Bioresource Technology 148 (2013) 443–452

Contents lists available at ScienceDirect

Bioresource Technology

journal homepage: www.elsevier .com/locate /bior tech

Optimisation of the two-phase dry-thermophilic anaerobic digestionprocess of sulphate-containing municipal solid waste: Populationdynamics

0960-8524/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.biortech.2013.09.002

⇑ Corresponding author. Tel.: +34 956 016423.E-mail address: soraya.zahedi@uca.es (S. Zahedi).

S. Zahedi a,⇑, D. Sales a, L.I. Romero b, R. Solera a

a Department of Environmental Technologies, Faculty of Marine and Environmental Sciences (CASEM), University of Cádiz, Pol, Río San Pedro s/n, 11510 Puerto Real, Cádiz, Spainb Department of Chemical Engineering and Food Technology, Faculty of Sciences, University of Cádiz, Pol, Río San Pedro s/n, 11510 Puerto Real, Cádiz, Spain

h i g h l i g h t s

� Biohythane was obtained from sulphate-containing organic industrial solid wastes.� Different organic loading rates and hydraulic retention times (HRTs) were carried out.� Maximum of 1.9 l H2/l/d and 5.4 l CH4/l/d were obtained in the process.� The lower HRT the greater acidogenic stage takes place in the second reactor.� Acetogens and Archaea were dominated over sulphate-reducing bacteria in second phase.

a r t i c l e i n f o

Article history:Received 12 July 2013Received in revised form 27 August 2013Accepted 1 September 2013Available online 11 September 2013

Keywords:Hydrogen productionMethane productionSulphide productionPopulation dynamicTwo-phase dry-thermophilic anaerobicdigestion

a b s t r a c t

Microbial population dynamics and anaerobic digestion (AD) process to eight different hydraulic reten-tion times (HRTs) (from 25 d to 3.5 d) in two-phase dry-thermophilic AD from sulphate-containing solidwaste were investigated. Maximum values of gas production (1.9 ± 0.2 l H2/l/d; 5.4 ± 0.3 l CH4/l/d and82 ± 9 ml H2S/l/d) and microbial activities were obtained at 4.5 d HRT; where basically comprised hydro-lysis step in the first phase (HRT = 1.5 d) and acidogenic step finished in the second phase as well as ace-togenic–methanogenic steps (HRT = 3 d). In the first phase, hydrolytic–acidogenic bacteria (HABs) wasthe main group (44–77%) and Archaea, acetogens and sulphate-reducing bacteria (SRBs) contents werenot significant; in the second phase (except to 2 d HRT), microbial population was able to adapt to changein substrate and HRTs to ensure the proper functioning of the system and both acetogens and Archaeawere dominated over SRBs. Decreasing HRT resulted in an increase in microbial activities.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

AD is one of the effective technologies used to recover energy re-sources from organic wastes, in addition to being a simple and effec-tive biotechnological means of reducing and stabilizing organicwastes. It is carried out by the coordinated action of various groupsof microorganisms and goes through several intermediate stages. Inthe first step (hydrolysis), complex organic polymers are hydrolyzedinto simpler soluble organic compounds by HABs; in this step largequantities of hydrogen (H2) are produced. In the second step (acido-genesis), HABs produce volatile fatty acids (VFA), alcohols, H2 andcarbon dioxide (CO2). In the third steps (acetogenesis), acid acetic,H2 and CO2 is produced by obligate H2-producing acetogens (aceto-gens), organisms that consume fermentation products, such as

propionate, butyrate, lactate, and ethanol. Syntrophobacter(propionate-utilizing acetogens (PUAs)) and Syntrophomonas(butyrate-utilizing acetogens (BUAs)) are the majority of the aceto-gens known (Mara and Horan, 2003). Acetogens requires very lowH2 partial pressure to favour the thermodynamics of the reactions(Boone et al., 1989). In the absence of external electron acceptorsanaerobic oxidation of butyrate and propionate occurs only in syn-trophic association with H2-utilizing methanogens (HUMs) (Booneet al., 1989; Liu et al., 2011). In the fourth step (methanogenesis),methane (CH4) is produced by methanogenic population. The meth-anogens are normally divided into two main groups based on theirsubstrate conversion capabilities. Acetate-utilizing methanogens(AUMs) are capable of converting acetate to CH4 and CO2 and HUMsconvert H2 and CO2 to CH4.

HABs, acetogens and methanogenic microorganisms differ, notonly in terms of their nutrition and pH requirements, but also withrespect to their physiology, growth, and nutrient uptake kinetics,

444 S. Zahedi et al. / Bioresource Technology 148 (2013) 443–452

and in their particular ability to with stand environmental changes,thus different requirements regarding reactor conditions, have ledto the development of two-phase AD processes (De la Rubia et al.,2009). The phase-separated anaerobic process including two-stagesystem normally comprises hydrolysis–acidogenesis in the firstphase and acetogenesis–methanogenesis process in the secondphase. A step forward of the common AD process, is the separatephase approach finalised to the production of H2 in the first phasereactor and CH4 in the second phase reactor. Obtained gases can beused separately or mixed together to obtain biohythane (Cavinatoet al., 2011). Methanogenic microorganisms grow more slowlythan HABs, at a rate similar to acetogens (3.6 d) and the optimumpH environment for methanogens and acetogens is in the range6.5–8.5 (Montero et al., 2009; De la Rubia et al., 2009).Consequently, conditions that are non-favourable to the growthof methanogens (short HRTs or/and low pH (5.2–6.5)) should beestablished in the first phase, while in the second phase, high HRTsand neutral or slightly basic pH should be imposed.

The organic fraction municipal solid waste (OFMSW) used tofeed proceeds of the municipal solid waste treatment plant ‘‘LasCalandrias’’ (Jerez de la Frontera, Cádiz-Spain) typically containedsulphate at concentrations higher than 5 g/kg. During AD ofsulphate-containing wastes, SRBs can lead to the undesirable pro-duction of hydrogen sulphide (H2S). The nutritional requirementsof SRBs are an inorganic electron acceptor, this is usually providedby sulphate ion and an electron donor, and essentially these consistof VFA or H2, and occasionally sugars and lounge chain fatty acids.Two stages of inhibition exist as a result of sulphate reduction(Chen et al., 2008); primary inhibition is due to competition forcommon organic and inorganic substrates from SRBs and second-ary inhibition results from the toxicity of sulphide to variousmicrobial groups.

Understanding the functioning of anaerobic reactors requiresquantitative information on microbial numbers, biomass, andactivities of the bacterial groups involved in the process (Monteroet al., 2009). The stability of the system depends on the activemicrobial groups involved in the process (Montero et al., 2009).Molecular tools as fluorescent in situ hybridization (FISH), basedon sequence comparison of small-subunit (SSU) ribosomal RNA(rRNA) molecules already have been used for the quantificationof population abundance in different anaerobic environments(McMahon et al., 2001; Montero et al., 2008, 2009).

This study aims to: (1) establish the optimal conditions (organicloading rate (OLR) or HRT) in order to maximise the gas production(GP); (2) investigate the population dynamics and (3) study the dif-ferent microbial activities, including SRBs in the two-phase dry-thermophilic AD process of sulphate-containing municipal solidwaste. No previous studies had been published about this. Forthese purposes, the effect of eight different OLRs (from 3 g to24 g TVS/l/d) or HRTs (from 25 d to 3.5 d) were tested.

The effect of the variations in operating parameters (HRT orOLR) on dissolved chemical oxygen demand (CODD), VFA, total vol-atile solids (TVS), hydrogen production (HP), methane production(MP), GP, sulphide production (SP), microbial population andmicrobial activities to improve the two-phase dry-thermophilicAD process of sulphate-containing municipal solid waste was stud-ied at laboratory scale. FISH was used to determine the maingroups involved in the anaerobic process.

2. Methods

2.1. Experimental equipment and operating condition

Two laboratory-scale continuously stirred tank reactors (CSTRs)were employed. The first reactor, dedicated to the HP (first phase),

had a 5.5 l working volume, while the second reactor (second phase)dedicated to the MP had a 5 l working volume, both heated byrecirculating water through a thermostatic jacket. A PRECISTERM6000142/6000389 (SELECTA S.A.) baths were used, with amaximum capacity of 7 l of water. The stainless steel reactors lidhave a diameter of 200 mm and contain three openings, one forthe biogas outlet, a feed inlet and another opening for the stirringsystem. The bottoms of the reactors have a discharge valve with a40 mm i.d., used for sampling. The biogas was collected in 40 lcapacity Tedlar (a polyvinyl fluoride plastic polymer) bags. Thebags are 29.8 cm wide and 45.7 cm long. The stirring systems con-sist of an IKA EUROSTAR Power Control visc-P4 overhead stirrercoupled to a stainless steel blade with scrapers which allowshomogenisation of the waste at a speed of 23 rpm. In CSTRs with-out recycling of solids, the solid retention time (SRT) and HRT areequal.

The experimental test was divided in eight periods or condi-tions (runs). A start up phase and a steady state condition pro-tracted for at least 3 consecutive HRTs were clearly defined forany experimental run (except at Run VIII, because the destabilisa-tion was observed). The operational conditions start up and steadystate condition periods are shown in Table 1. The whole experi-ment length was 398 d (Run I, 0–100; Run II, 101–189; Run III,190–257; Run IV, 258–303; Run V, 304–337; Run VI, 338–368;Run VII, 369–395; Run VIII, 396–398).

2.2. Inoculum, substrate and feeding

The seed used as inoculum for the methanogenic reactor wascollected from a single phase dry-thermophilic AD of OFMSW;while inoculum for the acidogenic reactor was collected from anH2-producing reactor. The total solid (TS) concentration and TVSin the methanogenic inoculum were 67 g/kg and 33 g/kg and inthe acidogenic inoculum these were 82 g/kg and 50 g/kg,respectively.

The tested substrate in the first phase was the OFMSW from the30 mm trommel of the municipal solid waste treatment plant inCadiz, Spain. The OFMSW was stored in 25 kg drums at �4 �C toavoid anaerobic degradation by the microorganisms found in thesolid waste itself. The TS concentration of the feed first reactorwas adjusted to 20% (which is characteristic of dry AD) by addingtap water. Characterisation of the substrate used in the assay isshowed in Table 2.

The tested substrate in the second phase was the effluent of thefirst phase.

In the first reactor NaOH 10 M was added to the substrate whenthe pH of the effluent was below 5.3. In the second reactor, anycontrol of pH was realised.

About the feeding regime, both reactors were fed once a day(semi-continuous).

2.3. Analytical methods

The analytical determinations made in this study can begrouped in two categories: physical–chemical analysis and micro-biological analysis.

2.3.1. Physical–chemical analysisFor the control of the reactors the following parameters were

determined: Total chemical oxygen demand (CODT), CODD, alkalin-ity, sulphate, TVS, pH, total volatile fatty acids (TVFA), acetic, buty-ric, propionic, volume and composition of the biogas (H2, CH4, CO2

and H2S). These determinations were performed according toAPHA (1995) and Fdez-Güelfo et al. (2010). The sulphates wereanalysed from the filtrate supernatant obtained by means effluentsample lixiviation (10 g of digested waste in 100 ml of Milli-Q

Table 1Operating conditions applied during the experimental test.

Run Operationtime (d)

Start upphase (d)

State conditionsphase (d)

Acidogenic phase Methanogenic phase Total performance

OLR(g TVS/l/d)

HRT (d) OLR(g TVS/l/d)

HRT(d)

TVSremoval (%)

CODD

removal (%)Sulphateremoval (%)

OLR(g TVS/l/d)

HRT(d)

TVSremoval (%)

I 100 25 75 9 10 3 15 72 85 27 3 25 84II 89 24 65 13 6.6 5 15 84 71 44 4 21.6 87III 68 18 50 13 6.6 7 10 76 69 56 5 16.6 80IV 46 13 33 20 4.4 10 6.6 73 67 53 8 11 80V 34 10 24 57 1.5 8 6.6 78 76 65 9 8.1 86VI 31 13 18 57 1.5 13 4.4 78 70 65 14 5.9 86VII 27 14 13 57 1.5 22 3 78 63 65 19 4.5 84VIII 3 a a 57 1.5 33 2 57 33 48 24 3.5 67

a Destabilisation was observed.

Table 2Characterisation physical–chemical and microbiological of the substrate used in thefirst phase reactor.

Parameter Average Range

pH 5.5 (0.7) 4.8–6.2CODD (g/l) 25 (5) 19–30CODT (g/l) 72 (2) 69–73TVFA (g acetic/l) 1.8 (0.4) 2.2–1.4Acetate (g/l) 1.7 (1.1) 2.8–0.6Propionate (g/l) 0.4 (0.7) 0.1–1.7Butyrate (g/l) 0.1 (0.0) 0.1–0.2TVS (g/kg) 85 (5) 79–89TS (g/kg) 120(10) 110–130TVS/TS (%) 71(6) 65–77Sulphate (g/l) 1.9 (0.7) 1.2–2.9Total population (108 cells/ml) 12.5 (1.0) 11.5–13.5Eubacteria (%) 97 (1) 96–98HABs (%) 90 (1) 89–91SRBsa (%) 4 (0) 4–4Acetogens (%) 6 (1) 5–7PUAs (%) 2 (0) 2–2BUAs (%) 4 (1) 3–5Archaea (%) 3 (0) 3–3AUMs (%) 1 (0) 0–1HUMs (%) 2 (0) 1–2

Average values are shown with standard deviations in parentheses.The number of replicas to pH analytical determination and microbiological analysiswere 398 and 18 respectively; and to physical analysis to TVS, TS, CODT, CODD,sulphate and VFA were 55.

a Percentages compared to total Eubacteria.

S. Zahedi et al. / Bioresource Technology 148 (2013) 443–452 445

water during 20 min). Samples of sulphates were further filteredusing a 1 lm pore size glass fibre filter. This parameter was mea-sured using a commercial kit (Merck Ref. 1.14791.0001).

The gas volume produced in the reactors was directly measuredusing a high-precision flow gas meter: Ritter� drum-type gas me-ter TG-01-Series (Wet-Type).

The biogas composition was carried out by gas chromatographyseparation (SHIMADZU GC-2010). The H2, CH4, CO2, O2 and N2 wereanalysed by means of a thermal conductivity detector (TCD) usinga Supelco Carboxen 1010 Plot column. A Supelco Supel-Q Plot col-umn and a flame photometric detector (FPD) were used for H2S.Samples were taken using a 1 ml Dynatech Gastight gas syringeunder the following operating conditions: Split = 100; constantpressure in the injection port (70 kPa); 2 min at of 40 �C; rampedat 40 �C/min until 200 �C; 1.5 min at 200 �C; detector temperature:250 �C; injector temperature: 200 �C. Helium was used as carriergas (266.2 ml/min); synthetic air (120 ml/min) and hydrogen(80 ml/min) as mixtures for the flame; and He (8 ml/min) as auxil-iary gas to make up. A commercial mixtures of H2, CH4, CO2, O2, N2

and H2S (Abelló Linde S.A.) were used to calibrate the system.Gas volume and composition were measured daily; in the

effluent, the pH was measured daily in all runs. TVS, CODD,sulphate, alkalinity and VFA were analysed two times a week for

Run I and II; three times a week for Run III and IV; five times aweek for Run V, VI and VII and daily for Run VIII. In the character-isation of substrate, the pH was measured daily. TVS, CODD,sulphate and VFA were analysed one time a week.

TVS removal was calculated as the percentage difference be-tween the TVS of the substrate and the TVS of the effluent withinthe substrate TVS. TVFA was calculated by addition of the individ-ual fatty acids.

The percentage of soluble organic matter, VFA and soluble sul-phate removal were calculated according to the followingequations:

ðCODD first phase � CODD second phaseÞ � 100=CODD first phase ð1Þ

ðVFAfirst phase � VFAsecond phaseÞ � 100=VFAfirst phase ð2Þ

ðSO�24 first phase � SO�2

4 second phaseÞ � 100=SO�24 first phase ð3Þ

2.3.2. Microbiological analysisFISH was used to count the microorganisms contained in the

substrate and reactors. The samples were collected from the dry-thermophilic anaerobic reactors and substrate into sterile univer-sal bottles. Absolute ethanol was added to the bottles in a volumeratio of 1 sample:1 ethanol. The samples were stored at �20 �C un-til they were fixed as described in the following section (within2 months). Prior to FISH analysis, the samples were pre-treated.The pre-treatment applied for microbiological count of high solidscontent samples was the addition of Tween 80 and 120 s of shaking(Montero et al., 2008, 2009). The main steps of FISH of whole cellsusing 16S rRNA-targeted oligonucleotide probes are cell fixation,consequent permeabilization and hybridization with the desiredprobe(s) (Montero et al., 2008, 2009). The 16S rRNA-targeted oligo-nucleotide, the incubated temperature and the hybridization time(isotonic moisture chamber) probes in this study, are shown in Ta-ble 3. All probes were labelled with 6-FAM at the 50 terminal. Thesamples were examined visually and cells counted using an AxioImager Upright epifluorescence microscope (Zeiss) with a 100 Wmercury lamp and an 100� oil objective.

The cellular concentration and percentages of Eubacteria,Archaea, BUAs, PUAs, SRBs, HUMs and AUMs were obtained byFISH. The total population was calculated as the sum of the relativeamounts of Eubacteria and Archaea, because the main anaerobicgroups in the anaerobic reactors are contained within these twodomains (Griffin et al., 1998). Acetogens were calculated as thesum of the relative amounts of PUAs and BUAs. HABs were calcu-lated as the difference in the relative amounts of Eubacteria andacetogens.

Regarding the sampling procedure: in the effluent, the microbi-ological analyses were measured in steady state conditions (exceptat 2 d HRT in the second phase, because destabilisation was

Table 3Oligonucleotides probes used in this study.

Probes Probe sequences (from 50 to 30) Target Formamide (%) Time (h) T (�C) References

S-D-Bact-0338-a-A-18 GCTGCCTCCCGTAGGAGT Eubacteria 20 1.5 46 Montero et al. (2008, 2009)S-D-Bact-0338-a-S-18 ACTCCTACGGGAGGCAGC None (negative control) 20 1.5 46 Montero et al. (2008, 2009)S-D-Arch-0915-a-A-20 GTGCTCCCCCGCCAATTCCT Archaea 35 1.5 46 Montero et al. (2008, 2009)S-F-Mbac-1174-a-A-22 TACCGTCGTCCACTCCTTCCTC Methanobacteriaceae (H2-utilizing) 35 1.5 46 Montero et al. (2008, 2009)S-F-Msae-0825-a-A-23 TCGCACCGTGGCCGACACCTAGC Methanosaetaceae (acetate-utilizing) 20 1.5 46 Montero et al. (2008, 2009)S-�-Srb-0385-a-A-18 CGGCGTCGCTGCGTCAGG SRBs 30 2 46 Zahedi et al. (2013)Synbac824 GTACCCGCTACACCTAGT Syntrophobacter sp (propoinate-utilizing) 10 2 46 Zahedi et al. (2013)S-F-Synm-0700-a-A-23 ACTGGTXTTCCTCCTGATTTCTA Syntrophomonadaceae (butyrate-utilizing) 30 2 52 Zahedi et al. (2013)

446 S. Zahedi et al. / Bioresource Technology 148 (2013) 443–452

observed): one time a week for Run I, II, III and IV; two time a weekfor Run V, VI, VII; and daily for Run VIII; and in the substrate, themeasurement was performed every 3 weeks.

To evaluate the biochemical activity on the OLR it has been con-sidered three microbial activity parameters, HAB activity, SRBactivity and Archaea activity.

HABs activity was calculated as the ratio of H2 volume gener-ated and the number of HABs inside the reactor by FISH staining.Archaea activity was calculated as the ratio of CH4 volume gener-ated and the number of Archaea inside the reactor by FISH staining.SRB activity was calculated as the ratio of H2S volume generatedand the number of SRBs inside the reactor by FISH staining (Soleraet al., 2001; Montero et al., 2009; Zahedi et al., 2013).

2.4. Hydrolysis and acidification yields

Hydrolysis and acidification yields parameters have been con-sidered to assess the performance and behavioural change intwo-phase dry-thermophilic AD.

Hydrolysis yield was defined as liquefaction or solubilisation oforganic matter. It was calculated according to the following equa-tion (Demirel and Yenigun, 2004; De la Rubia et al., 2009) :

Hydrolysis yield ¼ Ss

Si� 100 ð4Þ

where Si is the initial total substrate concentration (CODT) and Ss isthe soluble output CODD.

Acidification yield was calculated via soluble COD through thefollowing equation (De la Rubia et al., 2009):

Acidification yield ¼ STVFA

Ss� 100 ð5Þ

where STVFA is the concentration of TVFA generated, expressed asmg COD/l using the theoretical COD equivalents for each VFA.

3. Results and discussion

In this section, the effect of the variations in operating parame-ters (HRT or OLR) on CODD, VFA, TVS, HP, MP, GP, SP, microbialpopulation and microbial activities in the two-phase dry-thermo-philic AD process of sulphate-containing municipal solid waste isstudied. To this end, this section has been structured into fourparts: process stability, first phase process, second phase processand optimal conditions of the two-phase dry-thermophilic AD pro-cess of sulphate-containing OFMSW.

The results about the characterisation of reactors effluents andyields of the process, and characterisation of the different micro-bial populations in the reactors are shown in Tables 4 and 5,respectively. All the values correspond to the analytical determina-tions in steady conditions (except at second phase in Run VIII, be-cause destabilisation was observed).

3.1. Process stability

GP and pH value are important indicators in anaerobicreactors, specifically for the activities of acetogenic bacteria andmethanogens.

In the first phase pH remained approximately constant for alltested runs, with medium values of 5.3 ± 0.6 (Table 4). These pHvalues were optimal for enhanced HAB activity (Demirer and Chen,2004; De la Rubia et al., 2009). In the second phase, pH reached aconstant value of 7.8 ± 0.5, except at Run VIII (7.0 ± 0.4).

The change in operational conditions was applied when thereactor was stable, HP and MP being the parameter chosen to dem-onstrate the stability of the first phase and second phase, respec-tively (Fig. 1a and b). Stable operation was tested from Run I toRun VII. In Run VIII was observed a destabilisation in the secondphase, the results were a decrease in % CH4, MP, SP, specific methaneproduction (SMP), pH, removal percentage of TVS, microbial popula-tion (except to SRBs) and increasing CODD, sulphate and VFA.

3.2. First phase process

First phase dry-thermophilic AD process of sulphate-containingOFMSW was carried out at four different HRTs or four OLRs(Table 1). Alkalinity remained approximately constant for all testedruns, with medium values of 6 ± 2 g CaCO3/l.

The biogas produced was composed of H2 and CO2, with no CH4

or H2S being detected under any of the tested conditions. Increas-ing OLRs resulted in an increase in GP, HP, specific hydrogen pro-duction (SHP) and percentage of H2 in biogas (% H2) (Table 4).The HP and SHP tested were significantly higher than those ob-tained for Romero et al. (2013) during the acidogenic fermentationof the OFMSW.

The performance of hydrolytic-phase and acidogenic-phasedigestion was evaluated in terms of the hydrolysis yield and acido-genic yield respectively (Table 4). When the HRT decreased from10 d to 1.5 d or OLR increased from 9 g to 57 g TVS/l/d, the hydro-lysis yield in this reactor was increased from 40% to 54 ± 3%,according to the increasing sequence of CODD. The acidogenic yieldremains constant (51 ± 3%) between 10 d and 4.4 d HRT (OLR of be-tween 9 g and 20 g TVS/l/d) (Run I, II, III and IV), however, at anHRT of 1.5 d (OLR = 57 g TVS/l/d) (Run V, VI, VII and VIII) decreasingin acidogenic yield value (27%) was produced according to thedecreasing of TVFA concentration in the effluent. This indicatesthat HP at an HRT of 1.5 d was mainly due hydrolytic activityand it produces an effluent suitability for the acidification in thesecond phase. It was according to an increase in HABs in the secondphase.

Regarding microbial population dynamics in the first phase, asummary to the main microbial groups involved is showed inTable 5. Logically, the amount of Eubacteria were much higher thanArchaea content, and within these Eubacteria, HABs were dominant(44–77%). The average values of the relations of Eubacteria:Archaea

S. Zahedi et al. / Bioresource Technology 148 (2013) 443–452 447

in the first phase reactor were in the range 80–93:20–7. The con-tent of Archaea (7–20%) and SRBs (11–22%) in a stable H2-produc-ing reactor was no significant because the biogas produced wassulphide and methane-free for all tested runs (no SRB and Archaeaactivities were observed). That is also true concerning the aceto-gens populations (13–36%), because the degradation of butyrateor propionate to H2 and acetate is endergonic under hydrolytic-aci-dogenic conditions (Liu et al., 2011), and therefore, HP in the firstphase is due exclusively to HAB biochemical activity. To evaluatethe biochemical activity in the first phase, the HAB activity param-eter has been considered. Increasing OLR (from 9 g TVS/l/d to57 g TVS/l/d) or decreasing HRT (from 10 d to 1.5 d) resulted inan increase in HAB activity (from 7 ± 1 � 10�13 l H2/cell/d to30 ± 4 � 10�13 l H2/cell/d). In systems with no biomass retention,a decreased HRT is reflected by faster rates of dilution and henceby a greater number of microorganisms exiting the system dailyin the effluent. Consequently, a large amount of substrate is con-sumed to keep the population size steady as a result of a young,very active population inside the reactor. This situation is intensi-fied when the systems work under no limiting substrate conditions(Solera et al., 2001; Zahedi et al., 2013).

The results conclude on the one hand the content of Archaea,SRBs and acetogens in a stable H2-producing reactor was no signif-icant (no activity was observed) and the other hand that the bestcondition to HP was achieved at an HRT of 1.5 d (Run V, VI, VIIand VIII), obtaining higher average values of HP (1.6–2.1 l H2/l/d),GP (3.5–4.4 l/l/d), SHP (28–39 l H2/kg TVS), hydrolysis yield (52–57%) and HAB activity (26–34 � 10�13 l H2/cell/d), however thelowest values of acidogenic yield (27%) was obtained. This fact pro-duces an effluent suitability for the acidification in the secondphase, becoming an acidogenic–acetogenic–methanogenic reactorrather acetogenic–methanogenic reactor, as will be discussed later.At an HRT of 1.5 d (optimal condition) the average values of theratios of Eubacteria:Archaea, acetogens:HABs, PUAs:BUAs andHUMs:AUMs were 88:12, 19:69, 10:9 and 7:5, respectively andthe relative SRBs percentage was 15%.

3.3. Second phase process

MP in the second phase was carried out at six different HRTs oreight different OLRs (Table 1). Alkalinity was ranges 4–9 g CaCO3/lfor all tested runs. The decomposition of TVS was in the range 72–84% (Table 1), except at Run VIII (57%). They were according tothose obtained by Cuetos et al. (2008) in the anaerobic codigestionof the mixtures of slaughterhouse with OFMSW in mesophilic con-ditions for HRTs of between 25 d and 50 d (OLR of between 0.90 gand 3.70 g TVS/l/d). The decomposition of CODD was in the range63–85%, except at Run VIII (33%) (Table 1). These results are in linewith those obtained by Ueno et al. (2007) (79%) in two-phase ther-mophilic AD from organic waste for HRT of between 4.3 d and 6.8 d(OLR of between 12.4 g and 16.6 g COD/l/d). On the other hand, theremoval percentages of sulphate (Table 1) were increased from27% to a maximum of 65% according to increase in SP.

In relation to VFA consumption, the main VFA consumed wasbutyric acid. It was according to others studies about two-phaseAD of solid waste (Raynal et al., 1998). The removal percentagesof butyric, acetic and TVFA were ranges 86–100%, 73–97% and69–100% respectively, except at Run VIII, in which they were15%, 58% and 27% respectively. The amount of propionic generatedfor all OLRs assayed (0–2 g/l) did not produce an inhibitory effect.Non-toxicity of propionic generated and the high consumption ofbutyric acid, suggesting that PUAs and BUAs levels (Table 5)have been sufficient to achieve low propionic and butyricconcentrations in steady state conditions (except in 2 d HRT(OLR = 33 g TVS/l/d)). It should be noted that in steady conditionsthe increase of TVFA was accompanied by an increase of the total

microbial communities in the reactor (Table 4 and 5). The similarrelationship was found by Cardinali-Rezende et al. (2012).

The biogas produced was composed of CH4, CO2 and H2S, withno H2 being detected under any of the tested conditions. It wasdue to in the second phase the operational conditions (neutral orslightly basic pH) are favourable to the growth of HUMs and there-fore the H2 generated in acidogenesis and/or acetogenesis is con-sumed quickly by HUMs. The results of GP, percentage of CH4 inbiogas (% CH4), percentage of H2S in biogas (% H2S), SMP and SPare shown in Table 4. The percentages of CH4 were in the range61–71%, except at Run VIII (54 ± 1%), at which a decrease inArchaea content compared to Run VII was detected. The MP values(1.5 l CH4/l/d) obtained by Cavinato et al. (2011) in thermophilicAD of biowaste employing CSTRs at an OLR of 10 g TVS/l/d waslower than those obtained in the present study (2.3 ± 0.1 l CH4/l/d)at the same OLR.

Increasing OLRs resulted in an increase in MP and SP (Fig. 2) ex-cept at Run VIII. The highest values of GP (8.2 ± 0.9 l/l/d), SP(82 ± 9 ml H2S/l/d), MP (5.4 ± 0.3 l CH4/l/d) and SMP (249 ± 18 lCH4/kg TVS) were obtained to OLR of 22 g TVS/l/d or HRT of 3 d(Run VII). These results are similar to those obtained by Ueno et al.(2007) (5.7 l CH4/l/d) at 5.5 d HRT (OLR = 16.6 g COD/l/d).

A summary to the main microbial groups involved in the secondphase reactor is showed in Table 5. The amount of microorganismsin the second phase was higher than those obtained in the firstphase. It is due to the several conditions existent (low pH andHRT) in the first phase. The proportion of Eubacteria in the secondphase was lower than those obtained in the first phase, even lowerthan those obtained by Xiao et al. (2013) (90.4%) in the methano-genic phase of the two phase system.

At Run V was observed a large shift in behaviour on the secondphase caused by the HRT (1.5 d) assayed in the first phase is insuf-ficient to acidogenic stage finish in the first reactor, as conse-quently of this, it produces an effluent suitability for theacidification in the second phase. One could say that acetogenic–methanogenic reactor (first period) (Run I, II, III and IV) wastransformed into an acidogenic–acetogenic–methanogenic reactor(second period) (Run V, VI, VII and VIII). It was according toincrease in relation of Eubacteria:Archaea and percentages of HABsat the second period. HABs were from non-detectable (acetogenic–methanogenic reactor) to detectable (acidogenic–acetogenic–methanogenic reactor). Therefore, the shift in behaviour on thesecond phase was the reason why in the first period the averagevalues of the relation of Eubacteria:Archaea (values close to 1)was lower than those in the second period (>1) which were in linewith those obtained by Montero et al. (2008) in single phase ther-mophilic AD from OFMSW.

Increasing OLR resulted in an increase in Eubacteria (except atan OLR of 33 g TVS/l/d (HRT = 2 d), particularly it was due to an in-crease in acetogens and HABs levels in the first and second period,respectively. At an OLR of 33 g TVS/l/d (HRT = 2 d) the destabilisa-tion of the system takes place and the wash out of the microorgan-isms occurs (except to SRBs).

In Run V a significant decrease in the microbial population (ex-cept to HABs) took place as the result of the shift in behaviour onthe second phase. The observed increase in SRBs and Archaea be-tween the Run V and VII d was due to increased microbial activityallows these microorganisms to have higher growth rates. Thechange in behaviour on the second phase did not affect MP or SP.In fact, both the MP and SP increase with the applied OLR(Fig. 2). These results seem to show on the one hand, that the activ-ity of anaerobic microorganisms in the reactor could be more re-lated to the OLR than to microbial concentrations and, on theother hand that not only the stability, but also the adequatedynamics of the microbial community (flexibility to adapt inresponse to changes in environments in particular to change in

Table 4Characterisation of reactors effluents and yields of the process.

Parameter I II III IV V VI VII VIII

Characterisation of the first phase reactorpH 5.6 ± 0.2 5.4 ± 0.3 5.3 ± 0.4 5.5 ± 0.4 5.3 ± 0.4 5.5 ± 0.3 5.2 ± 0.4 5.0 ± 0.2CODD (g/l) 29 ± 2 33 ± 3 32 ± 2 34 ± 1 36 ± 2 36 ± 2 39 ± 1 39 ± 1TVS (g/kg) 51 ± 3 72 ± 3 70 ± 4 64 ± 2 55 ± 3 55 ± 3 65 ± 4 65 ± 4Sulphate (g/l) 1.5 ± 0.0 1.8 ± 0.1 1.8 ± 0.2 1.7 ± 0.2 2.0 ± 0.2 2.0 ± 0.6 2.3 ± 0.5 2.3 ± 0.5Alkalinity (g CaCO3/l) 5 ± 1 6 ± 1 6 ± 2 6 ± 2 4 ± 0 4 ± 0 4 ± 0 4 ± 0TVFA (g acetic/l) 12.2 ± 1.8 15.1 ± 1.2 14.3 ± 2.2 17.0 ± 0.9 8.8 ± 1.0 8.8 ± 1.0 9.6 ± 1.0 9.6 ± 1.0Acetate (g/l) 3.7 ± 0.3 3.5 ± 0.4 3.5 ± 0.4 4.6 ± 0.4 2.3 ± 0.5 2.3 ± 0.5 2.6 ± 0.7 2.6 ± 0.7Propionate (g/l) 0.6 ± 0.1 0.4 ± 0.1 0.4 ± 0.1 0.5 ± 0.1 0.1 ± 0.0 0.1 ± 0.0 0.1 ± 0.0 0.1 ± 0.0Butyrate (g/l) 3.8 ± 0.4 5.0 ± 1.0 4.5 ± 2.0 2.0 ± 0.1 3.8 ± 0.8 3.8 ± 0.8 4.0 ± 0.8 4.0 ± 0.8Hydrolysis yield (%) 40 45 45 47 52 52 57 57Acidogenic yield (%) 48 53 53 53 27 27 27 27

Characterisation of the second phase reactorpH 7.8 ± 0.1 8.0 ± 0.2 7.6 ± 0.1 7.6 ± 0.3 7.7 ± 0.2 7.5 ± 0.1 7.5 ± 0.1 7.0 ± 0.4CODD (g/l) 4 ± 0 10 ± 1 10 ± 1 12 ± 1 9 ± 1 10 ± 1 14 ± 0 26 ± 1TVS (g/kg) 14 ± 3 11 ± 1 17 ± 3 17 ± 3 12 ± 3 12 ± 2 14 ± 1 28 ± 2Sulphate (g/l) 1.1 ± 0.1 1.0 ± 0.4 0.8 ± 0.2 0.8 ± 0.2 0.7 ± 0.2 0.7 ± 0.1 0.8 ± 0.0 1.2 ± 0.4Alkalinity (g CaCO3/l) 6 ± 0 7 ± 1 7 ± 2 6 ± 2 5 ± 0 5 ± 0 4 ± 0 4 ± 0TVFA (g acetic/l) 0.0 ± 0.0 1.4 ± 0.3 2.2 ± 0.1 4.6 ± 0.3 2.2 ± 0.4 2.5 ± 0.1 3.0 ± 0.5 7.0 ± 1.9Acetate (g/l) 0.1 ± 0.0 0.8 ± 0.3 0.2 ± 0.1 1.2 ± 0.6 0.7 ± 0.2 0.2 ± 0.2 0.7 ± 0.1 1.1 ± 0.6Propionate (g/l) 0.0 ± 0.0 0.3 ± 0.1 0.3 ± 0.3 2.1 ± 0.2 0.8 ± 0.1 1.3 ± 0.2 1.2 ± 0.1 1.2 ± 0.1Butyrate (g/l) 0.0 ± 0.0 0.1 ± 0.0 0.7 ± 0.1 0.1 ± 0.0 0.1 ± 0.0 0.0 ± 0.0 0.0 ± 0.0 3.4 ± 1.1

First phase reactor yieldsGP (l/l/d) 0.4 ± 0.1 0.8 ± 0.2 0.7 ± 0.1 1.1 ± 0.2 4.2 ± 0.2 3.7 ± 0.2 4.0 ± 0.2 4.0 ± 0.2H2% 27 ± 4 30 ± 5 30 ± 5 38 ± 5 48 ± 2 49 ± 2 47 ± 3 47 ± 3HP (lH2/l/d) 0.1 ± 0.0 0.2 ± 0.0 0.2 ± 0.0 0.4 ± 0.0 2.0 ± 0.1 1.8 ± 0.2 1.9 ± 0.2 1.9 ± 0.2SHP (l H2/kg TVS) 13 ± 3 20 ± 4 20 ± 4 23 ± 4 35 ± 4 32 ± 4 33 ± 4 33 ± 4

Second phase reactor yieldsGP (l/l/d) 1.1 ± 0.1 1.7 ± 0.3 1.9 ± 0.2 3.7 ± 0.2 2.3 ± 0.2 4.2 ± 0.2 8.2 ± 0.9 6.7 ± 0.7CH4% 68 ± 1 64 ± 2 66 ± 3 62 ± 1 67 ± 4 65 ± 2 66 ± 2 54 ± 1H2S% 0.8 ± 0.2 0.8 ± 0.2 0.8 ± 0.2 0.9 ± 0.3 0.8 ± 0.1 1.1 ± 0.2 1.0 ± 0.2 0.7 ± 0.1MP (l CH4/l/d) 0.7 ± 0.0 1.1 ± 0.2 1.2 ± 0.2 2.3 ± 0.1 1.5 ± 0.2 2.7 ± 0.2 5.4 ± 0.3 3.8 ± 0.8SP (ml H2S/l/d) 9 ± 1 14 ± 1 15 ± 1 33 ± 8 18 ± 4 46 ± 8 82 ± 9 47 ± 5SMP (l CH4/kg TVS) 219 ± 21 238 ± 50 177 ± 49 235 ± 15 174 ± 24 194 ± 19 249 ± 18 116 ± 35

Two-phase yieldsGP (l/l/d) 1.5 ± 0.2 2.5 ± 0.5 2.6 ± 0.3 4.8 ± 0.4 6.5 ± 0.4 7.9 ± 0.4 12.2 ± 1.1 10.7 ± 0.9% H2 6.7 ± 1.0 10.8 ± 1.8 10.6 ± 1.7 8.8 ± 1.4 30.8 ± 1.3 23.3 ± 1.0 15.5 ± 1.0 17.7 ± 1.1% CH4 46.7 ± 0.8 44.0 ± 1.4 47.8 ± 2.0 47.9 ± 0.8 23.1 ± 1.4 34.2 ± 1.1 44.3 ± 1.3 35.5 ± 0.7% H2S 0.6 ± 0.2 0.6 ± 0.1 0.6 ± 0.1 0.7 ± 0.2 0.3 ± 0.0 0.6 ± 0.1 0.7 ± 0.1 0.4 ± 0.1% CO2 46.1 ± 1.9 44.6 ± 3.3 43.7 ± 3.8 42.6 ± 2.4 45.9 ± 2.7 41.9 ± 2.1 39.6 ± 2.5 46.4 ± 1.8

All the values corresponding to the analytical determinations in steady conditions (except at second phase in Run VIII). The number of replicas to biogas and pH analyticaldeterminations were 75, 65, 50, 33, 24, 18, 13 and 3 times for Run I, II, III,IV, V, VI, VII and VIII, respectively; and to VFA, CODD, sulphate, alkalinity and TVS analyticaldeterminations were 21, 19, 21, 14, 17, 13, 9 and 3 times for Run I, II, III,IV, V, VI, VII and VIII, respectively.

448 S. Zahedi et al. / Bioresource Technology 148 (2013) 443–452

substrate (caused by the HRT assayed in the first phase) and oper-ating conditions (HRT or OLR)) are important factors for the stableperformance of the reactors. The similar relationship can be foundin the literature (Ayala-del-Río et al., 2004; Fernández et al., 1999;Miura et al., 2007). For example, a study performed with a func-tionally stable methanogenic reactor revealed that microbial com-munities could be very dynamic; suggesting that stable function isnot always correlated with stable community structure (Fernándezet al., 1999).

In mentioning to acetogens, the average value of the proportionof BUA/PUA presented an increasing trend from 0.8 to 1.7 when theHRT was decreased or OLR was increased, which were in line withthose obtained by Zahedi et al. (2013) in single phase thermophilicAD from OFMSW. Even at an OLR of 33 g TVS/l/d (HRT = 2 d) wherethe destabilisation of the system takes place and the wash outof the microorganisms occurs (except to SRBs), the average valueof the proportion of BUA/PUA remained at 1.6. It was accordingto the high amount of butyric in the substrate at all runs tested.The low VFA values seem to indicate that acetogens (higher than26% at all runs) may have been sufficient to achieve low VFAconcentrations, except to Run VIII.

Regarding the Archaea population, the average values of the ra-tio HUM/AUM during all study was higher than 0.9 for all tested

conditions. This fact is due to two reasons, the first is the low HRTsassayed and the second is the high amount of butyric in the sub-strate. Acidogenic and acetogenic process produces large amountsof H2, which should be consumed quickly by HUMs because acet-ogens and AUMs do not grow well in presence of H2 in the system(Boone and Bryant, 1980; Montero et al., 2008). Therefore, low ra-tios HUM/AUM will indicate that the H2 generated would be accu-mulated in the system, preventing the activity of the AUMs andacetogens. In fact, has been demonstrated that increase in HUMsimprove the stability of the AD process (Ghanimeh et al., 2013).Ghanimeh et al. (2013) studied the potential of improving the sta-bility of thermophilic AD of source-sorted OFMSW by addingleachate and compost during inoculation, because the methano-genic populations in municipal waste compost and landfill leach-ate are dominated by thermophilic hydrogenotrophs.

The SRBs population was variable and mainly higher than thoseobtained by Zhang et al. (2011) after feeding sulphate in the reac-tors (28.6%) and Zahedi et al. (2013) (17.3–23.5%). Nevertheless, itshould be noted the high proportion of SRBs (60%) at the Run VIII,compared to those average values (between 26% and 44%) at theothers runs tested. The population growth of SRBs in the HRT of2 d (OLR = 33 g TVS/l/d) was due to two reasons, that cancels dailyremoval (wash) from the system. The first reason was the constant

Table 5Characterisation of the different populations in the first and second phase reactor.

I II III IV V VI VII VIII

First phase reactorTotal population (108 cells/ml) 2.9 ± 0.3 4.3 ± 0.8 4.1 ± 0.8 7.7 ± 0.4 8.3 ± 0.2 8.5 ± 0.3 9.1 ± 0.3 9.1 ± 0.3Eubacteria (%) 82 ± 2 84 ± 3 83 ± 3 91 ± 2 88 ± 2 88 ± 2 88 ± 2 88 ± 2HABs (%) 49 ± 5 54 ± 4 53 ± 3 75 ± 2 69 ± 2 69 ± 3 69 ± 1 69 ± 1Acetogens (%) 33 ± 3 30 ± 4 30 ± 4 16 ± 3 19 ± 2 19 ± 2 19 ± 2 19 ± 2BUAs 16 ± 3 14 ± 3 15 ± 3 8 ± 2 10 ± 3 9 ± 2 9 ± 1 9 ± 1PUAs (%) 17 ± 0 16 ± 1 15 ± 1 8 ± 2 10 ± 3 9 ± 2 10 ± 1 10 ± 1SRBsa (%) 21 ± 1 14 ± 0 14 ± 0 14 ± 3 16 ± 3 15 ± 3 14 ± 1 14 ± 1Archaea (%) 18 ± 2 16 ± 3 17 ± 3 9 ± 2 12 ± 2 12 ± 3 12 ± 1 12 ± 1AUMs (%) 10 ± 1 8 ± 1 8 ± 2 4 ± 1 5 ± 0 5 ± 1 5 ± 1 5 ± 1HUMs (%) 9 ± 1 8 ± 2 9 ± 1 5 ± 1 7 ± 2 7 ± 2 7 ± 0 7 ± 0

Second phase reactorTotal population (108 cells/ml) 9.3 ± 1.1 11.7 ± 1.5 19.0 ± 1.2 28.1 ± 3.6 16.9 ± 2.3 22.9 ± 4.2 29.4 ± 3.5 20.1 ± 2.2Eubacteria (%) 41 ± 7 50 ± 6 49 ± 4 47 ± 6 66 ± 5 67 ± 7 70 ± 4 75 ± 6HABs (%) 0 0 0 0 4 ± 1 17 ± 5 41 ± 1 32 ± 1Acetogens (%) 41 ± 6 50 ± 5 49 ± 4 47 ± 5 63 ± 4 50 ± 2 30 ± 5 43 ± 5BUAs 18 ± 3 22 ± 3 25 ± 3 26 ± 3 37 ± 3 27 ± 1 19 ± 3 26 ± 2PUAs (%) 23 ± 3 27 ± 2 25 ± 1 20 ± 2 25 ± 1 23 ± 1 11 ± 2 16 ± 3SRBsa (%) 34 ± 4 39 ± 5 37 ± 4 31 ± 4 30 ± 2 28 ± 2 29 ± 2 60 ± 6Archaea (%) 59 ± 7 50 ± 6 51 ± 4 53 ± 8 34 ± 5 33 ± 7 30 ± 4 25 ± 6AUMs (%) 29 ± 5 20 ± 1 19 ± 1 26 ± 4 17 ± 2 17 ± 3 14 ± 2 13 ± 3HUMs (%) 30 ± 1 30 ± 5 32 ± 3 28 ± 4 17 ± 2 15 ± 4 16 ± 2 13 ± 3

The number of replicas analysed for each condition in steady conditions (except at second phase in Run VIII) were 11, 9, 7, 5, 7, 5, 4 and 3 times for Run I, II, III,IV, V, VI, VII andVIII, respectively.

a Percentages compared to total Eubacteria.

S. Zahedi et al. / Bioresource Technology 148 (2013) 443–452 449

input of these microorganisms through the substrate and thesecond, the high duplication speeds of them.

To evaluate the biochemical activity in the second phase, theArchaea and SRB activity parameter have been considered, becausealthough the determination of the number of microorganisms isimportant in many microbial ecology studies (Zabriskie andHumphrey, 1978; Hanning et al., 2007), these papers do not assessthe activities associated with the total population. In fact, in thepresent paper, the highest values of SRBs and Archaea wereobtained at Run VIII and Run IV, respectively and it was not in linewith the maximum values of SP and MP (Run VII). There is a highcorrelation between OLR and both activities, except at the maxi-mum OLR tested (33 g TVS/l/d) (Fig. 3). This study shows that theincrease in microbial activity inside the reactor is directly propor-tional to the OLR (or inversely proportional to the HRT).

As commented in introduction section, two stages of inhibitionexist as a result of sulphate reduction (Chen et al., 2008); primaryinhibition is due to competition for common organic and inorganicsubstrates from SRBs and secondary inhibition results from thetoxicity of sulphide to various microbial groups. In referring tocircumstances of toxicity of sulphide, the greater value ofMP (5.4 ± 0.3 l CH4/l/d), SP (82 ± 9 ml H2S/l/d), Archaea activity(62 ± 3 � 10�13 l CH4/cell/d) and SRB activity (1.0 ± 0.1 �10�13 l H2S/cell/d) were obtained at the same OLR (22 g TVS/l/d)corresponding to 3 d HRT (lower than doubling time of acetogensand methanogens microorganisms); this suggests that the not inhi-bition from sulphide toxicity to microorganisms present in thereactor. In referring to circumstances of competition for substrates,the average values of the ratio acetogens/SRBs and Archaea/SRBsfrom Run I (15 d HRT or OLR of 3 g TVS/l/d) to Run VII (3 d HRTor OLR of 22 g TVS/l/d) were 1.5 ± 0.4 and 1.4 ± 0.3 respectively,indicating methanogens/acetogens domination over SRBs, how-ever, the presence of H2S in the sample gas was indicated thatsome H2 or VFA were consumed by SRBs rather than by acetogensor Archaea. At Run VIII (2 d HRT or OLR of 33 g TVS/l/d) the ratioacetogens/SRBs and Archaea/SRBs were decreased to 0.7 and 0.4,respectively. This indicates that, for HRTs much lower than the dou-bling time of acetogens and methanogens (2 d), SRBs were

dominant over those populations (acetogens and methanogens)and the reason was simply due to the shorter doubling time of SRBs.

The optimal HRT (3 d) in order to maximise the MP in the sec-ond phase of a two-stage thermophilic AD process of sulphate-con-taining municipal solid waste was the same to those obtained in asingle-phase dry-thermophilic AD under high OLRs of sulphate-containing municipal solid waste (Zahedi et al., 2013), however,under optimal conditions, some differences in the proportion ofmicroorganisms have been found between both. In a single phasethe average values of the ratios of Eubacteria:Archaea, HABs:aceto-gens, PUAs:BUAs and HUMs:AUMs were 87:13, 72:16, 6:10 and6:7, respectively and the relative SRBs percentage was 21% whilein the second phase of a two-stage the average values of the ratiosof Eubacteria:Archaea, HABs:acetogens, PUAs:BUAs and HUMs:AUMswere 70:30, 41:30, 11:19 and 16:14, respectively and the relativeSRBs percentage was 29%. It should be noted the high average valuesof the ratios of Eubacteria:Archaea and HABs:acetogens, in the singlephase, compared to those average values at the second phase of atwo-stage. It is due to substrate employed in the second phase waspreviously hydrolysed by HABs in the first phase and therefore thesecond phase of a two-stage required a lower proportion of HABsto ensure the proper functioning of the AD process.

It is necessary to emphasize that although the proportion ofmicroorganisms in the reactor is a key, to establish the optimalproportion in the AD process from real urban solid waste is verydifficult, because the OFMSW had a high variability. Changes inwaste composition can be produced by several factors, includingclimate, collection frequency and seasonal practices (Tchobanog-lous et al., 1997). These changes can affect to the microbial popu-lation inside the reactor (Zahedi et al., 2013). Every systempresents an optimum proportion of microorganisms and it isdependent on the waste characteristics, the source of inoculumused, the percentage in TS in the feed, HRT tested in the process,etc. and, hence, should be determined specifically in each case.

In fact, the present paper, as noted above, suggest that not onlythe stability, but also the adequate dynamics (‘‘flexibility’’) of themicrobial community structure are important for the stable perfor-mance of the reactors treating OFMSW.

Fig. 1. (a) Evolution of HP (l H2/l/d) and MP (l CH4/l/d) from t = 0 to t = 257 d. (b) Evolution of HP (l H2/l/d) and MP (l CH4/l/d) from t = 258 to t = 398 d.

Fig. 2. Relationship between OLR (g TVS/l/d) and MP (l CH4/l/d) and SP (ml H2S/l/d).Fig. 3. Relationship between OLR (g TVS/l/d) and Archaea activity (l CH4/cell/d) andSRB activity (l H2S/cell/d).

450 S. Zahedi et al. / Bioresource Technology 148 (2013) 443–452

3.4. Optimal conditions of the two-phase dry-thermophilic AD processof sulphate-containing OFMSW

Two-phase dry-thermophilic AD process to the sequential HPand MP of sulphate-containing municipal solid waste was carriedout at eight different HRTs (from 25 d to 3.5 d) or eight OLRs (from3 g to 24 g TVS/l/d) (Table 1). The biogas produced was composedof H2, CH4, CO2 and H2S (Table 4). With regards to decompositionof TVS, the average values were 83 ± 4% (Table 1), except at Run

VIII (67%). These results are slightly lower with those (95%) ob-tained by Chu et al. (2008) in the two-phase AD of food waste. Itis due to the diluted food waste employed by these authors, withTS concentration of 10.7–12.8%. In the present study, the TS con-centration of the feed first reactor was adjusted to 20%. GP pre-sented an increasing trend when the HRT was decreased or OLRwas increased (Table 4), except at Run VIII (destabilisation in thesecond phase was observed). As has been discussed in previoussections, behavioural differences in the first and second phase were

S. Zahedi et al. / Bioresource Technology 148 (2013) 443–452 451

found in the present study. On one hand in the first period, fromRun I to Run IV, the behaviour in the first phase was hydrolytic–acidogenic, while for the second phase was acetogenic–methanogenic. On the other hand in the second period, from RunV to Run VIII, the behaviour in the first phase was mainly hydrolytic,while for the second phase was acidogenic–acetogenic–methanogenic. It was according to microbial population dynamics.

To feasibility study for semi-continuous HP and MP in two-phasedry-thermophilic AD of sulphate-containing municipal solid wastehas been considered all steady conditions tested (except Run VIII).Obtained gases can be used separately or mixed together to obtainbiohythane (Cavinato et al., 2011). With regard to biohytane, severalstudies (Porpatham et al., 2007; Akansu et al., 2007; Rakopoulos andMichos, 2009) have demonstrated that the use of mixed gases allowsa better combustion with a reduced greenhouse gasses emissioncompared with fossil fuels. The amount of H2 in the mix must beabove 10–20%, major quantity could not assure the best perfor-mance of engine and of emissions. This indicates that all conditionstested in the present research can be taken into account for biohy-tane energy, although Run V and VI could not assure the best perfor-mance of engine and of emissions. Therefore, GP and microbialactivities, in particular, HAB and Archaea activities, were the param-eters selected to determine the optimum operating condition forsemi-continuous HP and MP in both case (for use separated gasesor mixed together). The highest values of GP (12.2 ± 1.1 l/l/d), HP(1.9 ± 0.2 l H2/l/d), MP (5.4 ± 0.3 l CH4/l/d), HAB activity(30 ± 3 � 10�13 l H2/cell/d) and Archaea activity (62 ± 3 �10�13 l CH4/cell/d) were obtained to OLR of 19 g TVS/l/d or HRT of4.5 d (Run VII), in which basically comprised hydrolysis step in thefirst phase (HRT = 1.5 d), where the average values of the ratios ofEubacteria:Archaea, acetogens:HABs, PUAs:BUAs and HUMs:AUMswere 88:12, 19:69, 10:9 and 7:5, respectively and the relative SRBspercentage was 15%; and the other three steps (acidogenic, aceto-genic and methanogenic) in the second phase (HRT = 3 d), wherethe average values of the ratios of Eubacteria:Archaea, aceto-gens:HABs, PUAs:BUAs and HUMs:AUMs were 70:30, 30:41, 11:19and 16:14, respectively and the relative SRBs percentage was 29%.At Run VII, the mixtures of gas obtained from the two reactors wereof 15.5% H2, 39.5% CO2, 0.7% H2S and 44.3% CH4.

In short, the novelties and relevancies of the present paper are:(1) the OFMSW used to feed proceeds of the municipal solid wastetreatment plant; these real urban wastes contain high amount ofsulphate, typically higher than 5 g/kg; (2) the overall duration ofthe experiments (398 d) has allowed considering the changes inwaste composition produced by several factors, including climate,collection frequency and seasonal practices; (3) the high microbialcontent of the OFMSW affect to AD process; (4) microbial contentinside the reactor, in both first and second phase not assess theactivities associated with the microbial population in the systems.This depends on the conditions of the medium (mainly HRT, OLR,pH, acidity and H2 partial pressure); (5) the contribution of indig-enous flora in the substrate and increasing in microbial activity(consequence of gradually OLR increasing or HRT decreasing) indi-cates that two-phase dry-thermophilic AD process of sulphate-containing municipal solid waste may be carried out from HRTslower than doubling time of methanogens and acetogens untilthe constant inputs of these microorganisms through substratewere not been high enough to compensate the microorganismsexiting the system daily in the effluent; (6) not only the stability,but also the adequate dynamics of the microbial community (flex-ibility to adapt in response to changes in environments, in partic-ular to change in substrate and operating conditions (HRT orOLR)) are important factors for the stable performance of the reac-tors and (7) optimal conditions (maximum of GP and microbialactivities) in the two-phase dry-thermophilic AD process of sul-phate-containing municipal solid waste were obtained when the

behaviour in the first phase was mainly hydrolytic, while for thesecond phase was acidogenic–acetogenic–methanogenic.

4. Conclusions

Optimal conditions (maximum of GP and microbial activities)comprised hydrolysis step and a not-finished acidogenic step inthe first phase (HRT = 1.5 d) in which the percentage of Eubacteriawas 88%. The acidogenic step finished in the second phase as wellas acetogenic–methanogenic steps (HRT = 3 d), where the relativepercentages of Eubacteria:Archaea; HUMs:AUMs; HABs:acetogensand SRBs were 70:30; 16:14, 41:30 (BUAs 19% and PUAs 11%)and 29%, respectively. In steady state conditions both acetogensand Archaea were dominated over SRBs. Decreasing HRT resultedin an increase in microbial activities. The inhibition by sulphatereduction seems to be not significant.

Acknowledgements

This work was supported by Spanish Ministry of Science andInnovation (MICINN) (CTM2007-62164), Spanish Ministry ofEconomy and Competitiveness (CTM2010-17654) and theInnovation, Science and Enterprise Department (AndalusiaRegional Government) (P07-TEP-02472). All projects co-financedby European Regional Development Fund (ERDF). S. Zahedi thanksfunding from Spanish Ministry of Science and Innovation (MICINN)(AP2008-01213).

References

Akansu, A.O., Kahraman, N., Ceper, B., 2007. Experimental study on a spark ignitionengine fuelled by methane–hydrogen mixtures. Int. J. Hydrogen Energy 32,4279–4284.

APHA, AWWA, WEF, 1995. Standard Methods for the Examination of Water andWastewater, 19th ed. Washington, DC, New York, USA.

Ayala-del-Río, H.L., Callister, S.J., Criddle, C.S., Tiedje, J.M., 2004. Correspondencebetween community structure and function during succession in phenol- andphenol-plus-trichloroethene-fed sequencing batch reactors. Appl. Environ.Microbiol. 70, 4950–4960.

Boone, R.D., Bryant, L., 1980. Propionate-degrading bacterium, Syntrophobacterwolinii sp. nov. gen. nov., from methanogenic ecosystems. Appl. Environ.Microbiol. 40, 626–632.

Boone, R.B., Johnson, R.L., Liu, Y., 1989. Diffusion of the interspecies electron carriersH2 and formate in methanogenic ecosystems and its implications in themeasurement of Km for H2 or formate uptake. Appl. Environ. Microbiol. 55,1735–1741.

Cardinali-Rezende, J., Colturato, L.F.D.B., Colturato, T.D.B., Chartone-Souza, E.,Nascimento, A.M.A., Sanz, J.L., 2012. Prokaryotic diversity and dynamics in afull-scale municipal solid waste anaerobic reactor from start-up to steady-stateconditions. Bioresour. Technol. 119, 373–383.

Cavinato, C., Bolzonella, D., Fatone, F., Cecchi, F., Pavan, D., 2011. Optimization oftwo-phase thermophilic anaerobic digestion of biowaste for hydrogen andmethane production through reject water recirculation. Bioresour. Technol.102, 8605–8611.

Chen, Y., Cheng, J.J., Creamer, K.S., 2008. Inhibition of anaerobic digestion process: areview. Bioresour. Technol. 99, 4044–4064.

Chu, C.F., Li, Y.Y., Xu, K.Q., Ebie, Y., Inamori, Y., Kong, H.N., 2008. A pH–temperaturephased two-stage process for hydrogen and methane production from foodwaste. Int. J. Hydrogen Energy 33, 4739–4746.

Cuetos, M.J., Gómez, X., Otero, M., Morán, A., 2008. Anaerobic digestion of solidslaughterhouse waste (SHW) at laboratory scale: influence of co-digestion with theorganic fraction of municipal solid waste (OFMSW). Biochem. Eng. J. 40, 99–106.

De la Rubia, M.A., Raposo, F., Rincón, B., Borja, R., 2009. Evaluation of the hydrolytic–acidogenic step of a two-stage mesophilic anaerobic digestion process ofsunflower oil cake. Bioresour. Technol. 100, 4133–4138.

Demirer, G.N., Chen, S., 2004. Effect of retention time and organic loading rate onanaerobic acidification and biogasification of dairy manure. J. Chem. Technol.Biotechnol. 79, 1381–1387.

Demirel, B., Yenigun, O., 2004. Anaerobic acidogenesis of dairy wastewater: theeffects of variations in hydraulic retention time with no pH control. J. Chem.Technol. Biotechnol. 79, 755–760.

Fdez-Güelfo, L.A., Álvarez-Gallego, C.J., Sales, D., Romero, L.I., 2010. Drythermophilic anaerobic digestion of organic fraction of municipal solid waste:methane production modeling. Waste Manage. 32, 382–388.

Fernández, A., Huang, S., Seston, S., Xing, J., Hickey, R., Criddle, C., Tiedje, J., 1999.How stable is stable? function versus community composition. Appl. Environ.Microbiol. 65, 3697–3704.

452 S. Zahedi et al. / Bioresource Technology 148 (2013) 443–452

Ghanimeh, S., El-Fadel, M., Saikaly, P., 2013. Improving the stability of thermophilicanaerobic digesters treating SS-OFMSW through enrichment with compost andleachate seeds. Bioresour. Technol. 131, 53–59.

Griffin, M.E., McMahon, K.D., Mackie, R.I., Raskin, L., 1998. Methanogenic populationdynamics during start-up of anaerobic digesters treating municipal solid wasteand biosolids. Biotechnol. Bioeng. 57, 342–355.

Hanning, C., Hanning, M., Rehmer, O., Braun, G., Hellwing, E., Al-Ahmad, A., 2007.Fluorescence microscopic visualization and quantification of initial bacterialcolonization on enamel in situ. Arch. Oral Biol. 52, 1048–1056.

Liu, P., Qiu, Q., Lu, Y., 2011. Syntrophomonadaceae-affiliated species as activebutyrate-utilizing syntrophs in paddy field soil. Appl. Environ. Microbiol. 77,3884–3887.

Mara, D., Horan, N., 2003. Handbook of Water and Wastewater Microbiology.Elsevier, Ltd..

McMahon, K.D., Stroot, P.G., Mackie, R.I., Raskin, L., 2001. Anaerobic codigestion ofmunicipal solid waste and biosolids under various mixing conditions—II.Microbial population dynamics. Water Res. 35, 1817–1827.

Miura, Y., Hiraiwa, M.N., Ito, T., Itonaga, T., Watanabe, Y., Okabe, S., 2007. Bacterialcommunity structures in MBRs treating municipal wastewater: relationshipbetween community stability and reactor performance. Water Res. 41, 627–637.

Montero, B., García-Morales, J.L., Sales, D., Solera, R., 2008. Evolution ofmicroorganisms in thermophilic-dry anaerobic digestion. Bioresour. Technol.99, 3233–3243.

Montero, B., García-Morales, J.L., Sales, D., Solera, R., 2009. Analysis of methanogenicactivity in a thermophilic-dry anaerobic reactor: use of fluorescent in situhybridization. Waste Manage. 29, 1144–1151.

Porpatham, E., Ramesh, A., Nagalingam, B., 2007. Effect of hydrogen addition on theperformance of a biogas fuelled spark ignition engine. Int. J. Hydrogen Energy32, 2057–2065.

Rakopoulos, C.D., Michos, C.N., 2009. Generation of combustion irreversibilities in aspark ignition engine under biogas–hydrogen mixtures fueling. Int. J. HydrogenEnergy 34, 4422–4437.

Raynal, J., Delgenes, J.P., Moletta, R., 1998. Two-phase anaerobic digestion of solidwastes by a multiple liquefaction reactors process. Bioresour. Technol. 65, 97–103.

Romero, M.A., Fdez-Güelfo, L.A., Álvarez-Gallego, C.J., Romero, L.I., 2013. Effect ofHRT on hydrogen production and organic matter solubilization in acidogenicanaerobic digestion of OFMSW. Chem. Eng. J. 219, 443–449.

Solera, R., Romero, L.I., Sales, D., 2001. Analysis of the methane production inthermophilic anaerobic reactors: use of autofluorescence microscopy.Biotechnol. Lett. 23 (22), 1889–1892.

Tchobanoglous, G., Hilary, T., Vigil, S.A., 1997. Integrated Solids Waste Management:Engineering Principles and Management Issues. McGraw-Hill, Inc..

Ueno, Y., Fukui, H., Goto, M., 2007. Operation of a two-stage fermentation processproducing hydrogen and methane from organic waste. Environ. Sci. Technol. 41,1413–1419.

Xiao, K., Guo, C.H., Zhou, Y., Maspolim, Y., Wang, J.Y., Ng, W.J., 2013. Acetic acidinhibition on methanogens in a two-phase anaerobic process. Biochem. Eng. J.75, 1–7.

Zabriskie, D.W., Humphrey, A.E., 1978. Estimation of fermentation biomassconcentration by measuring culture fluorescence. Appl. Environ. Microbiol.35, 337–343.

Zahedi, S., Sales, D., Romero, L.I., Solera, R., 2013. Optimisation of single-phase dry-thermophilic anaerobic digestion under high organic loading rates of industrialmunicipal solid waste: population dynamics. Bioresour. Technol. 146, 109–117.

Zhang, J., Zhang, Y., Quan, X., Liu, Y., An, X., Chen, S., Zhao, H., 2011. Bioaugmentationand functional partitioning in a zero valent iron-anaerobic reactor for sulfate-containing wastewater treatment. Chem. Eng. J. 174, 159–165.