Novel routes for urban bio-waste management: a combined acidic fermentation and anaerobic digestion process for platform chemicals and biogas production

A combined acidic fermentation and anaerobic digestion (AD) treatment has been developed on pilot scale for urban bio-waste conversion into volatile fatty acid (VFA) and biogas. The specific waste mixture was composed by the pre-treated organic fraction of municipal solid waste (OFMSW) and waste activated sludge (WAS), both produced inside the Treviso (northeast Italy) municipality. The effect of temperature (37°C and 55°C) was investigated in both steps. Only the mesophilic fermentation process provided a VFA-rich stream (19.5 g CODVFA/L) with stable physical-chemical features, with no need of chemicals addition for pH control. The sludge buffering capacity made this step technically feasible. The AD step was performed on the solidrich fraction of fermented bio-waste, after dilution with excess WAS. No relevant differences were observed under the two investigated temperature: in the steady state (organic loading rate of 2.5 kg VS/m3 d), the specific biogas production was 0.40 and 0.45 m3/kg VS at 37°C and 55°C respectively, with similar CH4 content (63-64% v/v). The scaled-up version of the system (in an average urban municipality of 170,000 Person Equivalent) revealed that the whole process is thermally sustainable if both reactors are operated at mesophilic temperature: 36% of surplus thermal energy and 13,03 MWh/d of produced electricity, which corresponds to a revenue of 609,605 €/year. In addition, 2,262 kg CODVFA/d are available for parallel purposes, such as the synthesis of bio-products with higher added value than bio-methane (e.g. biopolymers). 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 1 Novel routes for urban bio-waste management: a combined acidic fermentation and anaerobic digestion process for platform chemicals and biogas production Francesco Valentino, Giulia Moretto, Marco Gottardo, Paolo Pavan, David Bolzonella, Mauro


Introduction
Anaerobic digestion (AD) is currently considered the most feasible technology for biogas production from the treatment of high solid content streams, such as manure, organic fraction of municipal solid waste (OFMSW), energy crops, agro-industrial wastewater etc.. (Tyagi et al., 2018). The oldest and more commonly adopted application of AD is the treatment of biological sludge (Mata-Alvarez et al., 2014), even if minimized with membrane bioreactor (MBR) technology (Sepehri and Sarrafzadeh, 2018). In the two last decades, a particular attention has been devoted to the co-digestion processes, which have been designed for a simultaneous treatment of biological sludge and OFMSW (Mata-Alvarez et al., 2014). With regard to the specific case of biowaste produced in urban scenario, the increasing of global population has led to new concerns for public authorities dealing with the treatment and disposal of biological sludge produced in municipal wastewater treatment plants (WWTPs) and food waste (Ma et al., 2017).
Through the conventional AD processes, the degradation of organic matter contained in food and sludge waste streams is not even complete, being in the wide range of 35-60% in most cases (Yin et al., 2016a). This performance is strongly affected by the low efficiency of hydrolysis, which is the rate-limiting step for the urban bio-waste anaerobic degradation process (Ma et al., 2018). Many pre-treatment methods have been explored for enhancing bio-solid hydrolysis and improving chemical oxygen demand (COD) solubilisation in order to increase bio-methane yield; among others, thermal hydrolysis (Li et al., 2017) and aerobic micro-aeration (Montalvo et al., 2018) pretreatment.
Archea have been recognised as main responsible for methane production from volatile fatty acids (VFAs), as well as for ethanol and lactate (Stoyanova et al., 2017) in a typical AD process.
However, since different microorganisms with specific biochemical pathways are involved, many other valuable products can be obtained via acidogenic fermentation. These products can have a wide spectrum of possible valorisation. As an example, the VFAs and lactic acid were described as 1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60  61  62  63  64  65 5 polyhydroxyalkanoates (PHAs) (Kourmentza et al., 2017). Alcohols have been also sometimes reported as possible substrate for PHA synthesis (Beccari et al., 2009). This route allows converting platform chemicals in bio-products at higher added value than methane (Pagliano et al., 2017). In fact, PHAs are one of the most promising classes of biopolymers, which can replace part of the synthetic marketed materials, due to their biodegradability and thermoplastic properties. In the frame of mixed microbial culture (MMC) technology (Valentino et al., 2017), the use of complex and low cost substrates is a strategy to accomplish both bio-products production and bio-waste treatment following their acidic fermentation. For this reason, acidogenic fermentation can be considered an option to exploit the organic matter contained in sludge and food waste as renewable feedstock, decreasing the amount of bio-waste to be disposed off (Koller, 2017). The VFAs composition and fermentation yield can vary according to the process parameters and settings. A desirable feature for PHA production is the ratio between odd and even carbons in the VFAs, which determine the composition of PHA co-polymers to a large extent (Bengtsson et al., 2010). Several parameters affected the fermentation process outputs such as inoculum (Yin et al., 2016b), pH (Wu et al., 2017), temperature (Dahiya et al., 2015), hydrogen partial pressure (Agler et al., 2011). A previous study showed that a pH between 5.0 and 6.0 was preferable to reach the highest fermentation process yield (Gottardo et al., 2017). However, the accumulation of VFAs produced in the fermentative metabolism generally causes a pH decreasing below the optimal values. Apart form chemicals additions for pH control, some authors recently proposed the recirculation of the anaerobic digestion effluent, as buffering agent, to control the pH in the fermentative stage. This strategy was applied in the combined two-stage fermentation and anaerobic digestion process (Gottardo et al., 2017). The method was particularly efficient when applied to a high solid content waste such as the pre-treated OFMSW. The high content of putrescible solids facilitate their fermentation even at relatively low hydraulic retention time (HRT < 4.0 days); as a consequence, the high level of VFAs (easily achieved) brings the pH below 5.0, inhibiting a further fermentation activity and making the pH-control strategy strictly necessary.

6
In the particular case illustrated in this paper, both VFAs and biogas production has been assessed from the anaerobic treatment of the specific cocktail of urban bio-wastes produced in the Treviso (north-east Italy) municipality. The pre-treated OFMSW was mixed with thickened waste activated sludge (WAS) from Treviso WWTP. The produced VFAs from acidogenic fermentation process were separated from the slurry for other potential uses (such as the PHAs synthesis). The solid-rich waste overflow was originated from a solid/liquid separation unit. This fraction was opportunely diluted with the same thickened WAS (collected in Treviso WWTP), and it has been used as secondary waste feedstock for biogas production through AD process. This approach differs from the traditional WAS-OFMSW co-digestion, developed at the Treviso full scale WWTP (Bolzonella et al., 2006), which allows recovering biogas only. In a frame of urban bio-refinery concept, the renewable carbon source may be driven in different routes that include also VFA production followed by biopolymer synthesis in a high load aerobic process (Valentino et al., 2015). The present pilot scale investigation has the scope to verify the process technical feasibility aimed at maximizing the production of VFA and at recovering biogas from the overflows.

Rationale of the approach
The designed flowchart consisted of: a) hydrolysis and fermentation step for VFA production from the most putrescible matter of the feedstock; b) solid/liquid separation unit for the solids removal from the liquid stream at high VFA content (suitable for further valorisation); c) anaerobic digestion step for biogas production from the solid-rich stream overflow ("solid cake"). Both steps were performed under mesophilic and thermophilic conditions. The whole process was evaluated in terms of yields, namely maximum VFA content for the fermentation process and biogas production for the anaerobic co-digestion processes. A mass and energy balance was made in order to assess the technical feasibility and to understand whether the process is energy efficient and sustainable.

Thickened Waste Activated Sludge (WAS) and source-sorted Organic Fraction of Municipal Solid Waste (OFMSW)
The feedstock was weekly collected inside the Treviso municipal WWTP facility. Its composition was representative of influent stream usually treated in AD Treviso full-scale plant: thickened WAS and OFMSW were at volumetric fraction of 65-70% and 30-35% respectively. The WAS came from the static thickener of the WWTP sludge line. Source sorted OFMSW was collected in 50 districts of the Treviso Province, and transferred to the full-scale WWTP after its pre-treatment (squeezing and homogenization) with a screw-press. The WAS and OFMSW volumetric fractions adopted in this study were in the range reported in a recent review (Tyagi et al., 2018), where a comprehensive list of examples of waste sludge-OFMSW mixture co-digestion is provided; in general, the volumetric OFMSW fraction is always lower (10-40% v/v) than WAS fraction (60-90% v/v) (Tyagi et al., 2018).

Anaerobic pilot units
The acidogenic fermentation unit consisted of a continuous stirred tank reactor (CSTR) of 0.38 m 3 of operative volume, mechanically stirred and under temperature control by using a thermostatic jacket. The hydrolysis and fermentation of the organic content of WAS-OFMSW mixture into VFAs were carried out under both mesophilic (37°C) and thermophilic (55°C) conditions, in two separate trials. All the other operating conditions are listed as follows: hydraulic retention time (HRT) of 6 days; organic loading rate (OLR) 6.5-7.0 kg VS/m 3 d approximately; uncontrolled pH in the range 5.0-5.5. The CSTR fermenter was connected to a solid/liquid separation unit consisting of a coaxial centrifuge, equipped with 5.0 μm porosity nylon filter bag for solid (TS) removal. The solid-rich stream overflow was separated and then diluted with the excess WAS (collected from the full-scale plant) as feeding for anaerobic digester.
Stainless steel CSTR (AISI-304) of 0.23 m 3 operating volume was used as anaerobic digester.
Mechanical anchor stirrer ensured mixing and homogenization. Two separated trials were carried-8 out at 37°C and 55°C, according to the temperature of the previous fermentation stage. The temperature was maintained by hot water recirculation through external jacket. Anaerobic digested sludge inoculum was collected in the same WWTP, from the 2300 m 3 CSTR anaerobic digester, treating thickened WAS and squeezed OFMSW mixture under mesophilic condition at an OLR of 1.8 -2.0 kg VS/m 3 d. For both AD trials, the digested sludge was acclimatized for 35-40 days (more than 2 HRTs). Operational parameters in both mesophilic and thermophilic trials were similar: HRT of 15 d and OLR between 2.2 -2.5 kg VS/m 3 d. The WAS/cake ratio in the feedstock was calculated on the basis of the WAS and OFMSW specific production: 0.3 kg OFMSW/PE day (based on wet weight) and 0.06 kg TS/PE day for WAS (Metcalf & Eddy, 2003).

Sampling and analytical methods
Feedstock characteristics were determined once per week on fresh sample, immediately after its collection. The reactor effluents were monitored three times per week for total solids (TS), volatile solids (VS), chemical oxygen demand (COD), total Kjeldahl nitrogen (TKN) and phosphorous (P).
For the TS determination, a drying temperature of 105°C was adopted (Peces et al., 2014). The process stability parameters, i.e., the pH, volatile fatty acid content and distribution, total and partial alkalinity, ammonia, were measured two/three times per week. All analyses were performed according to the standard methods (APHA/ AWWA/WEF, 1998). The analyses of VFA were conducted using an Agilent 6890N gas chromatograph equipped with a flame ionization detector (T 100°C), a fused silica capillary DB-FFAP (15 m length, 0.53 mm x 0.5 mm film thickness) column, with hydrogen as carrier. The analysis was conducted by increasing the temperature from 80°C to 200°C (10°C/min). Samples were centrifuged at 4000 rpm for five minutes; the supernatant was filtered by using a 0.20 µm cellulose acetate filter before the analysis. The biogas production was monitored using a flow metre (Ritter CompanyTM); methane, carbon dioxide and oxygen in the biogas were continuously determined using a portable infrared gas analyser GA2000TM (Geotechnical InstrumentsTM), and once a day using a gas chromatograph 6890N (Agilent   1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60  61  62  63  64  65 9 TechnologyTM). The analyses were conducted with an HP-PLOT MOLESIEVE column (30 m length, 0.53 mm ID x 25 mm film thickness), using a thermal conductivity detector and argon as carrier (79 ml/min). H 2 , CH 4 , O 2 and N 2 were analysed using a thermal conductivity detector (TCD) at a temperature of 250°C. The inlet temperature was 120°C, with a constant pressure in the injection port (70 kPa). Samples were taken using a gas-type syringe (200 mL biogas). Once the entire sample was vaporized, separation of the peaks occurred within the column at a constant temperature of 40°C (8 min).

Calculation
All the parameters characterising reactor performances were calculated after steady states were achieved. Descriptive statistics and exploratory data analysis were performed using the open source program, R (The R Foundation for Statistical Computing, version 3.5.0).
For the fermentation stage, the process was considered sufficiently stable and under steady state after consistent VFA production was detected. Solubilisation and fermentation yield (Y VFA ) were determined with respect to the characteristics of initial unfermented feedstock, in particular total COD (COD TOT(0) ) TS (0) and VS (0) . For the anaerobic digestion step, steady state was considered achieved when the biogas production rate (GPR) was constant (below 10% of deviation from average values).
The mass and energy balance, as well as the energy yield, were performed by considering both stages in two combination of temperature: a) fermentation and digester reactors under mesophilic condition; b) fermentation reactor under mesophilic condition and digester reactor under thermophilic condition. The aim of this comparison was to evaluate which configuration is more profitable in terms of surplus thermal energy and electricity production.

Bio-waste composition and characteristics
The chemical features of the feedstock used for fermentation reactor were relatively constant during the whole operation period (almost 4 months). The average dry matter was 47 ± 2 g TS/kg, with 83% of VS. The nutrient level was more similar to that of biological sludge (WAS) instead of the typical OFMSW content (usually N-or P-limiting with respect to COD content), since the feedstock largely composed by thickened WAS (65-70% v/v). With regard to the soluble COD, nutrient level was 100/5.7/1.6 as COD/N/P. The monitored features of the feedstock mixture are listed in Table 2 (a more detailed version is given in Supplementary Material, Table S1). The thickened WAS had the following characteristics: 30 ± 2 g TS/kg, 21.5 ± 0.7 g VS/kg (72% of dry matter), 40 ± 1 g N/kg TS, 15.0 ± 0.2 g P/kg TS, 0.15 ± 0.08 g COD SOL /L, 48 ± 6 g COD/kg TS.

Acidogenic Fermentation
The WAS-OFMSW fermentation was performed without pH control using the sludge buffering capacity (Cabbai et al., 2016), and thereby avoiding process costs associated with the addition of chemicals. The process conducted at 37°C produced a fermented stream with stable chemical characteristics starting from the 20 th days (about 3.5 HRT) of operation ( Figure 1A). After this initial period, VFA level was constantly maintained close to 20 g COD VFA /L; correspondingly, pH trend did not show relevant fluctuations, being minimum values close to 5.0 and maximum values around 5.7 ( Figure 1B). Indeed, the WAS-OFMSW mesophilic fermentation did not require any strategy for pH control since the feedstock buffering capacity (402 ± 24 mg CaCO 3 /L, pH 5.7) counterbalanced the unavoidable pH decrease due to the stable VFA production. This was a relevant aspect, which demonstrated the robustness and simplicity of process operation at the same time. The combination of two parameters such as the relative WAS percentage in the mixture and the mesophilic temperature made the fermentation process technically feasible.
Fermentation performances as well as physical-chemical characteristics were affected by this process instability as indicated by the standard deviations of average values (sometimes one order of magnitude higher than those under mesophilic temperature) reported in Table 2.
Higher temperatures solubilised more solids and increased the COD SOL and nutrient release in the mixed liquor. The VFA concentration also increased above 20 g COD VFA /L; however, a not negligible fraction of COD SOL remained unconverted into VFA, being the average COD VFA /COD SOL ratio equal to 0.59. Feedstock fermentability at 37°C was found to be less performing for solids and COD solubilisation (0.41 ± 0.02 and 0.26 ± 0.01 on initial VS and COD TOT respectively) even though VFA yields were comparable to those measured at 55 °C.
The acidic WAS-OFMSW fermentation was found to be more performing to that WAS only, independently from the chosen process temperature. The VFA yields obtained in this work (0.41-0.44 g COD VFA /g VS (0) , 0.26-0.28 g COD VFA /g COD TOT(0) ; Table 1 The process stability is also an important aspect affecting the technical process feasibility: mesophilic fermentation appeared more technically attractive since the VFA-rich stream features were more stable over time and the COD VFA /COD SOL ratio was higher (t -test, alpha value = 0.01) than value obtained in thermophilic fermentation (Figure 2).

Figure 2.
A VFA-rich stream with consistent composition is also a desirable feature for PHA production since the balance of odd and even carbons in the VFA determines the polymer composition  (Table S1; Supplementary Material), most likely dependent on feedstock characteristics and not affected by fermentation temperature.
Regarding the solid cake (SC), the Table 3 summarizes the main characteristics of both SC overflows, obtained from mesophilic and thermophilic fermentation process. In spite of the observed differences for the features of mesophilic and thermophilic VFA-rich streams, the working  5.0 ± 0.6 6.0 ± 0.7 * Solid-rich cake produced after fermentation process

Anaerobic Digestion
Mesophilic and termophilic processes were started at relatively low OLR (0.7-0.8 kg VS/m 3 d) by using a feed solution mainly composed by thickened WAS. The SC content was gradually increased in order to achieve an OLR of 2.3 -2.5 kg VS/m 3 d approximately. The SC content strongly affected the operative OLR since it was characterized by TS of one order of magnitude higher than the thickened WAS.
In mesophilic condition, steady state was achieved in 40 days (roughly 2 HRTs). In the start up period, no relevant pH drop was observed even at the highest VFA concentration detected (close to 1.0 g COD/L at day 10 th after inoculum). Most of the readily fermentable COD was already converted into VFA in the previous hydrolysis/fermentation step; as a consequence, the process was easily controlled even in the transient conditions when the risk of methanogenic bacteria inhibition is usually high. Average pH value of steady state period was 7.6 ± 0.2, advantageous for the growth of methanogens (Gottardo et al., 2017). The pH remained constant because of the high buffering capacity of the system as indicated by the partial alkalinity (1,738 ± 138 mg CaCO 3 /L) and by the VFA/Partial Alkalinity (P.Alk) ratio (0.13 ± 0.04 mg acetic acid/mg CaCO 3 ) ( Figure 3A). In particular VFA/P.Alk ratio showed a maximum peak (1.14) in the first week after inoculum and then it decreased over the course of operation. In the steady state period, VFA/P.Alk ratio was constantly below 0.20 indicating high process stability and the negligible VFA accumulation. In 3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60  61  62  63  64  65 16 fact, it is widely recognized that VFA/P.AlK value greater than 0.30 may indicate a possible process upset or instability usually caused by organic matter overload and a consequently VFA concentration increase (Bolzonella et al., 2003). Ammonia release from proteins degradation led to an increase of total ammonium in the liquid phase from 0.46 ± 0.07 to 0.79 ± 0.06 g N-NH 4 + /L.
However, this value was abundantly below 3.0-3.5 g N-NH 4 + /L, which is the typical critical value for the inhibition of metanogenic bacteria (Chen et al., 2008).
The AD step under thermophilic conditions reached the steady state in 30 days after inoculum, approximately 0.5 HRT less than mesophilic process. Also in this case, the process was easily controlled since no relevant VFA accumulation was detected (maximum value 1.01 g COD/L, day 13) and pH was easily maintained above 7.5. Starting from day 22, VFA/P.Alk ratio was permanently below 0.20 mg acetic acid/mg CaCO 3 . Figure 3B shows the trends of total-partial alkalinity and VFA/P.Alk ratio along the experimentation: when the steady state was achieved, partial and total alkalinity had average concentrations of 2,524 ± 279 and 3,788 ± 314 mg CaCO 3 /L; VFA/P.Alk ratio was equal to 0.12 ± 0.03. Figure 3C shows the biogas production rate (GPR) and methane percentage in the produced biogas during AD mesophilic trial. Despite the fast increase of methane content (above 60% v/v after two weeks of operation or 1.0 HRT approximately), the GPR trend showed a slower increase in the course of process stabilization and in parallel to the progressive increase of the applied OLR (roughly up to 1.5 kg VS/m 3 d) during the start up phase. At the steady state, GPR was 1.38 ± 0.08 m 3 biogas/m 3 r d; the percentage of methane detected in the biogas was 63 ± 3%; most of remaining part was CO 2 (32 ± 3%). The specific gas production (SGP) was found to be equal to 0.40 ± 0.02 m 3 biogas/kg VS.
Similarly to mesophilic condition, in the AD thermophilic trial the OLR was increased from 0.8 to 1.5-1.8 kg VS/m 3 d during the start up, and then it was maintained close to 2.5 kg VS/m 3 d over the course of the run. The progressive increase of the OLR reflected the trend of GPR, which was 1.80 ± 0.09 m 3 biogas/m 3 r d in the steady state, while methane biogas content exceeded 60% in 0.5 HRT approximately ( Figure 3D).  19 Figure 4.

Overall, both mesophilic and termophilic digestions of SC overflow diluted with thickened WAS
were demonstrated to be technically feasible and robust with reference to the process parameters (pH, partial alkalinity, VFA concentration, biogas production and composition). Table 4 summarizes the main monitored parameters and quantified performances in the steady state of the two AD trials.

Mass and energy balance
The mass balance discussed in this paragraph is illustrated in detail in Figure S1 (Supplementary Material). The data analysis of each separate pilot reactor was transferred to an industrial plant of average size, identified as 170,000 PE. Based on the specific production reported in the paragraph 2.3, the inlet amount of WAS corresponds to 10,200 kg TS/d. A smaller amount is driven to the fermentation reactor (around 24%) and the rest to the AD step (76%) in order to dilute the SC overflow. The inlet amount of the pre-treated OFMSW is 6,120 kg TS/d, based on a dry matter of 15% and a recovery of 80% of TS from the OFMSW screw-press pre-treatment (Majone et al.,

2017).
For the fermentation process, only the mesophilic condition was considered technically feasible and the relative yield of 0.41 g COD VFA /g VS (0) was taken into account. The bio-waste mixture is conveyed to the mesophilic fermenter with a TS and VS level of 6.3% and 5.2% w/w. As a   1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60  61  62  63  64  65 21 consequence, the applied OLR was approximately 10 kg VS/m 3 d. The gaseous effluent flow rate out of the acidogenic fermenter is 712 Nm 3 /d, corresponding to 1,153 kg VS/d, roughly equal to the 16% of the VS influent amount. The solid/liquid separation unit allows obtaining two streams: the liquid fraction and the SC overflow. The liquid stream has a volumetric flow rate of 105,946 kg/d (almost 80% of the total), corresponding to 2,262 kg COD VFA /d, to be used for the synthesis of bioproducts with higher added value than methane (Tamis and Joosse, 2015). In fact, a production of 0.54 tons of marketable biopolymer (PHA) per year is potentially feasible, based on average yield of 0.40 COD PHA /COD VFA , typically quantified in the aerobic MMC-PHA production technology (Valentino et al., 2017).
The SC overflow has a flow rate of 30,054 kg/d, which is mixed with 297,108 kg/d of excess thickened WAS. This dilution is necessary to decrease the solid content of the SC overflow, approximately from 23% w/w to 4.5% w/w (TS basin), before being fed into the final AD step.
The first scenario provides for the AD step in mesophilic condition (

Mass and energy balance assessment
The calculation for the assessment of mass and energy balance were made by considering the size of the plant (170,000 Person Equivalent), and the WAS and OFMSW specific production: 0.3 kg OFMSW/PE day (based on wet weight) and 0.06 kg TS/PE day for WAS (Metcalf & Eddy, 2003). The distribution of these two streams in the fermentation and AD step was defined as experimentally made, as well as for the relative volumetric portions (the VFA-rich liquid fraction and the solid cake fraction) after the solid/liquid separation unit. The OFMSW stream was assumed to be totally conveyed to the acidogenic fermentation step, together with the 24% of the WAS stream. Indeed, the volumetric percentage (as shown in Figure S1) were: 30% v/v OFMSW and 70% v/v WAS as feedstock for the fermentation step. The rest of WAS (76%) was used for the dilution of the solid cake, recovered from the solid/liquid separation unit. The solid cake (30,054 kg/d) roughly corresponds to the 22% of the total influent acidogenic fermentation flow rate (136,000 kg/d). Indeed, the influent of AD step (327,161 kg/d) was the sum of the solid cake and excess WAS flow rates.
The VFA production as well as biogas production was quantified based on the experimental data obtained: specifically the fermentation yield (gCOD VFA /gVS (0) ) and the specific gas production (SGP, m 3 biogas/kgVS).
The energy balance was assessed taking into account the parameters reported below: a) the volume of both reactors, defined by the ratio between the VS flow rate (kgVS/d) and the fixed OLR (kgVS/m 3 .d). The OLRs were: 10 kg VS/m 3 .d and 2.1 kgVS/m 3 .d, respectively for the fermenter and the digester. The volumes were 712 m 3 (acidic fermenter) and 5,213 m 3 (digester); b) the reference parameters and boundary conditions defined in  Figure S1. Flow rates and mass balance of the scaled up anaerobic process: mesophilic acidogenic fermentation followed by mesophilic or thermophilic anaerobic digestion.