Lignocellulosic bioethanol production: prospects of emerging membrane technologies to improve the process – a critical review

2.1 Pretreatment processes

Pretreatment is the fundamental strategy to reduce the recalcitrance nature of lignocellulosic material to facilitate the enzymatic action on plant polysaccharides. The main purpose of pretreatment is to solubilize and separate one or more of the structural components, namely, hemicellulose, lignin, and cellulose, from biomass. Thus, the process increases substrate porosity with lignin redistribution and permits the maximum exposure of saccharification enzymes to cellulose and hemicellulose surfaces to achieve improved hydrolysis with minimal energy consumption (Zhu and Pan 2010). In the last few decades, several approaches have been evolved as economic and most useful pretreatment processes for lignocellulosic biomass. Pretreatment processes are broadly classified in four different groups, namely, (a) physical pretreatment, (b) chemical pretreatment, (c) physicochemical pretreatment, and (d) biological pretreatment. Different classifications of all these existing pretreatment methodologies are represented in Figure 1.

Figure 1:

Classifications of different pretreatment methodologies to reduce the recalcitrant nature and crystallinity of lignocellulosic biomass.

Under physical pretreatment, mechanical comminution is considered as the most common technique where, through the combination of chipping, grinding, and milling, lignocellulosic materials can be comminuted from 10 to 30 mm to 0.2 to 2 mm (Kumar et al. 2009, Balat 2011). Milling methods, specifically ball milling, two-roll milling, hammer milling, colloid milling, and vibro-energy milling, can play a significant role in reducing the particle size and degree of crystallinity (Taherzadeh and Karimi 2008) and improving the digestibility (Kumar et al. 2009) of lignocellulosic biomass. Energy consumption in mechanical comminution is totally dependent on the type of material used during the process and particle size required after pretreatment. Balat (2011) clearly reported that the energy consumption to reduce the particle size of 1.6 mm for hardwood material was 130 kWh/ton, and it was 14 kWh/ton for the same particle size reduction of corn stover. The specific energy consumption was reported to reduce the particle size of wheat straw material to 0.8 and 3.2 mm through hammer mill, which were 51.6 and 11.4 kWh/ton (Talebnia et al. 2010). Pyrolysis is another endothermic physical pretreatment process, which is well known for less energy consumption (Sarkar et al. 2012). In this method, the cellulosic material rapidly decomposes to gaseous products such as H₂, CO, and residual char through the exposure of temperature greater than 300°C. The residual char is further processed by leaching to get enough carbon sources such as glucose to support microbial growth (Sarkar et al. 2012). Leustean (2009) clearly pointed out that pyrolysis pretreatment was effective to achieve improved conversion of cellulosic material from ground material to glucose. Microwave and electron-beam irradiation are also feasible physical pretreatment approaches that are easy to operate (Bjerre et al. 2000). Microwave pretreatment uses high heating efficiency of a microwave oven whereas electron-beam irradiation is aligned with continuously changing magnetic field (Sarkar et al. 2012). Irradiation basically helps in other pretreatment processes as it cannot separate hemicellulose or lignin directly from lignocellulosic materials (Sun et al. 2016). Duarte et al. (2012) highlighted that electron beam or γ-rays combined with mechanical crushing can further enhance the enzymatic saccharification process of lignocellulosic biomass. High-energy radiations of γ-rays with deeper penetration power is successful in improving the digestibility of enzymes (Jahnavi et al. 2017). High-energy requirement and the expensive nature of such treatments hinders their applicability in industrial applications (Taherzadeh and Karimi 2008).

Alkaline pretreatment is regarded as the most commonly used chemical pretreatment method to digest lignin matrix and make cellulose and hemicellulose available for enzymatic degradation. A variety of alkaline reagents such as sodium hydroxide (NaOH), Ca(OH)₂, KOH, Na₂CO₃, and aqueous ammonia have been implemented for different lignocellulosic materials, and among them NaOH, Ca(OH)₂ is generally used in most of the cases (Sun et al. 2016). The cell wall of lignocellulosic materials is normally disrupted by solubilizing hemicelluloses, lignin, and silica and the crystallinity of cellulose is finally reduced due to swelling (Sarkar et al. 2012). It helps the substrates to be fractionated into alkali-soluble lignin, hemicelluloses, and residues. The advantages of alkaline pretreatment includes ambient operating conditions and the utilization of lower temperature and pressure than other treatment procedures (Mosier et al. 2005, Balat 2011). Cai and Zhang (2005) pointed out the efficiency of alkaline NaOH/urea solutions in dissolving, swelling, and fragmenting cellulose material at low temperature (about −15°C). Sun et al. (2016) suggested that low temperature at NaOH treatment leads to high accessibility of lignocellulosic material in saccharification stages by increasing the swelling of that material. The effectiveness of NaOH in increasing the digestibility of hardwood material from 14% to 55% and reduction of lignin content up to 55% was shown by Kumar et al. (2009). Sun et al. (1995) evaluated the most effecting NaOH concentration of 1.5% for separation of 60% lignin and 80% hemicellulose from wheat straw material when the material was treated at 20°C for 144 h. NaOH pretreatment was effective enough to retain the lignin content to even less than 26% for hardwood, wheat straw, switch grass, and softwood materials (Zhao et al. 2008). Less than 65% reduction of lignin from cotton stalk material was similarly reported by Silverstein et al. (2007) when the material was treated with 2% NaOH for 90 min at 121°C and 15 psi. Utilization of NaOH pretreatment and its combination with other pretreatment processes in the reduction of lignin and increasing accessibility of cellulose from kans grass material was demonstrated by Ashgar et al. (2016), Chaudhary et al. (2012), and Kataria and Ghosh (2014). The maximum loss of solid including lignin of 47.5% was reported by Kataria and Ghosh (2014), whereas total lignin removal of 40% and 74% were reported by Ashgar et al. (2016) and Chaudhary et al. (2012), respectively, during NaOH pretreatment. Ammonia pretreatment is also considered as noncorrosive and nontoxic alkaline treatment (Kim et al. 2016). Compared to NaOH pretreatment, lime (Ca(OH)₂) treatment is another efficient method that leads to increase the crystallinity index with the removal of lignin involving lower cost, less safety requirements, and better environmental benefits (Alvira et al. 2010, Balat 2011). Kim and Holtzapple (2005) reported that removal of 43.6%–48.4% original lignin was possible from corn stover-based lignocellulosic material by nonoxidative lime pretreatment, whereas lignin removal was 57.8%–87.5% from oxidative lime pretreatment at 25°C–55°C for 16 weeks.

Among all other chemical pretreatments, acid pretreatment is considered effective to achieve directly high yield of sugars from lignocellulosic materials. Even acid pretreatment is also applied to solubilize fractional hemicelluloses from lignocellulosic material; hence, it has been preferably used in the process of fractionating components of lignocellulosic biomass (Sun et al. 2016). The acids that are studied in such purposes are hydrochloric acid (HCl), sulfuric acid (H₂SO₄), phosphoric acid (H₃PO₄), and nitric acid (HNO₃; Menon and Rao 2012). Operating conditions that are followed for acid treatments are (a) high temperature and low acid concentration and (b) low temperature and high acid concentration. For a large array of biomasses such as switch grass, corn stover, spruce, poplar, kans grass, coconut fiber, and rice/wheat straw, H₂SO₄ has been commonly implemented in diluted conditions (Zhu et al. 2008, Wyman et al. 2009, Digman et al. 2010). However, the biggest drawback of acid treatment is the production of various microbial growth inhibitors such as acetic acid, furfural, and 5-hydroxymethyalfurfural, which are needed to be detoxified before fermentation. Generally, higher recovery of xylose and enhanced enzymatic digestibility can be achieved with higher pretreatment temperature and short residence time. Diluted H₂SO₄ concentration of (0.2–2%) and higher temperature (121–372°C) were implemented in all existing studies in the same research field as it facilitates most important functions such as hydrolysis of hemicellulose, exposure of cellulose for digestion, and solubilization of heavy metals (Balat 2011). Hernández et al. (2013) demonstrated the maximum recovery of cellulose up to 87% and xylan up to 50.2% from empty pods of Moringa oleifera at diluted H₂SO₄ treatment. Saha et al. (2005) achieved 74% yield in saccharification process of wheat straw material when it was treated with 0.75% (v/v) volume of H₂SO₄ at 121°C for 1 h. Solid acid catalyst utilization in the pretreatment process is another emerging technique that leads to better saccharification efficiency. In a saccharification study of macroalgae cellulosic material (Tan and Lee 2015), optimum yield of glucose was enhanced to 99.8% through soiled acid pretreatment technique [10% (w/v) biomass loading, 4% (w/v) catalyst loading, 30 min, 120°C] followed by enzymatic hydrolysis. In another experimental observation, Qi et al. (2018) developed carbon-based solid (C-SO₃H) acid catalyst by carbonization technique using microcrystalline cellulose and H₂SO₄. Such solid catalyst showed high catalytic activity during pretreatment of corncob-based lignocellulosic substrate and the initial xylose recovery yield was achieved 78.1%.

Recently, ionic liquids as green solvent consisting large organic cations and small inorganic anions are gaining importance for the pretreatment of lignocellulosic biomass. Mainly two classes of ionic liquids (simple salts and binary ionic liquids) have been implemented as powerful solvents to dissolve cellulose or lignocellulosic biomass (Menon and Rao 2012). Li et al. (2010) investigated the effect of phosphate and chloride group of ionic liquids in the pretreatment studies of corncob. For sugarcane bagasse, twofold higher enzymatic hydrolysis yield was observed when it was pretreated with 1,3-N-methylmorpholine-N-oxide ionic liquids (Kuo and Lee 2009). In a similar study (Saha et al. 2017a), better pretreatment of sugarcane bagasse was obtained with ionic liquid 1-ethyl-3-methylimidazolium acetate at the ratio of 1:20 (w/w) at 140°C for 120 min. Xu et al. (2015) pointed out that pretreatment for 24 h with similar ionic liquid 1-ethyl-3-methyl-imidazolium dimethylphosphate was effective to solubilize more than 40% lignin. Zhao et al. (2009) clearly indicated that lignocellulosic material pretreated with ionic liquids always shows much lower crystallinity and higher accessibility of enzymes than untreated biomass. Dadi et al. (2006) highlighted ionic pretreatment with Avicel-PH-101 was useful to increase the enzymatic hydrolysis rate and yield up to 50% and twofold, respectively, in comparison to the untreated one. Jahnavi et al. (2017) indicated the presence of four major groups of cations in ionic liquids to categorize ionic liquid. Menon and Rao (2012) clearly explained different aspects including advantages, disadvantages, and challenges of using ionic liquids in different pretreatment case studies. Some techniques such as ozonolysis, organosolv, wet oxidation, cosolvent fractionation, and γ-valerolactone have been newly emerged as potential chemical pretreatment process, but still the industrial feasibility of such processes will be dependent on their technoeconomic suitability for large-scale operations. Among all such mentioned techniques, oraganosolv (utilization of organic solvent and/or aqueous solvent) is a promising alternative method for the delignification of lignocellulosic material. The utilization of methanol, ethanol, acetone, ethyl glycol, and tetrahydrofurfural alcohol allows the isolation and recovery of pure lignin (Jahnavi et al. 2017). The processes are normally carried out between 150°C and 200°C to break hemicellulose bonds and such processes became efficient when they are combined with acid catalysts (Alvira et al. 2010).

Steam explosion, which is known as autohydrolysis, is the most promising method under physicochemical pretreatment. Lignocellulosic materials are subjected to high-pressure saturated steam (0.69–4.83 MPa) and temperature (160–260°C) for a short period of time like to several seconds to minutes and then suddenly the pressure is reduced to atmospheric pressure. Pan et al. (2005) pointed out that hemicelluloses and lignin materials can be released through stream pretreatment followed by NaOH treatment. Horn et al. (2011) showed that steam-treated (170–220°C) wheat straw sample became efficient during enzymatic hydrolysis. Residence time, temperature, biomass size, and moisture content are the most influencing parameters in such process (Jedvert et al. 2012). Due to some attractive features such as low environmental influence, utilization of less harsh chemicals, and high-energy efficiency, steam explosion has now been widely used before lignocellulosic hydrolysis (Alvira et al. 2010). The main drawback of this process is the production of inhibiting aromatic compounds or byproducts, which are required to be detoxified through vigorous water washing or other methods. Boussaid et al. (2000) indicated that both enzymatic digestion and recovery of total sugars have been improved with medium steam explosion technique. Addition of acids such as H₂SO₄ at low concentration or 0.3%–3% CO₂ improves the overall process by reducing temperature, increasing hydrolysis rate, and lowering the production of inhibitory compounds by the complete removal of hemicelluloses (Bondesson et al. 2013). Martín et al. (2002) explored the same advantage of the process with sugarcane bagasse material when it was impregnated with SO₂ and H₂SO₄ before steam explosion at 205°C for 10 min. In this study, higher yield of xylose was obtained from SO₂ impregnated sugarcane bagasse, whereas higher yield of glucose was achieved from H₂SO₄ impregnated sugarcane bagasse. Tomás-Pejó et al. (2008) observed that wheat straw became a useful substrate for SSF-based bioethanol production after steam explosion treatment at 200°C for 10 min.

Ammonia fiber explosion (AFEX) is another physicochemical pretreatment method which exposes lignocellulosic material to liquid ammonia at relatively moderate temperature of 90°C–100°C and at high pressure of 250–300 psi for 30–60 min (Kim et al. 2011). The salient feature of this process is the partial decrystallization of cellulose and the fractionation of lignin is caused by the swelling action of liquid ammonia. Although AFEX pretreatment method is effective to remove little lignin and hemicellulose from lignocellulosic material, it does not produce any inhibitors and the processing time is comparatively less. Parameters such as ammonia loading, temperature, water loading, blow down pressure, number of treatments, and time determine the efficiency of the process. The main drawback of such process is that it is less efficient for biomass containing higher lignin such as softwood newspaper (Sun and Cheng 2002). Alizadeh et al. (2005) achieved 93% glucan conversion from switch grass pretreated with liquid ammonia near 100°C for 5 min. Similarly Teymouri et al. (2005) achieved approximately 100% conversion of cellulose and 80% conversion of hemicellulose to fermentable sugars when corn stover was pretreated with AFEX at 90°C for 5 min followed by enzymatic hydrolysis. Harun et al. (2013) recovered 97% ammonia and it was reused for the subsequent batches of AFEX pretreatment with rice straw as substrate.

Under physicochemical pretreatment, liquid hot water (LHW) method is another useful strategy to hydrolyze the maximum amount of hemicellulose (Das Neves et al. 2007). The temperature normally maintained in this hydrothermal process is between 170°C and 230°C and at the pressure above 5 MPa for 20 min. Water acts as acid at higher temperature; therefore, it facilities the maximum hydrolysis of hemicellulose into liquid phase and the maximum recovery of xylose (88–98%) can be achieved without any other chemical treatment. Although the process is regarded as environmentally and economically attractive, it still suffers from the production of microbial growth-inhibiting compounds such furfural and carboxylic acid. In an experimental investigation of LHW with bamboo, Xiao et al. (2014) observed the maximum removal of hemicelluloses and less degradation of cellulose after different stages of pretreatment in temperature range of 140°C–200°C for 10–120 min. Yu et al. (2009) achieved the maximum xylose recovery of 86.4% through two-stage LHW treatment of Eucalyptus grandis. Zhuang et al. (2016) evaluated the comparative effectiveness of LHW process in lowering downstream pressure to increase the accessibility of cellulose. Lu et al. (2013) observed that LHW treatment at 180°C–210°C for 20 min was useful for fed-batch SSF process for bioethanol production from reed-based substrate.

Among all such pretreatment processes, biological pretreatment is considered as the most environment-friendly process as degradation of lignocellulosic complexes is carried out in mild conditions with the help of microorganisms such as brown rot, white rot, and soft rot fungi. Among such fungi, white rod fungi such as Phanerochaete chrysosporium, Ceriporia lacerata, Cyathus stercoreus, Ceriporiopsis subvermispora, Pycnoporus cinnabarinus, and Pleurotus ostreatus have been used successfully in the pretreatment process of different lignocellulosic biomass (Hatakka 1983, Prasad et al. 2007). Enzymes such as laccase, lignin peroxidase, and manganese peroxidases can play the most vital role in such pretreatment process by modifying lignin content (Floudas et al. 2012). Although the process has many advantages including less energy requirement, utilization of no added chemicals, and mild condition for treatment, long pretreatment time impedes its industrial application (Sun et al. 2016). Suhara et al. (2012) achieved high lignin degradation (>50%) from bamboo culms after 12 weeks of pretreatment using white rot basidiomycete Punctularia sp. In another study of biological pretreatment, Singh et al. (2008) achieved delignification efficiency of 60%–92% from fungus Trichoderma reesei and Aspergillus terreus. Canam et al. (2011) highlighted the potential of genetically modified microorganism such as the cellobiose dehydrogenase-deficient Trametes versicolor strain for such pretreatment studies.

A lot of research efforts have been focused for the successful conversion of lignocellulosic biomass to its valuable constituents through the development of effective pretreatment strategy. But still, obstacles in the existing pretreatment processes make the overall lignocellulosic bioethanol production commercially challenging. It includes the insufficient separation of cellulose and lignin, formation of byproducts that inhibit ethanol fermentation, high use of chemicals or energy, and considerable generation of wastes (Menon and Rao 2012). The conventional method of acid pretreatment with concentration below 4% and temperature greater than 160°C comes up with the formation of toxic inhibitors such as furfural from xylose and hydrothermal furfural (HMF) from glucose in addition to acidic acids and phenolics (Taherzadeh 1999). Separation and detoxification of acetic acid with concentration greater than 10 g/l is more difficult than furfural and HMF. Therefore, the estimation of inhibitory toxic levels is the major concern for evaluating the cost-effectiveness of pretreatment processes. Parameters such as toxic inhibitors formed per level of sugar recovered and ratio of energy consumption versus sugar yield mostly determines the overall cost-effectiveness of the process (Zhu and Pan 2010). Several pretreatment methods, including mechanical, chemical, or biological methods, have been evolved for reducing the recalcitrant nature of the cell wall, although some of them still remain economically unfeasible due to key technical issues. The efficient pretreatment process for one particular feedstock may not be suitable for another biomass. Biological pretreatment necessities, careful control of slow microbial growth conditions, and a large amount of space are required to carry out the treatment processes, and most of the lignolytic microorganisms solubilize cellulose and hemicellulose along with lignin (Eggeman and Elander 2005). As a result, the treatment became less attractive in a commercial standpoint. Only few pretreatment processes including steam explosion, LHW, and AFEX have been reported for technoeconomic analysis. As membrane processes permit the permeability of the selective components, it can play the most effective role for the separation of cellulose, hemicellulose, lignin, and inhibitory components from pretreated mixture. Hence, membrane-integrated treatment opportunities after immediate pretreatment stages will definitely helpful improve the overall process economics.

2.2 Hydrolysis method

Chemical hydrolysis and enzymatic hydrolysis are two most commonly applied methods used for the conversion of simple sugars from lignocellulosic materials before fermentation. Other hydrolysis methods including γ-rays or electron-beam irradiation or microwave irradiation can be implemented for such conversions, but the assessment of economic suitability for such processes is more important before industrial applications (Balat 2011).

Under chemical hydrolysis, diluted acid hydrolysis and concentrated acid hydrolysis are two conventional techniques where lignocellulosic materials are exposed to diluted acid or concentrated acid for a specific period of time and specific temperature for the conversion of simple sugars. Diluted acid hydrolysis is comparatively more advantageous as it facilitates the step-by-step hydrolysis for the separation of hemicellulose and cellulose compounds, whereas concentrated acid (70–90%) hydrolysis mainly facilitates the liberation of hemicellulosic sugars (Hayes 2009). Minimum sugar degradation and achieving maximum sugar yield up to 100% are the most positive features of such process, although the high cost of acid consumption and environmental and corrosion problems make the process unsuitable for commercial applications.

Enzymatic hydrolysis comes up with 100% selective conversion of hemicellulosic materials to simple sugars with fewer requirements of energy and mild environmental conditions (Ferreira et al. 2009). The recalcitrance nature of lignocellulosic materials due to the presence of lignin, high surface area, and cellulose crystallinity makes the enzymatic hydrolysis very slow (Pan et al. 2006). Mild environmental conditions (temperature 40–50°C, pH 4–5), less corrosion properties, high yield (75–85%) of reactions, and selective operations make the enzymatic hydrolysis more suitable than acid hydrolysis (Hamelinck et al. 2005). Enzymatic hydrolysis is substrate specific, as cellulase enzymes break down bonds of cellulose molecules and hemicellulase enzymes are only responsible for hemicellulose molecules. Cellulases are group of enzymes that synergistically hydrolyze cellulose to glucose monomers. Such enzymatic mechanism is carried out by endoglucanases, exoglucanases, or cellobiohydrolases and β-glucoside enzymes. Endoglucanase hydrolyzes intramolecular β-1,4-glucosidic bonds of cellulose chains randomly to produce new chain ends, whereas exoglucanases suitably cleave cellulose chains at the ends to release soluble cellobiose or glucose and β-glucosides hydrolyze cellobiose to glucose. Hemicellulases are the complex mixture of at least eight enzymes that hydrolyze hemicellulose to xylose and arbinose as five-carbon sugars and galactose, glucose, and mannose as six-carbon sugars. The generalized process for the pretreatment of lignocellulosic biomass to remove lignin and further enzymatic saccharification process to produce simple sugars is represented in Figure 2. A wide range of fungi and bacteria such as Trichoderma, Penicillium, Fusarium, Phanerochaete, Humicola, Schizophyllum, Aspergillus, and Bacillus spp. have been successfully reported for cellulase enzyme production (Cardona Alzate and Sánchez Toro 2006). In the last three decades, Trichoderma has been the most well-studied strains among various microorganisms for the production of cellulase and hemicellulase enzymes (Menon and Rao 2012). Although the enzymes produced by T. reesei are efficient, they suffer from low glucosidase activity and finally lead to incomplete hydrolysis of cellobiose in reaction (Menon and Rao 2012). Aspergillus is another well-known fungus for β-glucosidase production from complex lignocellulosic compounds (Taherzadeh and Karimi 2007). Sometimes, the enzyme mixture of both species plays the most efficient role in improved hydrolysis processes (Chen et al. 2008, Sateesh et al. 2012).

Figure 2:

Pretreatment of lignocellulosic biomass to remove lignin and further enzymatic saccharification process to produce simple sugars.

The most influencing process parameters such as temperature, pH, mixing rates, substrate concentration, and cellulase enzyme loading determine simple sugar yields from enzymatic saccharification processes (Olsson and Hahn-Hägerdal 1996, Alkasrawi et al. 2003). In terms of substrate concentration, high solid loading [>15% (w/v)] is effective to improve the process efficiency and economy (Zhang and Lynd 2010). Such environment-friendly high solid loading saccharification process comes up with lower capital investment costs due to the use of fewer reactors, less energy consumption during heating and cooling, and reduced disposal costs due to less usages of water (Zhang et al. 2013). However, the challenges associated with such high solid loading saccharification processes are increased viscosity due to difficulties in mixing, heat and mass transfer limitations, and inhibition of toxic components during fermentation (Alvira et al. 2013). Such drawback related to high solid loading could be dodged with the application of high enzyme dosages and novel process strategies. The fed-batch approach is considered as the more efficient process while dealing with high solid loading for various advantages (Hodge et al. 2009). Fed-batch feeding strategies provide sequential time-dependent addition of substrates to deal with the maximum concentration of substrate and to liquefy the substrate before next addition (Liu et al. 2017). The concept of fed-batch feeding strategies in high solid loading saccharification process is represented in Figure 3. Liu et al. (2015) applied fed-batch saccharification process with sugarcane bagasse with maximum 15% solid loading. In this study, viscosity-dependent sequential feeding was implemented with gradually less concentration (8%, 7%, and 6%) of sugars. Maximum glucose concentration was achieved with 134.9 g/l from total solid content of 36% and enzyme dosages of 8.5 FPU/g. In a similar fed-batch saccharification process, time-dependent optimal solid loading of alkali-pretreated sugarcane bagasse was implemented by Gao et al. (2014) with enzyme loading of 10 FPU/g. Fresh solids of 7% were consecutively added at the intervals of 6, 12, and 24 h to achieve maximum 33% solid loading, 129.50 g/l glucose concentration, and 56.03 g/l xylose concentration at the end of the process. Sugiharto et al. (2016) clearly highlighted that enzyme feeding strategies can play an important role in fed-batch enzymatic hydrolysis process. In this study, commercial Cellic CTec 2 enzyme with dosages of 185 FPU/ml from Novozymes was implemented on 25% solid loading for 40 h to increase the digestibility and produce glucose concentration up to 26% and 12%. Liu et al. (2017) evaluated the effect of solvent water on high solid saccharification process and it was concluded that solvent such as oleyl alcohol instead of water as solvent was inhibitory for cellulose hydrolysis. Wang et al. (2017) implemented an applied electric field to improve the saccharification of lignocellulosic substrate. Conventionally, batch saccharification process is considered as time taking as common time duration is 3–7 days and high solid loading is required in that process to achieve acceptable conversions (Selig 2008). As an suitable alternative to batch saccharification, continuous countercurrent saccharification was proposed by Zentay et al. (2016) while working with commercial α-cellulose. As the utilization of commercial enzyme in saccharification process is considered as expensive, production of enzymes from cheap substrate, immobilization of enzymes, and recycle and reuse of the enzymes are alternative solutions to make the enzymatic saccharification process commercially more viable. Aligned with the same concept of enzyme recycling, Pihlajaniemi et al. (2014) developed solid recycling process for improved enzymatic saccharification process. During saccharification process, cellulase enzyme can be present in soluble form and the maximum amount of insoluble cellulase enzyme can be adsorbed onto the solid substrate. The recycling of such enzyme adsorbed solid substrate after hydrolysis can be considered as an suitable approach for enzyme recovery and reuse (Pihlajaniemi et al. 2014). To improve the enzymatic hydrolysis process, Van Dyk and Pletschke (2012) emphasized that surfactants are normally used as common additives to prevent the nonproductive adsorption of enzymes on solid substrate.

Figure 3:

Concept of fed-batch feeding strategies in high solid loading saccharification process.

Although acid hydrolysis occurs comparatively in a shorter duration of time, certain drawbacks make the process industrially unfit. For diluted acid hydrolysis process, the biggest limitation is how to increase the glucose yields to more than 70% while minimizing glucose decomposition and maintaining high cellulose hydrolysis rate. On the contrary, expensive concentrated acid hydrolysis process leads to environmental and corrosion problems. Alternatively, as enzymatic hydrolysis is carried out at mild environmental conditions with regard to pH and temperature and involves low energy requirement, the utility cost of the process is considered much lower than acid hydrolysis (Al-Zuhair et al. 2013). However, the obstacles that have been observed in enzymatic hydrolysis lack an ideal reactor system, slow reaction rate, ineffective use of cellulase, and enzyme inhibition by cellobiose and glucose (Noble et al. 1990). Even xylo-oligosaccharides such as xylose produced during hemicellulose hydrolysis create a higher level of cellulase enzyme inhibition than xylin (Qing et al. 2010). Still now, batch reactors are mostly used for industrial enzymatic process. The difficulties that have been encountered with batch reactors are batch-to-batch oscillations, high cost of labor, frequent startup and shutdown problems, and the need to recover enzyme or preparation of enzyme after each batch (Prazeres and Cabral 1994). Again, using the free form of enzyme in a continuous process is economically unfeasible as free enzymes leave from the outlet of the continuous process. MBR is considered as one of the relevant approaches to address all those issues involved in enzymatic saccharification process. It facilitates the reuse and recycle of expensive enzymes while avoiding product inhibition problem by instant permeation of reaction products such as cellobiose and glucose. In the last few decades, continuous MBR has been developed to improve enzymatic hydrolysis operations (Mameri et al. 2000). On the contrary, MBR also facilitates saccharification reaction in the form of enzyme immobilization where membrane can act as solid surface where the chemical and physical attachment of the enzyme and product removal can be done simultaneously.

2.3 Conventional fermentation technology to produce bioethanol

In this process, depending on the compositions of lignocellulosic hydrolyzate materials, the conversion of sugars to ethanol is normally performed through a biological pathway specifically by bacteria or yeast. Parameters such as process economics, kinetic properties of the microorganism, and types of lignocellulosic hydrolyzate help to determine the suitable mode of fermentation among batch, fed-batch, and continuous modes.

For a commercially vibrant bioethanol production process, an ideal microorganism plays most the vital role. It should have broad substrate conversion capacities and high ethanol yield productivity and it must have the ability to withstand high processing temperature and high concentration of ethanol. Mostly, Saccharomyces cerevisiae commonly known as industrial yeasts and Zymomonas mobilis have been used for bioethanol production in brewery and wine industries (Limayem and Ricke 2012). The main limitation of such organisms is that they are capable of fermenting hexose sugars such as glucose to ethanol but unable to ferment pentose sugars such as xylose. Pentose substrates such as xylose can be metabolized by Pichia stipitis, Candida shehatae, and Candida parapsilosis to produce bioethanol. Recently, recombinant S. cerevisiae carrying heterologous gene for XR and XDH from P. stipitis and xylulokinase (XK) have been developed to produce bioethanol from both hexose- and pentose-based substrates (Katahira et al. 2006). Compared to conventional yeasts, bacteria such as Z. mobilis, Escherichia coli, and Klebsiella oxytoca are well known for their rapid fermentation of simple sugars. Saez-Miranda et al. (2006) clearly demonstrated that, compared to traditional S. cerevisiae, Z. mobilis can achieve 5% higher bioethanol yield and fivefold higher volumetric productivity of bioethanol from glucose-based solution. For the fermentation of pentose and hexose sugars from lignocellulosic hydrolyzate material, the utilization of mixed microbial population has been preferred more. Chandel et al. (2011) achieved a maximum 15±0.92 g/l bioethanol production from Saccharum spontaneum implementing cocultures of P. stipitis and thermotolerant S. cerevisiae. In a similar experimental investigation with S. spontaneum, Singh et al. (2014) developed a sequential coculture system with P. stipitis and Z. mobilis for better utilization of total sugars. Through the successful applications of metabolic engineering, the E. coli strain was developed with the ability to ferment a wide variety of sugars with no additional requirements (Doelle et al. 1989). Thermophilic microorganisms such as Clostridium thermocellum, Thermoanaerobacterium saccharolyticum, Thermoanaerobacterium thermosulfurigenes, Thermoanaerobacterium ethanolicus, Thermoanaerobacterium pseudoethanolicus, Thermoanaerobacterium mathranii, Thermoanaerobacter pentosaceus, Thermoanaerobacter brockii, Caloramator viberbiensis, Caloramator fervidus, and Paenibacillus spp. have been extensively investigated for improved production of ethanol at high temperatures (Arora et al. 2015). Kluveromyces marxianus emerge as the most promising yeast, which can grow at higher temperature of 45°C–52°C and can effectively produce bioethanol along with enzymatic saccharification process. In an experimental SSF study with sugarcane bagasse, Lin et al. (2013) achieved ethanol concentration and theoretical yield up to 24.6 g/l and 79%, respectively, using K. marxianus at 42°C in a rotary drum reactor.

The fermentative conversion process of biomass to ethanol can be carried out with three different process configurations such as SHF, SSF, and consolidated bioprocessing (CBP) approach. SHF is basically a two-stage process. First, using suitable enzymes, cellulose and hemicellulose materials are hydrolyzed to reducing sugars. In the next stage, such reducing sugars are then converted into ethanol by fermentation process using suitable microorganism. Both processes (i.e. enzymatic hydrolysis and fermentation) can be performed under their optimal conditions (temperature, pH, nutrient composition, and solid loading). It is considered as the main advantage of this configuration as the optimum operating conditions of both processes differ considerably. The optimum temperature for enzymatic hydrolysis is about 50°C, which differs from fermentation temperature at 28°C–37°C. Moreover, the overall performance of the process is not affected as two processes are conducted separately in different vessels and hence hydrolyzing enzymes are not influenced by the presence of ethanol. On the contrary, the biggest drawback of the process is the inhibition of enzyme activity by increasing concentrations of released simple sugars in the vessel. Such inhibition process slows down the rate of cellulose hydrolysis. There are many reports suggesting that the economic investment is normally increased for SHF due to the use of more than one vessel at different times. Another biggest drawback of SHF process is that, when it is performed under high solid loading condition, the activity of cellulolytic enzymes gets reduced due to inhibition problem. To improve the rate of cellulose hydrolysis, the utilization of various surfactants such as Tween, polyethylene glycol, or ionic liquids can play a vital role while reducing enzyme attachment to lignin (Eriksson et al. 2002). Dahnum et al. (2015) compared the performance of both SHF and SSF processes on empty fruit bunch and 4.74% ethanol was achieved at 72 h of SHF experiments, whereas 6.05% ethanol was obtained at 24 h of SSF process. In a similar comparative experimental observations, Kelbert et al. (2015) achieved 2.3-fold higher ethanol concentration from SHF process of non-nutrient-supplemented Eucalyptus globules and the result was comparatively unfavorable than the SSF process. In another comparative experimental study of SHF (Cotana et al. 2015) with the advanced SSSF process, the overall ethanol yield achieved was 13.17 g ethanol from 100 g Cynara cardunculus L., which was slightly lowered than the result (13.64 g from 100 g) achieved from the SSSF process.

SSF is a single-stage process where enzymatic hydrolysis and fermentation for bioethanol production is carried out in a combined manner in a same vessel. SSF is a more promising and advantageous approach with respect to SHF because of low production cost, less processing time, less reactor volume, higher ethanol productivity, lower requirement of enzyme, ability to overcome enzymatic inhibition by simultaneous endproduct removal, and lower requirement for sterile conditions as bioethanol is produced immediately with glucose conversion (Jahnavi et al. 2017). The saccharification process involving enzymatic reaction is normally carried out around 50°C, whereas the temperature normally maintained for ethanol fermentation is between 28°C and 37°C for most fermenting organisms. It is a difficult task to lower the optimum temperature of cellulases through protein engineering. Accordingly, high-temperature fermentation has been the major priority for the SSF process and thermotolerant yeast strains can play a major role in that process. In an experimental observation, Saini et al. (2015) showed that a thermotolerant yeast strain such as K. marxianus DBTIOC-35 was effective to produce a maximum 29 g/l bioethanol with 73% yield through SSF experiments carried out at 42°C–45°C using wheat straw as substrate. To deal with the difference between the optimal temperature of the enzyme action and microbial fermentation to produce bioethanol, different efficient and modified SSF strategies have been developed. The process configuration of SSF with short presaccharification or semisimultaneous saccharification (PSSF or SSSF) has been developed as a good alternative to enhance higher ethanol concentration, yield, and productivity (Gonçalves et al. 2014, Shahsavarani et al. 2013). Gonçalves et al. (2014) showed that a short presaccharification step at 50°C for 8 h in the SSSF process with three microorganisms (S. cerevisiae, P. stipitis, and Z. mobilis) had a positive effect to increase the overall yield of bioethanol from 79.27–84.64% to 85.04–89.15%. In a similar study with coconut fiber, Gonçalves et al. (2016) showed the hydrothermal pretreatment catalyzed with NaOH (HPCSH) followed by SSF (SSSF) process using S. cerevisiae, P. stipitis, and Z. mobilis microbes was useful to achieve high yields of bioethanol of 91.17% and 91.03%. Among the existing process of the modified SSF approach, simultaneous saccharification and cofermentation (SSCF) has been most feasible approach to convert both hexose and pentose sugars to increased concentration of lignocellulosic bioethanol. For the bioconversion of both types of pentose and hexose sugars in SSCF processes, ideal coculture fermentation with combinations of suitable microorganisms, which do not effect each other’s metabolic activities, are mainly required (Chandel et al. 2011, Singh et al. 2014). Singh et al. (2014) performed sequential coculture fermentation system with P. stipitis and Z. mobilis in a 7 l bioreactor and the average ethanol yield and productivity were 0.453 g/g and 1.580 g/l h from kans grass-based substrate. Chandel et al. (2011) made efforts toward coculture fermentation system with P. stipitis and thermotolerent S. cerevisiae with an appropriate ratio. The achieved yield and concentration of ethanol were 0.49 g/g and 15 g/l from wild sugarcane materials and the results were comparatively better than the reports where the same microorganisms were used separately for the same purpose. There are some studies of novel SSCF processes (Olofsson et al. 2010, Zhang and Lynd 2010) where the combination of enzyme and xylose feeding strategies is implemented with recombinant xylose fermenting strain to achieve high ethanol titers. Among all existing process configurations of SSF, repeated-batch fermentation with immobilized microorganism and fed-batch SSSF approach are the other two promising and advanced approaches. SSF with repeated-batch fermentation with immobilized microorganism is comparatively disadvantageous as it is time consuming and, due to transport limitation, conversion efficiency will be comparatively less, whereas fed-batch SSSF scheme largely reduces time and costs associated with the process (Hasunuma et al. 2013). To produce concentrated bioethanol at the end of the process, high substrate concentration in terms of solid loading is useful and it will be instrumental to reduce the operational cost and energy consumption during the final purification stages. Again, microbial cell viability and solid loading are totally correlated with cellulase dosages (Tomás-Pejó et al. 2008). To deal with the proper enzymatic dosage and suitable solid loading, appropriate feeding strategy for enzymes in fed-batch SSF appears to be effective to achieve higher ethanol yield (Menon and Rao 2012). Lu et al. (2010) achieved ethanol concentration of 49.5 g/l from 30% cumulative and time-dependent insoluble solid feeding in fed-batch enzymatic hydrolysis and fermentation approach. A similar experience of increasing ethanol yield was obtained by Hoyer et al. (2010) when pretreated spruce substrates as feed was added manually at different time intervals of fed-batch SSF process. High substrate concentration during fermentation became inhibitory for microbial growth and it leads to low ethanol yield. To deal with such problem, fed-batch SSSCF or PSSCF could be the best possible solution for ideal substrate conversion. Lu et al. (2013) clearly showed how substrate feeding strategies in different time intervals of enzymatic prehydrolysis periods in fed-batch SSSF process were helpful to achieve ethanol concentration of 39.4 g/l. Fed-batch SSSF process is efficient in achieving high ethanol concentration, but the effectiveness can be further strengthened through integrating cofermentation strategies into it. The concept of such advantageous fed-batch SSSCF approach is represented in Figure 4. Considering all existing process configurations of SSF approaches for lignocellulosic bioethanol production, the fed-batch SSSCF approach with proper enzymatic dosage and suitable solid loading is the most advanced biorefinery approach, which has not been reported so far.

Figure 4:

Concept of advantageous fed-batch SSSCF approach for lignocellulosic bioethanol production.

To improve microbial performance for efficient cellulose hydrolysis and increasing energy efficiency during fermentation process, enzyme production, saccharification, and fermentation can be further consolidated into a single process. Such single-stage CBP has been increasingly recognized as the most potential and promising strategy to reduce costs and environmental impacts of the bioethanol production process (Tomás-Pejó et al. 2008, Hasunuma et al. 2013). The biological conversion of pretreated lignocellulosic substrate by one or more efficient organisms into bioethanol while hydrolyzing biomass with enzymes, produced on its own and fermenting those material, could offer a large number of advantages while promoting lignocellulosic bioethanol closer to market (Olson et al. 2012). The major advantages of CBP process include the reduction of capital investment, elimination of requirements associated with enzyme production, simplification of total reactions, improving hydrolysis efficiency, and reduction of contamination risk (Hasunuma and Kondo 2012, Hasunuma et al. 2013). The factors that make CBP technology more attractive are (1) suitability of high levels of enzyme production and enzymatic hydrolysis rate, (2) production of minimum level of toxic components inhibitory for CBP microorganisms, (3) cofermentation of hexose and pentose sugars while minimal degradation of carbohydrate fractions, and (4) tolerating high levels of endproduct (Parisutham et al. 2014, Den Haan et al. 2015). The major challenging feature of CBP is saccharification and fermentation usually carried out at common higher temperature and such drawback could be dodged using thermophilic CBP microbes. The capabilities of various fungi, such as Aspergillus, Fusarium, Monilia, Rhizopus, Neurospora, Trichoderma, Phlebia, and Mucor spp. as ideal CBP microorganisms have been reported for the improved conversion efficiency of cellulose to ethanol (Anasontzis et al. 2011, Hasunuma et al. 2013). More specifically, among thermophilic CBP microbes, K. marxianus appear to be the most promising microorganisms, which can grow well at higher temperature (45–52°C) and can efficiently produce ethanol with higher titer (Hasunuma and Kondo 2012). Hong et al. (2007) identified K. marxianus as a high-efficiency recombinant strain that can directly produce 43.4 g/l ethanol from 100 g/l cellobiose at high temperature (45–50°C) CBP process.

Although engineering yeast, filamentous fungi, and bacteria have been used significantly as CBP microbes, process-ready improved microorganisms for the production of higher level of bioethanol is still under development. Among common category 1 CBP microorganisms, C. thermocellum (Lynd et al. 2005), T. mathranii (Taylor et al. 2009), and cellulolytic fungi such as T. reesei (Huang et al. 2014) are already recognized, but to achieve high yield and concentration of bioethanol, still intensive research is required in this field. Xu et al. (2009) acknowledged that T. reesei as an efficient producer of cellulase can be used for bioethanol production through CBP approach with low ethanol yield. For improving performance of overall CBP process with both cellulose and hemicellulose degradation and conversion in to bioethanol, Fusarium oxysporum is considered as the best filamentous organism for CBP process. Paschos et al. (2015) showed the efficiency of F. oxysporum in improved ethanol production process with final ethanol concentration of 58 g/l. In a case study, Mattila et al. (2017) reported the direct bioconversion of lignocellulosic waste materials by phlebioid fungal species without involving any pretreatment of substrate, whereas achieved bioethanol yield (10.4%) and concentration (5.9 g/l) were comparatively low. In another CBP-based experimental observation with Phlebia sp, Kamei et al. (2012) obtained a similar low ethanol concentration of 8.4 g/l from 20 g/l unbleached hardwood kraft pulp. To make CBP-based bioethanol production industrially acceptable, there must be a demand to improve the performance of CBP process with suitable process configuration involving ideal microorganism. Among advanced CBP processes, cell recycling batch fermentation (CRBF) as proposed by Matano et al. (2013) was instrumental to achieve enhanced bioethanol concentration with reduce time and costs. The concept of advanced CRBF under CBP is represented in Figure 5.

Figure 5:

Concept of advanced CRBF strategy under CBP for lignocellulosic bioethanol production.

The fermentative biochemical route for production of biofuel is considered advantageous, as it has a greater probability of cost reduction compared to thermochemical route. Chemicals such as phosphorous and nitrogen generated during thermochemical route are normally released to the atmosphere, which leads to additional eutrophication and acidification (Limayem and Ricke 2012). On the contrary, the biochemical route exhibited the requirement of higher quantity of water than the thermochemical route. In the near future, water availability could become a major concern if sustainable water utilization practices are not carried out in agricultural field. Even the utilization of higher water content causes lower concentration of fermentable sugars in prehydrolyzates, which finally leads to low concentration of ethanol after fermentation. Most of the conventional methods cannot concentrate sugar solution while separating inhibitors. Cell recycle MBR system can overcome this problem and enable high cell densities inside the reactor. High cell density fermentation system allows significantly faster fermentations than lower cell concentration system. Hence, higher production goals can be achieved in a comparatively smaller reactor and in shorter time duration (Westman and Franzén 2015). At the same time, such bioreactor allows easier separation of endproducts, inhibitors that can strongly inhibit the fermentation system. The implementation of such kind of process design will definitely be helpful to decrease the production cost of ethanol.

2.4 Conventional separation process

Distillation is the conventional separation and purification process of bioethanol from aqueous fermentation broth. Based on the different volatilities, ethanol is being separated from aqueous mixture by fractional distillation. Based on the boiling point of ethanol (78.3°C), which is lower than the boiling point of water (100°C), ethanol will be converted to steam before water during boiling and can be separated by condensation. Continuous distillation column system with multiple effects is normally used in biorefineries and in large-scale industries (Kent et al. 1994). Most volatile components from the liquid mixture are separated from the top of the column during distillation. The main limitation of the distillation process is to separate the components with the same boiling point from liquid mixture. Industrially applicable anhydrous ethanol, which contains 99.5% ethanol by volume and water content not more than 0.5%, can be produced through some advanced distillation processes such as adsorption distillation, azeotropic distillation, diffusion distillation, extractive distillation, membrane distillation, and vacuum distillation (Kumar et al. 2010).

4.1 Lignocellulosic substrates for membrane technology-based bioethanol production

The most commonly used lignocellulosic carbon sources for bioethanol production process are rice straw, wheat straw, wheat bran, rice husk, banana waste, newspaper, sugarcane bagasse, and grasses. The compositions in terms of cellulose, hemicellulose, and lignin in those lignocellulosic feedstock were represented by Menon and Rao (2012) and Van Dyk and Pletschke (2012). The substrates that are mainly used for membrane technology-based bioethanol production are mainly pure and expensive carbohydrate sources such as glucose, fructose, sucrose, and lactose (Cheryan and Mehaia 1983, Hoffmann et al. 1985, Melzoch et al. 1991, Ding et al. 2012), industrial waste (whey and molasses), and sludge-based materials (Mehaia and Cheryan 1990, Tin and Mawson 1993, Kaseno and Kokugan 1997, Lee and Yeom 2007). Time-consuming and expensive pretreatment processes are the main prerequisites to avoid membrane fouling while working with waste materials. Specific agricultural lignocellulosic waste substances such as rice straw and wheat straw (Ishola et al. 2013, 2015a,b, Ylitervo et al. 2014, Zahed et al. 2016) have been mainly exploited in experimental studies of membrane reactor facilitated lignocellulosic bioethanol production process. In some experimental studies (Lee et al. 2000, Ylitervo et al. 2014), general lignocellulosic hydrolyzate materials were investigated in MBR-based bioethanol production process. Information related to specific lignocellulosic substrates used until now for MBR facilitated bioethanol production, maximum product concentration or productivity achieved, and membrane materials or modules used in those studies are represented in Table 2. Among the diverse lignocellulosic biomaterials, so many biomaterials such as perennial grass (e.g. kans grass and S. spontaneum) and coconut mesocarp, which are believed to be novel and potential nonfood lignocellulosic feedstock with great prospective for bioethanol production, have not been implemented until now for MBR-based bioethanol production process.

4.2 Membrane-based advanced enzymatic saccharification

Cellulases, which are considered as the most important and key enzymes for saccharification processes, are normally inhibited by the products of the enzyme substrate reactions. Utilization of the enzymes during saccharification process in the free form is not suitable due to its expensive nature. The alternative solution of such problem will be to recycle and reuse the enzymes. The retention of enzymes in free form rather than immobilized form is implemented more (Ishola et al. 2013, 2015a), as the immobilized form restricts its ability to penetrate the solid substrate; hence, expected conversion efficiency will be comparatively less (Al-Zuhair et al. 2013). Similarly, batch mode of enzymatic hydrolysis process for cellulose conversion is often restricted by product inhibition problem caused by newly generated hexose and pentose sugars (Andrić et al. 2010). To overcome such problems, research efforts have been directed toward the instant removal of simple sugars after the enzymatic saccharification process to maintain and even increase the conversion rate (Abels et al. 2013). To reuse and recycle the soluble enzyme solutions without using any expensive protein separation unit, MBR, specifically UF-based MBR system, can be successfully used to separate and reutilize the enzymes. Such design facilitates in the minimization of substrate utilization, instant separation of excess enzymes, and reutilization of such enzymes after enzyme substrate reactions. The basic configuration of such MBR system for enzymatic saccharification is represented in Figure 6. Such UF membrane-integrated continuous stirred tank bioreactor system was studied by Henley et al. (1980) for enzymatic saccharification. The investigations were carried out with four different lignocellulosic substrates and cellulase enzymes from Trichoderma spp. at a temperature of 50°C in an acetate buffer solution of pH 4.8. Maximum conversion efficiency was achieved as 70% with powdered cellulose, which was used as one of the substrates among those four substrates. Comparatively less conversion efficiency of 53% was achieved by Bélafi-Bakó et al. (2006) while working with tubular MBR system. Commercial enzyme solution from T. reesei was used in this study at 50°C and at pH 4.8 in sodium acetate buffer solution. A similar type of UF-based MBR system was studied by Kinoshita et al. (1986) with cellulase enzymes obtained from Sporotrichum cellulophilum. Experimental investigations were carried out with increasing permeate flow rates and maximum glucose recovery was achieved at permeate flow rate of 15 ml/h up to 40 h of operation. Productivity was decreased after 40 h of operation due to the presence of proteases with crude enzymes, which finally leads to the deactivation of the enzymes. Similar correlation between glucose removal with increasing permeate flux was established by Alfani et al. (1982). Maximum conversion rate was increased from 2% to 5% at 25 h of operation interval with increasing permeate flux. Increasing permeate flux always leads to quick depletion of reaction volume inside of the reactor system and limited recovery of simple sugars. Such drawback could be dodged with an instant supply of fresh substrates and enzymes in fed-batch or continuous mode. Andrić et al. (2010) clearly explained that enzymatic conversation rate can be increased up to a certain extent with the addition of feed in increasing dilution rate. In the same direction, Kinoshita et al. (1986) observed that enzymatic conversion rate did not improve above feed flow rate of 15 ml/h when experiments were conducted with increasing feed flow rate of 3–30 ml/h. A slight contradictory result was obtained by Yang et al. (2006) when increased hydrolysis rate and increased glucose concentration were observed between low dilution rates of 0.057–0.075/h, whereas comparatively lower glucose concentration was achieved at higher dilution rates of 0.075–0.25/h. Such experimental observations clearly suggest that increasing dilution rates beyond a certain limit can influence adversely in conversion rates and in achieving maximum glucose concentration.

Figure 6:

Basic configuration of MBR system for the enzymatic saccharification of lignocellulosic substrate.

It appears that enzymatic conversion rates are always positively influenced with feeding strategies of cellulosic substrates in fed-batch fermentation processes with suitable dilution rates (Andrić et al. 2010). Semicontinuous feeding strategies of substrates with initial enzyme loading and intermittent product removal have been normally adopted in advanced fed-batch enzymatic saccharification processes. Authors such as Kuhad et al. (2010), Soares et al. (2017), and Sugiharto et al. (2016) adopted different feeding strategies of lignocellulosic substrates in normal bioreactor facilitated fed-batch saccharification processes. To enhance the saccharification efficiency, increasing solid loading approach [from 12% (w/v) to 33% (w/v)] were adopted in high solid loading fed-batch saccharification processes (Gao et al. 2014). The best result was optimized when saccharification process was initiated with 12% (w/v) solid loading and 7% fresh solids were gradually introduced at 6, 12, and 24 h to achieve final solid loading of 33%. However, among the existing studies in the same research field, only few researchers highlighted the advantages of membrane-based fed-batch saccharification strategies in their observations. Ohlson et al. (1984) achieved improved initial hydrolysis rate and improved degree of enzymatic conversion from 40% in batch mode to 95% in semicontinuous fed-batch hydrolysis mode in an UF MBR system. Kinoshita et al. (1986) observed the lowering actual glucose concentration during fed-batch operations and it was termed as negative “dilution effect”. MBR-based continuous hydrolysis process with continuous product removal is another advantageous strategy for improving saccharification efficiency. Advantages of such MBR-based continuous operations in lignocellulosic saccharification processes were already highlighted by Ghose (1969) 49 years ago. Ghose and Kostick (1970) explained in their report that, due to the rapid transport properties of reaction products, high conversion efficiency of 77% was achieved in continuous MBR system. With emerging features and advances in membrane materials, such MBRs can be scaled up from 300 ml to 10 l to facilitate saccharification processes, but still such kind of large-scale membrane processes deserves vigorous technoeconomic evaluation before implementation. UF membrane-integrated bioreactor system with working volume capacity of 2.5 l was investigated by Gan et al. (2002) for the enzymatic degradation of α-cellulose. Selection of suitable membranes with appropriate molecular weight cutoff facilitates the permeation of simple sugars such as glucose (molecular weight of 180 g/mol) and separation of fungal cellulase enzymes ranges from 35,000 to 65,000 g/mol (Andrić et al. 2010). Normally, UF membranes with MWCO of 10 kDa are more useful in the separation of glucose during hydrolysis and in achieving high conversion rate (Ohlson et al. 1984, Gan et al. 2002). Ohlson et al. (1984) pointed out that such MBR system is also successful for the enzymatic hydrolysis of lignocellulosic substrate without subsequent washing. Membrane cleaning and back-washing of membrane to remove deposited organic or inorganic material from its surface is one of the most important criteria to maintain the permeability of the membrane and it also determines the lifetime of a membrane. Membrane cleaning should be periodically done when permeate flow drops nearly about 10% and normalized pressure drop increases about 10%–15%.

4.3 MBRs for fermentative production of lignocellulosic bioethanol

The most effective strategy for high titer bioethanol production is membrane-based cell recycle fermentation system. Such cell recycle MBR system plays the most important role in fermentative production of lignocellulosic bioethanol through an ecofriendly and sustainable route (Saha et al. 2017b). Cell recycle system not only reduces the time and costs associated with inoculum preparation but at the same time also ensures high biomass density fermentation system (Matano et al. 2013). In certain cases, high-density fermentation system leads to the formation of high product concentration while fermenting inhibitory hydrolyzate. Another advantage of such membrane cell recycle fermentation system (MBR) is that the dilution rate is not dependent on the growth rate of the microorganism (Ylitervo et al. 2013). High dilution rates lead to maximum bioethanol productivity and yields (Ylitervo et al. 2014). Low dilution rates are the only possible operating conditions in which traditional continuous cultivation systems without cell recycle are successful (Brandberg et al. 2008). Taherzadeh et al. (2001) observed that, at low dilution rate of 0.1/h, continuous fermentation of diluted acid hydrolyzate was successfully carried out by yeast strains, whereas at comparatively higher dilution rate of 0.2/h fermentation fails. During lignocellulosic fermentations, separation of microbial cells by conventional cell collection procedures from unutilized lignocellulosic residue is a difficult task. Considering this difficulty, Matano et al. (2013) developed two-phase separation of lignocellulosic residues from yeast cells in high solid loading lignocellulosic bioethanol production process. High solid loading with high cell density microbial culture system in MBR can lead to membrane fouling, which finally declines the flux during long operational time and hence restricts its credibility in large-scale commercial applications. Soluble microbial products produced during fermentation normally get adsorbed and deposited on the membrane surface to form a biofilm layer. Several strategies are present to restrict membrane fouling problem, and among them, the selection of proper membrane material and proper operational conditions based on the application helps to mitigate the problem up to large extent. Operational conditions including low transmembrane pressure, low shear rate, and high cross-flow sweeping action in cross-flow operations rather than dead-end operations help to control membrane fouling (Wei et al. 2014). Basic membrane operations in cross-flow and dead-end modes are represented in Figure 7. In dead-end mode, feed flows directly perpendicular to the filter medium and resistances of the liquid flow increase with rapid cake formation. In cross-flow mode, feed flow is tangential to the filter medium. To avoid membrane fouling, Ishola et al. (2015a) implemented a novel concept of forward flow operation for every 4 min followed by the reverse flow of 1 min in 2.5 l Minifors, Switzerland MBR. Stepwise addition of the substrate was maintained in such advanced membrane-based SSFF strategies where both glucose and xylose were converted to bioethanol. Similar cofermentation strategy was also adapted by Mahboubi et al. (2016) in integrated permeate channel (IPC) MBR system. Such continuous membrane-based fermentation system comes up with total consumption of glucose, 83% consumption of xylose, and 83% theoretical bioethanol yield. While producing 99.28±0.68 g/l of bioethanol from sweet sorghum stem juice, Thani et al. (2017) clearly indicated that such kind of membrane-based continuous operations plays the most instrumental role in high titer production. While operating with lignocellulosic substrate such as sugarcane bagasse in similar membrane-integrated continuous fermentation system, comparatively remarkable bioethanol concentration (43.2 g/l) was achieved by Saha et al. (2017a). In this study, fermentation was carried out in 20 l pilot plant bioreactor system integrated with first-stage MF unit and second-stage NF unit. The concept of reverse MBR (rMBR) where cells are immobilized between immersed membrane layers and separated from actual substrate solution is recently introduced by Ishola et al. (2015b). Cells are encapsulated between synthetic membrane layers in the form of compact multilayer membrane column or IPC flat sheet membrane (Doyen et al. 2010). Such kind of advanced IPC rMBR configuration was used for bioethanol production from wheat straw hydrolyzate using recombinant yeast (Ishola et al. 2015b). Improved simultaneous utilization of xylose and glucose with in situ detoxification of inhibitory components was facilitated by such novel configuration.

Figure 7:

Mechanism of membrane filtration operations: (A) cross-flow mode and (B) dead-end mode.

4.4 Membrane processes for pretreatment and concentrating substrate solution

All existing studies of lignocellulosic bioethanol production suffer from low concentration of fermentable sugars and hence production of low concentration of ethanol, which may lead to high operational cost and energy consumption during subsequent purification stages. The conventional methods implemented for concentrating sugar solutions or detoxification include solvent extraction, evaporation, activated charcoal adsorption, and ion exchange procedures (Wei et al. 2014). High processing costs, generating additional waste materials, requirement of long processing time (Wei et al. 2014), and chances of losing sugars are the major disadvantages associated with such conventional processes (Parawira and Tekere 2011). Membrane-based selective separation systems can solve these issues while removing the inhibitors. To enhance the effectiveness of fermentation process and to achieve concentrated bioethanol, sugar solutions after saccharification process can be preconcentrated using suitable membranes, specifically NF membranes before fermentation. At the same time, the NF membrane scheme will be useful in the continuous separation of products during fermentation and helps to avoid product inhibition problem.

Similarly, membrane separations are also useful during the initial stages of pretreatment. Pentose and hexose sugars are mainly the endproducts after saccharification processes and there are certain microorganism that can ferment separately or both hexose and pentose sugars. As a result, separations of such hexose and/or pentose sugars sometimes become necessary during pretreatment process to accelerate fermentation efficiencies. In such cases, NF-based membrane separations have been successfully implemented for the separation of xylose from glucose before fermentation (Sjöman et al. 2007, Roli et al. 2017). Sometimes, side products such as furfural, which is toxic to fermentative microorganisms, are generated during various pretreatment processes. To achieve the maximum productivity of lignocellulosic bioethanol, it is necessary to remove such toxic materials before fermentations. The separation of toxic materials from fermentation broth through pervaporation membranes was highlighted by some researchers. Similar experiments were conducted by Ghosh et al. (2010, 2007) for the separation of furfural material from its mixture at different concentrations and at different temperatures through casted flat-sheet pervaporation membranes. The separation of toxic components such as phenols and furfural from biomass hydrolyzates via pervaporation mechanism was also studied by Sagehashi et al. (2007). Experiments were conducted with varying the thickness (40–200 μm) of silicone rubber pervaporation membrane. It was found that keeping the separation factor constant for every experiment at constant 60°C temperature, total permeate flux was inversely proportional to the membrane thickness. In an another case study, Cai et al. (2013) implemented polydimethylsiloxane (PDMS) made pervaporation membrane for the removal of furfural, which originated from diluted acetic acid pretreated sweet sorghum bagasse hydrolyzate. The process was quite effective to bring down the furfural concentration from 10 to 0.6 g/l.

4.5 Advanced membrane separation processes for recovery of bioethanol

There are some specific membranes separation processes that have been extensively studied in the area of bioethanol separation from fermentation broth in the most purified form and made their way into commercialization successfully. In typical lignocellulosic bioethanol production process, most of the fermentative microorganisms can withstand bioethanol concentrations up to 10 wt%. To overcome this limitation, the continuous removal of the bioethanol from fermentation broth became essential. To make the overall process energy saving and cost effective, it also necessary that produced bioethanol must be concentrated before the next refining process (Wei et al. 2014). Distillation is considered as a conventional method for the removal of ethanol from fermentation broth in concentrated form. But unluckily, the overall prospect of the process did not become attractive, as the energy requirement for the recovery of low concentration product is comparatively higher and it became difficult to directly integrate such process with fermentation because the operational temperature is much more higher (Wei et al. 2014). Therefore, such disadvantages associated with conventional distillation process lead to the selection of membrane-based advanced processes for the direct recovery of bioethanol from fermentation broth. Among these advanced processes, membrane distillation, membrane pervaporation, and membrane extractions have been extensively studied and some of them made their way into commercialization. Based on the difference of partial pressure of ethanol and water, ethanol vapor get separated through membrane distillation, whereas concentration difference between feed and permeate causes the main driving force for mass transfer in pervaporation and membrane extraction (Abels et al. 2013). In membrane extraction, two immiscible liquids are allowed to contact with each other in close vicinity of the membrane and low concentrated solutes are allowed to transfer through the membrane surface. Additional resistance of the membrane is considered as the limitation of the membrane extraction process (Abels et al. 2013). On the other side, phase change is required for membrane pervaporation mechanism to transfer the feed at the permeate side in the form of vapor phase. Mass transfer through this membrane process is not regarded as pressure driven and it can be explained with the solution-diffusion model (Wijmans and Baker 1995). Membrane pervaporation process is considered advantageous, as it allows the separation of even very low concentration of ethanol from aqueous solution. It can be implemented successfully in conventional process design, as it can minimize energy consumption of thermal unit operation such as distillation or dephlegmation.

Membrane distillation is already a well-developed process to separate bioethanol from fermentation broth. It is basically a thermally driven process where porous hydrophobic membrane supports water vapor transport to separate it. Such membrane-based process requires latent heat of evaporation to reach vapor-liquid equilibrium for separation and phase change from liquid to vapor. Temperature difference across the membrane comes up with vapor pressure gradient, which mainly acts as the driving force. Membrane distillation can be carried out at much lower temperature than conventional thermal distillation, as the driving force for membrane distillation does not involve directly thermal sources (Onsekizoglu 2012). So far, research focuses have been mainly made toward the identification of suitable membrane material for membrane distillation and selecting ideal process configurations. Polymeric membranes are the most preferred membranes rather than inorganic membranes for such process and have gained much more attention due to their possibility to modulate the intrinsic properties (Onsekizoglu 2012). Due to low surface tension values, polymeric membranes such as polytetrafluoroethylene, polypropylene (PP), and polyvinylidenefluoride (PVDF) are most commonly used for such process. Among different modules, plate and frame, spiral wound, tubular, capillary, and hollow-fiber membrane modules are usually implemented by different researchers (Onsekizoglu 2012). Ethanol productivity can be improved when membrane distillation operation for ethanol recovery can be coupled with fermentation process to remove the product in the form of inhibitor. Gryta et al. (2000) initially introduced the concept of direct contact membrane distillation (DCMD) for ethanol synthesis from lactose solution inside a Biotron reactor. The efficiency of ethanol production was improved through such DCMD, as it allows the selective removal of fermentation products from the reactor. Considering the partial pressure of ethanol, which is higher than the water, ethanol vapor can get transferred preferentially through the membrane pores. Membrane distillation is also considered as a sustainable alternative green technology approach because the conventional distillation technique to separate ethanol from water requires high temperature, electricity, and labor. It was reported that ethanol production can be increased by 15.5% while adopting continuous process, lowering the osmotic pressure in the fermentation broth, decreasing glycerol synthesis level, and increasing yeast cell concentration in fermentation broth (Lewandowicz et al. 2011). Membrane swelling can lead to increasing feed solution temperature, which finally helps in achieving increased membrane flux (Lewandowicz et al. 2011). It was found that the temperature of the feed has a large influence on the overall separation process. There are lots of experimental observations (Gostoli and Sarti 1989, Udriot et al. 1989, Tomaszewska and Białończyk 2013) where distillation is directly integrated with fermentation for better separation of ethanol. Among them, Udriot et al. (1989) achieved a remarkable 87% increase in ethanol productivity in such integrated system.

Similarly, just like membrane distillation, membrane pervaporation is another well-established strategy for the separation of alcohol from fermentation broth. The process consists of two stages. Solution is being permeated through the membrane and evaporated to vapor phase through the nonporous membrane. During the process, the upstream side of the membrane will be under ambient pressure, whereas the downstream side will be under vacuum pressure. Partial pressure difference of the components on the two sides plays the most important driving force for the process and solution-diffusion is the main mechanism involved in the process. Pervaporation membrane performance is dependent on the characteristics of the membrane such as membrane thickness and the characteristics of the feed solution such as feed solution temperature and ethanol concentration in the feed (Hui et al. 2012). A specific limitation of this process is that the process is susceptible to temperature-sensitive components such as microorganisms and requires high-temperature heat source and additional heat exchanger to recover the heat (Vane 2004). Based on the characteristics and limitations of the process, research investigations have been made for the selection of a suitable membrane, development of a long-lasting process, and better process integration. Based on the material used in the pervaporation process, pervaporation membranes are classified into two groups: hydrophobic and hydrophilic. Compared to hydrophobic membranes, hydrophilic membranes are effective enough to separate water from high concentration of alcohol (>85%; Abels et al. 2013). Due to superior temperature and mechanical stability, inorganic membranes gained lot of commercial attention. Recently, large numbers of polymeric membranes such as cellulose acetate butyrate membrane, PDMS, or polyimide (PI) membranes are commonly used for the pervaporation process. To deal with the pervaporation process with concentrated alcohol concentrations, zeolite-based membranes and/or polymeric composite membranes with polyvinyl alcohol or PI active layer can be implemented (Abels et al. 2013). It was found out that polystyrene (PS) is more hydrophobic than PDMS and also has a higher tensile and mechanical strength property, whereas some modified PDMS-PS membrane showed improved tensile and mechanical strength (Liang and Ruckenstein 1996). Huang et al. (2013) developed hydrophilic cellulose-ester pervaporation membrane coated with hydrophilic perfluoropolymer for high selectivity of water. Initially, water flux for water/ethanol mixture through cellulose-ester hydrophilic membrane was totally dependent on feed composition, and membrane performance was highly affected by swelling due to absorbed water. Pervaporation is a well-established process in the last three decades (Abels et al. 2013) and is available for commercial applications including a number of industrial processes. The biggest advantage of the process is its long-term stability, which has been reflected in different research investigations. Qureshi et al. (2001) investigated the performance of silicalite-silicone mixed-matrix pervaporation membrane when it was exposed to fermentation broth for 870 h. The integrated process was successful during 10 successive fermentations runs with continuous recycle from pervaporation unit. A similar long-term applicability of newly synthesized silicon rubber-coated pervaporation membrane was studied by Ikegami et al. (2003) for 40 h of continuous run. Another study (Nomura et al. 2002) related to the recovery of ethanol from fermentation broth by pervaporation membrane observed a linear increase of permeate flux with increasing feed concentration of ethanol. Long time operations of more than 48 h were useful to achieve ethanol concentration up to 80%, but it leads a significant decrease of permeate flux. In the same direction, Mulder and Smolders (1986) already suggested the suitability of pervaporation membrane for continuous long time recovery of bioethanol from fermentation broth. Dense polymeric silicon rubber membranes were used in this study for the separation of ethanol, i-propanol, and n-propanol from aqueous fermentation broth. To determine the effect of raw material and its pretreatment methods in the separation of bioethanol from fermentation broth through pervaporation, Gaykawad et al. (2013) conducted studies with different lignocellulosic substrates such as barley straw and wood chips. The raw materials were pretreated with varying concentrations of H₂SO₄. The maximum concentration of bioethanol achieved was 30 g/l after pervaporation operations, which were conducted with commercial PDMS membrane. The economic feasibility of pervaporation operations in commercial-level ethanol production plant was established by O’Brien et al. (2004). The concentrated ethanol concentration achieved was 150–170 g/l after pervaporation operations, conducted with a pervaporation membrane having cross-sectional area of 0.22 m². This process also has some specific drawbacks, as operations of such integrated processes at higher temperature could lead to fouling problem (Abels et al. 2013). In the same review, the concept of a useful membrane-integrated scheme consisting of MF or UF separation in the first stage to separate cells and pervaporation system containing both hydrophobic and hydrophilic membranes in the second stage to recover pure ethanol was represented. Hydrophobic membrane was useful to separate low concentration of ethanol from fermentation broth in the first stage, whereas hydrophilic pervaporation unit leads to maximum dehydration of ethanol. Huang and Vane (2006) presented a similar integrated strategy for energy-efficient recovery of bioethanol by a coupled dephlegmation/pervaporation process. In this process, two pervaporation units were coupled with one dephlegmator to achieve final ethanol concentration up to 99 wt%.

Among all existing separation processes, membrane extraction has also emerged as a novel and most promising technology that offers numerous advantages. In membrane extraction process, large surface area is normally provided by MF or UF membrane to bring an aqueous mixture in close vicinity to the extractant. Diffusive transportation of the product from aqueous solution into the extractant is facilitated through membrane pores. Hollow-fiber membrane modules and membranes made specifically by PVDF or PP are commonly applied in this operation (Abels et al. 2013). Separation process through membrane extraction offers better transport of the product through diffusion mechanism and stability in continuous operations due to the modularity of membrane elements (Abels et al. 2013). In an experimental investigation, Vatai and Tekič (1991) implemented artificial kidney-based permeable membrane and sec-octanol solvent for the membrane extraction of bioethanol from water. The conclusions made from this study were (1) as hollow-fiber modules have a large surface area, microporous hollow-fiber membrane may provide high mass transfer rate per unit volume, and (2) limitation of conventional liquid separation processes such as flooding and loading can be easily overcome by microporous membrane-based dispersion free solvent extraction mechanism. Such membrane operations have been largely implemented in the selective removal of specific components such as propionic acid, acetic acid, lactic acid, n-hexane, acetic acid, and gibberellic acid from fermentation broth, but still now no products have been commercialized and its applicability is still limited in the bioethanol separation field (Abels et al. 2013). Extraction of ethanol from fermentation broth through this technique was studied by Chang et al. (1992). In this study, diafiltration extraction membrane module with dibutylphtalat (DBP) extractant was used in the first stage of the membrane-integrated setup and maximum productivity achieved was 20 g/l h.

4.6 Overall analysis: membrane processes in lignocellulosic bioethanol production

Membrane separation processes are implemented successfully in different stages of bioethanol production process. More specifically, it is implemented in concentrating sugar solution before fermentation, separation of enzymes, separation of microorganism, and finally for separation and purification of bioethanol. There are number of studies where membrane-based preconcentration experiments were conducted with xylose solutions. NF membrane separations are most commonly used at lower pH range of solutions and overall rejection efficiencies were achieved more than 85% in all existing studies. Surprisingly, 99.5% rejection efficiency was achieved in a specific NF-based preconcentration experiments where the pH of the solution was maintained at 10 (Qi et al. 2011). Pressure-driven reverse osmosis (RO) process was also used to concentrate the solution where 99.67% rejection efficiency was achieved (Zhou et al. 2013). Similarly, like concentrating sugar solution, separation and recovery of enzymes were specifically carried out through UF membranes by different researchers. Mostly flat-sheet membranes are used in such operations. Even for the separation of bioethanol from fermentation broth, different membrane-based processes have been prioritized. Overviews of such studies related to the separation of different compounds through membrane separations are presented in Table 3.

In spite of major advantages associated with membrane technology to facilitate different stages of bioethanol production process, there are some key issues that always influence membrane performance. Concentration polarization and membrane fouling are two major concerns that affect membrane permeate flux with time. Operational feasibility of MBR does not directly support for high solid loading saccharification process as a result of low glucose concentration obtained in permeate and it finally leads to low percentage of ethanol conversion during fermentation process. High solid loading enzymatic saccharification process for lignocellulosic biomass in large-scale MBR encounters poor mixing, high viscosity, and mass transfer limitation problems (Andrić et al. 2010). Another drawback of such high solid loading enzymatic saccharification process in MBR is membrane fouling problem with untreated lignocellulosic substrate including lignin and partially recalcitrant cellulose (Andrić et al. 2010). However, most of those difficulties related to membrane operations can be resolved by effective process configuration and design configuration of membrane. Enzyme activity loss due to concentration polarization problem can be solved through the optimization of hydrolysis reaction rate and selection of suitable dilution rate (Andrić et al. 2010). It appears that enzymatic conversion efficiency inside MBR can be improved through fed-batch feeding of cellulosic substrate with intermittent product removal (Ohlson et al. 1984). Intermittent product removal through membrane processes during saccharification and fermentation processes enables to overcome product inhibition problem. On the other side, long-term performance of membrane in fouling free conditions can be ensured by virtue of sweeping flow in flat-sheet cross-flow modular operation of membrane (Dey and Pal 2012) or through sMBR system, which facilitates such operation with less energy involvement (Ylitervo et al. 2014).