The residual water entering WWTP undergoes different treatment procedures, starting with grit removal, where the large-size waste is withdrawn, followed by the removal of fat oil and grease (FOG) before primary sedimentation. The sludge from the settling tank undergoes a process of thickening and, after that, results in the PS. The treated water from the primary treatment stage withstands a biological process (resulting in waste-activated sludge), and the excess follows a thickening process by air flotation, resulting in the SS, which were the two types of sludge used to form the mixture (MSS) applied on the AD process throughout the different experimental assays. The PS and SS used in the trials were collected in a WWTP at Lisbon, Portugal, with a treating capacity of 50,000 m³/day, serving approximately 210,000 inhabitants’ equivalent and stored at 4 °C until further use. The mango peel waste was collected during the experimental trials according to the requirements for the assays in progress through the surrounding academic population consumption and the researchers involved in the study.

The PS:SS [v:v] ratio used to prepare the MSS in the laboratory was 60:40 to replicate what occurs in WWTPs. To homogenise the substrate, the MSS was agitated for 2 minutes (Velp Scientifica, 35 W and 230 V, with 1300 rpm).

Regarding the MPW, to use it as a co-substrate for AcoD, it was necessary to obtain a liquid fraction (MPL) which was obtained by blending the MPW with water in a proportion of 400 g of water to 100 g of MPW using a cooking blender (Selecline, 230 V, 400 W), on highest speed for 2 minutes. After blending, the pulp was sieved (mesh size of 2 mm), and the remaining liquid fractions were stored at 4 °C until further use. Figure 1 illustrates a schematic representation of the pre-treatment processes applied to MSS and MPW.

Figure 1.

Schematic representation of the pre-treatment processes applied to the MSS and MPW

It was assumed for the feeding combination a ratio of 90% MSS and 10% MPL [v:v], along with a mechanical pre-treatment, as shown in Figure 2, after analysing the physicochemical characteristics of MSS and MPL.

Figure 2.

Schematic representation of the co-digestion trials

A CSTR with an 11.3 L working volume, running under mesophilic conditions (36±1 °C), was used to perform the trials. A feeding pump (Watson Marlow, 60 rpm) is connected to the CSTR to feed the reactor, and an electric stirrer (Velp Scientifica ES) is inside the reactor to promote its agitation. The feeding mixtures and digestates for each HRT were characterieed (reactor input and output) to define the operational parameter to monitor AD/AcoD process.

The monitoring plan established to follow the continuous AD and AcoD process included daily procedures, like the control of temperature inside the reactor made by a thermal blanket, biogas production (registered by a flow meter from MilliGascounter, Ritter), methane production (measured with a portable analyser − LMSxi Multifunction Landfill Gas Analyser) and pH assessment of the feeding mixtures and digestates (measured using Multi 3430, WTW).

The total solids (TS), volatile solids (VS), VS/TS ratio, volatile suspended solids (VSS), total chemical oxygen demand (TCOD), soluble chemical oxygen demand (SCOD), SCOD/TCOD ratio, total Kjeldahl nitrogen (TKN) and total alkalinity (TA) were periodically analysed according to American Public Health Association [31]. Total organic carbon (C_org) was determined by dividing the organic matter content (expressed in VS) by 1.724, according to [32], and the carbon/nitrogen ratio (C/N) was calculated by taking the ratio between C _organd TKN.

Different parameters, namely the OLR, VS and TCOD removal efficiency, specific gas production (SGP), specific methane production (SMP) and specific energy loading rate (SELR), were evaluated to assess the process performance and stability of the HRTs.

The following procedures consisted in selecting the HRT with the best performance and process stability and establishing performance indicators to act as tools to simplify the discussion of the results achieved. Two indicators were defined based on AD and AcoD performance to help identify the increments and the advantages of selecting a co-substrate to improve AD technology.

All analyses were performed in two or three replicates, and the results are shown together with the means and standard deviations for each.

The GraphPad Prism Software (version 5.0) was used for the statistical analysis. The Tukey test was used to compare three or more samples, and the Student’s t-test was used to compare two samples, with a 95% level of confidence (p = 0.05). Differences were considered significant when p-values were less than 0.05.

The experimental trials’ design in the study followed the physicochemical characteristics of MPL referred to above, bearing in mind the spare capacity of digesters already implemented in real-scale WWTP [35].

Observing the above constraints established the definition of a correct feeding mixture ratio for lab-scale co-digestion trials to an up-scale design without infrastructure changes. Additionally, co-substrate selection needs a careful approach since carbohydrate-rich co-substrates can be easily biodegraded, inducing instability [35].

Regarding the previous statements, the authors proposed some assumptions: a ratio of 90% MSS and 10% MPL [v:v] for the feeding mixture, coupled with a mechanical pre-treatment, as illustrated in Figure 2; also an HRT of 15 days was selected, based on previous work with a similar composition of MSS and different type of co-substrates [36], [38]. It is important to cover a range of OLR values (low, medium and high) with two more HRTs (13 and 20 days) to evaluate the performance and stability of co-substrate incorporation. For each HRT, three cycles were performed after steady-state conditions were achieved.

Regarding the values presented in Table 2, HRT of 13 days had the highest pH value (5.56, compared with 5.33 and 5.27 from HRT 15 and 20 days, respectively). However, there were no representative fluctuations. A statistical analysis of the pH values of the feed and digestate was done to ascertain the impact of HRT variation. A box-and-whisker plot was created (Figure 3) to depict this analysis, with the whiskers extending to the minimum and maximum and a five-number summary: the minimum, the 25th percentile, the median, the 75th percentile, and the maximum. Although the pH of feeding mixtures was below the minimum recommended values [29], [34] (Figure 3a), the digestates showed values of pH (between 6.97±0.13 and 7.05±0.09) which are in line with the buffer capacity of the AD process, as illustrated in Figure 3b.

Figure 3.

Feeding mixtures (a) and digestates (b) pH pattern along trials; different letters in the graphic indicate significantly different results (p < 0.05) according to the Tukey test

Organic matter expressed in VS was higher for 13 days (25.14±2.30 g/L), showing a decrease of about 12% for the other HRTs. This behaviour is reflected in VS/TS ratio, where 13 days achieved the highest percentage of 88.

The highest SCOD/TCOD ratio (18%) was observed for HRT = 13 days, compared with 16% and 14% from HRT 15 and 20 days, respectively, indicating the presence of more easily available organic matter that can be biodegraded more rapidly. Comparing the C/N ratio of feeding mixtures, an improvement with the incorporation of MPL in MSS is visible, with an increment of 13% for HRT = 13 days, proven that using co-substrates such as MPL can, in fact, improve this parameter, as shown by other researchers [28], [38].

To support the decision of HRT selection for AcoD assays, Table 3 exhibits the results of the performance and stability parameters.

When decreasing HRT, OLR increases, resulting in a positive synergetic effect in the SMP parameter. Highlighting the OLR value of 1.94±0.02 g VS/L_reactorday for HRT = 13 days, the highest SMP value (425±17 mL CH₄/g VS) was achieved. It represents an increment of 16% compared with results obtained by Di Maria et al. [38] for similar operational parameters, with an OLR of 2.1 kg VS/m³day and an HRT between 10−14 days.

Over the different HRTs, there were no major changes in the VS content, with the highest removal rate (50%) registered at HRT = 20 days, reducing with the decrease in HRT, which is in agreement with the experiment described by Di Maria et al. [38], where it is suggested that VS removal percentages similarly reduced as HRT decreased and OLR increased. However, compared with the literature, even the lowest VS removal rate achieved throughout the trials had better results. For an OLR of 1.46 kg VS/m³day, with an HRT of 14 days, a VS removal of 22% was reported, less than half of what was attained across all of our HRTs [38].

Methane quality remained constant among trials (62−63%), with better results than the ones obtained by Di Maria et al. [38], where for OLR ranging from 1.46 to 2.8 kg VS/m³day the methane quality has ranged between 57−62%.

To evaluate the stability of the process, the following parameters were considered: SELR, TA, and pH of the digestate. SELR relates organic load (expressed in TCOD) fed into the reactor with the biomass inside the active volume of the reactor (expressed in suspended volatile solids mass). The anaerobic microbial consortium has a limited capacity to biodegrade organic matter, and when this limit is surpassed, microbial activity inside the reactor may become unstable, affecting the methanogenesis pathway. Therefore, the SELR values recommended must be below 0.4 1/day to ensure the stability of the AD process. Along the HRTs, the SELR results obtained (0.18 1/day, 0.22 1/day and 0.24 1/day) agree with those reported in the literature by other authors [39], [40]. SELR did not demonstrate significant differences (p > 0.05) between the several trials, proving that the process stability was unaffected by HRT variations.

Regarding the recommended values for the TA parameter, some authors [24] suggested that TA should be above 1500 mg CaCO₃/L to maintain a stable process but did not define the upper limits. However, other authors warned of instability in the reactor at higher alkalinity values, with a maximum recommended value of 4000 mg CaCO₃/L [41]. Therefore, for analysing the results obtained along the assays, the range: 1500 mg CaCO₃/L < TA < 4000 mg CaCO₃/L was defined to guarantee the stability of the AcoD process. The values obtained during the AcoD trial (2761, 2320, and 2973 mg CaCO₃/L) fulfil the recommended values previously considered.

Although the pH feeding values are similar between the trials, the HRT of 13 days significantly differs from the others (p < 0.05). Despite the variability of feeding mixtures along the trials, Figure 3b shows that there were no major variations in the pH values of digestate throughout all trials: the pH remained close to neutrality (6.97–7.05), which is an indirect indicator of stability and will sustain the buffering capacity of the process. In contrast to the feeding pH, there were no statistically significant variations between HRT 13 days and HRT 15 days (p > 0.05).

Based on the previously discussed, it is possible to conclude that all HRT trials were within limits established not to compromise the stability and buffer capacity of the AcoD process.

The SMP parameter allows us to assess the best HRT regarding AcoD performance and stability due to a good correlation with daily methane production and daily VS organic load in the feeding mixture. Analysing the values shown in Table 3, the HRT of 13 days was the trial with the best performance in terms of SMP, with values 36% and 8% higher than the HRT of 15 days and 20 days, respectively. It is also important to highlight the correlation with the digestate pH, TA, and SELR values, which remained within the recommended ranges to evaluate the AcoD process stability.

After selecting HRT = 13 days as the trial with the best performance and stability parameters, it was decided to compare it with a mono-digestion scenario with the same HRT in similar operational conditions.

To allow the comparison between AD and AcoD, the performance and stability parameters regarding the AD trial (which refers to authors’ unpublished data) are the following: SMP of 220 mL CH₄/g VS, with an OLR of 1.94 g VS/L_reactor.day, SCOD/TCOD of 11%, a C/N ratio of 8 and a SELR of 0.17 1/day. SCOD/TCOD ratio provides information on the bioavailability in the feeding load, indirectly justifying the increase in the SMP values recorded since there was a 64% increment in the bioavailability of the feeding mixture in the AcoD trial (Figure 4). It is shown that there are no significant differences between the SMP values obtained from the AD trial and the AcoD trial (p > 0.05), allowing the choice of the one that exhibits better results in terms of a specific production. This observation supports the idea that co-digestion can be used to produce bioenergy.

Comparing the C/N ratio between AD and AcoD trials, it is possible to register an increment of 38% from AD to AcoD, which agrees with the results obtained by other authors [28], [38], attesting to the benefits of incorporating the co-substrate.

Figure 4.

Comparison between AD and AcoD with HRT of 13 days; different letters in the diagram indicate significantly different results between AD₀ and AD₁ (p < 0.05) according to Student’s t-test

Using the SMI indicator, eq. (1), to allow the comparison of SMP between AD and AcoD, it is possible to assess a value of 93%, which means that AcoD practically doubled the results obtained in the AD process. The obtained results align with those published by other researchers: [42] reported an increment of 48% in SMI values when a co-substrate (FVW and food waste) was added, compared with mono-digestion. On the other hand, Karki et al. [34] reported a 27% increase in co-digestion trials when compared with AD for an OLR of 2.5 kg VS/m³day. These results support the thesis that implementing co-digestion in WWTPs can indeed increase biogas production and the performance of the AD process.

To corroborate the results obtained with SMI, an energy analysis was performed regarding EPRI. Eq. (3) described above, using the TS concentration of 28.04 kg/m³ (Table 1) and η_AD = 0.26 (which refers to an average achieved by personal data), yielded E ₀= 30.91 kWh/m³_{feed sludge} for AD trial. For AcoD trials, considering HRT = 13 days, using the TS concentration of 28.46 kg/m³ (Table 2) and η_AD = 0.40 (Table 3), the results obtained from eq. (4) were E ₁= 55.21 kWh/m³_{feed sludge}.

Thus, the incorporation of 10% MPL in MSS for an OLR of 1.94 kg VS/m³_reactor·day led to a 79% increase in the EPRI. Based on these results, it is possible to set trial HRT = 13 days at a real scale, coupling simple management modifications at WWTP facilities.

Nevertheless, the proposed indicator seems an important (even conservative) predicting tool to implement in real-scale new management scenarios. However, the feed regime should be carefully planned with a stepwise increase to the desired feed ratio to acclimatise the bacteria and prevent overloading.