Integrated in situ transesterification for improved biodiesel production from oleaginous yeast: a value proposition for possible industrial implication

Abstract

The conventional biodiesel production process using oleaginous yeast biomass often involves multiple energy intensive unit operations and processing steps that increase the overall cost and reduce the economic viability of the biodiesel product. Thus, this study attempts to design and optimize an in situ process for the direct conversion of lipids in disrupted wet biomass of oleaginous Pichia guilliermondii with an average total lipid content of 50 ± 2% [w/w, on dry cell weight (DCW) basis] to biodiesel, while bypassing important steps of biomass processing such as drying and lipid extraction. The in situ process involved applying sonication as an energy efficient cell disruption strategy that helped extract 44.5 ± 2.3% (w/w) neutral lipid on a dry cell weight (DCW) basis, using methanol-hexane as the most appropriate binary solvent system. Subsequently, the critical transesterification parameters such as biomass : methanol (w/v), catalyst concentration (v/v, %), reaction time and temperature that influence in situ biodiesel production were standardized. A maximum FAME (fatty acid methyl esters) yield of 92% (w/w of lipid), was achieved. This yield is comparable to that obtained by the ex-situ multistep transesterification process that requires approximately 7 h more than the in situ process, thereby resulting in greater productivity. The properties of the biodiesel product, as calculated from the FAME profile using empirical equations, conformed to the ASTM and CEN standards for it to qualify as an alternative to petro-diesel. Reports on direct transesterification of yeast biomass are scant. Thus, to the best of our knowledge, the developed in situ process integrating cell disruption, lipid extraction and transesterification is more energy efficient and productive as compared to those reported on yeast or algal feedstocks.

---

1. Introduction

Continuous usage of fossil fuels over the last few decades and its associated impact on global climate change has prompted world scientific communities to search for renewable and sustainable fuel sources. Bio-oil production from various lignocellulosic biomass has been studied but most of the conversion processes involve high temperature and pressure.¹ Among various alternative sources, single cell oil (SCO) from oleaginous microbes has been proven advantageous mainly because of ease of culturing and scale up, short life cycle and non-dependence of growth on external conditions of these yeasts. Oleaginous microbes are defined as those that can accumulate lipid more than 20% of their dry cell weight. Among oleaginous microbes, lipids from yeasts have gained more attention due to their shorter doubling time and their associated higher lipid titre, production rate and yield.² Several species such as Rhodotorula, Yarrowia, Rhodosporidium, Lipomyces, etc. have been reported to accumulate lipid up to 70% of their dry cell weight.^3,4 The lipids produced under growth-limited conditions are mostly in the form of triacylglycerol (TAG) containing long chain fatty acids, whose composition is well comparable to those obtained from food crops, thus rendering them as most desirable biodiesel feedstock.⁵

Although 2^nd generation biofuels offer several advantages over 1^st generation biofuels, the overall production economy is determined primarily by three sequential processes that are involved in making them, namely, cell disruption, lipid extraction and transesterification. So far, major research endeavours towards achieving a viable biofuel production process have continuously focussed on reducing the adverse impacts of using energy-intensive cell disruption process, longer lipid extraction procedure and large amount of toxic solvents.⁶

On the other hand, conversion of lipid-intact-wet biomass to FAME derivatives through direct transesterification, also called in situ transesterification circumvents the need for energy and time consuming dehydration process and also by combining lipid extraction and transesterification into a single step. It reduces the overall cost by eliminating the steps of extracting and refining the oil. Early attempts of in situ transesterification was investigated using plant based feedstock such as sesame seeds, jatropha seeds, sunflower flower seeds, which resulted in only poor efficiency with yields up to 20% of conventional process.⁷ However, subsequent researches based on optimization of critical parameters influencing in situ transesterification process improved the efficiency of the process to over 90%.⁸ Only very recently, this concept has gained considerable interests towards production of biodiesel from microalgal and yeast biomass.^9,10

The major bottle necks of in situ transesterification process that might cause reduced lipid extraction efficiency include ineffective contact between lipid and extracting solvent due to the usage of wet biomass and hydrolysis of some of fatty acid esters products back to alcohol and free fatty acids due to reversible nature of transesterification.¹¹ Despite these challenges at reaction level, the inherently simple in situ transesterification process if designed systematically may lead to a cost-effective biodiesel manufacturing process.

Among various factors influencing the in situ transesterification process, the choice of suitable binary solvent system plays a critical role, as it determines the extent to which the solvent can effectively extract the intracellular lipid for transesterification. In addition, the type of microbial species greatly influences the choice of the binary solvent system. Most often, a combination of a polar and a non-polar solvent has been used, with the former being relatively permeable through the cell membrane facilitates effective extraction of neutral lipids by the non-polar solvent. Other factors such as biomass to solvent ratio, catalyst concentration, temperature and reaction time have also demonstrated to have profound influence on transesterification efficiency.¹² It is presumed that any improvement in the downstream processing such as enhanced lipid recovery or implementation of in situ transesterification would improve the competitiveness of microbial biodiesel production. Till date, only a few studies have been reported to improve the process for maximizing lipid extraction towards optimal biodiesel production. In order to realize the commercial potential, evaluation of the current technologies for microbial oil (MO) production in terms of energy demand must be done. The current research challenges are as to how to reduce reaction time, minimize product loss by eliminating certain steps, develop less energy intensive drying process, avoid the use of large volumes of toxic solvents and catalysts that pose significant environmental and health threats.

Current study, therefore, attempts to improve the overall efficiency of biodiesel production from wet yeast biomass by integrating cell disruption-cum-lipid extraction with in situ transesterification reaction. This process integration helps in eliminating the use of cost and energy intensive equipment like dryer and lipid extractor. The next step would be to standardize the critical process parameters, such as catalyst concentration, biomass : methanol (w/v) ratio, reaction time and temperature for enhanced biodiesel production and characterize the biodiesel product in terms of major fuel properties to check for compliance with the ASTM and CEN specifications.

2. Materials and methods

2.1. Strain and culture conditions

Pichia guilliermondii, isolated from a Kharagpur-based oil mill, West Bengal, India was used as the candidate oleaginous microbe for biodiesel production. It was stored in 20% glycerol at −80 °C. The culture was revived by streaking a loop-full of culture in Nutrient agar Agar (Himedia, India) and then grown in MGYP broth in two stages at 28 °C and 180 rpm, each time adding 5% (v/v) inoculum. MGYP broth was then used to inoculate 100 mL Reader medium ((NH₄)₂SO₄, 3 g L⁻¹; KH₂PO₄, 1 g L⁻¹; K₂HPO₄·3H₂O, 0.16 g L⁻¹; MgSO₄·7H₂O, 0.7 g L⁻¹; NaCl, 0.5 g L⁻¹; Ca(NO₃)₂·4H₂O, 0.4 g L⁻¹) with 50 g L⁻¹ crude glycerol as C-source in 1 L shake flask with shaking condition and temperature similar to previous stage. After 72 h, the cells were harvested by centrifuging the culture broth at 25 °C, 8000 rpm for 10 min.

2.2. Conventional transesterification

The harvested biomass was dried in hot air oven, disrupted by sonication and the intracellular lipids (free fatty acids and triacylglycerols) were extracted by Bligh & Dyer¹³ method. The lipids were transesterified to their corresponding methyl esters by reaction with methanol using 2% sulphuric acid as the catalyst at 65 °C for 4 h in shaking condition. The free fatty acid content of the yeast lipid was measured to be relatively high around 22.3% by the classical titration method involving KOH, thus necessitating the transesterification reaction to be carried out with acid catalyst rather than base catalyst to avoid loss of free fatty acid due to neutralisation. The biomass : methanol ratio was kept at 1 : 5 (w/v). The hexane phase containing the FAMEs was recovered and analyzed in Gas Chromatography (Thermo Fisher Scientific, India, Chemito Ceres 800+).

2.3. Lipid extraction by different cell disruption techniques

About 0.5 g of harvested wet biomass (unless and otherwise stated) was subjected to different cell disruption procedures like bead beating, ultrasonication and homogenization, whose procedures are given below and the most suitable technique was chosen based on the percentage lipid recovered from each of these methods. Different cell-disruption procedures used were as follows:

2.3.1. Bead beating. Dry biomass (equivalent to 0.5 g wet biomass) was mixed with methanol (10 mL) and disrupted for 10 min in bead beater (Biospec Mini Bead Beater 16, Unigenetics Instrument, India) loaded with glass beads (0.2 mm) of about 50% of biomass. Rated power of the instrument was 851 W.

2.3.2. Sonication. As stated, the biomass was treated with 10 mL methanol and disrupted in sonicator (Q SONICA, Q125, U.S.A.) with a power rating of 125 W for 10 min in pulse mode (45 s On; 15 s off) at 50% amplitude.

2.3.3. Homogenization. Homogenization was carried out in a lab-scale homogenizer (SCILOGEX U.S.A., D-500), which disrupts cells by the shear force generated in the narrow space enclosed between the rotating cylindrical steel-shaft and the outer static hollow steel casing. The shaft was rotated at a speed of 20 000 rpm with similar on-off timing as that of sonication, however done manually and the suspension was maintained in ice-cold bath (15 min before and during homogenization) to prevent the loss of methanol due to evaporation. Rated power of the instrument was 500 W.

2.4. Lipid analysis

SEM and fluorescence microscopy techniques were used for qualitative assessment of lipids obtained by different cell-disruption procedures. The amount of actual lipid present in the sample was quantified by gravimetric assay.

2.4.1. Microscopy. For fluorescence microscopy (Olympus IX51, Japan), cells before and after disruption were subjected to nile red staining (1 μg mL⁻¹ nile red in acetone) and lipid content was analyzed semi-quantitatively based on the size and intensity of yellow-gold fluorescence emitted by the intact neutral lipid. The cells were viewed under 100× with blue light. The disrupted biomass was further analysed under analytical Scanning Electron Microscope (Zeiss, Oxford Instruments, England) to check the efficiency of each process.

2.4.2. Gravimetric analysis. Lipids obtained through different cell-disruption techniques were extracted with chloroform : methanol following the method of Bligh and Dyer¹³ in lipid extractor (J.P. SELECTA Spain, Extractor fat Det-Gras N 6SAM, no. 4002842). While the solvent phase was recovered, the lipid-rich phase was collected in the cups. Hexane was used to selectively extract neutral lipids from other lipid fractions comprising mainly of polar lipids. The amount of liquid obtained at each step was calculated gravimetrically by evaporating the accompanying solvent phase using either a hot-water bath or vacuum evaporator. All the samples were stored at 4 °C for further use.

2.5. Selection of binary solvent system

In this study, different binary solvent systems comprising hexane as non-polar solvent and ethanol, methanol, isopropanol and dichloromethane as polar solvents were standardized for maximum lipid yield from both dry and wet biomass. The polar and non-polar ratio of 1 : 1 (% v/v) was maintained for all binary solvents. A control experiment was conducted following the method of Bligh & Dyer.¹³ For all cases, 0.5 g dry biomass and equivalent wet biomass (75% moisture content) were subjected to disruption by sonication followed by extraction in lipid extractor. Since the solvent system is binary in nature, the boiling temperature is maintained for the higher boiling point component of the mixture. Lipids obtained were quantified by gravimetric analysis.

2.6. In situ transesterification

In situ transesterification was carried out with wet biomass. Wet-biomass slurry equivalent to 0.5 g of dry weight was used for this purpose, disrupted in sonicator and transesterified under the optimal transesterification parameters.

The critical process parameters for direct transesterification of wet yeast biomass were systematically standardized in this study which includes biomass to methanol ratio, catalyst concentration, time and temperature. Sulphuric acid was used as the catalyst for transesterification reaction and it was varied from 2% to 8% (v/v). The ranges of other critical parameters chosen are as follows: biomass to methanol ratio, 1 : 5 to 1 : 40 (w/v), reaction time, 4 to 12 h and temperature 40–70 °C. The FAME yield% (mg FAME/mg neutral lipid × 100) was calculated based on weight of neutral lipid obtained in control, since the biomass used were from same batch.

2.7. Energy consumption in conventional and in situ transesterification

Energy consumption at each step of the transesterification process was calculated based on the following equation

Energy Consumption by unit operation/process (kW h) = Power rating of the instrument (kW) × Time of operation (h) (1)

The detail calculation have been included in ESI.†

2.8. FAME analysis

The qualitative and quantitative determination of FAME was carried out using gas chromatography equipped with BPX 70 capillary column (30 m × 0.25 mm) and Flame Ionization Detector (FID), according to the protocol described by Dineshkumar et al.¹⁴ The operating conditions were maintained as: injector temperature, 260 °C; detector temperature 280 °C; oven temperature starting at 70 °C for 1 min, and increasing at 5 °C min⁻¹ to 180 °C for 10 min and 6 °C min⁻¹ to 220 °C for 11 min and a split ratio of 1 : 25. FAME mix 37 component (Supelco, Sigma) and methyl nonadecanoate (C19:0) was used as standard for quantification and as an internal standard respectively.

2.9. Calculation of biodiesel properties

The biodiesel properties were calculated based on the set of equations by Tanimura et al.¹⁵ The following properties were calculated using the equations:

Viscosity = −0.6316AU + 5.2065

Specific gravity = 0.0055AU + 0.8726

Cloud point = −13.356AU + 19.994

Cetane number = −6.6684AU + 62.876

Iodine number = 74.373AU + 12.71

HHV = 1.7601AU + 38.534

where, AU (average unsaturation) = ∑N × C_i, N = number of double bonds in unsaturated fatty acids, C_i = concentration (mass fraction) of the component.

3. Result and discussion

3.1. Effect of different cell disruption techniques on lipid extraction

Yeast cell wall contains approximately 1–2% of chitin microfibrils and thus requires disruption or destabilization of the cell wall prior to lipid extraction. Several techniques of yeast cell wall disruption have been reported in literature like ultrasound disintegration, osmotic shock, homogenization and application of several types of chemicals. However, mechanical methods of disruption (sonication and homogenization) were found to be more effective than other methods. Each of these methods adopts different mechanisms to disrupt cells, thus a systematic evaluation is often required before choosing a particular method. For instance, sonication, capable of generating high frequency ultrasound (above 16 KHz) breaks open the cell by inducing microbubble cavitations, while, bead milling destabilizes the cell membrane by high speed impacting of cells with fine beads and is considered to be an efficient technique for yeast cell disruption compared to bacteria.¹⁶ On the other hand, high speed homogenization disrupts cells by combination of shear force and cavity formation. Although non-mechanical means of disruption requires less energy, their application is restricted only to small scale operation due to their poor economics and efficiency.¹⁷ Moreover, the usage of chemicals is considered unsustainable because of challenges in scale up. So, the purpose of this study is to find out an effective mechanical cell disruption method specific to Pichia guilliermondii cells, which would also prove to be workable for larger scales.

Fig. 1 shows the morphology of both undisrupted and disrupted cells by different methods under scanning electron microscope (SEM). Each of these methods adopting different mechanisms to disrupt cells, as stated earlier, resulted in different degrees of cell disruption as evident from the microscopic images (Fig. 1 and 2). It could be seen that both sonication and bead beating (Fig. 1C and D) were more efficient in destabilizing the cells, resulting in almost complete disruption of cells. However, high-speed homogenization could not cause sufficient disruption of Pichia guilliermondii cells (Fig. 1B), probably due to the less magnitude/intensity of shear force generated by it. These findings were further corroborated by the results of fluorescence microscopy, which showed that cell-inbound-lipids seen in case of unbroken cells of control (Fig. 2A) were prominently seen lying outside the cells after disruption procedure, as shown in Fig. 2C and D corresponding to sonication and bead beating techniques. The fact that fungal cell wall structure is highly dynamic in nature, whose composition and mechanical properties not only varies among genera and species but also with stages in life cycle, makes the choice of cell disruption technique highly species-specific. For instance, maximum recovery of lipid from C. curvatus was achieved by sonication.¹⁸ In the current study, the extent of cell disruption by different techniques leading to the choice of the species-specific disruption technique was finally validated by gravimetric analysis based on the amount of neutral lipid extracted by hexane for individual method of cell disruption. Fig. 3 shows that sonication resulted in 40% (w/w) neutral lipids, followed by bead beating (35% w/w) and high-speed homogenization (12% w/w). The total amount of intracellular lipids in undisrupted cell were quantified to be 50 ± 2% by fluorescence spectroscopy (Fluoromax) using triolein and phosphatidylcholine as neutral and polar lipid standards respectively with nile red dye at specified excitation and emission wavelengths.¹⁹

Fig. 1 SEM images of (A) undisrupted (control), (B) homogenized, (C) sonicated and (D) bead beated yeast biomass.

Fig. 2 Fluorescence microscopic images of nile red stained cells (A) undisrupted (control), (B) homogenized, (C) sonicated, (D) bead beated, with a scale length of 10 μm.

Fig. 3 Lipid extraction efficiency by different cell disruption techniques.

3.2. Energy consumption by different cell disruption methods

As the cell disruption processes are energy intensive, a comparative assessment of energy requirement by different process is very important to design an optimal biodiesel production process. A brief survey of literature about energy requirement by individual cell disruption techniques considered in this study is presented here. Among the disruption techniques, sonication required less energy to disrupt a unit mass of cell. For disrupting 20 mL of 10 g L⁻¹ S. cerevisiae cell suspension, for about 20 min, sonication consumed 90 W, which is equivalent to 270 MJ kg⁻¹ dry biomass. The energy consumption was reduced by half for disrupting 200 mL of 8.5 g L⁻¹ of algal cells (Chlorococcum sp.) 5 min, consumed 750 W, which amounts to 132 MJ kg⁻¹ dry biomass. Energy consumption for bead milling is reported to be more at about 840 W for processing 100 mL algal cell suspension (5 g L⁻¹) for 5 min which is equivalent to 504 MJ kg⁻¹ dry biomass. High speed homogenization has been reported to consume energy 600 W (equivalent to 540 MJ kg⁻¹ dry biomass) when operated for 15 min for the disruption of 800 mL of cell suspension (10 g L⁻¹).²⁰ The energy consumption by different cell disruption processes of the current study is presented in Table 1. It can be seen that sonication is highly energy-efficient process, which consumes only 153 MJ kg⁻¹ of biomass. The next energy efficient technique, high-speed homogenization can be easily ruled out from consideration, as it resulted in only 12% neutral lipid extraction, as compared to 40% for sonication. Alternatively, the least energy efficient disruption technique, bead milling consumed about 1040 MW kg⁻¹, despite its recovery of 35% neutral lipids. However, it is strongly believed that energy consumption per unit biomass in bead milling can be considerably reduced by increasing the biomass charged per cycle of disruption, whose investigation is beyond the scope of the current study. Thus considering important factors such as scalability, scope for reduction in energy consumption and greater lipid recovery (of about 35% w/w) of bead milling process, it may act as economically viable cell disruption process for large scale production of biodiesel. However, through current study it is suggested that sonication is energy efficient only for small scale disruption of Pichia guilliermondii cells. Nevertheless, the comparative assessment made in the present study guided us choosing the most efficient cell-disruption technique, which further served as an important prerequisite to compare the efficiencies of erstwhile-used conventional transesterification process and novel in situ transesterification process for optimal biodiesel production, an important objective of the current investigation.

Table 1 Energy consumption in different cell disruption processes^a

Unit operations/processes	Energy consumption (MJ kg^−1 biomass)
a 1 kW h = 3.6 MJ.
Sonication	153
Homogenization	612
Bead beating	1040

---

3.3. Optimization of binary solvent system for maximum lipid yield

Neutral lipids in the form of triacylglycerols (TAGs) are the main feedstock for biodiesel production.²¹ So, the focus is to selectively extract neutral and other non-polar lipids from the cells by varying the solvent polarity in a single step. The fact “like dissolves like” is well known and has been successfully exploited here for efficient extraction of neutral lipids.

From Fig. 4 it can be seen that methanol : hexane system is second efficient in extracting lipid (44.5 ± 2.3% w/w of DCW) next to chloroform : methanol (44.8 ± 2%). A possible reason may be due to homogeneous catalysis since the later solvent system is miscible and chloroform with its higher polarity was conducive for cell wall rupture.²²

Fig. 4 Comparison of lipid yields (%) from dry and wet yeast biomass using different solvent systems.

However, chloroform has potential toxic effects and cannot be used in large scale. Thus, methanol : hexane was chosen over it, which is more environmental friendly and sustainable. It is also economical to choose hexane over chloroform since it is cheaper. Interestingly, methanol : hexane solvent system can act as a sole solvent system for extracting lipid as well as direct transesterification due its role in formation of methyl esters. Both the dry and wet biomass showed good lipid extraction efficiency, so we chose wet biomass for our further experiments. Direct use of wet yeast biomass is of considerable importance since the cost of biodiesel is reduced by removal of the drying step and it also simplifies the process.

3.4. In situ process for biodiesel production

Direct or in situ transesterification converts the lipids in wet biomass to FAMEs in a single step. The lipid extraction and transformation to FAME have been combined here to make the overall process less laborious and energy efficient. Reduction of steps also ensures lesser product loss. In order to determine the optimal conditions for the in situ process, a systematic series of experiments were carried out.

3.4.1. Effect of ​ biomass : methanol ratio. The effect of methanol loading on transesterification is of crucial importance from industrial perspective due to the fact that recycling the solvent is an expensive affair. Hence, optimum amount of methanol should be used in the process to achieve maximum conversion to FAME. Fig. 5A shows that the % conversion increased with increasing methanol concentration up to a threshold value (1 : 20 w/v) and decreased thereafter. This could be justified by the fact that increase in methanol concentration increases the polarity of the mixture and drives the reaction towards the formation of product, but increasing the methanol amount beyond the threshold level dilutes the biomass thereby reducing the conversion efficiency. Hence, a maximum FAME yield of 74% (w/w) was obtained for biomass : methanol ratio of 1 : 20 (w/v) (Fig. 5A). This result is corroborated with the study of Thliveros et al.²³ which obtained a higher FAME yield of 77.9%, upon optimization of biomass : methanol ratio for R. toruloides.

Fig. 5 Effect of biomass : methanol ratio and catalyst concentration (A), and reaction time and temperature (B) on % FAME yield.

3.4.2. Effect of catalyst concentration. In this study, four different acid concentrations as catalysts were tested. Fig. 5A shows the FAME conversion percentage for different catalyst concentration using biomass : methanol ratio of 1 : 20 (w/v). The FAME yield was found to improve with increasing the catalyst amount from 2% to 4%. However, the yield decreased slightly with further increase in catalyst amount. Both acid and base catalysts help further to destabilize the cell membrane in yeast and let the organic solvent access the intracellular lipid. Optimum usage of acid as catalyst have both industrial as well as scientific significance in the sense large use of acid is industrially not favourable and excess acid in reaction mixture cause side reactions leading to polymerization of unsaturated fatty acids.²⁴

3.4.3. Influence of reaction temperature. The effect of reaction temperature on biodiesel yield can be seen in Fig. 5B. It was observed that the FAME yield enhanced with increase in temperature from 40 to 60 °C. This is due to the fact that increasing the reaction temperature decreases the viscosity of lipids with enhanced reaction rate and reduced time²³ but only up to a critical level, beyond which the FAME yield starts reducing. Thus, it should be maintained at optimal for maximum FAME yield.

3.4.4. Influence of reaction time. Fig. 5B shows FAME yields for reaction time of 4 h to 10 h performed with 4% catalyst concentration, 1 : 20 (w/v) biomass : methanol ratio at 60 °C. The FAME yields were quite low during the initial hours and showed a maximum at 6 hours, after which the decline in yield is probably due to degradation of FAME because of the reversible nature of the transesterification reaction. A FAME yield of 92% was obtained after 6 h of acid catalysed reaction which is 11 hours faster than that reported by Thliveros et al.²³

3.5. Comparison of conventional and in situ transesterification

An almost equal percentage of lipids was extracted from both dry and wet biomass indicating that wet biomass can be directly used for in situ conversion to FAME (Fig. 4). According to the reports by Guldhe et al.²⁵ on biomass drying, oven drying (500 W) and lyophilisation (915 W) consumes about 6 kW h and 21.96 kW h in 12 h and 24 h respectively. Sun drying can be an alternative, but it is time consuming. Thus, significant amount of energy and time (around 7–8 h) can be saved by circumventing the drying step. The energy requirement of the lipid extractor (J.P. Selecta, Spain, Extractor fat Det-Gras N 6SAM, no. 4002842, 125 W) is about 34.72 kW h kg⁻¹ of biomass and the entire cycle of boiling, cooling and rinsing takes about 30 min. The transesterification step carried out on a hot plate-magnetic stirrer (Cole-Parmer, Model no. 88880001, 600 W) consumes about 3.6 kW h kg⁻¹ biomass (eqn (1)) when operated at 60 °C for 6 h. Hence, bypassing the drying step with direct transesterification step reduced the energy consumption by around 51.9% (eqn (2) and Table 2) with reduced process time of at least 7–8 h. The total energy consumption calculations for both conventional and in situ transesterification are included in ESI.† Further, the energy cost calculated for conventional and in situ process accounts for about 12.84 USD and 6.16 USD (calculated at the rate of 0.15 USD per kW h²⁶) respectively (Table 2) saving 6.68 USD for processing 1 kg biomass. In addition, capital cost can be reduced by avoiding usage of two major equipments, dryer and lipid extractor. A schematic of the entire in situ and conventional transesterification process has been shown in Fig. 6. An almost equal percentage of FAME yield obtained at the end of both the processes suggests that the in situ method can serve as a potential replacement for the conventional method. Thus, the in situ method can well be scaled up to pilot or industrial scale, as it is more sustainable and feasible in terms of solvent usage, energy efficiency, time requirement and reduced capital cost. A comparative FAME yield% obtained using different feedstock (such as microalgae, yeast, etc.) after in situ transesterification is presented in Table 3, which shows that our yeast strain can serve as a potentially better feedstock for biodiesel production.

Table 2 Comparative total energy consumption and cost involved in both conventional and in situ transesterification process

Technique involved	Energy consumption per unit operation/process (kW h kg^−1 biomass)				Total energy consumption (kW h)	Total cost to process 1 kg biomass (USD)^b
	Biomass drying	Cell disruption	Lipid extraction	Transesterification
a Energy consumption of hot air oven is 6 kW h kg^−1 biomass (Guldhe et al.^25).b On an average, cost of electricity is 0.15 USD per kW h (Dhanarajan et al.^26).
Conventional	6^a	42.5	34.72	2.4	85.62	12.84
In situ	—	42.5	—	3.6	41.1	6.16

---

Fig. 6 Schematic of the steps involved in conventional and direct/in situ transesterification.

Table 3 Comparison of FAME yields obtained after in situ transesterification from different species of microalgae and yeast

	Strain	Process parameters (catalyst concentration, solvent loading, time, temperature)	FAME yield (%)	Reference
Microalgae	Nannochloropsis sp.	2% (w/w) KOH; lipid : methanol, 1 : 400 (mol mol^−1); 4 h, 60 °C	90.94	Dianursanti et al.^27
	Chlorella sp.	10% v/v H2SO4, biomass : ethanol, 1 : 10 (w/v); 2 h, 90 °C	90.52	Zhang et al.^22
	Schizochytrium limacinum	7.5% H2SO4, 40 min, biomass : methanol, 1 : 7.4 (w/v), 90 °C	42.05	Johnson et al.^28
Yeast	Rhodosporidium toruloides	40% (w/v) NaOH, biomass : methanol, 1 : 20 (w/v); 10 h, 50 °C	97.7	Thliveros et al.^23
	Pichia guilliermondii	4% (v/v) H2SO4, biomass : methanol, 1 : 20 (w/v); 6 h, 60 °C	92	This study

---

3.6. FAME composition analysis

The fatty acid composition of the lipids obtained after transesterification is shown in Table 4. The saturated and unsaturated fatty acid contents play an important role in determining the suitability of its use as biodiesel. The greater the straight chain length of the biodiesel the better the fuel properties in terms of heating value and cetane number. An almost equal saturated : unsaturated fatty acid ratio in lipid is desirable for its use as biodiesel because higher percentage of saturated fatty acid can block fuel lines in winter due to solidification and higher unsaturation will lead to oxidative unstability. A higher percentage of oleic acid is most suitable for conventional biodiesel which matched with our results as obtained after in situ transesterification (Table 4).²⁹ In all cases, the major fatty acids were palmitic acid (16 : 0), stearic acid (18 : 0) and oleic acid (18 : 1) accounting more than 70% of the total fatty acids, similar to those reported by Zhang et al.³⁰ Yeast lipids in general have a fatty acid composition rich in C18 species, oleic acid ranging from 55.43–72.95%, stearic acid from 6–25% and linoleic acid (18 : 2) from 3.43–27.4%. A fatty acid composition of palmitic, stearic, oleic, linoleic and linolenic acid is considered ideal for biodiesel feedstock.¹⁵ Composition of fatty acids obtained from the algae C. curvatus is also similar in terms of percentage saturation (around 20%).⁶

Table 4 FAME composition of transesterified neutral lipids obtained by extraction with different binary solvents and in situ method

	Fatty acids	12 : 0	14 : 0	14 : 1	15 : 0	16 : 0	16 : 1	17 : 0	17 : 1	18 : 0	18 : 1	18 : 2	20 : 5	22 : 0	24 : 0
	Solvent systems
Conventional transesterification	Ethanol : hexane	ND	1.4	ND	2.8	42.8	ND	ND	ND	29.6	21.4	ND	ND	ND	2
	Methanol : hexane	0.6	ND	1.4	2.8	39.3	0.6	ND	1.3	22.3	26.4	2.8	ND	ND	2.4
	Isopropanol : hexane	3.7	ND	ND	3	40.1	ND	ND	ND	26.1	21.9	2.6	ND	ND	2.7
	DCM : methanol	ND	ND	ND	3.9	36.3	ND	ND	ND	23.7	30.8	3.7	ND	ND	1.5
	Chloroform : methanol	5.2	2	4.7	1.5	31.8	ND	2.4	1.8	22.1	25.9	ND	ND	ND	2.7
In situ		ND	0.2	2.5	2.1	30.5	4.5	0.6	4.4	11.2	34.8	7.1	ND	ND	2.1
References	Cui et al.^6	ND	0.868	ND	0.1	21.6	1.3	ND	0.075	12.2	45.2	13.4	ND	0.255	1.1
	Macías-Sánchez et al.^24	ND	3.4	ND	ND	17	17.3	ND	ND	0.4	3.6	7.2	22	ND	ND

---

3.7. Biodiesel properties

The physical and chemical properties of biodiesel can be predicted from its FAME composition profile. The biodiesel property varies significantly with the feedstock used. There are certain fuel standard setting organizations such as ASTM (in the U.S.) and European Committee for Standardization (CEN). ASTM D6751 and D7467 are the standards for biodiesel blendstock B100 and B6 to B20 respectively. The CEN standard for B100 is called EN 14214. Important properties of biodiesel which should be taken into account for its suitability as diesel fuel include viscosity, cetane number, cloud point, specific gravity, iodine value, heating value, cold filter plugging point and saponification value.

The measurement of biodiesel properties requires a considerable amount of oil sample and specialized equipment. This problem can be overcome by using a predictive set of equations which has been proven quite successful in estimating the physical properties of biodiesel.¹⁵ Table 5 shows the values of different biodiesel properties obtained from this study and successful comparison with ASTM D6751 and EN 14214 specifications. Higher heating value (HHV) corresponds with the energy content of the oil and usually a higher value is preferred. Cetane number (CN) indicates the ignition properties of fuel relative to cetane, which has a value of 100. The lowest value can be 15. Basically, it is the time delay in ignition of engine. The CN value depends on the chain length of fatty acids and increases with length. The minimum values for cetane number as per the ASTM and European standards are 47 and 51 respectively which matches with our result of 58.86. Iodine number gives a measure of the unsaturated fatty acids and is given by the mass of iodine consumed by 100 g oil. It depends greatly on the feedstock used. This property is considered as an important fuel property since a higher degree of unsaturation will lead to formation of deposits in the fuel line due to polymerization of glycerides upon heating. The effect increases with the degree of unsaturation. The European standard has the limit for this as 120 g I₂/100 g sample which has been met by our sample.³¹ Specific gravity of the biodiesel sample was found to be 0.875 kg L⁻¹ which exactly falls in the specified range of 0.86–0.9 kg L⁻¹ as per EN 14214 standard and also agrees with the values reported for Chicha seed oils (0.89 kg L⁻¹).³² Saponification value represents the number of milligrams of KOH required to saponify 1 g of fat and is given by the average molecular weight of the fatty acids present in the oil. Moreover, the biodiesel properties are in line with those obtained from vegetable oils such as rapeseed oil, jatropha oil¹⁵ and algal oil.³¹

Table 5 Performance evaluation of biodiesel produced through in situ transesterification vis-á-vis ASTM and European standards

Biodiesel properties	This study	Biodiesel from vegetable oil	Biodiesel from yeast lipid^a	ASTM D6751	EN 14214
a Tanimura et al.^15b Francisco et al.^31
Viscosity (mm s^−2)	4.74	4.4^a	4.4–4.7	1.9 to 6	3.5–5
Specific gravity (kg L^−1)	0.87	0.87^a	0.87	0.85	0.86–0.9
Cloud point (°C)	10.17	2.9–4.6^a	5.7–10.2	−3 to 12 °C	Not specified
Cetane number	57.97	54.3–55.2^a	57–58	47 min	51 min
Iodine number	67.41	98–107^a	66–92	Not specified	120 max
High heating value	39.82	40.5–40.7^a	39–40	Not specified	Not specified
Saponification value	204.06	217–266^b	Not specified	Not specified	Not specified

---

4. Conclusion

The novelty of this study lies in the systematic development of an in situ or direct transesterification process of disrupted oleaginous yeast biomass for biodiesel production. The process eliminates the use of two energy-intensive equipments, namely, dryer and lipid extractor and hence, is very efficient in terms of energy, time and capital cost requirements. The current study stands out in its approach towards the development of an improved in situ process for oleaginous yeast biomass based biodiesel production. Among the cell disruption methods employed, ultrasonication was found to be more effective and energy efficient in disrupting the yeast cells, as evident from the fluorescence and scanning electron micrographs and also indicated by the amount of lipid extracted. Methanol/hexane was found to extract significantly higher amount of lipids as compared to the other solvent systems. While methanol in combination with hexane destabilizes cell membrane, hexane extracts neutral lipids specifically. Further, methanol also acts as a reagent in transesterification reaction, thereby serving twin purposes. The optimized in situ process resulted in the maximum FAME yield of 92% (w/w) and the biodiesel properties conformed to the ASTM and CEN standards. Though the percentage conversion was comparable with the reported in situ processes, the optimized process requires lesser energy, time and capital expenditure and therefore, gains the economic and environmental advantages. Thus, this in situ process, if suitably scaled up can make a pragmatic value proposition for cost competitive manufacture of biodiesel.

Acknowledgements

The authors thankfully acknowledge Department of Biotechnology, Government of India (Project grant No.: BT/PR6909/PBD/26/391/2013, 21/03/2014) for the financial support. JC gratefully acknowledges the Ministry of New and Renewable Energy, Government of India for her fellowship and Department of Biotechnology, IIT Kharagpur for all the research facilities. RD is thankful to Department of Science and Technology (DST) INSPIRE, Government of India for his fellowship. JC is thankful to Dr Vivek Rangarajan and Dr Subhasish Das for their valuable technical inputs and suggestions.

References

1. A. W. Bhutto, K. Qureshi, R. Abro, K. Harijan, Z. Zhao, A. A. Bazmi, T. Abbas and G. Yu, RSC Adv., 2016, 6, 32140–32170.
2. F. Yang, C. Cheng, L. Long, Q. Hu, Q. Jia, H. Wu and W. Xiang, Energy Fuels, 2015, 29, 2380–2386.
3. Y. Li, Z. Zhao and F. Bai, Enzyme Microb. Technol., 2007, 41, 312–317.
4. X. Meng, J. Yang, X. Xu, L. Zhang, Q. Nie and M. Xian, Renewable Energy, 2009, 34, 1–5.
5. H. W. Yen, Y. T. Liao and Y. X. Liu, J. Biosci. Bioeng., 2016, 121, 209–212.
6. Y. Cui and Y. Liang, Appl. Energy, 2014, 119, 438–444.
7. K. J. Harrington and C. Darcyevans, J. Am. Oil Chem. Soc., 1985, 62, 1009–1013.
8. B. Kim, H. Im and J. W. Lee, Bioresour. Technol., 2015, 185, 421–425.
9. W. I. Suh, S. K. Mishra, T.-H. Kim, W. Farooq, M. Moon, A. Shrivastav, M. S. Park and J.-W. Yang, Algal Res., 2015, 12, 405–411.
10. H. Kakkad, M. Khot, S. Zinjarde, A. RaviKumar, V. Ravi Kumar and B. D. Kulkarni, Bioresour. Technol., 2015, 197, 502–507.
11. J. Y. Park, M. S. Park, Y. C. Lee and J. W. Yang, Bioresour. Technol., 2015, 184, 267–275.
12. S. Siler-Marinkovic and A. Tomasevic, Fuel, 1998, 77, 1389–1391.
13. E. G. Bligh and W. J. Dyer, Can. J. Biochem. Physiol., 1959, 37, 911–917.
14. R. Dineshkumar, S. K. Dash and R. Sen, RSC Adv., 2015, 5, 73381–73394.
15. A. Tanimura, M. Takashima, T. Sugita, R. Endoh, M. Kikukawa, S. Yamaguchi, E. Sakuradani, J. Ogawa and J. Shima, Bioresour. Technol., 2014, 153, 230–235.
16. A. P. Middelberg, Biotechnol. Adv., 1995, 13, 491–551.
17. M. Klimek-Ochab, M. Brzezinska-Rodak, E. Zymanczyk-Duda, B. Lejczak and P. Kafarski, Folia Microbiol., 2011, 56, 469–475.
18. X. Yu, T. Dong, Y. Zheng, C. Miao and S. Chen, Eur. J. Lipid Sci. Technol., 2015, 117, 730–737.
19. G. De Bhowmick, N. Vegesna and R. Sen, Bioresour. Technol., 2015, 198, 781–788.
20. A. K. Lee, D. M. Lewis and P. J. Ashman, Biomass Bioenergy, 2012, 46, 89–101.
21. M. H. Liang and J. G. Jiang, Prog. Lipid Res., 2013, 52, 395–408.
22. Y. Zhang, Y. Li, X. Zhang and T. Tan, Bioresour. Technol., 2015, 196, 712–715.
23. P. Thliveros, E. Uckun Kiran and C. Webb, Bioresour. Technol., 2014, 157, 181–187.
24. M. D. Macías-Sánchez, A. Robles-Medina, E. Hita-Peña, M. J. Jiménez-Callejón, L. Estéban-Cerdán, P. A. González-Moreno and E. Molina-Grima, Fuel, 2015, 150, 14–20.
25. A. Guldhe, B. Singh, I. Rawat, K. Ramluckan and F. Bux, Fuel, 2014, 128, 46–52.
26. G. Dhanarajan and R. Sen, in Biosurfactants. Production and utilization- processes,Technologies and economies, ed. F. V. S. Naim Kosaric, CRC Press, 2014, ch. 10, pp. 153–162,  DOI:10.1201/b17599-12.
27. Dianursanti, P. Religia and A. Wijanarko, Procedia Environ. Sci., 2015, 23, 412–420.
28. M. B. Johnson and Z. Wen, Energy Fuels, 2009, 23, 5179–5183.
29. S. K. Narwal and R. Gupta, Biotechnol. Lett., 2013, 35, 479–490.
30. X. Zhang, H. Shen, X. Yang, Q. Wang, X. Yu and Z. K. Zhao, RSC Adv., 2016, 6, 26752–26756.
31. É. C. Francisco, D. B. Neves, E. Jacob-Lopes and T. T. Franco, J. Chem. Technol. Biotechnol., 2010, 85, 395–403.
32. J. C. Gomes Filho, A. S. Peiter, W. R. O. Pimentel, J. I. Soletti, S. H. V. Carvalho and L. Meili, Ind. Crops Prod., 2015, 74, 767–772.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra14003c

---