The effect of anaerobic co-fermentation on acidification performance of food waste and cardboard waste Open Access

According to statistics, about 1.58 × 10⁸ tons of food waste (FW) were produced in China in 2017 (Xiaoming et al. 2019), and the total production of FW had achieved 2.10 × 10⁸ tons in 2018 (Kaijun et al. 2020). Although FW production increased annually, only less than 20% of FW could be recycled (Zhou et al. 2018), while a large amount of FW could not be effectively utilized. In the transportation and treatment of FW, it bred mosquitoes and contaminated water and the atmosphere (Longsheng et al. 2011). As the traditional treatment technique for FW, landfills led to a large amount of leachate, which caused serious environmental pollution (Laiqing et al. 2009). Meanwhile, the percentage of cardboard waste (CB) in municipal waste was around 20% (Wanwu 2018), accounting for 40% to 50% of landfill space (Yuan et al. 2012). A large part of CB was not effectively utilized, which showed excellent energy potential.

Anaerobic fermentation, widely utilized in recent years, was one of the best treatment technologies to reduce harmless and solid waste (Yue et al. 2018). In particular, dry anaerobic fermentation could deal with a large number of raw materials, save water resources, and have a high volume gas production rate (Jiadong et al. 2018). Moreover, fermentation of FW could be operated under high solid-state (10–20% total solid (TS)) (Lili 2016). Since the nutrition imbalance in the fermentation system efficiently contributed to acidification (Chen et al. 2011), it was easier to produce volatile fatty acids (VFAs) accumulation in high solid-state fermentation, which induced the acidification and unstable operation of the reaction system (Liu et al. 2006). The total solid concentration of FW was as high as 20%, in which volatile solids accounted for more than 80% (Gang 2016). These volatile solids were mainly carbohydrates (50–70%), easily degradable (Uçkun Kiran et al. 2014). With the action of amylase, the rapid degradation resulted in the continuous acidification of fermentation broth (Longsheng et al. 2012). In addition, the high protein content of FW caused malnutrition of fermentation substrate and low carbon nitrogen ratio (C/N) (Lei et al. 2016). To avoid these problems, mixing raw material fermentation was used to adjust C/N (Jinping et al. 2018), which could be applied in FW fermentation to slow the acidification process. In the co-fermentation of FW, sludge and straw, the C/N ratio was optimized, the gas production performance was improved (Xin et al. 2020), and the co-fermentation of sawdust and FW could also promote microbial colony formation (Oh et al. 2018). Yenigün et al. (Yeniguen & Demirel 2013) carried out the co-fermentation of seaweed CB to achieve the balance of C/N in the system and concluded that the optimal C/N was between 20 and 25. However, the optimal C/N ratio of anaerobic digestion with high solid content is 27–32 (Zeshan et al. 2012), making CB a new choice for high solid co-fermentation materials because of its suitable C/N, high TS content, and slow biodegradation characteristics (Capson-Tojo et al. 2018). Small amounts of CB promoted anaerobic fermentation of FW, which could form a better co-substrate for FW and CB fermentation.

Therefore, this study was based on the anaerobic digestion of FW, which focused on the effects of adding CB and different substrate concentrations on anaerobic digestion gas production and acidification in the fermentation process of FW. To effectively improve the gas production performance of raw materials and alleviate the impact of acidification, the reasonable ratio and optimal parameters of the two raw materials were obtained, which provided a technical basis for efficient utilization of anaerobic digestion of FW.

In this study, FW and CB were used as raw materials, and FW was taken from the fifth canteen of Shenyang Agricultural University. In addition to home meals, the canteen dishes also included fried snacks and local specialities, which were representative. The mixture was milled and blended to ensure its homogeneity. CB, obtained from waste express boxes, was shredded to less than 1 cm before using and kept in a well-ventilated room. The inoculum was collected from the supernatant of biogas slurry from an anaerobic biogas digester of No.36 greenhouse of Shenyang Agricultural University. The TS of the raw materials was determined after 8 hours' drying in the oven with the constant temperature at 105 °C; the volatile solid (VS) was measured after being burned in a muffle furnace at 550 °C for 2 hours; carbohydrates were extracted from the samples after grinding in a water bath; the C/N ratio of the samples was determined after air drying and decocting. The main characteristics of raw materials are shown in Table 1.

In this study, FW and CB were used as the mixed raw materials, and the ratio of FW to CB was set as 10:0, 9:1, 8:2 and 7:3 under the suitable C/N ratio for anaerobic fermentation. With four raw material ratios, the TS concentration was set at 12%, 15%, 18% after the initial TS concentration analysis.

According to the experimental conditions above, after adding the corresponding FW and CB into the container, biogas slurry was added to achieve the total mass of 500 g before mixing thoroughly. The 1-L glass reactor was operated at 35 ± 1 °C with working volume from 280 to 540 mL (according to the ratio of FW to CB). All the test conditions were repeated three times.

TS and VS were determined by the gravimetric method. Total nitrogen (TN) and total organic carbon (TOC) were determined according to organic fertilizer NY525-2012 standard. Carbohydrates were determined by the phenol sulfuric acid method. The volume gas production rate, pH value, VFAs and soluble chemical oxygen demand (SCOD) were measured daily. A portable pH meter (PHS-3G, Leici, China) was applied to detect the pH value in the fermentation process. The concentration of VFAs, namely acetic acid, propionic acid, n-butyric acid, isobutyric acid, isovaleric acid, n-valeric acid and n-hexanoic acid, was determined by gas chromatography (Agilent 6890N). SCOD was determined by spectrophotometry (Zhenhui 2003). Amylase was measured by calculating the amount of reducing sugar. The sample supernatant from each reactor was tested for amylase activity. The hydrolytic activity was determined by measuring the enzymatic release of maltose from starch. Determination of coenzyme F420 and calculation of the absorbance of the enzyme solution at 420 nm at different pH values were carried out through the formula used by Weimin & Qing (1984). The protease was determined by spectrophotometry, and the protease activity of the same supernatant was tested.

The pH value of the FW anaerobic fermentation acidification stage is a relatively intuitive parameter. If the pH value was not adjusted manually in the fermentation process, it would decrease in all reactors, as illustrated in Figure 1, and the change was more significant with a higher initial TS concentration. In the mono-digestion reactor, the pH value decreased with the increasing FW components, and the final pH value was lower than 5, which indicated that excessive acidification might significantly inhibit anaerobic fermentation (Zhang et al. 2015).

The pH value of co-digestion was relatively more stable than that of mono-digestion, which might be due to the addition of CB. As for the initial TS concentration of 12%, the ratio of FW to CB was 9:1, 8:2, and 7:3, and the corresponding initial pH values increased to 7.16, 7.57 and 7.97, respectively. The initial pH value followed the same phenomenon at other initial TS concentrations. The presence of CB in co-digestion resisted the initial pH drop and increased the buffer capacity, causing the acidification process to be more resistant to pH fluctuations, resulting in higher methane production than mono-digestion (Figure 3(a)). Li (Wanwu 2018) also found that the addition of CB could improve the initial buffer capacity of the system. The pH value of mono-digestion decreased rapidly to about 3.5, which was much lower than the pH value required for methanogens to survive (Juan et al. 2019). In the co-digestion reactor, with the increase of the percentage of CB, the pH value gradually increased to 6.11, which indicated that the percentage of CB had a significant influence on the pH value regulation in the anaerobic fermentation process.

Figure 2 shows the concentration of SCOD in the hydrolysate of anaerobic fermentation, where it could be found that the SCOD of all conditions increased rapidly in the next day. When the initial concentration of TS was 12%, the highest value was 18,600 mg/L on the second day of mono-digestion, and then the content decreased, maintaining a low degradation rate. With the ratio of FW: CB = 8:2, the highest content of SCOD was 29,433 mg/L, and the content of SCOD decreased significantly from the third day. When the initial TS concentration was 15%, the SCOD content in the reactor increased and maintained a high concentration. When the initial TS concentration was 18%, the changing trend of SCOD in the reactor was similar to that with other initial TS concentrations, but the range of change was not significant.

Soluble organic matter, the product of hydrolytic microorganisms, was consumed by fermentation microorganisms (Jing et al. 2018). In the hydrolysis stage, these fermentation microorganisms also consumed soluble substances, but their consumption rate was lower than that of hydrolysis microorganisms (Yuan 2011). In the early stage of mono-digestion, the content of SCOD was in the range of 6,100–18,600 mg/L, inducing the accumulation of organic matter and the inhibition of acidification (Yuan et al. 2016). The main material of CB was lignocellulose, which was hard to degrade. Conversion from lignocellulose to soluble organic matter was the rate-limiting step during anaerobic digestion of lignocellulose (Ruijie et al. 2020). Therefore, increasing the percentage of CB in co-digestion slowed down the hydrolysis rate of fermentation substrate. However, Figure 2 showed that with FW: CB = 9:1 and FW: CB = 7:3, the SCOD content was similar. This phenomenon might be because the excessive content of CB significantly limited the hydrolysis rate. In addition, the consumption of hydrolysates by fermentation microorganisms did not stop; thus, no hydrolysates could be used by fermentation microorganisms. This assumption was also supported by the maximum methane content in the co-digestion reactor. As illustrated in Figure 3, the maximum methane contents with FW: CB = 9:1 and FW: CB = 7:3 were similar when the initial TS concentration was the same. Therefore, adding an appropriate amount of CB contributed to reducing the consumption of soluble organic matter in the cultivation process of mixed raw materials, which can be employed for methane production. Hao Xin et al. (2020) observed that the cumulative gas production of 1:1 and 1:2 of FW and straw was significantly higher than that of the 2:1 group and mono-digestion. When the ratio was 1:1, the maximum gas production reached 868 mL, and the changing trend of the production was similar to that in this study. With the increase of straw content, the cumulative gas production firstly increased and then decreased.

The methane production is shown in Figure 3(a) and 3(b). High methane productions were achieved, with the highest value of 3.4 L/kg in the condition FW: CB = 8:2. In Figure 3(a), the methane production was linear with the initial TS concentration. However, methane production increased with the increase of TS concentration in the mono-digestion of FW and was negatively correlated with TS concentration in co-digestion reaction. High TS concentration has been found to reduce the hydrolysis rate of co-digestion, mainly composed of lignocellulose, also jeopardizing methane production (Brown et al. 2012). The decrease in methane production observed in the co-digestion system cannot be attributed to TS concentration. As displayed in Figure 3(b), when the initial TS concentrations were different, the methane yield was similar to the reaction of FW: CB = 9:1 and FW: CB = 7:3, and the yield achieved the highest level when the substrate ratio was FW: CB = 8:2. It can be inferred that addition of CB led to excessive amounts of cellulose, hemicellulose and lignin, which were complicated to degrade with much lower degradability than that of FW. The degree and speed of hydrolysis of organic matter in CB were also far lower than FW. Therefore, there was no significant contribution to the production of excess methane.

In the mono-digestion reactor, a possible explanation for the increase of methane production caused by higher TS concentration was that higher initial TS concentration increased the production of VFAs, which were the end product of hydrolysis acidification and the substrate mainly utilised by methanogens. There was no significant difference at the minimum pH value (3.44–3.64) in the hydrolysis stage. Hence, high TS concentration and high VFAs content also played a buffer role in the reactor and avoided pH decrease. This hypothesis was also supported by the pH value in the co-digestion reactor, as shown in Figure 1. In the same substrate ratio reactor, the pH value was lower with the decrease of TS concentration. Liu et al. (2011) observed the effect of VS concentration of FW on methane production and found that if VS concentration was less than 16 g/L, methane production increased with the increase of VS content, while if VS concentration was equal to 16 g/L, the maximum methane production reached 506.44 mL/g, and the changing trend of methane production was similar to that in this study. However, it must be considered that the above experimental studies aimed to explore the effect of pH regulation on biogas production from anaerobic digestion of FW, which was not the primary purpose of this study.

This study determined the concentration of VFAs and their components, including acetic acid, propionic acid, n-butyric acid, isobutyric acid, isovaleric acid, n-valeric acid, and n-hexanoic acid. As shown in Figure 4, the VFA yield was high, with the values for the total VFAs ranging from 6.12 to 17.04 g/L. The total VFAs yields mainly depended on the composition of the substrate, and the VFAs yield of the single fermentation of kitchen waste was similar to both the composition and the highest VFAs yield when a small amount of CB was added. Additionally, the higher percentage of CB in the substrate, the higher the TS concentration and the lower the VFAs yield. Two different behaviors (mono-digestion and co-digestion) were observed depending on the substrate. The higher TS concentration in the mono-digestion reactor resulted in a high concentration of FW, VFAs concentration increased rapidly, and pH value decreased, resulting in acidification of the anaerobic system. However, the co-digestion reactor had the opposite effect with an increase in TS concentration, contributing to a lower initial concentration of FW and increased buffer capacity. The major VFAs were similar in all reactions, including acetic acid and butyric acid, hexanoic acid, propionic acid, and valeric acid. Shen et al. (2016) found that the highest VFAs yield of about 0.763 g/(gCOD_removal) was obtained after 3% phosphoric acid pretreatment of FW; the main components of VFAs were acetic acid and butyric acid, followed by propionic acid and a small amount of valeric acid. It has also been found that acetic acid, butyric acid and propionic acid were the most common acidification products of food and paper in the organic components of municipal solid waste. The maximum yield of VFA at 42 °C was 21.5 g/L (Ghimire et al. 2015).

The composition of VFA components provided useful information about the metabolic pathways. Figure 5 shows the percentage of single VFA to total VFAs under the different substrate and initial concentration conditions during the 7-day fermentation time. The ratio of butyric acid to acetic acid could be used as the stability index of the anaerobic fermentation process (Soomro et al. 2020), and the value would increase at lower pH and higher VFA concentrations. Figure 5 displays that TS concentration positively correlated with acetic acid in the mono-digestion system, while butyric acid content firstly increased and then decreased with the increase of TS concentration, resulting in higher butyric acid yield and lower acetic acid yield at higher TS concentration, which implies that the butyric/acetic acid ratio increased at higher TS concentration conditions. In this experiment, the final concentration of VFA in the mono-digestion system improved with the increase of TS concentration, and the fermentation substrate was inhibited at lower pH, resulting in a shift of the acidification product to butyric acid. However, the increase in the butyric/acetic acid ratio did not cause significantly low methane yields. The higher initial TS concentration also resulted in more acid accumulation, suggesting that the final VFA concentration had a negative impact on the fermentation process.

On the other hand, in the co-digestion system, the final yields of propionic acid, butyric acid and valeric acid were less affected by the initial conditions, while the yields of acetic acid and caproic acid were relatively highly affected. In the three groups of solids concentration tests, the substrate percentage of CB was negatively correlated with the final pH, indicating that CB had a large buffering capacity. The percentage of acetic acid was relatively high, with low levels of CB percentage in the substrate and TS concentration. On the contrary, a higher percentage of CB in the substrate, a higher concentration of TS, and a higher percentage of hexanoic acid suggested that both variables might have promoted the synthesis of VFA. Because the high concentration of TS did not cause an increase in the production of hexanoic acid in the mono-digestion process of FW, it could be speculated that the addition of CB contributed to the increasing buffer capacity, which was more conducive to the formation of hexanoic acid under the condition that the pH was maintained at a high level (6.9–7.9). The percentage of hexanoic acid in the volatile acid increased to 26% in a reactor with a moderate percentage of CB (FW: CB = 8:2).

Anaerobic fermentation can be divided into various types based on the main VFA products except for acetic acid (Ren et al. 2015). As shown in Figure 5(a), in the acidification stage of mono-digestion of FW, the final concentrations of butyric acid and acetic acid gradually increased from 8.9 g/L to 11.1 g/L as the solids concentration increased, accounting for 83.88% of total acid at 12% solid concentration, 58.94% of total acid at 15% solid concentration, and 69.81% of total acid at 18% solid concentration, respectively, which was classified as butyric acid fermentation. As demonstrated in Figure 5(b)–5(d), the accumulation of butyric acid was more prominent with an increasing percentage of CB for mixed fermentation of the same solids concentration of FW. Especially if the initial concentration was 12%, the concentration of butyric acid reached 3.85 g/L on day 7 in the FW:CB = 7:3 group, which was 3.4 times higher than that of acetic acid. The other groups had different levels of butyric acid accumulation, with the proportion ranging from 40% to 80%, and propionic acid with less than 20%. Therefore, co-digestion could be considered as butyric acid type fermentation. In all reactors, the major products were acetic acid, propionic acid, isobutyric acid, and hexanoic acid. While the combined abundance of n-butyric acid, isovaleric acid and n-valeric acid was slight, there was no significant change in the fermentation process.

Experiments were carried out under conditions of different TS and ratios of FW to CB to study the amylase activity. Figure 6 shows that at the beginning of the co-digestion, the amylase activity of the FW:CB = 9:1, FW:CB = 8:2, and FW:CB = 7:3 groups ranged from 92.65 to 112.8 U/mL for different TS conditions, which was lower than mono-digestion. After hydrolysis occurred, amylase activity increased continuously, but remained below the level of the single fermentation of FW, resulting from the addition of CB limiting the rate of substrate hydrolysis. The maximum activity of 127.9 U/mL was obtained on day 4 at TS concentration = 18% and FW:CB = 7:3. From day 3 of the reaction, the amylase activity of the co-digestion gradually decreased to between 84.2 U/mL (TS concentration = 15%, FW:CB = 8:2) and 109.8 U/mL (TS concentration = 12%, FW:CB = 7:3) by the end of the reaction. Although there was a decreasing trend, the rate of decline was slower than in mono-digestion. The rate of amylase decline fluctuated with increasing proportion of CB in co-digestion, and at TS = 15%, FW:CB = 8:2, the amylase activity decreased from day 3 of fermentation to day 7 with a rate of 43.6 U/mL, which was lower than the others. This was because the content of CB limits the degradation of carbohydrates in FW while stabilizing its degradation environment. The results showed that adding an appropriate amount of CB in high-solids mixed anaerobic fermentation allowed the amylase to degrade rapidly and maintain the stability of degradation preferentially.

Protease activity was essential for the hydrolysis step of anaerobic fermentation, where proteases were used to achieve efficient biodegradation of proteins to intermediates or desired products (Kim et al. 2012). In this experiment, the protease activity of the filtered bulk methane in the reactor was determined by quantifying trichloroacetic acid. Figure 7 showed the relative changes in protease activity at each sampling point according to the fermentation time. In general, the protease activity of the co-digestion group was higher than mono-digestion. The protease activity of the FW:CB = 9:1, FW:CB = 8:2, and FW:CB = 7:3 groups with TS = 12% reached maximum values of 32.19 U/mL, 39.45 U/mL, and 43.79 U/mL on day 5; meanwhile, the co-digestion with TS = 15% and TS = 18% groups reached maximum values of 36.25 U/mL, 41.68 U/mL, 45.3 U/mL, 38.96 U/mL, 40.78 U/mL, and 46.21 U/mL on day 5. The protease activity in the co-digestion increased with the proportion of CB. The protease activity of TS = 15% and FW:CB = 8:2 decreased from 41.68 U/mL to 18.12 U/mL within 2 days, since the addition of CB regulated the C/N ratio and facilitated the degradation of proteins. There was a delay in reaching the maximum protease activity compared to the amylase activity of the same group. Protein degradation was slower than carbohydrates as the anaerobic degradation of proteins involved multiple microorganisms. Proteins were firstly hydrolyzed and degraded by protein hydrolases into peptides and individual amino acids, and then the peptides and amino acids were acidified into VFAs, ⁠, ammonium and reduced sulfur. VFAs were further converted to acetic acid and or CO₂ by acetic acid-producing bacteria, and both were finally converted to CH₄ by methanogenic bacteria. Initial hydrolysis was the rate-limiting step in protein degradation, with an overall slower degradation rate. This suggested that adding the appropriate amount of CB promoted the conversion of volatile acids, which accelerated protein degradation.

Coenzyme F₄₂₀ was widely found in methanogenic bacteria involved in methane formation. It could represent the activity of methanogenic bacteria and monitor the activity of methanogenic bacteria (Dong et al. 2010). Coenzyme F₄₂₀ could be sampled without maintaining anaerobic conditions and requiring time-consuming incubation tests and was relatively simple to analyze. The change of coenzyme F₄₂₀ in co-digestion is shown in Figure 8, with a more stable trend in coenzyme F₄₂₀ compared to the mono-digestion. The coenzyme F₄₂₀ concentration firstly decreased and then increased with the percentage of CB increasing at TS = 12% and TS = 15% conditions, and decreased to 0.059 mmol/L with an increasing percentage of CB at TS = 18% condition. The final coenzyme F₄₂₀ concentration was up to 0.11 mmol/L at TS = 15% and FW:CB = 8:2. This was because the addition of CB increased the buffer capacity in the reactor, providing conditions for the recovery of coenzyme F₄₂₀ activity and allowing the reaction to run properly. Stable coenzyme F₄₂₀ concentrations allowed coenzyme F₄₂₀ to act stably throughout the reaction, supported by the methane yield. Compared to mono-digestion, methane production increased, and there was no stagnation of gas production. Co-digestion could alleviate the problem of acid inhibition in high solids anaerobic fermentation of FW, which promoted methane production and led to increased methane production. These could suggest that adding the appropriate amount of CB could control the concentration of coenzyme F₄₂₀, which ultimately affected methane production.

In mono-digestion, the acidification was serious with the increase of initial TS concentration, and the maximum methane yield was 1.4 L/kg with high TS content. Adding CB into the co-digestion process caused lower methane production, but the acidification stage was stabilized by increasing the buffer capacity. The highest methane yield of 3.4 L/kg was obtained with a low percentage of CB. A high percentage of acetic acid could be achieved with a low percentage of CB in the substrate and a low TS concentration (also in mono-digestion reaction). On the contrary, a higher percentage of hexanoic acid could be achieved with a higher percentage of CB in the substrate and higher TS concentration. In the reactor with FW:CB = 8:2, the percentage of hexanoic acid in the volatile acid was up to 26%. The mono-digestion and the co-digestion could be classified as butyric acid fermentation. The ratio of FW to CB was the critical factor affecting acidification. Butyric acid was the main product of volatile acid during the fermentation process. Considering the yield of volatile acid and methanogenic efficiency, the optimum addition of CB was 20%, namely, FW:CB = 8:2. The production of volatile acids in mono-digestion was the highest, and the percentage of CB should be high enough to ensure the sustainability and balance of the anaerobic fermentation process. This study showed that the ratio of FW to CB could be used as an easily controlled parameter to produce high value-added products.

This work was supported by the Liaoning Provincial Natural Fund (Grant No. 2021-MS-228); the Liaoning Provincial Education Department Project (LJKZ0692).

All relevant data are included in the paper or its Supplementary Information.