Anaerobic digestion of wheat straw – Performance of continuous solid-state digestion

Bioresource Technology 146 (2013) 408–415

Contents lists available at ScienceDirect

Bioresource Technology journal homepage: www.elsevier.com/locate/biortech

Anaerobic digestion of wheat straw – Performance of continuous solid-state digestion Marcel Pohl ⇑, Kathrin Heeg, Jan Mumme Leibniz Institute for Agricultural Engineering Potsdam-Bornim e.V., Max-Eyth-Allee 100, 14469 Potsdam, Germany

h i g h l i g h t s  Thermophilic anaerobic digestion of wheat straw was investigated (55 and 60 °C).  Organic loading rate dependent behavior of the UASS system was analyzed.  The influence of the substrate’s chopping length on reactor performance was studied.  Hydrolysis kinetics from steady state operation has been determined.  Functional separation in the two-stage system for OLR P 8 gVS L

a r t i c l e

i n f o

Article history: Received 2 June 2013 Received in revised form 18 July 2013 Accepted 21 July 2013 Available online 27 July 2013 Keywords: Anaerobic digestion Solid-state Methane Wheat straw Organic loading rate

1

d1.

a b s t r a c t In this study the upflow anaerobic solid-state (UASS) reactor was operated at various conditions to optimize the process parameters for anaerobically digesting wheat straw in a continuous process. Additionally, particle size effects have been studied in the operation at 55 and 60 °C. Moreover, the incremental effect of the organic loading rate (OLR) to the system was examined from 2.5 to 8 gVS L1 d1. It was found that the UASS operating at 60 °C with a small OLR yields highest methane production, but the advantage over thermophilic operation is negligible. The rise in OLR reduces the systems yields, as expected. From OLR = 8 gVS L1 d1 a second stage is necessary to circumvent volatile fatty acids accumulation. Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction Renewable energies play substantial role to fight against the global warming and climate change, as they are capable of delivering energy in a CO2-neutral way. However, the production of photovoltaic and wind energy is strongly depending on meteorological conditions. Meanwhile, biomass delivers a steady energy output and the energy is therefore available for base-load power delivery. To circumvent the competition between energy and food production, the usage of waste materials for bioenergy production draws more attention. Wheat straw, for instance, is used for the production of bioethanol, biohydrogen and biogas. In a comparative study, the production of biogas from straw turned out to be the most energy efficient among those three (Kaparaju et al., 2009b) . Straw consists of cellulose and hemicellulose, linked to the anaerobically non-digestible lignin. Therefore, anaerobic digestion of straw usually requires long retention times and thus, delivers ⇑ Corresponding author. Tel.: +49 331 5699 921; fax: +49 331 5699 849. E-mail address: [email protected] (M. Pohl). 0960-8524/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.biortech.2013.07.101

comparably small methane yields with respect to time. Furthermore, straw addition increases the density of CSTR type reactors content and along the effort needed for mixing. To overcome these disadvantages of straw, a lot of effort is already placed into pretreatments for the improvement of bioavailability, even though it poses a major cost factor (Hendriks and Zeeman, 2009; Sapci, 2012; Nkemka and Murto, 2013). Straw’s poor biodegradability was shown in its applicability as a suitable biofilm carrier in two-phase anaerobic digestion processes (Andersson and Björnsson, 2002; Svensson et al., 2007), as its decomposition by bacterial enzymes is comparatively slow. Another disadvantage of lignocellulosic substrates for anaerobic digestion is their poor nutrients content, which is necessary for microbial growth. This has to be compensated by co-fermentation (Nkemka and Murto, 2013) with a supplementing co-substrate, or by chemical additives (Kridelbaugh et al., 2013). The UASS reactor, primarily presented by Mumme et al. (2010), was invented to provide high performance anaerobic digestion with little shear stress disturbing the biofilms developing in the solid-state bed and the anaerobic filter (AF), as it is known that

M. Pohl et al. / Bioresource Technology 146 (2013) 408–415

biofilm thickness decreases with the increase of shear stress (Paul et al., 2012). A two-stage system was used, as a spatial separation of hydrolysis and methanogenesis are known to improve AD performance. The UASS reactors, as well as the AF, were built without a stirrer, so the spatial vicinity of H2-producing and -consuming microorganisms can be promoted, which is thermodynamically beneficial (Kim et al., 2002). Using maize silage as feedstock, the UASS reactor was able to operate at OLRs of up to 17 gVS L1 d1 with methane yields of 312 L g1 VS (Mumme et al., 2010). The most important process parameter of continuous anaerobic digestion operation is the organic loading rate (OLR). Increasing the OLR will proportionally reduce the solids retention time (SRT) as well as the methane building potential of the substrate. In terms of efficiency, the optimization of OLR and methane yield needs to be identified. Moreover, high biogas yields also have the advantage of leaving fewer residues for subsequent disposal and thus allow smaller reactor sizes, which is economically more attractive (Tong et al., 1990). Previous experiments demonstrated the feasibility of the stable long-time anaerobic digestion of lignocellulosic substrate in the UASS system (Pohl et al., 2012). Those promising results encouraged to test the system for more operational parameters in order approach the optimal conditions. The emphasis of this study was to investigate the process behaviors and performances of the UASS system depending on the OLR with the aim to optimize process parameters. Additionally, as thermophilic anaerobic digestion in the UASS system showed higher methane yields than mesophilic fermentation (Pohl et al., 2012), an increase of temperature from 55 to 60 °C was tested for even further efficiency improvement. Moreover, the influence of the substrate’s particle size on the UASS reactor was also studied. 2. Methods 2.1. Substrate characteristics Wheat straw as a sole substrate was studied. After harvest, the straw was chopped on the field by a mobile chaff cutter (Ralle, Germany), afterwards milled in a straw mill (Himel, Germany) to its final average cutting length of 35 mm and removed from dust in a cyclone. Finally, the straw was pressed into bales. For comparative studies of the straw’s chopping length, it was further chopped into halves using a hammer mill. In the following, this variety is denoted as ‘‘short’’. The wheat straw used had a total solids (TS) content of 95.3%, 88.9% of volatile solids (VS) and a chemical oxygen demand (COD) of 1.19 g g1. Crude fiber content was determined to be 46.3% of the straw’s dry matter. The substrate contained 15.8 g kg1 of total nitrogen and 0.07 g kg1 of ammonium. Volatile fatty acids (VFAs) in the substrate, calculated as the sum of concentrations of C2- to C6-acids, were at 2.06 g kg1. For thermophilic anaerobic digestion, the wheat straw showed a maximum methane 1 yield of 304:29 L kgVS , which was determined according to the biochemical methane potential guideline VDI-Gesellschaft Energie und Umwelt (2006). 2.2. Reactor setup and operation The technical setup consisted of four identical systems, each of them containing an UASS reactor and an AF to prevent VFA accumulation. The UASS reactors had a working volume of 39 L, leaving 10 L of headspace. The AFs of 30 L were filled with PE biofilm carriers (‘‘Bioflow 40’’, RVT Process Equipment GmbH, Germany) with a surface of 305 m2 m3 for biofilm establishment. UASS reactors were built from stainless steel with an inspection window to con-

409

trol and measure the solid-state bed; AF reactors were built from transparent acrylic glass. The process liquid circulation was set to a flow rate of 2.6 L h1 using peristaltic pumps. The flowing regime for all reactors was from bottom to top. Both the UASS and AF reactors were heated by water jacked with LAUDA thermostats (Lauda, Lauda-Königshofen, Germany). To reduce heat loss, all process liquor lines and the AF reactors themselves were insulated. For inoculation, 3.5 kg of digestate from a previous experiment, digesting wheat straw as a sole substrate as well, were added to each UASS reactor. The inoculating digestate was characterized by a crude fiber content of 46.5%, volatile solids content of 107.7 gVS kg1, and a 1 pH of 9.1. The ammonium content was 0:56 g kgFM , the Kjeldahl1 nitrogen content was determined to be 1:45 g kgFM . With four UASS–AF systems available, each experiment was run in duplicates to improve statistical certainty. A detailed description and a scheme of the technical setup can be found elsewhere (Pohl et al., 2012). In contrast to prior experiments, where the stable long-time meso- and thermophilic operation at OLR = 2.5 gVS L1 d1 were shown (Pohl et al., 2012), several operational parameters have been modified for this study. The major alteration was the successive raise of the OLR throughout the experiment, which is following: 2.5, 3.5, 4.5, 6 and 8 gVS L1 d1 (equivalent to 3.1, 4.3, 5.6, 7.4 and 9.9 gCOD L1 d1). The subsequent OLRs ran for at least twice of the SRT to ensure steady state in terms of performance parameters, mainly indicated by constant biogas production. Earlier experiments also revealed that thermophilic anaerobic digestion of wheat straw at 55 °C in the UASS–AF system turned out to be more efficient than mesophilic operation (37 °C). Therefore, two of the four systems’ UASS reactors were set to 60 °C for 76 days. For better differentiability, operations at 60 °C will be referred to as ‘‘hyperthermophilic’’ throughout this manuscript. The connected AFs were perpetually operated at 55 °C, which is known to be a favorable temperature for methanogenic microorganisms. Additionally, it was previously shown that temperature effects on the acceleration of hydrolysis, which is known to be the rate limiting step in anaerobic digestion. The hypothesis of this study was, that an increase in temperature from 55 to 60 °C could increase the efficiency by widening the bottleneck of hydrolysis in the UASS reactor, while keeping the methanogens in the attached AF at their preferred temperature. After the experiments with hyperthermophilic temperatures in the UASSs at OLR = 2.5 and 3.5 gVS L1 d1, those systems were set back to thermophilic temperature and fed with wheat straw of a shorter chopping length. The reactors were fed once a day through a diagonal feeding tube to the bottom of the UASS reactor. Digestate removal was carried out once a week, but became necessary as much as twice for OLR P 4.5 gVS L1 d1 for the thermophilic systems with the unchopped straw and for OLR = 8 gVS L1 d1 for the chopped-straw systems. The higher frequency turned out to be necessary to reduce the height of the solid-state bed, which would have blocked the diagonal feeding tube otherwise. Clogging by decomposed substrate was not observed. The volume taken out during digestate removal was compensated with water addition. Temperature measurements at different heights and radial positions inside the solid-state bed revealed differences below 1 K, so a homogenous temperature distribution was assumed. Due to wheat straw’s poor concentrations of trace elements, they had to be supplemented along with the feeding. Following the recommendation of Abdoun and Weiland (2009), medium No. 144 of the ‘‘Germany collection of microorganisms and cell cultures’’ (Brunswick, Germany) was added to the process liquor on a daily basis. The dosage was adjusted according to the organic loading rate. Among other trace elements, medium No. 144 contains iron, calcium, copper, zinc and sodium.

410

M. Pohl et al. / Bioresource Technology 146 (2013) 408–415

2.3. Analytical methods Each of the reactors, UASSs as well as AFs, was equipped with a combined pH-temperature-probe (InPro4260, Mettler-Toledo, USA) for continuous online measurement. The biogas production rates were measured using a drum-type gas meter (Ritter, Germany). The biogas composition of each reactor was measured using an industrial biogas analyzer (SSM 6000, Pronova, Germany). As the biogas composition was only measured on working days, missing values for calculations, e.g. methane yields, were determined by linear interpolation of the surrounding measurements. Chemical analyses of the UASS’s and the AF’s effluents as well as of the digestates were carried out weekly. The samples were analyzed for electric conductivity, total solids, volatile solids, ammonia, nitrogen, Kjeldahl nitrogen, COD, volatile fatty acids (C2–C6) and trace elements. Volatile fatty acids were measured with a CP-3800 gas-phase chromatograph (Varian Inc., USA), equipped with a FFAP column (30 m  0.32 mm, film thickness 0.5 lm by Permabond, USA) and a flame ionization detector. Trace elements were measured using an iCAP 6000 Series ICP-MS (Thermo Fisher Scientific Inc., USA). C, N, S, and H fractions were analyzed with a vario EL III elemental analyzer (Elementar Analysensysteme GmbH, Germany), but have only been available for solid samples.

3. Results and discussion 3.1. Reactor start-up and operation Throughout the course of the experiments all four systems – eight reactors overall – ran without any significant technical disturbances or leakages for 158 days with feeding and three of those for another 36 days to determine the UASS’s constants of hydrolysis. Variation of temperatures and the straw’s chaff sizes had visible and measurable effects on the systems performance. However, not all parameters were influenced by the procedural alterations. Acetic acid showed a peak for all four UASS reactors at OLR = 8 gVS L1 d1, reaching 0.145 g L1 and 0.02 g L1 for the UASS reactors for unchopped as well as for chopped straw, respectively. Acetic acid concentration rose to 0.58 g L1 in average, which is still below the inhibiting concentration of 3 mg L1 (Borja et al., 1996). In the corresponding AFs process liquor, the acetic acid only raised up to 0.06 g L1 in average. Even at the highest OLRs, the remaining VFAs (C3–C6) stayed similar or below the detection limit of 0.02 g L1. As the system’s buffer capacities were sufficient, the increase of acetic acid was not reflected in the process liquors pH. It can be concluded that the UASS reactor is able to handle straw fermentation up to OLR = 6 gVS L1 d1 as a onestage system, whereas for higher OLRs a second stage for handling VFAs turned out to be necessary. Sugar, which is easily metabolized by the microorganisms in the system, was present in the sub1 strate at a concentration of 5:1 g kgVS , but was not detectable in the process liquor, or digestate. Trace element analysis revealed similar observations as in Pohl et al. (2012), that some trace elements tended to accumulate, whereas others were consumed in the process. In the process liquor Ca and K accumulated, whereas Cd, Co, Fe, Mn, and Ni, among others, stayed evenly concentrated. On the other hand, B, Cu, and Na were depleted over time. In the digestate samples analyzed, the distribution was different. B and K accumulated in the solidstate bed, whereas Co, Cr, Fe, Mn and Na were consumed. Others stayed on a nearly constant level, e.g. Al, Ca, Mg and P. Those observations could help to adapt the trace element solution to the UASS– AF system for the mono-fermentation of wheat straw. Especially the accumulating trace elements should be reduced to avoid the risk of reaching inhibitory concentrations. Overall, most trace ele-

ments were by one or more decimal powers higher concentrated in the digestate than in the process liquor, which stresses the filtering effect of the solid-state bed. As for mineral substances, Cl and P concentrations stayed steady throughout the experiments for all systems. Elemental carbon, hydrogen, nitrogen, and sulfur were unaffected by the technical alterations in the process, according to the elemental analysis. As expected, the digestate’s percentage of total solids (TS) increased with the progress of the experiments, as the time for degradation of the solids shortens at higher OLRs. As in the beginning TS of 11.92 ± 0.6% in average was measured, the final samples had an average TS of 14.2 ± 0.02%. This behavior is comparable for all process parameters, so neither the temperature nor the chaff size seemed to have an influence. Surprisingly, the volatile solids fraction of the digestate’s dry matter did not change with the SRT. It was expected that a longer retention time would improve the degradation of organic constituents. Analysis of the digestate’s acid detergent fiber fraction – the amount of cellulose and lignin – revealed that those structure building substances were not digested in the process, even not at high SRTs. The value was constant around 60% throughout the experiment. Whereas the fraction of raw fiber in the digestate – mainly consisting of easily degradable cellulose – steadily decreased as the OLR rose. For the TS concentration in the process liquors on the other hand the influence of the altered process parameters where visible. When the reactors were operated at 55 and 60 °C, a uniform increase in TS over the first 100 days from 0.68% to 1% was observed. However, switching to short straw led to TS fractions of 0.1% less in the process liquor. The volatile solids (VS) revealed the opposite behavior. For the digestate their fraction insignificantly decreased from 94.1% TS to 93.7% TS, while for the process liquor the VS content decreased from 53.5% TS down to 40.7% TS. Despite of the process liquor’s pH variation from 7.2 to 7.8 along with the OLR, the digestate’s pH values for all four UASS reactors were considerably stable in the course of the experiment, starting at 9.15 for the smallest OLR, moving steadily down to a pH of 9 at the highest OLR. Contrary to previous experiments, where no signs of nutrient deficiency were observed, this study was conducted without any addition of ammonium carbonate in the first place. As a result, a steady decline in the process liquor’s electrical conductivity and, in parallel, ammonium concentration was observed. As the C:N-ratio of the straw fed was measured to be 112:1 – recommended is 16:1 – 25:1 (Deublein and Steinhauser, 2010) – replenishment of nitrogen is short. Neither the OLR nor the substrate’s chaff size or the reactor’s temperature seemed to have a significant effect on the consumption of nitrogen in the process liquor. For all eight reactors of the four systems, the decline was nearly linear. Ammonia dissociates to ammonium with the increase of temperature and pH (Borja et al., 1996). However, it was found nearly constant throughout the experiment. The dissociation of the ammonium fraction was probably due to the establishment of biofilm in the UASS’s solidstate bed and on the AF’s biofilm carriers, as microorganisms consume ammonium during their growth. However, the major quantity of nitrogen was taken out from the UASS along with the digestate. In the course of the experiment UASS and AF reactor’s ammonia levels dropped from 690.1 and 707.8 mg L1, respectively, down below the detection limit of 2 mg L1. This suggests, that a long-time mono-fermentation of wheat straw needs an additional nitrogen source to function properly, as there are ammonium salts (Pohl et al., 2012; Kridelbaugh et al., 2013), urea or glutamine (Kridelbaugh et al., 2013). To allow a representative phase of decaying with active biomass, three reactor systems were supplemented with 20 g of

411

M. Pohl et al. / Bioresource Technology 146 (2013) 408–415 Table 1 Methane yields as averages from duplicates at different operational parameters. First values are for the combined UASS–AF systems, values in brackets give the methane yields of the UASS’s only. OLR (gVS L1 d1)

Solids retention time (d)

Thermophilic ðLCH4 g1 VS Þ

2.5 3.5 4.5 6 8

14–21

0.201 0.187 0.178 0.179 0.144

7–14 3–7

(0.179) (0.168) (0.143) (0.134) (0.066)

Chopped, thermophilic ðLCH4 g1 VS Þ 0.185 (0.144)

Hyperthermophilic ðLCH4 g1 VS Þ 0.207 (0.143) 0.191 (0.12)

0.175 (0.146) 0.174 (0.101)

ammonium carbonate each, as Gallert et al. (1998) reported that thermophilic cultures are most active at 1 gNH3 L1. 3.2. Influence of organic loading rate The most obvious procedural effect of the organic loading rate is on the SRT of the wheat straw in the UASS’ solid-state bed. In theory, they should be inversely proportional to each other. As the digestate was removed weekly, twice a week for higher OLRs, the SRT can only be given in time spans between digestate removals (Table 1). The restart of the experimental setup yielded more methane than the previous experiment at OLR = 2.5 gVS L1 d1. As previously 0:165 LCH4 g1 VS (Pohl et al., 2012) were reached, this later experiment had the advantage of starting with an adapted biocenosis. At identical procedural parameters, the thermophilic systems yielded 0:201 LCH4 g1 VS this time. As Table 1 shows, for all examined procedural conditions, the increase of the OLR results a decrease of methane yield, and therefore, efficiency of the anaerobic digestion process. This observation applied to the UASS reactors as well as to the combined systems (UASS + AF). With the increase of the OLR, a decrease in COD degradation was expected, as the solids retention time decreases proportionally. However, as Fig. 1 shows, COD degradation over OLR was neither linear nor continuous decreasing. When the whole systems

were taken into account, higher OLR shifts biogas production from the UASS to the AF. Nevertheless, raising the OLR from 2.5 to 6 gVS L1 d1 for the thermophilic systems decreased the SRT by more than half (60%), but COD degradation was only decreased by 12.8%. COD degradation was calculated for the combined system of UASS and AF, as the percolation was steady and too high to distinguish between the system’s reactors. For the OLR of 8 gVS L1 d1, and therefore most probably for higher ones as well, an accumulation of VFAs in the UASS reactors was observed (Fig. 2), which did not extend to the connected AFs. It can be concluded that for OLRs up to 8 gVS L1 d1 the UASS would have worked as a single-stage reactor as well and the AFs have not been necessary. However, for higher OLRs the separation of hydrolysis in the UASS and methanogenesis in the AF worked properly. As expected, the digestate’s CODs – reflecting the spend methane building potential – increased in the course of the experiments. This behavior was unaffected by temperature and chaff size, but reflects the shorter solids retention time at higher OLRs. Between OLR = 2.5 and 8 gVS L1 d1 the digestate’s COD rose from 0.15 to 0:18 g g1 TS . Thamsiriroj et al. (2012) reported that a shorter SRT lowers the methane fraction of the biogas produced. This behavior was also seen for the UASS reactors of the systems, whereas the biogas of the attached AFs showed higher methane fractions with increasing OLR (Table 2).

3.3. Influence of temperature The UASS–AF system, contrary to observations of Ahn and Forster (2000) and Mussoline et al. (2013), yielded as much as 36% more methane at thermophilic (55 °C) compared to mesophilic (37 °C) operation in a previous experiment. So, a further increase of temperature might improve the system’s performance. The increase of temperature from 55 °C to 60 °C reduced COD degradation at OLRs of 2.5 and 3.5 gVS L1 d1 from 44.4% to 44.1% and from 41.9% to 40%, respectively (Fig. 1). So the gain in performance seen at a temperature raise from 37 to 55 °C (Pohl et al., 2012) did not extend to higher temperature levels. A difference in the time needed for the start-up of the reactors – as reported to be faster for thermo – than for mesophilic anaerobic digestion (Li et al.,

50 55 °C 60 °C 55 °C, short

COD degradation [%]

40

30

20

10

0 2.5

3.5

4.5

6 -1

8

-1

OLR [g VS L d ] Fig. 1. COD degradation of the UASS–AF systems under varying operating conditions.

412

M. Pohl et al. / Bioresource Technology 146 (2013) 408–415

1.2

-1

VFA [g L ]

UASS (60 °C) AF (60 °C) UASS (short) AF (short)

UASS (55 °C) AF (55 °C)

1.0 0.8 0.6 0.4 0.2

0.24

UASS + AF (55 °C)

0.22 0.20 0.18 0.16 0.14 UASS + AF (60 °C) UASS + AF (short)

0.12 0.10 8.0

T

60

OLR

T [°C]

-1

-1

OLR [gVS L d ]

-1

methane yield [LCH4 gVS ]

0.0

6.0 4.5 3.5 2.5

55 0

20

40

60

80

100 120 140 160 0

days of operation

20

40

60

80

100 120 140 160

days of operation

Fig. 2. Process description. Strictly thermophilic systems on the left, systems with modified temperature and chopping length on the right. Methane yields are given as averages over 3 days.

Table 2 Methane fractions in the biogas for the UASS and AF reactors. Averages from duplicates, except  from single reactor system. OLR (gVS L1 d1)

0 (decay) 2.5 3.5 4.5 6 8

Thermophilic CH4,UASS (%)

CH4,AF (%)

62.6⁄ 56.4 55.5 52.5 52.1 45.5

75.4⁄ 65.6 63.9 67.9 72.2 73.7

Chopped, thermophilic

Hyperthermophilic

CH4,UASS (%)

CH4,AF (%)

CH4,UASS (%)

CH4,AF (%)

53.8

68.9

51.8 48.7

69.8 73.2

54.3 49.3

65.0 72.4

2011) – was not observed. Furthermore, adaptation to a new OLR under all process conditions was conducted within a week, judging from the stabilization of the UASS–AF system’s methane production. For hyperthermophilic operation, the biogas production rates of the UASS reactors were lower compared to their thermophilic counterparts. However, the biogas production rates of the AFs connected to the hyperthermophilic UASSs increased, meaning a shift in biogas production from UASS to AF, from hyperthermophilic to thermophilic milieu, took place. Hyperthermophilic UASS reactors showed a smaller methane fraction in the biogas formed, but regenerated after setting the temperature back to thermophilic. Therefore, the AFs of the hyperthermophilic UASSs had higher methane contents in the biogas

than those of the strictly thermophilic systems. This indicates that methanogens were suppressed at hyperthermophilic temperature, but were capable of quickly recovering as the temperature was reversed to thermophilic conditions (Fig. 3). Details on the methane content of the biogas under varying process conditions can be taken from Table 2, the remainder to 100% being mainly CO2. Other gases were analyzed, as there were H2 and H2S, and found that these were in the range of two- to three-digit ppm. As El-Mashad et al. (2004) showed, a rise in temperature from 50 to 60 °C does accelerate hydrolysis, but severely decelerates methanogenesis. This is in accordance with the findings described here and explains the shift of the methane production rate to the AFs. Nevertheless, as operating costs for hyperthermophilic operation are higher than for that of thermophilic operation and the productivity is comparable, thermophilic temperature is favorable. Methane yields for the overall systems consisting of UASS and AF were slightly better at hyperthermophilic operation, most likely due to higher enzymatic activity at higher temperature (Kim et al., 2012), which overweighs the inhibition of methanogens. As formulated by Chen and Hashimoto (1978), the maximum specific microbial growth rate (d1) is linearly dependent on the temperature. Following their model (Eq. (1)), the rise in temperature from 55 to 60 °C would increase the growth rate from 0.586 to 0.651 d1, a rise of 11%:

lm ¼ 0:013  T  0:129

ð1Þ

Unfortunately, complex systems like the biocenosis in an anaerobic digestion reactor do not always act linear. Clarens and Moletta (1990) found the optimal growth conditions for methane

413

M. Pohl et al. / Bioresource Technology 146 (2013) 408–415

UASSs to 60 °C

90

UASSs to 55 °C

methane fraction [%]

80

70

60

50 UASS 1 UASS 2 AF 1 AF2

40

30 0

20

40

60

80

100

120

140

160

days of operation Fig. 3. Methane contents of the biogas as developing in the course of the temperature change in the UASS reactors.

producing Methanosarcina sp. to be at the temperature of 55 °C, favoring thermophilic operation from the biological point of view. Additionally, higher temperatures move the equilibrium of ammonia and ammonium towards the cytotoxic ammonia.

chopping lengths were marginal; the influence of the OLR was more distinctive (Table 2).

3.4. Influence of particle size

The constant of hydrolysis (kH) was calculated from the cumulative methane production during the phase of decay using first order kinetics. Curve fitting was conducted using a non-linear least squares routine. As one thermophilic reactor system was emptied out before the decay phase for chemical analyses, there was only one value for kH for unchopped straw. Though, for the experiments with shorter chopped straw, duplicates were available. For the unchopped straw a constant of hydrolysis of 0.249 d1 was calculated, whereas the kH was 0.299 d1 for chopped straw. This reflects that the organic contents of the shorter chopped straw were more available for microorganisms due to the substrate’s higher surface. These values were significantly higher than the constants of hydrolysis determined for the same system with the same substrate decaying from OLR = 2.5 gVS L1 d1, where for the thermophilic reactors kH was at 0.138 d1 using first-order kinetics. The higher values in this later experiment are explained by the reactor’s history, which had a positive effect on the biocenosis. Due to the higher organic loading rate immediately before decaying, the UASS’s microorganisms were specialized on hydrolysis rather than on methanization, which was mainly carried out in the AFs at that time. Nevertheless, the values for kH found here are in accordance with findings of Veeken and Hamelers (1999), who found a linear Arrhenius-like relationship for the constant of hydrolysis and temperature. Their values for temperatures between 20 and 40 °C extrapolated to 55 °C would result in kH = 0.23 d1. On the other hand, Thamsiriroj et al. (2012) found a proportionality between SRT and kH, which would result in a higher kH for the experiment decaying from OLR = 2.5 gVS L1 d1. Additionally, hydrolysis is reported to slow down with falling pH (Veeken et al., 2000), which was noticeably lower when starting the phase of decay in the current experiment. On the hypothesis, hydrolysis is mainly taking place in the solid-state bed – as effluent characteristics of coupled UASS and AF reactors did not noticeably differ – the difference in kH can be explained by the solid-state

Particle size has an effect on the results of chemical analysis of straw (Sapci, 2012). Larger chops show higher contents in hemicellulose, cellulose and lignin, but lower ash contents. On the procedural side, the shorter particle size of the chopped straw prolongs their solids-retention time, as digestate removal was not gravimetrically controlled, but by the height of the solidstate bed, which was intended to stay constant. But an effect on methane production by the advantageous retention time was not observed. Contrary to expectations, at OLR = 3.5 gVS L1 d1 the COD degradation was smaller for the chopped straw, 40.4% compared to 41.9% (Fig. 1). This stays in conflict with the assumption, that chopping enlarges the surface area of the substrate and therefore the area available for enzymatic degradation. For OLR = 6 gVS L1 d1 on the other hand, COD degradation efficiency of unchopped and chopped straw were found similar (38.7% and 38.8%, respectively). The degradation efficiency proportions for OLR = 3.5 gVS L1 d1 were then turned around at OLR = 8 gVS L1 d1, where the chopped straw was degraded by 35.9%, whereas the unchopped straw was only digested by 28.8%. As Table 1 demonstrates, the particle size also had an effect on the methane yields of the UASS–AF system, in general backing the observation on COD degradation. For OLRs 3.5 and 6 gVS L1 d1 the unchopped straw yielded more methane per gVS, whereas for OLR = 8 gVS L1 d1 the efficiency was turned over to the benefit of the reactors fed with the chopped straw. Interestingly, the UASS’ methanization stayed more active for the chopped straw. 58% of the system’s methane was produced by the UASS with chopped straw, whereas for the unchopped straw the UASS only delivered 45% of the overall output. For smaller OLRs the UASS’ share of the system’s methane yield was around 80%. The differences in the methane fractions of the biogas produced for the different

3.5. Hydrolysis

414

M. Pohl et al. / Bioresource Technology 146 (2013) 408–415

beds pH, which was lower in the previous experiment, 8.9 compared to 9.2. For the dynamic periods of the experiment, the kH was calculated from the UASS’s COD balance. Taking into account, that hydrolysis is the rate limiting step in anaerobic digestion (Donoso-Bravo et al., 2009), the UASS reactors methane rate dynamics – converted into the methane’s COD equivalent – was balanced against the COD added by feeding (CODSubstrate) and extracted during digestate removal (CODDigestate), as shown in Eq. (2). CODSSB represented the intrinsic COD of the UASS reactor’s solid-state bed and CODDeg the degradable fraction of the solid-state bed:

CODDeg ¼ CODSSB  ð1  kH Þ þ CODSubstrate  CODDigestate

ð2Þ

The calculated CODDeg was used to calculate a methane equivalent, which was fitted against the methane actually produced using a non-linear least squares routine in order to obtain kH. The factor used for converting 1 LCH4 into 1 gCOD was 2.86. Fig. 4 shows the computed results for kH depending on the experimental variations. Generally, an increase in OLR involved a decrease in kH. This is probably due to the fact that with the higher SRT at low OLRs, the time for hydrolytic bacteria was sufficient enough to hydrolyze easily degradables as well as the harder degradable compounds of the substrate. For higher OLRs the time for hydrolyzing hardly degradable material was missing then. This assumption is supported by the observations described in Section 3.1 that with higher OLRs the systems TS contents increased, while the VS contents decreased. Furthermore, the retention time of microorganisms in the reactor was shorter for higher OLRs, assuming that their majority was located in the solid-state bed. This took the biocenosis higher efforts on reproduction, rather than the production on hydrolytic enzymes. Surprisingly, although methane yields were slightly higher, kH was lower at 60 °C than at 55 °C, which leads to the assumption that either the hydrolytic microorganisms or their enzymes were inhibited by the temperature increase. For the shorter chopped straw the kH values were lower as for the untreated straw, probably due to the better availability of organic compounds of the substrate, which made hydrolysis less time demanding. In contrast to the values derived from the reactor’s phases of decay, the constants of hydrolysis calculated from the dynamic process are in the same order of magnitude as previous results (Pohl et al., 2012).

3.6. Performance comparison Unfortunately, drawing a direct comparison of the UASS system with known reactor types – such as the CSTR or the UASB – is difficult, as the yield for biogas and methane are strongly dependent on the substrate. Especially for straw the range of possible yield is huge, as it is among others influenced by climate conditions and the vegetation state at time of harvest (Amon et al., 2007). For two different kinds of wheat straw, Tong et al. (1990) found a maximum CH4 conversion efficiency of 70% and 78% in a biochemical methane potential test, where the calculatory basis is comparable to the COD degradation shown in Fig. 1. Also in batch operation, Brown et al. (2012) report no significant difference in methane yields from wheat straw at liquid or solid-state anaerobic digestion. However, they were able to observe a 2- to 7-times higher volumetric productivity in the solid-state system. In further literature values for hydrolysates of straw are known (e.g. Kaparaju et al., 2009a), but not for straw in a monofermentation. Nkemka and Murto (2013) anaerobically digested wheat straw hydrolysate in an UASB type reactor, reach1 1 ing methane yields of 0:29 L g1 d , COD at an OLR of 2.5 gCOD L whereas the UASS system used in this study yielded 1 1 0:206 L g1 d , which is less but spares COD at an OLR of 3.1 gCOD L the pretreatment. Ghosh and Bhattacharyya (1999) report a COD degradation of 26% after 21 days of fermentation in a batch system at 30 °C, which is exceeded by far by the UASS system, even at higher OLRs, which comes with a shorter SRT (Fig. 1). For vegetable wastes, a COD degradation of <50% is reported for an OLR of 2.75 gVS L1 d1 (Babaee and Shayegan, 2011), with the constraint that food wastes are better digestible than lignocellulosic substrates. Deublein and Steinhauser (2010) for instance, reported that the maximum biogas yield of straw is obtained after 15–20 days, but without referring to the circumstances of the underlying experiments. Fernandes et al. (2009) determined COD degradation from straw of 50% in a CSTR after a hydraulic retention time (HRT) of 90 days, degrading less than 40% at a HRT of 21 days. For a mixture of maize silage and straw Mumme et al. (2010) measured a COD degradation of 86–93% in a thermophilic operated UASS system. This lack of data points out, that it’s difficult to digest straw in a CSTR type reactor as it raises the process liquor’s viscosity and

0.10 55 °C 60 °C 55 °C, short

0.08

-1

kH [d ]

0.06

0.04

0.02

0.00 2.5

3.5

4.5

6 -1

8

-1

OLR [g VS L d ] Fig. 4. Constant of hydrolysis in dependence of the parameters altered.

M. Pohl et al. / Bioresource Technology 146 (2013) 408–415

promotes the establishment of floating covers. Therefore, straw is usually used in small concentrations in co-digestion with slurry or manure. Additionally, due to the missing efficient reactor types for digesting straw continuously, most studies compare potentials in batch assays, which are difficult to compare to the system described here due to their theoretically infinite solids retention time. 4. Conclusions Despite of the disadvantages for anaerobic digestion, this study showed that it is feasible to ferment straw in the UASS system. The preferable operational parameters for digesting straw in the UASS system might be thermophilic temperatures, as the higher yields at 60 °C hardly compensate the higher effort for heating and isolation. Up to an OLR of 6 gVS L1 d1, the UASS was able to handle the mono-fermentation of lignocellulosic waste as a single-stage system, beyond a stabilizing AF turns out to be necessary. The substrate’s particle size was found to have little or no effect on the efficiency. Acknowledgements This work is funded by the German Federal Ministry of Education and Research (BMBF) in cooperation with Project Agency Jülich (PtJ). The authors would like to thank S. Engnath and T. Zenke for setting up and operating the lab scale experiments. Furthermore we would like to thank E. Janiszewski, L. Herklotz and C. Prautsch for conducting lab analyses. References Abdoun, E., Weiland, P., 2009. Optimization of monofermentation from renewable raw materials by the addition of trace elements. Bornimer Agrartechnische Berichte 68, 69–78. Ahn, J.-H., Forster, C.F., 2000. A comparison of mesophilic and thermophilic anaerobic upflow filters. Bioresour. Technol. 73, 201–205. Amon, T., Amon, B., Kryvoruchko, V., Machmüller, A., Hopfner-Sixt, K., Bodiroza, V., Hrbek, R., Friedel, J., Pötsch, E., Wagentristl, H., Schreiner, M., Zollitsch, W., 2007. Methane production through anaerobic digestion of various energy crops grown in sustainable crop rotations. Bioresour. Technol. 98, 3204–3212. Andersson, J., Björnsson, L., 2002. Evaluation of straw as a biofilm carrier in the methanogenic stage of two-stage anaerobic digestion of crop residues. Bioresour. Technol. 85, 51–56. Babaee, A., Shayegan, J., 2011. Effect of organic loading rates (OLR) on production of methane from anaerobic digestion of vegetables waste. In: Presented at the World Renewable Energy Congress, Linköping, Sweden. Borja, R., Sánchez, E., Weiland, P., 1996. Influence of ammonia concentration on thermophilic anaerobic digestion of cattle manure in upflow anaerobic sludge blanket (UASB) reactors. Process Biochem. 31, 477–483. Brown, D., Shi, J., Li, Y., 2012. Comparison of solid-state to liquid anaerobic digestion of lignocellulosic feedstocks for biogas production. Bioresour. Technol. 124, 379–386. Chen, Y.R., Hashimoto, A.G., 1978. Kinetics of methane fermentation. Biotechnol. Bioeng. Symp. (United States) 8. Clarens, M., Moletta, R., 1990. Kinetic studies of acetate fermentation by Methanosarcina sp. MSTA-1. Appl. Microbiol. Biotechnol. 33, 239–244.

415

Deublein, D., Steinhauser, A., 2010. Biogas from Waste and Renewable Resources: An Introduction. Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim. Donoso-Bravo, A., Retamal, C., Carballa, M., Ruiz-Filippi, G., Chamy, R., 2009. Influence of temperature on the hydrolysis, acidogenesis and methanogenesis in mesophilic anaerobic digestion: parameter identification and modeling application. Water Sci. Technol. 60, 9. El-Mashad, H.M., Zeeman, G., van Loon, W.K.P., Bot, G.P.A., Lettinga, G., 2004. Effect of temperature and temperature fluctuation on thermophilic anaerobic digestion of cattle manure. Bioresour. Technol. 95, 191–201. Fernandes, T.V., Klaasse Bos, G.J., Zeeman, G., Sanders, J.P.M., van Lier, J.B., 2009. Effects of thermo-chemical pre-treatment on anaerobic biodegradability and hydrolysis of lignocellulosic biomass. Bioresour. Technol. 100, 2575–2579. Gallert, C., Bauer, S., Winter, J., 1998. Effect of ammonia on the anaerobic degradation of protein by a mesophilic and thermophilic biowaste population. Appl. Microbiol. Biotechnol. 50, 495–501. Ghosh, A., Bhattacharyya, B.C., 1999. Biomethanation of white rotted and brown rotted rice straw. Bioprocess Eng. 20, 297–302. Hendriks, A.T.W.M., Zeeman, G., 2009. Pretreatments to enhance the digestibility of lignocellulosic biomass. Bioresour. Technol. 100, 10–18. Kaparaju, P., Serrano, M., Angelidaki, I., 2009a. Effect of reactor configuration on biogas production from wheat straw hydrolysate. Bioresour. Technol. 100, 6317–6323. Kaparaju, P., Serrano, M., Thomsen, A.B., Kongjan, P., Angelidaki, I., 2009b. Bioethanol, biohydrogen and biogas production from wheat straw in a biorefinery concept. Bioresour. Technol. 100, 2562–2568. Kim, H.-W., Nam, J.-Y., Kang, S.-T., Kim, D.-H., Jung, K.-W., Shin, H.-S., 2012. Hydrolytic activities of extracellular enzymes in thermophilic and mesophilic anaerobic sequencing-batch reactors treating organic fractions of municipal solid wastes. Bioresour. Technol. 110, 130–134. Kim, M., Ahn, Y.-H., Speece, R.E., 2002. Comparative process stability and efficiency of anaerobic digestion; mesophilic vs. thermophilic. Water Res. 36, 4369–4385. Kridelbaugh, D.M., Nelson, J., Engle, N.L., Tschaplinski, T.J., Graham, D.E., 2013. Nitrogen and sulfur requirements for Clostridium thermocellum and Caldicellulosiruptor bescii on cellulosic substrates in minimal nutrient media. Bioresour. Technol. 130, 125–135. Li, Y., Park, S.Y., Zhu, J., 2011. Solid-state anaerobic digestion for methane production from organic waste. Renewable Sustainable Energy Rev. 15, 821– 826. Mumme, J., Linke, B., Tölle, R., 2010. Novel upflow anaerobic solid-state (UASS) reactor. Bioresour. Technol. 101, 592–599. Mussoline, W., Esposito, G., Giordano, A., Lens, P., 2013. The anaerobic digestion of rice straw: a review. Crit. Rev. Environ. Sci. Technol. 43, 895–915. Nkemka, V.N., Murto, M., 2013. Biogas production from wheat straw in batch and UASB reactors: the roles of pretreatment and seaweed hydrolysate as a cosubstrate. Bioresour. Technol. 128, 164–172. Paul, E., Ochoa, J.C., Pechaud, Y., Liu, Y., Liné, A., 2012. Effect of shear stress and growth conditions on detachment and physical properties of biofilms. Water Res. 46, 5499–5508. Pohl, M., Mumme, J., Heeg, K., Nettmann, E., 2012. Thermo- and mesophilic anaerobic digestion of wheat straw by the upflow anaerobic solid-state (UASS) process. Bioresour. Technol. 124, 321–327. Sapci, Z., 2012. The effect of microwave pretreatment on biogas production from agricultural straws. Bioresour. Technol. 128C, 487–494. Svensson, L.M., Björnsson, L., Mattiasson, B., 2007. Enhancing performance in anaerobic high-solids stratified bed digesters by straw bed implementation. Bioresour. Technol. 98, 46–52. Thamsiriroj, T., Nizami, A.S., Murphy, J.D., 2012. Use of modeling to aid design of a two-phase grass digestion system. Bioresour. Technol. 110, 379–389. Tong, X., Smith, L.H., McCarty, P.L., 1990. Methane fermentation of selected lignocellulosic materials. Biomass 21, 239–255. VDI-Gesellschaft Energie und Umwelt (Ed.), 2006. Fermentation of Organic Materials – Characterisation of the Substrate, Sampling, Collection of Material Data, Fermentation Tests. Beuth Verlag, Berlin. Veeken, A., Hamelers, B., 1999. Effect of temperature on hydrolysis rates of selected biowaste components. Bioresour. Technol. 69, 249–254. Veeken, A., Kalyuzhnyi, S., Scharff, H., Hamelers, B., 2000. Effect of pH and VFA on hydrolysis of organic solid waste. J. Environ. Eng. 126, 1076–1081.