Anaerobic Digestion of Lignocellulosic Materials Using Ethanol-Organosolv Pretreatment

Abstract

Biomethane potential from the anaerobic digestion of rice straw, hazelnut skin, and cocoa bean shell was investigated after applying an organosolv pretreatment. Pretreatment of the three lignocellulosic materials was performed at 150°C and 180°C for 60 min using 50% (v/v) ethanol as organic solvent. Afterward, untreated and pretreated feedstocks were used in batch biomethane production tests under mesophilic conditions (37 ± 1°C). The highest efficiency was obtained pretreating the rice straw at 180°C, which resulted in a 42% higher biomethane production yield from the 235 mL CH₄/g volatile solid obtained with the untreated straw. A lower increase was achieved for hazelnut skin, whereas a partial inhibition of the biomethane production was observed for cocoa bean shell, compared with untreated materials. Compositional analyses showed that, after organosolv pretreatment, a reduction of the lignin and hemicellulose content was achieved in the three lignocellulosic residues, together with a relative increase of cellulose content.

Introduction

Biogas is a renewable energy source that can be obtained from the anaerobic digestion (AD) of various organic substrates. Compared to other biofuels, biogas results in a higher energy efficiency, which makes its commercial production advantageous (Murphy and Power, 2009). Lignocellulosic materials (LMs) have the potential to represent the main feedstock for biogas production worldwide, because of their abundance, relatively low cost, and sustainability (Deublein and Steinhauser, 2011). However, to date, LMs are not commonly employed in anaerobic digesters (Sawatdeenarunat et al., 2015).

Due to the recalcitrant structure of LMs toward biodegradation, the biogas yields are generally low (Taherzadeh and Karimi, 2008). In particular, the presence of lignin in the LM matrix represents the main limitation to the bacterial hydrolysis by hindering the access of microorganisms to the carbohydrate portion (Karimi et al., 2013). To overcome this drawback, a pretreatment step is required to enhance the digestibility of LMs (Hendriks and Zeeman, 2009). Several pretreatment methods have been investigated in recent years, including steam explosion, alkali, acid, ammonia fiber explosion, and organic solvents (Monlau et al., 2013). Among these techniques, organosolv is emerging as one of the most promising in efficiently removing lignin from the lignocellulosic structure (Hesami et al., 2015; Ostovareh et al., 2015).

The organosolv method is based on the pretreatment of LMs with organic solvents, such as methanol, ethanol, acetone, glycerol and organic acids, at temperatures in the range from 100°C to 250°C (Mancini et al., 2016a). This treatment leads to the breakdown of the lignin molecules by cleavage of ether linkages and their subsequent dissolution (Zhao et al., 2009). By employing low boiling point alcohols, several advantages can be obtained with the organosolv method compared to other conventional pretreatments, such as acid or alkaline (Park et al., 2010). After the pretreatment, the solvent can be easily recovered and recycled through a distillation stage (Taherzadeh and Karimi, 2008). Moreover, the recovery of a highly pure lignin fraction as an economically viable byproduct can be obtained at the end of the process (Doherty et al., 2011). Therefore, the organosolv technique represents a promising pretreatment within the development of LM biorefineries (Mesa et al., 2016).

Up to now, the organosolv pretreatment has mostly been investigated as a technique to improve the enzymatic hydrolysis for bioethanol production (Kabir et al., 2015). In the last couple of years, however, organosolv has also been tested as a method to improve the biogas production from different LMs, including forest and agricultural residues (Mancini et al., 2016a).

In this research, the effect of organosolv pretreatment was evaluated on the biomethane production yields from three different LMs, namely rice straw, hazelnut skin, and cocoa bean shell. Rice straw represents the main crop residue in the world, with a yearly production of around 700 million tons (Croce et al., 2016). Ninety percent of rice cultivation is concentrated in the developing countries of Eastern and Southeast Asia, where open field burning is still a common practice to dispose the unused straw, thus causing serious pollution cases (Contreras et al., 2012). Differently, employing rice straw to produce biogas through the AD process is considered one of the most environmentally friendly methods for converting this LM into renewable energy (Mussoline et al., 2013).

Environmental and economic problems can also arise in the food industry from the disposal of hazelnut skin (Piccinelli et al., 2016) and cocoa bean shell (Martínez et al., 2012). These byproducts, which represent the skin and shell of hazelnut and cocoa fruits, respectively, are obtained after the roasting process (Mancini et al., 2016b).

In this study, biomethane potential (BMP) tests were carried out to assess the methane production yields in batch assays containing the three untreated and organosolv-pretreated LMs. The effect of different pretreatment temperatures (i.e., 150°C and 180°C) was evaluated, employing 50% (v/v) ethanol as the organic solvent. The biomethane production yields were recorded for all the analyzed LMs and the experimental results were fitted using a first-order and a modified Gompertz model (Lay et al., 1997), providing information about the rates of methane production of the untreated and organosolv-pretreated LMs. The development of AD was further monitored by analyzing the volatile fatty acid (VFA) concentration profiles. Moreover, the composition of the three LMs was investigated to assess the extent of the pretreatment on the lignin and carbohydrate portions.

Materials and Methods

Feedstocks and inoculum

Rice (Oryza sativa) straw was harvested from agricultural fields in Pavia (Italy) and cut down to a particle size smaller than 4 mm. Hazelnut (Corylus avellana) skin and cocoa (Theobroma cacao) bean shell were obtained as byproducts of the conventional industrial roasting process. The two substrates, received from an Italian company manufacturing chocolate and confectionery products, were sieved through a 4 mm sieve. The physicochemical characterization of the three LMs is reported in Table 1.

Table 1.

Characteristics of Raw Lignocellulosic Materials

	Rice straw	Hazelnut skin	Cocoa bean shell
TS (%)^a	93.1 ± 0.1	91.8 ± 0.2	90.6 ± 0.3
VS (%)^a	76.8 ± 1.1	89.1 ± 0.8	81.2 ± 0.1
TKN (% TS)	0.7 ± 0.0	1.3 ± 0.2	3.3 ± 0.5
Protein content (% TS)	4.9 ± 0.3	8.3 ± 1.5	20.3 ± 3.2

TS and VS are expressed in terms of fresh matter.

TS, total solids; VS, volatile solids; TKN, total Kjeldahl nitrogen.

The digestate from a full-scale AD plant treating buffalo manure and milk whey from a mozzarella factory located in Capaccio (Italy) was used as the inoculum. The inoculum was degassed for 2 days at 37°C before using it in the experiments. The total solid (TS) and volatile solid (VS) content of the inoculum was 2.62 (±0.14)% and 1.67 (±0.03)%, respectively.

Organosolv pretreatment

The organosolv pretreatment was performed using a high-pressure stainless steel vessel (Sigma-Aldrich, Germany) with a working volume of 300 mL. An identical procedure was applied to pretreat all the three LMs. The reactor was first loaded with 15 g of LM and soaked in 150 mL of 50% (v/v) ethanol. The reactor was sealed and placed in a TCF 50 PRO convection oven (ArgoLab, Italy), equipped with a ramping program. The oven was heated at a rate of 3°C/min to the desired temperature (i.e., 150°C or 180°C), which was finally held for 60 min. Afterward, the reactor was cooled in an ice bath. The pretreated LM was removed, and washed with 100 mL fresh 50% (v/v) ethanol and subsequently with distilled water until pH 7.0 was obtained in the liquor. The LMs were left overnight to air dry and finally stored in plastic bags at room temperature until further use.

BMP tests

BMP batch tests were carried out in 125 mL sealed serum bottles (Wheaton) under mesophilic conditions (37°C ± 1°C). Biomethane production was measured by the water displacement method, according to the procedure described by Esposito et al. (2012), modified as in Mancini et al. (2016b). Each bottle was loaded with 50 mL of inoculum and 0.5 g VS of untreated or pretreated LM to obtain an inoculum to substrate ratio of 1.5. All the experiments were performed in triplicate and the daily biomethane production was recorded until it was lower than 1% of the cumulative volume of the produced biomethane (i.e., on day 43). For VFA analysis, 0.5 mL of the liquid phase was daily sampled from each bottle during the first 10 days of the experiment, except for days 1, 3, 8, and 9.

Biomethane production kinetics were evaluated for each BMP test to quantify whether the organosolv pretreatment caused an enhancement of the AD rates. Two models were applied, namely a first-order kinetic model and a modified Gompertz model (Lay et al., 1997), using Equations (1) and (2), respectively: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} G = {G_m} \left[ {1 - { \rm{ex}}{{ \rm{p}}^{ \left( { - {k_0}t} \right) }}} \right] \tag{1} \end{align*} \end{document} \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} G = { G_m } \ { \rm { exp } } \left\{ { - { \rm { exp } } \left[ { { \frac { { R_m } \ e } { { G_m } } } \left( { { \rm { \lambda } } - t } \right) + 1 } \right] } \right\} \tag { 2 } , \end{align*} \end{document}

where G (mL CH₄/g VS) is the cumulative volume of methane after a time t (d), G_m (mL CH₄/g VS) is the maximum cumulative volume at the end of the experimental run, k₀ (1/day) is the first-order kinetic constant, t (day) is the digestion time, e is the Euler's number (i.e., 2.7183), λ is the lag phase (day), and R_m (mL CH₄/g VS·day) is the maximum biomethane production rate. k₀, λ, and R_m were determined by curve-fitting using the software GraphPad Prism 6.0 (GraphPad Software, Inc.), based on the experimental data of cumulative methane production obtained in the BMP tests. The coefficient of determination r² was calculated for both the adopted models to evaluate the accuracy of the predictions.

Water retention value

Water retention value (WRV), also known as water swelling capacity, is the ability of a substrate to keep water molecules in the cell wall pores (Goshadrou et al., 2013). WRV is used as an indication of the accessible interior surface area and the consequent suitability of the LM to enzymatic hydrolysis. The WRV analysis is based on the principle that no enzyme can enter the pores of LMs if water cannot (Karimi and Taherzadeh, 2016). After the centrifugation of a water-saturated sample, the WRV is defined as the amount of water that can be retained per unit weight of dry material.

Approximately 1.0 g of each LM was mixed with deionized water in a bottle agitated at 150 rpm for 60 min. The mixture was then filtered using a 0.45 μm filter (Merck Millipore). The obtained cake was transferred into a nonwoven fabric material, which was soaked in deionized water for 2 h at room temperature. The fabric was wrapped, placed into a centrifuge tube with support to make space for water accumulation, and centrifuged at 3,000 g for 15 min. The substrate was collected and weighed before and after drying at 105°C for 24 h. The WRV was calculated as follows (Goshadrou et al., 2013): \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} { \rm { WRV } } = { \frac { { W_ { wet } } - { W_ { dry } } } { { W_ { dry } } } } \tag { 3 } , \end{align*} \end{document}

where W_wet and W_dry are the wet and oven dry mass of the LM, respectively.

Analytical methods

TS and VS of both untreated and pretreated LMs were determined by drying the samples to constant weight at 105°C and by igniting at 575°C, respectively (Sluiter et al., 2008a). Total Kjeldahl nitrogen (TKN) was measured using the Kjeldahl method (Pansu and Gautheyrou, 2007). Total proteins were calculated multiplying the TKN values by a correction coefficient of 6.25 (Hall and Schönfeldt, 2013). VFAs were analyzed using a Prominence LC-20A Series HPLC (Shimadzu, Japan) equipped with a Rezex ROA-Organic Acid H⁺ column (Phenomenex) heated at 40°C and an SPD-20A UV detector set at 220 nm. A 0.0065 mM H₂SO₄ solution was used as mobile phase at a flow rate of 0.6 mL/min. The detection limit was 0.1 mM for each compound analyzed. Before the analysis, the samples were centrifuged for 5 min at 8,000 rpm and filtered with 0.22 μm Millex cellulose membranes (Merck Millipore).

Content of structural carbohydrates and lignin of the pretreated and untreated LMs was analyzed according to the procedure described by Sluiter et al. (2008b). A two-step acid hydrolysis was applied using first concentrated (i.e., 72%) and then diluted (i.e., 4%) H₂SO₄. The acid-soluble lignin content of the LMs was determined using a Lambda 365 UV/Vis spectrometer (Perkin Elmer), whereas the acid-insoluble lignin content was determined gravimetrically after drying the samples at 575°C. The structural carbohydrates of the LMs were analyzed using the HPLC reported above, equipped with a RID-20A refractive index detector (Shimadzu, Japan) and a Rezex RPM-Monosaccharide Pb²⁺ (8%) column (Phenomenex), heated at 90°C. HPLC-grade water was used as the mobile phase at a flow rate of 0.6 mL/min.

Statistical analyses

Statistically significant difference between the biomethane production of the pretreated and the untreated substrates in the batch assays was determined by a paired t-test (Montgomery, 2008) using the software package Minitab 17.0 (Minitab, Inc., State College). The results were considered statistically significant when the p-value obtained was below 0.05.

Results and Discussion

Effect of the organosolv pretreatment on LM composition

Rice straw, hazelnut skin, and cocoa bean shell were pretreated using 50% (v/v) ethanol before AD, to improve the biogas production yield. The length of the pretreatment was set to 1 h, based on the results of previous studies, which showed a better efficiency compared to a shorter pretreatment length of 30 min (Mancini et al., 2016a). The untreated and pretreated materials were characterized in terms of carbohydrates and lignin content, and the results obtained are presented in Table 2. The composition of the raw substrates was significantly different between rice straw, hazelnut skin, and cocoa bean shell. Nevertheless, the pretreatment had similar impacts in changing the original composition of the three LMs.

Table 2.

Chemical Composition and Water Retention Value of Untreated and Organosolv-Pretreated Lignocellulosic Materials

Lignocellulosic material	Pretreatment conditions	Glucan (%)	Xylan (%)	Galactan (%)	Arabinan (%)	Mannan (%)	Lignin (%)	Water retention value (%)
Rice straw	Untreated	28.6 ± 0.2	14.3 ± 0.9	0.4 ± 0.1	4.8 ± 0.2	ND	17.3 ± 0.3	1.52 ± 0.03
	1 h, 150°C, 50% ethanol	29.7 ± 0.8	12.8 ± 0.1	0.3 ± 0.0	4.4 ± 0.3	ND	14.8 ± 0.4	1.68 ± 0.02
	1 h, 180°C, 50% ethanol	32 ± 0.1	11.3 ± 0.2	0.3 ± 0.0	4.4 ± 0.8	ND	14.1 ± 0.1	1.77 ± 0.12
	Hazelnut skin	Untreated	11.4 ± 0.4	2.9 ± 0.2	0.9 ± 0.2	2.1 ± 0.1	ND	34.4 ± 1.6	1.51 ± 0.04
1 h, 150°C, 50% ethanol		11.5 ± 0.1	2.6 ± 0.2	0.6 ± 0.2	1.7 ± 0.1	ND	32.6 ± 0.6	1.57 ± 0.05
1 h, 180°C, 50% ethanol		12.5 ± 0.1	2.5 ± 0.2	0.6 ± 0.2	1.6 ± 0.0	ND	32.5 ± 0.2	1.56 ± 0.04
Cocoa bean shell		Untreated	13.5 ± 0.1	0.9 ± 0.2	3.1 ± 0.1	1.5 ± 0.0	1.5 ± 0.1	29.9 ± 0.3	3.22 ± 0.04
	1 h, 150°C, 50% ethanol	13.9 ± 0.7	0.6 ± 0.1	3.0 ± 0.2	0.8 ± 0.1	1.7 ± 0.1	27.6 ± 0.2	5.26 ± 0.72
	1 h, 180°C, 50% ethanol	15.0 ± 0.3	0.5 ± 0.1	2.7 ± 0.1	0.5 ± 0.1	2.1 ± 0.2	26.3 ± 0.1	5.47 ± 0.13

ND, not detected.

Total lignin content, expressed as the sum of acid-insoluble lignin and acid-soluble lignin, was 34.4% and 29.9%, in terms of dry matter, for hazelnut skin and cocoa bean shell, respectively. The untreated rice straw had much lower lignin content (i.e., 17.3%). Depending on the temperature adopted, the pretreatment was able to reduce the lignin content by 14–15% for rice straw, 5–6% for hazelnut skin, and 8–12% for cocoa bean shell. A higher pretreatment temperature (i.e., 180°C) corresponded to a higher delignification compared to the milder operative condition (i.e., 150°C). The lignin removal observed in the three pretreated LMs was likely linked to both the cleavage of the bonds between lignin and carbohydrates and the solubilization of lignin (McDonough, 1992).

Carbohydrate content was significantly lower for hazelnut skin and cocoa bean shell than for rice straw. The glucan content, which refers to the cellulose amount in the LMs, was 28.6%, 11.4%, and 13.5% for the raw rice straw, hazelnut skin, and cocoa bean shell, respectively (Table 2). The sum of the other polysaccharides, constituting the hemicellulose portion, was 19.5% for rice straw, 5.9% for hazelnut skin, and 7.0% for cocoa bean shell (Table 2). The total carbohydrate content of hazelnut skin and cocoa bean shell was similar to that reported by Zeppa et al. (2015) and Martínez et al. (2012), respectively. Different from rice straw, the protein and fat content of hazelnut skin and cocoa bean shell are not negligible: the amount of proteins constituted 8.3% and 20.3% of the raw hazelnut skin and cocoa bean shell, respectively (Table 1). The fat content was not assessed in this study. However, the amount of fat for cocoa bean shell has been reported in the range 4–18% (Redgwell et al., 2003), whereas Zeppa et al. (2015) recorded a total fat content between 11% and 19% for hazelnut skin.

As a result of pretreatment, xylan and arabinan (i.e., the main constituents of hemicellulose) decreased for all the three LMs (Table 2). On the other hand, after the organosolv pretreatment, the glucan content increased by 4–12%, 1–10%, and 3–11% for rice straw, hazelnut skin, and cocoa bean shell, respectively. Although the cellulose hydrolysis would benefit from a decreased amount of hemicellulose in the matrix, a loss of hemicellulosic sugars can result in a lower biogas production (Gu et al., 2015).

Effect of organosolv pretreatment on LM WRV

Values of the water retention capacity obtained for the three LMs are reported in Table 2. The WRV of the untreated rice straw, hazelnut skin, and cocoa bean shell was 1.52, 1.51, and 3.22, respectively. The WRV of the untreated rice straw was in line with the results obtained by Teghammar et al. (2012), who measured a WRV of 1.45.

An increase of the WRV was observed for all the pretreated materials. In particular, for rice straw and cocoa bean shell, the enhancement was significant. The WRV of cocoa bean shell increased by 63–70% at the pretreatment temperature of 150°C and 180°C, respectively. For rice straw, the WRV was enhanced by 11–16%. A lower increase, in the range of 3–4%, was recorded for hazelnut skin. The increase of the WRV was probably related to the effectiveness of the organosolv pretreatment in removing lignin and hemicellulose, thus causing an increase of the accessible surface area and the pore volume (Zhao et al., 2009). A direct correlation between the carbohydrate accessibility to microorganisms and the digestibility of LMs was reported by several studies (Karimi and Taherzadeh, 2016; Mancini et al., 2018a).

Effect of organosolv pretreatment on biogas production from LMs

The cumulative methane production curves obtained from the AD of the three untreated and pretreated LMs are shown in Fig. 1, while the final production yields are reported in Table 3, as the average of triplicates. The organosolv pretreatment was particularly effective for rice straw, with a significant enhancement of the BMP (Fig. 1a). A 29% increase of the cumulative biogas production was obtained upon pretreating rice straw with the organosolv method at 150°C for 1 h, that is, from 235 (±12) to 303 (±19) mL CH₄/g VS. An increase of the pretreatment temperature to 180°C had a further beneficial effect on the methane production yield, which increased to 332 (±6) mL CH₄/g VS, corresponding to a 41% increase compared to the untreated straw. The enhancement obtained was associated with the dual benefit of the pretreatment in achieving a delignification of the straw (Table 2), together with an increase of the accessibility of the material to the microorganisms assessed by the WRV (Table 2). Previous studies showed an inverse linear relationship between the lignin content of a substrate and its BMP (Liew et al., 2012; Mirmohamadsadeghi et al., 2014). At the same time, an increase in the LM accessibility improves the hydrolysis stage (Chandra et al., 2009; Shafiei et al., 2014).

FIG. 1.

Cumulative methane production from AD of rice straw (a), hazelnut skin (b), and cocoa shell (c) ● untreated; ■: organosolv at 150°C; organosolv at 180°C) and modified Gompertz model fit with experimental data ( untreated; organosolv at 150°C; organosolv at 180°C). AD, anaerobic digestion.

Table 3.

Specific Methane Yield, and Specific Rate Constant and Coefficients of Untreated and Organosolv-Pretreated Lignocellulosic Materials

Lignocellulosic material	Pretreatment conditions	Specific methane yield (mL CH₄/g VS_in)	Specific rate constant k₀ (1/day)	Correlation coefficient^a r²	Specific rate constant R_m (mL CH₄/g VS·day)	Lag phase time λ (day)	Correlation coefficient^b r²
Rice straw	Untreated	235 ± 12	0.217 ± 0.031	0.961	23.68 ± 0.97	1.8	0.990
	150°C organosolv	303 ± 19	0.247 ± 0.035	0.958	26.99 ± 0.97	1.5	0.991
	180°C organosolv	332 ± 6	0.249 ± 0.034	0.960	30.43 ± 0.78	1.3	0.995
	Hazelnut skin	Untreated	261 ± 3	0.099 ± 0.013	0.970	21.93 ± 0.25	1.6	0.999
150°C organosolv		264 ± 4	0.135 ± 0.019	0.954	31.67 ± 0.71	1.5	0.997
180°C organosolv		288 ± 6	0.136 ± 0.017	0.963	35.14 ± 0.49	1.4	0.999
Cocoa bean shell		Untreated	231 ± 6	0.190 ± 0.011	0.993	31.30 ± 0.71	0.6	0.995
	150°C organosolv	181 ± 2	0.184 ± 0.019	0.977	26.23 ± 0.79	1.0	0.992
	180°C organosolv	219 ± 4	0.185 ± 0.017	0.983	29.97 ± 0.88	0.8	0.992

First-order model coefficient of determination.

Modified Gompertz model coefficient of determination.

A slight improvement in the biogas production from hazelnut skin was obtained after the organosolv pretreatment at 180°C (Fig. 1b). A biomethane production of 288 (±6) mL CH₄/g VS was achieved, which represented a 10% enhancement compared to the untreated LM. On the other hand, the cumulative methane yield obtained by pretreating hazelnut skin at 150°C was similar to that obtained with the untreated substrate, that is, 264 (±3) and 261 (±4) mL CH₄/g VS, respectively.

A negative impact of the organosolv pretreatment on the biogas yields from the cocoa bean shell was noticed during the BMP test (Fig. 1c). At both pretreatment temperatures, the final biogas production yield was lower than that achieved using the untreated LM, that is, 231 (±16) mL CH₄/g VS. This adverse result might be attributed to a potential loss of biodegradable matter that occurred during the pretreatment, namely proteins and fats, which constitute a significant part of cocoa bean shell. Moreover, despite the cocoa bean shell being repeatedly washed after the pretreatment, it is likely that the ethanol was not completely removed from the pretreated substrate, due to its higher WRV (Table 2). This could have resulted in a partial inhibition of the microorganisms, in particular the hydrolytic bacteria, which are known to be susceptible to inhibition by organic solvents (Sun and Cheng, 2002). This could represent a potential disadvantage in employing organosolv for the pretreatment of certain LMs such as cocoa bean shell, since the consumption of high amounts of water for washing the feedstock could make the whole process economically unprofitable.

To verify whether the biomethane production yields achieved after pretreating the LMs with organosolv were significantly different from those obtained with the untreated LMs, a statistical analysis was conducted using a paired t-test (Montgomery, 2008). The increase of the biogas production obtained after pretreating the rice straw with organosolv was statistically significant at both pretreatment temperatures (i.e., p-value was 0.023°C and 0.026°C at 150°C and 180°C, respectively). On the other hand, the cumulative values obtained for the pretreated hazelnut skin were not significantly different from the untreated material (p-value >0.05). The inhibition observed in the biomethane production yield using the cocoa bean shell pretreated at 150°C was statistically significant, with a p-value of 0.006.

Despite a technical-economic analysis not being conducted in this study, the results of previous investigations on the pretreatment of LMs to enhance the AD process concluded that the cost of the solvent used for the pretreatment represented the main part of the material costs (Kabir et al., 2015). Hence, an efficient recovery of the solvent is crucial for the feasibility of the industrial scale process. When low molecular weight organic solvents, such as ethanol, are used for the pretreatment, their recovery and reuse can be more efficient, thus leading to better economics of the whole process.

Methane production kinetics

First-order kinetic models are commonly used to determine the methane production rates during AD (Shana et al., 2013) and are particularly appropriate for complex substrates such as LMs, where hydrolysis is considered the limiting step (Stronach et al., 1986). Recently, the Gompertz equation has been employed successfully to model the biomethane production, with the assumption that the biomethane production rate is proportional to the microbial activity in the AD reactor (Krishania et al., 2013). In this study, the specific rate constants k₀ and the maximum biomethane production rates R_m were obtained by fitting the experimental results of the BMP tests with a first-order and a modified Gompertz model, respectively. The values of k₀ and R_m, estimated with a 95% confidence interval, are reported in Table 3. The fitting of the cumulative biomethane production curves by the Gompertz model is shown in Fig. 1.

Degradation process of rice straw was improved by employing a pretreatment temperature of 150°C (Fig. 1). An additional increase of the k₀ was obtained for the rice straw pretreated at 180°C (Table 3). The specific rate constant was 0.217, 0.247, and 0.249 1/day, for the rice straw untreated and pretreated at 150 and at 180°C, respectively. R_m was also enhanced by the pretreatment, increasing from 23.68 mL CH₄/g VS·day for the raw straw to 26.99 and 30.43 mL CH₄/g VS·day with the 150°C and 180°C pretreated straw, respectively. The significant acceleration of the AD process caused by the organosolv pretreatment was mainly due to the increase of the accessible surface area and the partial removal of lignin and hemicellulose (Table 2).

The highest enhancement of the kinetic rate was observed for hazelnut skin, with the k₀ increasing from 0.099 1/day for the untreated LM to 0.135 and 0.136 1/day for the pretreated LM at 180°C and 150°C, respectively (Table 3). After 11 days, 91% and 90% of the final biomethane production yields were achieved at these temperatures, whereas the methane production obtained from the untreated hazelnut skin was 77% on the same day. The removal of the lignin and hemicellulose fraction from the pretreated hazelnut skin, which was similar at both pretreatment temperatures (Table 2), led to a faster biomethane production (Fig. 1b), entailing a shorter AD time. In continuous systems, this result can be of major importance as it can lead to an optimization of the overall process and a decrease of the reactor volumes. Similar to k₀, the pretreated hazelnut skin showed higher Rm values (i.e., 31.67 and 35.14 mL CH₄/g VS·day at 150°C and 180°C, respectively) compared to the maximum biomethane production rate of the untreated LM, which was 21.93 mL CH₄/g VS·day.

Contrary to hazelnut skin and rice straw, the k₀ decreased from 0.190 1/day for the untreated cocoa bean shell to 0.184–0.185 1/day for the pretreated LM. Analogously, the Rm of the untreated cocoa bean shell (i.e., 31.30 mL CH₄/g VS·day) was higher compared to that of the pretreated LM (Table 3). This could indicate that an inhibition of the AD process occurred as a result of the organosolv pretreatment, as shown in Fig. 1c.

VFA production

VFAs are intermediates of the AD process, which are produced from the fermentation of carbohydrates, hydrolysis of long-chain fatty acids, and deamination of amino acids. VFAs are commonly used as an indicator of the system performance. A deeper understanding of the AD process can be obtained by monitoring the VFA concentration and speciation (Yeshanew et al., 2016). In this study, the trend of VFA production was followed during the first 10 days of each BMP test to assess the impact of the organosolv pretreatment on their accumulation (Fig. 2).

FIG. 2.

Evolution of VFA concentrations during first 10 days of AD of rice straw (a) untreated; (b) pretreated at 150°C; and (c) pretreated at 180°C; hazelnut skin (d) untreated; (e) pretreated at 150°C; (f) and pretreated at 180°C; and cocoa bean shell (g) untreated; (h) pretreated at 150°C; (i) and pretreated at 180°C. VFA, volatile fatty acids.

Pretreated rice straw (Fig. 2b, c) showed a higher VFA production than the untreated substrate (Fig. 2a). When the organosolv pretreatment was performed at 180°C, the total VFA concentration ranged from about 800–1,000 mg HAc/L from day 4–7. Subsequently, the VFA concentration dropped to 60 mg HAc/L on day 10, when more than 70% of the cumulative biomethane yield was produced.

Karthikeyan and Visvanathan (2013) reported that the biogas production might be inhibited when acetic acid accumulates above 2,000 mg/L and the overall VFA concentration exceeds 8,000 mg/L. Both the acetic acid and the total VFA concentration analyzed in this study remained constantly below this level (Fig. 2), probably entailing that no inhibition of the methanogenic stage occurred for any of the employed LMs.

A correlation between the VFA production and the hydrolysis rate has been reported for lignocelluloses in the literature (Mancini et al., 2018b) and can be observed for all the LMs used in this study (Fig. 2). In particular, the organosolv pretreatment slightly improved the hydrolysis of hazelnut skin at both pretreatment temperatures; thus, the VFA production of the pretreated hazelnut skin was moderately higher during the first 4 days of AD compared to that of the raw LM (Fig. 2d–f). Notwithstanding, the total VFA production from the untreated hazelnut skin was 120 mg/L on day 10 (Fig. 2d–f), whereas on the same day, it dropped to below the detection limit for the pretreated hazelnut skin, with consequent repercussions on the methane production obtained (Fig. 1). On the contrary, the total amount of VFA was lower in the first days of AD of the pretreated cocoa bean shell (Fig. 2g–i), especially at 150°C. This could indicate that the hydrolysis of cocoa bean shell was slightly hindered by the organosolv pretreatment.

Conclusion

Organosolv pretreatment was effective for rice straw, enhancing its biomethane production yield by 29–41%. These results were corroborated by the VFA and WRV analyses, which showed an increase of the VFA production and the accessible surface area by the pretreated rice straw. Compositional analyses showed that the organosolv pretreatment changed the chemical composition of the three LMs, causing a decrease of the lignin and hemicellulose and an increase of the cellulose content, which was more significant at the highest pretreatment temperature. A 10% increase of the biomethane production was obtained for the hazelnut skin pretreated at 180°C, whereas the pretreatment at 150°C did not enhance the cumulative biomethane yield, which remained at 261 mL CH₄/g VS. The BMP of the raw cocoa bean shell, that is, 231 mL CH₄/g VS, was negatively affected by organosolv pretreatment, most likely due to an inhibition of the hydrolysis step.

Footnotes

Acknowledgments

The authors would like to thank the European Commission for providing financial support through the Erasmus Mundus Joint Doctorate Programme ETeCoS³ (Environmental Technologies for Contaminated Solids, Soils, and Sediments) under the grant agreement FPA no. 2010-0009.

Author Disclosure Statement

No competing financial interests exist.

References

Chandra

R.P.

, Ewanick

S.M.

, Chung

P.A.

, Au-Yeung

, Del Rio

, Mabee

, and Saddler

J.N.

(2009). Comparison of methods to assess the enzyme accessibility and hydrolysis of pretreated lignocellulosic substrates. Biotechnol. Lett., 31, 1217.

Contreras

L.M.

, Schelle

, Sebrango

C.R.

, and Pereda

(2012). Methane potential and biodegradability of rice straw, rice husk and rice residues from the drying process. Water Sci. Technol., 65, 1142.

Croce

, Wei

, D'Imporzano

, Dong

, and Adani

(2016). Anaerobic digestion of straw and corn stover: The effect of biological process optimization and pre-treatment on total bio-methane yield and energy performance. Biotechnol. Adv., 34, 1289.

Deublein

, Steinhauser

. (2011). Biogas from Waste and Renewable Resources: an Introduction. 2nd ed. Weinheim: Wiley-VCH Verlag GmbH & Co. KGaA.

Doherty

W.O.S.

, Mousavioun

, and Fellows

C.M.

(2011). Value-adding to cellulosic ethanol: Lignin polymers. Ind. Crops Prod., 33, 259.

Esposito

, Frunzo

, Liotta

, Panico

, and Pirozzi

(2012). Bio-methane potential tests to measure the biogas production from the digestion and co-digestion of complex organic substrates. Open Environ. Eng. J., 5, 1.

Goshadrou

, Karimi

, and Lefsrud

(2013). Characterization of ionic liquid pretreated aspen wood using semi-quantitative methods for ethanol production. Carbohydr. Polym., 96, 440.

, Zhang

, and Zhou

(2015). Effect of Ca(OH)₂ pretreatment on extruded rice straw anaerobic digestion. Bioresour. Technol., 196, 116.

Hall

N.G.

, and Schönfeldt

H.C.

(2013). Total nitrogen vs. amino-acid profile as indicator of protein content of beef. Food Chem., 140, 608.

10.

Hendriks

A.T.

, and Zeeman

(2009). Pretreatments to enhance the digestibility of lignocellulosic biomass. Bioresour. Technol., 100, 10.

11.

Hesami

S.M.

, Zilouei

, Karimi

, and Asadinezhad

(2015). Enhanced biogas production from sunflower stalks using hydrothermal and organosolv pretreatment. Ind. Crops Prod., 76, 449.

12.

Kabir

M.M.

, Rajendran

, Taherzadeh

M.J.

, and Sárvári Horváth

(2015). Experimental and economical evaluation of bioconversion of forest residues to biogas using organosolv pretreatment. Bioresour. Technol., 178, 201.

13.

Karimi

, and Taherzadeh

M.J.

(2016). A critical review on analysis in pretreatment of lignocelluloses: Degree of polymerization, adsorption/desorption, and accessibility. Bioresour. Technol., 203, 348.

14.

Karimi

, Shafiei

, and Kumar

(2013). Progress in physical and chemical pretreatment of lignocellulosic biomass. In Gupta

V.K.

and Tuohy

M.G.

, Eds., Biofuel Technologies. Berlin Heidelberg: Springer-Verlag, p. 53.

15.

Karthikeyan

O.P.

, and Visvanathan

(2013). Bio-energy recovery from high-solid organic substrates by dry anaerobic bio-conversion processes: A review. Rev. Environ. Sci. Biotechnol., 12, 257.

16.

Krishania

, Vijay

V.K.

, and Chandra

(2013). Methane fermentation and kinetics of wheat straw pretreated substrates co-digested with cattle manure in batch assay. Energy, 57, 359.

17.

Lay

J.J.

, Li

Y.Y.

, and Noike

(1997). Influences of pH and moisture content on the methane production in high-solids sludge digestion. Water Res. 31, 1518.

18.

Liew

L.N.

, Shi

, and Li

(2012). Methane production from solid-state anaerobic digestion of lignocellulosic biomass. Biomass Bioenergy, 46, 125.

19.

Mancini

, Papirio

, Lens

P.N.L.

, and Esposito

(2016a). Solvent pretreatments of lignocellulosic materials to enhance biogas production: A review. Energy Fuels, 30, 1892.

20.

Mancini

, Papirio

, Lens

P.N.L.

, and Esposito

(2016b). Effect of N-methylmorpholine-N-oxide pretreatment on biogas production from rice straw, cocoa shell, and hazelnut skin. Environ. Eng. Sci., 33, 84.

21.

Mancini

, Papirio

, Lens

P.N.L.

, and Esposito

(2018a). Increased biogas production from wheat straw by chemical pretreatments. Renew. Energy, 119, 608.

22.

Mancini

, Papirio

, Riccardelli

, Lens

P.N.L.

, and Esposito

(2018b). Trace elements dosing Trace elements dosing and alkaline pretreatment in the anaerobic digestion of rice straw. Bioresour. Technol., 247, 897.

23.

Martínez

, Torres

, Meneses

, Figueroa

, Pérez-Álvarez

, and Viuda-Martos

(2012). Chemical, technological and in vitro antioxidant properties of cocoa (Theobroma cacao L.) co-products. Food Res. Int., 49, 39.

24.

McDonough

T.J.

(1992). The chemistry of organosolv delignification. Tappi J. 76, 186.

25.

Mesa

, Albernas

, Morales

, Corsano

, and González

(2016). Integration of organosolv process for biomass pretreatment in a biorefinery. In Mussatto

, Ed., Biomass Fractionation Technologies for a Lignocellulosic Feedstock Based Biorefinery. Amsterdam: Elsevier, p. 229.

26.

Mirmohamadsadeghi

, Karimi

, Zamani

, Amiri

, and Sárvári Horváth

(2014). Enhanced solid-state biogas production from lignocellulosic biomass by organosolv pretreatment. BioMed Res. Int., 2014, 1.

27.

Monlau

, Barakat

, Trably

, Dumas

, Steyer

J.-P.

, and Carrère

(2013). Lignocellulosic materials into biohydrogen and biomethane: Impact of structural features and pretreatment. Crit. Rev. Environ. Sci. Technol., 43, 260.

28.

Montgomery

D.C.

(2008). Design and Analysis of Experiments. New York: John Wiley & Sons.

29.

Murphy

J.D.

, and Power

(2009). Technical and economic analysis of biogas production in Ireland utilising three different crop rotations. Appl. Energy, 86, 25.

30.

Mussoline

, Esposito

, Giordano

, and Lens

(2013). The anaerobic digestion of rice straw: A review. Crit. Rev. Environ. Sci. Technol., 43, 895.

31.

Ostovareh

, Karimi

, and Zamani

(2015). Efficient conversion of sweet sorghum stalks to biogas and ethanol using organosolv pretreatment. Ind. Crops Prod., 66, 170.

32.

Pansu

, and Gautheyrou

(2007). Handbook of Soil Analysis: Mineralogical, Organic and Inorganic Methods. Berlin Heidelberg: Springer-Verlag,

33.

Park

, Kim

H.-Y.

, Koo

B.-W.

, Yeo

, and Choi

I.-G.

(2010). Organosolv pretreatment with various catalysts for enhancing enzymatic hydrolysis of pitch pine (Pinus rigida). Bioresour. Technol., 101, 7046.

34.

Piccinelli

A.L.

, Pagano

, Esposito

, Mencherini

, Porta

, Petrone

A.M.

, Gazzerro

, Picerno

, Sansone

, and Rastrelli

(2016). HRMS profile of a hazelnut skin proanthocyanidin-rich fraction with antioxidant and anti-Candida albicans activities. J. Agric. Food Chem., 64, 585.

35.

Redgwell

, Trovato

, Merinat

, Curti

, Hediger

, and Manez

(2003). Dietary fibre in cocoa shell: Characterisation of component polysaccharides. Food Chem. 81, 103.

36.

Sawatdeenarunat

, Surendra

, Takara

, Oechsner

, and Khanal

S.K.

(2015). Anaerobic digestion of lignocellulosic biomass: Challenges and opportunities. Bioresour. Technol., 178, 178.

37.

Shafiei

, Karimi

, Zilouei

, and Taherzadeh

M.J.

(2014). Enhanced ethanol and biogas production from pinewood by NMMO pretreatment and detailed biomass analysis. BioMed Res. Int., 2014, 1.

38.

Shana

, Ouki

, Asaadi

, Pearce

, and Mancini

(2013). The impact of intermediate thermal hydrolysis on the degradation kinetics of carbohydrates in sewage sludge. Bioresour. Technol., 137, 239.

39.

Sluiter

, Hames

, Ruiz

, Scarlata

, Sluiter

, Templeton

, and Crocker

(2008a). Determination of structural carbohydrates and lignin in biomass. National Renewable Energy Laboratory, NREL Technical Report No. NREL/TP-510-42618., 1

40.

Sluiter

, Hames

, Hyman

, Payne

, Ruiz

, Scarlata

, Sluiter

, Templeton

, and Wolfe

(2008b). Determination of total solids in biomass and total dissolved solids in liquid process samples. National Renewable Energy Laboratory, NREL Technical Report No. NREL/TP-510-42621, 1.

41.

Stronach

S.M.

, Rudd

, and Lester

J.N.

(1986). The biochemistry of anaerobic digestion. In Stronach

S.M.

, Rudd

and Lester

J.N.

, Eds., Anaerobic Digestion Processes in Industrial Wastewater Treatment. Berlin Heidelberg: Springer Berlin Heidelberg, p. 1; pp. 1–20.

42.

Sun

, and Cheng

(2002). Hydrolysis of lignocellulosic materials for ethanol production: A review. Bioresour. Technol., 83, 1.

43.

Taherzadeh

M.J.

, and Karimi

(2008). Pretreatment of lignocellulosic wastes to improve ethanol and biogas production: A review. Int. J. Mol. Sci., 9, 1621.

44.

Teghammar

, Chandra

, Saddler

J.N.

, Taherzadeh

M.J.

, and Sárvári Horváth

(2012). Substrate characteristics analysis for anaerobic digestion: A study on rice and triticale straw. Bioresources. 7, 3921.

45.

Yeshanew

M.M.

, Frunzo

, Lens

P.N.

, Pirozzi

, and Esposito

(2016). Mass loss controlled thermal pretreatment system to assess the effects of pretreatment temperature on organic matter solubilization and methane yield from food waste. Front. Environ. Sci., 4, 62.

46.

Zeppa

, Belviso

, Bertolino

, Cavallero

M.C.

, Dal Bello

, Ghirardello

, Giordano

, Giorgis

, Grosso

, and Rolle

(2015). The effect of hazelnut roasted skin from different cultivars on the quality attributes, polyphenol content and texture of fresh egg pasta. J. Sci. Food Agric., 95, 1678.

47.

Zhao

, Cheng

, and Liu

(2009). Organosolv pretreatment of lignocellulosic biomass for enzymatic hydrolysis. Appl. Microbiol. Biotechnol., 82, 815.