen solubility parameters, φ is volume fraction of each solvent in the mixture. We have computed the HSP characterizations for organic solvents and organic solvent mixtures commonly employed to extract lipids from microalgal biomass in Table 2. All of the Hansen solubility parameters of pure organic solvents were obtained from Hansen, 2007. Eq. (6) was used to calculate composite Hansen solubility parameters for organic solvent mixtures. Triacetin was used as a representative for all types of triacylglycerols. Ra was calculated with triacetin as a solute (Eq. (4)). Ro value of 12 MPa 1/2 indicates marginal solubility of a solute in a solvent and was assigned for RED estimation (Eq. (5)). As can be seen from Table 2, triacetin appears to be highly soluble in hexane/isopropanol mixture (3:2 v/v) (RED = 0.3) and chloroform (RED = 0.4). On the other hand, chloroform/methanol/water mixture (1:2:0.8 v/v/v) appears to be poor solvents for triacetin (RED = 1.2). Such a finding is contradictory to previous lipid extraction works which recommend the use of chloroform/methanol/water mixture (Molina Grima et al., 1994). We note that the current triacetinbased HSP characterizations must be treated with caution. Triacylglycerols with higher carbon number will have different Hansen solubility parameters to triacetin. Additionally, limited availability of Hansen solubility parameters for other desirable neutral-lipid derivatives, such as diacylglycerols, monoacylglycerols, and neutral lipids/polar lipids complexes, prevents the complete characterization of microalgal lipids. More research is needed before the mechanisms underlying organic solvent extraction of microalgal lipids (as outlined in Section 4.1.1.) can be fully explained with HSP characterization. 4.1.3. Selection of organic solvents In addition to satisfying the previously mentioned criteria for an ideal lipid extraction technology, the selected organic solvents should

Table 2 HSP characterizations for organic solvents and organic solvent mixtures commonly employed to extract lipids from microalgal biomass. Ra was calculated with triacetin as a solute. Hansen solubility parameters for triacetin: δD = 16.5 MPa1/2, δP = 4.5 MPa1/2, δH = 9.1 MPa1/2. Organic solvent or organic solvent mixture

Hansen dispersion parameter or δD (MPa1/2)

Hansen dipolar parameter or δP (MPa1/2)

Hansen hydrogen bond parameter or δH (MPa1/2)

Affinity of solvent with triacetin or Ra (MPa1/2)

Relative energy difference or RED

Solubility of triacetin in the solvent

Chloroform Methanol Water Chloroform/methanol (1:2 v/v) Chloroform/methanol/water (1:2:0.8 v/v/v) Hexane Isopropanol Hexane/isopropanol (3:2 v/v) Ethanol

17.8 15.1 15.6 19.0 15.9 14.9 15.8 15.3 15.8

3.1 12.3 16.0 9.8 10.7 0.0 6.1 2.4 8.8

5.7 22.3 42.3 17.7 22.1 0.0 16.4 6.6 19.4

4.5 15.6 35.2 11.2 14.5 10.6 7.6 4.1 11.2

0.4 1.3 2.9 0.9 1.2 0.9 0.6 0.3 0.9

Very good Poor Very poor Good Poor Good Good Very good Good

Table 3 Methods and results summary of recent studies investigating organic solvent extraction of microalgal lipids. Kates, 1986b (method for microalgae)

Guckert et al., 1988

Nagle and Lemke, 1990

Molina Grima et al., 1994

Lee et al., 1998

Fajardo et al., 2007 Halim et al., 2011

Microalgal species

N/A

Any

Chlorella sp.

Chaetoceros muelleri and Monoraphidium minutum

Isochrysis galbana

Botryococcus braunii

Phaeodactylum tricornutum

Chlorococcum sp.

State of microalgal biomass at the start of extraction

Concentrate Concentrate (residual water content = 90 wt.%)

Dried powder by lyophilization

Concentrate (residual Dried powder by lyophilization water content = 85 wt.%)

Concentrate

Dried powder by lyophilization

Concentrate (residual water content = 70 wt.%) or dried powder by thermal drying

Mass of dried microalgal biomass (g)

5 to 6

0.1 Not specified. Mass of wet paste = 1.5 g.

60

5

0.12

10

4

Organic solvents or organic solvent mixtures

Water/methanol/ chloroform (0.8:2:1 v/v/v)

Water/ isopropanol (1:5 v/v)

For Soxhlet: methylene chloride/ methanol (2:1 v/v). For batch extraction 1: chloroform/ methanol/50 mM phosphate buffer (35:70:28 v/v/v/). For batch extraction 2, n-hexane/IPA/ distilled water (70:47.7:3 v/v/v).

1-butanol, ethanol, hexane/2-propanol (2/3 v/v), water/ methanol/chloroform as a control system

Chloroform/methanol/water (1:2:0.8 v/v/v), hexane/ethanol (1:2.5 v/v), hexane/ethanol (1:0.9 v/v), butanol, ethanol, EtOH/water (1:1 v/v), hexane/ isopropanol (1:1.5 v/v)

Chloroform/methanol (2:1 v/v), Ethanol hexane/isopropanol (3:2 v/v), dichloroethane/methanol (1:1 v/v), dichloroethane/ ethanol (1:1 v/v), acetone/ dichloromethane (1:1 v/v)

Amounts of organic solvent or organic solvent mixture added

76 ml organic solvent mixture/ g dried microalgal biomass

16 ml organic solvent mixture/ g concentrate

For Soxhlet: 1000 ml organic solvent mixture/ g dried microalgal biomass. For batch extraction 1: 1330 ml organic solvent mixture/g dried microalgal biomass. For batch extraction 2: 1207 ml organic solvent mixture/ g dried microalgal biomass.

20 g organic solvent mixture/ g dried microalgal biomass

76 ml organic solvent mixture/ g 250 ml organic solvent dried microalgal biomass mixture/ g dried microalgal biomass

5 ml organic solvent mixture/g dried microalgal biomass

75 ml organic solvent mixture/ g dried microalgal biomass

Duration (min)

60

3

90 For Soxhlet: 180. For batch extraction 1: 2160. For batch extraction 2: overnight.

60

50

1440

450

Degree of agitation (rpm)

Not specified

Not specified

For Soxhlet: N/A. For batch extraction 1: not specified. For batch extraction 2: not specified.

High (not specified)

Not specified, but stated to be constant

High (not specified)

500

800

Extraction temperature (°C)

25

50–60

For Soxhlet: not specified. For batch extraction 1: not specified. For batch extraction 2: room.

Near boiling point of each solvent (not specified)

25

Not specified

25

25

N/A

11.9; Soxhlet with methylene chloride/methanol (2:1 v/v)

Not specified as results are expressed as efficiencies to a control system (assigned at 100%). 2nd Highest efficiency with 1-butanol at 94%

8.9; chloroform/methanol/water (1:2:0.8 v/v/v). 2nd Highest efficiency with ethanol at 8.0 wt.% of dried microalgal biomass

28.6; chloroform/ methanol (2:1 v/v)

6.3; ethanol

6.8; hexane/ isopropanol (3:2 v/v)

N/A Maximum final total lipid yield (wt.% of dried microalgal biomass); the organic solvent or the organic solvent mixture to achieve it

Hexane, hexane/ isopropanol (3:2 v/v)

R. Halim et al. / Biotechnology Advances 30 (2012) 709–732

Kates, 1986b (method for microorganisms)

719

R. Halim et al. / Biotechnology Advances 30 (2012) 709–732

preferably be volatile for low-energy distillation from the crude lipids (Kates, 1986b; Medina et al., 1998). Chloroform/methanol (1/2 v/v) is the most frequently used organic solvent mixture for lipid extraction from any living tissue. Using this organic solvent system, residual endogenous water in the microalgal cells acts as a ternary component that enables the complete extraction of both neutral and polar lipids. It is noted that this method does not require the complete drying of microalgal biomass. Once the cell debris is removed, more chloroform and water are added to induce biphasic partitioning. The lower organic phase (chloroform with some methanol) contains most of the lipids (both neutral and polar) while the upper aqueous phase (water with some methanol) constitutes most of the non-lipids (proteins and carbohydrates) (Medina et al., 1998). Extraction using chloroform/methanol (1/2 v/v) is fast and quantitative. Chloroform, however, is highly toxic and its usage is undesirable. The method was originally developed by Folch et al. (1951) for the isolation of total lipids from brain tissues. For this reason, its efficacy in extracting lipids from microalgal biomass still needs further assessment. In a study by Lee et al. (1998), the performance of five different organic solvent mixtures in extracting lipids from bead-beaten Botryococcus braunii cells was compared. As can be seen in Table 3, chloroform/methanol obtained the highest final total lipid yield at ~ 0.29 g/g dried microalgal biomass. On the other hand, dichloroethane-based organic solvent mixtures (dichloroethane/methanol and dichloroethane/ethanol), previously recommended for lipid extraction from the green algae Cladofora, were found to have limited efficacies when applied to B. braunii. Hexane/isopropanol (3/2 v/v) mixture has been suggested as a low-toxicity substitute to chloroform/methanol system (Halim et al., 2011). The mixture works in a similar fashion with chloroform/ methanol system. Upon biphasic separation, the upper organic phase (hexane with some isopropanol) contains most of the lipids (both neutral and polar) while the lower aqueous phase (water with some isopropanol) contains most of the non-lipids (proteins and carbohydrates). When evaluated for microalgal lipid extraction, hexane/isopropanol mixture was found to be more selective towards neutral lipids compared to chloroform/methanol system (Guckert et al., 1988; Lee et al., 1998; Nagle and Lemke, 1990). As previously mentioned, segregation of neutral lipid class at the lipid extraction step is highly desirable as it would allow microalgal biodiesel production to occur with minimal downstream purification. Guckert et al. (1988) attributed the neutral lipid selectivity of hexane/isopropanol mixture to its inability to extract the polar lipid constituents of microalgal membranes (chloroplast membranes contain glycolipids and cell membranes contain phospholipids). The hexane/isopropanol system, however, yielded a surprisingly low total lipid recovery when applied to B. braunii (Lee et al., 1998). Pure alcohol (such as butanol, isopropanol, and ethanol) is cheap, volatile, and has a strong affinity to membrane-associated lipid complex due to its ability to form hydrogen bonds. However, its polar nature is also a disadvantage as it limits interaction with freestanding neutral lipid globules. For this reason, when used as a microalgal lipid extraction solvent, alcohol is almost always combined with a non-polar organic solvent, such as hexane or chloroform, to ensure the total extraction of both forms of neutral lipids (freestanding globules and as membrane-associated complexes) (Halim et al., 2011). In their study, Nagle and Lemke (1990) evaluated the efficiencies of three organic solvents (butanol, hexane/2-propanol mixture, ethanol) in extracting crude lipids from Chaetoceros muelleri and compared them to a control water/methanol/chloroform mixture (Table 3). Even though the control polar/non-polar mixture was verified to be the most effective organic solvent system (assigned an arbitrary extraction efficiency of 100%), butanol (with an average extraction efficiency of 94%) was found to be highly promising with

a final total lipid yield consistently higher than hexane/2-propanol mixture or ethanol in all triplicates. The authors argued that, even though all of the organic solvents used were safe to handle, the consistency of butanol yields from one replicate to another (standard deviation of 3%) indicated a lower sensitivity to changes in the extraction procedure, a beneficial attribute if the process were to be scaled up. Due to its propensity to inactivate many phosphatidases and lipases, Kates (1986b) recommended the use of isopropanolcontaining organic solvent mixture to extract lipids from unicellular algal species that produces lipid degradative enzymes. 4.1.4. Operating variables The evolution of lipids during the organic solvent extraction on microalgal biomass is observed to follow a first-order kinetics equation (Halim et al., 2011; Harrison et al., 2003b):   −kt me ¼ ms;o 1−e

ð1Þ

where me is the amount of lipid extracted in the organic solvent at time t (g lipid/g dried microalgal biomass), ms,o is the amount of lipid originally present in the cells (g lipid/g dried microalgal biomass), k is the lipid mass transfer coefficient from the microalgal cells into the organic solvent (min − 1), and t is the extraction time (min). The parameter k itself is a function of several operating variables and can be described generally as: k ¼ f ðag; s=b; TÞ

ð2Þ

where ag is the degree of agitation (revolution per minute), s/b is the ratio of organic solvent to dried microalgal biomass (ml organic solvent/g dried microalgal biomass), and T is the extraction temperature (°C). Table 3 provides a summary of the levels of operating variables used in various studies investigating organic solvent extraction of microalgal lipids. Fig. 8 (Fajardo et al., 2007) shows typical extraction curves obtained when using an organic solvent to extract lipids from microalgal biomass. These curves conform to the model of Eq. (1), where the rate of lipid recovery is observed to decrease with extraction time. During the 24-hour extraction, a majority of the lipids is recovered within the first 8 h (60–70% of all extractable lipids) and extending the extraction time beyond 12 h does not seem to make any significant contribution to the final total lipid yield. This kind of asymptotic behavior is attributed to the diffusion-driven nature of lipid extraction where the rate of lipid evolution is controlled by the 100

Total lipid yield ( wt% of dried microalgal biomass)

720

90 80 70 60 50 40 30 20 10 0 0

4

8

12

16

20

24

Time (h) Fig. 8. Typical first-order extraction curves obtained for organic solvent extraction of microalgal lipids. Microalgae: Phaeodactylum tricornutum, organic solvent: ethanol. Ratio of organic solvent to dried microalgal biomass (s/b) was optimized. X 5 ml ethanol/g dried microalgal biomass, ▲ 10 ml ethanol/g dried microalgal biomass, ■ 15 ml ethanol/g dried microalgal biomass. Modified from Fajardo et al. (2007).

R. Halim et al. / Biotechnology Advances 30 (2012) 709–732

lipid concentration gradient between the microalgal cells and the organic solvent. Extraction is most rapid in the beginning when concentration gradient is at its steepest. As lipids are removed from the microalgal cells into the organic solvent, the concentration gradient diminishes and lipid extraction slows down. As can be seen in Table 3, every study indicates a different ratio of organic solvent to dried microalgal biomass (s/b). The appropriate s/b value for each microalgal strain varies depending on its lipid content and its intrinsic solvent-cellular interaction. During their investigation of lipid extraction from Phaeodactylum tricornutum, Fajardo et al. (2007) found extraction efficiency to increase with decreasing s/b ratio (Fig. 8). The final total lipid yields obtained with 10 and 15 ml ethanol per gram of dried microalgal biomass were around 80 wt.%. With 5 ml ethanol per gram of dried microalgal biomass, the final total lipid yield was roughly 90 wt.%. Even though this claim was rather counter-intuitive, the authors ascribed the higher total lipid yield at lower organic solvent ratio to its higher agitation intensity per volume unit. It is important to find the optimum s/b value for a specific microalgal strain. An s/b value that is too high results in excessive consumption of organic solvent, while a value that is too low leads to handling difficulties and incomplete extraction. Variation in the extraction temperature has been reported to influence lipid yield. Increasing temperature from 30 °C to 60 °C was observed to enhance lipid extraction rate from animal tissues (Fajardo et al., 2007). However, a rise beyond 70 °C led to an oxidative degradation of thermo-labile components which resulted in a lower lipid yield. In a study by Balasubramanian et al. (2010), increasing temperature during lipid extraction from Scenedesmus obliquus resulted in a significant increase in the final total lipid yield. The authors attributed this increase to enhanced mass transfer kinetics at higher temperature. The kinetics and the mechanism underlying organic solvent extraction of microalgal lipids are not yet well understood and require further research. It is noted that organic solvent extraction has several disadvantages. The method generally uses large amounts of toxic organic solvents, is slow, and requires energy-intensive evaporation for solvent removal. The extent of lipid extraction by a volume of

water out condenser water in

Soxhlet extractor with a thimble to hold the microalgal biomass

heating mantle

round-bottom flask containing the organic solvent Fig. 9. The Soxhlet apparatus. Modified from De Castro and Ayuso (2000).

721

organic solvent is also thermodynamically restricted by the lipid mass transfer equilibrium (Wang and Weller, 2006). Once lipid concentration between the bulk organic solvent and the cellular matrices has reached its partition equilibrium level, no further transfer of lipids from the cells to the organic solvent will take place.

4.1.5. Modifications to organic solvent extraction A majority of the laboratory-scale organic solvent extractions reported in the literature have been performed as a batch process. Even though batch extraction is limited by lipid mass transfer equilibrium, a continuous organic solvent extraction able to overcome this limitation requires a large amount of organic solvent and becomes too expensive. Through its ingenious cycles of solvent evaporation and condensation, the Soxhlet apparatus continuously replenishes cells with fresh organic solvent (hence evading equilibrium limitation) while simultaneously minimizing solvent consumption (Luque de Castro and Garcia-Ayuso, 1998; Wang and Weller, 2006). The apparatus is shown in Fig. 9 and has 3 compartments: a continuously heated roundbottom flask to store the extracting organic solvent, the Soxhlet extractor to hold the microalgal biomass (existing as concentrate or disrupted concentrate or dried powder), and the continuously cooled condenser. Organic solvent from the heated round-bottom flask enters the condenser and is immediately channeled into the Soxhlet extractor. The organic solvent comes in contact with the microalgal biomass and performs lipid extraction. The thimble in the extractor prevents the microalgal biomass from being carried away by the organic solvent flow and, as such, serves as a filter to remove cell debris. Once the organic solvent in the extractor reaches the overflow level, a siphon unloads the organic solvent-lipids mixture from the extractor back into the round-bottom flask. The organic solvent is heated and evaporates again while the extracted crude lipids remain in the round-bottom flask. This cycle is repeated until no more crude lipids are extracted in the Soxhlet extractor. Despite its advantageous design in avoiding equilibrium limitation, the Soxhlet apparatus suffers from high energy requirement for continuous distillation (Luque de Castro and GarciaAyuso, 1998; Wang and Weller, 2006). Independent studies by Guckert et al. (1988) and Halim et al. (2011) verified the superior efficacy of Soxhlet extraction when compared to a batch extraction. Among the three systems experimented by Guckert et al. (1988) to extract lipids from Chlorella sp., Soxhlet extraction using methylene chloride:methanol (2:1 v/v) mixture obtained the highest final total lipid yield (Table 3). In terms of the dry microalgal weight, the final total lipid recovered was approximately 11.9% by Soxhlet extraction using methylene chloride: methanol, 11.1% by batch extraction using chloroform/methanol/50 mM phosphate buffer, and 5.8% by batch extraction using n-hexane/isopropanol/distilled water. Halim et al. (2011) found Soxhlet operation of hexane extraction to be significantly more efficient than its batch counterpart when used to extract lipids from Chlorococcum sp. (final total lipid yield of batch extraction = 0.015 g lipid/g dried microalgal biomass, final total lipid yield of soxhlet extraction = 0.057 g lipid/g dried microalgal biomass). Despite its improved total lipid recovery, Soxhlet extraction potentially suffered from lipid degradation resulting from the use of elevated temperature throughout the process. Guckert et al. (1988) noted that the crude lipids recovered using a Soxhlet system contained less PUFAs then those obtained by batch extractions and ascribed this observation to potential thermal degradation due to the harshness of the Soxhlet method. A couple of modifications to organic solvent extraction have also been introduced: microwave-assisted organic solvent extraction and accelerated or subcritical organic solvent extraction. Each modification utilizes an auxiliary process that enhances the kinetics of lipid extraction by the organic solvent through speedy disruption of the cellular structures (Luque de Castro and Garcia-Ayuso, 1998; Wang and Weller, 2006). Modified organic solvent extraction synergistically

722

R. Halim et al. / Biotechnology Advances 30 (2012) 709–732

combines cell disruption pre-treatment (to be reviewed in Section 5) and lipid extraction as a single step. Microwave-assisted organic solvent extraction uses the aid of electromagnetic radiation within a specific frequency range to deliver large amount of thermal energy to the microalgal cells (Balasubramanian et al., 2010). When the cells receive this energy, local internal superheating occurs leading to instantaneous temperature rise within the matrices and rapid pressure effects on the cell wall/membrane structure. Cell structures are immediately ruptured forcing cell constituents to spill out. This effective expulsion of cell materials facilitates a more rapid diffusion of microalgal lipids into the extracting organic solvent. Microwave-assisted heating is substantially more rapid than conventional heating as heat is delivered via radiation rather than convection and conduction. Balasubramanian et al. (2010) examined the use of microwave-assisted hexane extraction to recover lipids from S. obliquus. Microwave-assisted hexane extractions were found to result in higher oil yields compared to conventionally waterheated hexane extraction control methods at all extraction temperatures and times. During accelerated or subcritical organic solvent extraction, lipid extraction is performed at an elevated pressure and temperature in order to accelerate extraction kinetics and to disintegrate cellular structures. Subcritical organic solvent extraction has some of the benefits of supercritical fluid extraction (described in Section 4.2), but is still performed below critical conditions in order to minimize operating cost (Herrero et al., 2005). Chen et al. (2011) examined the use of subcritical ethanol to extract lipids from wet microalgal paste of Nannochloropsis sp. and found the extraction process to be highly efficient (maximum final lipid recovery= 90.21% of total lipids). Neither of the modifications described above (microwave-assisted or subcritical organic solvent extraction) has been applied to an industrial scale due to their high energy requirements. It is also noted that there is currently limited understanding on the key variables affecting the performances of these modified extraction processes (Luque de Castro and Garcia-Ayuso, 1998; Wang and Weller, 2006). The scale-up potentials of organic solvent extraction and its modifications are examined in Table 1. 4.2. Supercritical fluid extraction Supercritical fluid extraction (SFE) is an emerging green technology that has the potential to replace traditional organic solvent extraction. 4.2.1. Basic principles When the temperature and the pressure of a fluid are raised over their critical values (Tc and Pc), the fluid enters the supercritical region (Fig. 10) (Pourmortazavi and Hajimirsadeghi, 2007; Reverchon and De Marco, 2006; Taylor, 1996). Supercritical fluid appears suitable to be used as an extraction solvent for lipid recovery from microalgal biomass due to the following reasons (Mendes et al., 1995, 2003, 2006; Taylor, 1996): • Tunable solvent power

Fig. 10. P–T phase diagram for carbon dioxide, showing the supercritical region.

Table 4 Physical properties of a typical fluid in different states. Modified from Taylor (1996).

Gas Supercritical fluid Liquid

Density (kg/m3)

Viscosity (μPa∙s)

Diffusion coefficient (mm²/s)

1 100–1000 1000

10 50–100 500–1000

1–10 0.01–0.1 0.001

The solvent power of a supercritical fluid is a function of its density which can be continuously adjusted by changing the extraction pressure and the extraction temperature. As such, solvent power of the fluid can be tuned such that it interacts primarily with neutral lipids (i.e. acylglycerols). • Favorable mass transfer As shown in Table 4, supercritical fluid displays physical properties intermediate to a liquid and a gas (Taylor, 1996). These transitional properties allow for rapid penetration of the fluid through cellular matrices, thus resulting in a higher total lipid yield and a shorter extraction time. • Production of solvent-free crude lipids Crude lipids obtained from supercritical fluid extraction are free from extraction solvent. Therefore, no energy is expended for extraction solvent removal. Supercritical carbon dioxide (SCCO2) is the primary solvent used in the majority of supercritical fluid extractions. Its moderate critical pressure (72.9 atm) allows for a modest compression cost, while its low critical temperature (31.1 °C) enables successful extraction of thermally sensitive lipid fractions without degradation. SCCO2 facilitates a safe extraction due to its low toxicity, low flammability, and lack of reactivity (Macias-Sanchez et al., 2007; Taylor, 1996). If the microalgal cells are to be cultivated at a coal-fired power station, the CO2 required for supercritical conversion can be conveniently obtained from the scrubbed flue gas of the station. This paper will focus on the use of SCCO2 for microalgal lipid extraction. Fig. 11 shows a lab-scale supercritical carbon dioxide (SCCO2) extraction unit used for the recovery of microalgal lipids (Applied separations, 2007). A mixture of microalgal biomass (existing as concentrate or disrupted concentrate or dried powder) and packing materials (normally diatomaceous earth or diatoms) in a specific ratio is placed inside the extraction vessel equipped with a heating element. A feed pump delivers CO2 from its source to the extraction vessel at a pressure greater than Pc. As soon as the vessel is heated (T > Tc), the compressed CO2 is converted to its supercritical state and performs lipid extraction on the microalgal biomass (Fig. 12). During the lipid extraction process, the microalgal biomass and diatoms are packed tightly inside the cylindrical extraction vessel. The supercritical carbon dioxide travels on the surface of the packed mixture and lipids are desorbed from the microalgal biomass. Immediately upon dissolution, the SCCO2 encloses the lipids to form a SCCO2–lipids complex. The complex, driven by concentration gradient, diffuses across the static SCCO2 film and enters the bulk SCCO2 flow. The frits placed at both ends of the extraction vessel prevent the mixture of microalgal biomass and diatoms from being carried away by the SCCO2 flow. As such, the frits serve as a filter to remove cell debris. The SCCO2–lipids mixture, as well as some residual water (if lipid extraction was performed on concentrate or disrupted concentrate), then leaves the extraction vessel to enter the collection vessel, where a micrometering valve is used to rapidly depressurize the incoming fluid (Fig. 11). Upon complete depressurization, the SCCO2, together with any residual water, returns to gaseous state and the extracted crude lipids precipitate in the collection vessel. As such,

R. Halim et al. / Biotechnology Advances 30 (2012) 709–732

723

Fig. 11. Schematic diagram of a laboratory-scale SCCO2 extraction system. The unit is used to extract lipids from microalgal biomass. Modified from Applied separations (2007).

SFE-derived crude lipids are free from any extraction solvent and do not need to undergo an extraction solvent removal step. Even though SCCO2 extraction can be operated as either a batch (static) or a continuous (dynamic) process, it is generally exercised as a continuous extraction as this often results in an improved yield (Taylor, 1996). The process described above is based on a dynamic extraction. The feasibility of applying SCCO2 process to extract microalgal lipids has been demonstrated (Andrich et al., 2005, 2006; Canela et al., 2002; Cheung, 1999; Halim et al., 2011; Herrero et al., 2006; Mendes et al., 1995, 2003, 2006; Mendiola et al., 2007; Sajilata et al., 2008; Thana et al., 2008).

4.2.2. Operating variables Operating variables which influence the performance of SCCO2 extraction of microalgal lipids include pressure, temperature, modifier addition, and fluid flow rate (Pourmortazavi and Hajimirsadeghi, 2007). Table 5 summarizes the levels of operating variables and the findings of recent studies investigating SCCO2 extraction of microalgal lipids.

Fig. 12. Schematic diagram of the proposed supercritical carbon dioxide extraction mechanism. Microalgal biomass and diatoms are packed tightly as a mixture inside the cylindrical extraction vessel. Supercritical carbon dioxide flows on the surface of the packed mixture. lipids, ○ SCCO2. Static SCCO2 film enclosing the packed mixture is formed due to the interaction between SCCO2 and microalgal biomass. The mechanism can be described in 3 steps. Step 1: desorption of lipids from the microalgal biomass into the static SCCO2 film. Step 2: solubilization of the released lipids by SCCO2. Step 3: formation of a SCCO2-lipids complex. Step 4: diffusion of the SCCO2lipid complex across the static SCCO2 film into the bulk SCCO2 flow.

The evolution of lipids during SCCO2 extraction on microalgal biomass can be described by the following first-order kinetics equation (Goto et al., 1993; Halim et al., 2011; Ozkal et al., 2005):   −kt me ¼ ms;o 1−e

ð3Þ

where me is the amount of lipid extracted by the SCCO2 at time t (g lipid/g dried microalgal biomass), ms,o is the amount of lipid originally present in the microalgal cells (g lipid/g dried microalgal biomass), k is the lipid mass transfer coefficient from the microalgal cells into the eluting SCCO2 (min − 1), and t is the extraction time (min). The parameter k itself is a function of several operating variables and can be described generally as: k ¼ f ðP; T; ½M; Q Þ

ð4Þ

where P is the extraction pressure (bar), T is the extraction temperature (°C), [M] is the concentration of polar modifier (mol% of CO2), Q is the SCCO2 flow rate (l/min). Fig. 13 (Andrich et al., 2005) shows typical extraction curves obtained when using SCCO2 to extract lipids from microalgal biomass. These curves conform to the model proposed in Eq. (3), where the rate of lipid recovery is observed to decrease with extraction time. During their studies investigating lipid extraction from Nannochloropsis sp., Andrich et al. (2005) found majority of the total lipids (>80%) to be extracted within 5000 s. Continuing the extraction run beyond 10,000 s did not seem to dramatically increase the total lipid yield. This kind of asymptotic behavior is ascribed to the diffusion-driven nature of lipid extraction where the rate of lipid evolution is controlled by the lipid concentration gradient between the microalgal biomass and the SCCO2. In our previous study evaluating the feasibility of SCCO2 process to extract lipids from Chlorococcum sp. (Halim et al., 2011), we found the first-kinetic model described by Eq. (3) to accurately describe experimental data (minimum r 2 value for all extraction curves = 0.92). Solvent power of SCCO2 during lipid extraction is a direct function of the extraction pressure and the extraction temperature (Taylor, 1996). Higher extraction pressure leads to a higher fluid density and, thus, to an increase in solvent power. However, increasing extraction pressure also increases operating cost, lowers selectivity, and often results in the co-extraction of unwanted cellular components. Temperature increase leads to two competing phenomena. The decrease in fluid density lowers SCCO2 solvent power while a simultaneous increase in lipid volatility enhances the lipid mass transfer into the bulk SCCO2 flow (Cheung, 1999; Soares et al.,

724

R. Halim et al. / Biotechnology Advances 30 (2012) 709–732

Table 5 Methods and results summary of recent studies investigating supercritical carbon dioxide extraction of microalgal lipids. Special attention is given to the effect of pressure change and of temperature change on the total lipid yield. Optimum condition is defined as the experimental condition that produces the highest final total lipid yield. Study

Microalgal species

SCCO2 flow rate; Polar modifier; Results and optimum condition Extraction Extraction temperature extraction quantity of pressure polar modifier or P (bar) or T (°C) duration (min)

Sajilata et al., 2008

Spirulina platensis

316, 350, 400, 450, 484

Andrich et al., 2005 Mendes et al., 2003

Final total lipid yield at the optimum condition (wt.% of dried microalgal biomass)

40

0.7 l/min; 26.4, 40, 60, 80, 94

Ethanol; 9.64, 11, 13, 15, 16.36 ml

Total lipid yield increased with P. Optimum condition was found at 400 bar, 60 min, and 13.7 ml ethanol.

8.6

Nannochloropsis 400, 550, sp. 700

40, 55

0.17 kg/min; 360

None

At constant T, lipid extraction rate increased with P. At constant P, lipid extraction rate slightly increased with T. Final total lipid yield was the same at any T and P.

25.0

Spirulina maxima

100, 250, 350

50, 60

Not specified; not specified

ethanol; At constant T, total lipid yield increased with P. 3.1 10 mol% of CO2 At constant P, total lipid yield decreased with T. At constant T and P, polar modifier addition significantly increased total lipid yield. Optimum condition was found at 350 bar, 60 °C with ethanol addition (10 mol%).

Cheung, Hypnea 1999 charoides

241, 310, 379

40, 50

1 l/min; 120

none

At constant T, total lipid yield increased with P. At low P (241 bar), total lipid yield decreased with T. At medium to high P (310 and 379 bar), total lipid yield increased with T. Optimum condition was found at 379 bar and 50 °C.

6.7

200, 350

40, 55

0.4 l/min; 500

none

At constant T, total lipid yield increased with P. At low P (200 bar), total lipid yield decreased with T. At high P (350 bar), total lipid yield increased with T. Optimum condition was found at 350 bar and 55 °C.

13.0

Mendes et al., 1995

Chlorella vulgaris

2007; Taylor, 1996). Table 5 compiles findings from previous studies on the effect of pressure change and of temperature change on SCCO2 lipid extraction from microalgal biomass. Because of its non-polar nature, SCCO2 is unable to interact with either polar lipids or neutral lipids that form complexes with polar lipids. The addition of a polar modifier, often referred to as co-solvent or entrainer, enhances the fluid affinity towards polar lipids as well as lipid complexes that contain both neutral lipids and polar lipids. It further facilitates complete lipid extraction by diminishing the fluid viscosity and allowing the SCCO2 to rapidly permeate through the cellular matrices (Pourmortazavi and Hajimirsadeghi, 2007; Taylor, 1996). Common polar modifiers include methanol, ethanol, toluene and methanol–water mixture. Mendes et al. (2006) demonstrated that the addition of methanol to SCCO2 significantly increased the extraction of γ-linolenic acid (from 0.05 wt.% of dried microalgal biomass to 0.44 wt.%) from Spirulina maxima. The flow rate of SCCO2 through the extraction vessel affects lipid extraction kinetics. Even though increasing SCCO2 flow rate enables

Fig. 13. Typical first-order extraction curves obtained for SCCO2 extraction of microalgal lipids. Microalgae: Nannochloropsis sp., all extractions were performed at a constant temperature (40 °C). Pressure was optimized. ● 70 MPa, ■ 55 MPa, ♦ 40 MPa. Modified from Andrich et al. (2005).

more effective contact between the extraction fluid and the lipids, it often results in uneven fluid penetration and dead volumes within the extraction vessel (Pourmortazavi and Hajimirsadeghi, 2007). The density in which microalgal biomass is packed to form a fixed bed within the extraction vessel plays an important role in influencing extraction efficiency (Pourmortazavi and Hajimirsadeghi, 2007). In the case of extraction from dried microalgal powder, packing density is directly related to microalgal powder particulate size and the volumetric ratio of packing materials (normally diatomaceous earth or diatoms) to microalgal powder. Even though higher packing density increases the amount of lipids in the vessel, it reduces the vessel's porosity and can adversely affect the extraction kinetics via fluid channeling effects (Pourmortazavi and Hajimirsadeghi, 2007). 4.3. Comparison between organic solvent extraction and SCCO2 extraction Table 6 provides a comparison between organic solvent extraction and SCCO2 extraction when they are used to extract lipids from microalgae. Despite having low reactivity with lipids and being effective when directly applied to a wet feedstock (concentrate or disrupted concentrate), organic solvent extraction is slow and uses large amounts of expensive/toxic solvents. It has a limited selectivity towards biodiesel-desirable lipid fractions (acylglycerols containing mainly cis-unsaturated fatty acids with less than 4 double bonds) and requires energy-intensive liquid–liquid separation method (such as distillation) to remove the organic solvent from the lipids. On the other hand, SCCO2 extraction is rapid and non-toxic. It has high selectivity towards biodiesel-desirable lipid fractions due to SCCO2 tuneable density and produces solvent-free crude lipids. It is also non-reactive with the lipids. It remains effective when applied to a wet feedstock (concentrate or disrupted concentrate) though high residual water content in the microalgal biomass tends to lead to flow impedance and restrictor plugging. High installation costs of the extraction pressure vessel as well as unfavorable energy requirements for the fluid compression and heating remain the primary obstacles for scaling-up SCCO2 extraction (Crespo and Yusty, 2005;

Table 6 Comparison between organic solvent extraction and SCCO2 extraction for microalgal lipid extraction. * * *: good. *: poor. Organic solvent extraction

SCCO2 extraction

** Selectivity is not easily tuned. When non-polar organic solvent is used, only limited amount of neutral lipids can be extracted. When non-polar/polar organic solvent mixture is used, both neutral lipids and polar lipids are extracted.

*** SCCO2 tunable selectivity when combined with flexible polar modifier arrangement should enable specific extraction of acylglycerols and minimize co-extraction of contaminants (polar lipids and non-acylglycerol neutral lipids).

Total lipid yield

**

*** Due to its intermediate liquid–gas properties, SCCO2 can penetrate through cellular matrices rapidly and produces a higher total lipid yield.

Extraction time

** Lipid extraction rate is slow and lipid extraction requires a long time for completion.

*** Due to its intermediate liquid-gaseous properties, SCCO2 can penetrate through cellular matrices rapidly. As such, lipid extraction rate is fast and lipid extraction can be completed within a short period.

Energy requirement

** It consumes little energy as lipid extraction is conducted near ambient conditions. However, organic solvent needs to be removed from the lipids via energy-intensive liquid–liquid separation method (such as distillation).

* It is highly energy-intensive as fluid compression and heating are needed to convert CO2 to supercritical state. However, crude lipids are free from extraction solvent and do not need to undergo an extraction solvent removal.

Installation and operating (non-energy related) cost

**

*

Expensive organic solvent is needed. Not all of the organic solvents can be recycled.

The pressure vessel needed for SCCO2 extraction can be extremely expensive to install.

***

*

Lipid extraction can be applied to microalgal concentrate without additional pre-treatment step and with minimal loss of efficiency.

High residual water content within the microalgal biomass results in flow impedance and restrictor plugging.

Hazard and toxicity

* Toxic organic solvents are used.

***

Reactivity with lipids

** Organic solvent is non-reactive with lipids. However, distillation carried out to remove the organic solvent from the lipids exposes the lipids to high temperature and possible artifact formation.

*** SCCO2 is non-reactive with lipids.

Applicability to microalgal concentrate or wet feedstock

R. Halim et al. / Biotechnology Advances 30 (2012) 709–732

Criteria Neutral lipids selectivity

725

726

R. Halim et al. / Biotechnology Advances 30 (2012) 709–732

cloud point and a high pour point), while biodiesel made from PUFA tends to be volatile and has a low oxidation stability. Schematic diagrams for an industrial-scale organic solvent extraction system and an industrial-scale SCCO2 extraction system are proposed in Fig. 15. For SCCO2 extraction, compressor is used to pressurize the fluid to a supercritical state. The scale-up potential of each lipid extraction technology is examined in Table 1. 5. Effect of cellular pre-treatment on lipid extraction

Fig. 14. Comparison between dynamic SCCO2 extraction and dynamic hexane extraction (using a Soxhlet apparatus). (a) total lipid yield. (b) FAME composition of the crude lipids. Microalgae: Chlorococcum sp. In (a), ■ SCCO2 extraction, ▲ Hexane extraction (using a Soxhlet apparatus). In (b), the letter t after the fatty acid name (C16:1t) denotes trans-isomerism. When no letter t appears, fatty acid is of cis-isomerism. For SCCO2 extraction, mass of microalgal dried powder = 20 g, dried powder : diatomaceous earth = 2/1 w/w, T = 60 °C, P = 30 MPa. For hexane extraction (using a Soxhlet apparatus), mass of microalgal dried powder = 4 g, total number of cycles or equilibrium establishments = 55. Modified from Halim et al. (2011).

Halim et al., 2011). For a more extensive evaluation than Table 6, a thorough understanding of mass transfer mechanisms and of kinetic parameters involved in lipid extraction is required. In our previous study extracting lipids from dried Chlorococcum sp. (Halim et al., 2011), we compared the performance of dynamic SCCO2 extraction with that of dynamic hexane extraction (using a Soxhlet apparatus). As shown in Fig. 14, SCCO2 extraction was found to be more efficient than hexane extraction. Eighty minutes of SCCO2 extraction (total lipid yield = 0.058 g lipid/g dried microalgal biomass) achieved a higher total lipid yield than 5.5 h of Soxhlet extraction (total lipid yield = 0.032 g lipid/g dried microalgal biomass). This outcome was expected since supercritical fluid has more favorable physicochemical properties and facilitates more rapid cellular permeation (Cheung, 1999; Mendes et al., 2003). Andrich et al. (2005) reported similar results. Despite demonstrating that both extractions eventually obtained equivalent final total lipid yields from Nannochloropsis sp., they measured a lower lipid mass transfer coefficient for extraction with hexane than with SCCO2. In terms of fatty acid composition (Fig. 14), Chlorococcum crude lipids extracted by SCCO2 comprised a substantially higher quantity of C18:1 than the corresponding crude lipids extracted by hexane. Such selectivity was beneficial since C18:1, as a cis-unsaturated fatty acid with less than 4 double bonds, is highly desirable for biodiesel production. As previously mentioned (Section 2), biodiesel derived from saturated fatty acids often has disadvantageous cold flow properties (a high

The effects of cellular pre-treatment on microalgal lipid extraction have not been investigated extensively. As previously described, the pre-treatment process can take alternative pathways depending on the desired biomass alterations (Fig. 4). The process can be performed in a single step or multiple steps. It is noted that most of the pre-treatment steps (such as thermal drying for complete water removal or high-pressure homogenization for cell disruption) are energy intensive and should only be carried out if they substantially enhance the efficiency of microalgal lipid extraction. Based on the combination of technologies available in Table 1, the pre-treatment process can alter the following conditions of the microalgal biomass: degree of cell disruption, residual water content, and, in the case of dried microalgal powder, particulate size. The efficiency of microalgal lipid extraction is known to increase with the degree of cell disruption. When intact cells are disintegrated during cell disruption, intracellular lipids are liberated from the cellular structures and released into the surrounding medium (Chisti and Moo-Young, 1986; Gouveia et al., 2007; Lee et al., 2010; Mendes-Pinto et al., 2001). During subsequent lipid extraction, the eluting extraction solvent can directly interact with these free lipids without penetrating into the cellular structures. The lipid extraction process is thus no longer restricted by the transportation of extraction solvent and lipids across the cell membrane. It is completed more rapidly and results in a higher lipid recovery. It is noted that most cell disruption methods (such as bead milling, ultrasonication, and high-pressure homogenization) require certain degree of water in the microalgal biomass for their successful operation. For this reason, cell disruption step in a pre-treatment pathway is always performed before the drying step (as illustrated in Fig. 4). Additionally, most cell disruption methods will not be able to process microalgal concentrate with exceedingly low water contents (i.e. microalgal paste or pellet). Laboratory-scale cell disruption methods (Fig. 16) are classified based on the manner in which they achieve microalgal cellular disintegration: mechanical or non-mechanical (Chisti and Moo-Young, 1986; Harrison et al., 2003a). Mechanical methods include bead mill, press, high-pressure homogenization, ultrasonication, autoclave, lyophilization, and microwave, while non-mechanical methods often involve lysing the microalgal cells with acids, alkalis, enzymes, or osmotic shocks (Chisti and Moo-Young, 1986). Bead mill, high-pressure homogenization, and ultrasonication are three of the most widely used methods for laboratory-scale microalgal cell disruption (Chisti and Moo-Young, 1986; Harrison et al., 2003a). Bead mill achieves cellular disruption by physically grinding the microalgal cells against the solid surfaces of glass beads in a violent agitation. Among the myriad of cell disruption methods, bead mill appears most suitable for large-scale application due to its low operating cost (Chisti and Moo-Young, 1986). High-pressure homogenization pumps microalgal concentrate through narrow orifice of a valve under high pressure. It then releases the concentrate into a low-pressure chamber. Cellular disintegration is thus achieved through high-pressure impingement of accelerated cellular jet on the stationary valve surface as well as through pressure-drop induced shear stress that the microalgal concentrate experiences as it passes from the valve to the chamber (Chisti and Moo-Young, 1986). Ultrasonication disrupts microalgal cells via transmission of sonic

R. Halim et al. / Biotechnology Advances 30 (2012) 709–732

727

Fig. 15. Proposed schematic diagrams of (a) an industrial-scale organic solvent extraction system and (b) an industrial-scale SCCO2 extraction system. The systems are intended for lipid extraction from microalgal biomass.

waves. These waves create a series of microbubble cavitations on the cell surface and eventually disintegrate the cell membrane/wall (Chisti and Moo-Young, 1986). Detailed working mechanisms of bead mill, high-pressure homogenization and ultrasonication can be found elsewhere (Chisti and Moo-Young, 1986; Harrison et al., 2003a).

Lee et al. (1998) assessed the effect of prior mechanical cell disruption on lipid extraction from the species B. braunii. They used chloroform/methanol mixture (2/1 v/v) as an extraction solvent and found completely disrupted microalgal cells to yield almost twice the amount of crude lipids of intact microalgal cells. Also, among the different mechanical cell disruption methods investigated

Fig. 16. The classification of laboratory-scale cell disruption methods.

728

R. Halim et al. / Biotechnology Advances 30 (2012) 709–732

Botryococcus sp. 30

20

10

Final total lipid yield (wt% of dried microalgal biomass)

0

Chlorella vulgaris

30

20

10

0

Scenedesmus sp.

30

20

10

0 s ve wa o r ic

g

tin

lav

toc Au

d

n-

No

ing

on

pti

u isr

ea

a

b d-

Be

M

k oc

on

ati

ic on

S

c oti

sh

m

Os

Fig. 17. Effect of prior cell disruption on total lipid yield of organic solvent extraction. Three microalgal species (Botryococcus sp., Chlorella vulgaris, and Scenedesmus sp.) were investigated. Organic solvents: chloroform–methanol (1/1 v/v) mixture. Modified from Lee et al. (2010).

Final total lipid yield (g lipid / g dried microalgal biomass)

(sonication, homogenization, high-pressure French press, bead beating or bead mill, and lyophilization), mechanical shearing with bead mill obtained the highest final total lipid yield. In a different study by Lee et al. (2010), the effect of prior cell disruption on lipid 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 A

B

C1

C2

D1

D2

Organic solvent Fig. 18. Effect of residual water content within the microalgal biomass on total lipid yield of organic solvent extraction. Microalgae: Chlorococcum sp. A: Hexane extraction of microalgal dried powder. B: Hexane extraction of microalgal concentrate. C: Hexane/ isopropanol (3/2 v/v) extraction of microalgal dried powder (C1: organic phase, C2: aqueous phase). D: Hexane/isopropanol (3/2 v/v) extraction of microalgal concentrate (D1: organic phase, D2: aqueous phase). For A and C: mass of dried powder = 4 g. For B and D: mass of concentrate = 13.3 g, residual water content = 70 wt.% of concentrate. For A, B, C, D: mass of dried microalgal biomass = 4 g, volume of organic solvent mixture = 300 ml, duration = 7.5 h. Modified from Halim et al. (2011).

extraction from three microalgal species (Botryococcus sp., Chlorella vulgaris, and Scenedesmus sp.) was evaluated (Fig. 17). Chloroform– methanol (1/1 v/v) mixture was used as an extraction solvent in all cases. Among the cell disruption methods assessed (autoclave, bead beating, microwave, sonication, and osmotic shocks), microwave obtained the highest final total lipid yield and appeared to be the most efficient for all three microalgal strains. For Botryococcus sp., bead beating and microwave obtained the highest final total lipid yields with respectively 0.281 and 0.286 g lipid/g dried microalgal biomass, while sonication seemed to be the least efficient at 0.088 g lipid/g dried microalgal biomass. For C. vulgaris, autoclave and microwave appeared to be the most efficient, whereas bead beating produced a low final total lipid yield at 0.079 g lipid/g dried microalgal biomass. With Scenedesmus sp., microwave was again found to show the highest extraction efficiency, while yields from the other methods were similar. For all three microalgal species, prior disruption of the cells by any of the assessed method was found to improve final total lipid yield during the lipid extraction step (refer to Fig. 17 and compare all of the methods with the control non-disruption). The mechanism in which residual water in the microalgal biomass affects lipid extraction is not well understood and warrants future investigation. One hypothesis speculates that the presence of residual water in the microalgal biomass will adversely affect lipid extraction efficiency. Water forms a barrier that prohibits effective lipid mass transfer from the cells to the extraction solvent. As such, drying of microalgal concentrate is non-optional and has to be performed prior to the lipid extraction. On the other hand, another hypothesis postulates that the presence of residual water in the microalgal biomass will improve lipid extraction efficiency. Water swells the cells and facilitates better solvent access to the lipids. Drying of microalgal concentrate prior to lipid extraction is deemed unnecessary and may hinder lipid mass transfer. Various microorganisms (bacteria, yeasts, and viruses) have been successfully extracted in their wet state (~ 90 wt.% water) using non-polar/polar organic solvent mixtures (Kates, 1986b; Medina et al., 1998). With regards to SCCO2 extraction, abundance of residual water in the microalgal biomass results in a highly compacted particle bed within the extraction vessel. This often leads to flow impedance and restrictor plugging (Pourmortazavi and Hajimirsadeghi, 2007; Schwartzberg, 1997). During their investigation of lipid extraction from Chlorococcum sp., Halim et al. (2011) assessed the effect of residual water content within the microalgal biomass on total lipid yield. As shown in Fig. 18 (modified from Halim et al., 2011), the presence of residual water in the microalgal biomass did not appear to substantially affect total lipid yield. Hexane extraction of concentrate (final total lipid yield = 0.010 g lipid/g dried microalgal biomass) obtained a slightly lower lipid recovery than its dry powder counterpart (final total lipid yield = 0.015 g lipid/g dried microalgal biomass), while hexane/isopropanol extraction of concentrate (final total lipid yield = 0.123 g lipid/g dried microalgal biomass) surprisingly obtained a higher total lipid yield than hexane/isopropanol extraction of dried powder (final total lipid yield = 0.068 g lipid/g dried microalgal biomass). These findings were encouraging particularly since the organic solvent extraction of wet biomass did not require any additional pre-treatment step. Among the drying technologies that can be applied to microalgal concentrate, freeze drying is preferred for its mild operating conditions. Thermal drying, though commonly used in laboratory practice, is not recommended as it degrades thermolabile lipids, results in evaporative loss of volatile lipids, and yields powder with nonuniform particulate size (Pourmortazavi and Hajimirsadeghi, 2007). Once dried, microalgal biomass forms powder (or agglomeration) that can be milled into different particulate size. Reducing the particulate size of microalgal powder prior to lipid extraction generally enhances lipid recovery as it increases the interfacial surface area

R. Halim et al. / Biotechnology Advances 30 (2012) 709–732

729

Fig. 19. Alternative process flow diagram showing the downstream processing steps needed with a simultaneous extraction and transesterification step to produce biodiesel from microalgae.

available for biomass-solvent contacts and shortens the diffusion pathway of the extraction solvent. However, exceedingly small particulate size of the microalgal powder may lead to a higher tendency of lipid re-adsorption, fluid channeling effects in the extraction vessel (for SCCO2 extraction), and inhomogeneous lipid extraction (Pourmortazavi and Hajimirsadeghi, 2007). Sabio et al. (Pourmortazavi and Hajimirsadeghi, 2007) conducted a study on SCCO2 extraction of oil from tomato skins and verified that smaller tomato skins resulted in higher oil recoveries. 6. Simultaneous extraction and transesterification of microalgal lipids Recent studies investigating biodiesel production from microalgae have focused their efforts on the development of an alternative

downstream processing step termed simultaneous extraction and transesterification (Wahlen et al., 2011). This step, also known as direct transesterification or in-situ transesterification, combines lipid extraction and transesterification in a single step, thereby simplifying the downstream pathway required for biodiesel production from microalgal biomass (Fig. 19). The method involves the simultaneous addition of acid catalyst and pure methanol to microalgal biomass (generally in the form of dried powder). The methanol extracts the lipids from the microalgal biomass and, catalyzed by the acid, concurrently transesterifies the extracted lipids to produce fatty acid methyl esters. Similar downstream processing steps as those required in the traditional pathway are then followed, where the reaction mixture (consisting of methanol, biodiesel, glycerol, reformed acid catalyst, un-transesterified lipids, and cell debris) undergoes cell-debris removal and post-transesterification purification

730

R. Halim et al. / Biotechnology Advances 30 (2012) 709–732

Fig. 20. Schematic diagram of a laboratory-scale OriginOil Single-Step Extraction™ system. Extracted from OriginOil, 2010.

(Fig. 19). Filtration is used for cell-debris removal. A laboratory-scale post-transesterification purification consists of multiple steps. The reaction mixture is first distilled to remove methanol. It is then left to settle under gravity to induce biphasic partitioning (top biodiesel/ un-transesterified lipids phase and bottom glycerol phase). The biodiesel/un-transesterified lipids phase is decanted off and washed repeatedly with water to eliminate any acid catalyst. It is noted that studies investigating direct transesterification of microalgal biomass have, so far, only used acid catalysts (acetyl chloride and H2SO4). Direct transesterification was originally derived by Lepage and Roy (1984) as a rapid method to analyze the fatty acid contents of adipose tissues and milk. The method is highly desirable as it reduces the amount of downstream manipulation required for biodiesel production from any given biomass. Compared to the conventional lipid-extraction-followed-by-transesterification approach, direct transesterification has been shown to improve biodiesel yields of various animal and plant tissues. However, the efficacy of direct transesterification on microalgal biomass has not been sufficiently investigated. Parameters that affect the performance of direct transesterification include the ratio of methanol to dried microalgal biomass (ml methanol/g dried microalgal biomass), the reaction temperature (°C), Wahlen et al. (2011) examined the effect of residual water content within the microalgal biomass on the kinetics of direct transesterification. Chaetoceros gracilis biomass with residual water contents ranging from 9 to 50% of microalgal paste weight was subjected to direct transesterification. Increasing residual water content within the microalgal biomass was found to progressively decrease FAME yield. For biomass with residual water content equal to 50% of the microalgal paste weight, FAME yield was half that obtained from the direct transesterification of dried powder (biomass with residual water content = 0%). 7. Microalgal biorefinery The cost of producing microalgal biodiesel can theoretically be offset by revenues generated from other co-products of the microalgal

biomass. Microalgae contain significant quantities of proteins and carbohydrates as well as smaller amounts of high-value functional ingredients (astaxanthin, canthaxanthin, carotenes, chlorophylls, Ω3 free fatty acids, and γ-linolenic acid). Each of these cell components can be appropriately utilized to co-generate a useable product in a biorefinery. Recent studies have concluded that industrial-scale production of microalgal biodiesel can only be made economically sustainable if a biorefinery based production strategy is pursued (Wijffels et al., 2010). In a biorefinery, the crude lipids are to be fractionated into high-value functional ingredients and lipids for biodiesel (acylglycerols). Functional ingredients have been linked with the promotion of anti-oxidant, anti-inflammatory, as well as anticarcinogenic activities in the human bodies and are typically used as food supplements. If the microalgal species contains a high level of proteins, the residual biomass from biodiesel production processes can be used as livestock feeds. If the species has high carbohydrate contents, the residual biomass can be fermented to produce bioethanol. As such, microalgal biorefinery will simultaneously produce biodiesel, high-value products, livestock feeds, and bioethanol. Unfortunately, the combination of technologies needed to implement a microalgal biorefinery is still in the early stages of development. Milder cell disruption/lipid extraction process needs to be explored to ensure that the functionalities of different cell components are retained. 8. OriginOil Single-Step Extraction of microalgal lipids OriginOil, Inc has established a novel method for microalgal lipid extraction (OriginOil, 2010). Instead of following the traditional sequence-based downstream processing pathway outlined in Fig. 4, the method devised by OriginOil performs three simultaneous functions (dewatering, cell disruption, and lipid extraction) in a single downstream step (Fig. 20). This approach is referred to as OriginOil Single-Step Extraction™. Within a single step (OriginOil, 2010), microalgal concentrate is exposed to Quantum Fracturing™, a patent-pending technology based on the science of fluid fracturing, combined with pulsed

R. Halim et al. / Biotechnology Advances 30 (2012) 709–732

electromagnetic fields and pH modification (Fig. 20). Microalgal cells are instantly disrupted and intracellular lipids are released from the microalgal biomass. The microalgal lipids rise to the top of the gravity clarifier for skimming, lipid fractionation, and transesterification to biodiesel, while the cell debris settles to the bottom of gravity clarifier for further processing as fuel and other valuable by-products. The water (or growth medium) in the middle of the gravity clarifier can be decanted and recycled for microalgal cultivation. There are three components to OriginOil Single-Step Extraction™ (OriginOil, 2010): 1) CO2 injection lowers pH of the growth medium to optimize electromagnetic fields delivery and to assist in cell disruption, 2) Quantum Fracturing™ creates fluid effect to mechanically stress microalgal cells, 3) Pulsed electromagnetic fields deliver the force that disrupt the microalgal cells. OriginOil Single-Step Extraction™ has several benefits (OriginOil, 2010). By combining three functions (dewatering, cell disruption, and lipid extraction) in a single downstream step, it substantially reduces the energy expenditure required to produce biodiesel from microalgal biomass. The lipid extraction method does not use any toxic extraction solvent and, thus, does not require a solvent recovery step. Finally, the lipid extraction method is highly effective when directly applied to wet feedstock (concentrate). 9. Conclusions The downstream technologies needed for industrial-scale production of microalgal biodiesel are still in the early stages of development. Lipid extraction from microalgal biomass has not received sufficient attention and represents one of the many bottlenecks hindering economic industrial-scale production of microalgal biodiesel. Future research on microalgal biodiesel should focus on developing an effective and energetically efficient lipid extraction process. A fundamental understanding of lipid mass transfer mechanisms from the microalgal biomass to the extraction solvent is needed to scale up the lipid extraction process. An ideal lipid extraction technology for microalgal biodiesel production needs to be not only lipid specific in order to minimize the co-extraction of non-lipid contaminants (such as protein and carbohydrates) but also selective towards acylglycerols in order to reduce downstream fractionation/purification (Fajardo et al., 2007; Medina et al., 1998). Additionally, the selected technology should be efficient (both in terms of time and energy), non-reactive with the lipids, relatively cheap (both in terms of capital cost and operating cost), and safe (Kates, 1986b). Microalgal culture is harvested as a dilute aqueous suspension (from 0.1 to 2 g dried microalgal biomass/L culture) and is substantially concentrated. However, dewatering the microalgal biomass beyond a paste consistency (200 g dried microalgal biomass/L culture) is energetically prohibitive. For this reason, it will be economically beneficial if the selected lipid extraction technology can be directly applied to relatively wet feedstock, i.e. concentrate or disrupted concentrate with concentrations between 10 and 200 g dried microalgal biomass/L culture (Halim et al., 2011). Studies in the past decade have demonstrated the use of organic solvents and the use of supercritical carbon dioxide to extract lipids from microalgal biomass. However, each technology has its merits and limitations. Despite having low reactivity with lipids and being directly applicable to relatively wet feedstock, organic solvent extraction is slow and uses a large amount of expensive/toxic solvents. On the other hand, supercritical carbon dioxide extraction is a promising green technology that can potentially be used for large-scale microalgal lipid extraction. It is rapid, non-toxic, has high selectivity towards acylglycerols, and produces solvent-free lipids. Its main disadvantages are associated with the high capital cost and the high energy requirement for supercritical fluid compression. Several studies have shown that the disruption of microalgal cells prior to the lipid extraction step enhances the extraction efficiency.

731

Acknowledgments This work was supported by an Australian Research Council (ARC) Linkage grant between the Department of Chemical Engineering in Monash University (Victoria, Australia) and Bio-Fuel Pty Ltd (Victoria, Australia).

References Andrich G, Nesti U, Venturi F, Zinnai A, Fiorentini R. Supercritical fluid extraction of bioactive lipids from the microalgae Nannochloropsis sp. European Journal of Lipid Science and Technology 2005;107:381–6. Andrich G, Zinnai A, Nesti U, Venturi F, Fiorentini R. Supercritical fluid extraction of oil from microalga Spirulina (Arthrospira) platensis. Acta Alimentaria 2006;35(2): 195–203. Applied separations. http://www.appliedseparations.com/Supercritical/Lab_Inst/ Analytical/SFE_2.asp2007. [Accessed in Dec 2009]. Balasubramanian S, Allen JD, Kanitkar A, Boldor D. Oil extraction from Scenedesmus obliquus using a continuous microwave system — design, optimization, and quality characterization. Bioresource Technology 2010. doi:10.1016/j.biortech.2010.09.119. Brown MR, Jeffrey SW, Volkman JK, Dunstan GA. Nutritional properties of microalgae for mariculture. Aquaculture 1997;151:315–31. Canela APRF, Rosa PTV, Marques MOM, Meireles MAA. Supercritical fluid extraction of fatty acids and carotenoids from the microalgae Spirulina maxima. Industrial and Engineering Chemistry Research 2002;41:3012–8. Chen M, Chen X, Liu T, Zhang W. Subcritical ethanol extraction of lipid from wet microalgae paste of Nannochloropsis sp. Journal of Biobased Materials and Bioenergy 2011;5(3):385–9. Cheung P. Temperature and pressure effects on supercritical carbon dioxide extraction of n-3 fatty acids from red seaweed. Food Chemistry 1999;65:399–403. Chisti Y. Research review paper: Biodiesel from microalgae. Biotechnology Advances 2007;25:294–306. Chisti Y, Moo-Young M. Review: disruption of microbial cells for intracellular products. Enzyme and Microbial Technology 1986;8:194–204. Christie WW. Methylation of fatty acids — a beginner's guide. http://www.lipidlibrary. co.uk/topics/methests/index.htm 2007. [Accessed in December 2009]. Crespo MOP, Yusty MAL. Comparison of supercritical fluid extraction and Soxhlet extraction for the determination of PCBs in seaweed samples. Chemosphere 2005;59:1407–13. Danquah MK, Gladman B, Moheimani N, Forde GM. Microalgal growth characteristics and subsequent influence on dewatering efficiency. Chemical Engineering Journal 2009;151:73–8. De Castro MD, Ayuso LE. Soxhlet extraction. Environmental Applications 2000:2701–5. Demirbas A. Comparison of transesterification methods for production of biodiesel from vegetable oils and fats. Energy Conversion and Management 2008;49: 125–30. Demirbas A, Karslioglu S. Biodiesel production facilities from vegetable oils and animal fats. Energy Sources Part A 2007;29:133–41. Dunstan GA, Volkman JK, Barrett SM, Garland CD. Changes in the lipid composition and maximisation of the polyunsaturated fatty acid content of three microalgae grown in mass culture. Journal of Applied Phycology 1993;5:71–83. Fajardo AR, Cerdan LE, Medina AR, Fernandez FGA, Moreno PAG, Grima EM. Lipid extraction from the microalga Phaedactylum tricornutum. European Journal of Lipid Science Technology 2007;109:120–6. Folch J, Ascoli I, Lees M, Meath JA, Lebaron FN. Preparation of lipide extracts from brain tissue. The Journal of Biological Chemistry 1951:833–41. Goto M, Sato M, Hirose T. Extraction of peppermint oil by supercritical carbon dioxide. Journal of Chemical Engineering of Japan 1993;26(4):401–7. Gouveia L, Nobre BP, Marcelo FM, Mrejen S, Cardoso MT, Palavra AF, et al. Functional food oil colored by pigments extracted from microalgae with supercritical CO2. Food Chemistry 2007;101:717–23. Guckert JB, Cooksey KE, Jackson LL. Lipid solvent systems are not equivalent for analysis of lipid classes in the microeukaryotic green alga, Chlorella. Journal of Microbiological Methods 1988;8:139–49. Gupta MN, Batra R, Tyagi R, Sharma A. Polarity index: the guiding solvent parameter for enzyme stability in aqueous-organic cosolvent mixtures. Biotechnology Progress 1997;13:284–8. Guzman HM, Valido AJ, Duarte LC, Presmanes KF. Estimate by means of flow cytometry of variation in composition of fatty acids from Tetraselmis suecica in response to culture conditions. Aquaculture International 2010;18:189–99. Halim R, Gladman B, Danquah MK, Webley PA. Oil extraction from microalgae for biodiesel production. Bioresource Technology 2011;102(1):178–85. Hansen CM. Appendix A: Table A.1. Hansen solubility parameters: a user's handbook. 2nd ed. Boca Raton, Florida: CRC Press; 2007. Hansen CM. Review article: polymer science applied to biological problems: prediction of cytotoxic drug interactions with DNA. European Polymer Journal 2008;44: 2741–8. Harrison RG, Todd P, Rudge SR, Petrides DP. Chapter 3: cell lysis and flocculation. Bioseparations sciences and engineering. London: Oxford University Press; 2003a. p. 83–93. Harrison RG, Todd P, Rudge SR, Petrides DP. Chapter 6: extraction. Bioseparations sciences and engineering. London: Oxford University Press; 2003b. p. 169–89.

732

R. Halim et al. / Biotechnology Advances 30 (2012) 709–732

Herrero M, Senorans FJ, Cifuentes A, Ibanez E. Optimization of accelerated solvent extraction of antioxidants from Spirulina platensis microalga. Food Chemistry 2005;93:417–23. Herrero M, Cifuentes A, Ibanez E. Sub- and supercritical fluid extraction of functional ingredients from different natural sources: plants, food-by-products, algae and microalgae, a review. Food Chemistry 2006;98:136–48. Huang GH, Chen F, Wei D, Zhang XW, Chen G. Biodiesel production by microalgal biotechnology. Applied Energy 2010;87:38–46. Kates M. Definition and classification of lipids. Techniques of lipidology isolation, analysis, and identification of lipids. Amsterdam: Elsevier Science Publisher; 1986a. Kates M. Lipid extraction procedures. Techniques of lipidology isolation, analysis, and identification of lipids. Amsterdam: Elsevier Science Publisher; 1986b. Lang X, Dalai AK, Bakhshi NN, Reaney MJ, Hertz PB. Preparation and characterization of bio-diesels from various bio-oils. Bioresource Technology 2001;80:53–62. Lee SJ, Yoon BD, Oh HM. Rapid method for the determination of lipid from the green algae Botryococcus braunii. Biotechnology Techniques 1998;7:553–6. Lee JY, Yoo C, Jun SY, Ahn CY, Oh HM. Comparison of several methods for effective lipid extraction from microalgae. Bioresource Technology 2010;101:S75–7. Lepage G, Roy CC. Improved recovery of fatty acid through direct transesterification without prior extraction or purification. Journal of Lipid Research 1984;25: 1391–6. Luque de Castro MD, Garcia-Ayuso LE. Soxhlet extraction of solid materials: an outdated technique with a promising innovative future. Analytica chimica acta 1998;369:1-10. Lv X, Zou L, Sun B, Wang J, Sun MY. Variations in lipid yields and compositions of marine microalgae during cell growth and respiration, and within intracellular structures. Journal of Experimental Marine Biology and Ecology 2010;391: 73–83. Macias-Sanchez MD, Mantell C, Rodriguez M, de la Ossa EM, Lubian LM, Montero O. Supercritical fluid extraction of carotenoids and chlorophyll a from Synechococcus sp. Journal of Supercritical Fluids 2007;39:323–9. Mathews CK, van Holde KE. Lipids, membranes, and cellular transport. Biochemistry. Menlo Park, California: The Benjamin/Cummings Publishing Company, Inc.; 1996. p. 316–58. Medina AR, Grima EM, Gimenez AG, Ibanez MJ. Downstream processing of algal polyunsaturated fatty acids. Biotechnology Advances 1998;16(3):517–80. Mendes RL, Coelho JP, Fernandes HL, Marrucho IJ, Cabral JM, Novais JM, et al. Applications of supercritical CO2 extraction to microalgae and plants. Journal of Chemical Technology and Biotechnology 1995;62:53–9. Mendes RL, Nobre BP, Cardoso MT, Pereira AP, Palavra AF. Supercritical carbon dioxide extraction of compounds with pharmaceutical importance from microalgae. Inorganica Chimica Acta 2003;357:328–34. Mendes RL, Reis AD, Palavra AF. Supercritical CO2 extraction of gamma-linolenic acid and other lipids from Arthrospira (Spirulina) maxima: comparison with organic solvent extraction. Food Chemistry 2006;99:57–63. Mendes-Pinto MM, Raposo MFJ, Bowen J, Young AJ, Morais R. Evaluation of different cell disruption processes on encysted cells of Haematococcus pluvialis : effects of astaxanthin recovery and implications for bio-availability. Journal of Applied Phycology 2001;13:19–24. Mendiola JA, Jaime L, Santoyo S, Reglero G, Cifuentes A, Ibanez E, et al. Screening of functional compounds in supercritical fluid extracts from Spirulina platensis. Food Chemistry 2007;102:1357–67. Molina Grima E, Robles Medina A, Gimenez Gimenez A, Sanchez Perez JA, Garcia Camacho F, Garcia Sanchez JL. Comparison between extraction of lipids and fatty acids from microalgal biomass. JAOCS 1994;71(9):955–9. Molina Grima E, Belarbi E-H, Acien Fernandez FG, Robles Medina A, Chisti Y. Recovery of microalgal biomass and metabolites: process options and economics. Biotechnology Advances 2003;20:491–515. Nagle N, Lemke P. Production of methyl ester fuel from microalgae. Applied Biochemistry and Biotechnology 1990;24:355–61.

Nelson DL, Cox MM. Lipids. Lehninger Principles of Biochemistry3 rd ed. ; 2000. OriginOil. Algae harvesting, dewatering and extraction a breakthrough technology to transform algae into oil. Amsterdam, The Netherlands: Worldbiofuels Markets; 2010. 15-17 March. Ota M, Kato Y, Watanabe H, Watanabe M, Sato Y, Smith R, et al. Fatty acid production from a highly CO2 tolerant alga, Chlorocuccum littorale, in the presence of inorganic carbon and nitrate. Bioresource Technology 2009;100:5237–42. Ozkal SG, Salgin U, Yener ME. Supercritical carbon dioxide extraction of hazelnut oil. Journal of Food Engineering 2005;69:217–23. Pourmortazavi SM, Hajimirsadeghi SS. Supercritical fluid extraction in plant essential and volatile oil analysis — review. Journal of Chromatography. A 2007;1163:2-24. Ramadan MF, Asker MHS, Ibrahim ZK. Functional bioactive compounds and biological activities of Spirulina platensis lipids. Czechoslovak Journal of Food Sciences 2008;26(3):211–22. Rao AR, Dayananda C, Sarada R, Shamala TR, Ravishankar GA. Effect of salinity on growth of green alga Botryococcus braunii and its constituents. Bioresource Technology 2007;98:560–4. Reverchon E, De Marco I. Review: supercritical fluid extraction and fractionation of natural matter. Journal of Supercritical Fluids 2006;38:146–66. Rosenberg JN, Wilkinson L, Betenbaugh MJ. A green light for engineered algae: redirecting metabolism to fuel a biotechnology revolution. Current Opinion in Biotechnology 2008;19:430–6. Sajilata MG, Singhal RS, Kamat MY. Supercritical CO2 extraction of γ-linolenic acid (GLA) from Spirulina platensis ARM 740 using response surface methodology. Journal of Food Engineering 2008;84:321–6. Schwartzberg HG. Mass transfer in a countercurrent, supercritical extraction system for solutes in moist solids. Chemical Engineering Communications 1997;157:1-22. Sheehan J, Dunahay T, Benemann J, Roessler P. A look back at the US Department of Energy's Aquatic Species Program: biodiesel from algae. In: Laboratory NRE, editor. National Renewable Energy Laboratory: US Department of Energy, 1998. pp. 1-100. Shenk P, Thomas-Hall S, Stephens E, Marx U, Mussgnug J, Posten C. Second generation biofuels: high-efficiency microalgae for biodiesel production. BioEnergy Research 2008;20. Snyder LR. Classification of the solvent properties of common liquids. Journal of Chromatography 1974;92:223–30. Soares BMC, Gamarra FMC, Paviani LC, Goncalves LAG, Cabral FA. Solubility of triacylglycerols in supercritical carbon dioxide. Journal of Supercritical Fluids 2007;43:25–31. Taylor LT. Supercritical fluid extraction. New York: John Wileys & Sons, Inc.; 1996. Thana P, Machmudah S, Goto M, Sasaki M, Pavasant P, Shotipruk A. Response surface methodology to supercritical carbon dioxide extraction of astaxanthin from Haematococcus pluvialis. Bioresource Technology 2008;99:3110–5. Uduman N, Qi Y, Danquah MK, Hoadley AFA. Marine microalgae flocculation and focused beam reflectance measurement. Chemical Engineering Journal 2010;162: 935–40. Volkman JK, Jeffrey SW, Nichols PD, Rogers GI, Garland CD. Fatty acid and lipid composition of 10 species of microalgae used in mariculture. Journal of Experimental Marine Biology and Ecology 1989;128:219–40. Wahlen BD, Willis RM, Seefeldt LC. Biodiesel production by simultaneous extraction and conversion of total lipids from microalgae, cyanobacteria, and wild mixedcultures. Bioresource Technology 2011;102:2724–30. Wang L, Weller CL. Recent advances in extraction of nutraceuticals from plants. Trends in Food Science and Technology 2006;17:300–12. Widjaja A, Chien C-C, Ju Y-H. Study of increasing lipid production from fresh water microalgae Chlorella vulgaris. Journal of the Taiwan Institute of Chemical Engineers 2009;40. Wijffels RH, Barbosa MJ. An outlook on microalgal biofuels. Science 2010;329(5993):796–9. Wijffels RH, Barbosa MJ, Eppink MHM. Microalgae for the production of bulk chemicals and biofuels. Biofuels, Bioproducts and Biorefining 2010;4(3):287–95.