Extraction of Lipids from Chlorella Alga by Supercritical Hexane and

Sep 26, 2016 - ... Hachemi , Andre Rudnäs , Annika Smeds , Kari Eränen , Atte Aho , Narendra Kumar , Jarl Hemming , Markus Peurla , Dmitry Yu Murzin...
1 downloads 0 Views 3MB Size
Download PDF
Article pubs.acs.org/IECR

Extraction of Lipids from Chlorella Alga by Supercritical Hexane and Demonstration of Their Subsequent Catalytic Hydrodeoxygenation Hoang S. H. Nguyen,†,‡ Imane Hachemi,‡ Andre Rudnas,‡ Paï vi Mak̈ i-Arvela,‡ Annika Smeds,‡ Atte Aho,‡ Jarl Hemming,‡ Markus Peurla,§ and Dmitry Yu. Murzin*,‡ †

Lappeenranta University of Technology, Lappeenranta, 53850, Finland Åbo Akademi University, Turku, 20500, Finland § University of Turku, Turku, Finland ‡

S Supporting Information *

ABSTRACT: Extraction of lipids from Chlorella algae with supercritical hexane resulted in the high lipids yield of approximately 10% obtained at optimum conditions in terms of extraction time and agitation compared to the total content of lipids being 12%. Furthermore, an easiness of hexane recovery may be considered as economically and ecologically attractive. For the first time, in the current work catalytic hydrodeoxygenation (HDO) of Chlorella algal lipids was studied over 5 wt % Ni/SiO2 at 300 °C and under 30 bar total pressure in H2. The conversion of lipids was about 15% as the catalyst was totally deactivated after 60 min. The transformation of lipids proceeded mostly via hydrogenation and hydrogenolysis with formation of free fatty acid (FFA). Lower activity might be attributed to deactivation of catalysts caused by chlorophylls and carotenoids. Even though the conversion is low, future studies in HDO of lipids extracted from other algae species having higher lipid content could be proposed. A coke resistant catalyst might be considered to improve catalytic activity.

1. INTRODUCTION Algae have recently been identified as attractive feedstocks for biofuels and biochemicals. First, algae are able to adapt to many living environments such as seawater or wastewater. More importantly, algae can also be used to remove carbon dioxide from the atmosphere via carbon fixation in photosynthesis.1 Carbon fixation makes algal biomass based fuel advantageous by excluding emitted CO2 known as neutral carbon.2 The diversity of feedstocks with algae containing proteins, carbohydrates, and lipids offers many opportunities for refined chemical products and biofuels. For example, proteins mainly contain amino acids, which could be utilized for amino alcohols production, whereas carbohydrates and lipids are the feedstocks for bioethanol or hydrogen and biodiesel or diesel production, respectively.1 Hitherto, numerous techniques for lipid extraction have been developed, especially industrial-scale applications have been attractive due to the growth of algae-based biofuel industry. However, algal lipid extraction at a commercial level is still at its infancy.3 One of the most common difficulties regarding the lipid extraction is breaking cell walls of algae, or so-called cell disruption. Various methods have been employed for cell disruption such as microwaves, sonification and bead-beating, high pressure homogenization, and heat disruption, etc.3 Traditionally, lipid extraction is widely carried out via the Bligh and Dyer’s method.4 However, the involvement of toxic solvents such as chloroform and methanol is the main barrier © 2016 American Chemical Society

for their large scale application concerning health risks and environmental issues. Although hexane is less efficient for lipid extraction from algae than the above-mentioned solvents,5 minimal nonlipid contaminants and higher selectivity toward neutral lipids fraction could be achieved by hexane.6 To penetrate the thick cell walls of algae, supercritical fluids of carbon dioxide have been studied for lipid extraction.7−9 Similar to carbon dioxide, the supercritical state of hexane might be advantageous for lipid extraction at the industrial scale because of its higher selectivity for neutral lipids and easiness of solvent recovery. However, knowledge on lipid extraction with supercritical hexane is still limited. In a previous study, lipids were extracted from Scenedesmus sp. with supercritical hexane for biodiesel production,6 in which fatty acid methyl ester (FAME) yields were considered as a comparative parameter. It should be noted that the main focus of the current work is on utilization of the lipids fraction, while high temperature treatment during the extraction in supercritical hexane might be detrimental for some other fractions present in algae, in particular proteins, which could be destroyed after exposure to high temperature. Received: Revised: Accepted: Published: 10626

August 9, 2016 September 21, 2016 September 26, 2016 September 26, 2016 DOI: 10.1021/acs.iecr.6b03041 Ind. Eng. Chem. Res. 2016, 55, 10626−10634

Article

Industrial & Engineering Chemistry Research

are raised above their critical values. In the present work, hexane was used as a solvent to extract nonpolar lipids because hexane is less toxic than chloroform. The supercritical state of hexane is achieved when temperature and pressure are raised over its critical point (234.6 °C and 30.3 bar).3 The supercritical lipid extraction carried out in the autoclave is shown in Supporting Information, Figure S1. In a typical experiment, 5 g of algae was added into the reactor followed by introduction of 100 mL of hexane. Argon was then purged to the reactor removing air to avoid even a minimal risk of explosion. The reaction mixture was heated to the desired temperature of 235 °C. The pressure was obtained by the high vapor pressure of hexane of approximately 34 bar at this temperature. Both temperature and pressure were monitored using temperature and pressure controllers (Parr 4843). The agitation was set to the desired speed. When extraction was complete, the mixture was cooled to room temperature by using a water cooling coil and depressurized. The algal biomass was separated by filtration. Nonpolar lipids were obtained by removal of hexane with a rotavapor. The extracted lipids were weighed gravimetrically and analyzed by a series of analytical methods including transmission electron microscopy (TEM), size exclusion chromatography (SEC), gas chromatography− mass spectroscopy (GC−MS), gas chromatography-flame ionization detector (GC-FID) and thermogravimetric analysis (TGA). The experimental investigation of the reaction conditions was identically made in duplicate. Figure S1 panels a and b present the different steps in the supercritical hexane extraction of algal biomass and the reactor setup, respectively. 2.3. Analysis of the Lipid Phase. The components in the liquid phase containing hydrocarbons, free fatty acids (FFA), and fatty acid methyl esters (FAME) were identified by comparison of their retention times with those of authentic standards and quantified by the internal standard method. The sample preparation for analysis is described below. The samples after being collected in glass vials were stored in a refrigerator at 4 °C. Thereafter they were silylated by the following procedure: 0.1 mL of a sample was diluted in 1 mL pyridine (Sigma-Aldrich, ≥ 99%). After that 0.1 mL of eicosane (10 mg/mL) (Acros Organics, 99%) was added as an internal standard associated with 0.1 mL of BSFTA (Acros Organics, ≥ 98%) and 0.05 mL of TMCS (Sigma-Aldrich, ≥ 98%). The samples were subsequently silylated in an oven at a temperature of 70 °C for 60 min. Finally, the samples were naturally cooled down before being injected to a gas chromatograph with a flame ionization detector (GC-FID) under the conditions of an injection volume of 1 μL, He pressure of 14.5 psi with a flow velocity of 75 mL/min, a capillary column (ZB-5HT Inferno, length of 30 m, internal diameter of 0.32 mm and film thickness of 0.25 μm), oven program from initial temperature of 70 to 205 °C at 5 °C/min hold 18 min, and then to 300 °C at 7 °C/ min hold 15 min. The signals were recorded by FID. The duplicate analyses were carried out based on duplicate extracted lipid samples. Since the triglyceride fraction cannot be detected by GC-FID with a long capillary column, it was treated and analyzed by GC-FID equipped with a short column−an Agilent J&W HP1/SIMDIST column (cut length of 7 m, internal diameter of 0.53 mm, and film thickness 0.15 μm). The collected samples were filled by 2 mL of methyl tert-butyl ether (MTBE) containing the following internal standards (0.04 mg/internal standard): heneicosanoic acid (C21:0, Sigma-Aldrich, 99%), betulinol (purified in our lab), cholesteryl heptadecanoate

Lipids contained in algae can be classified into neutral or nonpolar lipids and polar lipids. The former comprises acylglycerols (mono-, di-, and triglycerides (MDTG)), free fatty acids (FFA), and pigments (chlorophyll and carotenes). The latter consists of phospholipids and glycolipids.6,10 Depending on the solvent polarity used in algal lipid extraction methods, the corresponding lipids would be separated. Biodiesel is widely used for algae-based biofuel production11,12 since it has higher cetane number and does not contain sulfur or aromatics and more favorable combustion emission profile.13 However, it suffers from inherent drawbacks such as unfavorable cold flow properties and poor storage stability of fatty acid ester.14 In addition, because of the high oxygen content of ester, biodiesel has a lower heat content compared to conventional petroleum-based fuels.15 Hydrodeoxygenation has been studied as an alternative pathway to utilize the lipids derived from algae to overcome the aforementioned drawbacks of biodiesel. Microalgae oil mainly containing triglycerides has been chosen as a model compound for hydrodeoxygenation.16 Scheme 1 shows the reaction pathways of microalgal oil (triglycerides) transformations to alkanes. Scheme 1. Reaction Pathways in Hydrodeoxygenation of Microalgal Oil to Alkanesa

a

Figure adapted with permission from ref 17. Copyright 2013 The Royal Society of Chemistry.

To our best knowledge, there are no previous reports concerning direct hydrodeoxygenation of lipids extracted from algae. The aim of this work is to provide a comprehensive study and to demonstrate for the first time hydrodeoxygenation of lipids extracted from Chlorella algae.

2. MATERIALS AND METHODS 2.1. Raw Material. Chlorella algae purchased from Luontaistukku, Finland, were used without any pretreatment. According to the manufacturer, Chlorella algae contain over 60% protein and 12% lipids. The moisture content was not reported by the manufacturer. In the previous work of the authors the content of moisture was ca. 7% in dry Chlorella algae.11 The catalysts used for HDO experiments have been prepared and characterized previously in our lab.18 Other chemicals used in this work are of analytical grade. 2.2. Supercritical Lipid Extraction. The supercritical region is obtained when the temperature and pressure of fluids 10627

DOI: 10.1021/acs.iecr.6b03041 Ind. Eng. Chem. Res. 2016, 55, 10626−10634

Article

Industrial & Engineering Chemistry Research

deviation of selectivity were calculated on the basis of duplicate lipid extraction at identical reaction conditions.

(Sigma-Aldrich, > 95%), and 1,3-dipalmitoyl-2-oleoglycerol (Sigma-Aldrich, 99%). Docosanol was concentrated by evaporating MTBE at 40 °C under a nitrogen stream. The evaporation was repeated after adding 1 mL of acetone; 150 μL of the silylation reagent mixture including pyridine, BSTFA, and TMCS (1:4:1 v/v/v) was added, and the final mixture was silylated in an oven at 70 °C for 45 min. The silylated samples were analyzed by PerkinElmer Clarus 680 gas chromatograph equipped with a FID and a short column mentioned above, the oven program starting at the initial temperature of 100 °C, holding 0.5 min, and then heating to 340 °C at 12 °C/min and holding for 5 min. 2.4. Analysis of Extracted Lipids. 2.4.1. Determination of Lipid Yield and Selectivity. Lipids extracted from Chlorella algae by supercritical hexane contain nonpolar lipids. To identify the optimum conditions for supercritical lipid extraction, the extracted lipids were weighed gravimetrically to obtain the lipid yield which was calculated in the following way. The average and the standard deviation (SD) of the lipids yield were made based on duplicate lipid extraction with the same reaction conditions. Y [%] =

OL 100 F

S[%] =

MDTG + FFA × 100 OL

(2)

Where MDTG is the mass of mono-, di- and triglycerides and FFA (g) and OL is the mass of obtained lipids (g). 2.4.2. Fatty Acid Profile. To obtain fatty acids profile of the extracted lipids, the lipid sample was hydrolyzed by sulfuric acid as described by Li-Beisson et al. with modification.22 The transesterified sample was silylated as described in section 2.3 before being analyzed by GC−MS and GC−FID. In a typical measurement, 100 mg of extracted lipids was placed in an 8 mL vial, and the vial was filled with 2 mL of 2.5% H2SO4 in methanol (v/v). Thereafter the sample was heated in an oven at 80 °C for 1 h. Afterward, the sample was naturally cooled to room temperature. A 2 mL aliquot of pentane was added into the vial followed by 0.9% NaCl (w/v) to extract FAME. The vial was vigorously shaken, and the phases were separated by a centrifuge. The upper layer-FAME fraction in pentane was carefully transferred into another vial. The FAME was concentrated by evaporation of the pentane and dissolved by dodecane in a 10 mL volumetric flask prior to the analysis by GC−MS and GC-FID. The preparation procedure for GC−MS and GC-FID analysis was described in section 2.3. 2.5. Chlorophylls and Carotenoids Measurements. As chlorophylls and carotenoids being nonpolar lipids could be extracted together with MDTG and FFA, the measurement of their concentration is essential. Chlorophylls a and b and carotenoids were determined by a method described by Wellburn.23 Two droplets of lipids sample were pipetted into a 12 mL vial and weighed and dissolved in 10 mLof 80% acetone (80% acetone and 20% deionized water). The sample concentration was in the range of 1−1.7 mg/mL. The diluted samples were scanned from 400 to 800 nm with a sample interval of 0.1 nm using a Shimadzu UV-2550 UV−vis spectrophotometer. The concentrations of Chlorophyll a and b and carotenoids for 80% acetone as a solvent were calculated following eqs 3 to 5. All measurements were made on duplicate lipid samples extracted at identical reaction conditions.

(1)

where OL is mass of obtained lipids (g) and F is the mass of feedstock (g) Extracted lipids were further analyzed by HPLC-SEC to identify the selectivity to mono-, di- and triglycerides (MDTG) and FFA. The details of the analytic procedure has been reported previously.19−21 Tetrahydrofuran was used as an eluent. In a typical measurement, 2−3 droplets of a sample were pipetted into a 12 mL vial. The vial was weighed and filled with 10 mL of tetrahydrofuran. The sample concentration was in the range of 1.5−3.0 mg/mL. The samples were analyzed and quantified by known concentrations of calibration samples of soybean oil. The prepared samples were filtered with 0.2 μm syringe filters (containing a PTFE membrane). The samples were then analyzed by a HPLC-SEC system composed of a LT-ELSdetector (Low Temperature Evaporative Light-Scattering Detector, Sedex 85, Sedere LT-ELS-detector) with additional components including degasser (DGU-14A), the gradient pump (FCV-10ALVP), the fraction collector (Pharmacia LKB-HeliFrac), and the system controller (SCL-10AVP). HPLC-SEC was also equipped with three different columns, two similar Jordi Gel DVB 500a (300 mm × 7.8 mm) columns and one Guard column (50 × 7.8 mm). The two similar columns were used to get a better separation. These three columns were used for separation of components, and the peaks were measured with an integrator Shimadzu Class- VP 8v.6.12 SP5. In general, higher molecular components would elute faster than lighter molecular compounds. Typically, the detected components eluted in the following order: triglycerides, diglycerides, monoglycerides, FFA, and sterols. A typical SEC chromatogram of lipid samples extracted with supercritical hexane at reaction conditions of 2 h and 600 rpm is shown in Figure S2. It is noteworthy that a large number of samples should be analyzed in the same day to achieve representative data. The amounts of MDTG and FFA were calculated using a calibration curve of soybean oil. The selectivity to MDTG and FFA was calculated with eq 2. The average and standard

Chlorophyll a (Ca) (μg/mL) = 12.25A 663.2 − 2.79A 646.8

(3)

Chlorophyll b (C b) (μg/mL) = 21.5A 646.8 − 5.1A 663.2

(4)

Total carotenoids (Cx + c) (μg/mL) = (1000A471 − 1.82Ca − 85.02C b)/198 (5)

2.6. Thermogravimetric analysis. The thermogravimetric analysis (TGA) was performed by a Pyris-TGA, PerkinElmer instrument under nitrogen flow of 20 mL/min to avoid oxidation of sample. The lipid sample was inserted in a platinum pan and heated from 25 to 800 °C in synthetic air. 2.7. Scanning Electron Microscopy Analysis and EDXA Analysis. The morphology of solid samples was determined by a Zeiss Leo 1530 Gemini scanning electron microscope instrument equipped with a ThermoNORAN vantage X-ray detector. The energy dispersive X-ray analysis was also performed with the same instrument. The measurements 10628

DOI: 10.1021/acs.iecr.6b03041 Ind. Eng. Chem. Res. 2016, 55, 10626−10634

Article

Industrial & Engineering Chemistry Research were made at different points of the residue obtained from TGA. 2.8. Organic Elemental Analysis. The quantification for organic elements including N, C, H, and S was carried out for the lipid sample using an organic elemental analyzer Flash 2000 from Thermo Scientific. Duplicate measurements were performed. In a typical measurement, approximately 2 mg of samples and standards was carefully weighed into tin cups. The measurements were carried out under He and O2. Methionine was used as a calibration standard, while sulfanilamide was used as a check standard. The standards were purchased from Elemental Microanalysis Limited. Two measurements were carried out, and the standard deviation was calculated on the lipid sample extracted at optimum conditions. 2.9. Fourier Transform Infrared Spectroscopy (FTIR). The infrared spectrometry analysis was performed with an ATI Mattson FTIR. A drop of the Chlorella sample was placed between two thin KBr disks. The spectrum was recorded in the range 4000−400 cm−1 with a resolution of 2 cm−1 using 64 scans per second. Two scans were conducted with the lipid sample extracted at optimum conditions. 2.10. Transmission Electron Microscopy (TEM). High resolution TEM images were made from raw and spent algae, which were treated with 5 wt % KOH in ethanol. The treated algae samples were then fixed with the 5% glutaraldehyde solution in 0.16 mol/L s-collidine buffer and then followed by 2% OsO4 containing 3% potassium ferrocyanide in 2 h. The samples were dehydrated with different ethanol concentrations of 70%, 96%, and twice at 100% ethanol and embedded in 45359 Fluka Epoxy Embedding Medium kit. A 70 nm thick section of the samples was made with an ultramicrotome and stained with 1% uranyl acetate and 0.3% lead citrate. Those thin sections were scanned by a JEOL JEM-1400 Plus TEM operated at 80 kV acceleration voltage integrated with a Quemesa 11 MPix bottom mounted digital camera providing a resolution of 0.38 nm. 2.11. Hydrodeoxygenation Experiment. The catalyst was reduced prior to being added into the reactor. The solvent was added into the bubbling unit in which it was preheated to 100 °C to decrease its viscosity; this could help to avoid solvent losses during the transfer into the reactor. Thereafter, the reactor was thoroughly flushed with argon. The reaction mixture was heated to the desired temperature at 10 °C/min and pressurized with H2 AGA 99.999% to 30 bar. After reaching the desired temperature, the stirring started (1200 rpm) and reaction time was set to zero. In a typical experiment, 1 g of reactant, 100 mL of dodecane, and 0.25 g of catalyst were used. The samples were taken at the following intervals: 0, 30, 60, and 360 min for kinetic analysis. All collected samples were analyzed by gas chromatography.

Figure 1. (a) Selectivity to MDTG and FFA and lipids yield and (b) pigments (chlorophylls and carotenoids) concentration at different extraction times in supercritical hexane extraction from Chlorella algae. Conditions: 5 g of algae, 100 mL of hexane, 235 °C, 34 bar, and 300 rpm stirring speed. (The average and SD of yield and chlorophyll and carotenoids concentration obtained from duplicate lipid extractions at the same reaction conditions were presented. Only the average of selectivity was presented.)

content provided by a manufacturer). Figure 1a presents the yield of lipids extracted from Chlorella algae at different reaction times. The calculation was made for two samples of duplicate supercritical lipid extractions. The lipids samples obtained at different extraction times were then analyzed by SEC to identify the selectivity to mono-, di-, and triglycerides (MDTG) and FFA. The analysis showed that the selectivity to these components decreased when the reaction time increased as illustrated in Figure 1a. The reason is that longer extraction time could lead to efficient diffusion of lipids through cell walls of algae; however, also other nonpolar lipids (such as chlorophyll and carotenoids) and nonpolar fractions could be extracted together with MDTG and FFA. In addition to MDTG and FFA, chlorophyll a and b and total carotenoids were determined by UV−vis spectroscopy. As presented in Figure 1b, prolonging the duration of extraction could lead to the decrease of chlorophyll a and b. Because of low lipid yields in the range of 0.014−0.02 lipids/g algae extracted at 5 and 10 min, the presence of even small amounts of chlorophyll a and b in the extract could lead to their high concentration. Because of this reason, the standard deviation of chlorophyll a, for example, is larger. Meanwhile, the concentration of total carotenoids seems to be constant. The average concentration of chlorophyll a in the current work is approximately 1000 μg/g, which is somewhat similar to the one obtained by Iqbal and Theegala who used Nannochloropsis sp.24

3. RESULTS AND DISCUSSION 3.1. Lipid Extraction. 3.1.1. Effect of the Extraction Time Using Supercritical Hexane as a Solvent. To study the effect of extraction time, supercritical lipid extraction using hexane as a solvent was performed at different times of 5, 10, 20, 30, 60, 120, and 360 min. The temperature and stirring speed were set to 235 °C and 300−320 rpm, respectively. The results in Figure 1a show that the longer extraction time gave higher yields of lipids, which increased 6-fold after 2 h as shown in Figure 1a. The yield of approximately 10% remained nearly constant even when the extraction time was prolonged up to 6 h. This value is close to the theoretical yield of approximately 12% (the lipids 10629

DOI: 10.1021/acs.iecr.6b03041 Ind. Eng. Chem. Res. 2016, 55, 10626−10634

Article

Industrial & Engineering Chemistry Research 3.1.2. Effect of Stirring Speed. The effect of agitation was investigated at three different speeds of 300, 600, and 1200 rpm. In a previous study, 300 rpm was set as a stirring speed for lipids extraction from Scenedesmus sp. with supercritical hexane.6 The temperature was still set to 235 °C, while experimental duration was 2 h. Figure 2a shows the yield of

FFA were predominantly extracted and the impurities were only extracted to a minor extent at higher agitation. The difference in the concentrations of chlorophyll a and b at different agitations is not significant. Total carotenoids concentration, however, decreased slightly with increasing stirring speed as shown in Figure 2b. The lipids yield of 8.96% obtained at optimum conditions of 300 rpm for 2 h (HCH-2-300) is lower than that in the previous finding in ref6 when Scenedesmus sp. was also extracted by supercritical hexane. This comparison is in fact not completely fair since higher lipids yield obtained in ref 6 can be attributed to a higher lipid content of Scenedesmus algae species which could be up to 40%,1 while the total lipids content from Chlorella algae provided by manufacturer is only 12%. This content corresponds to nonpolar and polar lipids. The former is composed of acylglycerols (MDTG), free fatty acids (FFA), and pigments (chlorophyll and carotenoids). The latter consists of phospholipids and glycolipids. In the optimized conditions almost 75% of the initial lipid content has been extracted. The Chlorella and Scenedesmus and Nannochloropsis algae species belong to a group of thick-cell wall algae species. This might cause difficulties in their lipid extraction.27 Efficiency of cell disruption was analyzed by comparison of TEM analyses for raw algae and solid residues after treating them in supercritical hexane at 300 rpm for 2 h. As shown in Figure 3, intact cell walls and intracellular organs could be clearly seen

Figure 2. (a) Selectivity to MDTG and FFA and lipids yield and (b) pigments (chlorophylls and carotenoids) concentration at different stirring speeds in supercritical hexane extraction. Conditions: 5 g of algae, 100 mL of hexane, 235 °C, 34 bar, and 2 h extraction time. (The average and SD of yield and chlorophyll and carotenoids concentration obtained from duplicate lipid extractions at same reaction conditions were presented. Only average of selectivity was presented.)

lipids at different stirring speed in the current study. Interestingly, the higher was the stirring speed, the lower was the yield of lipids. In fact, the yield of lipids decreased approximately by 30% when stirring speed reached 1200 rpm. The fall in the yield at high stirring speed is attributed to low diffusion of lipids, which happens across the cell wall, since diffusive extraction is an influencing factor for lipid extraction and is dependent on the magnitude or intensity of bulk convection in the media.25 In addition, a gentle stirring speed has been found to accelerate cell lysis and to enhance extraction efficiencies.26 The calculation was made on duplicate samples for supercritical lipid extraction. SEC analysis was conducted to investigate selectivity to MDTG and FFA of lipids samples extracted with supercritical hexane using different stirring speeds as shown in Figure 2a. The results demonstrate that higher selectivity was observed with higher stirring speeds of 600 and 1200 rpm compared to that at 300 rpm. A higher selectivity was attributed to a lower yield of lipids. At low stirring speed of 300 rpm, nonpolar components seem to be extracted efficiently, giving as a result higher amount of impurities. On the other hand, MDTG and

Figure 3. Transmission electron microscopy images of Chlorella samples: (a) fresh Chlorella; (b and c) treated Chlorella by supercritical hexane at 300 rpm for 2 h.

for fresh Chlorella (Figure 3a), whereas, only cell walls were observed for the treated sample confirming the diffusive extraction mechanism (Figure 3b). For the same sample of the treated algae, an interesting finding was witnessed namely that the cell walls seemed to collapse together forming coke which mainly remained after extraction (Figure 3c). This is understandable as algae was treated at high temperature and pressure of 235 °C and 34 bar, respectively. The main extraction mechanism being diffusion in the current work is also similar to the one obtained in ref 28 in which subcritical cosolvents (hexane and ethanol) were used to extract lipids from microalgae pastes of Nannochloropsis sp; after extraction, the algal cell shrunk and collapsed, and some microholes were formed. Diffusive extraction has been reported as a main mechanism in lipid extraction from algae.25 The image of 10630

DOI: 10.1021/acs.iecr.6b03041 Ind. Eng. Chem. Res. 2016, 55, 10626−10634

Article

Industrial & Engineering Chemistry Research

Scenedesmus species could reach up to 40%.1 However, the fatty acid yield is quite similar to the one obtained by Olkiewicz et al.12 who used hydrated phosphonium ionic liquid for lipid extraction from Chlorella vulgaris. In addition to GC and SEC analyses, the HCH-2-300 lipid sample was also analyzed by FTIR. As shown in Figure S4, the FTIR displays peaks corresponding to various bond vibrations in lipids at wavenumbers of 3008, 2954, 2925, 2854, and 1739 cm−1. The obtained results are quite similar to those of other studies as presented in Table 2. More importantly, it is consistent with SEC and GC analysis results, in which most of the components observed were MDTG and FFA, respectively, corresponding to the total yield of fatty acids of 5 wt %. Other components such as polysaccharides, carbohydrates, and amides were also identified in the HCH-2-300 sample. The organic elemental analysis (OEA) showed that carbon and hydrogen are the main components in the lipid sample HCH-2-300 as shown in Table S1, which displays two measurements along with the average. The high content of carbon could result in a high heating value. On the other hand, the sulfur content higher than a maximum limit of 50 ppm (0.005%) for sulfur in diesel in EU might be unfavorable for biofuel production used in diesel engine without hydrodesulfurization. Small amounts of nitrogen represent proteins which were also observed by FTIR. Thermogravimetric analysis was performed for the lipid sample (HCH-2-300) as shown in Figure S5 featuring three decomposition stages. Such analysis was aimed to identify lipid components (MDTG, FFA, chlorophyll, and carotenoids) released predominantly in a temperature range of 250−500 °C. The first stage released 15−20% of the total volatiles components ranging from 170 to 250 °C which may represent aldehydes and ketones33 potentially formed during extraction via autoxidation. The second stage within 250−500 °C released over 60% of the volatiles which may correspond to fatty acids components. Two main peaks detected at 350 and 400 °C in the HCH-2-300 lipid sample are consistent with the results of SEC and UV−vis for MDTG, FFA, chlorophylls, and total carotenoids, in which 60% of these components were detected. The final stage in the range of 500−800 °C released 5% of volatiles, and only 0.45% of the char residue was formed. The obtained result was slightly different from the one obtained by Kebelmann et al.33 in which the two main peaks obtained from TGA of Chlamydomonas algae species were observed at 300 and

extracted algae residue in Figure 3b might be considered as treated algae at an intermediate extraction stage and the coke was formed as a final treatment when the holes were nearly destroyed as seen in Figure 3c. 3.1.3. Characterization of the Lipid Sample. The lipid sample produced via supercritical hexane extraction at optimum conditions of 2 h and 300 rpm (HCH-2-300) was transesterified according to the procedure described in section 2.8.2 and analyzed by GC−MS and GC-FID for identification and quantification of the fatty acid profile, respectively. Figure S3 presents the chromatograms of HCH-2-300 analyzed by GC− MS. Additionally, Table 1 shows the percentage of fatty acid Table 1. Fatty Acid Profile of HCH-2-300a crude lipids yield fatty acids composition (% FAME) myristic acid (C14:0) 7,10-hexadecadienoic acid (C16:2) 7,10,13-hexadecatrienoic acid (C16:3) 9-hexadecenoic acid (C16:1) hexadecanoic acid (C16:0) 15-methyl hexadecanoic acid methyl ester (C17:0) cis-10-heptadecenoic acid (C17:1) heptadecanoic acid (C17:0) linoleic acid (C18:2) oleic acid (C18:1) stearic acid (C18:0) fatty acids content (%) fatty acid yieldb (wt %)

8.96 ± 0.17 1.03 ± 0.030 0.38 ± 0.350 9.37 ± 0.023 11.69 ± 0.065 22.35 ± 0.066 0.66 ± 0.101 0.58 ± 0.033 3.88 ± 0.055 42.81 ± 0.163 3.06 ± 0.013 4.19 ± 0.129 53.59 ± 0.17 4.8

a

All measurements were made in duplicate and the standard deviation was calculated. The value was presented as average ± SD. bFatty acid yield (wt %) = yield of crude lipids (wt %) × fatty acid content (%)/100.

components in HCH-2-300 which have been derived from the corresponding FAME shown in Figure S3. As presented in Table 1, the unsaturated fatty acids are the main components being approximately 68%. The high content of unsaturated fatty acids is quite similar to the one obtained in ref 11 where nearly 65% unsaturated fatty acids were obtained from another Chlorella alga originating from China. The fatty acid yield in the current work of 4.8% is nearly 2fold lower than that obtained in the finding of Shin et al., who used supercritical hexane to extract lipids from Scenedesmus sp. algae.6 It is understandable since the lipid content in

Table 2. Frequencies and Band Assignment for FTIR Obtained from the HCH-2-300 Samplea

a

main peak (cm−1)

assignment

component

wavenumber range (cm−1)

3008 2954 2925 2854 1739 1708 1671 1461 1378 1273 1173 968

νasCH2 νasCH2/νasCH3 νasCH2 νasCH2 ν(CO) ν(CO) ν(CO) νasCH2 νs(C−O) νas(PO) ν(C−O−C)/νas(PO) ν(C−O−C)/νas(PO)

lipids lipids lipids lipids lipids esters protein (amide I) protein (amide III) carboxylic acids nucleic acids, phosphoryl groups polysaccharides polysaccharides

2809−3012 2809−3012 2809−3012 2809−3012 1745−1734 1538−1709

1191−1356 1200−900 1200−900

ref 29 30, 29, 29 29, 30 32 30, 30, 29, 29, 29,

31 30, 32 31

31 31 32 30 30

Two scans were made on HCH-2-300 sample and identical results were obtained. 10631

DOI: 10.1021/acs.iecr.6b03041 Ind. Eng. Chem. Res. 2016, 55, 10626−10634

Article

Industrial & Engineering Chemistry Research 400 °C and 78% of the volatiles represented fatty acids components. The char residue formed in TGA analysis was further characterized by EDXA. Four spots on the char residue were randomly chosen for the analysis shown in Figure S6. The results revealed that the background of the char is mainly made from Al2O3, since Al and O are the main elements in the char as presented in Table S2. The presence of Mg in the char is understandable since chlorophylls present in lipids contain Mg. In addition, a lower content of Ca is advantageous as this element is a poison for catalytic hydrodeoxygenation. Compared to SEM-EDXA results of char residue obtained from the spent algae presented in Figure S7 and Table S3, the absence of P representing phospholipids (polar lipid) in the lipid sample HCH-2-300 is in accordance with the obtained results when the supercritical lipid extraction using hexane was efficient mainly toward nonpolar lipids. The lower content of P might be advantageous since phospholipids could cause catalyst deactivation.34 3.2. Demonstration of Hydrodeoxygenation. Lipids extracted under supercritical hexane at 300 rpm for 2 h were studied in catalytic hydrodeoxygenation with 5 wt % Ni/SiO2. Figure 4 showing kinetics of the lipids transformation,

and presulfided NiMo/Al2O3 in a continuous flow microreactor,16 the conversion in this work is low. This might be attributed to catalyst deactivation caused by chlorophylls and carotenoids or sulfur present in the algal lipid sample. It was reported that organic sulfur compounds are more poisonous for nickel-based catalysts than inorganic sulfur compounds.34 In the present study, the sulfur content is 0.25% higher than the maximum limit of 50 ppm (0.005%) allowed in EU for sulfur in diesel. The transformation of lipids occurs mainly by hydrogenation and hydrogenolysis with the formation of FFA. In comparison, conversion in hydrodeoxygenation of stearic acid over 5 wt % Ni/SiO2 obtained in our previous study,18 was 100%, while selectivity to hydrocarbons was 96.9% as shown in Table 3. This is understandable since the amount of impurities detrimental for catalysis is higher in extracted lipids as compared to neat stearic acid. The HDO experiment was carried out only once since the reproducibility of the HDO experiment has been well studied.35−37

4. CONCLUSIONS Lipids extraction from Chlorella algae in supercritical hexane was conducted in the current work varying the extraction time and the stirring speed. Such extraction was revealed to be feasible since a lipids yield of 10% was obtained under the optimum conditions of 300 rpm and 2 h duration compared to the total nominal lipid contents of 12%. Furthermore, an environmentally friendly extraction with supercritical hexane is attractive due to easy solvent recovery. For the first time, hydrodeoxygenation of raw Chlorella algal lipid was studied over 5 wt % Ni/SiO2 at 300 °C and under 30 bar total pressure in H2. The conversion of lipids was about 15% after 6 h as 5 wt % Ni/SiO2 was totally deactivated after 60 min caused by chlorophylls and carotenoids. The selectivity of 5 wt % Ni/SiO2 toward hydrocarbon (C15−C18) is low at approximately 6%. Transformation of lipids proceeded mostly via hydrogenation and hydrogenolysis with formation of free fatty acids. Even though the lipid conversion is rather low, future studies in hydrodeoxygenation of lipids extracted from other algae species having higher lipid content could be proposed. Coke resistant catalysts would be more preferable than purification of lipids since MDTG and FFA could be lost during the purification process. Hydrodesulfurization of algal lipid is also suggested to diminish the amount of sulfur, which plays a key role in catalyst deactivation.

Figure 4. Kinetics of lipid transformation on 5 wt % Ni/SiO2. Conditions: 1 g of reactant and 0.25 g of catalyst in 100 mL of dodecane, 300 °C, 30 bar total pressure in H2, and 1200 rpm stirring speed.

illustrates that the concentration of MDTG decreased with time since they were hydrogenated and hydrogenolyzed to FFA. However, the concentration of FFA was nearly constant after 1 h which can be related to complete inactivation of 5 wt % Ni/SiO2. The products were various oxygenated compounds, while the selectivity to C15−C18 was a rather low 5.9% as presented in Table 3. Compared to a previous study in which microalgae oil extracted from Nannochloropsis salina strain was hydrodeoxygenated over 1 wt % Pt/Al2O3, 0.5 wt % Rh/Al2O3

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.iecr.6b03041.

Table 3. Conversion and Selectivity in Hydrodeoxygenation of Lipids on 5 wt % Ni/SiO2 (ID 1) and Stearic Acid on 5 wt % Ni/ SiO2 (ID 2) selectivity (%) after 6 h

a

ID

conversion (%) after 6 h

hydrocarbons (C15−C18)

FFA

MG

DG

TG

sterols + steryl esters

1 2a

15.25 100

5.90 96.9

41.65

13.49

33.38

2.17

3.42

The result has been published in ref 18. 10632

DOI: 10.1021/acs.iecr.6b03041 Ind. Eng. Chem. Res. 2016, 55, 10626−10634

Article

Industrial & Engineering Chemistry Research

■

of Biodiesel via in Situ Transesterification of Chlorella Alga and Its Catalytic Deoxygenation to Renewable Diesel. Fuel 2015, 155, 144. (12) Olkiewicz, M.; Caporgno, M. P.; Font, J.; Legrand, J.; Lepine, O.; Plechkova, N. V.; Pruvost, J.; Seddon, K. R.; Bengoa, C. A Novel Recovery Process for Lipids from Microalgae for Biodiesel Production Using a Hydrated Phosphonium Ionic Liquid. Green Chem. 2015, 17 (5), 2813. (13) Hoang, N. S. H.; Vinh, L. X.; Hien, T. T. Biodiesel Production from Waste Shortening Oil from Instant Noodle Factories. J. Mater. Cycles Waste Manage. 2016, 18 (1), 93. (14) Morgan, T.; Santillan-Jimenez, E.; Harman-Ware, A. E.; Ji, Y.; Grubb, D.; Crocker, M. Catalytic Deoxygenation of Triglycerides to Hydrocarbons over Supported Nickel Catalysts. Chem. Eng. J. 2012, 189−190, 346. (15) Santillan-Jimenez, E.; Crocker, M. Catalytic Deoxygenation of Fatty Acids and Their Derivatives to Hydrocarbon Fuels via Decarboxylation/decarbonylation. J. Chem. Technol. Biotechnol. 2012, 87, 1041. (16) Zhou, L.; Lawal, A. Hydrodeoxygenation of Microalgae Oil to Green Diesel over Pt, Rh and Presulfided NiMo Catalysts. Catal. Sci. Technol. 2016, 6, 1442. (17) Zhao, C.; Bruck, T.; Lercher, J. A. Catalytic Deoxygenation of Microalgae Oil to Green Hydrocarbons. Green Chem. 2013, 15 (7), 1720. (18) Hachemi, I.; Jenistova, K.; Mäki-Arvela, P.; Kumar, N.; Eränen, K.; Hemming, J.; Murzin, D. Y. Comparative Study of Sulfur-Free Nickel and Palladium Catalysts in Hydrodeoxygenation of Different Fatty Acid Feedstocks for Production of Biofuels. Catal. Sci. Technol. 2016, 6, 1476. (19) Tolvanen, P.; Mäki-Arvela, P.; Kumar, N.; Eränen, K.; Sjöholm, R.; Hemming, J.; Holmbom, B.; Salmi, T.; Murzin, D. Yu. Thermal and Catalytic Oligomerisation of Fatty Acids. Appl. Catal., A 2007, 330, 1. (20) Kuuluvainen, V.; Mäki-Arvela, P.; Rautio, A.-R.; Kordas, K.; Roine, J.; Aho, A.; Toukoniitty, B.; Ö sterholm, H.; Toivakka, M.; Murzin, D. Yu. Properties of Adsorbents Used for Bleaching of Vegetable Oils and Animal Fats. J. Chem. Technol. Biotechnol. 2015, 90, 1579. (21) Kuuluvainen, V.; Mäki-Arvela, P.; Eränen, K.; Holappa, A.; Hemming, J.; Ö sterholm, H.; Toukoniitty, B.; Murzin, D. Yu. Extraction of Spent Bleaching Earth in the Production of Renewable Diesel. Chem. Eng. Technol. 2015, 38 (5), 769. (22) Li-Beisson, Y.; Shorrosh, B.; Beisson, F.; Andersson, M. X.; Arondel, V.; Bates, P. D.; Baud, S.; Bird, D.; DeBono, A.; Durrett, T. P.; Franke, R. B.; Graham, I. A.; Katayama, K.; Kelly, A. A.; Larson, T.; Markham, J. E.; Miquel, M.; Molina, I.; Nishida, I.; Rowland, O.; Samuels, L.; Schmid, K. M.; Wada, H.; Welti, R.; Xu, C.; Zallot, R.; Ohlrogge, J. Acyl-Lipid Metabolism. In The Arabidopsis Book; The American Society of Plant Biologists, 2013; p e0161. (23) Wellburn, A. R. The Spectral Determination of Chlorophylls a and b, as Well as Total Carotenoids, Using Various Solvents with Spectrophotometers of Different Resolution. J. Plant Physiol. 1994, 144 (3), 307. (24) Iqbal, J.; Theegala, C. Microwave Assisted Lipid Extraction from Microalgae Using Biodiesel as Co-Solvent. Algal Res. 2013, 2 (1), 34. (25) Ranjan, A.; Patil, C.; Moholkar, V. S. Mechanistic Assessment of Microalgal Lipid Extraction. Ind. Eng. Chem. Res. 2010, 49 (6), 2979. (26) Yang, F.; Xiang, W.; Sun, X.; Wu, H.; Li, T.; Long, L. A Novel Lipid Extraction Method from Wet Microalga Picochlorum Sp. at Room Temperature. Mar. Drugs 2014, 12 (3), 1258. (27) Borowitzka, M. A. Algae for Biofuels and Energy; Borowitzka, A. M., Moheimani, R. N., Eds.; Springer: Dordrecht, The Netherlands, 2013; pp 77−89. (28) Chen, M.; Liu, T.; Chen, X.; Chen, L.; Zhang, W.; Wang, J.; Gao, L.; Chen, Y.; Peng, X. Subcritical Co-Solvents Extraction of Lipid from Wet Microalgae Pastes of Nannochloropsis Sp. Eur. J. Lipid Sci. Technol. 2012, 114 (2), 205. (29) Liu, J.; Mukherjee, J.; Hawkes, J. J.; Wilkinson, S. J. Optimization of Lipid Production for Algal Biodiesel in Nitrogen

Schematic diagram and picture of reactor used in supercritical lipid extraction, SEC chromatogram of lipids extracted from Chlorella alga at reaction conditions of 2 h and 300 rpm (HCH-2-300), a chromatogram of HCH-2-300 analyzed by GC−MS, FTIR, and TGA of HCH-2-300, SEM of char residues from TGA of HCH2-300 and from TGA of spent algae, element composition of HCH-2-300 from organic elemental analysis, weight percentage of elements in char residues of HCH-2-300 and spent algae (PDF)

AUTHOR INFORMATION

Corresponding Author

*E-mail: dmurzin@abo.fi. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work is a part of the activities in the Johan Gadolin Process Chemistry Centre. The authors acknowledge Dr. Peter Backman, Linus Silvander, and Sten Lindholm for performing TGA, SEM-EDXA, and OEA analysis. The authors are grateful to Dr. K. Eränen for his help with the experimental work.

■

REFERENCES

(1) Trivedi, J.; Aila, M.; Bangwal, D. P.; Kaul, S.; Garg, M. O. Algae Based biorefineryHow to Make Sense? Renewable Sustainable Energy Rev. 2015, 47, 295. (2) Searchinger, T. D.; Hamburg, S. P.; Melillo, J.; Chameides, W.; Havlik, P.; Kammen, D. M.; Likens, G. E.; Lubowski, R. N.; Obersteiner, M.; Oppenheimer, M.; Roberston, G. P.; Schlesinger, W. H.; Tilman, G. D. Fixing a Critical Climate Accounting Error. Science 2009, 326, 527. (3) Theegala, C. S. Algal Cell Disruption and Lipid Extraction: A Review on Current Technologies and Limitations. In Algal Biorefineries: Vol. 2: Products and Refinery Design; Prokop, A., Bajpai, K. R., Zappi, E. M., Eds.; Springer International Publishing, 2015; pp 419−441. (4) Bligh, E. G.; Dyer, W. J. A Rapid Method for Total Lipid Extraction and Purification. Can. J. Biochem. Physiol. 1959, 37, 911. (5) Lee, S. J.; Yoon, B.-D.; Oh, H.-M. Rapid Method for the Determination of Lipid from the Green Alga Botryococcus Braunii. Biotechnol. Tech. 1998, 12, 553. (6) Shin, H.-Y.; Ryu, J.-H.; Bae, S.-Y.; Crofcheck, C.; Crocker, M. Lipid Extraction from Scenedesmus Sp. Microalgae for Biodiesel Production Using Hot Compressed Hexane. Fuel 2014, 130, 66. (7) Mariod, A. A.; Matthäus, B.; Ismail, M. Comparison of Supercritical Fluid and Hexane Extraction Methods in Extracting Kenaf (Hibiscus Cannabinus) Seed Oil Lipids. J. Am. Oil Chem. Soc. 2011, 88, 931. (8) Mouahid, A.; Crampon, C.; Toudji, S.-A. A.; Badens, E. Supercritical CO2 Extraction of Neutral Lipids from Microalgae: Experiments and Modelling. J. Supercrit. Fluids 2013, 77, 7. (9) Taher, H.; Al-Zuhair, S.; Al-Marzouqi, A. H.; Haik, Y.; Farid, M.; Tariq, S. Supercritical Carbon Dioxide Extraction of Microalgae Lipid: Process Optimization and Laboratory Scale-Up. J. Supercrit. Fluids 2014, 86, 57. (10) Li, Y.; Ghasemi Naghdi, F.; Garg, S.; Adarme-Vega, T. C.; Thurecht, K. J.; Ghafor, W. A.; Tannock, S.; Schenk, P. M. A Comparative Study: The Impact of Different Lipid Extraction Methods on Current Microalgal Lipid Research. Microb. Cell Fact. 2014, 13 (1), 14. (11) Viêgas, C. V.; Hachemi, I.; Freitas, S. P.; Mäki-Arvela, P.; Aho, A.; Hemming, J.; Smeds, A.; Heinmaa, I.; Fontes, F. B.; da Silva Pereira, D. C.; Kumar, N.; Aranda, A. A. G.; Murzin, D.Yu. A Route to Produce Renewable Diesel from Algae: Synthesis and Characterization 10633

DOI: 10.1021/acs.iecr.6b03041 Ind. Eng. Chem. Res. 2016, 55, 10626−10634

Article

Industrial & Engineering Chemistry Research Stressed Cells of Dunaliella Salina Using FTIR Analysis. J. Chem. Technol. Biotechnol. 2013, 88 (10), 1807. (30) Benning, L. G.; Phoenix, V.; Yee, N.; Tobin, M. Molecular Characterization of Cyanobacterial Silicification Using Synchrotron Infrared Micro-Spectroscopy. Geochim. Cosmochim. Acta 2004, 68 (4), 729. (31) Dean, A. P.; Martin, M. C.; Sigee, D. C. Resolution of Codominant Phytoplankton Species in a Eutrophic Lake Using Synchrotron-Based Fourier Transform Infrared Spectroscopy. Phycologia 2007, 46 (2), 151. (32) Ponnuswamy, I.; Madhavan, S.; Shabudeen, S. Isolation and Characterization of Green Microalgae for Carbon Sequestration, Waste Water Treatment and Bio-Fuel Production. Int. J. Bio-Sci. BioTechnol. 2013, 5, 17. (33) Kebelmann, K.; Hornung, A.; Karsten, U.; Griffiths, G. Intermediate Pyrolysis and Product Identification by TGA and PyGC/MS of Green Microalgae and Their Extracted Protein and Lipid Components. Biomass Bioenergy 2013, 49, 38. (34) Irandoust, S.; Edvardsson, J. Poisoning of Nickel-Based Catalysts in Fat Hydrogenation. J. Am. Oil Chem. Soc. 1993, 70 (11), 1149. (35) Şenol, O. I.̇ ; Viljava, T.-R.; Krause, A. O. I. Hydrodeoxygenation of Methyl Esters on Sulphided NiMo/γ-Al2O3 and CoMo/γ-Al2O3 Catalysts. Catal. Today 2005, 100 (3−4), 331. (36) Şenol, O. I.̇ ; Viljava, T.-R.; Krause, A. O. I. Hydrodeoxygenation of Aliphatic Esters on Sulphided NiMo/γ-Al2O3 and CoMo/γ-Al2O3 Catalyst: The Effect of Water. Catal. Today 2005, 106 (1−4), 186. (37) Şenol, O. I.̇ ; Ryymin, E.-M.; Viljava, T.-R.; Krause, A. O. I. Reactions of Methyl Heptanoate Hydrodeoxygenation on Sulphided Catalysts. J. Mol. Catal. A: Chem. 2007, 268 (1−2), 1.

10634

DOI: 10.1021/acs.iecr.6b03041 Ind. Eng. Chem. Res. 2016, 55, 10626−10634