Biohythane and organic acid production from food waste by two-stage anaerobic digestion: a review within biorefinery framework

Abstract

The two-stage anaerobic digestion (AD) is gaining popularity because of the process stability and possibility of recovering multiple-resources such as biohydrogen and organic acids from the first stage dark fermentation (DF) and methane in the AD as the second stage while treating the organic waste. As the performance of two-stage processes is influenced by the type of substrate and operational conditions, there have been several experiments at laboratory and pilot scales to determine the optimum conditions. The main objective of this review is to provide an updated overview of advancements in biohythane and organic acids production from food waste (FW) in the two-stage DF-AD process. Likewise, this work also provides an insight into the economic and future prospective of utilizing organic acids for different biochemicals such as polyhydroxyalkanoates, polylactate, and microalgal biomass production. The integration of optimum operational parameters, pretreatment methods, types of bioreactors is essential in combined DF-AD processes. The parameters and reactor configuration have to be optimized depending upon the targeted end-products. More research into the techno-economic analysis of different bioreactor configurations for long-term operations in an integrated DF-AD process with FW as a feedstock is needed to realize its viability for commercial application.

Graphical abstract

Introduction

With the increase in energy demand, depletion of fossil fuel, and growing concern over global warming, the interest in green technologies for resource recovery is increasing. The current global energy demand is expected to double by 2050. More than 80% of the current energy demand is fulfilled by fossil fuels which are limited resources. Moreover, the by-products from burning fossil fuels are the major source of greenhouse gas emissions. The need for renewable sources to produce bioenergy and biochemicals that have positive environmental impacts is driving the bio-based economy (Cherubini et al. 2009; Bastidas-Oyanedel et al. 2019). The bio-based economy promotes the technologies that recover energy and chemicals from biomass through the biological degradation process.

Anaerobic digestion (AD) is one of the widely used biological processes for the production of renewable energy as biogas and nutrients from organic waste (Jain 2019). The benefits of converting the combustible methane (CH₄) gas to heat and electricity through combined heat and power (CHP) plant and the use of digestate as fertilizer has made AD a popular technology. However, these factors may not upgrade both material recovery efficiency and the economic viability of commercial biogas facilities (Kaparaju et al. 2009; Sawatdeenarunat et al. 2015).

The pursuit of optimization of the AD process for multiple-resource recovery has led to realizing a two-stage AD process. The two-stage AD process differs from the single-stage AD in terms of microbial pathways, growth kinetics, and environmental condition for acidogens and methanogens which is carried out in two separate bioreactors. Besides, several studies show that the two-stage process is much more beneficial than single-stage processes in terms of performance, efficiency, and stability (Liu et al. 2006; Park et al. 2010; Luo et al. 2011; Ghimire et al. 2021). Moreover, the fermentation time and reactor volume are also reduced in a two-stage AD process with possibilities for multiple-resource recovery. As reported by Roy and Das (2016), the total gaseous energy recovery from the starchy wastewater for the two-stage process was 53.6% whereas only 28% of gaseous energy was recovered from the single-stage biohydrogen (H₂) production. The two-stage AD process also shows stability in the long run which opens up possibilities for upscaling (Cavinato et al. 2012; Massanet-Nicolau et al. 2015).

The two-stage AD process simultaneously produces H₂ in the first stage and CH₄ in the second stage. The combination of H₂ (10–25% v/v) and CH₄ (75–90% v/v) is termed biohythane which has been considered a high-grade fuel, better than methane in terms of flammability, flame speed, and easier ignition with less energy input (Bolzonella et al., 2019). In addition, the by-products from the two-stage AD process such as Volatile Fatty Acid (VFA), Lactic Acid (LA), and alcohol are useful biochemicals (O-Thong et al. 2018). The two-stage AD has a good potential to become a sustainable biorefinery approach for multiple fuels and biochemicals production. Various processes such as dark fermentation (DF), photo-fermentation (PF) within the two-stage biorefinery process are available for H₂ production. In particular, DF is the most studied and promising technology because it has the potential to produce H₂ along with biochemicals from a wide variety of feedstock (Nasr et al. 2012; Alibardi and Cossu 2016). Furthermore, combining DF with AD can be an appropriate option to create H₂ and CH₄, thus, biohythane (Meena et al. 2020).

Current research interests in “biorefinery” are growing as shown by an increasing number of publications of 2,153 scientific articles and book chapters in the last decade (2010–2020) compared to 402 numbers (before the decade) in Scopus. The research and commercial interest in anaerobic biorefinery is increasing due to the versatility of the process for the production of multiple fuels and chemicals (K. C. et al. 2015; Agler et al. 2011; Tedesco and Stokes 2017). As such DF-based biochemical process can be a promising route for achieving the bio-based products and energy (Clomburg and Gonzalez, 2013; Bastidas-Oyanedel et al. 2015; Motte et al. 2015; Sarma et al. 2015; Luongo et al. 2017; Nizami et al. 2017; Moscoviz et al. 2018).

Conversion of Food waste (FW) to produce bioenergy (biofuels) and by-products (fertilizers, chemicals) has been considered as one of the efficient approaches for not only energy production but also solid waste management (Dahiya et al. 2018; Paritosh et al. 2017; Strazzera et al. 2018). About 57% and 53% of the waste composition is organic waste in low and middle-income countries, respectively (Kaza et al. 2018). The major portion of organic waste is FW. Globally, FW amounts to 1.6 billion tonnes and much of it ends up in landfill sites leading to the emission of greenhouse gases (GHGs). FW is estimated to produce 3.3 billion tonnes of carbon dioxide (CO₂) emissions per year (FAO 2013). The theoretical CH₄ yield and electricity production potential of FW is 0.300 m³/kg VS and 473.8 kWh, respectively (Suhartini et al. 2019). Therefore, tapping the energy potential of FW via a biorefinery framework could be a sustainable approach for organic waste management and the production of multiple-resources for commercial use.

This paper aims to provide an extended overview of the DF and AD process for biohythane and biochemical production (particularly VFA) from FW within the biorefinery approach. This paper discusses how pretreatment, different operational parameters, and the choice of bioreactors in the DF-AD affect the substrate degradation and metabolic pathways for H₂, CH₄ and organic acids production from FW. Likewise, the optimum conditions for biohythane and VFA production, economic and future prospective of utilizing the VFA for different biochemical such as polyhydroxyalkanoates, polylactate, and microalgal biomass has been discussed.

FW as a feedstock for biohythane and biochemical production

The FW which is generally rich in carbohydrates has been widely used for H₂ and biochemical production. FW comprises mostly carbon chains and has a high energy content (Fatima et al. 2020). The average energy content in the mixed food sample (meat, fruits, and vegetables) was found to be 14.31 MJ/kg (Tanai 2016). Every year, approximately one-third of the food produced for human consumption gets wasted which is likely to increase with the global population (Gustavsson et al. 2011). However, the characteristics of FW vary depending upon the consumption pattern of different countries (Paritosh et al. 2017). Globally, the highest FW constituent is cereal (more than 80,000 million tonnes) whereas considering only Asia, vegetable (60,000-kilo tonnes) contributes to the maximum FW (Paritosh et al. 2017). Thus, this abundant FW can be extensively used in two-step DF-AD (Bolzonella et al. 2019; Dahiya et al. 2018).

Generally, carbohydrate-rich FW is regarded as the most ideal substrate for H₂ and CH₄ production in a two-stage process, whereas lipids and protein-rich foods are the least preferred (Alibardi and Cossu 2016). A 20 fold increase in biological H₂ production was observed when using carbohydrate-rich substances compared to lipid and protein-rich substances (Fatima et al. 2020).

Cieciura-Włoch & Borowski (Cieciura-Włoch and Borowski 2019) found out that plant-based waste such as fruit and vegetable waste is most suitable for hydrogen production with the highest yield of 280 L H₂/kg VS. They also found out that slaughter house and kitchen waste have high methane yield but unfavorable for hydrogen production because of high protein and fat. The protein-rich substrate with a low C: N ratio increases ammonia concentration during the AD process leading to inhibition of methanogenic microorganisms (Bolzonella et al. 2019). Similarly, the lipid is degraded into long-chain fatty acids (LCFA) which have an inhibitory effect on acetogens and methanogens (Dasa et al. 2016).

Besides the nutrients, the high total solid (TS) content of around 20–30% in FW makes it favorable for operating the first stage reactor under dry conditions (Capson-Tojo et al. 2017). According to the TS content of the FW was also found to be adjusted to 4.3% and 10% in tests carried out by Algapani et al. (2018) and Akhlaghi et al. (2019), respectively. The TS content greater than 15% can result in a decrease in substrate conversion and lactate is produced as a major fermentation product (Ghimire et al. 2017). Similarly, methane yield was found to be higher in the FW of TS 15–20% than in the TS of 5–10% (Chen et al. 2014). FW has a high VS content (21 to 27% VS) and is particularly suitable for the co-production of biohydrogen and platform molecules such as short organic acids and/or alcohols (Uçkun Kıran et al. 2015).

FW can be co-digested with substrates such as anaerobic sludge (can function as inoculum), waste activated sludge, wheat straw, chicken manure, and grass to enhance biohythane and VFA production (Ghimire et al. 2017; Show et al. 2018; Wang et al. 2014; Esteban-Gutiérrez et al. 2018). According to Wang et al. (2014), the VFA production from FW with the addition of aerobic and anaerobic activated sludge was recorded to be 0.482 g/g VSS_removal and 0.918 g/g VSS_removal, respectively. Similarly, the co-digestion of FW with brown water (feces without urine) formed acetic and butyric acids as major metabolites which are the most appropriate for methe and useable by-products production (Rajagopal et al. 2014).

FW is rich in organic content due to which there is a possibility of hydrolysis inhibition by the accumulation of ammonia, rapid acidification, and formation of LCFA (Menzel et al. 2020). This inhibits the optimum energy recovery from FW in single-stage AD. Therefore, besides co-digestion, two-stage AD, and pretreatment have also been considered as the efficient approach for enhancing the solubilization and hydrolysis of FW (Uçkun Kiran et al. 2015; Rodríguez-Valderrama et al. 2020; Menzel et al. 2020).

Two-stage AD process as a pathway for biohythane and biochemicals

Generally, the conversion of organic waste during AD consists of four steps: (i) hydrolysis (ii) acidogenesis (iii) acetogenesis, and iv) methanogenesis. In the two-stage AD process, these processes are physically separated by two reactors under controlled operating conditions (Fig. 1).

Fig. 1

Schematic diagram of the two-stage AD process

Stage I: dark fermentation (DF) process

Hydrolysis and acidogenesis phases are carried out in the first stage reactor. In this reactor, the complex organic compounds are hydrolyzed and further degraded by acidogenic microorganisms to produce C2-C5-based VFA, H₂, CO_2, and alcohols in the DF process (Camacho et al. 2019). Even though FW is readily biodegradable, hydrolysis is the rate-limiting step during the AD process (Zhang et al. 2014). However, for the two-stage AD process of organic waste, methanogenesis is the rate-limiting step (Meena et al. 2020) because of the slower methanogenic kinetics and longer lag phase in the second-stage reactor (De Gioannis et al. 2014).

During the DF process, most of the energy contained in the organic substrate is converted into VFAs whereas only 7.5–15% of the total energy is converted to H₂ (Bolzonella et al. 2019). The possible pathways for H₂ production are presented in Table 1 Eqs. (1) to (4), which shows that the maximum theoretical yield from the DF process can be 4 mol H₂/mol glucose (Saratale et al. 2019). However, this cannot be achieved due to the complex nature of the FW which may contain a different fraction of carbohydrates, proteins, and fat.

Table 1 Equation representing metabolic pathway for H₂ production

Stage II-anaerobic digestion (AD) process

The DF effluent from the first stage can be fed into the second stage in which acetogenesis and methanogenesis occur. The VFAs with higher molecular weight and other intermediates are further oxidized by H₂-producing acetogens as shown in Eqs. 5–7 (Van et al. 2020).

$${\text{CH}}_{{3}} {\text{CH}}_{{2}} {\text{COO}}^{ - } + {\text{ 2H}}_{{2}} {\text{O }} \leftrightarrow {\text{ CH}}_{{3}} {\text{COO}}^{ - } + {\text{ CO}}_{{2}} + {\text{ 3H}}_{{2}}$$

(5)

$${\text{CH}}_{{3}} \left( {{\text{CH}}_{{2}} } \right)_{{2}} {\text{COO}}^{ - } + {\text{2H}}_{{2}} {\text{O }} \leftrightarrow {\text{ 2CH}}_{{3}} {\text{COO}}^{ - } + {\text{ 2H}}_{{2}} + {\text{ H}}^{ + }$$

(6)

$${\text{CH}}_{{3}} {\text{CH}}_{{2}} {\text{OH }} + {\text{ H}}_{{2}} {\text{O }} \leftrightarrow {\text{ CH}}_{{3}} {\text{COO}}^{ - } + {\text{ 2H}}_{{2}} + {\text{ H}}^{ + }$$

(7)

Approximately, 25% of acetate and 11% of H₂ are formed in acetogenesis (Anukam et al. 2019). Though H₂ is produced during acetogenesis, it is not recommended to extract H₂ from this stage (Van et al. 2020). Because acetogenic microorganisms are compatible with methanogens (Roy and Das 2016). And, this acid phase product (acetate and H₂ and CO₂) is consumed by methanogenic microorganisms for CH₄ production (Meegoda et al. 2018). The undigested or unhydrolyzed food residues can also be converted to CH₄ in the second-stage reactor. In the two-stage AD process, AD is mostly used for the recovery of CH₄ rather than organic acids.

In acetoclastic methanogenesis, acetate is converted into CH₄ and CO₂ which is accountable for two-thrid of CH₄ production (Eq. 8). Alternatively, reduction or CO₂ to CH₄ can also take place, which is termed as hydrogenotrophic methanogenesis which is accountable for the remaining one-third CH₄ (Eq. 9) (Fenchel et al. 2012).

$${\text{CH}}_{{3}} {\text{COOH }} \to {\text{ 2CO}}_{{2}} + {\text{ CH}}_{{4}}$$

(8)

$${\text{4H}}_{{2}} + {\text{2CO}}_{{2}} \to {\text{ CH}}_{{4}} + {\text{ H}}_{{2}} {\text{O}}$$

(9)

The physical separation benefits the growth of required microorganisms within that particular process. For example, hydrolysis is limited by fermentative bacteria whose optimum pH is 5.5, whereas the pH of 7.5–8.5 is needed for methanogenic microorganisms (Sivagurunathan et al. 2018). In DF, the hydrolysis and acidogenesis process is maintained by lowering the hydraulic loading rate (HRT), controlling the pH around 5.5—6.5, and inoculum pretreatment to enrich H₂ producers (Chu et al. 2008; Xia et al. 2016; Rafieenia et al. 2017).

Operational parameters affecting the two-stage AD process producing biohythane.

The optimization of operational parameters such as pH, temperature, hydraulic retention time (HRT), substrate loading rates, pretreatment methods, recirculation of effluent from the second stage to the first stage (recirculation ratio), types of bioreactor is essential in the integrated two-stage DF-AD process. The operational parameters affecting the biohythane production in the two-stage AD process are shown in Fig. 2. The effect of varying operating parameters in a two-stage AD process has been summarized and presented in Table 2.

Fig. 2

Operating parameters affecting the biohythane production

Table 2 Two-stage AD process using FW under varying operating conditions

pH

The pH is one of the most important parameters in biohythane production because it determines the subsistence of acidogenic or methanogenic microorganisms. It directly affects the metabolic pathways involved in the process of anaerobic fermentation and digestion. The formation of undissociated organic acids when pH is lower than 4 ceases H₂ production in DF (Bolzonella et al. 2019).

While operating a two-stage process for biohythane production from FW, Cavinato et al. (2011) reported 15% of H₂ production in the first phase and 65% of CH₄ in the second phase at pH of 4.32 and 7.68, respectively. Moreover, around 35% of H₂ production was observed at pH 3.51. Maintaining pH at acidic and neutral alkaline conditions is suitable for DF and AD reactors, respectively. This creates a favorable growth environment for H₂ producing microorganisms in DF and methanogens in AD (Meena et al. 2020).

Anaerobic fermentation (AF) or DF process have been used by the different researchers for VFA production under controlled pH varying from pH of 6 to 8 (Table 3). However, the VFA production was improved in alkaline conditions (pH 9) in comparison to acidic conditions (pH 6) for FW in batch tests (Cheah et al. 2019). The certain value of pH also affects the concentration of targeted VFAs. The culture pH of 6 was found to be suitable for the production of acetate whereas, pH 7 facilitated the production of butyrate from FW (Hussain et al. 2017).

Table 3 Production of VFA from different types of fermentation

The pH inside the first stage keeps on fluctuating due to the formation of VFAs. This inhibits the H₂ production pathway as H₂ and CO₂ is consumed by acetogens to produce acetate (Cavinato et al. 2011). Therefore, the pH of the first stage reactor can be controlled by external chemical addition (Micolucci et al. 2014), the use of co-substrate with high alkalinity or buffering capacity (Yeshanew et al. 2016; Ghimire et al. 2016), and continuous operation for CH₄ production could also avoid pH drop as high VFA concentration is buffered by high total ammonia concentration (Capson-Tojo et al. 2016). However, these external methods and continuous processes result in high costs and instability of operation. Therefore, a cost-effective solution in two-stage AD could be recirculating the digestate of the CH₄ reactor to the H₂ reactor (Cavinato et al. 2012, 2011; Micolucci et al. 2014).

Temperature

An optimum temperature plays a vital role in the growth of microorganisms as it influences biochemical reactions and their metabolism. Microorganisms that are responsible for H₂ production in the first stage and the CH₄ production in the second stage are often found to be grown at mesophilic (30–35 °C) to, thermophilic (55–60 °C) to higher thermophilic condition (70–90 °C) (O-Thong et al. 2018).

Most of the researchers have kept the optimum temperature at 55 °C for biohythane production as evident in Table 2. According to O-Thong et al. (2018), higher thermophilic temperatures are more favorable for H₂ production during DF than thermophilic and mesophilic conditions. Microorganisms such as Caldicellulosiruptor sp. survive in extreme temperatures and have hydrolytic enzymes which can utilize various substrates such as cellulose for more H₂ production. Moreover, higher temperature accelerates their metabolism and degradation efficiency that reduces the retention time for biohythane production. Furthermore, thermophilic condition decreases the solubility of CH₄ and CO₂ along with the destruction of pathogens (O-Thong et al. 2018). However, Ghimire et al. (2021) recorded the highest H₂ (53.5 mL H₂/g VS_added) and CH₄ (307.5 L CH₄/kg VS_added) production from FW in the DF-AD process at mesophilic and thermophilic temperature, respectively. The probable reason for this distinction is the quicker adaptation of heat-treated inoculum at the mesophilic condition as it is sourced from the mesophilic digester (Ghimire et al. 2021).

The most commonly used process temperature for VFA production from the AF process is 35℃ (Table 3). According to Cho et al. (2015), the VFA concentration from AF increases with the increase in temperature. VFA produced at 55 °C was 3.2 times greater than the VFA produced at 35 °C. Similarly, the concentration of propionic acid increased by 15% but acetic acid decreased to 65.8% with the increase in temperature. However, Jiang et al. (2013) found 45℃ as the most favorable temperature for the highest VFA production (47.89 g/L). The value was only slightly higher than the VFA (41.34 g/l) produced at 35 ℃. According to Jiang et al. (2013), the higher temperature leads to higher solubilization but lower acidogenesis of FW thus, leading to lower VFA production.

Hydrogen partial pressure (HPP)

The hydrogen partial pressure (HPP) in the liquid phase is an essential parameter for the DF (Ding and Zhao 2018; Beckers et al. 2012). Because high HPP (> 60 Pa) shifts the equilibrium reaction to the H₂ producing reaction (Ding and Zhao 2018) by deactivating the enzymes such as hydrogenase and NADH (Ramírez-Morales et al. 2015). However, no correlation was reported between HPP, liquid, and the NADH/NAD⁺ ratio (De Kok et al. 2013). But, lowering HPP affected pH and microbial growth rate in DF (Ding and Zhao 2018). The H₂ and VFA production in DF decreases with the increase in HPP (Ramírez-Morales et al. 2013). Besides the concentration, HPP also influences the composition of VFAs though propionic acid concentration is not directly affected by ppH₂ (Ding and Zhao 2018).

A study conducted by Beckers et al. (2012) presented an increase of 7% of biohydrogen yields on the decrease total pressure 0.11 bar and a decrease of 19.5% yield on the lowering of pressure by 0.18 bar. The researchers experimented on an anaerobic biodisc reactor (ABR) that fixes the biomass and favors liquid to gas transfer leading to efficient H₂ production (Beckers et al. 2012). An identical study by Cazier et al. (2015) altered HPP from 0 to 1.557 bar which led to a substantial decrease in CH₄ yield and substrate degradation. It was seen that high HPP led to a decrease in degradation of the substrate as well as accumulation of metabolites.

The HPP can be controlled by venting out excess H₂ gas. Another solution could be introducing different gases such as N₂, CO₂, CH₄, or even biogas and creating a vacuum on the headspace of the reactor (De Kok et al. 2013; Dionisi et al. 2019). However, this leads to diluted H₂ production. Ramírez-Morales et al. (Ramírez-Morales et al. 2015) suggested integrating membrane technology to remove excess H₂ during a continuous fermentation process. The most suitable option would be to vent out the H₂ produced at regular intervals or continuously.

Hydraulic retention time (HRT)

In bioreactor systems without sludge recirculation, HRT can refer to the time length that the biomass remains in the bioreactor (Strazzera et al. 2018) which depends upon the volume of the reactor and the flow rate of the feed. Therefore, it is also associated with capital cost. The optimum HRT for the first reactor is 1–3 days whereas for the second reactor is 10–15 days (O-Thong et al. 2018). This is because the H₂ producing bacteria have a faster growth rate compared to methanogenic bacteria.

Similarly, the longer HRT for the second reactor allows enough time for substrate degradation. The highest methane production rate of 2041.7 L CH₄/m³/day from an anaerobic fixed bed reactor (AFBR) was possible even at the shortest HRT of 1.5 days given the maximum organic loading rate (OLR) of 6.1 kg COD/m³.day (Yeshanew et al. 2016). This implies that the use of different biofilm reactor technologies such as AFBR that prevent washing out methanogenic organisms. It helps in the reduction of HRT without compromising the reactors ‘performance. Though methanogens such as Clostridiaceae are highly reduced at low HRT (4–8 h), other homoacetogens like Clostridium ljungdahlii, were difficult to remove due to the granular structure of Up-flow Anaerobic Sludge Blanket (UASB) and packed bed reactor (PBR) (Si et al. 2015).

Vo et al. (2019) found out that the optimum H₂ (714 mL/L.d) and CH₄ (254 mL/L.d) were produced at an HRT of 2 days. The study experimented with a single-stage AF reactor which consists of two compartments each for H₂ and CH₄ production. The effect of HRT and OLR along with pH on VFA production from biowaste has been explored. Lim et al. (2008) studied the effect of 3 different HRTs (4, 8, and 12 days) at controlled pH 5.5 and OLR of 5 g TS/ L.d. There was an increase in the VFAs concentration (5.5 g/L at 4 days, 13 g/L at 8 days, 22 g/L at 12 days) with an increase in HRT. At shorter HRTs, acetate was the dominant product while at longer HRT (12 days) propionic acid was higher.

Han and Shin (2002) also studied the effect of HRT (0.25, 0.33, 0.50, and 1 d) on VFA production from FW on Leachate bed reactor (LBR) under controlled temperature (35℃) and pH. At an HRT of 1 day, the maximum VFAs concentration of 202 and 181 mmol/L was determined with rumen and anaerobic bacteria, respectively. However, at 0.25 days of HRT, the least VFAs concentration was detected. According to Strazzera et al. (2018), lower HRT (< 10 days) favors VFA production because the longer HRT facilitates methanogenic microorganisms to convert VFA to CH₄.

Organic loading rate (OLR)

Organic loading rate (OLR) refers to the amount of substrate fed into the reactor in a day per unit of working volume. It is a deterministic parameter for the failure of the system due to bulking and acidifications. For CH₄ production from vegetable waste, the optimum OLR for a CH₄ production rate of 0.26m³ CH₄/kg VS was 1.4 kg VS/m³.d, and as the OLR increased VS degradation and biogas yield decreased (Jalil et al. 2019). High OLR reduces the hydrolysis and acidogenesis process due to imbalanced osmotic pressure and increased viscosity that limits mass transfer and metabolic activities (Tang et al. 2016).

However, some experiments have also proved better biohythane production even at high OLR under the condition of low HRT (Cavinato et al. 2011; O-Thong et al. 2018; Yeshanew et al. 2016). A stable H₂ (56.6 mL H₂/g VS) and CH₄ (248 mL CH₄/g VS) production were observed under high OLR (> 10 g VS/ L.d) for a two-phase thermophilic continuous stirred tank reactor (CSTR). But under similar OLR the one-phase thermophilic CSTR failed because of acetate and propionate accumulation in the methane reactor (Luo et al. 2010). Another example from the pilot-scale (50L) two-stage AD process, the gas production from FW increased up to 13,000–15,000 L/d due to an increase in OLR at a low HRT of 3.9 days (Lee and Chung 2010).

Jiang et al. (2013) studied the VFA production from FW at 3 different OLR (5, 11, and 16 g TS/L) and constant temperature (35 °C), pH (6.0), and HRT (5 days). The researcher found that the VFA production increased sharply at the initial days (7–14 days) then remained stable for OLR of 5 and 16 g TS/L. However, at OLR of 16 g TS/L, the VFA production increased by 12 days then decreased sharply. Acetate and Butyrate were the dominant (60%) VFA produced at all OLR. Though there is an insignificant change in VFA fraction at high OLR (Paudel et al. 2016; Cheah et al. 2019), Amha et al. (2019) also recorded an increase in VFA production while increasing OLR. Because increasing OLR is attributed to the diverse microbial community (Srisowmeya et al. 2020).

Nevertheless, the two-stage AD process for FW can operate at high OLR but within controlled HRT and pH (Lim et al. 2008; Srisowmeya et al. 2020). High OLR provides excess carbon for degradation inducing acidic conditions in the reactor in the long-term. Maintaining the pH (5–6.5) for DF and low HRT (2–3 days) that helps to wash-out methanogenic microorganisms provides an optimum condition for biohythane and VFA production at high OLR in a two-stage AD process (Cavinato et al. 2016; Paudel et al. 2016).

Recirculation ratio (RR) of second stage-effluents

The recirculation ratio (RR) refers to the ratio of the returned volume of the second-stage reactor effluent to the volume of the first stage reactor influent. The digestate from the methanogenic stage is recirculated to maintain the pH of the first phase reactor (Algapani et al. 2019). This innovative strategy utilizes the residual buffer capacity of digestate to supply nutrients in the first stage reactor, i.e., DF reactor. This has proved to enhance the biohythane production for two-stage reactors and improves its cost-effectiveness (Micolucci et al. 2014; Cavinato et al. 2011).

In a study conducted by Yeshanew et al. (2016), the supernatant effluent from AFBR was fed to CSTR at a RR ranging from 0.24–0.48, 0.5–0.8, and 0.6–1 for three different periods, respectively. An improved and stable hydrogen production rate of 178.2, 253.5, and 391.7 L H₂/m³.d in periods I, II, and III, respectively was obtained (Yeshanew et al. 2016). Similarly, Algapani et al. (Algapani et al. 2019) stated that the RR reduces the need for external chemicals to maintain the pH in the H₂ reactor by 54%. It also increased the H₂ by 8% and decreased the CH₄ production by 3%. However, there was no difference in the total energy production (Algapani et al. 2019).

According to Chinellato et al. (2013), the first stage reactor produced CH₄ at a high OLR of 20 kg TVS/m³/d instead of H₂. However, with the RR of 2.9, the highest H₂ production from FW was obtained at 20 kg TVS/m³/day. Similarly, for the VFA production from LBR, improving the leachate recirculation rate improved the hydrolysis performance by 10–16% and the acidification yield increased to 340 g COD/kg TVS added (Hussain et al. 2017).

Although many such experiments have proved to enhance the efficiency of biohythane production (Chu et al. 2008; Chinellato et al. 2013; Cavinato et al. 2016), the increased ammonia concentration in the first reactor could confine the performance of anaerobic microorganisms known for both H₂ and CH₄ production (Cavinato et al. 2012; Algapani et al. 2019). Therefore, it is recommended to daily observe the ammonia concentration and remove the recirculation of the second reactor effluent once the ammonia concentration reaches its maximum stable value (Cavinato et al. 2012).

Based on the literature, the optimum conditions in the two-stage AD process for H₂, VFA, and CH₄ production from the FW are summarized in Table 4. H₂ production from DF was found to be suitable within the acidic range unlike for VFA and CH₄ production. The selection of pH value for VFA also depends upon the targeted organic acid. Similarly, the temperature range is within the mesophilic to thermophilic conditions. From the limited available studies, a low range of HPP (below atmospheric pressure) is suitable for the DF-AD process. The OLR and RR vary depending upon the HRT and the reactor size.

Table 4 Optimum range of operational parameters for H₂, VFA, and CH₄ production from FW using DF-AD process

DF-AD process integration

The coupling of DF and AD as two-staged biorefinery is observed to be more productive in terms of energy recovery not only because of the production of biohydrogen but also because of higher yields of CH₄ compared to an uncoupled system (Ruggeri et al. 2015; Ghimire et al. 2021). This is seen as a result of biological pretreatment offered by DF by breaking the nutrients in the substrate down to VFAs (Malave’ et al. 2015; Malavè et al. 2018). The reactor type and configuration strongly influence the coupled system, especially through biomass retention (Show et al. 2011).

It is seen that the DF-AD system is integrated widely in CSTRs connected in series with a batch, continuous or semi-continuous feeding, and effluent extraction mechanism (Table 2). In batch mode, the substrate is added to the digester and inoculated for complete digestion, whereas, in continuous feeding mode, the substrate is fed continuously or semi-continuously, and biogas is collected continuously (Stalin and Prabhu 2007). The CSTR is suitable for FW with a TS of 2–12% (Liu et al. 2013). Moreover, two CSTRs have been combined vertically that produced maximum H₂ and CH₄ content of 8.6% and 48%, respectively, from the FW at HRT of 2 days (Vo et al. 2019). The popularity of CSTR is attributed to its simplicity, uniform mixing, and suitability with any kind of substrate (Saratale et al. 2019). However, at HRTs lower than 2.5 days, CSTR enabled mass cell wash-out from the reactor (Kongjan and Angelidaki 2010).

For the two-stage process stability and better performance, CSTR has also been combined with high state reactors like AFBR, UASB, Up-flow Anaerobic Packed Bed (UAPB), and Expanded Granular Sludge Bed (EGSB) (Lay et al. 2010; Liu et al. 2013; Cisneros-Pérez et al. 2017). These wet reactors are suitable for liquid substrate (TS < 2%) and facilitate biomass retention for a longer time which is appropriate for slow-growing methanogenic organisms; hence, mostly used as the second-stage reactors in the two-stage AD process. (Liu et al. 2013; Nualsri et al. 2016; Van et al. 2020).

A study by Ren et al. (2010) has concluded that attached sludge CSTR is more stable than suspended sludge CSTR. A study comparing CSTR and AFBR in immobilized and suspended cell systems concluded that DF is enhanced in reactors that aid biofilms (Qureshi et al. 2005). The recirculation strategy in these reactors' configurations allows maximum H₂ production from DF even at HRT of less than 2 days (Yeshanew et al. 2016).

A settler tank can also be added in between CSTR and UASB, or similar reactors for solid–liquid separation. Lee and Chung (2010) developed the first pilot-scale system consisting of CSTR (500 L) as DF, UASB (2300 L) as AD fed with FW liquid using anaerobic sludge as the inoculum, and a fuel cell fed with the purified biohydrogen. However, CSTR and UASB have been successfully integrated for biohythane recovery from sugarcane without the need for a separation tank in between (Nualsri et al. 2016).

Another novel two-stage AD process has been an integration of dry reactors like LBR with UASB for biohythane (Han and Shin 2004) and VFA production (Browne et al. 2013; Yan et al. 2019) from FW. The dry reactor is suitable for FW with TS ≥ 15%(Liu et al. 2013; Van et al. 2020). Han and Shin (2004) used BIOCELL (LBR-UASB) which has VS conversion efficiency to H₂ was 28.2% and CH₄ 69.9%, respectively, from FW. The BIOCELL demonstrated stability through the resource recovery process. LBR has several advantages to CSTR as LBR doesn’t require substrate dilution reducing process water, no stirrer for mixing that saves energy and the leachate can be recycled without passing it through the solid–liquid separation unit (Browne et al. 2013; Hussain et al. 2017). Similarly, the bacterial dynamics in LBR can degrade even the resistant dietary fibers present in FW (Xiong et al. 2019).

Various types of membrane filtration reactors (MFR) such as microfiltration, nanofiltration, evaporation, electro-dialysis, and ultrafiltration have been used to separate the VFAs (Zacharof and Lovitt 2014; Sasiradee et al. 2017). Short Chain Fatty Acids (SCFA) yield of 7453 mg COD/ L was recorded from the integration of alkaline sludge fermenter and membrane separation unit (Longo et al. 2015). The combination of anaerobic membrane filtration technologies is more efficient than the single step in recovering the high quality of targeted organic acid (Sasiradee et al. 2017; Sikder et al. 2012). Besides, the appropriate selection of adsorption media is essential in membrane filtration technology as it affects the removal efficiency (Uslu 2009). Out of the four different ion exchange resins (IRA-900, IRA-400, IRA-96, and IRA-67), Luongo et al. (2019) recorded IRA-67 recovered 97% of LA in batch test whereas the desorption efficiency in fixed bed reactor was only 68%. Similarly, Amha et al. (2019)found ceramic membrane favorable for FW treatment in long-term without irreversible fouling.

The working volume and flow rates of DF and AD systems vary according to the operating mechanism and the speed of reactions. The difference between the equivalent working volumes of the different stages of anaerobic is symbolized by the coupling ratio. Contorted coupling ratio could result in reactor failure or inefficient operation (Srisowmeya et al. 2020). From Table 2, it can be observed that the size of the second reactor is either equal to or greater than that of the first reactor in all the cases. As DF has the potential to process higher OLR than AD, the reactor size for AD is correspondingly seen to be higher than that of DF (Ren et al. 2011).

The DF-AD reactor configurations for FW experimented at laboratory or pilot scale have been shown in Fig. 3. The second-stage reactor configuration depends upon the characteristics of effluent from the first stage reactor (Liu et al. 2013). Similarly, the performance of each reactor along with its drawback is summarized in Table 5.

Fig. 3

Available DF-AD reactor configuration for FW in literatures

Table 5 Performance, microbial characteristic, and drawbacks of DF-AD bioreactors

Overall, the selection of DF-AD reactor configuration depends upon the TS content of the feedstock, targeted VFA, and process stability required. In addition to a combination of two-stage reactors, an additional unit like aerobic denitrification tank (Lee et al. 2010) and zeolite adsorption unit (Petracchini et al. 2018) for further treatment of leachate before recycling to the first reactor. This is done to avoid the inhibition by ammonia in the first reactor and ensure long-term stability.

Pretreatment of FW and inoculum

FW composition is heterogeneous which may cause technical instability in the two-stage AD process (Cesaro and Belgiorno, 2014), especially during the start-up phase. The carbohydrate-rich FW is favorable for DF whereas, protein and lipid are difficult to degrade. To catalyze the decomposition process in DF, the inoculum is added. The inoculum obtained from the sludge of other waste treatment plants and breweries contains mixed culture (Alibardi and Cossu 2015; Cappai et al. 2018; Algapani et al. 2019). The mixed culture induces ecological interactions and competition between various H₂ consuming and H₂ producing microorganisms for the same organic substrate which ultimately reduces the performance of the reactor. The H₂ producing microorganisms have a better chance to survive the harsh conditions during the pretreatment of inoculum than the non-spore-forming bacteria such as methanogens and can germinate when favorable conditions are provided to them (Li and Fang 2007).

The pretreatment of FW is usually done to improve the hydrolysis of hardly biodegradable compounds, while inoculum pretreatment is done to enrich H₂ producing microorganisms and inactivate the consumers like methanogens during the start-up of the DF process (Wang and Yin 2017) and improve H₂-CH₄ yield in two-stage AD process (Lee and Chung 2010; Yin et al. 2014). The inoculum pretreatment may be only necessary for the start-up phase while the FW pretreatment is a continuous process. The common pretreatment measures adopted by the researchers for the FW and the enrichment of H₂ producers are explained in the following sections.

Heat treatment

Heat treatment of inoculum for the enrichment appears to be common, simple, inexpensive, and effective (Li and Fang 2007; Wang and Wan 2008). During the heat-shock treatments, the inoculum is heated at 100° C for 30 min under atmospheric pressure (Jariyaboon et al. 2015; Akhlaghi et al. 2019; Sun et al. 2019). The acidogenic bacteria can form spores and ensure their survival whereas methanogens are deactivated in those conditions (Li and Fang 2007). The results after heat treatment did not show any traces of CH₄ in the gas composition obtained after DF. Wang and Wan (2008) found out that heat-shock treated digested sludge increased the substrate degradation efficiency by 97.2%. It further improved H₂ yield (221.5 mL/g glucose) and biomass concentration (2739 mg/L). In DF, Ghimire et al. (2016) reported that H₂ yield from heat-shock treated activated sludge increased by two folds than the anaerobic sludge.

The substrates can be pretreated to enhance the conversion of substrates to intended products. During the heat pretreatment such as steam explosion, the substrate is heated under high temperature (160–260 °C) and pressure (7–50 bar) for a short duration (30 s–20 min) (Keskin et al. 2019). Then, the pressure is quickly released which causes an explosive effect. This pretreatment is suitable for substrates with high lignocellulose (Eg: wheat straw, corn starch) that is commonly found in food industry waste. The explosion improves its digestibility by 90% than steam pretreatment without explosion (Pielhop et al. 2016). This method has been commonly used in the solubilization of sewage sludge (Donoso-Bravo et al. 2015; Dereix et al. 2006) but can be also utilized for the FW pretreatment. However, the steam explosion pretreatment of FW makes it more dilute decreasing the acetic acid and lactic acid concentration (Svensson et al. 2018).

Another similar heat pretreatment method under high temperature and pressure is known as hydrothermal pretreatment. In this method, at high temperature and pressure, the ionized products of water are increased which hydrolyze the macromolecules resulting in a noticeable improvement in FW fermentation (Yin et al. 2014). The suitable temperature was reported to be 160 °C, which gave a VFA yield of 0.908 g/g VS_removal, 47.6% higher than the control. However, higher temperatures can result in the production of toxic compounds that inhibit microbial activities. Similarly, Ding et al. (2017) recorded 140 °C as the optimum temperature for hydrothermal heat pretreatment of FW as it favored efficient solubilization of carbohydrates and protein. The H₂ and CH₄ yield from two-stage AD was recorded to be 43.0 mL/g VS and 511.6 mL/g VS. The energy conversion efficiency of pretreated FW increased by 31.7% than the untreated FW.

Chemical treatment

Methanogens are strict anaerobes and are very sensitive to many chemicals. It has been found that CH₄ production drops sharply at a pH of below 6.3 or above 7.8 (Li and Fang 2007). The chemical treatment involves the introduction of acidic/alkali compounds to maintain the desired pH of the reactor. Acid pretreatment is efficient in solubilizing carbohydrates whereas alkali treatment is suitable for solubilizing protein, lipids, and lignin (Parthiba Karthikeyan et al. 2018).

Jang et al. (2015) also recorded the highest H₂ production by alkali-shock treatment of FW with 6 N KOH at pH 11 and 12 and without using any co-substrate. Vavouraki et al. (2013) also recorded that pretreatment of a kitchen waste under 1.12% HCl for 94 min or 1.17% HCl for 86 min (at 100 °C) increased the soluble sugar concentration by 120% compared to the untreated kitchen waste. According to Lamaison et al. (2015), acid treatment of the anaerobically digested sludge from sugarcane vinasse treatment plant is better than the heat-treated sludge in long-term. Because the acid pretreated sludge favored the growth of Clostridiaceae microbes which consumes the lactate concentration that is known for H₂ inhibition. Likewise, the alkaline pretreatment of raw rice straw using 4% and 8% NaOH at 55 °C for 24 h, increased the H₂ yield by 26 and 57-fold, respectively (Ghimire et al. 2016).

The chemicals like sodium 2-bromoethasulfonic acid (BESA), iodopropane, chloroform, and acetylene are commonly used as methanogenic chemical inhibitors (Wang et al. 2011; Li and Fang 2007; Zhu and Beland 2006; Venkata et al. 2008; O-Thong et al. 2009). BESA inhibits the activity of the co-enzyme M reductase complex which is a chief component for methanogenesis (Venkata et al. 2008; Zhu and Beland 2006). According to Zhu and Beland (2006), BESA is the most cost-effective pretreatment method for inhibiting methanogens without impacting the H₂ producing microorganism in digested sludge.

Another chemical pretreatment with ozone is quite popular for waste-activated sludge (Carrère et al. 2010) but it is least explored with FW. Ariunbaatar et al. (2014) obtained a negligible increase in CH₄ production from the ozone-treated FW. This could be mainly because a high amount of fermentable sugar lost at a high ozone dose.

The chemical pretreatment has been combined with heat and ultrasonic treatment for the effective conversion of FW to H₂ and CH₄. Monlau et al. (2012) compared seven different types of thermo-chemical pretreatment of sunflower stalk. Chemicals (NaOH, H₂O₂, Ca (OH)₂, HCl, and FeCl₃) were used to vary the pH from 2.3 to 12.2, and the temperature was maintained at 55 °C and 170 °C. Alkaline pretreatment (4% NaOH) at 55℃ for 24 h is recommended for the lignocellulose substrate (Monlau et al. 2012). Similarly, Elbeshbishy et al. (2011) found that the combination of ultrasonic sound and acid produced the highest H₂ yield of 118 mL/g VS_initial and VFA concentration 16,900 mg COD/L.

The addition of chemicals is found to be quite an expensive pretreatment method and hence, least preferred. Moreover, concentrated acid pretreatment is required for the efficient hydrolysis of biomass with high cellulose content. As acids are highly corrosive, specific materials are needed to construct the reactor. This adds up to the capital cost making it an expensive option (Keskin et al. 2019). Similarly, the FW or inoculum needs to be neutralized after the chemical treatment which makes it complicated (Banu et al. 2020).

Other pretreatment methods

In the AD process, oxygen is avoided as it inhibits the activity of anaerobic fermentation microorganisms. However, aeration pretreatment of FW in single-stage AD process has been proven to enhance microbial diversity leading to improved hydrolysis and acidogenesis process (Lim and Wang 2013; Girotto et al. 2016). Even in a two-stage process, pre-aeration of FW has increased VFA production (Xu et al. 2014). According to Rafieenia et al. (2017), the pre-aerated FW produced less H₂ than the non-aerated FW. But the CH₄ production increased in the second reactor, especially the protein-rich FW produced the highest CH₄ (351 ml/gV). The effect of aeration pretreatment depends upon the FW composition, aeration time, intensity, and air introduction method (Girotto et al. 2016; Rafieenia et al. 2017).

Moreover, researchers have also studied the effect of aeration pretreatment of inoculum on H₂ production (Wang and Wan 2008; Zhu and Beland 2006). The comparison of the results of pretreatment methods (i.e., acid, base treatment, heat shock, aeration, and chloroform) showed that the inoculum pretreated with aeration has higher cumulative H₂ production than the inoculum treated with chloroform and untreated inoculum (Wang and Wan 2008).

Load shock as a pretreatment was also carried out to obtain high glucose loading of 50 g/L at the beginning of fermentation (Jariyaboon et al. 2015). Different pretreatment methods of inocula for optimizing hydrogen production conducted by (Chang et al. 2011) consisted of acid, base, heat-shock, aeration, chloroform, and 2-bromoethanesulfonate treatment. Acid pretreatment at pH 3 yielded the best result with maximum hydrogen production of 22.81 mmolH₂/gVSS from activated waste sludge.

Cisneros-Pérez et al. (2017) compared a cell wash-out treatment and heat pretreatment of the inoculum in an anaerobic fluid bed reactor (AFBR). This study showed that cell wash-out treatment had higher H₂ volumetric production rates (7 L H₂/L-d) and yields than thermal treatment (3.5 mol H₂/mol hexose). The effect of pretreatment on H₂ production is different based on different sources of inoculum. However, the selection of the pretreatment method should be based on the use of effluent from dark fermentation. For example, the selection of chemical treatment methods such as using BESA could create a problem when the effluents from dark fermentation are meant to be used for methane production in the second stage.

For FW with a high TS loading rate, ultrasonic pretreatment could be a good option (Gadhe et al. 2014). Ultrasonic pretreatment mechanically disrupts the physical, biological, and chemical properties of a substrate/ inoculum under varying frequency (Pilli et al. 2011). Gadhe et al. (2014) recorded 75% enhanced H₂ production by ultrasonically treating the FW of TS content72,500 mg/L at 1200 W for 5, 10, and 15 min depending upon the specific energy input. Elbeshbishy and Nakhla (2011) also recorded increased efficiency in FW degradation along with an increase in H₂ and CH₄ production rates. But, ultrasonic pretreatment is energy-intensive and expensive hence, requires optimization of operational parameters to set up a full-scale plant (Pilli et al. 2011).

Moreover, commercial enzymes (glucoamylase, proteases, viscozyme, and lipases) have also been used to enhance hydrolysis (Moon and Song 2011; Uçkun Kiran et al. 2015; Donoso-Bravo et al. 2015). According to a study, a mixture of enzymes resulted in higher VFA production than the use of a single enzyme (Kim et al. 2006a). But, the cost of treating 1 ton of FW using glucoamylase is around USD 120 which is expensive (Uçkun Kiran et al. 2015). Therefore, microorganisms such as fungal biomass that is rich in glucoamylase and protease have been used for the efficient production of biomethane from FW (Uçkun Kiran et al. 2015).

Several researchers have reviewed and compared various pretreatment technologies for substrate and inoculum within the biorefinery framework (Galbe and Wallberg 2019; Bhatia et al. 2020; Karthikeyan et al. 2018; Banu et al. 2020). Although thermal pretreatment of FW has been commonly practiced because of its simplicity and cost-effectiveness, not a single pretreatment method could be termed as the “best”(Galbe and Wallberg 2019). The choice of pretreatment technologies depends upon the desired end product and the FW composition (Banu et al. 2020). Some of the researchers have also integrated more than two pretreatment methods (such as acid-thermal) to optimize the performance of the two-stage AD process.

Utilization of VFAs for polyhydroxyalkanoates (PHAs), polylactate, and microalgal biomass production.

With the increasing concern for sustainability and the environment, bio-plastics are becoming popular. By 2025, global demand for bio-based plastics will increase to 2.87 million tonnes from 2.11 million tonnes in 2020 (European Bioplastics 2018). One of the precursor chemicals used for biodegradable plastic production is polyhydroxyalkanoates (PHAs). PHAs are a type of bio-polyesters accumulated by different bacterial cells in the form of granules inside the cytoplasm. The unique thermal and mechanical behavior makes it a completely biodegradable and suitable alternative to plastics (Reis et al. 2011). Producing PHAs from VFA under low temperature, HRT conditions to prevent methanogenic activity has been known to be an affordable process (Beccari et al. 2009). Besides, using mixed microbial culture and VFA rich substrates like FW from AF and sewage sludge reduces the PHA production cost while optimizing PHA yield at the industrial level (Van Aarle et al. 2015; Serafim et al. 2004). Perez-Zabaleta et al. (2021) obtained the highest PHA concentration (3.3 g/L) and content (43.5% w/w) from VFA effluent rich in caprionic and acetic acid, respectively.

Within VFA, butyrate was recorded to produce 57% of polyhydroxybutyrate (PHB) which is higher than with acetate at 54% in sequence batch reactor (SBR), making it the desirable substrate for PHBs (Marang et al. 2013). The production of PHA from organic acid derived from the OFMSW has also been studied for a pilot-scale anaerobic percolation bio cell reactor (100 L). The study showed a PHA production of 223 ± 28 g/kg with a hydroxybutyrate/hydroxyvalerate (%) ratio of 53/47 under optimized acidogenic fermentation (Colombo et al. 2017). Though FW is a potential source to produce PHA, it could be an expensive method considering the pretreatment of FW, transportation cost, and microbial strain selection (Nielsen et al. 2017).

The DF effluents containing VFA can be also used for the production of microalgal biomass (Fei et al. 2015; Turon et al. 2015; Ren et al. 2014). Microalgae are unicellular microorganisms that can be a rich source of carbon compounds. The use of microalgae is increasing especially in the industrial sector as biofuels, by-products, health supplements, pharmaceuticals, and cosmetics (Chalima et al. 2017). They have also been applied for wastewater treatment and CO₂ sequestration (Krishna et al. 2012). According to Lacroux et al. (2020), the growth of a mixed group of microalgae depends upon the concentration of VFA determined in the ranges of 71–207 mg/L and 13–25 mg/L as threshold concentrations for microalgae growth in acetic acid and butyric acid, respectively.

VFA is a byproduct derived from the fermentation of FW and thus, is known to be a cost-effective carbon source. This could be a possible alternative to expensive refined glucose for large-scale microalgal production. Fei et al. (2011) carried out a preliminary cost analysis on using VFA for cultivating C. albidus for biodiesel production. The study showed that VFAs-based biodiesel production was affordable compared to the use of agricultural products for lipid accumulation. In another study, Fei et al. (2015) used VFA as a carbon source to yield 0.187 g/g lipid coefficient and 48.7% lipid content of Chlorella protothecoides.

Moreover, the microalgae can directly convert the VFA into acetyl-CoA by acetyl coenzyme-A synthetase which could be used for fatty acid biosynthesis and lipid accumulation. Therefore, the use of VFA for microbial mass production could be favorable in terms of time and money. However, limited researches are available on the use of VFA derived from organic waste for microbial cell growth and especially lipid production. Most of the researches has focused on VFA production from microalgae. Therefore, there is a huge scope for further process development on using VFA for microbial-derived protein production for animal feed and energy biomass.

Economic prospects of biohythane and biochemicals

Besides the optimization of the two-stage AD process, some studies covered the economic viability for the production of biohythane and biochemical (Urbaniec and Grabarczyk 2014; Han, et al. 2016a, b; Bastidas-Oyanedel and Schmidt 2018). Bastidas-Oyanedel and Schmidt (2018) compared the economic advantage of combining DF with LA fermentation technologies for the conversion of 50 tonnes/day of FW into CH₄, power generation, LA, PLA, H₂, acetic acid, and butyric acid. Besides the power generation, all other alternatives were profitable. The highest profit of 296 $/tonne VS was obtained from DF with the separation and purification of butyric acid.

Similarly, the novel biohythane process (pretreatment, DF with Caldicellulosiruptor saccharolyticus and AD) resulted in high H₂ productivity with 6.1 L/L/d of pure biohythane and 69% of energy recovery from the sugar fraction of the wheat straw. But, the sparging of DF has high energy demand leading to the increased operational cost. Moreover, the addition of nutrients for high H₂ productivity also added to the total production cost of 160 Euro/GJ biofuel (Willquist et al. 2012). A two-stage AD of sugarcane (vinnase) showed 20–30% better economic performance than single-stage AD (Fuess et al. 2018). The net present value (NPV) was found to be $208.58–219.86 million. Besides, optimizing the alkalization of methanogenic systems improves both the economic and environmental performance of AD plants by reducing operating costs and the risks associated with human toxicity and freshwater eutrophication (Fuess et al. 2018).

Another large scale (50 m³) DF was found to be economically feasible as its return on investment was higher than the DF of 10 m³. The net present value, payback period, and internal rate of return with a scale of 50 m³ were USD 526,551, 6.9 years, and 9.25%, respectively. However, the major drawback of DF was the low hydrolysis rate and nutrient conversion efficiency (Han, et al. 2016a, b). Besides AD, DF has been integrated with solid-state fermentation (SSF) reactor to produce H₂ from FW at a large scale. Considering the waste from the food industry, the H₂ production cost is comparatively low (Table 6). The unit cost of an H₂ production plant with a capacity of 10 tonnes/day and a lifetime of 10 years was USD 2.29/m³ (25.45 USD/kg) which was less than the market price of USD 2.7/m³ (30.0 USD/kg) (Han, et al. 2016a, b).

Table 6 H₂ production cost within food industry based on different processes and pretreatment

Similarly, Urbaniec and Grabarczyk (2014) compared the H₂ production cost under three different scenarios: base, update, and optimistic whose values are based on knowledge, research, and considering two-stage will be fully developed in future, respectively. The H₂ production cost (9.3 Euro/kg) was found lowest in the optimistic case.

According to Ljunggren and Zacchi (2010), the main contributors to the H₂ production cost are the capital costs and the nutrients added. However, with cheap FW as substrate and effluent recirculation in two-stage AD can remove these hurdles. Recently, Li et al. (2020) demonstrated that a large-scale two-stage AD process that treats 216 tons of FW daily is technically and economically feasible. The research team estimated the total investment cost of large scale to be 24.34 million Yuan (1 Yuan is equivalent to 0.15 USD). And, with the selling price of biogas as 0.46 USD/m³, the total investment can be paid back in 1.5 years. Similarly, the advantage of the biorefinery system converting waste to multiple products (energy, biochemical, & nutrients) and its environmental benefit, trade-off the cost in long run (Krishnan et al. 2019). Nevertheless, more research on the techno-economic study of the two-stage conversion of FW has to be realized in future.

Currently, the industrial production of VFAs is mainly conducted using chemical processes like the oxidation or carboxylation of precursors such as aldehyde and alkenes. However, the continual use of energy-intensive processes contributes to CO₂ emissions and ecological imbalances (Sivagurunathan et al. 2018). The feasibility study of DF conducted in Abu Dhabi showed DF as a promising biorefinery technology not only for the treatment of OFMSW but also for VFA production. The VFA is a versatile chemical as it can be further developed into biofuel or bioplastic in the downstream process. Given a waste tipping fee of 22 USD/tonnes of OFMSW_, VFA could be used in downstream processes without any charge. This could make advancement and investment in technology a viable option. The maximum cost of VFA production from DF was calculated to be 15 USD/m³ effluent (Bonk et al. 2015).

The market size of organic acids in 2013 was reported to be 2.9 million tonnes with a market size of 3.5 billion USD. The market annual growth rate (AAGR) until 2020 is 8.8% (Sun et al. 2019). Table 7 shows the bulk market price and size of the direct DF products and value-added products that can be produced utilizing the DF effluents. Likewise, the bio-polymers (polylactic acid/PLA, PHA) are showing moderate market growth of 9–30% per annum with good future demand for the year 2023 (Sun et al. 2019).

Table 7 Bulk prices and market size of major DF products (Adapted and modified from (Bastidas-Oyanedel et al. 2019))

The microbial pathway for biochemical production also influences the cost and yield (Manandhar and Shah 2020; Francois et al. 2020). A study for large-scale (10,000 metric tons per annum) production of lactic acid (LA) from corn determined the production cost of $1181, $1251, and $844 through bacteria, fungi, and yeast pathways were, respectively. The lowest production cost from yeast was due to the lower requirement of chemicals, equipment, and utilities for neutralization and recovery of LA. Moreover, the influencing factors for LA production cost are sugar-to-lactic-acid conversion rates, grain price, plant size, annual operation hours, and gypsum use (Manandhar and Shah 2020). Similarly, the chemical structure of feedstock is another factor affecting the production cost. Based on the study by Daful and Görgens (2017), the cellulose fraction of sugarcane bagasse and brown leaves (SCBL) is economically and environmentally more viable than the hemicellulose fraction of SCBL.

Although FW is a potential source to produce organic acids and biohydrogen, it could be an expensive method considering the pretreatment of food waste, transportation cost, and microbial strain selection (Nielsen et al. 2017; Yun et al. 2018). However, Monlau et al. (2015)suggested that alkaline pretreatment could be economically sustainable if it is applied at a high substrate concentration and/or maximum heat is recovered during pretreatment. Similarly, Colombo et al. (2017) recommended that optimizing the acidic fermentation process (Eg: recirculating the digestate) could support the organic acid formation from OFMSW at a large scale. Also, DF consumes less energy than other conventional technologies for H₂ and VFA production (Bonk et al. 2015). Therefore, the two-stage AD biorefinery process from FW can be a sustainable route to achieving economic benefits with the least environmental impact (Dahiya et al. 2018).

Conclusion

Biorefinery framework considering two-stage AD process is a promising technology especially DF for H₂ production along with biochemicals that have economic values. The selection of operational parameters depends upon the targeted VFA. Acetate and butyrate are the most abundant VFA produced from FW. Other parameters, OLR and RR also affect the stability of the reactor, biohythane and biochemical production. Generally, OLR and HRT are relative which means low HRT and high OLR enhance the biohythane and VFA production. To enhance the performance of two-stage AD, FW can be pretreated. However, the choice of pretreatment methods and optimum operational conditions depend upon the desired product. The DF-AD process integration also depends upon the TS content of the substrate, type of bioreactors, and coupling ratio.

Despite the high capital investment in two separate bioreactors systems, the DF-AD process can be economically viable if multiple fuels and biochemicals are recovered. Moreover, the market size for VFA production is expanding with an 8.8% growth rate. The two-step process can be a sustainable route to produce biohythane and biochemicals such as butyrate and PHA that have high economic values trades off the cost. Besides the use of biohythane in the industrial sector, the VFA can also be used to produce bioplastic and cultivate microalgae for further energy recovery. Furthermore, various bioreactors configuration within the DF-AD process for FW that can withstand the fluctuation in operational parameters for a longer-term needs to be explored. Similarly, a techno-economic analysis of biohythane and biochemicals recovery from the different FW compositions while optimizing the two-stage AD process needs to be conducted. This would help to determine the viability of a two-stage AD process with FW as a feedstock at a commercial scale.