The impact of inoculum source, inoculum to substrate ratio and sample preservation on methane potential from different substrates

Biomass and Bioenergy 83 (2015) 474e482

Contents lists available at ScienceDirect

Biomass and Bioenergy journal homepage: http://www.elsevier.com/locate/biombioe

Research paper

The impact of inoculum source, inoculum to substrate ratio and sample preservation on methane potential from different substrates Veronica Moset*, Nawras Al-zohairi, Henrik B. Møller Aarhus University, Department of Engineering, Blichers All
e 20, DK 8830, Tjele, Denmark

a r t i c l e i n f o

a b s t r a c t

Article history: Received 20 June 2014 Received in revised form 16 June 2015 Accepted 16 October 2015 Available online 9 November 2015

Batch experiments were conducted to evaluate the impact of inoculum sources, inoculum to substrate (IS) ratio and storage conditions on the potential and production rate of methane (CH4) from different substrates: wheat straw, whole crop maize, cattle manure, grass and cellulose. The results of the test with four inocula and four substrates indicated that inoculum source could have a significant impact on both CH4 potential (BMP) and the kinetics parameters of different substrates. The two inocula showing the highest BMP and production rates in each period were those coming from a feeding with more than 70% of animal manure under thermophilic conditions. The impact of the IS ratio in the range 0.25e2.5, in terms of g volatile solids (VS) substrate/g VS inoculum, depended on substrate type. Maize silage was more affected to changes in the IS ratio than wheat straw. The optimal IS ratio range for maize was 1.0e1.5, however, a wider IS range can be used in wheat straw (0.5e2.5). The impact of freezing and drying depended on biomass type. Freezing, drying and ensiling of grass increased the CH4 yield compared to fresh grass. Drying of maize had no impact while freezing reduced the CH4 potential. Drying and freezing had no impact on straw. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Ultimate methane potential Methane production rate Inoculum source Inoculum/substrate ratio Sample preservation

1. Introduction The total global energy consumption is expected to grow by one third during the next 25 years which will increase the pressure on fossil energy and the requirement for a higher contribution from renewable energy. There is an increasing interest in renewable energy, such as anaerobic digestion, due to its economic and environmental benefits [1,2]. Ultimate methane (CH4) potential (B0), which is the maximum CH4 yield that can be obtained from a substrate, is a widely used parameter for correct design and budgeting of full-scale biogas plants as well as for assessment of the influence of biogas plants on greenhouse gas (GHG) emissions. In addition, B0 is a key parameter when calculating CH4 emission from animal slurry according with the IPCC [3] procedure. However important differences are shown in literature among B0 when compared within the same substrate, probably because differences in methodologies. Owen et al. [4], Angelidaki et al. [5], and ISO-11734 1995(E) [6] recommended different procedures for B0 determination but no

* Corresponding author. E-mail address: [email protected] (V. Moset). http://dx.doi.org/10.1016/j.biombioe.2015.10.018 0961-9534/© 2015 Elsevier Ltd. All rights reserved.

procedure has been defined as a standard method [7,8]. Therefore, a precise determination of the B0, and how this value can be affected according to different methodologies used during the determination in the laboratory is essential. In order to choose a procedure, it is important to know the effect of each parameter on B0 determination. On this regard several works have been publish evaluating the effect of specific parameters on B0 determination [7,8]. From these works it is concluded that one of the most important parameters affecting B0 is the inoculum, not only the inoculum source, but also the amount of the inoculum added, what is called inoculum to substrate ratio (IS ratio). It is clearly shown that IS ratio can affect not only the biodegradability but also the CH4 production rate or hydrolysis rate, calculated from first-order kinetics models [7]. However, it is not clear how this factor affects the results, and if this depends on substrate composition. Sample preservation before the B0 test is also another factor affecting biodegradability and thus B0 results [9,10]. However, scarce information about this issue can be found in literature. In this work we aimed to study the effect of three of the major points of controversy among B0 methodologies: inoculum source, IS ratio and sample preservation, by means three anaerobic batch experiments. In each batch experiment, one of the three factors was evaluated by using different substrates in a factorial design, in order

V. Moset et al. / Biomass and Bioenergy 83 (2015) 474e482

to determine whether substrate composition could influence the results. Inocula from four biogas plants in Denmark working with different temperatures and substrates were used to investigate the impact of inoculum source. The impact of the IS ratio on volatile solids (VS) basis was evaluated in the range 0.25e2.5 VS of substrate/g VS inoculum. The storage methods used to test the effect of sample preservation were drying, freezing and ensiling. The substrates used were wheat straw, whole crop maize, cattle manure, grass and cellulose.

2. Materials and methods 2.1. Substrates and inocula Table 1 shows the origin, sample preservation conditions and characteristics of the substrates used. Whole crop maize (Artist) was harvested and chopped at Research Centre Foulum (Aarhus University, Denmark) in October 2011. It was ensiled for one year in a plan silo in Foulum (Denmark). Fresh whole crop maize (Adept) was harvested and chopped at Research Centre Foulum (Aarhus University, Denmark) in October 2012. This maize was stored in a refrigerator at 5
C for a couple of days without additives before use. Wheat straw was harvested in eastern Jutland, Denmark, in the summer of 2011. The straw was homogenized by hammer-milling (Cormall HDH 770, Sønderborg, Denmark). The size of the straw particles was in the range 0.5e3 cm. The baled wheat straw was stored in a bale at room temperature. The grass sample contained 17% clover and 83% ryegrass and was harvested in October 2012 at Research Centre Foulum (Aarhus University, Denmark). The grass was chopped to particle size 0.5e10 cm with a machine from Landtechnik Weihenstephan (Versuchshacksler no. 008, Austria). Cattle manure (Holstein) was collected from the livestock building at Research Centre Foulum (Aarhus University, Denmark). The cows were fed 9.76% wheat, 65.85% late grass silage and 24.39% rapeseed meal. Cattle manure was stored at 20
C and before use it was thawed at room temperature. The total VFA was 2.5 ± 0.3 g/L. The cellulose was powdered cotton linters with a particle size of 50 mm (Sigma Aldrich, USA). The inocula were collected from four different biogas plants. All inocula were collected on the same day, and pre-incubated for 15 days at their corresponding initial temperatures in order to deplete the residual biodegradable organic material (degasification), according to the recommendation of Angelidaki et al. [5]. After degasification, they were filtered with a 1-mm screen. Table 2

475

shows the origin, feed composition storage conditions and chemical composition of the inocula used. 2.2. Experiments 2.2.1. Impact of inoculum source (experiment 1) This experiment was conducted in a factorial design in which four inocula (Horsens, Bånlev, Foulum and Thorsø) and four substrates (wheat straw, maize silage, cattle manure and cellulose) were used. The dry matter (DM) dose of substrate to each bottle was fixed (4 ± 0.1 g DM per bottle), resulting in an IS ratio in the range 0.79 (Bånlev) 1.21 (Foulum). 2.2.2. Impact of IS ratio (experiment 2) This experiment was conducted in a factorial design in which six IS ratios (0.25, 0.5, 1, 1.5, 2 and 2.5) were tested with two different substrates (wheat straw and fresh whole crop maize) using inoculum from Foulum whose characteristics are shown in Table 2. The IS range was chosen according to literature [7]. The two substrates tested in this experiment were selected based on the fact that wheat straw and maize are products commonly added to anaerobic digestion plants in Denmark and also on the results from experiment 1 by using inoculum from Foulum. 2.2.3. Experiment 3: impact of sample preservation This experiment was conducted in a factorial design in which three different storing methods were used: Drying, freezing and vacuum ensiling. These methods were selected based on normal procedure in real plants and laboratories. Storing biomass samples in refrigerator and freezer and drying are commonly used methods in laboratories, before a B0 analysis; while ensiling is widely used for storing of crops for biogas production. Three substrates were used: fresh whole crop maize, wheat straw and grass. These substrates were selected because they are commonly used as cosubstrates in anaerobic digestion plants. Drying was carried at 50 ± 1
C for about 69 h. Freezing consisted of a storage period of 6 day at 18 ± 1
C in plastic bags. Vacuum ensiling was only used on grass, using the Webomatic CT100 model I22 vacuum machine (Genpack A/S, Denmark). The fresh chopped grass was ensiled in a plastic bag under vacuum for 33 days at 5 ± 0.5
C. The batch experiment was carried at thermophilic conditions (53 ± 1
C) for 91 days using an inoculum from Foulum. In this experiment the IS ratio in terms of gVS/gVS was 1 ± 0.02.

Table 1 Origin, sample preservation conditions and characteristics in terms of dry matter (DM), volatile solids (VS) and pH and experimental distribution of the substrates used.

Substrate

Origen

Sample preservation

DM %

Maize

Fresh Frozen Dry Ensiled Baled Frozen Dry Fresh Frozen Dry Ensiled Frozen Dry

30.7 30.5 96.1 35.8 90.1 92.2 96.9 19.8 20.3 95.4 20.3 11.33

Wheat straw

Grass

Cattle manure Cellulose ┼

% percentage on wet weight basis.

┼

VS % 28.8 29.5 93.0 32.4 85.8 88.0 92.4 17.1 17.5 82.5 17.4 9.96

pH

6.84

Experiment 23 3 3 1 123 3 3 3 3 3 3 1 1

476

V. Moset et al. / Biomass and Bioenergy 83 (2015) 474e482

Table 2 Origin, substrate process temperature and chemical composition of the inocula used. Horsens Animal manure Industrial waste Maize and grass Sludge municipal WWTP Temperature Dry matter Volatile solids pH

┼

% ┼ % ┼ %


C ┼ % ┼ %

100 35 2.8 1.6 7.97

Bånlev

Foulum

Thorsø

75 25

80 20

75 23 2

53 3.4 2.2 7.98

53 3.1 2.0 8.40

35 2.7 1.5 8.35

┼

% percentage on wet weight basis.

2.3. Analytical methods and batch experiment

BMPt ¼ B0 exp {exp [mm e/B0 (l  t) þ 1]}

The analyses were carried out according to APHA [11] for DM (105
C) and VS (550
C). Knick Type 911(Germany) was used to measure the pH. The methodology followed for determination of B0 was carried out according to the procedure described by Møller et al. [12]. The test was conducted in 500-ml infusion bottles. The bottles were closed with butyl rubber stoppers, sealed with aluminum crimps, flushed with N2 for 2 min and incubated for more than 90 days at 35
C or 53
C depending on the original temperature. The volume of biogas produced was measured by the water displacement method where the pH in the water bath was below 4. Gas samples were collected by connecting the test bottles to the sample bottles using a plastic tube attached to a needle. Biogas samples were analyzed for O2, CO2 and CH4 concentration with an Agilent 7890A gas chromatograph with GC sampler 80 (Agilent Technologies, USA) equipped with a thermal conductivity detector and an Alltech CTR 1 double column. The injector, detector and oven temperatures were 110, 150 and 40
C, respectively. Helium was the carrier gas. The bio-methane potential (BMP) at 46 days (42 days in experiment 1), 60 days (62 days in experiment 1), and 90 days were expressed as the cumulative production (L) per gram of VS from the substrate introduced to the bottles. The average of BMP at each period was calculated at standard temperature and pressure (LSTP) (273 K and 1 atm.), which is the sum of the STP liter yield of CH4 after subtraction of the CH4 yield from adding only inoculum, per kg VS added (Eq. (1)).

where BMPt denotes the cumulative CH4 yield (L CH4 kgVS1) at time (t) expressed in days; B0 is the ultimate CH4 yield (L CH4 kgVS1); mm: maximum CH4 production rate expressed in reciprocal of time [day1], e: Euler's number, ʎ: duration of the lag phase [days]. Root mean squared error (RMSE) was used to calculate the k, mm, ʎ values. Squared correlation coefficient (R2) was used to evaluate the precision of the model fit. Data were analyzed using SAS System Software (SAS Inst. Inc., Cary, NC) [15]. Differences in BMP at around 40, 60 and 90 days (B42/B46, B60/B62 and B0 respectively) over the experimental period was analyzed using a repeated measures analysis (PROC MIXED) of SAS® [15]. Differences in k were tested by analysis of variance using the GLM procedure of SAS® [15]. Specifics factors in each experiments were inoculum source in the first experiment, IS ratio in the second and sample preservation in the third experiment. In the three experiments substrate was also considered as a class factor in the model; and therefore the interaction between specific factors in each experiment and substrate was tested in a factorial design.

BMPt

  P
VCH4 calculated ½STPL STPL CH4 ¼ VS sample ½kg kg VS

(1)

The BMP obtained at different times designed as BMPt; the BMP after 90 days is regarded as the B0.

2.4. Calculation methods and statistical test The bio-methane potential was modeled as a function of time by fitting the experimental data to two non-linear regression models. The first is a first-order kinetics model proposed by Hashimoto [13] (Eq. (2)): BMPt ¼ B0 (1  expk$t)

(2)

where BMPt denotes the cumulative CH4 yield (L CH4 kg VS1) at time (t) expressed in days; B0 is the ultimate CH4 yield (L CH4 kgVS1); k is the BMP rate constant, expressed in reciprocal of time (day1), which was substrate-specific and gave information about the time required to achieve a certain fraction of B0; t: time [days]. The second model is based in a modification of modification of Gompertz equation [14] (Eq. (3)).

(3)

3. Results and discussion 3.1. Impact of inoculum source In experiment 1 when comparing the least squares means (lsmeans) of BMP from substrates regardless inoculum source, cellulose was the substrate showing the highest BMP (328.7 ± 5.21 LSTP CH4 kg VS1, P < 0.05), followed by maize silage (278.3 ± 5.52 LSTP CH4 kgVS1, P < 0.05), wheat straw (241.0 ± 5.31 LSTP CH4 kgVS1, P < 0.05) and cattle manure (205.70 ± 5.21 LSTP CH4 kgVS1, P < 0.05). The average BMP of cellulose obtained in this work is similar to that obtained in literature [7,16]. The comparison of maize silage, wheat straw and cattle manure with literature, results more complicated than cellulose due to the complex composition and the high variability of these substrates among works. Indeed, different concentrations of protein, fat, cellulose and hemicellulose can be found in literature for maize silage, wheat straw and cattle manure, and therefore important differences in BMP can be expected. Amon et al. [10] stated that crude protein and crude fat concentration contributes most to the total CH4 energy of maize silage, and these factors are very dependent on harvesting time and maize variety. These authors also founded that composition and then BMP in cattle manure is mainly dependent of animal diet and performance. Similarly, a wide BMP range was found in literature for wheat straw (from 150 to 370 L CH4 kg1 VS) [12,17] mainly explained by differences in chemical composition of the different wheat straws and the pretreatment used among works. Concerning the ls-means of BMP from different inoculum sources regardless substrate, they ranked in the order: Thorsø  Foulum  Horsens > Bånlev, meaning that Thorsø and Foulum, both fed with more than 70% of animal manure (Table 2),

V. Moset et al. / Biomass and Bioenergy 83 (2015) 474e482

were the two inocula showing the highest BMP in each period. These results are in contrast with literature, where inoculum from WWTP has been identified as the most suitable for BMP assays, in detriment of inoculum from agricultural wastes treatment plants [8,and18]. This apparent better BMP performance with inocula from Foulum and Thorsø could also be related with the fact that these are thermophilic inocula and Horsens and Bånlev are mesophilic. However according with literature [19e21], no difference might be expected on B0 between temperature ranges (mesophilic vs. thermophilic). Table 3 shows the ls-means of the BMP at different periods (BMP42, BMP62 and B0) for each substrate and inoculum tested in experiment 1, and the statistical significant differences among lsmeans. The interaction between inoculum and substrate was found statistical significant (P < 0.001) in this experiment (data not shown). This fact indicates that although the numerically lower BMP was obtained in cattle manure in all inocula tested, BMP from the other substrates behaved differently depending on the inoculum source. In inoculum from Horsens, Bånlev, and Thørso the BMP of the different substrates ranked in the order: cellulose  maize Table 3 Least squares means and the statistical significant differences among means of the cumulative methane yield (BMP) at 42, 60 and 90 days and the models parameters calculated in experiment 1 with four different inocula. Parameter

┼

BMP42

BMPz62

B*0

k‫ﺍ‬

m: m

ʎ,

┼

Substrate

Inoculum source Horsens

Bånlev

Foulum

Thorsø

IS ratio

0.91

0.79

1.21

0.93

Maize silage Wheat straw Cattle manure Cellulose Maize silage Wheat straw Cattle manure Cellulose Maize silage Wheat straw Cattle manure Cellulose Maize silage Wheat straw Cattle manure Cellulose Maize silage Wheat straw Cattle manure Cellulose Maize silage Wheat straw Cattle manure Cellulose

287.0abA 251.9aAB 213.5aB 298.4bA 293.9aA 268.7aAB 225.3aB 321.0bA 299.1aA 280.6aAB 233.5aB 328.2bA 0.112bA 0.049bcB 0.051cB 0.055bB 21.61bA 11.06abB 8.25bcB 14.37cB 0.006b 1.950b 0.218b 2.102b

195.9cB 139.0bC 111.1bC 308.2bA 212.0bB 160.8bC 126.0bC 321.9bA 219.4bB 176.6bBC 136.2bC 327.9bA 0.057cA 0.034cB 0.035cB 0.057bA 14.33cB 6.78bC 5.33cC 28.28bA 4.360aB 5.297aAB 5.740aAB 6.894aA

324.7aA 278.1aA 226.8aB 283.6bA 330.6aA 293.8aA 237.5aB 284.7bAB 334.4aA 300.8aAB 245.1aC 282.4bBC 0.170aA 0.082aC 0.114aB 0.156aA 37.74aA 16.20aB 16.26aB 36.48aA 0.117b 0.000b 1.372b 1.492b

273.3bBC 227.1aB 225.0aB 391.7aA 281.6aB 248.8aB 238.2aB 397.5aA 287aB 265.7abB 249.0aB 399.4aA 0.055cB 0.055bB 0.074bA 0.071bAB 16.86cB 10.21abC 12.96abBC 27.00bA 4.860aA 0.021bB 0.371bB 3.616bA

BMP42 Cumulative methane yield at 42 days in terms of LSTP CH4 kg VS1, LSTP: liter at standard conditions. z BMP60 Cumulative methane yield at 62 days in terms of LSTP CH4 kg VS1, LSTP: liter at standard conditions. B*0 Cumulative methane yield at 91 days in terms of LSTP CH4 kg VS1, LSTP: liter at standard conditions. k‫ ﺍ‬is the BMP rate constant expressed in terms of reciprocal of time (day1), calculated from model fit of the cumulative specific bio-methane potential (BMP) curve measured in L CH4 kg VS1 as a function of time BMPt ¼ B0 (1  expk$t) [22]. , m: are the parameters calculated from model fit of the cumulative specific biom ʎ methane potential (BMP) curve measured in L CH4 kg VS1 as a function of time based in a modification of Gompertz equation [13] BMPt ¼ B0 exp {exp [mm e/B0 (l  t) þ 1]}. Where m: m is the volumetric rate of methane production expressed in terms of LCH4 kg VS1 day1 and ʎ, is the duration of the lag phase expressed in terms of time (day). aec mean values bearing different superscripts in the same row in each period are significantly different (P < 0.05). AeC means values bearing different subscripts in the same column in each period are significantly different (P < 0.05).

477

silage  wheat straw  cattle manure with different statistical significant differences among substrates depending on the inoculum source. In Bånlev and Thorsø, BMP of cellulose showed the highest BMP in each period (P < 0.05); lower significant differences on BMP determined with these two inocula were obtained among the rest of substrates. In Horsens, although cellulose showed the highest numerical BMP, this factor only resulted in a statistical significant higher (P < 0.05) than the BMP of cattle manure in any period. In Foulum however, the ranking of the BMP at different substrates was: maize silage  wheat straw  cellulose > cattle manure. Therefore the main differences between inoculum from Foulum and the rest of inocula is the ability to degrade cellulose. Cellulose has been used as a reference substrate in many BMP tests [7,16] because it has a simple and robust structure. The theoretical CH4 yield (Bu) calculated from Bushwell's formula [22] is 415 LCH4 kg VS1, which means 96% of Bu was reached by using Thorsø inoculum, 79% by using inoculum from Horsen and Bånlev and only 69% of Bu was reached with inoculum from Foulum. According to literature, from 80% to 90% of Bu can be converted into CH4 when cellulose is digested, assuming around 15% of Bu to form new cells and cell metabolism [7]. This fact could indicate a lower ability of the inoculum form Foulum to completely degrade cellulose. However, no statistical significant differences were observed in BMP of cellulose among Foulum, Horsens and Bånlev, and only Thorsø inoculum showed the highest BMP of cellulose (P < 0.05) in all periods. Concerning other differences among inoculum sources within substrates, Bånlev showed the lowest B0 (P < 0.05) for maize silage and cattle manure, but no statistical significant differences were obtained in B0 from maize and cattle manure among the other inocula. Inoculum from Bånlev also showed the lowest BMP for wheat straw in all periods, however in this case, statistical significant differences between the B0 obtained in Bånlev and Thorsø were not observed. The explanation of the differences in the apparent ability to degrade the substrates among inoculum sources cannot be explained with the results obtained in this work. A microbial characterization of inocula could help to understand their different performances at similar substrates. Concerning the modeling, the ls-means of parameters from both models and the statistical significant differences are shown in Table 3 for each substrate and inoculum tested. The Hashimoto kinetic rate [13] of the different substrates used ranged from 0.055 to 0.156, these values are lower than those reported by Raposo et al. [7], but in the same range than those reported by Gunaseelan [23], especially compared to those obtained by using inoculum from Horsen, Bånlev and Thorsø are used. In both models, Foulum showed the highest numerical rate in all substrates and periods, these differences were only statistical significant in the case of Hashimoto [13] model (P < 0.05). In Gompertz model [14], no statistical significant differences were observed on mm between Foulum, Thorsø and Horsens in the wheat straw and between Foulum and Thorsø when cattle manure was digested. These results explain the fact that samples made with inoculum from Foulum reached the highest percentage of B0 at 42 days in all substrates types including cellulose. This could indicate that hydrolysis is not the cause of the apparent lower ability to degrade cellulose of this inoculum, and therefore a longer incubation time wouldn't increase the B0 of cellulose obtained with inoculum from Foulum. Vedrenne et al. [8] also stated differences on the time to reach 90% of the B0 among inoculum sources. In their work however, the time to reach 90% of B0 was significantly reduced with the use of inoculum from WWTP digester and anaerobic digester of winery waste compared to inoculum from the laboratory reactors treating swine slurry. Veeken and Hamelers [24] and Raposo et al. [7]

478

V. Moset et al. / Biomass and Bioenergy 83 (2015) 474e482

reported that although anaerobic biodegradability not depend on temperature, the rate constants of the process increased at higher temperatures. In this case however, temperature cannot explain differences on k and mm since inoculum from Thorsø (thermophilic) showed lower rates than the other thermophilic inoculum (Foulum) and similar rates than those obtained by using inoculums from Horsens and Bånlev (mesophilic). Regarding differences on model rates among substrate composition, the rates obtained from both models showed a similar behavior. In both cases, model rate resulted dependent of inoculum source. In inoculum from Horsens the highest (P < 0.05) k and mm were both obtained in maize silage. In Bånlev, the highest mm (P < 0.05) was obtained in cellulose, however not statistical significant differences were obtained on k between maize silage and cellulose in when inoculum from Bånlev was used. Similar result was obtained in Foulum, where the highest rates (k and mm) were obtained when maize silage and cellulose were digested. In Thørso, the highest mm (P < 0.05) was obtained in cellulose, however not statistical significant differences were obtained on k between cattle manure and cellulose when inoculum from Thørso was used. The comparison among B0 and these rates results useful to extrapolate these results to continuous reactors, where retention times are generally lower than 30 days. For instance, although no statistical significant differences were obtained in B0 between maize and wheat straw in any inoculum source tested, maize showed a higher k and mm (P < 0.05) than wheat straw in all inoculum sources except Thorsø, meaning that lower time is required in maize to achieve B0 compared to wheat straw. Concerning the lag phase obtained from Gompertz model [14], although all substrates showed the numerical higher lag phase when digesting cellulose, not clear statistical significance differences were obtained on lag phase among substrates in any inoculum. Comparing among inoculum sources, the highest lag phase was observed in Bånlev for all substrates except in maize silage where Thorsø showed the highest numerical lag phase, only statistical similar (P < 0.05) to that obtained in Bånlev. This result reinforces the above hypothesis about Bånlev as the worse adapted inoculum to these substrates, which could indicate that in order to reach similar B0 in these tested substrates by using inoculum from Bånlev, longer incubation times could be required. In fact, as shown in Fig. 1a, 1b, 1c and 1d, the BMP in Bånlev showed the longer time to reach the plateau in all substrates tested. 3.2. Impact of IS ratio In this experiment no statistical significant differences in lsmeans of BMP were found between maize (337.7 ± 4.38 LCH4 kg VS1) and wheat straw (328.7 ± 4.38 LCH4 kg VS1) when averaged per substrate, regardless of the IS ratio. These values are in the same range than those obtained in experiment 1. As in this case, in experiment 1 no significant differences in BMP were found in any period between maize silage and straw when, as in this experiment, inoculum from Foulum was used. Concerning differences among IS ratios regardless of the substrate, no statistical significant differences were obtained in BMP in the IS ratio range 1.5e2; however, the interaction between IS ratio and substrate was found statistical significant (P < 0.001) in this experiment (data not shown). This fact indicates that BMP calculated with different IS ratios could behave differently depending on the substrate composition. Therefore, substrate composition should be considered in the election of the IS ratio. Table 4 shows the ls-means of the BMP at different periods (BMP42, BMP62 and B0) for each IS ratio (0.25e2.5) and substrate (maize and wheat straw) tested in experiment 2 and the statistical significant differences among ls-means. In maize, the lowest

(P < 0.05) BMP was obtained with the IS ratio of 0.25 in all periods, probably because this ratio caused a strong inhibition of BMP of maize. In fact, at the end of the incubation period, the average pH value in bottles with maize and IS ratio of 0.25 was approximately 5 and the average concentration of VFAs was: 4.21 g/L acetic acid, 7.69 g/L butyric acid and 0.46 g/L propionic acid (data not shown). That means that probably an IS ratio of 0.25 in maize was enough to assure acidogenic and acetogenic populations, however no for methanogenic populations. According to previous studies, an excessively low IS ratio may be toxic for the microorganisms [13,25], but our study indicates that this depends on the substrate type. An IS ratio of 0.25 in maize was found to inhibit CH4 production completely, while it only affected the CH4 production rate and CH4 potential of wheat straw to a smaller extent. In the range 0.5e2 a higher IS ratio gives a higher BMP in maize; in fact, the highest BMP was obtained with IS ratio of 2 (397.4 LCH4 kg VS1), although no statistical significant differences were found in BMP in the range 2e2.5 in any period. The BMP of maize obtained in this work when IS ratio higher than 2 is much higher than the range obtained in literature (280e366 LCH4 kg VS1) [10]. This fact can be explained by an overestimation of BMP, due to the residual CH4 production of the inoculum. In fact, Hansen et al. [16] and Angelidaki and Sanders [26] stated that although a high amount of inoculum guarantees a fast CH4 production decreasing the required incubation time, the high CH4 production of the inoculum can increase the uncertainty of the results. In fact, Vedrenne et al. [8] recommended a low inoculum to substrate ratio to avoid an overestimation of the B0 value. However, Raposo et al. [25] and Raposo et al. [27] didn't find important differences in B0 in a wide range of IS ratios tested: 1e3 for maize silage and 0.5 to 3 in sunflower oil cake, respectively. No significant differences were found in BMP in the range 0.5e1.5 for maize in any period. Concerning wheat straw, BMP of wheat straw increased gradually when the IS ratio increased from 0.25 to 1.5 in any period, decreasing thereafter. However no statistical significant differences were obtained in BMP of wheat straw among IS ratios in the range 0.5e2.5 in BMP45 and B0. In BMP60 an IS ratio higher than 2 resulted in higher yields than that obtained with 1.5. The slight decrease of BMP when IS ratio exceed 1.5 might indicate that an increased IS ratio for slowly degradable substrates such as wheat straw might reduce BMP. This could be explained again by the uncertainty of the results when high amount of inoculum is used [16,26], which could affect the BMP results differently (overestimation or underestimation) depending substrate composition. Regarding the models parameters, the ls-means of parameters from both models and the statistical significant differences are shown in Table 4 for each substrate and IS ratio tested. These values are in the same range than those obtained from maize and wheat straw in experiment 1 when inoculum from Foulum was used. Except in maize samples at IS ratio of 0.25, which showed the least fit to the non-linear first-order equations due to inhibition, and therefore model parameters of this treatment was not considered in statistical analysis. Similarly to results obtained in experiment 1 using inoculum from Foulum, although not differences were obtained in BMP between substrates (maize vs. wheat straw) in this experiment in the IS range between 0.25 and 1.5, maize showed higher rates (k and mm) than wheat straw in the IS range between 1 and 2.5, this result confirms the faster biodegradability at low retention times of maize compared to wheat straw. This fact can be clearly observed when comparing Fig. 2a and b, where the cumulative CH4 yields, and the obtained modeling equations and R2 values of the observed and simulated CH4 production of the different substrates and IS ratio

V. Moset et al. / Biomass and Bioenergy 83 (2015) 474e482

479

Fig. 1. Cumulative methane production (BMP) in terms of LSTP CH4/kg VS (LSTP ¼ liter at standard temperature and pressure: 273 K, 1 atm) of A) maize silage, B) wheat straw, C) cattle manure and D) cellulose as a function of time from the four inoculums. The lines representing the non-linear regression models calculated from model fit of the cumulative specific bio-methane potential (BMP) curve measured in L CH4 kg VS1 as a function of time BMPt ¼ B0 (1  expk$t) [13].

Table 4 Least squares means (ls-means) of the cumulative methane yield (BMP) at 42, 62 and 91 days and the BMP rate constant (k) of maize and wheat straw at different inoculum substrate (IS) ratios tested in experiment 2. Parameter

┼

BMP42 BMPz60 B*0 k‫ﺍ‬

m: ʎ,

Substrate

IS ratioº 0.25

0.5

1

1.5

2

2.5 (2.4)

Maize Wheat Maize Wheat Maize Wheat Maize Wheat Maize Wheat Maize Wheat

straw

16.1dB 278.7bA 14.4dB 287.3cA 12.1dB 298.5bA

straw

0.060

straw

12.48

straw

0.44

343.3c 324.2ab 347.0c 331.2abc 352.8c 341.6ab 0.098b 0.075 28.49c 16.92 1.44 0.00

375.6c 340.0a 378.7bc 345.0ab 396.9bc 355.9a 0.160aA 0.081B 38.11bcA 18.90B 0.00 0.00

387.8cb 354.2a 390.1bc 359.1a 397.4bc 376.3a 0.155aA 0.076B 40.01abcA 18.97B 0.00 0.00

457.5aB 314.3abB 458.7aA 320.3bcB 465.5aA 346.4abB 0.169aA 0.073B 50.79aA 17.24B 0.00 0.19

428.5abA 302.6abB 428.1abA 307.5bcB 437.5abA 334.3abB 0.160aA 0.075B 45.22abA 17.16B 0.00 0.17

straw straw

┼ BMP42 Cumulative methane yield at 42 days in terms of LSTP CH4 kg VS1, LSTP: liter at standard conditions. BMPz60 Cumulative methane yield at 62 days in terms of LSTP CH4 kg VS1, LSTP: liter at standard conditions. B*0 Cumulative methane yield at 91 days in terms of LSTP CH4 kg VS1, LSTP: liter at standard conditions. 1

k‫ ﺍ‬is the BMP rate constant expressed in terms of reciprocal of time (day ), calculated from model fit of the cumulative specific bio-methane potential (BMP) curve measured in L CH4 kg VS1 as a function of time BMPt ¼ B0 (1  expk$t) [22]. , m: are the parameters calculated from model fit of the cumulative specific bio-methane potential (BMP) curve measured in L CH4 kg VS1 as a function of time based in a m ʎ modification of Gompertz equation [13] BMPt ¼ B0 exp {exp [mm e/B0 (l  t) þ 1]}. Where m: m is the volumetric rate of methane production expressed in terms of LCH4 kg VS1 day1 and ʎ, is the duration of the lag phase expressed in terms of time (day). aec mean values bearing different superscripts in the same row in each period are significantly different (P < 0.05). AeC means values bearing different subscripts in the same column in each period are significantly different (P < 0.05).

are shown. In maize, no statistical significant differences were obtained in k in the range 1e2.5 and in mm in the range 1.5e2.5. Concerning wheat straw, not statistical significant differences were obtained in k and mm in the range evaluated in this study (0.25e2.5). The CH4 production of the six IS ratios fitted the first-order equations very well in wheat straw. The least correlation in this case was obtained in bottles with an IS ratio of 2.5. Hashimoto [13] and Raposo et al. [27] found increased k values with increased IS ratio in ball-milled straw and sunflower oil cake respectively, however, our work only found this relation with maize. As in experiment 1 when inoculum from Foulum was used, in this experiment no statistical significant differences were obtained

in the lag phase calculated from Gompertz model [14]. However, numerical higher lag phases were observed in both maize and wheat straw when lower IS ratios, being numerically higher in maize than in wheat straw. Which means the microbial population needs more time to start degrading the biomass at lower IS ratios, and this fact is more evident when easy biodegradable substrates like maize. 3.3. Impact of sample preservation A clear effect of sample preservation on BMP was not observed in this experiment, in fact, a strong interaction was found between sample preservation and substrate (P < 0.001), meaning that the

480

V. Moset et al. / Biomass and Bioenergy 83 (2015) 474e482

Fig. 2. Cumulative methane production (BMP) in terms of LSTP CH4/kg VS (LSTP ¼ liter at standard temperature and pressure: 273 K, 1 atm) of A) maize and B) wheat straw as a function of time at different inoculum substrate (IS) ratios: 0.25, 0.5, 1.0, 1.5, 2.0 and 2.5 in terms of g volatile solids (VS) of substrate/g VS inoculum. The lines representing the nonlinear regression models calculated from model fit of the cumulative specific bio-methane potential (BMP) curve measured in L CH4 kg VS1 as a function of time BMPt ¼ B0 (1  expk$t) [13].

effect of sample preservation on BMP depends on substrate composition. Therefore, the impact of sample preservation on the potential and BMP rate depend on substrate type, and therefore substrate composition should be considered in the election of the appropriate sample preservation. The explanation about the interaction between sample preservation and substrate is not evident, and no information about this regard has been found in literature. Therefore, more research works are needed before a clear explanation can be drawn. Table 5 shows the ls-means of the BMP at different periods (BMP46, BMP60 and B0) for each sample preservation and substrate tested and the statistical significant differences among ls-means. The values from maize and wheat straw are in the same range than those reported in experiment 1 and 2 when similar substrates where digested by using inoculum from Foulum. In grass, although ensiled samples showed the numerical higher BMP in all periods, no statistical significant differences were observed on BMP46, BMP60 and B0 among frozen, dry and ensiled samples. The higher BMP of ensiled samples could be explained because during the ensilage process, some methanogenic precursors like lactic acid, acetic acid, methanol, alcohols, formic acid and Hþ are formed from crude fiber pre-decomposition [10]. Another reason however, could be an underestimation of VS. As Kreuger et al. [28] explained, volatile compounds are formed during ensiling which could be potentially lost during the VS determination [11]. In fact, drying at 60e85
C and correction for VS loss is recommended in VS determination, especially at lower pH [15,25]. According to Kreuger et al. [28] and Zubr [29] and, 73e93% weight/weight of the total organic acid content in an ensiled substrate is lactic acid. In our study the lactic acid was not determined and correction for VS was not possible. Fresh grass samples showed a lower (P < 0.05) BMP than ensiled grass in all evaluated periods, and lower than dry grass at BMP46. This result is in accordance with Lowman et al. [30] who found that fresh perennial ryegrass (Lolium perenne) produced significantly low gas than dried grass in an in vitro study digestibility. These authors suggested that drying could alter the plant cell making them easier to degrade. In maize, freezing has a negative impact (P < 0.05) on BMP comparing with fresh samples; drying however, did not affected BMP, since no statistical significant differences on BMP were observed between fresh and dried samples. The effects of freezing were noted by Kohn and Allen [9] who reported that the freezing of forage affected neutral detergent fiber, acid detergent fiber, lignin and ash contents in different ways, depending on forage type and duration of freezing. In wheat straw no statistical significant differences on BMP were observed among sample preservation types. Concerning the modeling, the ls-means of the models

Table 5 Least squares means (ls-means) of the cumulative methane yield (BMP) at 46, 60 and 91 days and the BMP rate constant (k) of grass, maize and wheat straw at different sample preservation types tested in experiment 3. Parameter ┼

BMP46

BMP60

B*0

k‫ﺍ‬

m: ʎ,

Substrate

Fresh

Frozen

Dry

Ensiled

Grass Maize Wheat Grass Maize Wheat Grass Maize Wheat Grass Maize Wheat Grass Maize Wheat Grass Maize Wheat

359.6bB 425.4aA 337.0B 365.0bB 428.2aA 341.4B 367.5bB 431.2aA 346.0B 0.163A 0.153A 0.081B 40.970A 42.935A 19.064B 0.197 0.000 0.285

382.7abA 357.45bAB 308.7B 388.15bA 360.25bAB 314.2B 396.2ab 363.2b 316.8 0.152A 0.147A 0.076B 41.354A 43.156A 16.991B 0.216 0.000 0.535

411.1aA 396.8abA 295.0B 412.5abA 399.8abA 298.9B 416.2abA 402.9abAB 304.1B 0.151A 0.153A 0.076B 42.036A 39.705A 17.597B 0.144 0.000 0.236

422.5a

straw

straw

straw

straw

straw

straw

423.0a

428.1a

0.177

49.175

0.002

┼

BMP46 Cumulative methane yield at 46 days in terms of LSTP CH4 kg VS1, LSTP: liter at standard conditions. BMPz60 Cumulative methane yield at 60 days in terms of LSTP CH4 kg VS1, LSTP: liter at standard conditions. B*0 Cumulative methane yield at 91 days in terms of LSTP CH4 kg VS1, LSTP: liter at standard conditions. k‫ ﺍ‬is the BMP rate constant expressed in terms of reciprocal of time (day1), calculated from model fit of the cumulative specific bio-methane potential (BMP) curve measured in L CH4 kg VS1 as a function of time BMPt ¼ B0 (1  expk$t) [22]. , m: are the parameters calculated from model fit of the cumulative specific biom ʎ methane potential (BMP) curve measured in L CH4 kg VS1 as a function of time based in a modification of Gompertz equation [13] BMPt ¼ B0 exp {exp [mm e/B0 (l  t) þ 1]}. Where m: m is the volumetric rate of methane production expressed in terms of LCH4 kg VS1 day1 and ʎ, is the duration of the lag phase expressed in terms of time (day). aec mean values bearing different superscripts in the same row in each period are significantly different (P < 0.05). AeC means values bearing different subscripts in the same column in each period are significantly different (P < 0.05). IS ratio ¼ 1 in all bottles tested.

parameters for each substrate and sample preservation method tested and the statistical significant differences among treatments are shown in Table 5. No statistical significant differences were observed in k among sample preservation types in any substrate tested. As expected, in all sample preservation types wheat straw showed the lowest k and mm (P < 0.05), indicating again that wheat straw was more slowly fermentable than grass and maize. In addition, lag phase calculated from Gompertz model [14] was numerically lower in maize than in wheat straw and grass, meaning

V. Moset et al. / Biomass and Bioenergy 83 (2015) 474e482

481

Fig. 3. Cumulative methane production (BMP) in terms of LSTP CH4/kg VS (LSTP ¼ liter at standard temperature and pressure: 273 K, 1 atm) of A) grass, B) maize and C) wheat straw as a function of time at the different sample preservation types tested in experiment 3 (fresh, dry and frozen and ensiling). The lines representing the non-linear regression models calculated from model fit of the cumulative specific bio-methane potential (BMP) curve measured in L CH4 kg VS1 as a function of time BMPt ¼ B0 (1  expk$t) [13].

again that microorganisms began to degrade faster the maize than the wheat straw or the grass at any sample preservation. This faster biodegradability of maize can be observed in Fig. 3, where the BMP and the obtained modeling equations and R2 values of the observed and simulated BMP production of the different substrates and samples preservation are shown.

[4]

[5]

4. Conclusions

[6]

Thorsø and Foulum, were the two inocula showing the highest BMP in each period both inocula came from a feeding with more than 70% of animal manure under thermophilic conditions. Foulum inoculum showed the highest degradation rates in all substrates and periods. However, a lower ability to degrade cellulose compared to other inocula was shown by this inoculum. This lower ability was not related with the degradation rates. An interaction was found between IS ratio and substrate composition, mainly due to the different behavior between substrates and IS ratio at higher or lower ratios. Maize silage was more affected to IS ratio than wheat straw. Based on BMP and degradation rates, the optimal IS ratio range for maize was 1.0e1.5, lower IS ratio caused inhibition and higher caused an overestimation. The optimal IS ratio for wheat straw was 0.5e2.5. The effect of sample preservation on BMP depends of substrate composition. Ensiling and drying of grass increased the CH4 yield, but on the other hand drying of maize had no significant impact. In maize freezing reduced the BMP. In straw the storing method had no significant influence on the BMP. The determination of kinetics parameters can supplement the information obtained from in a BMP test in terms of hydrolysis rate and microbial adaptation to a specific substrate.

[7]

[8]

[9] [10]

[11]

[12] [13] [14]

[15] [16]

References [1] K.-o. Energiministeriet, Energi Strategi 2050-fra kul, olie og gas til grøn energi. Sammenfatning [From coal, oil and gas to green energy. Summary], Regeringen, Copenhagen, Denmark, 2011. [2] Association EB, A Biogas Road Map for Europe, Report, European Biomass Association AEBIOM, 2009. [3] IPCC, Emissions from livestock and manure management, in: IPCC Guidelines

[17]

[18]

for National Greenhouse Gas Inventories, Agriculture, Forestry and Land Use, vol. 4, 2006. Kanagawa, Japan. W. Owen, D.C. Stuckey, J.B. Healy Jr., L.Y. Young, P.L. McCarty, Bioassay for monitoring biochemical methane potential and anaerobic toxicity, Water Res. 13 (6) (1979) 485e492. I. Angelidaki, M. Alves, D. Bolzonella, L. Borzacconi, J.L. Campos, A.J. Guwy, S. Kalyuzhnyi, P. Jenicek, J.B. van Lier, Defining the biomethane potential (BMP) of solid organic wastes and energy crops: a proposed protocol for batch assays, Water Sci. Technol. 59 (5) (2009) 927e934. ISO 11734:1995, Water Quality e Evaluation of the “Ultimate” Anaerobic Biodegradability of Organic Compounds in Digested Sludge e Method by Measurement of the Biogas Production first ed., 13, BSI, 1996.
ndez-Cegrí, M.A. De la Rubia, R. Borja, F. Be
line, C. Cavinato, F. Raposo, V. Ferna
ndez, M. Fern
G. Demirer, B. Ferna andez-Polanco, J.C. Frigon, R. Ganesh,
ndez, G. Menin, A. Peene, P. Scherer, M. Torrijos, P. Kaparaju, J. Koubova, R. Me H. Uellendahl, I. Wierinck, V. de Wilde, Biochemical methane potential (BMP) of solid organic substrates: evaluation of anaerobic biodegradability using data from an international interlaboratory study, J. Chem. Technol. Biotechnol. 86 (2011) 1088e1098.
line, P. Dabert, N. Bernet, The effect of incubation conditions F. Vedrenne, F. Be on the laboratory measurement of the methane producing capacity of livestock wastes, Bioresour. Technol. 99 (1) (2008) 146e155. R.A. Kohn, M.S. Allen, Storage of fresh and ensiled forages by freezing affects fibre and crude protein fractions, J. Sci. Food Agric. 58 (2) (1992) 215e220. T. Amon, B. Amon, V. Kryvoruchko, W. Zollitsch, K. Mayer, L. Gruber, Biogas production from maize and dairy cattle manuredinfluence of biomass composition on the methane yield, Agric. Ecosyst. Environ. 118 (1) (2007) 173e182. APHA, in: A.E. Greenberg, L.S. Clesceri, A.D. Eaton (Eds.), Standard Methods for the Examination of Water and Wastewater, 21th Ed., American Public Health Association, Washintong DC, USA, 2005. H.B. Møller, S.G. Sommer, B.K. Ahring, Methane productivity of manure, straw and solid fractions of manure, Biomass Bioenergy 26 (5) (2004) 485e495. A.G. Hashimoto, Effect of inoculum/substrate ratio on methane yield and production rate from straw, Biol. Wastes 28 (4) (1989) 247e255. €, T.Y. Pai, C.F. Chiang, K.P. Chao, H.M. Lo, T.A. Kurniawan, M.E.T. Sillanp€ aa M.H. Liu, S.H. Chuang, C.J. Banks, S.C. Wang, K.C. Lin, C.Y. Lin, W.F. Liu, P.H. Cheng, C.K. Chen, H.Y. Chiu, H.Y. Wu, Modeling biogas production from organic fraction of MSW co-digested with MSWI ashes in anaerobic bioreactors, Bioresour. Technol. 101 (16) (2010) 6329e6335. SAS, SAS User's Guide: Statics, Ver. 9.0., SAS Institute Inc., Cary, N.C., 2001. T.L. Hansen, J.E. Schmidt, I. Angelidaki, E. Marca, C. Jansen J la, H. Mosbæk, T.H. Christensen, Method for determination of methane potentials of solid organic waste, Waste Manag. 24 (2004) 393e400. I. Angelidaki, L. Ellegaard, Codigestion of manure and organic wastes in centralized biogas plants. Status and future trends, Appl. Biochem. Biotechnol. 109 (2003) 95e105. C. Mateescu, I. Constantinescu, Comparative analysis of inoculum biomass for biogas potential in the anaerobic digestion, UPB Sci. Bull. Ser. B 73 (3) (2011) 99e104.

482

V. Moset et al. / Biomass and Bioenergy 83 (2015) 474e482

[19] A.G. Hashimoto, V.H. Varel, Y.R. Chen, Ultimate methane yield from beef cattle manure: effect of temperature, ration constituents, antibiotics and manure age, Agric. Waste 3 (4) (1981) 241e256. [20] J. Palatsi, A. Gimenez-Lorang, I. Ferrer, X. Flotats, Start-up strategies of thermophilic anaerobic digestion of sewage sludge, Water Sci. Technol. 59 (9) (2009) 1777e1784. [21] V. Moset, M. Poulsen, R. Wahid, O. Højberg, H.B. Møller, Mesophilic versus thermophilic anaerobic digestion of cattle manure: methane productivity and microbial ecology, Microb. Biotechnol. 8 (5) (2015) 787e800. [22] A.M. Buswell, H.F. Mueller, Mechanics of methane fermentation, J. Ind. Eng. Chem. 44 (1952) 550e552. [23] V.N. Gunaseelan, Biochemical methane potential of fruits and vegetable solid waste feedstocks, Biomass Bioenergy 26 (2004) 389e399. [24] A. Veeken, B. Hamelers, Effect of temperature on hydrolysis rates of selected biowaste components, Bioresour. Technol. 69 (1999) 249e254. [25] F. Raposo, C.J. Banks, I. Siegert, S. Heaven, R. Borja, Influence of inoculum to

[26] [27]

[28]

[29] [30]

substrate ratio on the biochemical methane potential of maize in batch tests, Process Biochem. 41 (6) (2006) 1444e1450. I. Angelidaki, W. Sanders, Assessment of the anaerobic biodegradability of macropollutants, Rev. Environ. Sci. Bio/Technol. 3 (2004) 117e129.
n, InF. Raposo, R. Borja, M.A. Martín, A. Martín, M.A. de la Rubia, B. Rinco fluence of inoculumesubstrate ratio on the anaerobic digestion of sunflower oil cake in batch mode: process stability and kinetic evaluation, Chem. Eng. J. 149 (2009) 70e77. €rnsson, Ensiling of crops for biogas production: E. Kreuger, I.A. Nges, L. Bjo effects on methane yield and total solids determination, Biotechnol. Biofuels 4 (2011) 44. J. Zubr, Methanogenic fermentation of fresh and ensiled plant materials, Biomass 11 (3) (1986) 159e171. R.S. Lowman, M.K. Theodorou, D. Cuddeford, The effect of sample processing on gas production profiles obtained using the pressure transducer technique, Anim. Feed Sci. Technol. 97 (3) (2002) 221e237.