Accepted Manuscript

Energy conversion of biomass crops and agroindustrial residues by combined

biohydrogen/biomethane system and anaerobic digestion

Elisa Corneli, Federico Dragoni, Alessandra Adessi, Roberto De Philippis,

Enrico Bonari, Giorgio Ragaglini

PII: S0960-8524(16)30437-0
DOI: http://dx.doi.org/10.1016/j.biortech.2016.03.134
Reference: BITE 16324

To appear in: Bioresource Technology

Received Date: 29 January 2016

Revised Date: 22 March 2016
Accepted Date: 25 March 2016

Please cite this article as: Corneli, E., Dragoni, F., Adessi, A., De Philippis, R., Bonari, E., Ragaglini, G., Energy
conversion of biomass crops and agroindustrial residues by combined biohydrogen/biomethane system and
anaerobic digestion, Bioresource Technology (2016), doi: http://dx.doi.org/10.1016/j.biortech.2016.03.134

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Title

Energy conversion of biomass crops and agroindustrial residues by combined

biohydrogen/biomethane system and anaerobic digestion

Authors

Elisa Cornelia,*, Federico Dragonia, Alessandra Adessib, Roberto De Philippis b,c,

Enrico Bonaria,d, Giorgio Ragaglinia, d

Affiliations
a
Institute of Life Sciences, Scuola Superiore Sant’Anna, Piazza Martiri della Libertà

33, 56127 Pisa, Italy

b
Institute of Chemistry of Organometallic Compounds (ICCOM), CNR, Florence, Italy
c
Department of Agrifood Production and Environmental Sciences, University of

Florence, Italy
d
CRIBE – Centro di Ricerche Interuniversitario Biomasse da Energia, Via Vecchia

Livornese 748, 56122 Pisa, Italy

Abstract

Aim of this study was to evaluate the suitability of ensiled giant reed, ensiled maize,

ensiled olive pomace, wheat bran for combined systems (CS: dark fermentation +

anaerobic digestion (AD)) producing hydrogen-rich biogas (biohythane), tested in batch

under basic operational conditions (mesophilic temperatures, no pH control). Substrates

were also analyzed under a single stage AD batch test, in order to investigate the effects

of DF on estimated energy recovery (ER) in combined systems. In CS, maize and wheat

bran exhibited the highest hydrogen potential (13.8 and 18.9 NL kgVS-1) and wheat

bran the highest methane potential (243.5 NL kgVS-1). In one-stage AD, giant reed,

maize and wheat bran showed the highest methane production (239.5, 267.3 and 260.0

* Corresponding author: E-mail address: e.corneli@sssup.it

NL kgVS-1). Butyrate/acetate ratio properly described the dark fermentation, correlating

with hydrogen production (r=0.92). Wheat bran proved to be a promising residue for CS

in terms of hydrogen/methane potential and ER.

Keywords: maize, giant reed, olive pomace, wheat bran, dark fermentation.

Abbreviations: eGR: ensiled biomass of Arundo donax L.; eZM ensiled biomass of Zea

mays L.; eOP: ensiled olive pomace; WB: wheat bran; CS: combined system; AD:

anaerobic digestion; DF: dark fermentation; CS_DF: first stage of combined system;

CS_AD second stage of combined system; ER: energy recovery; Rmax: maximum daily

production rate; T50: time, expressed in days, when 50% of methane production was

reached; TS: total solids; VS: volatile solids; WSC: Water Soluble Carbohydrates;

VFAs: short chain volatile fatty acids; BBP: Biochemical Biogas Potential; BMP:

Biochemical Methane Potential; BHP: Biochemical Hydrogen Potential; MC: Methane

Concentration; HC: Hydrogen Concentration; NDF: Neutral Detergent Fiber; ADF:

Acid Detergent Fiber; ADL: Acid Detergent Lignin.

1 Introduction

Anaerobic digestion (AD) represents a mature technology for biofuels production,

alternative to fossil fuels (Weiland, 2010). In AD, biomethane is the main metabolite

produced and it can be used as biogas for power and heat production. Otherwise

upgraded biomethane can be used as gaseous vehicle fuel or injected into natural gas

grids, replacing natural gas (Weiland, 2010; Adessi et al., 2012). AD fermentation can

be divided into two main stages (Weiland, 2010; Adessi et al., 2012). In the first stage,

the hydrolysis of complex organic substrates and their acidogenic fermentation occur,

producing hydrogen, carbon dioxide and organic end products, like short chain volatile

fatty acids (VFAs), lactic acid and ethanol (Weiland, 2010; Adessi et al., 2012; Ghimire

2
et al., 2015a). Then, in the second stage, acetogenesis and methanation of organic

products take place and main final products are methane (50-70%) and carbon dioxide

(30-50%) (Weiland, 2010). An interesting opportunity for AD process is represented by

the spatial separation into a two-stage system, i.e. Combined System (CS), composed of

a first stage, named dark fermentation (DF), where hydrogen is obtained, and a second

one that is mostly methanogenic (Argun and Kargi, 2011; Adessi et al., 2012; Ghimire

et al., 2015a). Considering the utilization of CS biogas, it can provide benefit in the

form of increased combustion efficiency of the mixed biofuel, hydrogen and methane,

known as biohythane. It improves engine performances, reducing nitrogen oxide and

CO2 emissions, as compared to sole CH4 combustion (Hallenbeck et al., 2009; Cavinato

et al., 2012). In CS, hydrogen production of dark fermentation (CS_DF) from

heterotrophic bacteria is enhanced and assessed and the hydrolysis of biomass can be

improved compared to one-stage AD (Hawkes et al., 2007; Giordano et al., 2011), thus

optimizing the following second stage of anaerobic digestion (CS_AD), in which

biomethane is produced (Adessi et al., 2012; Ghimire et al., 2015a).

In CS_DF, the hydrogenase enzyme can reduce protons to hydrogen, neutralizing the

electrons coming from the organic compounds oxidation (Ghimire et al., 2015a). Since

the main fermentation end products of CS_DF are acetic acid and butyric acid, with

butyric acid in excess of acetic acid, the process is also called acetate/butyrate-type

fermentation and these acids allow the highest theoretical conversion of glucose to H2

comparing with the other acids. In particular, the conversion of 1 mol of glucose to

acetic acid involves the production of 4 mol of H2 and the conversion to butyric acid of

2 mol of H2 (Hawkes et al., 2007; Abo-Hashesh and Hallenbeck, 2012; Ghimire et al.,

2015a). Moreover, since the inoculum is a mixed non-sterile culture, like biogas plant

sludge, a pre-treatment, as heat shock, acid or alkaline treatment and sonication, is

3
recommended in order to reduce the activity of H2 consumer microorganisms, like

homoacetogens, methanogens, nitrate and sulfate reducing bacteria, which are in

syntrophic association with H2 producers (Argun and Kargi, 2011; Ghimire et al.,

2015a). The use of non-sterile mixed cultures is generally preferred over a pure culture

since it is easily available, low cost, not requiring aseptic environment and it allows

robust process, taking advantage of syntrophic relations among microbial communities

(Ghimire et al., 2015a).

Temperature and pH are important factors influencing DF efficiency (Guo et al., 2010;

Abo-Hashesh and Hallenbeck, 2012). Thermophilic temperatures generally result in

higher H2 yields, since they facilitate the hydrolysis of lignocellulosic component and

speed up the process. The correction of the initial pH values to 5-7 would increase the

performances of DF, since in mixed cultures it increases the activity of hydrogen

producing bacteria, while inhibiting hydrogen consuming microorganisms (Guo et al.,

2010; Argun e Kargi, 2011; Abo-Hashesh and Hallenbeck, 2012).

At plant scale, a combined system composed by a first stage of dark fermentation and

the second stage of anaerobic digestion could be feasible based on conventional two-

stage biogas plants (Schievano et al., 2014). However, with the view of developing rural

economies, farmers and biogas plant owners would take advantage of using simple

operational conditions based on the most widespread technology nowadays. Thus,

features such as thermophilic temperature and pH correction, which can advance the

process, could imply additional costs. The use of local and low cost substrates, suitable

for H2 production in a conventional biogas-like plant could be a feasible approach to

this technology, provided that these substrates are actually suitable for conversion into

hydrogen and methane. In particular, it has been reported that substrates for dark

fermentation are preferably carbohydrate-rich materials, characterized by high sugar and

4
low lignin contents, since generally, the higher is the biodegradability of the substrate,

the higher is the H2 production rate (Ghimire et al., 2015a).

Biomass crops and residues of the agroindustry are promising substrates for

fermentation systems (Ghimire et al., 2015a). Giant reed (Arundo donax L.), maize (Zea

mays L.), olive pomace and wheat bran were selected as vegetable biomasses for the

present work, since they are still poorly studied in biohydrogen/biomethane CS. To the

best knowledge of the authors, this is the first time that ensiled giant reed and ensiled

olive pomace are tested in a two-stage batch CS. Giant reed is an interesting perennial

rhizomatous grass, which represents a promising feedstock for bioenergy supply chains,

like AD (Corno et al., 2014; Ragaglini et al., 2014; Dragoni et al., 2015). It is a low

input crop with high biomass production and great adaptability to marginal land, thus

potentially reducing the competition with food production (Corno et al., 2014). On the

other hand, another opportunity consists in the use of agro-industrial residues that are

abundant, economic, renewable and biodegradable biomasses (Ghimire et al., 2015b).

Olive pomace and wheat bran were selected being widespread residues in the

Mediterranean area. They come from olive oil and wheat milling industry, respectively.

Italy is the second biggest EU producer of olive oil and thus it is the second EU

producer of olive pomace, with an annual output of 2.5-3 million tons (Mantovi et al.,

2013). Olive pomace is a substrate rich in fiber components, lipids and proteins. In

order to use this substrate for biogas production, the concentration of lignin-rich parts,

mainly represented by pits, should be reduced and the pretreatment indicated is pitting

(Mantovi et al., 2013; Mateo and Maicas, 2015). Wheat bran is a matrix consisting of

structural carbohydrates, starch and proteins (Pan et al., 2008) and in Italy it has an

annual output of 1.1- 1.4 million tons (ISTAT, 2012), considering that wheat bran

accounts about 14-19% of the grains’ weight (Pan et al., 2008). In this study, maize was

5
considered as a widespread bioenergy crop, which is an annual food species widely

distributed, requiring relevant input and high fertile soils (Amon et al., 2007; Dragoni et

al., 2015). One of the most important issues representing a key challenge for the

bioenergy use of biomasses is the storage, and ensiling is a widespread technology for

storing wet feedstocks with total solid content ranging from 25 to 35% (Weiland, 2010;

Liu et al., 2015).

The aim of the present work was to investigate the suitability of biomasses that are

relevant in the Mediterranean area (ensiled giant reed, ensiled maize, ensiled olive

pomace, wheat bran) in a batch dark fermentation system, in order to evaluate their

hydrogen potential. Aiming to compare the considered substrates under basic

operational conditions, mesophilic temperatures were tested, without pH control. Dark

fermentation test was then followed by a batch second stage of anaerobic digestion,

forming a combined system, in order to assess the residual potential of methane

production and the overall energy recovery of the combined system. The same

substrates were also analyzed under a single stage AD batch test. The two processes (CS

and one-stage AD) were then compared in terms of energy recovery from each different

biomass.

2 Materials and methods

2.1 Feedstocks and ensiling

Four different substrates were chosen for assessing the production of H2 and CH4 in the

two-stage CS: ensiled biomass of A. donax (eGR), ensiled biomass of Z. mays (eZM),

ensiled olive pomace (eOP) and wheat bran (WB). The CS, composed by a first stage of

hydrogenic dark fermentation (CS_DF) and a second stage of methanogenic anaerobic

digestion (CS_AD), was compared with a one-stage AD process producing CH4.

6
Three samples of aboveground biomass were harvested from giant reed and maize

grown in experimental fields located in San Piero a Grado, Pisa, Italy (43° 40’ 49.21’’

North, 10° 20’ 47.15’’ East; 1 m above mean sea level and 0% slope). Giant reed was

cut in July 2013, before crop senescence, since juvenile traits of the crop have been

positively correlated with Biochemical Methane Potential (BMP) (Ragaglini et al.,

2014). Maize was harvested in September 2013, at milky-wax ripeness. Entire plants of

giant reed and maize were chopped in an electric power shredder AL-KO H1600 (AL-

KO Maschinenfabrik GmbH, Obdach, Austria). Pitted olive pomace and wheat bran

were collected from a three-phase olive processing plant and from a wheat milling

industry, respectively. The biomass of giant reed, maize and olive pomace was

separately ensiled in triplicate at lab scale in plastic buckets (5 L), while wheat bran,

having a low moisture content, did not need to be ensiled for storage. The ensiled

material was treated with a commercial bacterial ensiling inoculant, Pioneer® 11CH4,

according to manufacturer guidelines (Pioneer Hi-Bred Italia Srl, Cremona, Italy). This

inoculant contains Lactobacillus buchneri LN40177, which is an obligate

heterofermentative bacterium, used to enhance the aerobic stability in silages; it

decreases the concentration of lactic acid and increases acetic acid (Driehuis et al.,

1999) in order to make the substrate more suitable for AD. Anaerobic conditions were

obtained by pressing the biomass into the buckets and using silicone to seal the

connection between the lid and the bucket. The dispersion of gas produced was

controlled through the application of bubblers in water, allowing to prevent the

penetration of air into the silage. Buckets were stored at room temperature for 60 days.

2.2 Inocula

Inocula were collected from a two-stage mesophilic commercial biogas plant, fed with

energy crops, agroindustrial residues, cattle and poultry manure. The CS_DF inoculum

7
was collected from the first stage of the plant; it was sieved and heat-treated for 1 hour

at 100 °C in oven, in order to enrich the acidogenic microflora with spore-forming H2

producers and to inhibit microorganisms that consume H2 (Argun and Kargi, 2011). The

total solids (TS) content of CS_DF inoculum after pretreatment was 119.6 g kg-1 and the

volatile solids (VS) content was 91.1 g kg-1. Methanogenic inoculum used both for

CS_AD and for AD was collected from the second stage of the biogas plant. In this case

the inoculum was sieved and incubated at 37°C for five days long in order to deplete the

biodegradable organic matter still present in it (Ragaglini et al., 2014). The

methanogenic inoculum had a TS content of 102.6 g kg-1 and a VS content of 75.8 g kg-
1
.

2.3 Batch experiments

Batch assays were established in order to perform both the CS and the AD of the four

substrates (eGR, eZM, eOP, WB). The experimental biogas plant was composed by

static batch reactors (2 L) operating in mesophilic conditions (37 ± 1 °C). Temperature

(Pt100) and pressure (piezoresistive transducers) were automatically and continuously

measured and recorded every 3 minutes by a Programmable Logic Controller (PLC)

connected to a PC, as described by Ragaglini et al. (2014). During the assays, three

replicates of each inoculum were digested alone, as blank experiments.

In the CS_DF assay, each reactor was filled with biomass and 100 g of inoculum with a

substrate to inoculum ratio of 2:1 based on VS and ultrapure water was added up to a

volume of 650 mL. Thus, reactors were sealed and flushed with N2 in order to substitute

the air in the headspace, allowing to obtain anaerobic conditions and to reduce the

partial pressure of hydrogen (pH2) (Abo-Hashesh and Hallenbeck, 2012). CS_DF batch

assay was realized in triplicates for each biomass and was kept for four days, about 10

times shorter retention time compared to the CS_AD (Pakarinen et al., 2008; Schievano

8
et al., 2014). The assay was made without any pH control, allowing the pH to be

dependent on the process natural conditions. At the end, samples of CS_DF effluents

(200 g) were collected from each replicate for chemical analysis. Then, methanogenic

inoculum was added to each replicate up to the starting volume of 650 mL (Pakarinen et

al., 2008) and reactors were closed and flushed again with N2, to allow anaerobic

conditions. The CS_AD was kept for 40 days.

Conventional AD assay was carried out in triplicate for each biomass by adding in each

reactor the biomass and 200 g of inoculum with a substrate to inoculum ratio of 2:1

(VS/VS) and ultrapure water was added up to a volume of 650 mL. Thus, reactors were

sealed and flushed with N2 as in the combined system assay. The entire process had a

retention time of 40 days.

2.4 Biogas parameters

The Biochemical Biogas Potential (BBP) was calculated according to the Ideal Gas Law

and to the Molar Volume of Ideal Gases at Standard Temperature and Pressure

conditions (1 bar, 273.15 K). The composition of biogas was measured at discrete

intervals by gas chromatography (micro-GC Agilent 3000, Agilent Technologies, Inc.,

Shangai, China): once at the end of CS_DF (the fourth day); four times for both the

CS_AD and the AD, more frequently at the beginning of the fermentation to better

assess the kinetics (at time intervals of 8, 15, 29 and 40 days after the beginning of the

assay). Both the pressure reduction due to biogas removal at each sampling time and the

biogas composition of the sampled gas were considered for estimating the cumulative

methane production in each batch (Ragaglini et al, 2014).

Then, the Biochemical Methane Potential (BMP) and the Biochemical Hydrogen

Potential (BHP) of the substrates, as well as the Methane Concentration (MC) and the

Hydrogen Concentration (HC) of the biogas, were determined. BBP, BMP, BHP of

9
inocula were subtracted from the biogas production of the biomasses tested, as reported

by Ragaglini et al. (2014). Moreover, the Energy Recovery (ER) of biomasses was

determined according to the superior heat of combustion of methane (35.16 MJ Nm-3)

and hydrogen (12.74 MJ Nm-3), with the view to compare the two-stage combined

system with the one-stage anaerobic digestion (Schievano et al., 2014).

Kinetics of CS_AD and of AD were examined by regression on time of the daily-

cumulated methane measured in each reactor, using a five parameters Modified

Gompertz function (Ragaglini et al., 2014 and references therein):

y = A * exp[Mr/Dr * exp(-Dr * x) – Ms/Ds * exp(-Ds * x)] (1)

where: A represents the upper limit; M the initial relative growth rate; D the relative

growth rate at inflection; r the rapid initial phase; s the slow final phase.

Goodness of fit was assessed for each substrate, by means of R2adj and Root Mean

Square Error (RMSE). Then, the function was used to calculate the maximum daily

production rate (Rmax) through the analysis of model derivative, and the time, expressed

in days, when 50% of methane production was reached (T50). Curve fitting and model

parameterization were performed using the R software, version 3.0.1, and the nmle

package. Details of the model parameterization and the goodness of fit are reported in

Appendix.

2.5 Chemical analysis of substrates

Samples for chemical analyses were prepared by milling the biomass of different

substrates in a Retsch SM1 rotor mill (Retsch, Düsseldorf, Germany) equipped with a 1

mm grid. TS and VS were determined in triplicate, according to standard methods

(Ragaglini et al., 2014 and references therein). TS were estimated by drying the material

at 105°C to constant weight of the sample and VS were estimated by determining the

ash content of dry samples in a muffle furnace at 550°C. TS and VS values of silages

10
were corrected according to volatilization coefficients of lactic acid, total VFAs and

total alcohols (0.375, 0.892, 0.975, respectively) as reported by Porter and Murray

(2001). Elemental analysis was carried out with the LECO CHN Elemental TruSpec

Determinator (LECO Corporation, St. Joseph, Michigan, USA), in order to determine

total nitrogen and C/N ratio. Neutral Detergent Fiber (NDF), Acid Detergent Fiber

(ADF) and Acid Detergent Lignin (ADL) were determined with Van Soest method

using the FiberCap ™ 2021/2023 system (FOSS Analytical AB, Höganäs, Sweden),

while non-structural carbohydrates, Water Soluble Carbohydrates (WSCs) and starch,

were determined as described in Ragaglini et al. (2014) and references therein.

Chemical characterization of substrates is showed in Table 1.

Further analyses, specific for the three ensiled substrates, were made. pH value was

determined with a CyberScan pH 110 pH meter (Eutech Instruments, Singapore).

Concentration of ethanol, lactic acid and VFAs (acetic acid, propionic acid, butyric acid,

isovaleric acid and valeric acid) was determined according to Bianchi et al. (2010).

Volatile components characterization of silages is reported in Table 2.

2.6 Chemical analysis of CS_DF effluents, CS_AD digestates and AD digestates

The pH value of CS_DF effluents and digestates was obtained with a CyberScan pH

110 pH meter (Eutech Instruments, Singapore). The concentration of ethanol, lactic acid

and VFAs was determined with HPLC chromatograph as described in Bianchi et al.

(2010) and the ammonium concentration was determined with Nessler method (Bianchi

et al., 2010). The molar ratio between butyrate and acetate (B/A) present in each

CS_DF effluents was calculated as a quantitative indicator of substrate metabolism and

H2 production by anaerobic microflora during CS_DF (Kim et al., 2006; Hawkes et al.,

2007). The characterization of CS_DF effluents, CS_AD digestates and AD digestates

is reported in Table 3, Table 4 and Table 5, respectively.

11
2.7 Statistical analysis

Biogas parameters (BBP, BMP, BHP, MC and HC) of different substrates were

analyzed one by one with one-way analyses of variance (ANOVA), considering the

substrate as independent variable. For each substrate, the ER of CS and AD and the

kinetic parameters (Rmax and T50) of CS_AD and AD were assessed one by one with

two-way ANOVA, considering the fermentation system and the substrate as factors.

The difference between each pair of means were determined by honest significant

difference test (Tukey’s test). Pearson correlation coefficient (r) was determined in

order to assess the relationship between two parameters, B/A and BHP in CS_DF.

Statistical analyses were performed with the 3.0.1 R software.

3 Results and discussion

3.1 Effect of substrate on dark fermentation and on second stage of anaerobic

digestion

Under the CS_DF process, WB showed the highest BBP (88.4 ± 20.6 NL kgVS-1),

comparable with eGR and eZM (62.5 ± 4.3 and 63.4 NL ± 1.8 kgVS-1, respectively) and

significantly higher than eOP (41.6 ± 1.1 NL kgVS-1) (P < 0.01). In general, eZM, WB

and eGR reached a plateau in biogas production, while eOP showed an early stage of

cumulative biogas production, not reaching the plateau in 4 days of CS_DF (Fig. 1).

Significant differences were observed for hydrogen production, in fact eZM and WB

showed higher values of BHP (13.8 and 18.9 NL kgVS-1 respectively), than eGR and

eOP (< 0.1 NL kgVS-1). HC in eZM and WB, 21.7 and 21.1% respectively, was

significantly higher than in eGR and eOP, where it was less than 0.2 %. On the

contrary, eGR and eOP showed the highest BMP, 19.8 and 20.7 NL kgVS-1

respectively, with a MC of 31.7 and 49.6 %, respectively. Those values were

12
significantly higher than those of eZM and WB (3.0 and 2.3 %, respectively), that

showed a BMP less than 2.0 NL kgVS-1 (Fig. 2).

Concerning the CS_AD, WB produced the highest value of BBP (387.7 NL kgVS-1) and

eOP produced the lowest value (204.1 NL kgVS-1), while eGR and eZM were

comparable, with BBP equal to 347.4 and 357.7 NL kgVS-1, respectively. WB showed

the highest BMP value (243.5 NL kgVS-1), while eZM, eGR and eOP showed lower

results (216.7, 195.7 and 136.6 NL kgVS-1 respectively). eOP had the highest MC (67.3

%), while eZM and WB had similar values, equal to 60.7 and 63.3 % respectively, and

eGR showed the lowest MC (56.5 %) (Fig. 3).

In relation to the CS_DF process, the two better performing substrates in hydrogen

production, eZM and WB, had the lowest methane potential and they reached the

plateau for biogas cumulative production, showing that 4 days of dark fermentation

were enough to reach the maximum hydrogen potential of these biomasses in the

present experimental conditions. Regarding wheat bran, BHP of the present study was

in line with the results obtained by Noike and Mizuno (2000) in a DF batch assay, at

35°C in shaking conditions and using fermented soybean meal as inoculum (10-43 NL

kgVS-1). On the other hand, Pan et al. (2008) reported a maximum BHP value of 128.2

NL kgVS-1, using a pretreated wheat bran in mesophilic mixed batch test with activated

sludge of paper mill, as inoculum, and with initial correction of pH value at 5.0. Also

Giordano et al. (2011) assessed wheat bran performance in a lab scale CS, reporting

higher BHP for both common wheat and durum wheat (53 and 90 NL kgVS-1

respectively), using as inoculum a granular mesophilic sludge from a biogas plant in a

mesophilic batch test with the initial correction of pH at 7. Schievano et al. (2014)

carried out the CS_DF of maize silage and olive pomace in semicontinuously operated

reactors, under thermophilic conditions with liquid swine manure as inoculum,

13
assessing nine different combinations of organic matter concentration and hydraulic

retention time: regarding maize silage, their BHP ranged from 12.5 to 120.9 NL kgVS-1,

according to the different operating conditions, while olive pomace showed an

hydrogen potential ranging from 0.1 to 13.0 NL kgVS-1. Another study, analyzing the

DF of olive pomace in a mesophilic batch, with a chemical-treated sludge of an AD

plant as inoculum and a retention time of 6 days, reported a higher BHP value, equal to

54.9 NL kgVS-1 (Ghimire et al., 2015b). Furthermore, concerning the retention time of

olive pomace in batch tests, Borja et al. (2005) analyzed the anaerobic acidogenic

fermentation of olive pomace at lab scale in mesophilic mixed batch, highlighting that a

retention time of 12 days is the optimum in the view of obtaining the maximum

production of acetate and butyrate. Thus, it could be hypothesized that a retention time

longer than 4 days would be necessary for olive pomace to achieve a higher hydrogen

production. Concerning eGR, Guo et al. (2014) reported the BHP of giant reed leaves

and stems separately, giving values of 25.1 and 34.2 NL kgTS-1 respectively, assessed

in a mesophilic batch test, using as inoculum a sludge from a semi-industrial anaerobic

digester treating vinasses. However, the study of Guo et al. (2014) differs from the

present one, since the giant reed biomass was not ensiled and the harvest time was not

indicated.

After four days of DF, no CS_DF effluent had pH values lower than 5, which is the

minimum threshold that most heterotrophic bacteria can tolerate (Abo-Hashesh and

Hallenbeck, 2012). In general, it has been reported that substrates rich in carbohydrates

have higher hydrogen production, because they imply acidification of the medium,

sustaining optimal conditions for the hydrolysis of biomasses and for the inhibition of

hydrogen consumers (Ghimire et al., 2015b). This is in accordance with this study,

where the two substrates showing higher BHP, eZM and WB, had a Non-Structural

14
Carbohydrates concentration (NSCs = starch + WSCs) equal to 12.4 % and 8.6 %,

respectively, and probably this implied an effective acidification of the medium,

resulting pH values of the final CS_DF effluents equal to 5.12 and 5.66, respectively.

On the other hand, NSCs in eGR was 1.7 % while in eOP starch was not detected and

WSCs was 0.1 % (Table 1). eOP did not produce hydrogen, while it showed a

continuously increasing biogas production during the 4 days of the trial and a final pH

value of 7.62, above the inhibitory threshold of methanogenic archaea (Weiland, 2010).

Therefore, chemical characteristics of eOP, under the tested conditions, did not allow

the development of a hydrogenogenic stage. Analogously, in eGR H2 production did not

occur, while the trend of biogas production and the final pH, equal to 5.35, suggested

that methanogenesis developed until it was inhibited by acidic conditions. Probably the

low performance of giant reed in CS_DF can also be related both to the early harvest

and to the ensiling process. In fact, the crop at early stages generally has lower NSCs

content than the mature crop (Ragaglini et al., 2014) and NSCs are metabolized to

VFAs during ensiling (Liu et al., 2015). Acetate concentration (3.37 g L-1) in CS_DF

effluents of eGR was higher than in the other substrates, even if this substrate had low

BHP. In fact, high concentrations of acetate do not imply a high H2 production if the

homoacetogenesis is not completely inhibited in DF batch tests (Guo et al., 2014).

According to the present results, the higher is the butyrate to acetate ratio (B/A), the

higher is the BHP. This observation is in line with the results reported by Ghimire et al.

(2015b). In fact, eZM and WB showed the highest B/A, 3.4 and 3.5 respectively, while

eGR and eOP were less than 0.4 (Table 3). Kim et al. (2006) found that a 2.6 is a B/A

threshold value beyond which the production of H2 is enhanced. Actually, B/A and BHP

were significantly and positively correlated, with a Pearson’s r value equal to 0.92 (P <

0.001). Lactic acid and ethanol were not detected in any of CS_DF effluents.

15
Ammonium concentrations of CS_DF effluents varied from a minimum of 0.35 g L-1 in

eOP to a maximum of 0.66 g L-1 in WB (Table 3) and resulted to be a not inhibiting

value for the methanogenic activity in the following CS_AD stage (Chen et al., 2008).

By the literature, Schievano et al. (2014) reported BMP of maize silage in a CS_AD

stage ranging from 304 to 504 NL kgVS-1, which is higher than BMP of this study

(216.7 NL kgVS-1), while they did not obtain methane from olive pomace. Regarding

wheat bran, Giordano et al. (2011) reported BMP of CS_AD ranging from 230 to 289

NL kgVS-1, that is comparable to the value obtained in this study (243.5 NL kgVS-1).

To the best knowledge of the authors, the investigation in a combined system of ensiled

olive pomace and ensiled giant reed have not yet been assessed. At the end of CS_AD

batch tests, the total VFAs concentration in digestates (acetate, propionate and butyrate)

was minimum in eZM (0.08 g L-1) and higher in the other substrates, varying from 1.05

g L-1 in eGR to 1.42 g L-1 in eOP. Ammonium analyses showed concentrations from a

minimum of 1.08 g L-1 in eGR to a maximum of 1.32 g L-1 in WB, that were consistent

with not inhibitory values for methanogenic activity (Schievano et al., 2012; Corno et

al., 2015). pH was similar for all digestates, comprised in a range from 8.27 to 8.44, that

is above the minimum threshold of methanogenic activity inhibition (7.0) (Weiland,

2010).

3.2 Effect of substrate on anaerobic digestion

Biogas parameters of one-stage AD were analyzed. eGR, eZM and WB showed BBP

values (341.5, 394.4 and 377.4 NL kgVS-1, respectively) significantly higher than eOP

(206.7 NL kgVS-1). Similar results were obtained for BMP, since eGR, eZM and WB

showed significantly higher values (239.5, 267.3 and 260.0 NL kgVS-1, respectively)

than eOP (172.0 NL kgVS-1). eOP had MC (83.0 %) higher than the other substrates,

i.e. eGR, eZM and WB (70.2, 67.7 and 68.9 %, respectively) (Fig. 4). Compared with

16
the literature, MC values were quite high for all substrates. In eZM and WB, this could

be related to a relatively low content of starch (Table 1) (Amon et al., 2007; Pan et al.,

2008), while in eOP it could be related to the lipid content, typically higher than in the

other substrates (Weiland, 2010; Mateo and Maicas, 2015). By the literature, few

studies analyzed the BMP of giant reed harvested in summer. Ragaglini et al. (2014)

reported a BMP value of fresh giant reed harvested in July of 325.2 NL kgVS-1, while

Dragoni et al. (2015) reported a similar value (323.0 NL kgVS-1) for ensiled giant reed

cut at the same time. Both these values are higher than what observed in the present

study. However, the methane potential of giant reed of this study is higher than those

reported by Liu et al. (2015) and Corno et al. (2015), 160 and 185 NL kgVS-1

respectively. The latter result was obtained in a continually stirred tank reactor (CSTR)

fed with ensiled giant reed, while Liu et al. (2015) analyzed the BMP of ensiled giant

reed in a mesophilic batch test. In both cases the harvest time of the crop was not

indicated. Methane potential of maize silage of the present study is consistent with BMP

reported in the literature by Corno et al. (2015) (261 NL kgVS-1). Higher BMP values

are reported in the literature. Amon et al. (2007) obtained BMPs ranging from 312 to

365 NL kgVS-1, in a mesophilic batch test using four different ensiled maize varieties

harvested at milk ripeness, and Schievano et al. (2014) obtained values ranging from

379 to 431 NL kgVS-1, in a semi-continuously operated reactor test, in which olive

pomace resulted unsuitable for AD. Concerning olive pomace, Fantozzi and Buratti

(2009) reported a BMP of 110 NL kgVS-1 using a mixed mesophilic batch test, which is

a lower value than the methane potential of the present study. Another study (Alagöz et

al., 2015) stated a BMP value of shredded olive pomace in mesophilic fed batch reactor,

equal to 180 NL kgVS-1, which is similar to the value obtained in this study. Finally,

Kolbl et al. (2014), comparing the BMP of wheat bran and maize silage under

17
mesophilic batch conditions, reported results consistent with those of the present study

(330.2 NL kgVS-1 and 256.1 NL kgVS-1, respectively).

VFAs concentration in AD digestates, calculated as the sum of acetate, propionate and

butyrate, varied from 0.3 g L-1 in WB to 0.21 g L-1 in eOP. Ammonium analyses

showed concentrations from a minimum of 0.65 g L-1 in eGR to a maximum of 0.99 g

L-1 in WB. pH was similar for all of digestates, comprised in a range from 8.32 to 8.50,

higher than the minimum threshold of 7.0, that inhibits the methanogenic activity

(Weiland, 2010) (Table 5).

3.3 Effect of substrate and fermentation system on energy recovery

Energy recovery (ER, MJ kgVS-1) was calculated in order to compare the overall

bioenergy potential of the substrates (eGR, eZM, eOP, WB) in the two fermentation

systems (CS, AD). The interaction of the two factors (substrates and fermentation

system) was not significant. The effect of the fermentation system was significant, with

lower ER of CS (7.5 MJ kgSV-1) compared with AD (8.3 MJ kgSV-1). The effect of

substrates was also significant: the highest ER values were obtained from WB (9.0 MJ

kgSV-1) and eZM (8.6 MJ kgSV-1). The latter was similar to eGR (8.0 MJ kgSV-1),

while the ER achieved from eOP was significantly the lowest (5.8 MJ kgSV-1) (Fig. 5).

Schievano et al. (2014) reported a semi-continuous assay, where ER of maize silage in

CS ranged from 12 to 19 MJ kgSV-1 and in AD ranged from 13 to 15 MJ kgSV-1, higher

than the ER reported in this study for eZM in CS (7.9 MJ kgSV-1) and in AD (9.4 MJ

kgSV-1). In general, they observed a better performance of CS compared with AD.

Nonetheless, in the present study the CS resulted to be less efficient than AD in terms of

ER. Probably, in a batch assay, the hydrogen potential can be assessed, while the

advantages of the overall CS process cannot be appreciated, because of the non-steady-

state conditions. Conversely, in continuous reactors, an adequate organic loading rate

18
provides a constant feeding composition for methanogenic archaea and an overall

greater stability (Hawkes et al., 2007; Schievano et al., 2012, 2014). In general, even if

the ER in CS was lower than the ER in AD, the final biogas composition of CS was

more suitable for combustion than the one of AD, since the mixture of methane with

hydrogen, i.e. hythane, allows better engine performances and a lower release of

nitrogen oxide and CO2 into the atmosphere (Hallenbeck et al., 2009; Cavinato et al.,

2012). Considering the best performing substrates, i.e. WB and eZM, the obtained

biohythane would potentially have about 7 % of biohydrogen, complying with the best

characteristics for biohythane mixture, where hydrogen content is never below 5%

(Cavinato et al., 2012). Evaluating the amount of CO2 released per each MJ produced

by biogas combustion, WB and eZM could reduce the CO2 production of -2.5 %

compared to eGR and eOP in CS and to all substrates in AD.

Considering the kinetics of CS_AD and of AD, the interaction of the two factors

(substrates and fermentation system) was significant for both T50 (P < 0.05) and Rmax (P

< 0.001) (Appendix). In CS_AD, the 50% of the methane production was reached in 2.9

days in eOP, which was not significantly different from the T50 value of WB (3.7 days),

that did not differ from eZM and eGR (4.4 and 4.6 days, respectively). In AD, all the

T50 values were similar one to the other, ranging from 4.7 days in WB to 5.5 days in

eGR, and were similar to eGR in CS_AD (Fig. 6C and 6D). Concerning the maximum

daily production rate, Rmax values of eZM, eOP and WB in CS_AD (32.4, 32.2 and 34.9

NL kgVS-1 day-1, respectively) were significantly higher than those of eGR in CS_AD

(24.9 NL kgVS-1 day-1) and of all substrates in AD: eGR, eZM, eOP and WB (26.8,

25.0, 22.9 and 23.5 NL kgVS-1 day-1, respectively) (Fig. 6A and 6B). Compared to AD,

the maximum rates in CS_AD were increased of 22.8, 28.9 and 32.8 % in eZM, eOP

and WB, respectively. On the contrary, in eGR a decrease in Rmax (-7.6 %) was

19
observed in CS_AD compared with AD. Kinetics of fermentations are related to

biomass biodegradability, which depends on chemical composition of substrates

(Weiland, 2010; Ragaglini et al., 2014). In eZM and in WB the increase of Rmax in the

CS_AD compared with the AD system can be attributed to the high concentration of

VFAs which were obtained in DF (Table 3). The high Rmax and the low T50 of eOP in

CS_DA could be attributed to the highly biodegradable components remaining from the

partial fermentation occurred in CS_DF, as showed by the absence of H2 and a curve of

CH4 production, that was still far from reaching the plateau. Conversely in eGR, Rmax of

CS_AD was significantly lower than in the other substrates, likely due to the fact that

readily fermentiscible components were more depleted during the CS_DF stage.

eGR and eOP resulted to be not so suitable for a combined fermentation system under

these experimental conditions, likely due to the abundance of more recalcitrant

components like hemicellulose, cellulose and lignin, which are slowly degraded (Corno

et al., 2014). A possible enhancement of the process for lignocellulose matrix could be

obtained by pretreatments, like hydrolysis, eventhough they require an additional cost of

the process and the production of inhibitory compounds, like furfural, can negatively

affect the process (Abo-Hashesh and Hallenbeck, 2012).

In general, it is worth stressing that in combined systems, DF has low hydrogen yields

owing to the inability of bacteria to overcome the thermodynamic barrier for a complete

conversion of substrate to hydrogen (Hallenbeck et al., 2009). Therefore, it can be of

interest to develop other kind of combined systems in order to enhance hydrogen

production, such as photofermentation or microbial electrolytic cells, where energy

supply is mandatory since they require light and electricity, respectively (Adessi et al.,

2012; Ghimire et al., 2015a).

4 Conclusions

20
The obtained results showed that the substrate significantly influenced H2 and CH4

potentials of the fermentation systems. Indeed, the higher content of non- structural

carbohydrates made WB and eZM more suitable than eGR and eOP for biohythane

production in a two-stage process (i.e. sequential dark fermentation and second stage of

anaerobic digestion), even under mesophilic conditions and without pH correction.

Conversely, eOP and eGR resulted unsuitable for a combined system performed under

these conditions, due to their lower biodegradability.

5 Acknowledgements

The authors wish to thank Sergio Cattani, Fabio Taccini and Cristiano Tozzini for their

help in building, setting-up and managing the batch fermentation systems.

6 References

Abo-Hashesh, M., & Hallenbeck, P. C. (2012). Fermentative hydrogen production. In

Microbial Technologies in Advanced biofuels Production (pp. 77-92). Springer US.

Adessi, A., De Philippis, R., & Hallenbeck, P. C. (2012). Combined systems for

maximum substrate conversion. In Microbial Technologies in Advanced Biofuels

Production (pp. 107-126). Springer US.

Alagöz, B. A., Yenigün, O., & Erdinçler, A. (2015). Enhancement of anaerobic

digestion efficiency of wastewater sludge and olive waste: Synergistic effect of co-

digestion and ultrasonic/microwave sludge pre-treatment. Waste Management, 46, 182-

188.

Amon, T., Amon, B., Kryvoruchko, V., Zollitsch, W., Mayer, K., & Gruber, L. (2007).

Biogas production from maize and dairy cattle manure influence of biomass

composition on the methane yield. Agriculture, Ecosystems & Environment, 118(1),

173-182.

21
Argun, H., & Kargi, F. (2011). Bio-hydrogen production by different operational modes

of dark and photo-fermentation: an overview. International Journal of Hydrogen

Energy, 36(13), 7443-7459.

Bianchi, L., Mannelli, F., Viti, C., Adessi, A., & De Philippis, R. (2010). Hydrogen-

producing purple non-sulfur bacteria isolated from the trophic lake Averno (Naples,

Italy). International Journal of Hydrogen Energy, 35(22), 12216-12223.

Borja, R., Sánchez, E., Rincón, B., Raposo, F., Martın, M. A., & Martın, A. (2005).

Study and optimisation of the anaerobic acidogenic fermentation of two-phase olive

pomace. Process Biochemistry, 40(1), 281-291.

Cavinato, C., Giuliano, A., Bolzonella, D., Pavan, P., & Cecchi, F. (2012). Bio-hythane

production from food waste by dark fermentation coupled with anaerobic digestion

process: a long-term pilot scale experience. international journal of hydrogen energy,

37(15), 11549-11555.

Chen, Y., Cheng, J. J., & Creamer, K. S. (2008). Inhibition of anaerobic digestion

process: a review. Bioresource technology, 99(10), 4044-4064.

Corno, L., Pilu, R., & Adani, F. (2014). Arundo donax L.: A non-food crop for

bioenergy and bio-compound production. Biotechnology advances, 32(8), 1535-1549.

Corno, L., Pilu, R., Tambone, F., Scaglia, B., & Adani, F. (2015). New energy crop

giant cane (Arundo donax L.) can substitute traditional energy crops increasing biogas

yield and reducing costs. Bioresource technology, 191, 197-204.

Dragoni, F., Ragaglini, G., Corneli, E., o Di Nasso, N. N., Tozzini, C., Cattani, S., &

Bonari, E. (2015). Giant reed (Arundo donax L.) for biogas production: land use saving

and nitrogen utilisation efficiency compared with arable crops. Italian Journal of

Agronomy, 10(4), 192-201.

22
Driehuis, F., Elferink, S. J. W. H., & Spoelstra, S. F. (1999). Anaerobic lactic acid

degradation during ensilage of whole crop maize inoculated with Lactobacillus buchneri

inhibits yeast growth and improves aerobic stability. Journal of Applied Microbiology,

87(4), 583-594.

Fantozzi, F., & Buratti, C. (2009). Biogas production from different substrates in an

experimental Continuously Stirred Tank Reactor anaerobic digester. Bioresource

technology, 100(23), 5783-5789.

Ghimire, A., Frunzo, L., Pirozzi, F., Trably, E., Escudie, R., Lens, P. N., & Esposito, G.

(2015a). A review on dark fermentative biohydrogen production from organic biomass:

Process parameters and use of by-products. Applied Energy, 144, 73-95.

Ghimire, A., Frunzo, L., Pontoni, L., d'Antonio, G., Lens, P. N., Esposito, G., & Pirozzi,

F. (2015b). Dark fermentation of complex waste biomass for biohydrogen production

by pretreated thermophilic anaerobic digestate. Journal of environmental management,

152, 43-48.

Giordano, A., Cantù, C., & Spagni, A. (2011). Monitoring the biochemical hydrogen

and methane potential of the two-stage dark-fermentative process. Bioresource

technology, 102(6), 4474-4479.

Guo, X. M., Trably, E., Latrille, E., Carrère, H., & Steyer, J. P. (2010). Hydrogen

production from agricultural waste by dark fermentation: a review. International

journal of hydrogen energy, 35(19), 10660-10673.

Guo, X. M., Trably, E., Latrille, E., Carrere, H., & Steyer, J. P. (2014). Predictive and

explicative models of fermentative hydrogen production from solid organic waste: role

of butyrate and lactate pathways. International Journal of Hydrogen Energy, 39(14),

7476-7485.

23
Hallenbeck, P. C., Ghosh, D., Skonieczny, M. T., & Yargeau, V. (2009).

Microbiological and engineering aspects of biohydrogen production. Indian journal of

microbiology, 49(1), 48-59.

Hawkes, F. R., Hussy, I., Kyazze, G., Dinsdale, R., & Hawkes, D. L. (2007).

Continuous dark fermentative hydrogen production by mesophilic microflora: principles

and progress. International Journal of Hydrogen Energy, 32(2), 172-184.

ISTAT (2012)

http://agri.istat.it/sag_is_pdwout/jsp/dawinci.jsp?q=plC010000010000012000&an=201

2&ig=1&ct=867&id=15A|18A|25A Accessed 09.11.2015.

Kim, S. H., Han, S. K., & Shin, H. S. (2006). Effect of substrate concentration on

hydrogen production and 16S rDNA-based analysis of the microbial community in a

continuous fermenter. Process Biochemistry, 41(1), 199-207.

Kolbl, S., Paloczi, A., Panjan, J., & Stres, B. (2014). Addressing case specific biogas

plant tasks: Industry oriented methane yields derived from 5L Automatic Methane

Potential Test Systems in batch or semi-continuous tests using realistic inocula,

substrate particle sizes and organic loading. Bioresource technology, 153, 180-188.

Liu, S., Ge, X., Liew, L. N., Liu, Z., & Li, Y. (2015). Effect of urea addition on giant

reed ensilage and subsequent methane production by anaerobic digestion. Bioresource

technology, 192, 682-688.

Mantovi, P., Fabbri, C., & Soldano, M., (2013). Si ottimizza la filiera del biogas se la

sansa viene pretrattata. L’informatore agrario, 47, 39-43.

Mateo, J. J., & Maicas, S. (2015). Valorization of winery and oil mill wastes by

microbial technologies. Food Research International. 73, 13-25.

Noike, T., & Mizuno, O. (2000). Hydrogen fermentation of organic municipal wastes.

Water Science & Technology, 42(12), 155-162.

24
Pakarinen, O., Lehtomäki, A., & Rintala, J. (2008). Batch dark fermentative hydrogen

production from grass silage: the effect of inoculum, pH, temperature and VS ratio.

International Journal of Hydrogen Energy, 33(2), 594-601.

Pan, C., Fan, Y., & Hou, H. (2008). Fermentative production of hydrogen from wheat

bran by mixed anaerobic cultures. Industrial & Engineering Chemistry Research,

47(16), 5812-5818.

Porter, M. G., & Murray, R. S. (2001). The volatility of components of grass silage on

oven drying and the inter‐relationship between dry‐matter content estimated by different

analytical methods. Grass and Forage Science, 56(4), 405-411.

Ragaglini, G., Dragoni, F., Simone, M., & Bonari, E. (2014). Suitability of giant reed

(Arundo donax L.) for anaerobic digestion: effect of harvest time and frequency on the

biomethane yield potential. Bioresource technology, 152, 107-115.

Schievano, A., Tenca, A., Scaglia, B., Merlino, G., Rizzi, A., Daffonchio, D., Oberti,

O., & Adani, F. (2012). Two-stage vs single-stage thermophilic anaerobic digestion:

comparison of energy production and biodegradation efficiencies. Environmental

science & technology, 46(15), 8502-8510.

Schievano, A., Tenca, A., Lonati, S., Manzini, E., & Adani, F. (2014). Can two-stage

instead of one-stage anaerobic digestion really increase energy recovery from biomass?.

Applied Energy, 124, 335-342.

Weiland, P. (2010). Biogas production: current state and perspectives. Applied

microbiology and biotechnology, 85(4), 849-860.

25
Figure captions

Fig.1. Cumulative production of biogas, in terms of Biochemical Biogas Potential

(BBP), of ensiled giant reed (eGR), ensiled Zea mays (eZM), ensiled olive pomace

(eOP) and wheat bran (WB) in the first stage of dark fermentation (CS_DF).

Fig.2. Biochemical Methane Potential (BMP), Biochemical Hydrogen Potential (BHP),

Methane Concentration (MC) and Hydrogen Concentration (HC) of the first stage of

dark fermentation (CS_DF) of ensiled giant reed (eGR), ensiled Zea mays (eZM),

ensiled olive pomace (eOP) and wheat bran (WB). Significance levels of ANOVA are

shown: * (p < 0.05), ** (p < 0.01), *** (p < 0.001).

Fig.3. Biochemical Biogas Potential (BBP), Biochemical Methane Potential (BMP) and

Methane Concentration (MC) of the second stage of anaerobic digestion (CS_AD) of

ensiled giant reed (eGR), ensiled Zea mays (eZM), ensiled olive pomace (eOP) and

wheat bran (WB). Significance levels of ANOVA are shown: * (p < 0.05), ** (p <

0.01), *** (p < 0.001).

Fig.4. Biochemical Biogas Potential (BBP), Biochemical Methane Potential (BMP) and

Methane Concentration (MC) of the anaerobic digestion (AD) of ensiled giant reed

(eGR), ensiled Zea mays (eZM), ensiled olive pomace (eOP) and wheat bran (WB).

Significance levels of ANOVA are shown: * (p < 0.05), ** (p < 0.01), *** (p < 0.001).

Fig.5. (A) Energy recovery (ER) from ensiled giant reed (eGR), ensiled Zea mays

(eZM), ensiled olive pomace (eOP) and wheat bran (WB) considering the mean

between the combined system (CS) and the anaerobic digestion (AD). (B) Energy

recovery from the combined system (CS) and the anaerobic digestion (AD) considering

the mean of the considered substrates. Significance levels of ANOVA are shown: * (p

< 0.05), ** (p < 0.01), *** (p < 0.001).

26
Fig.6. Kinetics of fermentation of ensiled giant reed (eGR), ensiled Zea mays (eZM),

ensiled olive pomace (eOP) and wheat bran (WB). Rmax and daily methane production

rate estimated as the first derivative of cumulate production curves of CS_AD (A) and

of AD (B); cumulative methane production along 40 days and T50 of CS_AD (C) and of

AD (D).

Table captions

Table 1 Total solids (TS), volatile solids (VS), total nitrogen (N), carbon to nitrogen

ratio (C/N), Neutral Detergent Fiber (NDF), Acid Detergent Fiber (ADF), Acid

Detergent Lignin (ADL), starch and Water Soluble Carbohydrates (WSCs) of ensiled

giant reed (eGR), ensiled Zea mays (eZM), ensiled olive pomace (eOP) and wheat bran

(WB).

Table 2 pH, ethanol, lactic acid and volatile fatty acids (acetic acid, propionic acid,

butyric acid, isovalerian acid and valerian acid) of ensiled substrates (ensiled giant reed

(eGR), ensiled Zea mays (eZM), ensiled olive pomace (eOP)).

Table 3 pH, ammonium, acetic acid, propionic acid, butyric acid, volatile fatty acids

(VFAs, as the sum of acetic acid, propionic acid and butyric acid) and butyrate to

acetate molar ratio (B/A) of dark fermentation (CS_DF) effluents of ensiled giant reed

(eGR), ensiled Zea mays (eZM), ensiled olive pomace (eOP) and wheat bran (WB).

Table 4 pH, ammonium, acetic acid, propionic acid, butyric acid and volatile fatty acids

(VFAs, as the sum of acetic acid, propionic acid and butyric acid) of the second stage of

anaerobic digestion (CS_AD) digestates of ensiled giant reed (eGR), ensiled Zea mays

(eZM), ensiled olive pomace (eOP) and wheat bran (WB).

Table 5 pH, ammonium, acetic acid, propionic acid, butyric acid and volatile fatty acids

(VFAs, as the sum of acetic acid, propionic acid and butyric acid) of anaerobic digestion

27
(AD) digestates of ensiled giant reed (eGR), ensiled Zea mays (eZM), ensiled olive

pomace (eOP) and wheat bran (WB).

28
 100

90
eGR
80 WB
70 eZM
eOP
60
BBP (NL kgVS-1)

50

40

30

20

10

0
1 10 20 30 40 50 60 70 80 90
Time (h)
29
30
31
32
33
 Table 1 Total solids (TS), volatile solids (VS), total nitrogen (N), carbon to nitrogen

ratio (C/N), Neutral Detergent Fiber (NDF), Acid Detergent Fiber (ADF), Acid

Detergent Lignin (ADL), starch and Water Soluble Carbohydrates (WSCs) of ensiled

giant reed (eGR), ensiled Zea mays (eZM), ensiled olive pomace (eOP) and wheat bran

(WB).

Substrate TS VS N C/N NDF ADF ADL Starch WSCs

(%) (%TS) (%TS) (%TS) (%TS) (%TS) (%TS) (%TS)
eGR 26.5 94.7 0.9 54.7 78.3 51.7 8.0 0.5 1.2
eZM 21.0 92.7 1.5 31.3 56.8 31.6 3.8 6.2 6.2
eOP 39.0 97.5 1.4 42.1 64.9 54.8 38.9 udla 0.1
WB 92.5 95.4 2.6 16.9 32.3 10.8 3.5 5.8 2.8
a
udl = under detection limit
Table 2 pH, ethanol, lactic acid and volatile fatty acids (acetic acid, propionic acid,

butyric acid, isovalerian acid and valerian acid) of ensiled substrates (ensiled giant reed

(eGR), ensiled Zea mays (eZM), ensiled olive pomace (eOP)).

Substrate pH Ethanol Lactic Acetic Propionic Butyric Isovaleric Valeric

(%TS) acid acid acid acid acid acid
(%TS) (%TS) (%TS) (%TS) (%TS) (%TS)
eGR 4.33 1.47 0.83 1.92 0.05 0.10 0.38 udla
eZM 3.72 1.55 13.64 6.60 0.40 udla udla udla
eOP 5.05 0.02 13.96 2.09 0.14 0.26 0.20 0.62
a
udl = under detection limit

2
 Table 3 pH, ammonium, acetic acid, propionic acid, butyric acid, volatile fatty acids

(VFAs, as the sum of acetic acid, propionic acid and butyric acid) and butyrate to

acetate molar ratio (B/A) of dark fermentation (CS_DF) effluents of ensiled giant reed

(eGR), ensiled Zea mays (eZM), ensiled olive pomace (eOP) and wheat bran (WB).

CS_DF pH NH4 + Acetic Propionic Butyric VFAs B/A

effluent (g L-1) acid acid acid (g L-1)
(g L-1) (g L-1) (g L-1)
eGR 5.35 0.43 ± 0.02 3.73 ± 0.04 0.59 ± 0.02 2.15 ± 0.10 6.47 ± 0.05 0.4
eZM 5.12 0.53 ± 0.01 3.05 ± 0.06 0.61 ± 0.04 15.37 ± 0.09 19.02 ± 0.06 3.4
eOP 7.62 0.35 ± 0.00 0.77 ± 0.01 0.28 ± 0.03 0.46 ± 0.03 1.50 ± 0.06 0.4
WB 5.66 0.66 ± 0.02 2.37 ± 0.12 0.44 ± 0.03 12.31 ± 0.14 15.12 ± 0.27 3.5

3
Table 4 pH, ammonium, acetic acid, propionic acid, butyric acid and volatile fatty acids

(VFAs, as the sum of acetic acid, propionic acid and butyric acid) of the second stage of

anaerobic digestion (CS_AD) digestates of ensiled giant reed (eGR), ensiled Zea mays

(eZM), ensiled olive pomace (eOP) and wheat bran (WB).

CS_AD pH NH4 + Acetic Propionic Butyric VFAs

digestate (g L-1) acid acid acid (g L-1)
(g L-1) (g L-1) (g L-1)
eGR 8.27 1.08 ± 0.02 0.20 ± 0.01 0.33 ± 0.08 0.52 ± 0.01 1.05 ± 0.08
eZM 8.44 1.16 ± 0.00 0.02 ± 0.00 0.02 ± 0.00 0.04 ± 0.01 0.08 ± 0.01
eOP 8.32 1.18 ± 0.03 0.11 ± 0.03 0.67 ± 0.05 0.64 ± 0.04 1.42 ± 0.04
WB 8.32 1.32 ± 0.00 0.37 ± 0.13 0.28 ± 0.02 0.48 ± 0.01 1.13 ± 0.16

4
Table 5 pH, ammonium, acetic acid, propionic acid, butyric acid and volatile fatty acids

(VFAs, as the sum of acetic acid, propionic acid and butyric acid) of anaerobic digestion

(AD) digestates of ensiled giant reed (eGR), ensiled Zea mays (eZM), ensiled olive

pomace (eOP) and wheat bran (WB).

AD pH NH4 + Acetic Propionic Butyric VFAs

digestate (g L-1) acid acid acid (g L-1)
(g L-1) (g L-1) (g L-1)
eGR 8.32 0.69 ± 0.00 0.06 ± 0.01 0.09 ± 0.05 0.07 ± 0.03 0.21 ± 0.07
eZM 8.41 0.65 ± 0.00 0.04 ± 0.00 0.02 ± 0.00 0.02 ± 0.00 0.08 ± 0.00
eOP 8.44 0.88 ± 0.00 0.02 ± 0.00 0.01 ± 0.00 0.01 ± 0.01 0.04 ± 0.01
a
WB 8.50 0.99 ± 0.00 0.02 ± 0.01 udl 0.01 ± 0.00 0.03 ± 0.02
a
udl = under detection limit

5
Highlights

• Biomass rich in non-structural carbohydrates was more suitable for combined

system

• Wheat bran showed the highest H2/CH4 potential and energy recovery

• Olive pomace and giant reed were unsuitable for combined H2/CH4 production

39