Relationship between the synergistic/antagonistic effect of anaerobic co-digestion 1 and organic loading

: Results from this study reveal a notable relationship between the 11 synergistic/antagonistic performance of sewage sludge – food waste anaerobic co-digestion 12 (AcoD) and organic loading. At the same sewage sludge content, biomethane potential (BMP) 13 assays show an increasing specific methane yield as the content of food waste increased to the 14 optimum organic loading of 15 kg VS/m 3 . Under these conditions, the specific methane yields 15 experimentally measured in this study were considerably higher than those calculated by 16 adding the specific methane individual co-substrates during mono-digestion. On the other hand, 17 at above the optimum organic loading value, the antagonistic effect (i.e. lower specific methane 18 yield compared to mono-digestion) was observed. The relationship between synergistic 19 performance of AcoD and organic loading was also evidenced in the removal of volatile solids 20 as well as chemical oxygen demand. Further analysis of the intermediate products show that 21 methanogenesis was the rate limiting step during AcoD at a high organic loading value. As the 22 organic loading increased, the digestion lag phase increased and the hydrolysis rate decreased. 23


26
Sewage sludge is a solid by-product from municipal wastewater treatment. Because sewage 27 sludge is rich in biodegradable organics and pathogenic agents, adequate treatment is necessary 28 prior to disposal or any form of land applications (Semblante et al., 2014). Given the large 29 amount of sewage sludge generated each day, sewage sludge management has become a major  Other benefits include the dilution of toxic compounds, improve nutrition balance, and load 53 increase of the biodegradable organic matter (Sosnowski et al., 2003). 54 A range of organic wastes is available for AcoD operation. Among them, food waste is 55 arguably the most abundant substrate that is also rich in energy (i.e. carbon) and nutrient 56 content (Thi et al., 2016). In general, food waste consists of 10-30% readily biodegradable 57 organic materials (Ratanatamskul & Manpetch, 2016;Zhang et al., 2016;Zhang et al., 2007). 58 Given the high organic content of food waste, AD has been identified as an ideal solution for 59 energy recovery from food waste. In addition to the many benefits of AcoD discussed above, reported a decrease in methane production by more than 40% during thermophilic AcoD of 67 sewage sludge and grease waste when the content of grease waste increased from 27 to 37% at 68 the same organic loading. Their results demonstrate an antagonistic effect possibly due to fatty 69 acid inhibition (Silvestre et al., 2014). In another study, Silvestre et al. (2015) did not observe 70 any changes in the specific methane yield during mesophilic AcoD of sewage sludge and crude 71 glycerol at more than 1% (v/v) co-substrate addition. Given the inconsistency in the literature 72 regarding synergistic effect during AcoD, it is hypothesised here that organic loading can play 73 a major role in governing the specific methane yield.

74
In practice, organic loading is a key parameter in the continuous operation of AcoD (Mata-75 Alvarez et al., 2014). In a batch process, organic loading can be defined as the ratio of either 76 VS or COD content over volume. In a continuous process, the retention time is taken into 77 account and the organic loading rate (OLR) can be used instead. Mono-digestion of sewage 78 sludge at WWTPs is usually operated at an OLR of less than 1 kg VS/(m 3 .d) (Nghiem et al.,79 2017). On the other hand, given the high organic content of the co-substrate (particularly food 80 waste), AcoD is operated at a much higher OLR value of up to 4.6 kg VS/( m 3 .d) (Nghiem et   The aim of this study is to explore the relationship between organic loading and the synergistic     Table 1. When the substrate 123 volume was less than 300 mL, Milli-Q water was added to obtain the total volume of 750 mL.

124
After filling with inoculum and substrates, the BMP bottles were flushed with N2 again, sealed 125 with rubber stopper instantly, and placed in the water bath, which was maintained at 35 °C.

126
The gas valves were then opened to allow biogas from entering to the gas collection gallery.

127
The BMP experiments were terminated when the daily methane production during three 128 consecutive days was less than 10 mL. All BMP bottles were mixed manually twice a day.

129
The BMP protocol used in this study is broadly consistent with the standard procedure Where P is the maximum methane potential (mL); M is the cumulative methane production 144 (mL); Rmax is the maximum methane production rate (mL/d); λ is the lag phase (d); e is Euler's 145 number (≈2.71828); and t is the time (d).   177 Where Ysp is the specific methane yield (mL); Ysub is the total methane production from the 178 substrate (mL); Yin is the total methane yield from the inoculum, which was 1145 mL, and

179
VSadded is the mass VS added from the substrate in the BMP bottle (g).

180
The calculated methane yield from a mixture of sewage sludge and food waste could also be 181 obtained from the specific methane yield of each individual substrate without taking into 182 account any synergistic effect: Where Ycp is the calculated methane yield (mL methane/g VSadded); VSFW is the VS added from     Further evidence of the synergistic effect of food waste and sewage sludge co-digestion as well 237 as the dependence of the synergistic effect of co-digestion on organic loading can also be seen 238 in Table 3. The specific methane yield of co-digestion between sewage sludge with either 30 or 239 70 g experimentally obtained in this study was 30-40% higher than the calculated value from 240 mono-digestion of each individual substrate by ignoring the synergistic effect (Eq. 4). Organic 241 loading is a major factor under these experimental circumstances. It is noted that I/S ratio and 242 pH may also impact the specific methane yields (Hashimoto, 1989;Jayaraj et al., 2014). By 243 contrast, antagonistic effect was observed for 110 g and 150 g food waste co-digestion with 244 sewage sludge due to organic overloading. In these two BMP tests, due to organic overloading, 245 the specific methane yield was even lower than that from mono-digestion (section 3.3.3). A    The modified Gompertz model was used to simulate the digestion process. As noted in section 292 2.3.1, the lag phase (λ) and the ultimate specific methane yield could be obtained by fitting data 293 presented in Figure 2 to the Gompertz model. The ultimate specific methane yields obtained 294 from the Gompertz model (Table 4) were consistent with experimentally obtained values 295 previously presented in Table 3. Similar results at a comparable organic loading level have also 296 been reported by Xie et al. (2017).   TS, VS and soluble COD removals in the co-digestion bottles were higher than those in the . It is noted that the production and consumption of soluble COD can occur 315 simultaneously, thus, data in Table 5 represent the overall balance of soluble COD in the   The Kh of the hydrolysis process was determined using Eq. (2) and cumulative methane 326 production data presented in Figure 2a. K h decreased as the organic loading increased (Table 6).

327
In other words, the hydrolysis rate decreased with increasing organic content. These results are

334
This study shows that the synergistic/antagonistic performance of AcoD between sewage 335 sludge and food waste was dependent on organic loading. At the same sewage sludge content, 336 the specific methane yield increased as the content of food waste increased to the optimum 337 organic loading of 15 kg VS/m 3 . At or below this optimum organic loading, the experimentally 338 obtained specific methane yields were notably higher than those values calculated by adding 339 the specific methane yields of individual co-substrates during mono-digestion. On the other 340 hand, at an excessive organic loading value, the antagonistic effect (i.e. lower specific methane 341 yield compared to mono-digestion) was observed. The interplay between synergistic 342 performance of AcoD and organic loading could also be seen in the removal rates of VS as 343 well as COD. Results from intermediate product analysis also suggests that methanogenesis 344 was the rate limiting step during AcoD.