Year-round biogas production in sugarcane biorefineries: Process stability, optimization and performance of a two-stage reactor system

Energy Conversion and Management 168 (2018) 188–199

Contents lists available at ScienceDirect

Energy Conversion and Management journal homepage: www.elsevier.com/locate/enconman

Year-round biogas production in sugarcane biorefineries: Process stability, optimization and performance of a two-stage reactor system

T

⁎

Leandro Jankea, , Sören Weinricha, Athaydes F. Leiteb, Heike Sträuberb, Claudemir M. Radetskic, Marcell Nikolauszb, Michael Nellesa,d, Walter Stinnera a

Department of Biochemical Conversion, Deutsches Biomasseforschungszentrum gemeinnützige GmbH, Torgauer Straße 116, 04347 Leipzig, Germany Department of Environmental Microbiology, Helmholtz Centre for Environmental Research – UFZ, Permoserstraße 15, 04318 Leipzig, Germany c Laboratório de Remediação Ambiental, Universidade do Vale do Itajaí, Rua Uruguai 458, 8320 Itajaí, Brazil d Faculty of Agricultural and Environmental Sciences, Chair of Waste Management, University of Rostock, Justus-von-Liebig-Weg 6, 18059 Rostock, Germany b

A R T I C LE I N FO

A B S T R A C T

Keywords: Sugarcane waste Anaerobic digestion Process design Methane potential Capacity factor

The concept of year-round biogas production to increase the capacity factor of anaerobic digestion (AD) plants in sugarcane biorefineries was investigated for the first time in semi-continuous feeding mode. To simulate the use of sugarcane vinasse during the sugarcane season and sugarcane filter cake (SFC) during the off-season period, a two-stage reactor system based on an acidogenic continuous stirred-tank reactor (1st stage) followed by solid–liquid separation and an upflow anaerobic sludge blanket (UASB) reactor (2nd stage) to convert the CODrich liquid fraction into biogas was operated. Additionally, to optimize the biogas production from SFC, the effects of its thermo-chemical pre-treatment on AD were investigated in a parallel reactor set-up. The saponification effect provided by autoclaving the substrate with sodium hydroxide improved the hydrolysis/fermentation of SFC in the acidogenic reactor, which in turn resulted in a 28% higher volumetric methane production in the methanogenic reactor (p < 0.05). However, the methane yields observed during operation of the two-stage reactor system were markedly lower than previously found in biochemical methane potential tests using SFC. In this case, the feed-in with low suspended solids required by UASB reactors prevented the utilization of the nonhydrolyzed/fermented solid fraction of SFC (> 60% of the substrate’s methane potential). Nevertheless, the capacity factor of the AD plants in sugarcane biorefineries could be increased from 0.55 up to 0.69 when considering a 200 d a−1 sugarcane season (0.66–0.83 for a longer season of 240 d a−1), representing an increase of 25.7%. The average capacity factor for biogas combined heat and power and upgrading units of around 0.91 (8000 h a−1) would be reached if further developments could improve the solubilization of non-hydrolyzed/ fermented solids or alternatively allow their direct use in the methanogenic reactor.

1. Introduction The anaerobic digestion (AD) process has been proven to be an alternative biomass conversion pathway to diversify the product portfolio of sugarcane biorefineries by recovering methane-rich biogas, promoting sustainable waste management practices and reducing greenhouse gas emissions that usually occur during temporary storage, transportation and application of sugarcane waste to the soil for water and nutrient recycling [1–3]. Among the different types of waste generated during sugarcane processing, sugarcane vinasse (SCV) and filter cake (SFC) are the most suitable substrates for biogas production due to their high availability, relatively easy degradability and favorable balance of nutrients. In addition, no competition with the current practice of soil application

⁎

Corresponding author. E-mail address: [email protected] (L. Janke).

https://doi.org/10.1016/j.enconman.2018.04.101 Received 2 March 2018; Received in revised form 26 April 2018; Accepted 27 April 2018 0196-8904/ © 2018 Elsevier Ltd. All rights reserved.

would occur, since the AD process is able to maintain the mineral content of the biomass in form of digestate allowing its proper use as organic fertilizer [4]. However, the seasonal characteristic of sugarcane crop limits the availability of the substrate to around 200–240 days per year, which in turn results in an energy system with a low capacity factor if SCV and SFC would be used for biogas production only during the sugarcane season. The low incentives to produce bioenergy in countries like Brazil (major sugarcane producer) and the insufficient profitability of the biogas projects with such characteristics has not encouraged the adoption of the AD technology by the sugarcane biorefineries in the recent years. In countries like Germany supporting the AD of energy crops, such as maize, sugar beet, and grass, the biomass naturally containing

Energy Conversion and Management 168 (2018) 188–199

L. Janke et al.

substrates provided fresh digestate, which was used as inoculum for the BMP tests and the acidogenic CSTRs. Lab-scale UASB reactors operated with SCV over 300 days provided adapted seed sludge used for the methanogenic UASB reactors [9].

moisture is harvested and conserved/stored by ensiling to be used as substrate throughout the whole year. This concept allows the operation of AD plants at an average capacity factor of around 0.91 (8000 h per year) often achieved by biogas combined heat and power (CHP) and upgrading units [5]. Thus, the profitability of biogas projects is improved since a major share of the capital expenditures (20–45%) of an agricultural AD plant is derived from the post-biogas producing facilities [6]. The concept of using SCV during the sugarcane season and conserving/storing SFC to be used as substrate during the off-season period was previously assessed by our research group based on biochemical methane potential (BMP) tests [7]. Despite the differences in methane potential between SCV derived from annexed and autonomous biorefineries (4.1–5.7 m3 CH4 tcane−1), SFC could maintain during the offseason period up to 85.6% of the daily methane production of SCV, thus demonstrating the potential of this concept for increasing the capacity factor of an AD plant, especially when considering that some of the obligatory plant downtime for maintenance could be planned to occur during the period with less methane production [7]. However, several challenges to implement such a biogas production concept remain. Regardless of the influence that different ensiling techniques for conserving and storing SFC could have on the final methane potential, the reactor configuration plays a key role. On the one hand, for substrates with low solid content (e.g. SCV) usually highrate anaerobic reactors are recommended, such as upflow anaerobic sludge blanket (UASB), expanded granular sludge bed (EGSB) or fixed bed. This is due to the fact that the immobilized biomass in form of biofilms allows shorter hydraulic retention times (HRT) resulting in lower reactor volumes. On the other hand, for substrates with high solid content (e.g. SFC), fully mixed anaerobic reactors are better suited, such as continuous stirred-tank reactors (CSTR), to allow a proper time to solubilize complex particulate organic matter [8]. The use of a different reactor type for each substrate in different periods of the year would simply transfer the idleness from the CHP/ upgrading units to the biogas producing process, most likely not improving the profitability of the projects. One possible strategy to overcome the drawbacks of reactor configuration is the use of an UASB reactor as a high-rate anaerobic reactor for SCV during the sugarcane season and decouple the AD of SFC into two steps during the off-season period, where SFC would be initially hydrolyzed/fermented in an acidogenic CSTR at short HRT of 3–5 days followed by solid-liquid separation and the existing UASB reactor would be used as a methanogenic reactor for biogas production from the separated liquid fraction of fermented SFC. Therefore, the concept of year-round biogas production in the sugarcane industry was experimentally assessed in the present study with the aims: (a) to investigate the process stability during gradual substrate substitution in the methanogenic UASB reactor; (b) to compare the methane potentials of different fractions of fermented SFC (liquid/ solid) with values from previous studies based on BMP tests; (c) to optimize the methane production from SFC by a thermo-chemical pretreatment method, and (d) to assess the increase of the capacity factor by using SFC as substrate during the sugarcane off-season. This approach can provide important inputs for an optimal process design leading to a more profitable utilization of these agricultural residues and thus facilitating the dissemination of the AD technology in the sugarcane sector.

2.2. Thermo-chemical pre-treatment SFC pre-treatment was conducted in 500 mL glass flasks with an alkaline reagent concentration of 6 g NaOH/100 g SFC based on fresh matter (FM). The substrate’s total solid (TS) content was 83 gTS L−1. SFC and NaOH solution were manually mixed and autoclaved for 30 min at 121 °C at 1 bar overpressure in a semi-automatic benchtop autoclave 2540 ML (Tuttnauer, Netherlands). After pre-treatment, SFC was neutralized with hydrochloric acid and stored at 4 °C until its use. 2.3. Semi-continuous experiment Two lab-scale CSTRs with 5 L total volume (3 L working volume) were used as acidogenic reactors being continuously stirred (100 rpm) using a central stirrer with vertical shaped blades to reduce the formation of floating layers. Fresh digestate from the CSTRs was daily centrifuged at 17,700 × g for 10 min at 10 °C in a lab-scale centrifuge Sorvall RC 6 plus (ThermoFisher, USA) for solid-liquid separation. The liquid fraction (hereafter referred to as liquid SFC) was used as substrate in two lab-scale methanogenic UASB reactors with 1.5 L total volume each (1.3 L working volume). To improve the substrate contact with the granular biomass in the UASB reactors, digestate was continuously re-circulated (5 mL min−1) by a peristaltic pump TU 200 (Medorex, Germany). The operation temperature in both CSTRs and in the UASB reactors was kept at mesophilic conditions (40 ± 1 °C) by recirculating hot water through the double-walled reactors. The schematic diagram of the two-stage reactor system is presented in Fig. 1. The experiment was conducted over 75 consecutive days in three different phases to simulate the SCV substitution with liquid SFC until reaching a technical steady-state during the last phase of the experiment [10]. For comparison, both CSTRs were fed with the same feeding frequency (once per day), organic loading rate (OLR) and HRT, only differing on substrate pre-treatment (control versus experimental reactor). The UASB reactors were automatically fed (20 times per day) by using a peristaltic pump PD 5201 (Heidolph, Germany) with the same HRT of 3.4 d during the whole experiment, but differing in OLR due to the effect of organic matter solubilization provided by the substrate pretreatment. For phase I of the experiment (days 0–12), the CSTRs were strategically fed with SFC to washout methanogens from the inoculum and to overload the reactors with organic acids. During this period, the UASB reactors were only fed with SCV to simulate the sugarcane season. For phase II (days 13–32) and phase III (days 33–75), the HRT was kept stable in the CSTRs (5 d), only differing in substrate input in UASB reactors. In this case, a gradual increase by 25% a week in SFC (fresh mass) was performed in phase II to avoid major disturbances in the microbial community due to the substrate substitution. Thus, during phase III only the liquid SFC was used as substrate in the methanogenic reactors to simulate the sugarcane offseason. For AD of SCV a combined supplementation of urea and trace elements was performed as described elsewhere [9]. The nutritional supplementation during AD of SFC was conducted in two steps: (a) during the CSTRs feeding a nutrients solution composed of 0.6 g S, 0.9 g Mn, 4.9 mg Co, 20.9 mg Cu, 16 mg Mo, 12 mg Ni, 5 mg W, 285 mg Ni and 2 mg Se per kg of TS was used and (b) during the UASBs feeding an urea solution of 2 g L−1 was applied. Detailed information about different feeding rates, OLR and HRT in each phase of the experiment is listed in Table 1.

2. Material and methods 2.1. Substrate and inoculum Samples of SCV and SFC were obtained from a distillery plant in the state of Goiás (Brazil) during the 2014/2015 season, transported to Germany in sealed plastic containers and stored at 4 °C until its use. A large-scale biogas plant that uses maize silage and cattle manure as

2.4. Biochemical methane potential tests The BMP of the solid digestate fraction from the acidogenic reactors (hereafter referred to as solid SFC) was determined according to VDI 189

Energy Conversion and Management 168 (2018) 188–199

L. Janke et al.

Fig. 1. Schematic diagram of the two-stage reactor system used during the semi-continuous experiment. Note: CSTR – continuous stirred-tank reactor. GLS – gasliquid-solid. UASB – upflow anaerobic sludge blanket. SCV – sugarcane vinasse. SFC – sugarcane filter cake.

substrate based on volatile solids - VS). The pH value in each batch reactor was measured before and after the BMP tests. Considering the different model derivations presented by Brulé et al. [11], an exponential two-pool one-step model (model C) was used to evaluate the methane production kinetics of the BMP tests. This modelling approach differentiates between rapidly and slowly degradable

(2016) using an Automatic Methane Potential Test System II (Bioprocess Control, Sweden) under mesophilic temperature (38 ± 1 °C) during 30 days. Prior to the BMP tests, the inoculum was degassed at 38 °C to reduce non-specific biogas production. To prevent inhibition the ratio of substrate/inoculum (gVS/gVS) was set to 0.23 ± 0.01 (i.e. around four times higher amount of inoculum than Table 1 Overview of the semi-continuous experiment. Reactor

Phase

Period (d)

SFCa (g d−1)

Water (mL d−1)

SCVb (mL d−1)

HRTc (d)

OLRd (gVS L−1 d−1 or gCOD L−1 d−1)

Acidogenic

I

0–2 3–4 5–6 7–8 9–10 11–12 13–32 33–77

300 250 220 190 165 150 150 150

900 750 650 565 490 450 450 450

– – – – – – – –

2.5 3.0 3.4 4.0 4.6 5.0

18.0 15.0 13.0 11.3 9.8 9.0

0–12 13–19 20–26 27–32 33–77

– 23.81 47.62 71.43 95.25

– 71.43 142.87 214.31 285.75

381 285.75 190.5 95.25 –

3.4

7.2 5.8 5.0 4.7 4.2

II III Methanogenic

I II

III

Note: ctrl means control reactor; expt means experimental reactor. a Sugarcane filter cake. b Sugarcane vinasse. c Hydraulic retention time. d Organic loading rate. 190

(ctrl); (ctrl); (ctrl); (ctrl);

6.8 6.7 6.0 6.1

(expt) (expt) (expt) (expt)

Energy Conversion and Management 168 (2018) 188–199

L. Janke et al.

fractions (two-pool) of the available substrate. Thus, four model parameters and constants needed to be adjusted to depict the respective measurement results: the total methane potential SBMP (mL gVS−1), the ratio of rapidly degradable substrate to total degradable substrate α, and the two first-order reaction constants for the degradation of rapidly degradable substrate kF (d−1) and slowly degradable substrate kL (d−1). The model implementation as well as the numeric parameter identification (Levenberg-Marquard algorithm) was performed in the software environment Matlab (Mathworks, USA) as previously described elsewhere [12].

benchtop spectrophotometer DR 3900 (Hach-Lange, Germany).

2.5. Analytical methods

2.7. Assessment on capacity factor

TS and VS were determined by drying the samples for 24 h at 105 °C in a drying oven (Binder, Germany) and further reducing the organic content to ashes for 2 h at 550 °C in a high temperature oven (Carbolite, UK). The TS content of SCV and SFC was corrected to account for losses of volatile compounds during oven drying according to Weiβbach [13]. The COD of non-centrifuged SCV was analyzed using a LCK 014 COD kit (Hach-Lange, Germany) according to the manufacturer’s protocol. Crude protein, crude lipids, crude fiber and nitrogen-free extracts of SFC were determined according to the Weender method [10]. To determine the macro element composition of SCV and SFC (C, H, N, S), about 30 mg of the substrates and 30 mg of WO3 were weighted and pressed in an aluminum foil for subsequent burning at 1150 °C catalytically with oxygen. The combustion gases were directed through a reduction tube where NOx is reduced to N2. The remaining gases (CO2, H2O, SO2) passed through three different adsorption columns and were detected with a thermal conductivity detector (C, H, N) and an infrared spectroscopy detector (S) by using a Vario Macro Cube (Elementar, Germany) [14]. To determine the trace element composition, dried samples were pre-treated with a mixture of HNO3/H2O2/HF and latter neutralized with H3BO3, and the resulting clear solution was analyzed by inductively coupled plasma atomic spectrometry – ICP-OES (ThermoFisher iCAP6200) [15]. The daily biogas production in both CSTRs and UASB reactors was measured by a drum-type gas meter TG 05 (Ritter, Germany) and corrected to standard temperature and pressure conditions (273.15 K and 101.325 kPa). The composition of the biogas (CH4, CO2, O2 and H2S) was measured twice a week at the headspace of the CSTRs and in a gas bag installed in the UASB reactors (after gas volume measurement) using a GA2000 Landfill Gas Analyzer (Geotechnical Instruments, UK). The specific methane production (SMP) is presented in milliliters per g of VS or COD (mL gVS−1 or mL gCOD−1) and the volumetric methane production (VMP) calculated in liters of methane per liter of reactor per day (L L−1 d−1). Five days per week the pH value of fresh digestate of the CSTRs was measured immediately after sampling (before feeding) with a pHelectrode Sentix 41 (WTW, Germany). With the same equipment and frequency the pH value of the digestate of the UASB reactors was measured directly in the digestate storage bottle. Twice a week digestate samples from both reactor types were centrifuged at 17,700 × g for 10 min at 10 °C. The supernatant was used after filtration for subsequent analyses, including the measurement of organic acids (OA) and the determination of the ratio of total inorganic carbonate to calcium carbonate (OA/alkalinity ratio, gOA gCACO3−1) by using a Titration Excellence T90 titrator (Mettler-Toledo GmbH, Switzerland). COD was analyzed using the same equipment as previously described. Volatile fatty acids (VFA), including acetic, propionic, n-butyric, iso-butyric, nvaleric, iso-valeric and caproic acid, were determined using a 5890 series II gas chromatograph (Hewlett Packard, USA) equipped with an HS40 automatic headspace sampler (Perkin Elmer, USA) and an Agilent HP-FFAP column (30 m × 0.32 mm × 0.25 µm) according to a method described elsewhere [16]. The ammonium-nitrogen concentration (NH4-N in g L−1) was determined from 500 µL filtered supernatant diluted with distilled water (1:500) with the Nessler method using a

To evaluate the benefits of using SFC as substrate during the offseason period, an assessment of the capacity factor of an AD plant integrated to an annexed sugarcane biorefinery was performed. The main input values are presented in Table 2. For simplification, the daily methane production that can be achieved using SCV as substrate during the sugarcane season times the maximum number of operation days in a year is assumed as nominal capacity of an AD plant [17]. Thus, the capacity factor of the different AD plant configurations assessed in this study was defined according to the Eq. (1) as follows:

2.6. Statistical analysis An analysis of variance (one-way ANOVA) was performed to verify whether statistical differences could be observed as an effect of the thermo-chemical pre-treatment on VMP and SMP in both acidogenic and methanogenic reactors during the semi-continuous experiment with a 95% confidence level. The calculations were run with the software MS-Excel 2016 (Microsoft, USA).

T

cf =

∑t CH 4 (t ) (1)

CH 4SCV × T

where: – cf is the capacity factor of the AD plant (–); – CH4 (t) is the methane produced in a certain time period T (m3 a−1); – CH4SCV is the methane production possible to be achieved from SCV during the sugarcane season (m3 d−1); – T is the maximum number of operation days in a year (T = 365 d a−1).

Table 2 Main input values used to assess the benefits on the capacity factor. Parameters

Value

Units

Milling capacity of the biorefinery Sugarcane seasona Sugarcane off-seasona SCV specific generationb Total SCV generation SCV COD concentrationc SCV specific methane productiond SFC specific generationa Total SFC generation SFC TS contentd SFC VS contentd SFC specific methane production in control CSTRd COD concentration in control liquid SFCd SFC specific methane production in control UASBd BMP of control solid SFCd SFC specific methane production in experimental CSTRd COD concentration in experimental liquid SFCd SFC specific methane production in experimental UASBd BMP of experimental solid SFCd

2 × 106 200–240 125 450 900 × 103 30 224.4 35 70 × 103 35.2 51.0 23.1 14.0 247.6 228.4 24.1

tcane a−1 d a−1 d a−1 L tcane−1 m3 a−1 g L−1 mL gCOD−1 kg tcane−1 t a−1 % of FM % of TS mL gVS−1 g L−1 mL gCOD−1 mL gVS−1 mL gVS−1

20.9 260.4

g L−1 mL gCOD−1

359.3

mL gVS−1

a

Approximate values [7]. Approximate value for annexed sugarcane biorefineries [1,7]. c Average value for Brazilian conditions [1]. d Obtained from the current experiment. The SFC specific methane production in control UASB was considering the values between days 33 and 53 due to the process imbalance observed in this reactor from day 60. b

191

Energy Conversion and Management 168 (2018) 188–199

L. Janke et al.

3.2. Start-up period of the semi-continuous experiment (phase I)

Table 3 Composition of sugarcane waste used during the biochemical methane potential tests and the semi-continuous experiment. Parameters

Sugarcane vinasse

Sugarcane filter cake

Units

Total solids (TS) Volatile solids (VS) Carbon (C) Nitrogen (N) Phosphorus (P) Sulfur (S) Iron (Fe) Calcium (Ca) Sodium (Na) Potassium (K) Magnesium (Mg) Nickel (Ni) Cobalt (Co) Molybdenum (Mo) Tungsten (W) Manganese (Mn) Copper (Cu) Selenium (Se) Zinc (Zn) Chemical oxygen demand (COD) Crude protein (XP) Crude lipids (XL) Crude fiber (XF) Nitrogen-free extracts (NFE) Ash

2.72 ± 0.40 71.9 ± 1.38 39.3 ± 1.05 3.53 ± 0.16 0.47 ± 0.01 1.58 ± 0.06 546 ± 3.80 7735 ± 20.30 491 ± 3.87 43,227 ± 872 5851 ± 159 0.79 ± 0.04 0.84 ± 0.03 0.86 ± 0.05 n.d 93.67 ± 1.45 9.58 ± 0.11 n.a 81.3 ± 0.49 24.70 ± 1.46

35.2 ± 0.44 51.0 ± 0.22 48.6 ± 2.01 2.44 ± 0.08 0.49 ± 0.04 0.18 ± 0.01 30,260 ± 1295 20,304 ± 1094 22.1 ± 2.69 1642 ± 32.5 3436 ± 68.2 17.9 ± 0.18 5.08 ± 0.47 n.d n.d 598 ± 49.2 59.1 ± 0.75 0.09 ± 0.02 115 ± 4.64 n.a

% FMa % TS % TS % TS % TS % TS mg kgTS−1 mg kgTS−1 mg kgTS−1 mg kgTS−1 mg kgTS−1 mg kgTS−1 mg kgTS−1 mg kgTS−1 mg kgTS−1 mg kgTS−1 mg kgTS−1 mg kgTS−1 mg kgTS−1 g L−1

n.a n.a n.a n.a 296 ± 11.4

125 ± 1.14 50.7 ± 1.45 225 ± 5.92 120 ± 10.75 477 ± 2.23

g kgTS−1 g kgTS−1 g kgTS−1 g kgTS−1 g kgTS−1

3.2.1. Process stability To provide optimal conditions for VFA production the initial OLR of the CSTRs was set to 18.0 gVS L−1 d−1 and gradually decreased to 9.0 gVS L−1 d−1 while the HRT was increased from 2.5 d to 5 d during phase I (days 0–12). Although the same trend of inoculum washout in both CSTRs was observed during this period (depicted from TAN reduction), the reactors behaved differently in terms of process overloading (Figs. 2 and 3). While the control CSTR showed a transient VFA increase to 4.6 g L−1 followed by a decrease to 2.1 g L−1 (pH value of 6.8 at day 12), the experimental CSTR presented a successful start-up for an acidogenic reactor with a constant VFA increase from 0.03 g L−1 to 8.5 g L−1 (pH value of 5.7 at day 12). Interestingly, in a previous study with a similar start-up strategy for VFA production from SFC without pre-treatment and trace elements supplementation, no adverse effects for the reactor start-up were reported [24]. In the present study, the reason for the reduced VFA production in the control CSTR can be found in the supplementation of trace element, which are well known to stimulate methanogenesis resulting in the conversion of VFA to biogas, and thus preventing a drop in the pH value. In contrast, the experimental CSTR showed a satisfactory start-up (even with trace elements supplementation) due to the additional COD release (rich in VFA) as result of the thermo-chemical pre-treatment which might have helped to keep the pH below the optimum value for methanogenesis (< 6.5) [25,26]. During the same period, the UASB reactors were operated in parallel using SCV as substrate with a constant OLR of 7.2 gCOD L−1 d−1 and a HRT of 3.4 d. Under these conditions, the AD process in both reactors was considered stable with an average VFA concentration of 0.13 ± 0.06 g L−1 (OA/alkalinity ratio of 0.13 ± 0.03 g L−1) and a pH value of 8.2 ± 0.1. Based on the previous long-term operation of these reactors with the same substrate (over 300 days), the length of phase I (12 days) and the HRT (3.4 d), it was assumed that the reactors were in steady-state [10].

Note: values are presented in mean (n = 3); ± represents the standard deviation. FM – Fresh matter. n.d – not detected. n.a – not analyzed.

3. Results and discussion 3.2.2. Methane production The methane production monitored in the CSTRs and UASB reactors is shown in Fig. 4. Significant differences (p < 0.05) in VMP and SMP between control and experimental CSTRs were observed mainly due to the reduced VFA accumulation observed in the control CSTR. The experimental CSTR presented a gradual decrease of the VMP to 0.5 L L−1 d−1 at day 12, whereas the control CSTR showed a constant VMP of 0.9 L L−1 d−1 in this experimental phase. Additionally, important differences were found in the biogas composition, which has shown a shift from methane to carbon dioxide as the major gas compound produced by the experimental CSTR and a constant methane concentration of 60.7% (v/v) found in the control CSTR (Fig. 5). Indeed, these results reinforce the argument of higher methanogenic activity in the control than in the experimental CSTR. In this case, methane could still be constantly produced even at relatively low HRT of 5 d in a CSTR due to the supplementation of trace elements and slower hydrolysis/fermentation of the substrate without pre-treatment. The average SMP in the UASB reactors reached 224.4 ± 7.3 mL gCOD−1 which is 6% lower than in our previous study in a mesophilic UASB reactor using SCV from an autonomous sugarcane mill and in the lower range of 220–269 mL gCOD−1 for mesophilic BMP reported elsewhere [7,9,27].

3.1. Substrate composition The main characteristics of SCV and SFC used in this study are presented in Table 3. The detailed assessment of both substrates has already been performed by our research group [18–20]. The most relevant observations on SCV were: (a) the low TS content of 2.7 ± 0.40% FM from autonomous mills (without sugar production), suggesting the necessity of AD reactors with biomass immobilization to avoid excessive reactor volumes; (b) the high sulfur content (1.58 ± 0.06% TS) of around 8 times higher than the recommended value as a nutrient which could potentially promote sulphidogenesis instead of methane formation, induce the precipitation of metal sulfides that reduce the availability of trace elements for microbial uptake and contaminate biogas with corrosive H2S, and (c) the acidic pH of around 3–4 (data not shown) requiring the development of alkalizing strategies to avoid reactor acidification [5,21–23]. Regarding SFC the high TS (35.2 ± 0.44% FM) and low VS (51.0 ± 0.22% FM) of the sample used in this study were notable. In this case, metal-based coagulants/flocculants (e.g. aluminum sulphate or ferric chloride) and neutralization agents (e.g. calcium hydroxide) added during sugarcane juice clarification could have contributed to increase the TS and reduce the VS content. Additionally, seasonal aspects such as humidity/drought that influences the amount of soil particles carried with the sugarcane during harvesting could also have contributed to these characteristics, since part of these inorganic particles are removed during SFC generation process.

3.3. Transition period of the semi-continuous experiment (phase II) 3.3.1. Process stability For phase II (days 13–32), the use of the liquid SFC derived from the CSTRs was gradually increased alongside with SCV in the UASB reactors to simulate the transition period to the sugarcane off-season. During this period, the OLR and HRT of the CSTRs were kept stable at 192

Energy Conversion and Management 168 (2018) 188–199

L. Janke et al.

Fig. 2. Main parameters monitored during the semi-continuous experiment. (a) volatile organic acids (OA) in CSTRs; (b) OA in UASB reactors; (c) ratio of organic acids with total inorganic carbonate to calcium carbonate (OA/alkalinity ratio) in CSTRs; (d) OA/alkalinity ratio in UASB reactors; (e) total ammonium-nitrogen (TAN) in CSTRs; (f) TAN in UASB reactors; (g) pH value in CSTRs; (h) pH value in UASB reactors.

9.0 gVS L−1 d−1 and 5 d, respectively. Under these conditions, the average VFA production in the experimental CSTR reached 10.1 ± 0.3 g L−1 (pH value of 5.6 ± 0.1). However, the control CSTR still presented a lower VFA concentration (2.3 ± 0.1 g L−1) due to its higher pH value of 7.0 ± 0.1. Therefore, a single dose of hydrochloric acid was added to the control CSTR at day 18 to adjust the pH to the value of 5.5. Immediately after this intervention, the concentration of VFA increased to an average of 7.8 ± 0.4 g L−1 at a stable pH value of 5.4 ± 0.1. The further conversion of these organic acids together with SCV into biogas in the UASB reactors was considered stable since no major VFA accumulation was observed (VFA < 0.1 g L−1). However, a slight decrease in the pH value was noticed as a result of a reduction in TAN concentrations, which has reduced the buffering capacity of the reactors [28,29].

3.3.2. Methane production As expected, the experimental CSTR showed negligible VMP (0.1 ± 0.05 L L−1 day−1) and SMP (< 25 mL gVS−1) during the transition period while the control CSTR could only achieve such low methane production levels after artificial controlling of the pH by the addition of hydrochloric acid. The SMP did not show significant differences (p > 0.05) between control and experimental UASB reactors, keeping similar values to those found during phase I. However, the lower VFA production of the control CSTR in comparison to the experimental one resulted in different VMP in the UASB reactors (p < 0.05). In addition, due to the lower COD production from both CSTRs in comparison to the COD of raw SCV (24.7 ± 1.46 gCOD L−1), a gradual decrease of the VPM during this experimental phase was observed in both UASB reactors. Furthermore, the biogas composition in the UASB reactors markedly changed during substrate transition. In this case, a trend of increasing 193

Energy Conversion and Management 168 (2018) 188–199

L. Janke et al.

Fig. 3. Volatile fatty acids (VFA) monitored during semi-continuous experiment. (a) control CSTR; (b) control UASB reactor; (c) experimental CSTR; and (d) experimental UASB reactor.

3.4. Steady-state period of the semi-continuous experiment (phase III)

methane concentration from 60.4 ± 0.9 (% v/v) to 69.5 ± 0.5 (% v/ v) was observed. In the meantime, the hydrogen sulfide concentration was drastically reduced from more than 5000 ppm (upper detection limit of the gas analyzer) to around 600 ppm at the end of this phase. Such increase in methane concentration could have occurred by different reasons: (a) sulfur-reducing bacteria outcompete methanogens for hydrogen and acetate due to thermodynamic advantages, thus reducing the amount of sulfur-rich SCV would result in more methane and less hydrogen sulfide; (b) by separating the AD process in two phases carbon dioxide is primarily produced in the acidogenic reactor, thereby methane-enriched biogas is found in the methanogenic reactor [30,31].

3.4.1. Process stability To provide consistent data for comparing the different reactor setups under steady-state condition no changes in OLR or HRT were made during phase III (days 33–77) of the experiment. The control CSTR showed an average VFA production of 7.7 ± 1.1 g L−1 (pH value of 5.6 ± 0.3), resulting in a VFA yield of 0.29 gVFA gVS−1. Based on the average COD solubilized during this period (14.0 ± 1.8 g L−1), a conversion of approximately 56% of the soluble compounds produced by hydrolysis into VFA were assumed (VFA:COD). In the meantime, the experimental CSTR showed a very similar pH

Fig. 4. Methane production monitored during the semi-continuous experiment. (a) volumetric methane production (VMP) in the CSTRs; (b) VMP in the UASB reactors; (c) specific methane production (SMP) in the CSTRs; and (d) SMP in the UASB reactors. 194

Energy Conversion and Management 168 (2018) 188–199

L. Janke et al.

Fig. 5. Biogas composition monitored during the semi-continuous experiment. (a) control CSTR; (b) control UASB reactor; (c) experimental CSTR; and (d) experimental UASB reactor.

different patterns in terms of biogas composition. In this case, the major gas compound was methane (54.4 ± 1.4% v/v) in the control CSTR, while the most dominant gas produced in the experimental CSTR was carbon dioxide (55.1 ± 2.2% v/v). Interestingly, both CSTRs presented an average SMP of 23.6 ± 0.7 mL gVS−1, which is higher than the negligible SMP observed in our previous study due to the stimulating effect of methanogenic activity provided by the trace elements supplementation [24]. Due to the higher COD solubilization in the experimental CSTR and the improved process stability observed in the experimental UASB reactor, the NaOH pre-treatment applied to SFC was responsible for a significant (p < 0.05) increase in VMP and SMP of the experimental UASB reactor during phase III of the experiment. Despite the reduced methane production in the control UASB as a consequence of the momentary process imbalance, a general trend of increasing SMP was observed in both UASB reactors. In this case, around 32% higher SMP was observed in the last HRT of the experimental UASB reactor in comparison to the values found during phase I and II. This improved performance is related to the higher concentration of intermediate products such as VFA found in liquid SFC (10.5 ± 0.35 g L−1) than in raw SCV (4.2 ± 0.03 g L−1) which are faster converted to methane than the more recalcitrant suspended solids contained in SCV [9,37].

value (5.6 ± 0.1) to the control CSTR, but with a 38% higher VFA production (10.7 ± 0.2 g L−1). Such improved performance resulted in a VFA yield of 0.40 gVFA gVS−1 and a slightly lower VFA:COD ratio 0.51. These values indicate that the NaOH pre-treatment enhanced primarily the hydrolysis phase, therefore reinforcing hydrolysis as the rate-limiting step during AD of lignocellulosic substrates [32]. In the UASB reactors, the trend of the TAN concentration reduction was confirmed, reaching the lowest values observed during the whole experiment (0.7 ± 0.1 g L−1). Nevertheless, this nitrogen depletion did not impair the stability of the AD process neither as buffer agent nor as nutrient source. In fact, the urea addition during the whole experiment (2 g L−1) was responsible for providing sufficient alkalinity to counteract a VFA accumulation of up to 3.2 g L−1 (41% of acetic, 27% of nbutyric, 14% of propionic, 11% of iso-butyric and 7% of other organic acids) since a stable pH value of 7.8 ± 0.1 was observed between days 53–77 in the control UASB reactor (Figs. 2 and 3). Interestingly, the AD process in the control UASB reactor was able to recover without any intervention, since towards to the end of the experiment the accumulated VFA was converted to biogas reaching similar concentration found in the experimental UASB reactor. The addition of water to maintain the HRT in the acidogenic reactors may have diluted important ions such as Na+ and Cl−, thus negatively affecting both the conductivity in the solution and the exchange of electrons between syntrophic microorganisms, ultimately causing the momentary process imbalance observed in the control UASB [33]. In this case, the NaOH pre-treatment followed by the pH neutralization with HCl could have contributed to the direct interspecies electron transfer (DIET) and/or providing sufficient Na+ as a nutrient for microbial growth since no process imbalance was observed in experimental UASB reactor during the whole experiment [34–36].

3.5. Biochemical methane potential tests of the acidified solid fraction of SFC Due to the high concentration of solids in the digestate obtained from the acidogenic CSTRs (7.2 ± 0.7% FM), a solid-liquid separation was performed to reduce its suspended solids concentration producing an acidified liquid fraction (COD-rich) suitable to be used as substrate for biogas production in the UASB reactors. To assess the methane potential of the separated solids, BMP tests were conducted and the results are presented in Table 4 and Fig. 6. Generally, the utilized twopool first-order kinetic model can depict the methane progression of the

3.4.2. Methane production The VMP and SMP did not show significant differences (p > 0.05) between control and experimental CSTRs. However, they presented 195

Energy Conversion and Management 168 (2018) 188–199

L. Janke et al.

Table 4 Results of the biochemical methane potential tests of the solid fractions from the acidogenic reactors fitted to the two-pool one-step model. Substrate

SBMP (mL gVS−1)

α (–)

kF (d−1)

kL (d−1)

R2 (–)

SBMPc (mL gVS−1)

αc (–)

R2c (–)

Increase Sc (%)

Digestate – control CSTRa Digestate – experimental CSTRb

228 ± 19.0 355 ± 21.6

0.43 ± 0.01 0.38 ± 0.03

0.37 ± 0.02 0.44 ± 0.04

0.05 ± 0.01 0.05 ± 0.01

1.00 0.99

228 ± 4.30 359 ± 15.4

0.43 ± 0.01 0.41 ± 0.04

1.00 0.99

– 63.56

Note: SBMP – total methane potential (approximated at infinite retention time). α – ratio of rapidly degradable substrate components to total degradable substrate. kF – first-order reaction constant for rapidly degradable substrate components. kL – first-order reaction constant for slowly degradable substrate components. R2 – coefficient of determination. a Values are presented in mean (n = 2); ± represents the standard deviation. b Values are presented in mean (n = 3); ± represents the standard deviation. c Optimized values (curve fitting) for constant values of kF = 0.37 (d−1) and kL = 0.05 (d−1).

Fig. 6. Cumulative methane yields obtained from the batch experiment and fitted to the two-pool one-step model. (a) Solid fraction from the control CSTR and (b) solid fraction from the experimental CSTR.

BMP tests with a high degree of precision (R2 > 0.99). The total methane potential (SBMP) of solid SFC from the control CSTR reached 228 ± 26.9 mL gVS−1, which is similar to values previously reported for raw SFC in a 30 days batch experiment (231.3 ± 10.6 mL gVS−1) and lower than values found in 35 days batch experiments (245–281 mL gVS−1) [7,27,28]. In the meantime, a markedly higher methane potential of 355 ± 26.8 mL gVS−1 was observed for the solid SFC from the experimental CSTR, representing an increase of 63.5% in comparison to the control sample. For fixed kinetic constants (kF = 0.37 d−1 and kL = 0.05 d−1) the model revealed almost identical shares of rapidly and slowly degradable components for the control and experimental CSTR (α = 0.43 ± 0.02 and 0.41 ± 0.03 respectively), indicating that substrate pretreatment did not affect the final distribution of slowly and rapidly degradable components in the solid phase of the digestate. The above values suggest that the 5 d of HRT used in both acidogenic CSTRs was not sufficient to hydrolyze/ferment all organic matter since the solid fraction after digestate separation still presented considerable methane potential. Interestingly, the solid SFC from the experimental CSTR showed a methane potential even higher than previously reported for raw SFC (317 mL gVS−1) pre-treated with the same concentration of NaOH used in this experiment (6 g NaOH/100 g SFCFM), suggesting that the acidogenic CSTR might have provided a second biological pre-treatment effect, resulting in a combined thermochemical-microbial pre-treatment method [38]. In addition, the solidliquid separation method used in this experiment could also have contributed to such increased performance since the active biomass most likely stayed in the solid fraction after 10 min centrifugation at 17,700 × g.

3.6. Impact on capacity factor To evaluate the relevance of the proposed AD concept for full-scale application an energy assessment focused on the improvement of the capacity factor of different AD plant configurations in an annexed sugarcane biorefinery (2 × 106 tcane a−1) was performed (Table 5). Thus, considering the use of SCV for biogas production directly in the UASB reactor, it would be possible to produce 30,294 m3 CH4 d−1 in a 200 d a−1 sugarcane season, or alternatively 25,245 m3 CH4 d−1 in a longer season of 240 d a−1. The use of untreated SFC for biogas production in the two-stage reactor system during sugarcane off-season would result in 6540 m3 CH4 d−1 (8633 m3 CH4 d−1 for a 240 d a−1 season), distributed in 25% for the acidogenic reactor and 75% for the methanogenic reactor. Such off-season methane production would allow an increase by 17.8% in the capacity factor of the biogas plant from 0.55 to 0.69 and from 0.66 to 0.77 for 200 d a−1 and 240 d a−1 season, respectively. The thermo-chemical pre-treatment confirmed to be a promising strategy to further improve the utilization of the CHP and/or upgrading units since, in comparison to the untreated SFC, an additional 45% of off-season methane production could be achieved thanks to the alkaline pre-treatment (9421 m3 CH4 d−1 and 12,436 m3 CH4 d−1 for 200 d a−1 and 240 d a−1 season, respectively). In this case, 18% of these methane production were derived from the acidogenic reactor and 82% from the methanogenic reactor owing to the enhanced VFA production in the experimental CSTR (described in Section 3.4.1). Thus, the proposed AD concept for year-round biogas production would result in an AD plant with a capacity factor of 0.69 for 200 d a−1 season (0.83 for 240 d a−1 season), which represents for both season lengths an increase by 25.7% 196

Energy Conversion and Management 168 (2018) 188–199

L. Janke et al.

Table 5 Methane production during sugarcane season and off-season periods and its impact on the capacity factor. Parameters

Season SCV

Off-season Untreated SFC

UASB

CSTR

UASB

CSTR + UASB

Solida

Totalb

CH4 production (m3 d−1) Capacity factor (–) Increase (%)

30,294 (25,245) 0.55 (0.66) –

1657 (2188) 0.57 (0.69) 4.5

4883 (6445) 0.62 (0.74) 13.3

6540 (8633) 0.65 (0.77) 17.8

10,686 (14,106) 0.71 (0.85) 29.1

17,226 (22,739) 0.80 (0.97) 46.9

Parameters

Season SCV

Off-season Pre-treated SFC

UASB

CSTR

UASB

CSTR + UASB

Solida

Totalb

30,294 (25,245) 0.55 (0.66) –

1731 (2287) (0.69) 4.7

7691 (10,152) 0.66 (0.80) 20.9

9421 (12,436) 0.69 (0.83) 25.7

16,193 (21,375) 0.79 (0.95) 44.1

25,614 (33,811) 0.93 (> 1.00c) 69.8

CH4 production (m3 d−1) Capacity factor (–) Increase (%)

Values in parentheses are for 240 days of sugarcane season. c Capacity factor > 1.00 because the daily methane production during the off-season period would be higher than the daily methane production during the sugarcane season. a Only solid fraction derived from the CSTR. b CSTR + UASB and solid fraction derived from the CSTR.

In contrast to our concept, a recent study has proposed the co-digestion of SCV and SFC in a sequential two-stage thermophilic UASB reactor system [39]. Even though this set-up would be advantageous for the AD process in terms of improved balance of nutrients and alkalinity allowing high OLRs and low HRTs, the biogas production would have to occur necessarily during the sugarcane season due to the fact that ensiling substrates with low solid content as SCV is technically and economically challenging. Thus, unless the produced biogas is intended to be used as co-fuel in the existing bagasse-based co-generation systems, such co-digestion concept would amplify the idleness of biogas CHP and/or upgrading units. In our experiment, the lignocellulosic recalcitrance of SFC and the low acceptance of suspended solids by UASB reactors are the major factors limiting the performance of the proposed AD concept. In this case, more effective pre-treatment techniques to enhance the solubilization of SFC would increase the overall performance of the system. Alternatively, more suitable solid-liquid separation systems could be used in combination with methanogenic reactors able to operate with higher suspended solids concentration. Furthermore, specifically designed ensiling systems for SFC conservation and storage should be developed. It is known that the organic acids produced during ensiling processes can act as a pre-treatment method to breakdown the lignocellulosic structure and make cellulose and hemi-cellulose better accessible for hydrolytic enzymes, potentially increasing the methane potential of AD substrates [40,41]. Besides the conventional ensiling techniques of bailing or semiburied silos, another interesting alternative for SFC could be a wetsystem where the TS of SFC would be adjusted with SCV producing a pumpable mixture able to be stored in low-cost open or covered lagoons under adequate pH (pH of SCV: 3–5). Such lagoon-based silo could provide positive effects for the substrate hydrolysis since the retention time would reach the whole season (200–240 d), possibly also making the use of a further acidogenic CSTR unnecessary as performed during the current study. This system still needs to be technically proven, in particular concerning possible emissions if open lagoons are used and consequent energy and carbon losses in form of methane, carbon dioxide or hydrogen [42].

in comparison to producing biogas only from SCV during the sugarcane season. It is important to note that independent of the biomass pre-treatment, a larger share of methane potential is not being utilized in the form of solid acidified SFC (> 60%). Therefore, under the conditions applied in this study, the proposed two-stage reactor system was able to recover between 40% and 57% of the SMP obtained from raw SFC (231.3 mL gVS−1) in previous studies based on BMP tests [38]. In case the hydrolysis/fermentation of the complex solid particles could be enhanced in future developments, the proposed AD concept could potentially provide an off-season methane production of 17,226 m3 CH4 d−1 in a 200 d a−1 season (22,793 m3 CH4 d−1 for 240 d a−1 season) for untreated SFC (capacity factor of 0.80 and 0.97 for 200 d a−1 and 240 d a−1 season, respectively) and 3 −1 25,614 m CH4 d in a 200 d a−1 season (33,811 m3 CH4 d−1 for 240 d a−1 season) for pre-treated SFC (capacity factor of 0.93 and > 1.00 for 200 d a−1 and 240 d a−1 season, respectively). Thus, exceeding the average capacity factor of biogas CHPs and upgrading units of around 0.91 (8000 h per year). 4. Final remarks An innovative AD concept to overcome crop seasonality in sugarcane biorefineries was proposed by using SFC as substrate for methane production during the off-season period. This concept is based on an acidogenic CSTR followed by solid-liquid separation allowing the subsequent UASB reactor to be used as high-rate methanogenic system for both SCV and liquid acidified SFC, thus requiring less reactor volume as these waste streams were treated in independent reactor systems. Another possible reactor configuration is the use of low-rate onestage reactor systems, such as CSTRs, partial mixed (sprinkler recirculation) or covered anaerobic lagoons. The use of these reactors could make the AD operation simpler and allow the full utilization of SFC if the reactor’s hydraulic systems would be designed to operate both with low and high suspended solids content. However, a much larger reactor volume would be required to avoid the washout of slowgrowing methanogens (HRT > 10 d). Nevertheless, the additional costs of larger reactors could be compensated with construction techniques able to take advantage of the economy of scale since in medium size sugarcane biorefineries (2 × 106 tcane a−1) a relatively large AD plant of around 4.2 MWel could be installed based on the SCV potential during 240 d a−1 of sugarcane season (25,245 m3 CH4 d−1).

5. Conclusions The present study evaluated a possible reactor configuration for year-round biogas production in sugarcane biorefineries. A thermochemical pre-treatment based on autoclaving with NaOH improved the 197

Energy Conversion and Management 168 (2018) 188–199

L. Janke et al.

hydrolysis/fermentation of SFC in the acidogenic CSTR, resulting in a 28% higher volumetric methane production in the methanogenic UASB reactor (p < 0.05). However, the UASB as a methanogenic reactor limited the use of the fiber fraction of SFC, considerably reducing the potential of this substrate for methane production during the off-season period. Nevertheless, an increase of up to 25.7% in the capacity factor from 0.55 up to 0.69 was achieved when considering a 200 d a−1 sugarcane season (0.66–0.83 for a longer season of 240 d a−1). Further developments focusing on SFC ensiling, pre-treatment, solid-liquid separation and reactor configuration can allow the operation of biogas CHPs and upgrading units at their average capacity factor of 0.91 (8000 h a−1) since when both liquid and solid fractions of acidified SFC are considered such value was exceeded.

plasma optical emission spectrometry (ICP-OES). Berlin: Beuth Verlag GmbH; 2011. [16] Sträuber H, Bühligen F, Kleinsteuber S, Nikolausz M, Porsch K. Improved anaerobic fermentation of wheat straw by alkaline pre-treatment and addition of alkali-tolerant microorganisms. Bioengineering 2015;2:66–93. http://dx.doi.org/10.3390/ bioengineering2020066. [17] Tu YL, Chang TJ, Hsieh CI, Shih JY. Artificial neural networks in the estimation of monthly capacity factors of WECS in Taiwan. Energy Convers Manag 2010;51:2938–46. http://dx.doi.org/10.1016/j.enconman.2010.06.035. [18] Janke L, Leite AF, Nikolausz M, Radetski CM, Nelles M, Stinner W. Comparison of start-up strategies and process performance during semi-continuous anaerobic digestion of sugarcane filter cake co-digested with bagasse. Waste Manag 2016;48:199–208. http://dx.doi.org/10.1016/j.wasman.2015.11.007. [19] Leite A, Janke L, Lv Z, Harms H, Richnow H-H, Nikolausz M. Improved monitoring of semi-continuous anaerobic digestion of sugarcane waste: effects of increasing organic loading rate on methanogenic community dynamics. Int J Mol Sci 2015;16:23210–26. http://dx.doi.org/10.3390/ijms161023210. [20] Leite AF, Janke L, Harms H, Richnow H-H, Nikolausz M. Lessons learned from the microbial ecology resulting from different inoculation strategies for biogas production from waste products of the bioethanol/sugar industry. Biotechnol Biofuels 2016;9:144. http://dx.doi.org/10.1186/s13068-016-0548-4. [21] Gustavsson J, Shakeri Yekta S, Sundberg C, Karlsson A, Ejlertsson J, Skyllberg U, et al. Bioavailability of cobalt and nickel during anaerobic digestion of sulfur-rich stillage for biogas formation. Appl Energy 2013;112:473–7. http://dx.doi.org/10. 1016/j.apenergy.2013.02.009. [22] Fuess LT, de Araújo Júnior MM, Garcia ML, Zaiat M. Designing full-scale biodigestion plants for the treatment of vinasse in sugarcane biorefineries: how phase separation and alkalinization impact biogas and electricity production costs? Chem Eng Res Des 2017;119:209–20. http://dx.doi.org/10.1016/j.cherd.2017.01.023. [23] Fuess LT, Kiyuna LSM, Ferraz ADN, Persinoti GF, Squina FM, Garcia ML, et al. Thermophilic two-phase anaerobic digestion using an innovative fixed-bed reactor for enhanced organic matter removal and bioenergy recovery from sugarcane vinasse. Appl Energy 2017;189:480–91. http://dx.doi.org/10.1016/j.apenergy.2016. 12.071. [24] Janke L, Leite A, Batista K, Weinrich S, Sträuber H, Nikolausz M, et al. Optimization of hydrolysis and volatile fatty acids production from sugarcane filter cake: effects of urea supplementation and sodium hydroxide pretreatment. Bioresour Technol 2016;199:235–44. http://dx.doi.org/10.1016/j.biortech.2015.07.117. [25] Janke L, Weinrich S, Leite AF, Terzariol FK, Nikolausz M, Nelles M, et al. Improving anaerobic digestion of sugarcane straw for methane production: combined benefits of mechanical and sodium hydroxide pretreatment for process designing. Energy Convers Manag 2016. http://dx.doi.org/10.1016/j.enconman.2016.09.083. [26] Appels L, Baeyens J, Degrève J, Dewil R. Principles and potential of the anaerobic digestion of waste-activated sludge. Prog Energy Combust Sci 2008;34:755–81. http://dx.doi.org/10.1016/j.pecs.2008.06.002. [27] Leite AF, Janke L, Harms H, Zang JW, Fonseca-Zang WA, Stinner W, et al. Assessment of the variations in characteristics and methane potential of major waste products from the brazilian bioethanol industry along an operating season. Energy Fuels 2015;29:4022–9. http://dx.doi.org/10.1021/ef502807s. [28] Janke L, Weinrich S, Leite AF, Schüch A, Nikolausz M, Nelles M, et al. Optimization of semi-continuous anaerobic digestion of sugarcane straw co-digested with filter cake: Effects of macronutrients supplementation on conversion kinetics. Bioresour Technol 2017;245:35–43. http://dx.doi.org/10.1016/j.biortech.2017.08.084. [29] Schmidt T, Ziganshin AM, Nikolausz M, Scholwin F, Nelles M, Kleinsteuber S, et al. Effects of the reduction of the hydraulic retention time to 1.5 days at constant organic loading in CSTR, ASBR, and fixed-bed reactors – performance and methanogenic community composition. Biomass Bioenergy 2014;69:241–8. http://dx.doi. org/10.1016/j.biombioe.2014.07.021. [30] Shakeri Yekta S, Lindmark A, Skyllberg U, Danielsson Å, Svensson BH. Importance of reduced sulfur for the equilibrium chemistry and kinetics of Fe(II), Co(II) and Ni (II) supplemented to semi-continuous stirred tank biogas reactors fed with stillage. J Hazard Mater 2014;269:83–8. http://dx.doi.org/10.1016/j.jhazmat.2014.01.051. [31] Linke B, Rodríguez-Abalde Á, Jost C, Krieg A. Performance of a novel two-phase continuously fed leach bed reactor for demand-based biogas production from maize silage. Bioresour Technol 2015;177:34–40. http://dx.doi.org/10.1016/j.biortech. 2014.11.070. [32] Weinrich S, Nelles M. Critical comparison of different model structures for the applied simulation of the anaerobic digestion of agricultural energy crops. Bioresour Technol 2015;178:306–12. http://dx.doi.org/10.1016/j.biortech.2014. 10.138. [33] Kamcev J, Sujanani R, Jang E-S, Yan N, Moe N, Paul DR, et al. Salt concentration dependence of ionic conductivity in ion exchange membranes. J Memb Sci 2018;547:123–33. http://dx.doi.org/10.1016/j.memsci.2017.10.024. [34] Barua S, Dhar BR. Advances towards understanding and engineering direct interspecies electron transfer in anaerobic digestion. Bioresour Technol 2017;244:698–707. http://dx.doi.org/10.1016/j.biortech.2017.08.023. [35] Baek G, Jung H, Kim J, Lee C. A long-term study on the effect of magnetite supplementation in continuous anaerobic digestion of dairy effluent – magnetic separation and recycling of magnetite. Bioresour Technol 2017;241:830–40. http:// dx.doi.org/10.1016/j.biortech.2017.06.018. [36] Jha P, Schmidt S. Reappraisal of chemical interference in anaerobic digestion processes. Renew Sustain Energy Rev 2017;75:954–71. http://dx.doi.org/10.1016/ j.rser.2016.11.076. [37] Wilkie AC, Riedesel KJ, Owens JM. Stillage characterization and anaerobic treatment of ethanol stillage from conventional and cellulosic feedstocks. Biomass Bioenergy 2000;19:63–102. http://dx.doi.org/10.1016/S0961-9534(00)00017-9. [38] Janke L, Leite A, Batista K, Weinrich S, Sträuber H, Nikolausz M, et al. Optimization

Acknowledgements The authors would like to acknowledge the technical assistance of the students Karoline Rodrigues and Gabriell Moura. This work was supported by the Brazilian National Scientific Counsel (CNPq) under the Program Science without Borders, individual grants of Leandro Janke (237938/2012-0) and Athaydes Leite (202024/2012-1). This work was also partially financed by the i-NOPA Project “Sustainable bioeconomy in Brazil: Bioenergy from biogas using various types of waste substrates from the Brazilian bioethanol industry”. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.enconman.2018.04.101. References [1] Moraes BS, Junqueira TL, Pavanello LG, Cavalett O, Mantelatto PE, Bonomi A, et al. Anaerobic digestion of vinasse from sugarcane biorefineries in Brazil from energy, environmental, and economic perspectives: profit or expense? Appl Energy 2014;113:825–35. http://dx.doi.org/10.1016/j.apenergy.2013.07.018. [2] Christofoletti CA, Escher JP, Correia JE, Marinho JFU, Fontanetti CS. Sugarcane vinasse: environmental implications of its use. Waste Manag 2013;33:2752–61. http://dx.doi.org/10.1016/j.wasman.2013.09.005. [3] Oliveira BG, Carvalho JLN, Chagas MF, Cerri CEP, Cerri CC, Feigl BJ. Methane emissions from sugarcane vinasse storage and transportation systems: comparison between open channels and tanks. Atmos Environ 2017;159:135–46. http://dx.doi. org/10.1016/j.atmosenv.2017.04.005. [4] Lukehurst C, Frost P, Seadi T Al. Utilisation of digestate from biogas plants as biofertiliser. Oxfordshire, United Kingdom: Report for the International Energy Agency, Task 37; 2010. [5] FNR. Guide to biogas: from production to use. Gülzow: Fachagentur Nachwachsende Rohstoffe e.V (FNR); 2010. [6] FNR. Biogas-Messprogramm II - 61 Biogas Anlage im Vergleich. 1st ed. Gülzow: Fachagentur Nachwachsende Rohstoffe e.V (FNR); 2009. [7] Janke L, Leite A, Nikolausz M, Schmidt T, Liebetrau J, Nelles M, et al. Biogas production from sugarcane waste: assessment on kinetic challenges for process designing. Int J Mol Sci 2015;16, doi:http://dx.doi.org/10.3390/ijms160920685. [8] Batstone DJ, Jensen PD. 4.17 - Anaerobic processes. Treatise Water Sci 2011:615–39, doi:http://dx.doi.org/10.1016/b978-0-444-53199-5.00097-x. [9] Janke L, Leite AF, Batista K, Silva W, Nikolausz M, Nelles M, et al. Enhancing biogas production from vinasse in sugarcane biorefineries: Effects of urea and trace elements supplementation on process performance and stability. Bioresour Technol 2016;217:10–20. http://dx.doi.org/10.1016/j.biortech.2016.01.110. [10] Verein Deutscher Ingenieure (VDI). VDI 4630 - Fermentation of organic materials: Characterisation of the substrate, sampling, collection of material data, fermentation tests. Düsseldorf: Beuth Verlag GmbH; 2016. [11] Brulé M, Oechsner H, Jungbluth T. Exponential model describing methane production kinetics in batch anaerobic digestion: a tool for evaluation of biochemical methane potential assays. Bioprocess Biosyst Eng 2014:1–12. http://dx.doi.org/10. 1007/s00449-014-1150-4. [12] Janke L, Weinrich S, Leite AF, Terzariol FK, Nikolausz M, Nelles M, et al. Improving anaerobic digestion of sugarcane straw for methane production: combined benefits of mechanical and sodium hydroxide pretreatment for process designing. Energy Convers Manag 2016;141:378–89. http://dx.doi.org/10.1016/j.enconman.2016. 09.083. [13] Weißach F, Strubelt C. Correcting the dry matter content of maize silages as a substrate for biogas production. Landtechnik 2008;63:82–3. [14] DIN. DIN EN 1504 - Determination of total content of carbon, hydrogen and nitrogen - Instrumental methods. Berlin: Beuth Verlag GmbH; 2011. [15] DIN. DIN EN 16170 - Determination of trace elements by inductively coupled

198

Energy Conversion and Management 168 (2018) 188–199

L. Janke et al. of hydrolysis and volatile fatty acids production from sugarcane filter cake: effects of urea supplementation and sodium hydroxide pretreatment. Bioresour Technol 2016;199:235–44. http://dx.doi.org/10.1016/j.biortech.2015.07.117. [39] de Barros VG, Duda RM, da Vantini JS, Omori WP, Ferro MIT, de Oliveira RA. Improved methane production from sugarcane vinasse with filter cake in thermophilic UASB reactors, with predominance of methanothermobacter and methanosarcina archaea and thermotogae bacteria. Bioresour Technol 2017;244:371–81. http://dx.doi.org/10.1016/j.biortech.2017.07.106. [40] Gallegos D, Wedwitschka H, Moeller L, Zehnsdorf A, Stinner W. Effect of particle

size reduction and ensiling fermentation on biogas formation and silage quality of wheat straw. Bioresour Technol 2017;245:216–24. http://dx.doi.org/10.1016/j. biortech.2017.08.137. [41] Zhao Y, Yu J, Liu J, Yang H, Gao L, Yuan X, et al. Material and microbial changes during corn stalk silage and their effects on methane fermentation. Bioresour Technol 2016;222:89–99. http://dx.doi.org/10.1016/j.biortech.2016.09.113. [42] Kreuger E, Nges I, Björnsson L. Ensiling of crops for biogas production: effects on methane yield and total solids determination. Biotechnol Biofuels 2011;4:44. http://dx.doi.org/10.1186/1754-6834-4-44.

199