**A Review on Insights for Green Production of Unconventional Protein and Energy Sources Derived from the Larval Biomass of Black Soldier Fly**

**Sabrina Hasnol 1, Kunlanan Kiatkittipong 2,\*, Worapon Kiatkittipong 3, Chung Yiin Wong 1, Cheng Seong Khe 4, Man Kee Lam 5, Pau Loke Show 6, Wen Da Oh 7, Thiam Leng Chew <sup>8</sup> and Jun Wei Lim 1,\***


Received: 3 January 2020; Accepted: 20 February 2020; Published: 28 April 2020

**Abstract:** The purpose of this review is to reveal the lipid and protein contents in black soldier fly larvae (BSFL) for the sustainable production of protein and energy sources. It has been observed from studies in the literature that the larval lipid and protein contents vary with the rearing conditions as well as the downstream processing employed. The homogenous, heterogenous and microbial-treated substrates via fermentation are used to rear BSFL and are compared in this review for the simultaneous production of larval protein and biodiesel. Moreover, the best moisture content and the aeration rate of larval feeding substrates are also reported in this review to enhance the growth of BSFL. As the downstream process after harvesting starts with larval inactivation, various related methods have also been reviewed in relation to its impact on the quality/quantity of larval protein and lipids. Subsequently, the other downstream processes, namely, extraction and transesterification to biodiesel, are finally epitomized from the literature to provide a comprehensive review for the production of unconventional protein and lipid sources from BSFL feedstock. Incontrovertibly, the review accentuates the great potential use of BSFL biomass as a green source of protein and lipids for energy production in the form of biodiesel. The traditional protein and energy sources, preponderantly fishmeal, are unsustainable naturally, pressingly calling for immediate substitutions to cater for the rising demands. Accordingly, this review stresses the benefits of using BSFL biomass in detailing its production from upstream all the way to downstream processes which are green and economical at the same time.

**Keywords:** black soldier fly larvae; protein; lipid; biodiesel; substrate; transesterification

#### **1. Introduction**

Fossil fuel holds the position of being the main source of energy consumed in the world. According to the World Energy Forum, the reserves of fossil-based oil, gas and coal, used mainly in the transportation, agriculture, domestic and industrial sectors, will be exhausted in less than a decade. As this main source of energy is rapidly diminishing at an alarming rate, it has accelerated the demands to find an alternative source that serves the same functions. This has lead researchers to consider renewable energy, offering not only improved energy security, but also a chance for the planet to reduce carbon emissions while providing much cleaner air. This in turn will permit the future generation to have a more sustainable green footing in regard to the environment. According to Barnwal and Sharma [1], fuels that are of biological origin, originating from vegetable oils, alcohol, biomass and biogas, are some of the alternatives presented from these past few years as sustainable fuels. Some of these fuels can be used directly, while others may need further modification before the fuels can be used. Biodiesel, one of the alternative fuels that originates from vegetable oils, animal fats and microorganisms such as microalgae, yeast, bacteria and fungi, shows promising results in becoming the main source of energy. For maximum yield, a transesterification process is carried out on the glyceride of the oily sources with alcohol in the presence of a catalyst to form fatty acid alkyl esters and glycerol [2]. However, biodiesel has challenges in implementation due to its high cost and limited availability of resources rising from the food versus fuel issue [3]. This is because the sources were limited to plant and animal feedstock, thereby competing with a food source needed for consumption. Microorganisms then became a new interest in synthesizing biofuels, making microbes such as bacteria, fungi and microalgae the next generation of biodiesel [4]. It was determined that microalgae contained the highest lipid content, over 75% measured relative to dry biomass weight [5]. However, this new source has led to the other problems such as extensive time consumption of medium preparation and intensive energy requirement for harvesting as microalgae are more buoyant and difficult to settle [6].

Thus, to generate biodiesel in a more favorable condition, researchers have suggested to derive the sustainable fuel from insects. Fuels derived from insects through insect farming allow several biochemical products and byproducts to be obtained, including proteins. Biodiesel production from insects has become more favorable since it has been found that insect breeding is economically and environmentally viable. Certain species of insects can easily degrade organic matter, converting organic waste into insect biomass. Insect breeding space is not large compared to the large land areas required for crops such as soybeans or to the large water footprint required for microalgae production. This new alternative has become more feasible, especially for countries with limited space and highly populated areas that need to devote their land for food-source production [7]. Insect larvae can accumulate lipids as their fat body and are able to stimulate the metabolic reserves needed, especially during their immature stages such as larva, pupa and nymph. Insects possesses a nutrient storage system that is used in the metamorphosis process, a structure called the "fat body". This structure is able to accumulate the lipids in the body as fat, which is used as an energy reserve and plays a role in the intermediary metabolism. From the research work conducted by Leong et al. [8], the *Hermetia illucens* larvae, or the black soldier fly larvae (BSFL), has become the ideal candidate in biodiesel production during its larval stage because the adult of the fly has been reported to be missing the mouthparts to feed and relies on food reserves, unlike common houseflies. This means that the black soldier fly is not a vector that can transmit diseases or parasites when feeding. Thus, this species of fly is not considered as a harmful pest, feeding on only kitchen waste, spoiled feed and manure. Recently, this fly, which can be commonly found in poultry- and pig-rearing units, has been found to be able to reduce unpleasant smells as it feeds on the manure or compost, efficiently reducing the polluting compounds from manures and compost. Undesirable bacteria are also reduced by the modification of the bacterial

microflora by the BSFL during feeding. The BSFL is a sustainable source for biodiesel production, as the chemical composition of this species is able to accumulate fat, depending on its feeding medium during its rearing process. Upon the lipid extraction for biodiesel production, the residual is a protein-rich larval biomass and can be used as the animal feed to replace fishmeal, which is not sustainable for the long term. Various research studies have been conducted on employing the BSFL biomass as the animal feed for farming of land animals as well as for aquaculture. Figure 1 presents the flow of the present review, starting from the BSFL substrate preparations all the way until the conditions for larval biodiesel production.

**Figure 1.** Flow of review, encompassing the larval substrate, rearing and biodiesel production conditions.

#### **2. Homogenous Substrate**

With a wide dietary range [9], BSFL has been evaluated for the precision and easy incorporation in formulating its diets that allow sufficient amount of lipid for biodiesel and protein production. Studies had been conducted on feeding BSFL with two different types of feeding mediums, namely, homogenous and heterogenous substrates. The homogenous substrates contain only one kind of organic matter, while the heterogenous substrates incorporate a mixture of two different types of organic matters or more before feeding the BSFL. For the homogenous substrates, there are various type of mediums that have been used to feed BSFL in order to assess its lipid content, biodiesel yield and protein content. Those single substrates are manure, animal feed/food, waste and nutritional meal.

Manure is basically an organic matter originating from the feces of animals, mainly used to fertilize crops. Different type of animals have different consumption of feeds in their diets, affecting the nutritional content of their manure. According to Li et al. [10], the use of cattle manure to feed BSFL would generate the extracted lipid content of 38.2 g, yielding 29.9% amount of fat. The biodiesel produced was 35.6 g and the BSFL that was fed with the cattle manure was able to yield 93% of biodiesel. When pig manure was used, the amount of lipid produced was 60.4 g with the yield of 29.1%, while the biodiesel produced was 57.7 g with the yield of 96%. The amount of lipid produced when chicken manure was used however gave the amount of 98.5 g with the yield of 30.1% and the biodiesel production of 91.4 g with yield of 93%. According to this study, the BSFL fat-based biodiesel fuel properties were comparable to a crop-based fuel, rapeseed. With the amount of crude fat as well as biofuel yield from the transesterification process, the results from this study show that BSFL fat has the potential as a feedstock in biodiesel production.

Other studies were also conducted by Newton et al. [11] in comparing the lipid and protein contents between poultry manure and swine manure. According to their studies, it was found that the lipid content of BSFL was slightly higher when fed with poultry manure, with the yield of 34.8%, while BSFL was able to yield 28% when fed with swine manure. The protein content of BSFL was higher when it was fed with swine manure, with yield of 43.2%, than when it was fed with poultry manure (42.1%). This study showed that BSFL contained a high concentration of oil that would yield as much energy as the methane fermentation that used the same type of manure. The difference in the

BSFL lipid and protein contents when reared by different types of manure reflects how the variation of diet affects the lipid and protein concentration, as it was tested that the other nutrients, except for phosphorus, can be found in slightly higher concentrations when fed with poultry manure. With its high level of oil in the BSFL, it would be likely best to not use BSFL as a bulk protein supplement for animal feedstock, but instead to use it as the potential energy source. According to the study conducted by Lalander et al. [9], poultry manure was fed to BSFL and the crude protein content obtained was 22.8%. It could be deduced that the development of the BSFL growth was dependent on the concentration of the protein of the BSFL. When the feed provides the BSFL with enough protein to accumulate, it will be used as part of the its development, making it consume less energy from its lipid content. However, it will result in a much smaller larva. In the same study, poultry feed gave the protein content of 17.3% and dog food gave 33.9%.

Lalander et al. [9] also investigated the effects on the concentration of crude protein of BSFL when they were fed waste materials. When food waste from local restaurants was used, 22.2% was obtained. Abattoir sheep waste gave 56.3%, human feces gave 35.5%, dewatered wastewater sludge gave 16.9%, sewage sludge gave 31.5% and the digested sludge gave the protein content of 14.7%. According to the protein conversion ratio, pure abattoir waste can have the potential to obtain a higher protein ratio if more carbon was added to allow nutrients in the substrates of the waste to be balanced. The nitrogen content of the waste can also be improved with added carbon as it allows the BSFL to utilize the protein content in a much higher usage during its development. The sludge may have low protein content as it has too few volatile solids. The feeding rate that was regulated to dry matter in this investigation was affected. Human feces has a high ratio, and this may be due to its biomass conversion ratio.

Another type of homogenous substrate was flour protein, as carried out by Arango Gutiérrez et al. [12], which contains proteinic ingredients and high digestibility that has the qualities that make it suitable for providing the right nutritional value in the animal's feed. According to the analysis, it is found that when the flour protein was fed to the BSFL, the larvae had the lipid content of 18.82% with the protein content of 36.98%. This research shows that the feed has potential ingredients to provide energy content.

#### **3. Heterogenous Substrate**

Lately, it has been found that the oxidation from the fiber of plants or crops is important factor that contributes greatly towards the metabolic activity of the BSFL. As reported by Li et al. [13], the fibers that exist provide the black soldier fly larvae the sufficient materials and energy required for life activities. Therefore, a balanced nutrition is required in the BSFL diet to ensure that the total conversion efficiency is enhanced; this will, in turn, assist the black soldier fly larvae's digestion of the materials. With a better nutrient balance, a higher yield is due to the synergy of the biological growth established being highly positive. According to Wu Li et al. [14], when corn cob residue was soaked in restaurant wastewater at the optimal soaking condition of 75 ◦C for 5 h, 23.34% lipid content was able to be produced from the BSFL. The restaurant wastewater was used to soak the corn because of its acidification properties, allowing cellulose hydrolysis which allows the lignocellulose of the corn to degrade easily. Different concentrations of xylose and glucose of a fibrous plant or crops in the BSFL feed greatly influences the insect's dry weight and the lipid content [13]. With xylose being the most abundant carbohydrate derived from fibers, especially corn, it became of great importance to extract the xylose to be able to produce lipid. Without any time lag, BSFL is able to consume both xylose and glucose of a plant, easily transforming it to lipid. When a standard feed with a mixture of 8% of glucose was added, 34.31% lipid was able to be yielded from the BSFL. If 6% xylose was mixed with the standard feed, 34.60% lipid was able to be yielded. This shows that both xylose and glucose are able to yield a good amount of lipid. Thus, when 0.3% glucose and 0.8% xylose were used in the mixture with standard feed, the lipid yield became 33.78%. On the other hand, 97.3% of the glucose and 93.8% of the xylose were able to be successfully converted to lipid as the dynamics changed between the three substrates within only 14 days. Another study showed that the types of substrate that are usually fed

to the BSFL have a high concentration of cellulose, hemicellulose and lignin, as the animal's main diet consists of crops or plants. The BSFL do possess guts with microbiome symbioses that are able to digest the cellulose consumed. With the right enzymes available in the BSFL, the cellulose, hemicellulose and lignin can be degraded. The main challenge, however, is when a feed with a high amount of crude fiber is used as the main diet of the BSFL, such as dairy manure. More energy is required to break down the cellulose of the fiber materials, thus reducing the lipid yield for biodiesel production. Therefore, a lower fiber content, such as chicken manure, is used together with the dairy manure as co-digestion with different ratios for the BSFL. The study conducted by Rehman et al. [15] evaluated the performance of the BSFL digestion and with the data obtained was able to develop a co-digestion mixture between dairy manure and chicken manure. With the ratio between dairy manure and chicken manure being 40:60, it was found that this ratio of co-digestion resulted in the larvae with the richest nutrient content, enhancing waste conversion efficiency of the BSFL. This study has shown that the use of organic waste in co-digestion must focus on implementing the process of mixing high fiber content with less-fibrous materials and explore the mechanisms as well as the magnitudes of the effect on the BSFL to ensure biodiesel production.

#### **4. Microbial-Treated Substrates**

A study has been conducted with the substrate mixture consisting of dairy manure, chicken manure and bacteria. The use of exogenous bacteria, *Bacillus* strains, assists the BSFL gut microbiome development in more efficiently reducing the waste capacity, utilizing the nutrients of the wastes and enhancing the production in the larval biomass. The ratio of dairy manure to chicken manure was 2:3 and this resulted in the lipid yield of 47.7% and protein yield of 53.9%. However, there was a significant increase in both of the yields when bacteria were added. The lipid yield was 67.8% while the protein yield was 71.2%. This shows that the usage of treatment with microbes utilizes a higher amount of lipid and protein compared to the controlled feed that contains only dairy manure and chicken manure. With the help of cellulose-degrading bacteria, a higher biomass promotes for a higher fat yield is promoted, as they enhance the digestion of the waste materials. Therefore, it is important for the selection of the bacteria in assisting the BSFL to ensure that the lignocellulose-rich waste is able to be managed successfully [16]. Soybean curd residue was also used as the feed of BSFL with the addition of a bacteria, *Lactobacillus buchneri.* The results shown by Somroo et al. [17] indicated that the lipid yields differed between when the feeds were only soybean curd residue (26.1%) or artificial feed (24.3%) and when bacteria were added to the feed, which resulted in an increase of the lipid yield (up to 30%). This gave a similar result for the protein yield as the insect–bacteria symbiosis increased the protein yield from 52.9% (soybean curd residue only) and 50.4% (artificial feed) to a much higher value of 55.3%. With this positive interaction, there is great benefit in the availability of the nutrients, playing a major role in the growth of BSFL, the development of the BSFL gut microbiota and the BSFL's production for digestive enzymes. This also shows that the use of symbiotic bacteria allows the success of the BSFL to adapt to new environments and new food sources while still being able to obtain positive growth and reproduction. When the treated rice straw with 39.7 g of glucose and 25.9 g of xylose underwent a fermentation process with *Saccharomyces cerevisiae,* the residues were mixed with enzymes containing hydrolyzed residues, such as lignocellulose, proteins and reducing sugars. The residue was then fed to the BSFL which underwent lipid extraction to yield a total of 5.2 g of lipid from 200 6-day-old BSFL. Additionally, 4.3 g of biodiesel was able to be produced from 200 BSFL. This shows that the nutritional source from BSFL diets consisting of lignocellulosic biomass can be another potential in lipid as well as biodiesel production. Having similar qualities as plant-based biodiesel, BSFL-based biodiesel is proven to be another alternative source of renewable energy [13]. Restaurant waste is heavily concentrated with lipids and protein. However, this substrate lacks the lignocellulose that the rice straw does not lack. If rice straws are used alone as the feed for the BSFL, the growth will be stilled because of the absence of nutrition. Therefore, using the ratio of restaurant waste to rice straw of 7:3 [18], a mixture was made. Rid-X contains natural bacteria that has the main

function of breaking down the cellulase, lipase, protease and amylase of the rice straw as well as the solid waste of the restaurant waste. This helps and increases the efficiency of the conversion for BSFL, degrading the cellulose and hemicellulose much faster. More nutrition from the both of the substrates is available for consumption, and the digestion of the food is aided by the microbes. A total of 43.8 g of biodiesel was able to be produced from 2000 BSFL. The properties of the biodiesel were also investigated, and it was found that the fatty acids of the biodiesel were similar to rapeseed-based biodiesel. Thus, it is shown that the quality of the biodiesel, despite originating from different sources, can still hold a high quality in terms of performances.

From these results, although BSFL contains the microbes that can hydrolyze the cellulose content of the feed, the amount of the microbes in the gut may not be sufficient to digest a much larger amount of feed. Research must continue to test various types of microorganisms in undergoing treatment with the feed of BSFL that contains high amount of fiber. This is to observe the conversion efficiency of the bacteria to obtain a high quality fuel for biodiesel production.

#### **5. Substrate Moisture Contents**

According to Barry et al. [19], the conversion of waste to biomass of BSFL can be achieved if the consumption of food waste is given a particular care and attention in ensuring its efficiency. Therefore, different parameters need to be investigated, preponderantly the moisture content of the larval feeding substrate, in pushing towards a successful bioconversion. It was found that when study was carried out using almond hull as the main medium for BSFL, alteration of the moisture content in the hull could directly impact the growth of BSFL [20]. The results showed positive effects on the dry weight (0.013 to 0.46 mg/larvae) as well as the yield of harvested larvae (3.7 <sup>×</sup> <sup>10</sup>−<sup>4</sup> to 0.11) as the moisture content was increased from 480 to 680 g kg<sup>−</sup>1. However, it was found that the larval consumption of hull decreased (from 15% to 13%) with increasing of the moisture content. Other studies that reported the effects of manipulating the moisture content of substrates on larval development presented opposite results. The BSFL had been found to grow bigger in terms of weight and needed less medium for consumption as the moisture content was increased [21]. The larval growth rate was also greatly affected when moisture content was reduced, as reported by Cheng et al. [22]. Using almond hull, as reported by Palma et al. [20], showed much different results, perhaps due to the decrease of diffusion of oxygen into the medium. Oxygen diffusion was limited when the pores of the hull were not air-filled and were blocked by the moisture. This directly impacted the growth of bacterial activity and would disrupt the synergistic potential of microorganisms that contributed in the conversion of hull to larval biomass. Therefore, more study needs to be conducted to observe the trends that affect the larval growth and substrate consumption resulting from manipulation of the moisture content, as the role of microorganisms plays a significant role in bioconversion of insect biomass.

#### **6. Substrate Aeration**

Managing the substrate aeration while growing the BSFL can improve the overall larval growth through engineering to acquire the right and suitable environment for medium digestion by larvae. Significant effects when manipulating the aeration content towards the development of larvae had been determined by experiment by Palma [20], using waste from almond as the BSFL feeding medium. It was found that increasing the aeration rate gave rise to a positive fit to the nonlinear regression of the BSFL weight and yield. Accordingly, the maximum larval weight and yield could be achieved at 95% at aeration of 0.57 and 0.05 mL/min g dry weight, respectively; this also contributed greatly towards the consumption of almond hull. It could be deduced that the aeration had a direct impact towards the growth of larvae, and the bed depth of the substrate may play a major role as well. This was because the anaerobic condition occurred from the oxygen utilization by both BSFL and microorganisms surpassing the oxygen being supplied from diffusion from the bed surface. This caused the larvae to migrate elsewhere to obtain nutrients that otherwise could be obtained when the larvae migrated to a deeper depth of substrate. When aeration rate was dropped, the larval consumption of hull was negatively

affected as well, showing that the presence of microorganisms could heavily impact the environment for rearing BSFL. Therefore, the oxygen content supplied by aeration should be considered of the great importance in insuring that the growth of BSFL does not negatively affect the rate of waste conversion. Moreover, the calcium content was also investigated, and it was found that increasing oxygen content during larval growth would generally increase the larval calcium content. The consumption of hull by BSFL impacted the uptake of calcium from the almond hull and later affected the larval biomass compositions. According to Liu et al. [23], calcium as the mineral element of BSFL was needed for the cuticle formation. Another study, conducted by Wong et al. [24], showed that the harvesting of BSFL at different calcium or chitin levels could directly affect the lipid content since the accumulated body fat tissues were needed during metamorphosis.

#### **7. Inactivation Methods**

Various ways that the BSFL could be inactivated were reported by Larouche et al. [25]. Grinding was the first method of mechanical disruption in which the larvae were homogenized at 15,000 rpm in their study. The larvae were packed under 95% vacuum on high hydrostatic pressure with 600 MPa of pressure treatment. The next method involved heating the BSFL, i.e., via blanching, where the larvae were immersed in boiling water for 40 s. Desiccation was another method of heating the BSFL. This approach required the larvae to be located in the air oven with the temperature set to 60 ◦C for 30 min. The other type of larval inactivation method was freezing, where the larvae were either frozen at the temperature of −20 ◦C or −40 ◦C for one hour. Freezing the larvae also could be completed by using liquid nitrogen in a vacuum package for 40 s. Finally, asphyxiation was the last inactivation method reported by Larouche et al. [25], in which the larvae were initially vacuum packaged and subsequently stored at the temperature of 27 ◦C for 120 h. The atmosphere was modified either to contain 100% carbon dioxide or nitrogen gas. The larvae were then stored at the temperature of 27 ◦C for 120 or 144 h. The ether extract for each of the larval samples was conducted using petroleum ether as extraction solvent. It was found that the larval lipid contents were higher when the inactivation method of asphyxiation (CO2 = 15.9%; N2 = 16.6%; vacuum = 15.9%) was used as opposed to heating (desiccation = 13.4%; blanching = 14.5%), freezing (−20 ◦C = 12.8%; −40 ◦C = 12.4%; liquid nitrogen = 12.6%) or mechanical disruption (grinding = 11.9%; high hydrostatic pressure = 12.0%).

The inactivation method of BSFL was lately studied even further, as it was important to be explored to ensure the investigation of the composition of larval lipid could be exploited for biodiesel production and also support a higher value of uses. The characteristics of larval lipid distributions during the processing and storage of BSFL biomass must be unveiled since there is currently a lack of this information. Caligiani et al. [26] directly related the inactivation method of BSFL with extracted lipid characteristics. The inactivation methods reported were blanching and freezing. In their study, half of the first sample was provided in a frozen condition and stored at the temperature of −20 ◦C, while the other half was ground and freeze-dried until it reached residual moisture of 10% before being kept at the temperature of −20 ◦C. The next set of larval samples were obtained alive. The larvae were killed by blanching the prepupae in hot water at the temperature of 100 ◦C for 40 s before storing at the temperature of −20 ◦C prior to the lipid analyses. Another live larval sample was stored directly at the temperature of −20 ◦C until use. During the extraction, the Soxhlet lipid extractor that either used diethyl ether or petroleum ether as the solvent was compared with the use of chloromethane solvent. The results of the first sample inactivated by using freezing before the arrival showed that there was no significant difference of using the different extraction solvents via the Soxhlet method. The next set of samples were obtained alive, and diethyl ether was used, resulting in the lipid yield was 13.0 ± 1% when inactivated by freezing; while using blanching, the lipid yield was 13.3 ± 0.8% from BSFL biomass. This also proved that the BSFL was a good source of lipid, unlike other type of insects. However, as compared with the method employing chloromethane, the larval lipid extracted was slightly lower (9.11%). In the live larval sample that was also inactivated via freezing, it was found that both of these samples had a high free fatty acids content. This also explained that the low

lipid yield when using the chloromethane for extraction was due to the loss of fatty acid salts in the aqueous phase. However, when the live BSFL were blanched before being frozen, the loss of fatty acid was negligible. When the freezing method was applied towards the BSFL, the amount of acyl glycerols was drastically reduced, most likely due to the activation of the lipase, releasing the free fatty acids. However, the free fatty acids were not used for biodiesel production as a reaction with acyl glycerols was needed during the transmethylation process. When the BSFL was inactivated by blanching, a thermal pretreatment method, the lipid fraction was observed to be stable as it was mainly composed of triacylglycerols. This may be due to the thermal environment deactivating the lipase activity in the BSFL, as it did not damage or influence the lipid fraction conspicuously, preserving it for transmethylation process in producing biodiesel.

#### **8. BSFL-Based Biodiesel**

Black soldier fly larval biomass has become an attractive candidate as a renewable source of energy due to its high lipid content. Transesterification is a process for biodiesel production from larval biomass in which the extracted lipid will react with alcohol. It has become essential to ensure that the lipid conversion during biodiesel production is at its highest efficiency. Surendra et al. [27] carried out an investigation to determine the fatty acid compositions of BSFL fats or lipids for biodiesel production. It was found that the BSFL had a very high amount of lauric acid (44.9% ± 1.5%) as compared with the crop-based biodiesel such as soybean oil (negligible) and palm oil (0.1%), a trait that was significant in terms of biodiesel production. The saturated fat was found to be 67% of total fatty acid while soybean was known to only have 11% and palm oil to have only 37% of total fatty acids. On the other hand, the BSFL had a proportion of 28% fatty acids being of unsaturated fat, lesser than that of soybean (85%) and palm oil (55%). The quality of the biodiesel was known to be greatly affected by the composition of fatty acids in a substrate. In this case, the BSFL-based biodiesel was shown to have a significant amount of saturated fatty acids and a low concentration of unsaturated fatty acids, making it an ideal substrate for a high quality of biodiesel production. Thus, the biodiesel would have a viscosity with much lower value and a more stabilized property in terms of its oxidative state. Additionally, the process of transesterification of larval oil that has been extracted must be efficiently processed in ensuring the biodiesel production is of the highest quality.

#### **9. Transesterification of Larval Lipids**

An optimum condition for executing the transesterification of BSFL lipids was investigated by Li et al. [28]. The harvested larval biomasses were initially fed with three different substrates individually, namely, cattle manure, pig manure and chicken manure. There is a two-step process during the conversion of larval lipids into fuel. The first step was the acid-catalyzed esterification of fatty acids. This step was to decrease the amount of acids in the BSFL lipids that were extracted and that acted as the pretreatment for the conversion process. The next step was the typical alkaline-catalyzed transesterification. One of the reaction conditions that was optimized was the esterification temperature for 1 h of reaction time using the methanol to larval lipid ratio of 8:1. It was found that as the temperature was increased from 55 to 85 ◦C, the conversion of fatty acids in the crude lipids to biodiesel increased from 73% to 92%. This demonstrated a positive relationship between temperature and conversion of fatty acids, as it could be directly related to the efficiency of mass transferred with increasing temperature that caused the crude lipid to be more soluble. Accordingly, the temperature of 75 ◦C was found to be optimal temperature for the esterification process. Another reaction condition investigated was the molar ratio of methanol to larval crude lipid. It was found that the maximum conversion of fatty acids was achieved at 90% when the optimum molar ratio used was 8:1. When a much lower ratio was used instead, the conversion was found to be incomplete. The reaction time was another reaction condition investigated, and the converted fatty acids were found to be 73% to 90% at the reaction times of 30 and 60 min, respectively. The biodiesels produced from the crude lipids of BSFL fed with chicken manure, pig manure and cattle manure were 91.4%, 57.8% and 35.6%, respectively, through the optimum transesterification condition of 30 min at 65 ◦C with molar ratio of methanol to lipid at 6:1 while using 0.8% NaOH as the catalyst of the reaction. The biodiesel was tested, and it contained a high percentage of saturated fatty acids at 67.6%. This meant the biodiesel produced would have a high oxidative stability value, an excellent trait for biodiesel storage. The optimal transesterification conditions were further tested and used during the experimentation with BSFL-based biodiesel derived from waste grease of restaurants [29]. The two-step process which consisted of the acid-catalyzed esterification and alkaline-catalyzed transesterification was carried out using 1% H2SO4 as the reaction catalyst with reaction temperature set at 75 ◦C, molar ratio of methanol to lipid at 8:1 and 1 h of reaction time. For the alkaline-catalyzed transesterification, the methanol-to-lipid ratio was kept at 6:1 with 0.8% NaOH as the reaction catalyst. The biodiesel produced was 23.6 g from 1000 g of solid residual fraction of restaurant waste fed to 1000 BSFL. The conversion rate of free fatty acids attained was 91.9%, with the total yield of biodiesel of 2.4%. Li et al. [10] also investigated the conversion of BSFL fat to biodiesel using the dairy manure as the main larval substrate. After conducting the two-step transesterification process, 15.8 g of biodiesel was able to be produced from 1.2 kg of dairy manure. The larval biodiesel also contained 58.2% saturated fatty acids and 39.8% of unsaturated fatty acids with overall quality satisfying the EN 14214 standard.

#### **10. Co-Solvent for Transesterification of Larval Lipids**

The conventional way of biodiesel production has generally presented some problems, such as consuming a high amount of energy, that make the process costly. Therefore, a direct transesterification involving fewer steps was suggested and investigated by Nguyen et al. [30]. Methanol was used in prior studies as both the solvent for lipid extraction and reactant for lipid transesterification. However, an excess amount of methanol could weaken the function of catalyst, reducing the yield of biodiesel. In the study by Nguyen et al. [30], co-solvents such as n-hexane, n-pentane, chloroform, acetone and petroleum ether were individually mixed with methanol during the direct transesterification process. With the capability to dissolve long chain triglycerides, these co-solvents showed high potential efficiency in extracting the larval lipid, yielding higher amounts of biodiesel. This would prevent lipid loss during the process as less solvent and energy were consumed. When the solvents were mixed with the methanol at the volume ratio of 1.17:1, a high yield of biodiesel was observed. The use of the mentioned solvent should bring positive effects, as the co-solvents are generally capable in dissolving the lipid and short-chain alcohol used as the homogenous catalyst. It was found that by using n-hexane as the co-solvent, the highest yield of biodiesel (63.37%) could be obtained as compared with acetone (54.83%), chloroform (48.50%) and petroleum ether (35.67%). The effects of volume ratios between the n-hexane and methanol were also investigated and it was observed that biodiesel could be yielded at a higher amount when a much lower volume ratio was used. This was probably because the high methanol content in the reaction would lead to a higher molar ratio between the methanol and lipid, leading to the higher reaction yield. Thus, the optimum volume ratio of n-hexane to methanol was determined to be 1:2. The methanol to biomass ratio of the reaction would increase, promoting the conversion yield. However, the frequency of collision between the lipid and methanol would decrease if excess solvent was used in the reaction. This would, overall, result in the increase of heat and mass transfer resistance, decreasing the conversion yield. Other than that, in a similar study with the presence of free fatty acids from the BSFL fat, the catalyst selected was sulfuric acid. As the catalyst loading was increased from 0.4 to 1.2 mL, the yield of biodiesel had also increased from 48.93% to 65.87%, respectively. The polymerization of unsaturated fatty acids was activated with the presence of excess sulfuric acid at high reaction temperature and long reaction time. The increase in reaction temperature would result in the increase of biodiesel yield (87.67% at temperature of 130 ◦C). The optimal temperature was found to be 120 ◦C as, although the extraction efficiency increased with enhancement of reaction rate, there was no significant difference between 120 and 130 ◦C. The biodiesel yields also increased with the reaction time and managed to reach the equilibrium state at reaction time longer than 90 min. Therefore, the highest biodiesel yield was able to be obtained at 94.14% with

the reaction temperature of 120 ◦C using n-hexane:methanol volume ratio of 1:2, the solvent at 12 mL, the catalyst loaded at 1.2 mL and reaction time of 90 min. The biodiesel yield was also observed to be increased from 4.73% with no co-solvent to 63.37% when n-hexane was used. Therefore, n-hexane was proven to be the suitable co-solvent for the reaction involving larval lipid transesterification.

#### **11. Biological Lipase Catalyst for Transesterification of Larval Lipids**

The two-step process of transesterification involves acid-catalyzed esterification and alkali-catalyzed transesterification. However, some complications have been presented when sulfuric acid and sodium hydroxide were used in the process, such as damaging the equipment and complications in removing the dissolved catalysts. Therefore, in the study by Nguyen et al. [31] for biodiesel production, methanol catalyzed with lipase was used during the transesterification of BSFL lipid. In the biodiesel production, several lipases were tested such as Novozym 435, Lipozyme TL-IM, Lipozyme RM-IM and lipase PS. Transesterification was then carried out using 10% lipase mixed with methanol and lipid at the molar ratio of 3:1 and reaction temperature and time of 20 ◦C and 4 h, respectively. According to the results, the Novozym 435 presented a biodiesel yield of 56.78% as opposed to Lipozyme RM-IM (42.18%), porcine pancreas lipase (23.46%), *Candida rugosa* lipase (22.82%), *Rhizopus oryzae* lipase (21.56%), amano lipase G (13.85%) and amano lipase PS (12.56%). This showed that Novozym 435 had the highest catalytic conversion property and also could be used repeatedly. With the enzyme loading at 20%, the maximum yield for biodiesel could be achieved with the methanol:lipid molar ratio of 6.33. Lipase deactivation, however, may occur when the methanol level exceeded the required amount and would cause a reversal in the aforementioned trend. The yield of the reaction would then decrease. In terms of the reaction temperature, the highest yield could be achieved at 40 ◦C, as higher temperatures would deactivate the activities of enzymes. Enzyme loading played another part in the reaction, as low loading would affect the temperature and later the biodiesel yield. On the other hand, at high level of enzyme loading the temperature did not bring any significant effect towards the yield. Therefore, any problem rising from the temperature could be overcome by increasing the concentration of enzyme in the reaction. In conclusion, to obtain a maximum yield of biodiesel of 97.65%, the molar ratio of methanol to lipid was set at 6.33:1, with 20% enzyme loading and 26 ◦C as the reaction temperature for 9.48 h. This experiment had shown positive results, and this should be an encouragement in using the green enzyme-catalyzed process to produce biodiesel.

However, a high level of methanol and ethanol would still cause the deactivation of lipase functions. This was because the absorption of alcohol on the surface immobilized the lipase. Therefore, to overcome this problem, Nguyen et al. [32] investigated the effects of methyl acetate, which is an acyl acceptor that would increase the rate of reaction during transesterification. It was observed that using a high amount of methyl acetate had no effects on the stability as well as the activity of enzymes in replacing the alcohol. To obtain a high amount of biodiesel yield using lipase-catalyzed transesterification (Novozym 435 chosen as catalyst) using methyl acetate as the acyl acceptor for BSFL in the production of biodiesel, several reaction conditions were tested and observed for their effects. The reaction conditions observed were the molar ratio of methyl acetate to lipid and enzyme loading. Biodiesel yield decreased with low ratio, as this increased the loading of enzyme in the reaction. This was because it caused the polymer beads to aggregate with the immobilized enzymes, and this disrupted the mass transfer, enabling the enzyme to react with the oil–water interface flexibly. The conversion yield would be lowered as a result. However, at a higher molar ratio, the enzyme loading would increase, thus increasing the biodiesel yield. The temperature of the reaction was also observed, and it was found that that there was no significant increase of biodiesel yield with the increase in temperature. The optimal ratio of methyl acetate to lipid on the other hand was found to be 12:1. In between the enzyme loading and temperature, the yield would increase as the temperature was increased with any amount of loaded catalyst. However, the deactivation of enzyme occurred at high temperature. Therefore, for Novozym 435, biodiesel yield was produced at the highest amount at 39.5 ◦C reaction temperature. Concisely, in order to obtain a maximum yield of biodiesel, 12 h

of reaction time, 14.64:1 molar ratio of methyl acetate to lipid, 17.58% enzyme loaded and 39.50 ◦C reaction temperature should be used in a lipase-catalyzed transesterification reaction using Novozym 435 as the catalyst. With the proven high biodiesel yield using the optimized conditions mentioned, BSFL lipid has become a reliable source of energy and can be further developed in future [33,34].

#### **12. Conclusions**

In short, this review has demonstrated that the BSFL biomass can be the source of protein and lipid for energy. The rearing conditions of BSFL can be systematically optimized to allow the accumulation of more larval protein and lipid in the fat body. Upon the harvesting, the lipid in the form of larval fat can be extracted and transesterified for producing a mixture of fatty acid methyl esters of biodiesel. In this regard, the extraction and transesterification processes can be optimized as well to maximize the BSFL-based biodiesel production. The residual BSFL biomass after the lipid extraction is a protein-rich larval biomass and can be served as the feedstock for animal feed production. In determining the optimum various larval processing conditions, the preparation of substrates, rearing of BSFL and eventually biodiesel production are vital in ensuring the maximum yield from BSFL biomass. The mixture of substrate added with 6% of xylose prompted 34.60% lipid content from the BSFL biomass. Microbial-treated substrate such as dairy manure and chicken manure mixture inoculated with *Bacillus* strains would yield 67.8% lipid and 71.2% protein. Addition of Rid-X to a mixture of restaurant wastes and rice straw could get 43.8 g of biodiesel from 2000 BSFL. To ensure the optimum conditions during rearing period, the moisture content of the substrate should be in the range of 480 to 680 g/kg. The increase in moisture content would result in the decrease in feed consumption. The aeration of the substrate should also be at 95% to achieve maximum larval weight and yield of dry weight as well as a positive growth of the larvae. To inactivate the mature BSFL for maximum yield, it has been shown that if petroleum ether was used as the extraction solvent during lipid extraction, asphyxiation would result in higher lipid content. If diethyl ether was used in the Soxhlet method, both blanching and freezing inactivation methods could be employed. The optimum transesterification conditions were determined to be at 30 min at 65 ◦C with molar ratio of methanol to lipid at 6:1 while using 0.8% NaOH as the catalyst in the reaction. During direct transesterification, hexane was recommended as the co-solvent, and sulfuric acid as the catalyst. The reaction temperature should be at 120 ◦C using hexane:methanol volume ratio of 1:2, the solvent at 12 mL, the catalyst loaded at 1.2 mL and reaction time of 90 min. Novozym 435 can be added in a lipase-catalyzed transesterification reaction with methyl acetate added to replace the methanol and ethanol. Studies have shown that with 12 h of reaction time, 14.64:1 of molar ratio between methyl acetate to lipid, 17.58% enzyme loaded and 39.50 ◦C of reaction temperature could ensure obtaining the maximum yield of biodiesel. Therefore, since the detailed laboratory-proven information with regard to the upstream and downstream of BSFL biomass production is currently accessible, the mass production of this feedstock at industrial scale should be assessed to unveil its feasibility concerning the cost and long-term environmental sustainability. Lastly, the authors of this review believe that the BSFL biomass could potentially arise as the new and unconventional feedstock for protein in replacing the traditional fishmeal if it is not used for biodiesel.

**Author Contributions:** Conceptualization, S.H. and J.W.L.; resources, K.K. and W.K.; writing—original draft preparation, S.H. and C.Y.W.; writing—review and editing, M.K.L. and P.L.S.; visualization, W.D.O.; supervision, T.L.C. and C.S.K.; funding acquisition, J.W.L., C.S.K. and K.K. All authors have read and agreed to the published version of the manuscript.

**Funding:** Financial supports from Yayasan Universiti Teknologi PETRONAS via YUTP-FRG with the cost center 015LC0-282 and Ministry of Education Malaysia under HICoE with the cost center of 015MA0-052 are gratefully acknowledged.

**Acknowledgments:** One of the authors, Kunlanan Kiatkittipong, wishes to thank the financial support received from the King Mongkut's Institute of Technology Ladkrabang, KMITL with the Grant no. KREF046209.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Review* **Liquid Biphasic System: A Recent Bioseparation Technology**

**Kuan Shiong Khoo 1, Hui Yi Leong 2, Kit Wayne Chew 3, Jun-Wei Lim 4, Tau Chuan Ling 5, Pau Loke Show 1,\* and Hong-Wei Yen 6,7,\***


Received: 16 December 2019; Accepted: 17 January 2020; Published: 23 January 2020

**Abstract:** A well-known bioseparation technique namely liquid biphasic system (LBS) has attracted many researchers' interest for being an alternative bioseparation technology for various kinds of biomolecules. The present review begins with an in-depth discussion on the fundamental principle of LBS and this is followed by the discussion on further development of various phase-forming components in LBS. Additionally, the implementation of various advance technologies to the LBS that is beneficial towards the efficiency of LBS for the extraction, separation, and purification of biomolecules was discussed. The key parameters affecting the LBS were presented and evaluated. Moreover, future prospect and challenges were highlighted to be a useful guide for future development of LBS. The efforts presented in this review will provide an insight for future researches in liquid-liquid separation techniques.

**Keywords:** liquid biphasic system; aqueous two-phase system; aqueous biphasic system; purification; separation; recovery; biomolecules

#### **1. Introduction**

The most trending research in the downstream biotechnology industries focuses on the production of various bio-based products from renewable sources. Examples of these sources are microalgae, fruit, lignocellulose biomass, secondary product, and crop waste. Separation and purification techniques for the recovery of biomolecules (e.g., proteins, carotenoids, and lipids) requires a precise operating condition to ensure high value end-products can be obtained [1,2]. Established extraction techniques such as membrane separation, chromatography-based method, ultrafiltration, and precipitation usually involve multiple step operations, complex pathways, time consuming operations, high energy inputs, and high cost for the recovery and extraction processes [3–5]. With that said, researchers are putting tremendous efforts in developing a new separation and purification techniques which can be performed in a one-step extraction process within a shorter period of time. On the other hand, the extraction solvents in a process that can be reused and recycled will lower the overall processing cost [6]. As for

food and pharmaceutical applications, this requires an alternative non-toxic and environmentally friendly extraction solvents [7].

A well-established bioseparation technology namely liquid biphasic system (LBS) has attracted numerous researchers' attentions in the separation and purification of biomolecules [8]. It is also known as liquid-liquid extraction technology in the downstream processing. The concerns associated from the conventional extraction method has been overcome by using liquid biphasic separation techniques. The liquid biphasic extraction technology is comprised of two liquids which is separated by an interfacial layer when the mixture of two incompatible liquids is beyond the critical condition. Generally, the characteristics of the phase-forming components creates a physico-chemical interaction which can easily acclimatize the target biomolecules to be partitioned to either the top or bottom phase depending on the selectivity of the components. Furthermore, various assisted technologies such as bubbling, ultrasound, and electrolysis have been incorporated into the LBS to enhance the effectiveness of biomolecules separation [9–12]. The application of the LBS has been applied for the extraction, separation, and purification of proteins, lipids, and carotenoids from microalgae [2,13].

This review article strives to summarize the cognitive knowledge and previous experimental research dealing with LBS for extraction and purification of various biomolecules. This review begins with the principles and fundamentals of LBS, followed by the various type of biphasic systems were presented. Recent works related with advance technologies such as bubble-, ultrasound, and electricity-assisted LBS were evaluated and assessed. Additional information regards to the quantification (i.e., partition coefficient, selectivity, separation efficiency, and recovery yield) and composition of LBS were tabulated with provided references in Sections 2 and 3. Each section in this review would allow the readers to understand the development of LBS technologies. Moreover, the key parameters affecting the extraction efficiency in LBS, advantages and drawbacks of LBS were comprehensively discussed. In addition, future prospect and challenges associated with LBS were also discussed. This review article has a significant impact on the liquid-liquid extraction and purification for various biotechnological products, which serve as a resourceful tool for researchers dealing with extraction of biomolecules using LBS.

#### **2. Liquid Biphasic System**

Liquid biphasic system (LBS) or commonly known as aqueous two-phase system (ATPS) has been long introduced for the separation, recovery, and purification of biomolecules, and it is the current research trend adopted in the separation and purification technology. It was started back in 1896 when Martinus Willem Beijerinck accidentally mixed an aqueous starch solution with gelatin and found that an immersible layer was formed between both the aqueous solutions [14,15]. This idea of LBS as an analytical separation technique was sparked by Per-Åke Albertsson in the 1960s who discovered the phenomenon by mixing two different polymers (e.g., polyethylene glycol and dextran) resulting in an aqueous medium containing two separable phases [14,16,17]. This application was then extended to several generations of scientists and engineers who have been working in the industrial biotechnology field. Figure 1 shows a schematic diagram of the principles of LBS.

The LBS is well-known for the extraction of different biotechnological materials such as proteins, lipids, and carotenoids [1,18,19]. The specialty of LBS compared to traditional organic solvent extraction techniques is the composition of the phase-forming components which contains large amount of water while maintaining a low interfacial layer that separates both phases. It can be either used to separate proteins from cellular debris or to purify targeted proteins from contaminated proteins. Likewise, LBS has the capability of directing the target biomolecules by partitioning them to the top phase for extraction [20]. Conventional polymer-based LBS which possess a low ionic system is generally used for the separation and purification of biomolecules which are sensitive toward ionic condition [16]. Nevertheless, polymer-based LBS was neglected due to lack of compatibility between high ionic strength biomolecules, expensive phase-forming components, and its high viscosity system. Further development in LBS using different phase-forming components such as alcohol-,

ionic liquids-, deep-eutectic solvent- and surfactant-based was utilized to replace the conventional polymer-based LBS.

**Figure 1.** Schematic diagram of the principle in the liquid biphasic system (LBS).

The selective partitioning of the LBS allows the extraction of biomolecules to be operated in a single-step process compared to traditional extraction techniques which require multiple operation steps. LBS possess an environmental-friendly, inexpensive, ease of scaling-up, rapid, and efficient techniques for recovery and purification of biomolecules. During the planning stage, it is crucial to understand the complexity of the physical and chemical interaction reaction throughout the partitioning process in the LBS [21]. The selection of various parameters which are compatible to the system properties are important to achieve an optimal extraction, recovery, and purification condition. It is also important to evaluate the interactions during the selection of various parameters (e.g., salt precipitation, crystallization, and absence of biphasic system) as it may affect the findings. Lastly, is to assess the effect of each process parameters on the product recovery and purity [21].

Fundamental principles for the formation of LBS requires a phase diagram or also called the binodal curve where these provide a set of information regarding the two-phase formation and their required concentration in the top and bottom phases [22]. A detailed study has been evaluated previously by Iqbal et al., (2016) on the tie line length (TLL) and slope tie line (STL) for the construction of phase diagrams [23]. Binodal curves can be constructed using three methods namely, turbidometric titration, cloud point, and node determination method for predetermined phase diagram [22,24,25]. Moreover, the partition coefficient (K) LBS is to evaluate the equilibrium relationship between the top and bottom phase in the LBS. However, there is still lack of studies reporting on the theory and chemistry of these phase forming mixtures in the LBS which is a gap to-be-filled. Apart from that, factors that affect the partition coefficient can be manipulated using electrical, hydrophobicity-phase forming components, bio-specific affinity, molecular size, and surface area to understand the physico-chemical properties of the partitioning mechanism in the LBS.

#### *2.1. Polymer-Based LBS*

The conventional polymer-based LBS is typically made up of two polymers (e.g., polyethylene glycol (PEG) and dextran) and PEG-salt combinations (e.g., phosphate-, sulphate-, and citrate-based) as the phase-forming components. The purpose of using polymer-based LBS is that the chemical composition of a non-ionic characteristics toward an ionic environment is compatible towards biomolecules having low ionic strength [16]. Aside from that, the phase forming component from polymer-based has the ability to be recycled and reused for subsequent extraction process and this reduces the cost of polymers phase-forming component [26]. Polymer-based LBS are commonly used for protein extraction due to its poor hydrophilic and hydrophobic interaction in polymer/salt-based

LBS. However, it is important to maintain concentration of salt solution as high salt concentration may denature and damage the fragile protein in the system.

In most work, conventional polymer-based LBS has been replaced by using thermo-separating polymers as the phase-forming component to overwhelm the limitation of polymer-based LBS such as high viscosity and difficulties in recycling process [27,28]. Thermo-separating polymers are random, di-block, and tri-block co-polymers of ethylene oxide (EO) and propylene oxide (PO) [29]. Thermo-separating polymers have a low cloud point temperature (≤47 ◦C) which is suitable to achieve temperature-induced phase separation where a target protein can be recovered from the polymer [30]. Generally, a back-extraction process such as ultrafiltration, diafiltration, and crystallization is needed to separate the target protein from the polymer. However, an in-depth understanding on the mechanism by the polymer phase-forming component for the recovery of biomolecules is still poorly understood. This shows a gap for future researchers to further explore the fundamental principles of this LBS extraction technique.

Several studies have been conducted involving cyclodextrin glycosyltransferase (CGTase) from *Bacillus cereus*. Ng et al., (2012) reported that the TLL of 41.2% (*w*/*w*), volume ratio (VR) of 1.25, pH 7, and crude loading (*w*/*w*) of 20% were the optimal conditions to recover cyclodextrins using polymer-based LBS with ethylene oxide–propylene oxide (EOPO) 3900 and two phosphate salts [31]. This experiment showed that the highest CGTase was purified up to 13.1-fold with a yield of 87% recovered in the EOPO-rich top phase. However, this experiment did not discuss the time period in cyclodextrins recovery. Another research carried out by Lin et al. [32] with modified method using flotation technique and the combination of PEG 8000 and potassium phosphate salt. The optimum conditions in cyclodextrins (CDs) recovery was optimized at 18% (*w*/*w*) PEG 8000 and 7.0% (*w*/*w*) potassium phosphate with TLL of 27.2% (*w*/*w*), VR of 3.0, pH 7, and crude loading (*w*/*w*) of 20%. The experiment showed that the recovery of CDs was affected by alternating each of the parameters such as TLL, VR, and pH where the purification factor (PFT), which corresponded to the highest CGTase purity up to 21.8 with a yield of 97.1%, was recovered in the PEG-rich top phase within a short period [32].

A similar approach utilizing polymer-based LBS was employed for the recovery of lignin peroxidase from *Amauroderma rugosum* (Blume and T. Nees) [33]. However, this experiment used a lower molecular weight (PEG 600) for a high purification of lignin peroxidase. Generally, this approach showed that a higher molecular weight polymer reduces the purification factor of lignin peroxidase due to the interaction of PEG and hydrophobic enzyme. An optimal condition in lignin peroxidase recovery was optimized at 15% (*w*/*w*) PEG 600 and 16% (*w*/*w*) dipotassium phosphate with highest purification factor of 1.33 ± 0.62 and recovery yield of 72.18 ± 8.50%.

#### *2.2. Organic Solvent-Based LBS*

Organic solvent-based LBS consists of various water-miscible alcohols (e.g., methanol, ethanol, 1-propanol, and 2-propanol) and inorganic salts. This form of LBS has been utilized to overcome the limitation of polymer-based LBS to improve the recovery of biomolecules from the phase-forming component [7]. The use of alcohol as the phase-forming components can easily recover the biomolecules by evaporating the alcohol from the top phase. A recent study also showed a greener approach using food grade alcohol such as ethanol and 2-propanol compared to the conventional polymer-based LBS for the extraction and recovery of carotenoids from microalgae [7]. Additionally, the phase-forming component can reduce the cost of the process by recycling and reusing the alcohol using rotary evaporator for the next extraction process. Despite its advantages, the drawbacks of using alcohol, especially methanol, as the phase-forming component is the toxicity and hazardous effects towards the environment.

Ooi et al. (2009) reported a study on purification of lipase from *Burkholderia pseudomallei* using alcohol/salt-based LBS [19]. The best lipase recovery was achieved in LBS composed of 16% (*w*/*w*) of 2-propanol, 16% (*w*/*w*) of potassium phosphate and 4.5% (*w*/*v*) sodium chloride with a purification factor of 13.5 along with the yield of 99%. The presence of alcohol component in LBS also did not inhibit the enzymatic activity of purified lipase. The effect of NaCl on lipase partitioning was found to generate an electrical potential difference in the LBS [34]. An increase in the salt concentration could generate an electrostatic potential that strongly expelled the negatively charge biomolecules toward the water-miscible alcohol in top phase, thus resulting in a high recovery yield.

Lin et al., (2013) conducted a study using alcohol/salt-based LBS to recover the intracellular human recombinant interferon-α2b (IFN-α2b) from *Escherichia coli* [34]. A different variety of combinations between alcohol-based top phase (ethanol, 1-propanol and 2-propanol) and salt phase (ammonium sulfate, dipotassium hydrogen phosphate, and monosodium citrate) were conducted. LBS composed of 18% (*w*/*w*) of propanol and 22% (*w*/*w*) ammonium sulfate in 1% (*w*/*w*) sodium chloride was reported to be the optimal conditions for the purification of IFN-α2b achieving a purification factor of 16.2 with the yield of 74.6%. Ammonium sulfate salt was selected due to its high level of pH in the system which provided a high purification factor of IFN recovery. As the pH environment in LBS increased, the contaminant protein and IFN protein were partitioned toward water-miscible alcohol top phase. This is mainly due to the negatively charge protein which tends to partition to the top phase and repels from the salt-rich bottom phase [34].

A recent study conducted on a recyclability test utilizing 1-propanol and ammonium sulfate system for the phlorotannin recovery from *Padina australis* and *Sargassum binderi* [35]. The highest recovery of phlorotannin were 76.1% and 91.67% with purification factor of 2.49 and 1.59 from *Padina australis* and *Sargassum binderi,* respectively. A consistent recovery of phlorotannin was obtained after conducting two cycles of the system. This showed a feasible and eco-friendly approach of utilizing the alcohol-based LBS for biomolecules extraction.

#### *2.3. Ionic Liquid-Based LBS*

A new trend of research by using ionic liquids (ILs) have been an alternative organic compound and non-volatile green solvent in the downstream processes. Their remarkable properties such as negligible vapor pressure, low melting point and high thermal stability have received numerous attention from researchers [36,37]. ILs are composed with tuneable physico-chemical properties of cationic and anionic ions [38]. The cationic part of ILs usually consists of choline cation, ammonium cation, quaternary ammonium or phosphonium, and guanidium cation. As for the anionic part, it consists of environmentally friendly sources such as carboxylic acid, amino acid and biological buffers. Thus, replacing ILs as the phase-forming component in LBS would be beneficial for the extraction and purification of specific target biomolecules from complex crude extract [39]. Additionally, ILs have also been employed for various applications such as electrolytes (e.g., fuel cells, batteries and sensors), CO2 capture, lubricants, and fuel additives. The cost of reactant for the synthesis of ILs are expensive. Therefore, it is important for ILs to be recycle- and reuse-able to ensure that ILs-based LBS are more feasible and applicable in the bioprocessing industries for the next extraction processes. A review by Ostadjoo et al., (2017) revealed the green and environmentally friendly, 1-ethyl-3-methylimidazolium acetate ([C2mim][OAc]) for its potential features in the field of lignocellulose biomass dissolution and biopolymer processing [40–42]. Yet, there is still insufficient studies related to their toxicity and eco-friendliness on scaling up these ILs, especially imidazole- and pyridinium-based ILs. Here we recommended that these ILs need to be further fabricated by replacing environmentally-friendly anionic part such as carboxylic acid, amino acid and biological buffers in order to minimize their toxicity in various application.

Gutowski et al. (2003) reported that by mixing imidazole-based ILs and a kosmotropic salt (i.e., K3PO4) would lead to the formation of a biphasic system [43]. This research had gained interest investigating the phase separation behavior of IL-based LBS. The study on protein extraction using IL-based LBS in a single step was conducted by Du et al. (2007). The researchers had successfully extracted the protein from human urine into the IL-rich top phase with a distribution of 10 and enrichment factor of 5 [44]. Apart from that, Ng et al. (2014) investigated the purification of CGTase from *Bacillus cereus* fermentation broth in IL/salt LBS, composing of 35% (*w*/*w*) of (Emim)BF4 and

18% (*w*/*w*) of sodium carbonate with the addition of 3% (*w*/*w*) of NaCl [45]. The optimized operating conditions showed that the IL-based LBS was a promising approach for the purification and recovery of CGTase in a single step operation attaining a high purification factor of 13.86 and yield of 96.23%. Ng et al. (2014) also reported that it was crucial in the selection of salt such as citrate and carbonate ions as they played an important role in LBS formation and was able to attract water molecules toward them by forming strong intermolecular interaction [45].

Chang et al., (2018) used a series of alkyl bromide imidazole for the extraction of C-phycocyanin (CPC) from *Spirulina platensis* and found that the longer the alkyl chain, C8MIM-Br enhanced the extraction efficiency of CPC [46]. The results indicated that by using C8MIM-Br/salt LBS the maximum extraction efficiency, partition coefficient, and separation factor of CPC were 99.0%, 36.6, and 5.8 respectively. ILs-based LBS demonstrated an efficient and feasible separation technique for the extraction of various biomolecules from complex crude extract. This was supported by a recent study that evaluated the protein partitioning in ILs-based LBS composed of Iolilyte 221 PG and citrate salts was found to be feasible but complex depending on various factors such as concentration of phase-forming component, pH, temperature, ionic strength, and chemical nature of the target biomolecules [47]. Proteins are negatively charged particles therefore it favours a system pH (≥6.50) higher than the isoelectric point of protein. Moreover, the partition coefficient for tie-line length within 38–76% were reference points for specific protein (e.g., bovine serum albumin and rubisco) to be partitioned at the top phase.

#### *2.4. Deep-Eutectic-Solvent-Based LBS*

Deep-eutectic-solvents (DESs) are defined as a subclass from ILs because of their similarity in physical and chemical properties of ILs [48]. The behavior exhibited from DESs are contributed from hydrogen bonding, whereas ILs are dominated by ionic interactions [49]. DESs are more environmentally friendly as compared to ILs (e.g., imidazole- and pyridinium-based ILs) which are toxic and non-biodegradable. The synthesis of DESs is by combining hydrogen bond acceptors (e.g., quaternary ammonium and phosphonium salts) and hydrogen bond donors (e.g., alcohols, carboxylic acid, and amide). A major advantage from DESs are their charge delocalization properties which are responsible for the decrease in melting point of mixture relative to the raw material [50]. The bottleneck from using ILs such as high cost and complex synthesis route have been solved by these DESs. By having the similar characteristic as ILs and exhibiting some distinguishing features, including ease of synthesis, low cost, and valuable for industrial application, DESs have gained interest in many fields especially in LBS [51].

Choline chloride (ChCl) is a convention quaternary salt used to synthesize DESs. ChCl-based DESs have the same advantages with ILs besides showing excellent biodegradability and low toxicity [52]. Zeng et al. (2014) had performed the extraction of bovine serum albumin (BSA) using four different kind of DESs, namely, choline chloride (ChCl)-urea, tetramethylammonium chloride (TMACl)-urea, tetrapropylammonium bromide (TPMBr)-urea, and ChCl-methylurea [53]. The extraction efficiency of BSA under the optimum LBS conditions composed of 0.7 g mL−<sup>1</sup> ChCl-urea and 2.0 mL dipotassium phosphate, K2HPO4 could reach up to 100.5% that collectively highlighted the advantages of the DES-based LBS for the extraction of protein. Unfortunately, this work was unable to back-extract the target protein free from the DES-LBS because of the hydrophilicity characteristic of DES in the aqueous solution.

A similar work with different DESs was investigated by Pang et al. (2017) using DES-based LBS which composed of choline chloride-polyethylene glycol (ChCl-PEG or DES) and sodium carbonate were applied for the extraction of specific protein (i.e., BSA and papain) [52]. ChCl-based DES was prepared by mixing two compounds, 0.68 g mL−<sup>1</sup> ChCl and 0.1 g mL−<sup>1</sup> PEG 2000 at the molar ratio of 20:1, stirring up to 100 ◦C until a homogenous colorless liquid was formed. The result showed that the DES-NasCO3 LBS under the optimum condition had successfully obtained a high extraction efficiency of BSA (95.16%) and papain (90.95%). Moreover, the back-extraction of target protein was performed

by extracting 1 mL DES top phase followed by the addition of ammonium sulfate (NH4)2SO4 and 0.45 mL ethanol to form a new LBS. However, it was found that by increasing the concentration in the salt-rich bottom concentration would lower the efficiency of the back extraction.

A modified DES-based LBS using ultrasonic-assisted were employed for the extraction of ursolic acid from Cynomorium songaricum Rupr [54]. This approach was compared to the convention ultrasonic-extraction method. The recovery yield of ursolic acid was comparable. However, the presence of LBS promotes a higher purification of ursolic acid. The recovery yield of ursolic acid was 22.10 ± 0.44 mg/g with purification factor of 42.41 ± 0.84% as compared to conventional ultrasonic-extraction method where the recovery yield was only 20.9 ± 0.79 mg/g with a low purification factor of 20.17 ± 0.77%.

#### *2.5. Surfactant*/*Detergent-Based LBS*

Surfactant-based LBS is the transformation of phase-forming component from conventional polymer-based LBS. The surfactant-based LBS is formed when both cationic and anionic surfactants are separated into two immiscible liquid phases which consist of a high concentration than critical micelle concentration (CMC) and at certain molar ratio of cationic and anionic surfactant composition. This novel approach of surfactant-LBS has gained interest mainly due to the combination phase which exist in many different forms (i.e., spherical micelles, rod-like micelles, or vesicles) by simply alternating different composition and concentration of surfactants [55]. The principle of surfactant-based LBS used the cloud point extraction (CPE) system in which the non-ionic surfactant is heated above the cloud point temperature, causing dehydration of detergent for the phenomenon of phase separation to occur [56]. The surfactant-LBS consists of one surfactant-rich phase and the other is the surfactant-dilute phase. The organic contaminant will partition into the surfactant-rich phase and will then aggregate and concentrate at that phase. The presence of small amount of remediated water in the contaminant will remain in the surfactant-dilute phase. Surfactant-based LBS is commonly used to separate hydrophobic and amphiphilic molecules by solubilization and partitioning of membrane-bound substances.

Surfactant-based LBS composed of 24% (*w*/*w*) Triton X-100 and 20% (*w*/*w*) xylitol was used for the purification of lipase from pumpkin seeds [57]. The results showed that the surfactant-based LBS had the ability to partition the lipase into the top surfactant-rich phase and leave the impurities at the bottom xylitol-rich phase. The proposed optimized method had successfully recovered the enzyme with purification factor of 16.4 and yield of 97%. This study also demonstrated that the recovery phase component could be recycled up to five runs with a high percentage of recovery of 97%. However, it was noted that there was a significant decrease in recovery of the phase component after the fifth cycle in which could be mainly due to the accumulation of impurities present in the phase component.

An example of surfactant-based LBS extraction was conducted by Sankaran et al. (2018) using surfactant and xylitol under the optimum operation condition of 25% *w*/*w* of xylitol concentration, 15% (*w*/*w*) Triton X-100, 80% *w*/*w* of crude lipase, 4 mL of top phase, 35 mL of bottom phase, pH 7, and 15 min of flotation time showed the maximum lipase extraction and efficiency of 3.63 and 86.46% [58]. In addition, the recyclability of both components in surfactant-LBS extraction makes this an excellent process, as this innovative method was practical and feasible to be applied in the biotechnology industry for extraction of other biomolecules. Table 1 summarizes the extraction of biomolecules using various types of phase-forming component in LBS.



#### **3. Advance Technologies Integrated with LBS**

#### *3.1. Bubble-Assisted LBS*

Bubble-assisted LBS or known as liquid biphasic flotation (LBF) is the combination of LBS and solvent sublation (SS), in which the biphasic medium composed of organic solvent and aqueous salt solution is aerated by air bubbles (e.g., nitrogen and oxygen) in promoting the adsorption of target biomolecules during the separation process [8] (refer to Figure 2a). SS is an adsorptive bubble separation technique introduced by Sebba who suggested that the use of an immiscible thin organic solvent layer overlaid on top of the liquid bulk as a modification of ion flotation [59]. LBF has accommodated the ease for extraction of high value biomolecules such as protein, lipase, astaxanthin, and betacyanin [8,10,60,61]. The theory of LBF system is the phenomenon of surface-active biomolecules having a sorption mechanism between the air bubbles surfaces. The bubbles then arise and dissolve in an organic solvent phase on top of the aqueous solution in the system [11]. With the presence of bubble-assistance in LBS, this could intensively strengthen the adsorption mechanism produced by the bubble transportation; thus, this system is feasible for separation and extraction of biomolecules. Figure 2a illustrates the set-up of bubble-assisted LBS.

A pilot-scale LBF consisting of 0.9 L of 50% (*w*/*w*) of 1-propanol and 1.5 L of 250 g/L ammonium sulfate salt, (NH4)2SO4 had been developed for direct recovery of lipase derived from *Burkholderia cepacian* [62]. The purpose of this study was to conduct a comparison between the recovery of lipase on pilot-scale and small-scale LBF processes. Preshna et al., (2016) had reported that the pilot-scale alcohol/salt LBF system acquired a purification factor of 12.2, efficiency of 88%, and a recovery yield of 93.27% which was feasible for purification of lipase to be implemented into the industrial scale processes [62].

Leong et al. (2018) utilized LBF which composed of 10 mL of 100% ethanol, 20 mL of 200 g/L K2HPO4 salt solution, 1 g FE (peel or flesh of red-purple pitaya), and 15 min flotation time for betacyanins extraction [8]. The results under the optimum conditions of LBF revealed that the betacyanins extractions from 1 g FE of peel in alcohol-rich top phase was 95.989 ± 1.708% with separation efficiency and partition coefficient of 88.361 ± 1.708% and 24.168 ± 2.949%, respectively. The recovery from 1 g FE of flesh was 95.488 ± 0.213% with separation efficiency and partition coefficient of 94.886 ± 0.060% and 21.195 ± 1.030%, respectively [8]. The objective of this work showed that the LBF has a great potential in bioseparation technology as compared with other extraction techniques such as diffusion extraction, ultrafiltration, and reverse osmosis in which only able to recover 70–75% of betacyanins [63].

Rather than using alcohol-based LBS, a recent study had showed the extraction of α-Lactalbumin from whey used a different phase forming component (i.e., PEG 1000 and citrate salts) along with bubble-assisted technologies showed a separation efficiency and purification fold of 87.54% and 5.33 [64]. The advantages of this study had showed the feasibility of bubble-assisted technology compared to conventional liquid-liquid extraction providing a low processing cost, rapid, and good separation yield. However, a further study is required to fulfil the gaps in the bubble-assisted technology. This is to ensure a better understanding regarding the mass transfer and the development of kinetics model of LBF in the separation of biomolecules.

**Figure 2.** Schematic diagram of (**a**) bubble-assisted LBS, (**b**) ultrasound-assisted LBS, and (**c**) electricity-assisted LBS.

#### *3.2. Ultrasound-Assisted LBS*

In the biotechnology processes, cell disruption is considered as the most important process for higher extraction and recovery yield. Ultrasound-assisted LBS is an integrated technique which has been extensively acknowledged by researchers due to its effective properties of cell disruption [65,66]. The advantages of ultrasound-assisted LBS includes low operating cost, less energy consumption and short period of time requirement [67]. The fundamental of ultrasound irradiation is the high shear forces produced from cavitation bubbles of ultrasonic waves and mechanical shears which enhanced the cell disruption for effective biomolecules extraction [68]. Figure 2b shows a schematic set-up of ultrasound-assisted LBS.

A recent study conducted by Sankaran et al. (2018) utilized the application of ultrasound-assisted LBS for extraction of protein from *Chlorella vulgaris* FSP-E microalgae [12]. The authors found that the ultrasound-assisted LBS had the ability to break down the rigid cell wall, followed by the release of protein for extraction. The maximum efficiency and yield of protein were 75% and 65.4%, respectively [6]. An integrated system of ultrasound and LBF was used to compare the effectiveness recovery of the release protein into the solution for extraction [69]. It was reported that the ultrasound-assisted LBF had better advantages over the ultrasound-assisted LBS, driven by its higher concentration coefficient and a better separation efficiency. This was mainly due to the presence of air bubbles which enabled the adsorption of surface-active proteins from the bottom phase to the top phase. As a result, this led to a higher separation efficiency and recovery yield. This integrated sugaring-out ultrasound-assisted LBF under the optimum conditions composed of 100% (*w*/*w*) acetonitrile, 200 g/L glucose concentration, biomass concentration of 0.6% with 5 min of 5 s ON/10 s OFF pulse mode, and at a flow rate of 100 cc/min had given rise to the protein separation efficiency and recovery yield of 86.38% and 93.33%, respectively.

Aside from that, ultrasound-assisted extraction has also been widely employed for the cell disruption of lignocellulose biomass from plants [70]. The extraction of phenylethanoid glycosides (e.g., echinacoside and acteoside) from *Cistanche deserticola* stems using ultrasonic-assisted LBS successfully recovered 27.56 and 30.23 mg/g, respectively [71]. This approach showed that ultrasonic-assisted LBS were efficient, eco-friendly and cheap method for extracting and enriching biomolecules from lignocellulose biomass. However, it is crucial to monitor the process temperature when dealing with ultrasonic irradiation. The high shear forces produced from the cavitation bubbles of sonic wave would generate a high temperature process which will degrade or deform the target biomolecules resulting in an unfavorable low extraction yield. Another supporting research of using the application of ultrasound-assisted LBS was the extraction and separation of antioxidants such as xylooligosaccharides (sugar) and phenolic compound from wheat. In ultrasound-assisted LBS composed of 23.8% (*w*/*w*) ammonium sulfate, 24.3% (*w*/*w*) ethanol, 1.2% (*w*/*w*) biomass loading with ultrasound wave (30 Hz, 500 W, 10 min), extraction yielded the highest recovery of sugar and phenols were 16 mg/g and 2.67 mg/g dry material [72]. This showed that implementation of ultrasound improved the efficiency of extraction of wheat chaff in LBS yielding 1.3–2 times higher, respectively than those without ultrasound.

#### *3.3. Electricity-Assisted LBS*

Electricity-assisted LBS (see Figure 2c) is a promising mild cell disintegration extraction technique for recovery of biomolecules. For instance, the electricity treatment such as pulsed electric field (PEF) demonstrates the conceptualization of the initiation of short electrical pulses in the order of magnitude of ms or μs subjecting the charge in the cell membrane which is sufficient to perform a rearrangement or disruption of membrane and lead to the pore formation. This process is also known as electroporation. However, an optimum condition is required as PEF is dependent on the intensity of the treatment and cell characteristics in which pore formation is reversible or irreversible [73–75]. PEF treatment also increased the mass transfer energy of the system. By combining both PEF and LBS would be an advantage for an efficient extraction of treated sample. This combination is known as an electropermeabilization where the presence of electric and extractive solvent improves the release of

intracellular compound from treated sample [76]. Moreover, electricity treatment not only provides higher extraction efficiency of biomolecules but also a greener approach in the biotechnology industries.

Lam et al., (2017) investigated the operating condition required to release selective proteins from the cell wall of *Chlamydomonas reinhardtii* (cc-124) strain and the cell wall deficient mutant strain (cc-400) using PEF treatment without the presence of LBS [77]. The results showed that after PEF treatment, with operating condition of 5–7.5 kV/cm, 1–10 pulses, and a pulse length of 0.05–0.2 ms on the cell wall, deficient mutant (cc-400) was on average three times higher than cell wall strain (cc-124) with average protein yield of 31 ± 6% protein and 11 ± 3% protein. Additional experiments utilizing PEF treatment with low energy input (range between 0.01 and 0.5 kWh/kgDW) were also conducted on cell wall deficient mutant strain (cc-400) with a maximum recovery of 30% at 0.04 kWh/kgDW. Furthermore, the results obtained from PEF treatment with low energy input was compared with bead beating which only obtain an average of 34 ± 4.2% proteins.

A recent work conducted by Leong et al. (2019) on betacyanins extraction from peel and flesh of red-purple pitaya using the liquid biphasic electric flotation (LBEF) [76] had reported that this new integration process of electricity supplied in LBF system could cause an electropermeabilization of red-purple pitaya membrane structure and improve the betacyanins extraction from red-purple pitaya. An optimum system composed of 100% (*w*/*w*) ethanol, 200 g/L of dipotassium hydrogen phosphate (K2HPO4) with 15 min floatation time (flow rate of 20–30 cc/min), and applied up to 3 V of voltage using graphitic electrodes showed the highest separation efficiency of betacyanins concentration (98.383 ± 0.215% for peel and 96.576 ± 0.0083% for flesh, respectively) [76]. Table 2 summarizes the advance technologies integrated with LBS for the extraction of biomolecules.



#### *Processes* **2020**, *8*, 149

#### **4. Key Parameters A**ff**ecting LBS**

#### *4.1. Type and Molecular Weight of Polymer*

In polymer-salt based LBS, the polymer phase component is crucial as it exhibits different degrees of hydrophobicity on target biomolecules partitioning. As the molecular weight of polymer increases, the hydrophobicity also increases due to the long hydrocarbon chain of monomers. This effect causes a reduction in free volume of the polymer-rich top phase, forcing the target biomolecules to be partitioned to the bottom phase. On the other hand, low molecular weight polymer will decrease the purification factor for target biomolecules as it will be partitioned together with contaminant proteins at the polymer-rich top phase [31,32]. Therefore, it is important in selecting an optimum condition for the hydrophobicity of polymers to obtain the maximum recovery of target compounds.

The effect of molecular weight has been discussed with the used of polymers such as PEG and potassium phosphate salt for the recovery of cyclodextringlycosyltransferase (CGTase) from *Bacillus cereus* [32]. In this work, the different molecular weights of PEG (e.g., PEG 4000, 6000, 8000, 10,000, and 20,000) were used in the LBF system for the CGTase extraction at a constant crude extract to volume ratio of 1.0:3.0. It was found that the maximum purification factor of 7.26 and 97.1% recovery of CGTase were achieved composed of 18.0% (*w*/*w*) PEG 8000 and 7.0% (*w*/*w*) potassium phosphates LBS. Well, as for the lowest molecular weight, PEG 4000 and highest molecular weight, PEG 20,000 showed a purification factor of 2.25 and 3.23, respectively. This indicated that the low molecular weight polymer (PEG 4000) withdraw contaminant biomolecules to the polymer-rich top phase and the high molecular weight (PEG 20,000) would engender a more viscous phase, resulting in the decrease of free volume of polymer-rich top phase caused by volume exclusion effect. In most cases, it is recommended to start with a low molecular weight, depending on the product compatibility while optimizing the partitioning condition.

In addition, one of the limitations of using PEG and salt as the phase-forming component in LBS is that most of them cannot be recycled for the next process. The non-recyclable phase-forming component makes the overall LBS in downstream processes to be unfavorable as it causes environmental pollution and increases cost operation [78]. To improve the recyclability of phase-forming component in the LBS process, another similar research replaced using thermo-separating polymer (EOPO) as the phase component for the purification and recovery of CGTase [31]. The recovery of EOPO after recyclability was more than 80% verifying the viability of recyclable characteristics. This simple, rapid and recyclable feature show that the LBS process is a promising and attractive approach for the recovery and purification of target biomolecules.

#### *4.2. Type and Concentration of Alcohol*

The use of different alcohols (e.g., methanol, ethanol, 1-propanol, and 2-propanol) with different concentrations in the LBS will affect the overall recovery yield of target biomolecules. The exposure of active site from the implementation of organic solvent helps to maintain the enzyme's open conformation and bind the target compounds to the alcohol-rich top phase. A larger amount of alcohol is favorable as it will enhance the target biomolecules buoyancy and stability towards the interface layer.

Santos et al. (2016) conducted an experiment on extraction of caffeine from coffee bean and guaraná seed and reported the possibilities to manipulate the partitioning of caffeine to either the alcohol-rich top phase and salt-rich bottom phase [79]. For caffeine to be partitioned at alcohol-rich top phase, an increase in the concentration of 2-propanol caused the increment in the "caffeine-water" interaction. This effect will promote the biomolecules to be partitioned at the alcohol-rich top phase. Meanwhile, methanol was selected for caffeine to be partitioned at the salt-rich bottom phase. The purpose of selecting methanol was due to its low partition coefficient; therefore, increasing the tendency of caffeine to be partitioned at the salt-rich bottom phase.

A recent study on recovery of glycyrrhizic acid (GA) and liquiritin (LQ) from Chinese licorice root (*Glycyrrhiza uralensis Fisch*) reported that 87% GA and 94% LQ were successful obtained at alcohol-rich top phase under the optimum condition of 25% (*w*/*w*) ethanol and 30% (*w*/*w*) K2HPO4 in the LBS [80]. The effect of alcohol concentrations from 14 to 34% (*w*/*w*) and the extraction efficiency and partition coefficient were studied. By increasing the alcohol concentration to 26% in the system, the extraction efficiency and partition coefficient increased for both GA and LQ biomolecules. However, the extraction efficiency and partition coefficient decreased when the alcohol concentration was increased to 34%. This was due to the large amount of water-soluble alcohol in the alcohol-rich top phase interacting with the water molecules and causing the biomolecules to be partitioned to the salt-rich bottom phase [81]. This term was also referred as "volume exclusion" effect. In general, the selection of alcohol is mainly dependent on the target biomolecules from the complex crude extract. Each target biomolecule has their respective physico-chemical properties and therefore, it is difficult to govern a specific optimum condition for extraction and separation in LBS.

#### *4.3. Type and Concentration of Salt*

In the LBS, it is critical in selecting the type of salts as the phase-forming component since it can significantly affect the solubility and interaction of the target biomolecules. When the salt is added into a solution, the surface tension of water will increase which then leads to the increase of hydrophobic interaction between protein and water [82]. Few studies had shown that a high saturation level of salt concentration will cause a reduction in solubility of target biomolecules due to the higher salting-out ability of salt [20,36]. Lu et al., (2016) reported that the ability of salt solution and hydrophilic alcohol solution to form a biphasic system was mainly dependent on the Gibbs free energy of salt hydration [83]. The alteration in environmental phase system and behavior of biomolecules partitioning is utilized by the different salt components [84]. Different salts used for the LBS were based on their capability to support hydrophobic interaction between biomolecules [85]. According to the Hofmeister series, the salting-out ability of anions are arranged in the following order: SO4 <sup>2</sup><sup>−</sup> > HPO4 <sup>2</sup><sup>−</sup> > citrate3<sup>−</sup> > F− > Cl− > Br− >I <sup>−</sup> > NO−<sup>3</sup> > ClO4<sup>−</sup> [25]. However, an optimum condition is required in order to obtain the maximum recovery of target biomolecules. It is also important to select a biodegradable and eco-friendly salt to ensure a more sustainable green approach in utilizing the LBS.

The effect of various salts used has been studied with the use of potassium dihydrogen phosphate (KH2PO4), magnesium sulfate (MgSO4) and ammonium sulfate ((NH4)2SO4) for the extraction of protein from *Chlorella sorokiniana* microalgae [10]. In this study, the salt concentration of 250 g/L were selected for each salt (KH2PO4, MgSO4, and (NH4)2SO4) as an optimum condition in the LBF. It was found that the KH2PO4, MgSO4, and (NH4)2SO4 exhibited high separation efficiency of 97.85%, 97.74%, and 97.74%, respectively. However, an observation was found using KH2PO4 solution where a white solid was formed and deposited around the interface at flotation time of 1.5 min, showing its incapability for the separation process. This formation happened when the properties of salt having a low solubility. Thus, an addition process is required to melt the solid salt solution. Another observation found using MgSO4 solution was the absence of interface in the LBF after a flotation time of 4 min. In contrast, it was observed that only (NH4)2SO4 solution could clearly render the highest recovery yield and purification values of 56.06% and 68.99%, respectively. The possible explanation was (NH4)2SO4 has a lower molecular weight as compared with KH2PO4 and MgSO4. As a conclusion, the extraction of protein is more favorable in the alcohol-rich top phase with increasing partitioning coefficient (K) when a low molecular weight salts is used [18,34]. However, the selection of various salts is still dependent on the compatibility of LBS and interaction among biomolecules.

The study of salt concentration was continued by using ammonium sulfate at the concentration range of 100 to 300 g/L. The effect of increasing salt concentration tends to increase the protein recovery yield. As supported by Phong et al. (2016), it was stated that the salting-out effect would occur at a higher salt concentration, the presence of ions tended to decrease the solubility of protein in the salt-rich bottom phase [9]. A further increase in salt concentration would decrease the protein recovery percentage. It was recommended to start with a minimum salt concentration of 20% (*w*/*w*) until the optimum condition was obtained rather introducing a high salt concentration abruptly.

#### *4.4. pH System*

The partitioning of target biomolecules can be affected by the pH system in LBS, due to a change in charges and solute properties of solute. The net charge of the target biomolecule becomes negative when the pH value is greater than the isoelectric point (pI) and positive when pH value is lower than the pI. If the net charge is equal to zero, both pH and pI values are equal [86]. Generally, it is found that in higher pH system would induce a positive dipole moment causing the partition coefficient to increase; therefore, favor the partitioning of negatively charge target biomolecules towards the polymer-rich top phase [87,88].

The partitioning of polyhydroxyalkanoate (PHA) from *Cupriavidus necator* H-16 in the thermoseparating-based LBS showed a good setup in altering the pH system as compared with conventional PEG-based LBS [89]. PHA showed a purification factor and recovery yield of 3.67% and 63.5%, respectively, at the pH 6 which was better than the conventional PEG-based LBS that had zero recovery of PHA in the top phase when pH was less than 7. However, there was a sudden drop in PHA recovery yield of 46.4% when the pH was adjusted to 8.0 to 8.8 in the system. Another study of extraction of BSA had shown that the different pH values could alter the net charge of targeted compound [90]. It was reported that the pH value increased from 6.0 to 9.0 which was larger than the isoelectric point of BSA (pI = 4.8) resulted in a maximum recovery yield of 84.32%. However, the high pH value is not favorable in the LBS since it can induce the protein denaturation.

Another experiment of antioxidants (i.e., xylooligosaccharides and phenol) extraction from wheat chaff explored the effect of pH on LBS [72]. The influence of pH value ranging from 2.5 to 7.0 was studied in the case of partitioning parameters of antioxidant such as recovery and partition coefficient. A maximum recovery of sugar ranging from 96% to 99% was obtained at pH 7.0 but the recovery of phenol decreased which could be explained by the phenol compound having a low pKa value of 4.5. In extend, at pH values near pKa such as pH 4.0 was reported that the partitioning of xylooligosaccharides was more towards the ethanol-rich top phase and phenol was more toward salt-rich bottom phase at the highest recovery of 75% and 77%, respectively. Hence, it is important to examine the effect of pH at the optimum condition to enhance the purification factor and recovery yield of the target biomolecules as it could be damaged or denatured by varying the sensitivity of pH conditions.

#### *4.5. Temperature*

The effect of temperature is dependent on the type of phase-forming components used in the LBS and stability of target biomolecules from denaturation. A change in temperature also affects the viscosity and density of the interface in the LBS. In most cases, the optimum temperature within the range of 20 to 40 ◦C was utilized for maximum recovery and partitioning of target biomolecules. The effect of temperature on the extraction efficiency of CPC from *Spirulina platensis* microalgae was studied and the maximum extraction efficiency up to 99.0% was achieved near the temperature range of 308 K [46]. It was found that lowering the temperature to 298 K caused the rate of CPC recovery to decrease, resulting in a low extraction efficiency. The influence of temperature on extraction efficiency study of BSA and papain was evaluated [52]. However, the studies showed that the extraction efficiency of both BSA and papain decreased when the temperature was increased. This phenomenon was due to the increasing temperature which could inhibit the interaction of amino acid and surface water of protein, resulting in less efficiency of protein extraction [89]. Hence, the effect of temperature should be taken into consideration as the extraction efficiency of the biomolecule is dependent on the range of temperature in the LBS.

#### **5. Future Prospect and Challenges of LBS Application**

The use of LBS is a promising separation technique for extraction of valuable biomolecules. The LBS can serve as an analytical tool to understand the chemical properties and behavior of target biomolecules. However, developing LBS as an alternative way for separation and purification for large-scale industrial application does encounter some key challenges that have to be re-addressed. One of the major concerns regarding the LBS is the partitioning coefficient (K) of the biomolecules into the top phase which is mainly dependent on the key parameters. It is time-consuming and crucial to investigate each of the key parameters in order to determine the optimum condition for maximum purification and recovery of biomolecules. The selection of phase-forming components should also be made concerning to their biocompatibility, hazards and biodegradability. Therefore, this favors an alternative phase-forming component which is more environmentally friendly and highly biodegradable in the aquatic environment.

Another challenge that needs to be addressed in the LBS is the extraction of biomolecules from natural sources. Regardless of various studies reporting the efficiency of LBS in the extraction of biomolecules from natural sources and microbial fermentation broths, it is still difficult to understand the biomolecules partition behavior, particularly when a complex crude feedstock is added into the LBS. Moreover, the contaminants might have the similar characteristics with the target biomolecules in the crude feedstock. This will cause lower extraction efficiency during the interaction with the extraction medium in the LBS. The lack of understanding on the partition behavior remains a challenge in utilizing the LBS for the recovery of biomolecules.

Furthermore, there is still some unexplored technologies such as magnetic and microwave approach which can be integrated with LBS for enhancing biomolecules extraction efficiency. The implementation of these new advanced technologies would be beneficial to enhance the knowledge in LBS. Moreover, it has been proven that the present assisted technologies using bubble-, ultrasound-, and electricity-assisted technologies showed a promising prospect in the recovery and purification of biomolecules. However, these assisted technologies required an in-depth study due to the lack of knowledge between its physico-chemical mechanism aligned with the LBS.

To maximize the large-scale use of LBS requires an ideal optimization technique at where the system can perform at its best desired performance for various application. A recent review by Torres-Acosta et al., (2019) has comprehensively evaluated the strategies to incorporate the LBS technologies in the industry [91]. One of the most frequent optimization techniques used is univariate optimization or known as one-factor-at-a-time (OFAT) analysis is where a single parameter at the time after the other is selected based on its best performance. Aside from that, response surface methodology (RSM) is another optimization technique found in most literature studies. This optimization technique composed of a statistical design that allows a simultaneous variation of several parameters compared to OFAT which depends on a single parameter at a time. Lastly, genetic algorithms (GA) is less frequently used compared to RSM however can deliver excellent results. The fundamental of GA involves the natural genetic inheritance (genotype) which relate specifically to the raw information of LBS components such as concentration of alcohol, salt, pH system, and temperature and then interprets the results (i.e., recovery yield, partition coefficient, and separation efficient) based on the characteristic of the LBS. The advantages of this GA compared to RSM is that GA does not require a regression or model tool as the optimizing approach, as the LBS is based on the previous results. In general, these optimization strategies are one of the best-selling points to make LBS to be implemented at industrial scales.

Aside from that, another strategy which could beneficial to the LBS is to study the recyclability and reusability of the phase-forming components in LBS. This is to ensure that the LBS not only can be employed as a separation and purification technique, but also promoting a sustainable low-cost process in the downstream processes. On top of that, the implementation of extractive technologies such as fermentation, cell disruption, bioconversion, crystallization, distillation, and metallurgy can be proposed along with the LBS to allow the production and purification tasks to occur in one-step process. The advantages of this extractive technologies prevent the inhibition of the product and enhance the stability of biomolecules in the production stage. These benefits from extraction technologies should be further explored for the future development of LBS.

#### **6. Conclusions**

LBS is a simple, selective, scalable, and efficient tool to be utilized in downstream processing for the purification and recovery of biomolecules. However, it is still yet favorable to be used at the commercial scale as the complexity of the partitioning mechanism is difficult to predict. The challenges associated with the LBS techniques such as economic feasibility and the understanding of partition behavior need to be addressed to ensure the applicability in biotechnology industries. It is believed that more development along with various kind of technologies integrated in LBS will be discovered in the future. Hence, promoting the LBS to be used in commercial applications in recovering various high value bio-based products.

**Author Contributions:** Writing—original draft, K.S.K.; Writing—review & editing, K.S.K., H.Y.L., K.W.C. and J.-W.L.; Conceptualization, P.L.S.; Supervision & Funding acquisition, P.L.S., T.C.L. and H.-W.Y. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by the Fundamental Research Grant Scheme, Malaysia [FRGS/1/ 2019/STG05/UNIM/02/2].

**Conflicts of Interest:** The authors declare that they have no conflicts of interests.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Review*
