Dissolved organic matter and bacterial population changes during the treatment of solid potato waste in a microbial fuel cell

Potato is the most important food crop after wheat, corn and rice, and is the staple food for 1.3 billion people (Stokstad 2019). In 2018, the global production of potato exceeded 368 million tons (FAOSTAT 2019). Large amounts of potato waste are generated every year across the world during cultivation, storage, processing and consumption (Mulugeta et al. 2020; Yan et al. 2020). Processed products such as French fries, chips, hash browns, purée and frozen food cause waste in the form of potato peelings ranging from 15% to 40% of the original fresh weight (Sampaio et al. 2020). The potato starch, flour and canning industries are also responsible for the production of large amounts of peel waste, whose disposal raises great environmental concerns (Pathak et al. 2018). Potato waste contains a high content of biodegradable organic matter, which is a renewable biomass source that can be used for energy generation. A microbial fuel cell (MFC) is a device that utilizes microorganisms as a catalyst to convert chemical energy in the biodegradable organic matter into electricity (Moharir & Tembhurkar 2018). Our previous study on the direct treatment of solid potato waste by MFC illustrated that the size of potato pieces is an important factor that greatly affects the performance of an MFC (Du & Li 2016). Large pieces of solid potato require a longer hydrolysis time, thus leading to an obvious time lag in electricity generation and significant reductions of current density and columbic efficiency (Du & Li 2016). Effective methods and operating conditions that can enhance the hydrolysis of solid potato and its treatment efficiency by MFC are needed which in turn must be tested experimentally.

Waste activated sludge (WAS) is a major byproduct of biological wastewater treatment and its generation is increasing due to increasing population and rapid urbanization. WAS has a high organic matter content and requires proper handing due to being biodegradable and its adverse impacts on the environment. More effective approaches that can replace the conventional ones (such as landfilling, incineration, direct dumping, etc.) to realize simultaneous WAS stabilization and energy and/or resource recovery are greatly needed (Shen et al. 2015; Cai et al. 2016).

WAS is mainly composed of microorganisms. Since anaerobic culturing may increase the presence of hydrolytic bacterial species in the rich microbial consortia of WAS (Du & Li 2017a), it is reasonable to infer that, if anaerobically cultured WAS is added to the solid potato as a mixed feedstock for an MFC, the hydrolytic microbes in the sludge may join anaerobic bacteria present in the MFC to promote hydrolysis of solid potato. To validate this, a comparative study under well-controlled experimental conditions is necessary.

Meanwhile, it is worth noting that, in addition to the positive effect expected from the addition of WAS, a concurrent negative effect may also occur, because the rich heterotrophic bacteria from the sludge can probably compete with the electrogenic bacteria cultured in advance in the anodic chamber for the hydrolysis products of solid potato, thus causing a decrease in the efficiency of electricity generation. This is very likely since, like anaerobic digestion that involves bacteria species with different functions (Foladori et al. 2010a; Du & Li 2017a), different bacterial species may exhibit different uptake rates for bioavailable organic substances in the reactor (Du & Li 2016). It is thus conceivable that when an MFC is used to treat solid potato mixed with WAS, its performance may differ with their mixing ratios. This means that whether the products of the enhanced hydrolysis of the solid potato by the hydrolytic bacteria in the added sludge are preferentially used by the electrogenic bacteria or by the large number of ordinary heterotrophic bacteria brought into the reactor by the added sludge may determine the overall performance of the MFC, judged by the efficiency in both waste stabilization and electricity generation. The effect of WAS on the treatment efficiency of solid potato was evaluated in terms of electricity generation as well as the occurrence and behavior of dissolved organic matter (DOM) in our previous study (Du & Li 2017b). But further clarification is still required before MFCs can be used in the practical treatment of solid organic waste and WAS.

DOM in the liquid phase of the anodic chamber of an MFC during the treatment of solid potato and WAS is a complicated mixture of different organic substances. Understanding the composition of DOM is very important for selection of optimal physicochemical methods for the treatment of the final solution remaining in the reactor (Ramesh et al. 2006; Du & Li 2017a). As a quick and relatively new analytical approach, fluorescence excitation-emission matrix (EEM) spectroscopy is increasingly used in the characterization of DOM in water and wastewater before and after various physicochemical and biological treatment processes (Chen et al. 2003; Ramesh et al. 2006; Hambly et al. 2015; Du & Li 2017a). By applying this analytical approach, more findings related to the hydrolysis and use of solid potato in MFCs may also be expected.

Accordingly, the main objective of this study was to investigate the effect of mixed feeding of anaerobically cultured WAS on the performance of an MFC in the treatment of solid potato waste. For this, four two-chamber MFCs inoculated and cultured with electrogenic bacteria in advance were operated in parallel for nearly 90 days to treat solid potato alone, WAS after anaerobic culturing alone and their mixtures at two different mixing ratios, respectively. The effect was evaluated in terms of the composition variation of DOM analyzed by fluorescence EEM spectroscopy, as well as the densities of viable and total bacteria remaining in the anodic solution of the reactors at the end of the operation quantified by flow cytometric analysis. Principal component analysis (PCA) was also performed to evaluate the interrelations among current density, DOM concentration and the fluorescence EEM-based DOM constituting fractions. To the authors' knowledge, after an intensive literature review, this is the first study focusing on the effect of WAS on the performance of an MFC treating solid potato waste; the results obtained may thus be of great reference value for the treatment of solid potato waste by MFC.

White potato used in our previous study (Du & Li 2016) was used in this study. Soon after being purchased, the fresh potato was cut into 5 mm cubes for direct use based on the results of our previous study on the treatment of potato cubes of different sizes by MFC (Du & Li 2016).

The returned activated sludge of a municipal wastewater treatment plant in the city of Gifu, Japan was used as the source of WAS for this study (Du & Li 2017b). The fresh sludge obtained was allowed to settle for 4 hours and the settled sludge was anaerobically cultured using glucose, sodium phosphate monobasic (NaH₂PO₄) and ammonium chloride (NH₄Cl) in the dark without aeration for 1 week. Immediately after culturing, the sludge was washed using pure water with five times' repeated mixing and centrifugation to remove residual nutrients. It was then fed to the corresponding MFC reactors. Pre-culturing caused a slight reduction of organic matter content in the sludge from 75.7% to 75.5% (as measured after incineration in an oven at 600 °C).

Four two-chamber MFCs assembled by Feng et al. (2010), hereafter called MFC 1 to MFC 4, were used (with a total volume of 250 mL and the working volume of 240 mL for each chamber). For both the chambers, there was an inlet, an outlet and a sampling port and a port for aeration was added to the cathode chamber. The schematic diagram of the MFC configuration is provided in Figure 1. Two-chamber MFCs were chosen because, even if the efficiency in electricity generation is generally reported to be lower than for single-chamber MFCs, they are considered to be better for investigating the fundamentals relating to reactions and operation conditions (Feng et al. 2010; Venkata Mohan et al. 2010); and also for the reason that the cathode chamber can allow the simultaneous removal of some pollutants through oxidative reactions via aerobic bacteria or reductive reactions via electrons transferred from the anode chamber, such as ammonium or its oxidized product, nitrate (Virdis et al. 2010; Feng et al. 2015).

Anaerobic bacterial consortia were cultured in advance with sodium acetate after inoculation from an MFC operated in our previous study (Du & Li 2016). The feeding conditions for potato, WAS and their mixtures in the four MFCs are shown in Table 1. MFC 1 and MFC 2 were fed with 5.0 g (wet weight) of potato and WAS, respectively. For MFC 3 and MFC 4, in addition to 5.0 g of potato (wet weight), 2.5 g and 1.25 g of WAS (wet weight) were added, respectively, to make the initial mixing ratio of the WAS to potato in these two reactors 1:2 and 1:4. The moisture content of potato and WAS was 96.0% and 94.6%, respectively. The initial chemical oxygen demand (COD) loading in the anodic chambers of the four MFCs were 3360, 1,569, 4,145 and 3,752 mg/L, respectively.

All four MFCs were operated in parallel with an external resistance of 100 Ω for 87 days under the controlled temperature of 30 °C in an incubator. Every 4–6 days, the pH of the anodic solution was adjusted to about 7.0 using 1 N hydrochloric acid (HCl) or sodium hydroxide (NaOH). The cathode chamber of the MFCs was fed with the phosphate buffer solution (PBS) and an air pump provided continuous aeration (Du & Li 2016).

The fluorescence EEM of DOM in the anodic chamber was analyzed for collected samples after filtration through 0.2 μm membrane filter by using a spectrofluorometer (RF-5300, Shimadzu, Japan). The excitation and emission scans were performed between 220 and 550 nm at increments of 5 nm and a bandwidth of 1 nm. The fluorescence EEM obtained was normalized into the quinine sulfate unit (QSU) by dividing the fluorescence intensity values of all samples with the fluorescence intensity value of 10 ppb quinine sulfate (in 0.05 M sulphuric acid (H₂SO₄) solution) at the designated excitation/emission wavelengths (E_x/E_m) of 350/450 nm (Lee et al. 2015). To minimize the inner filter effect, all samples were diluted using pure water to total organic carbon below 2 mg/L (Lee et al. 2015).

PCA was performed by using SPSS software (IBM SPSS Statistic 21.0, IBM Corp., Armonk, NY) to better understand the interrelations among current density, the indexes reflecting the DOM concentration and composition, including soluble chemical oxygen demand (SCOD) and the fluorescence intensity of the major fluorescence peaks appearing in the fluorescence EEM during the whole MFC treatment process. For this, all measurement results of each MFC were used as the pooled data for analysis after standardizing the obtained data for each index.

The bacterial population in the anodic solution at the end of the MFC operation was analyzed in triplicate using a flow cytometer (CyFlow^R Cube 6, PARTEC, Germany). The samples of the anodic solution were gently vortexed to disaggregate the cells before analysis. SYBR Green I and 5(6)-carboxyfluorescein diacetate (CFDA) were used separately as the dyeing agents for staining total and viable cells, respectively, by following the protocol documented by Foladori et al. (2010a). The concentration of dead bacteria was the difference between the total and the viable cells.

The changing profiles of current density are plotted in Figure 2. For MFC-1 fed with the solid potato alone, current density increased gradually and reached its optimum value of 36 mA/m² after 12 days' operation. This level was maintained for about 20 days before current density gradually decreased until the end of the operation.

Compared to MFC-1, the current density of MFC-2 fed with the sludge alone was very low (the optimum being about 5 mA/m²) and was observed only for the initial 5 days. For MFC-3 and MFC-4 fed with the mixtures of solid potato and sludge at two different ratios, it is very interesting to notice that the current density of MFC-4 fed with 1:4 of the sludge was markedly higher than MFC-1 fed with the potato alone, indicating that some heterotrophic microorganisms in the cultured WAS were involved in the hydrolysis of the solid potato, and hence generated more hydrolysis products for electricity generation by electrogenic bacteria. The current density of MFC-3 fed with 1:2 of the sludge was much lower, with its optimum value of 10 mA/m². This demonstrates that mixed feeding of WAS could enhance the electricity generation of solid potato, and that an optimal mixing ratio may exist. A lower proportion of WAS could significantly promote electricity generation; while a higher one could significantly inhibit electricity generation. For MFC-4, the current density reached the optimum value of 150 mA/m² soon after the start of the operation and then maintained it at around this level for nearly 52 days before a sharp drop was observed after the long stable electricity generation period.

The significantly lower current density observed for MFC-2 fed with the WAS alone may imply that the bacteria brought into the reactor from the sludge, rather than the pre-inoculated electrogenic bacterial species in the reactor, might have preferentially utilized the hydrolysis products of the sludge, leaving a very limited amount available for electrogenic bacteria (Du & Li 2017b). The significant differences in current density observed for MFC-3 and MFC-4 treating the solid potato mixed with different amounts of WAS may also serve to support this inference. For MFC-3, with the larger amount of WAS, the increased hydrolysis products were probably preferentially used by the larger number of the sludge-originating bacteria, rather than by the electrogenic ones. However, for MFC-4, with the smaller amount of WAS, the increased hydrolysis products were probably more effectively used by electrogenic bacterial species, thus leading to obvious current density and power density increases.

The possibility that some bacterial species of the anaerobically cultured WAS had the ability to inhibit the growth and activity of electrogenic bacteria in the anode chamber could not be excluded. This is worth considering since Nor et al. (2015) confirmed in their study the presence of rich bacteria in the anode chamber fed with anaerobic palm oil mill effluent (POME) sludge as the sole bacterial source for the treatment of POME by MFC. An optimum mixing ratio probably exists in the treatment of solid potato mixed with WAS by MFC, a topic with potentially great application significance requiring further clarification in future studies.

For all four MFCs, SCOD showed a trend of increases in the initial period and then decreases until the end of the treatment (data not shown), showing very similar trends in the changes in current density. This indicated that increasing SCOD enhanced the use of organic matter by electrogenic bacteria and increased current density, while decreasing of SCOD led to less organic matter being used for electricity generation. Based on the observed profiles of current density and the profiles of COD, the columbic efficiency of the four MFCs was estimated as 6.1, 0.3, 0.9 and 31.1%, respectively. As shown in Table 2, comparisons with previous studies in the literature indicate that the COD removal by the MFCs in the present study was higher than that of the MFC treating raw activated sludge (Suor et al. 2014), and covered the removal rates of the MFCs treating composite food waste (Hou et al. 2016) and food waste leachate (Moharir & Tembhurkar 2018). The energy generation efficiency of Moharir & Tembhurkar (2018) and the present study were similar.

The fluorescence EEM spectra of DOM in the anodic solution at three different time points of the MFC operation are shown in Figure 3. Throughout the whole treatment process, a total of five peaks, namely Peaks 1–5, were detected. These peaks corresponded to the E_x/E_m wavelengths of 225/340, 280/344, 240/436, 405/468 and 405/468 nm, and were related to the tyrosine-like aromatic amino acids (Peak 1), aromatic protein-like substances (Peak 2), fulvic acid-like substances (Peak 3) and humic acid-like substances (Peak 4 and Peak 5), respectively (Chen et al. 2003).

At the beginning (day 0), Peak 1 and Peak 2 were the two main peaks that emerged, with their overall fluorescence strength being most obvious for MFC-2 fed with the WAS alone, followed by MFC-3 and MFC-4 fed with the solid potato mixed with the two different ratios of the WAS, implying that mixing WAS into solid potato enhanced the presence of the tyrosine-like aromatic amino acids and aromatic protein-like substances from the beginning of the operation. The appearance of Peak 1 and Peak 2 at the beginning of the operation was due to the dissolution of organic matter on the surface of the solid potato and/or WAS. On day 18, when the current density reached or stayed at the highest levels for the two MFCs (MFC-1 and MFC-4) with obvious electricity generation (Figure 2), the EEM images of MFC-1, MFC-3 and MFC-4 treating potato with and without WAS were distinctly different from those of MFC-2 treating the WAS alone; this is mainly reflected by the obvious appearance of Peak 2, Peak 3 and Peak 5. Since Peak 3 and Peak 5 were reported to represent fulvic acid-like and humic acid-like substances, their appearance may suggest that effective humification of the solid potato had occurred inside these three reactors (Geng et al. 2020). The more obvious appearance of Peak 2, reported to reflect the presence of aromatic protein-like substances in MFC-3 and MFC-4 but not in MFC-1, may also suggest hydrolysis of the solid potato promoted by the mixed anaerobically cultured sludge (Du & Li 2017b). This supported the fact that aromatic proteins and tryptophan- and tyrosine-like aromatic amino acids, as the hydrolysates of proteins and other substrates, accounted for a large proportion of DOM in solid potato and WAS (Pang et al. 2014) At the end of the experiment, for MFC-3 (treating the mixture with the higher amount of WAS at a ratio of 1:2), Peak 5 diminished and Peak 2 also showed a distinct decrease. For MFC-4 (treating the mixture with the smaller amount of WAS at a ratio of 1:4), Peak 2 and Peak 5 were getting much stronger, indicating enhanced hydrolysis and humification by the mixed WAS (Du & Li 2017b).

The changes in the fluorescence intensity of all five detected peaks during operation are shown in Figure 4. For the three MFCs (MFC-1, MFC-3 and MFC-4) treating potato with and without WAS, Peak 2 showed a general trend of increases, Peak 3 showed a general tread of gradual increases in the initial 10 days and then slight variations, and Peak 5 showed a general trend of increases and then decreases (for MFC-1 and MFC-3) or only slight variations (for MFC-4). This indicated that a smaller proportion of humic and fulvic acid-like substances (less or not biodegradable hydrophobic acids) were formed from the hydrolysis of solid potato and WAS (Pang et al. 2014). However, for MFC-2 (fed with the WAS alone), apart from Peak 3 showing fewer apparent changes throughout the whole operation, the fluorescence intensity of both Peak 2 and Peak 5 revealed a trend of consistent decreases. The obvious differences in the changing trends of these peaks between the three MFCs (MFC-1, -3 and -4) and the MFC-2 indicate the differences in the composition of the DOM, a product of hydrolysis and decomposition or consumption that occurred simultaneously in the reactors. However, special attention should be paid to the fact that, for all four MFCs, particularly for MFC-1 and MFC-4 where distinct changing profiles in current density were observed, the extent of changes in the fluorescence intensity of the detected EEM peaks was apparently lower than the extent of changes in SCOD (Du & Li 2017b). This suggests that the fluorescence EEM could reflect the composition of a part of the DOM present in the anodic chamber of the MFCs treating solid potato, WAS or their mixtures, with some organic constituents of the products of hydrolysis of the solid biomass probably not being detected, which may include such organic acids as citrate, isobutyrate, acetate, and propionate, that have been found in the anaerobic digestion of WAS and vegetable waste (Chen et al. 2013; Du & Li 2016, 2017a). This needs further clarification in future studies.

The results of PCA based on the data of current density, SCOD and the fluorescence intensity of all five detected EEM peaks of DOM in the anodic solution of all four MFCs during the whole treatment process are shown in Figure 5. For MFC-1 treating the solid potato alone, as the first principle component, the loadings of current density (0.90) and SCOD (0.73) together with Peak 3 (0.76), Peak 4 (0.72) and Peak 5 (0.62) were higher, indicating that the current density was strongly related to the hydrolysis product of SCOD and its further decomposition and humification. As the second principle component, Peak 1, Peak 2 and Peak 3 were three constituting organic fractions that had comparatively higher loadings (0.60, 0.72 and 0.54, respectively). Their distribution suggests that the biodegradable hydrolysis products of tyrosine-like aromatic amino acids and aromatic protein-like substances behaved in a manner similar to the less-biodegradable humic acid-like substances.

In terms of the principal component scores, the plots showed that the score of the first component increased gradually until day 24, then stayed at a same level from day 24 to day 37 and then decreased gradually from day 37 to day 70. The score of the second principal component increased dramatically from day 0 to day 12, decreased slightly from day 12 to day 24, and then showed dramatic decreases from day 24 to day 70. This indicates that, for this reactor, as hydrolysis of the solid potato and the accompanying consumption of the resulting biodegradable constituents by electrogenic bacteria progressed over time, electricity was generated; and the consumption of the biodegradable constituents by ordinary heterotrophic bacteria also occurred, leading to the formation and accumulation of less-biodegradable and non-biodegradable organic substances (Chen et al. 2013; Geng et al. 2020).

For MFC-2, treating the sludge alone, in contrast to the MFC-1, treating the solid potato alone, the plots of the loadings were widely distributed and the scores fluctuated in a less-regular manner. As the first principle component, Peak 1 (0.93) and Peak 4 (0.92) showed higher loadings, together with SCOD (−0.64), Peak 2 (−0.68) and Peak 5 (−0.85) in the opposite direction. As the second principle component, however, SCOD also showed comparatively higher loadings (−0.56) and was found to exist in the opposite direction from Peak 2 (0.62) and Peak 3 (0.66), suggesting the behavior of SCOD differed with its constituting substances, reflected by these two fluorescence peaks. A general trend of decreases in the score of the first principal component as the treatment progressed from day 0 to day 70 was mainly reflective of the behaviors of the products of hydrolysis of aromatic protein-like substances and the degradation products of humic acid-like substances, which showed more obvious changes in their concentrations than the other three fractions of organic substances reflected by Peak 1, Peak 3 and Peak 4. The sharp decrease in the score of the second principal component from day 5 to day 12 probably indicated the increase of SCOD that resulted from hydrolysis of the sludge after a lag period from day 0 to day 5; and the sharp increase of the score from day 24 to day 37 was probably reflective of the ready decrease of SCOD. This may be a result of preferential consumption of the biodegradable constituents by bacteria from the pre-cultured sludge rather than by the electrogenic bacterial species inoculated before the start of the experiment as well as the formation and accumulation of the degradation product of fulvic acid-like substances reflected by Peak 3 (Du & Li 2017b).

When a larger amount of the sludge was mixed into the solid potato, the plots of both loadings and scores of MFC-3 indicated that the treatment from day 0 to day 24 was a process related to the gradual increases of SCOD. This may be as a result of the enhanced hydrolysis of the solid potato by the bacteria brought into the reactor from the larger amount of pre-cultured WAS, and the formation and accumulation of the less-biodegradable fulvic acid-like (Peak 3) and humic acid-like substances (Peak 5). This is due to the likely preferential degradation and consumption of the biodegradable hydrolysis products by the bacterial species from the sludge rather than by the electrogenic species pre-inoculated into the reactor. From day 24 to day 70, there was a trend of distinct decreases in the score of the first principle component. This, together with the observed result of no effective electricity generation under this mixed feeding condition, indicated that this period was mainly associated with the process of SCOD decreases, i.e., the preferential consumption of the biodegradable products of the solid potato by the bacterial species from the sludge rather than by the electrogenic species pre-inoculated into the reactor (Du & Li 2017b).

For MFC-4, where a marked effect of the mixed feeding of the sludge on electricity generation was observed, the PCA results were quite unique and different from those of other three MFCs. As the first principle component, the loadings of all five fluorescence peaks were higher (all above 0.7); and as the second principle component accounted for 32.4% of the variance, the loadings of SCOD and current density were clearly higher (0.977 and 0.84, respectively). The trend of general increases in the score of the first principle component was reflective of the observed increases in the concentration of the organic substances indicated by the five fluorescence peaks, suggesting their occurrence and behavior as a combined result of the promoted hydrolysis of the solid potato by the bacteria from the anaerobically cultured sludge and the degradation/consumption by bacterial species inside the chamber (including electrogenic bacteria).

The distinct increasing trend of the second principle component score from day 0 to day 12 reflected the increases of SCOD as a result of promoted hydrolysis and the enhanced electricity generation due to the likely preferential use of the biodegradable constituents by electrogenic bacterial species. The distinct decreasing trend of the score from day 24 to day 54 mainly indicated the decreases of SCOD due to the continued preferential consumption of its biodegradable constituents by electrogenic bacteria, which led to the current density being maintained at around the highest level until day 54, when SCOD decreased nearly to the lowest level observed at the end of the experiment.

The flow-cytometric density plots of total and viable bacteria in the anodic solutions of the MFCs at the end of the operation are shown in Figure 6. For viable bacteria, except the MFC-2 fed with the WAS alone, for which the red fluorescence were widely plots scattered, MFC-1, MFC-3 and MFC-4 produced similar cytograms, each with two obvious peaks. For total bacteria, MFC-1 and MFC-4, fed with the solid potato alone and the mixture of the solid potato with the smaller amount of sludge (mixing ratio of 1:4, showed distinct peaks and the plots were comparatively more gathered and bundled; however, for MFC-2 and MFC-3, fed respectively with the sludge alone and the mixture of the solid potato with the larger amount of sludge (mixing ratio of 1:2), the plots were more widely distributed, probably reflecting the likely differences in size, i.e., the volume of cells of the activated sludge. Meanwhile, the changes in the morphology and geometry due to the lysis and disruption of bacteria when biodegradable organic constituents in the reactors were consumed were also a likely reason that contributed partly to the observed wide distribution, as could be inferred from the results of previous studies based on scanning electron microscope observations (Du & Li 2017a) and flow cytometric analysis (Foladori et al. 2010a, 2010b).

The differences in the concentrations of total and viable bacteria in the final anodic solutions of all four MFCs are shown in Figure 7. Compared to the concentration of viable bacteria in MFC-2, that in MFC-1 was obviously higher, indicating more viable bacteria remained in the reactor fed with the solid potato alone than in the reactor fed with the WAS alone. For the two MFCs (MFC-3 and MFC-4) fed with the mixture of the solid potato and WAS, the concentration of viable bacteria was obviously higher in MFC-4, fed with the smaller amount of the WAS (mixing ratio of 1:4), indicating that heterotrophic bacteria in WAS enhanced hydrolysis of solid potato and the hydrolyzed product was sufficient for the survival of both heterotrophic bacteria brought in with the WAS and electrogenic bacteria. The lower viable bacterial concentration, together with the higher concentration of total bacteria observed for MFC-3, indicated that more bacteria remaining at the end of the operation of this reactor were dead, probably due to the faster consumption and depletion of biodegradable organic substances from the hydrolysis of the solid potato and also from the disintegrated cells of the sludge as a combined result of endogenous respiration and bacterial competition for available substrate. This explanation that can be inferred, taking also into consideration the observed result shown in our previous study (Du & Li 2017b), is that, for this reactor, the detected SCOD values were obviously lower and the time point from which distinct concentration changes of SCOD terminated was earlier (day 37).

Compared to MFC-3, the concentration of total bacteria in the reactor of MFC-2 fed with the sludge alone was significantly lower. This may indicate that, under the treatment condition without the presence of other available substrate sources (the solid potato in this study), the disintegration of the cells due to their lysis and disruption occurred more intensively, leading to a significant reduction in the number of intact cells from the sludge. Since no distinct electricity generation was observed in this reactor, as was the case in the reactor of MFC-3, where a larger amount of sludge was mixed, as shown in our previous study (Du & Li 2017b), it is reasonable to infer that biodegradable substances in the detected SCOD resulting from the disintegration of certain cells in the sludge (for MFC-2) and the hydrolysis of the solid potato and also the disintegration of certain cells from the sludge (for MFC-3) were preferentially consumed by the large number of bacteria from the sludge, rather than by the electrogenic bacteria inoculated into and cultured inside the reactor before the experiment. The distinct enhancement in electricity generation observed for MFC-4, where the sludge was mixed to the solid potato in a smaller ratio of 1:4, suggests that an optimum mixing ratio may exist when WAS is used for the purpose of accelerating the hydrolysis of solid vegetation waste and combined treatment by MFCs. This needs further comparative studies under more different mixing conditions, with detailed evaluation of the changes in the composition and structure of the bacterial community involved in the whole MFC treatment process.

SCOD displayed solubilization/hydrolysis and degradation of solid potato and WAS, showing very similar trends in terms of the current density. This indicated that hydrolysis of solid potato and WAS contributed to more DOM being used by electrogenic bacteria for current density generation, and their degradation led to less DOM available for current density generation. Fluorescence EEM showed that large amounts of tyrosine-like aromatic amino acids and aromatic protein-like substances were hydrolyzed and degraded/utilized partially by electrogenic bacteria for current density generation. And the lower mixing ratio of WAS supported various viable heterotrophic bacteria that could significantly enhance hydrolysis of solid potato and supply sufficient hydrolyzed products for electrogenic bacteria to generate current density.

Mixing WAS into solid potato enhanced the presence of the tyrosine-like aromatic amino acids and aromatic protein-like substances, and promoted hydrolysis and humification of the solid potato. At the end of the operation, more viable bacteria were detected for the two MFCs fed with the solid potato alone and its mixture with the smaller amount of WAS, where distinct electricity generation was observed, than for the MFCs fed with the WAS alone and the mixture of the solid potato with the larger amount of the WAS. The results of this study will be of great value for the treatment of solid potato waste by MFCs.

This research is supported by the Natural Science Youth Foundation of Jiangsu Province (Project Number: BK20171017).

All relevant data are included in this paper.