Chemical Engineering Journal 219 (2013) 443–449

Contents lists available at SciVerse ScienceDirect

Chemical Engineering Journal

journal homepage: www.elsevier .com/locate /cej

Effect of HRT on hydrogen production and organic matter solubilizationin acidogenic anaerobic digestion of OFMSW

1385-8947/$ - see front matter � 2013 Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.cej.2012.12.090

Abbreviations: AD, anaerobic digestion; ASC, acidogenic substrate as carbon;CSTR, continuous stirred tank reactor; DAC, dissolved acid carbon; DOC, dissolvedorganic carbon; HP, hydrogen production; HRT, hydraulic retention time; MBT,mechanical–biological-treatment; WWTP, wastewater treatment plant; MW,molecular weight; OFMSW, organic fraction of municipal solid waste; OLR, organicloading rate; SHP, specific hydrogen production; TCD, thermal conductivitydetector; TS, total solids; VFA, volatile fatty acids; VS, volatile solids; MSW,municipal solid waste.⇑ Corresponding author. Tel.: +34 956016379; fax: +34 956016411.

E-mail address: luisisidoro.romero@uca.es (L.I. Romero García).

M.A. Romero Aguilar, L.A. Fdez-Güelfo, C.J. Álvarez-Gallego, L.I. Romero García ⇑Department of Chemical Engineering and Food Technology, Faculty of Science, University of Cádiz, 11510 Puerto Real, Cádiz, Spain

h i g h l i g h t s

" The maximum hydrogen production is reached at 1.9-days HRT." The maximum solubilized organic carbon concentration is reached at 1.9-day HRT." The ratio ASC/DOC allows further interpretation of process limitations.

a r t i c l e i n f o

Article history:Received 12 September 2012Received in revised form 14 December 2012Accepted 30 December 2012Available online 11 January 2013

Keywords:AcidogenicHydraulic retention timeHydrogenOFMSW

a b s t r a c t

The main objective of this work has been to analyze the effect of hydraulic retention time (HRT) on thehydrogen production from the organic fraction of municipal solid waste (OFMSW) coming from a full-scale mechanical–biological-treatment (MBT) plant. Furthermore, it has also been studied, simulta-neously, the effect of HRT on the solubilization of organic matter.

Experiments were conducted in an anaerobic continuous stirred tank reactor (CSTR) operating at ther-mophilic-dry conditions (55 �C and 20% in total solids concentration respectively). A decreasing sequenceof nine HRTs, from 15 days to 1.5 days, was imposed to evaluate its influence on the hydrogen production(HP), the specific hydrogen production (SHP) and the solubilized organic matter expressed in form of aci-dogenic substrate as carbon (ASC), dissolved organic carbon (DOC) and dissolved acid carbon (DAC).

The results have shown that the best results were obtained at 1.9-day HRT with a feeding regime oftwice a day (type-C). At these conditions, the HP and the SHP was 1.077 L H2/Lreactor day and 24.3 mLH2/g VSadded, respectively. Maximum concentrations obtained of solubilized organic matter were:1201 mg/L for ASC, 1936 mg/L for DOC and 735 mg/L for DAC.

As novelty, the parameter ASC and, specially, the ratio ASC/DOC have shown to be adequate for analyz-ing and interpreting the behavior of the process and, concretely, to determine if hydrolysis and acidogen-esis are coupled (stable process) or decoupled (transitory stage).

� 2013 Elsevier B.V. All rights reserved.

1. Introduction

The current generation of municipal solid waste (MSW) and thehigh consumption rate of fossil fuels are environmental problemstypical of areas with high population density. Nowadays, the

MSW generation represents a serious problem since most of themare deposited in landfills causing negative environmental effects.In Spain, 588 kg per capita are annually collected and this amountis increasing every year [1]. On the other hand, the current reservesof fossil fuels are being depleted very fast due to the growing en-ergy needs [2].

In this sense, many works demonstrated the possibility of cou-pling hydrogen generation with the use of several organic sub-strates including waste materials (such as MSW), industrialwastewaters and agro-industrial wastes. This generation of hydro-gen from wastes may simultaneously offer environmental and eco-nomic benefits in order to meet the growing demand for renewableenergy [3–5].

Among these works, anaerobic digestion (AD) processes areconsidered one of the best options to treat MSW since they may

Effluent

40-L Tedlar Bag

(Biogas)

Influent

7-L Heating bath(55ºC)

10-L Reactor

Stirrer

Fig. 1. Diagram of the laboratory-scale CSTR.

444 M.A. Romero Aguilar et al. / Chemical Engineering Journal 219 (2013) 443–449

solve two problems at once: to reduce the pollution caused by itsdeposition and, simultaneously, it is possible to produce a cleanalternative fuel (hydrogen) through its treatment. It must be notedthat the hydrogen is considered one of the most promising ener-getic vectors of the future due to its lack of environmental impactand its ability to be used in fuel cells. In fact, the role of replacingfossil fuels has been commonly assigned to this gas in order to basethe future on the ‘‘hydrogen economy’’ [6,7].

In particular, it has been reported that the two-stage AD, versusthe traditional single-stage AD process, allows to differentiate thebiofuel production (bio-hydrogen, bio-methane or both (bio-hythane)), improving the overall yield of gas and allow highermethane levels in the biogas generated in the second reactor anddecreasing its purification cost. However, today most of full-scalebiogas plants in Europe are operating on single-stage processand, hence, the two-stage technology remains unusual in this field[8,9]. This fact may be mainly due to that the process stability ofthe dark fermentation (acidogenic stage of the AD) and the maximi-zation of bio-hydrogen production yields are still uncertain.

In recent literature, complex substrates have been often used indark fermentation process for hydrogen production, such asOFMSW, sewage sludge, wheat straws or sugar beets, generallymixed with different inoculums such as sewage sludge, anaerobi-cally digested sludge from wastewater treatment plants (WWTPs)or cow and pig manures. Either, the improvement of the hydrogengeneration implies that the bioactivity of hydrogen-consumersbacteria (homoacetogens and hydrogenotrophic methanogens) ofthese inoculums must be inhibited and, in this case, it may be ex-pected that the anaerobic treatment of biowaste will present asuitable potential to generate this gas.

In order to inhibit the hydrogen consumers, the system shouldbe operated at low hydraulic retention times (HRTs), to promotethe washing-out of methanogens, and a pH around 5.5.

HRT is a critical design parameter since it determines the micro-bial/substrate reaction time and thus the removal efficiency of thesubstrate. Values of HRT between 8 and 12 h are considered suit-able for continuous hydrogen production from solubilized organicmatter [12]. The use of high-solids anaerobic digestion was origi-nally developed to reduce liquid management issues with agricul-tural wastes [13], but also offers opportunity for high volumetricproduction rates [14,15] due to higher substrate concentrations.

Besides, in order to select the acidogenic bacteria from the over-all microbiota and to reach the inhibition of hydrogen-consumersbacteria it is very important to establish pH values near to 5.5which inhibits growth of methanogens but is considered optimumfor acidogenic bacteria [16]. Therefore, pH may be considered an-other basic control parameter of the dark fermentation.

About the hydrogen yield, it generally ranges from 16.26 L H2/kg VSadded for sewage sludge [10] to 257 L H2/kg VSadded of house-hold wastes [11] even if at extreme thermophilic temperature. Inthis context it was demonstrated that, in general, the carbohy-drate-rich feedstock presents higher hydrogen yields. In relationto the best conditions of the AD for enhance the hydrogen produc-tion, it is very important to highlight that in the case of food waste,similar to the OFMSW, the acidogenesis stage in thermophilic re-gime of temperature (55 �C) shows greater volumetric productionrates and efficiencies versus the mesophilic (35 �C) option [12].In addition, the most common type of bioreactor used for hydrogenproduction by dark fermentation is the continuous stirred tankreactor (CSTR) [17].

According to the above statements, the main aim of this studyhas been to enhance the hydrogen production from OFMSW com-ing from a full-scale mechanical–biological-treatment (MBT) plantby means of dark fermentation at thermophilic-dry conditions in alaboratory-scale CSTR. To reach this general objective, the HRT wasstudied and its effect on the solubilized organic matter was also

analyzed. In addition, new indirect parameters (Acidogenic Sub-strate as Carbon, ASC and Dissolved acid Carbon, DAC) recentlypublished has been used for a better understanding of the hydro-lytic and acidogenic phases behavior in the dark fermentation ofOFMSW.

Finally, it must be noted that the AD process for waste stabiliza-tion, as it was mentioned before, is very attractive option in man-agement of organic solid wastes and, hence, a lot of full-scalebiomethanization plants are operating today around the world.However, this research is focused on the thermophilic-dry acido-genic anaerobic digestion (dark fermentation) of the OFMSW withhigh particle size (15 mm) coming from a full-scale MBT plant,which is an innovative aspect of the paper.

2. Materials and methods

2.1. Continuous stirred tank reactor (CSTR)

A 10-L CSTR (5.5 L of working volume) without biomass recy-cling was used (Fig. 1). From a thermostatic water bath (7 L of vol-ume), hot water was pumped through the jacket of the reactor tomaintain the thermophilic conditions (55 �C). The reactor wasequipped with a discharge ball valve and several input/outputports located at the top: a stirring paddle (stirring rate of12 rpm), biogas outlet and feed inlet.

In this type of reactor, each HRT has been calculated by a fixedinflow rate. In this sense, a decreasing sequence of nine HRTs, from15 days to 1.5 days, was imposed to evaluate the influence of thisparameter on the organic matter solubilization rate, the hydrogendaily generation and the specific hydrogen production (SHP). EachHRT was maintained three periods in order to reach stableoperation.

In addition, the feeding regime was modified according to thedecreasing sequence of HRT imposed to the system (Table 1). Inthis sense, the feeding regime was increased as the HRT was de-creased in order to minimize microorganism washing-out at shortHRTs. In the last case, higher feeding frequencies are necessary tominimize that the washing-out induced by semicontinuous regime

Table 1Feeding regimes for each HRT.

Feeding regime Acronym HRT (day)

Once every 2 days: every 2 days, double dailyfeeding is supplied to the reactor

A 15; 10; 6.6

Once a day: daily feeding is supplied to the reactor B 6.6; 4.42.9; 1.9

Twice a day: half of daily feeding is supplied twicea day

C 1.9; 1.5

M.A. Romero Aguilar et al. / Chemical Engineering Journal 219 (2013) 443–449 445

(versus a real continuous regime) was more important that theHRT influence on the process performance. When the system isoperated in semicontinuous operation regime and low HRTs, if itis compared with a continuous regime, the washing-out effect ofthe microorganisms on the process yield can be higher than the ef-fect of the imposed HRT. This is a consequence of operating proto-col at semicontinuous regime with a feed frequency of one doseper day, since the total volume of daily feeding corresponding toa HRT is fed once and, therefore, it may cause the mentioned wash-ing-out.

Finally, it must be noted that this paper aims to determine theeffect of the HRT for enhancing the hydrogen production. Thus, thecriterion to decide which is the optimal HRT for this work has beenthat SHP was maximum.

As known, the organic loading rate (OLR) depends on both HRTand the concentration of organic matter in the feeding. However, itcan be pointed out that, in this work, the characteristics of thefeeding used have remained fairly constant so that the analysiscould be done either on the basis of OLR or HRT.

2.2. Inoculum and feedstock

Methanogenic inoculum was obtained from the effluent of a pi-lot-scale CSTR (300 L of volume) placed on a full-scale MBT plantcalled ‘‘Las Calandrias’’ which is located in Jerez de la Frontera(Cádiz-Spain). The pilot-scale CSTR was operated at themophilic-dry (55 �C and 20% in total solids) conditions and at 15-day HRT.The pH, alkalinity and total volatile fatty acids (VFA) of the inocu-lum were 7.88, 5.51 gCaCO3/L and 5294 mg AcH/L respectively.Although the pilot-scale reactor was methanogenic, the inoculumwas taken during an episode of acidification so that the methano-genic microorganisms were mainly inhibited and, therefore, acido-genic bacteria were selected. For this reason the acetic acidconcentration in the inoculum is abnormally high.

The OFMSW used as feedstock to supply the laboratory-scaleCSTR came from a 15-mm trommel placed on the full-scale MBTplant Las Calandrias. This OFMSW was collected and stored in 25-L drums at �4 �C to avoid its degradation at room temperatureand to ensure that the composition of the OFMSW along the assayswas approximately constant. The characterization of the OFMSW isshown in Table 2.

Before feeding the reactor, the total solids concentration of thefeedstock was adjusted to 20% (typical of the AD at dry condition,

Table 2Characterization of the raw OFMSW.

Parameter Range of values

pH 5.8–6.0Density (g/L) 0.79–0.81Alkalinity (gCaCO3/L) 5.97–6.36Ammonia (gNH3–N/L) 0.016–0.019TS (g/g sample) 0.54–0.58VS (g/g sample) 0.22–0.24Dissolved organic carbon (mg/g sample) 25.78–26.31Total volatile fatty acids (mg AcH/g sample) 9.89–10.25

[18]) by adding of tap water. Next, if the pH of the feeding was low-er than 5, 10 M-NaOH solution was added to adjust the pH value toapproximately 5.5, at which methanogenic archaeas are inhibited.

Finally, it must be noted that the first three HRTs (15–6.6 days)at type-A feeding regime (see Table 3) may be considered as astart-up period necessary for adaptation of the acidogenic microor-ganisms of the inoculum to the typical operational conditions ofthe acidogenic AD (low HRTs and type-B feeding regime) and sub-strate. Therefore, these first three stages, which are not character-istic of the dark fermentation due to the high HRT imposed to thesystem, should not be considered to compare the gas performancein order to determine the optimal HRT in which the maximum SHPis obtained.

As a consequence, along the stages 1–3, was needed the dailyacidification of the reactor to pH 5.5 using 10 M-H3PO4 solutionto inactivate H2-utilizing methanogenic archaeas. After this start-up period, it was observed that the pH was self-regulated by thesystem to 5.5 approximately and there was not methane on thebiogas.

2.3. Analytical techniques

The following analytical determinations were used for wastecharacterization and process monitoring and control: total solids(TS), volatile solids (VS), alkalinity, pH, dissolved organic carbon(DOC) and volatile fatty acids (VFA). The determinations were per-formed according to Standard Methods [19]. TS, VS and pH weredetermined directly from effluent samples. The pH was measureddaily using a Crison� 20 Basic pH-meter, and TS, VS, alkalinity,DOC and VFA were usually analyzed three times a week for HRTsfrom 15 to 4.4 days and daily for HRTs between 4.4 and 1.5 days.The DOC and VFA concentration were analyzed from the filtratesupernatant obtained by means of a lixiviation procedure (10 g ofdigestate in 100 mL of Milli-Q water during 20 min) of the effluentsamples. Samples for the DOC analysis were further filtered thougha 0.7 lm glass fiber filter. On the other hand, the samples for theVFA analysis were further filtered through a 0.22 lm Teflon filterand they were analyzed by gas chromatography as it is describedby Fdez-Güelfo et al. [20].

The biogas generated in the reactor was collected into 40-L Ted-lar� gas bags and its volume was directly measured using a high-precision drum-type gas meter, (Ritter� TG5). Gas volume reportedis standardized at 0 �C and 1 atm.

The gas composition (hydrogen, methane and carbon dioxide)was determined by gas chromatography (Shimadzu� GC-14B) witha stainless steel column packed with Carbosieve� SII and a thermalconductivity detector (TCD). For the system calibration was used acommercial mixture of H2, CH4, CO2, O2 and N2 (Abelló Linde�, S.A.).

2.3.1. Acidogenic substrate as carbon (ASC)ASC is the fraction of solubilized organic matter that has not

been transformed into VFA and therefore, may be used to studythe behavior of the hydrolytic and acidogenesis phases [21]. Thus,ASC may be used in this study to describe how affects HRT on thesephases and, therefore, on the H2 production. As high ASC concen-tration results, higher amounts of solubilized organic matter couldbe transformed to VFA during the acidogenesis step and, therefore,higher H2 production could be expected.

ASC is calculated indirectly from classical parameters (Eqs. (1)and (2)) according to Fdez-Güelfo et al. [21].

ASCðM=L3Þ ¼ DOCðM=L3Þ � DACðM=L3Þ ð1Þ

DACðM=L3Þ ¼Xi¼7

i¼2

AiHðM=L3Þ � ni � 12MWi

" #ð2Þ

Table 3Operational conditions and main measured and calculated average parameters.

Stage HRT(days)

Operationtime (days)

OLR(g VS/Lreactor day)

Feedingregime

HP(L H2/Lreactor day)

SHP(L H2/g VSadded)

DOC(mg/L)

DACa

(mg/L)ASCb

(mg/L)ASC/DOC(no units)

Biogas composition (%)

H2 CO2

1 15 45 6.0 A 0.034 ± 0.002 0.00567 ± 0.0005 652 ± 12 150 503 0.771 36.42 ± 1.12 63.58 ± 1.362 10 30 8.2 0.071 ± 0.003 0.00866 ± 0.0010 905 ± 18 243 663 0.733 34.03 ± 1.25 65.97 ± 0.473 6.6 20 12.1 0.069 ± 0.005 0.00570 ± 0.0011 842 ± 15 235 607 0.721 27.76 ± 2.05 72.24 ± 2.05

4 6.6 20 12.0 B 0.095 ± 0.003 0.00792 ± 0.0009 1587 ± 19 593 994 0.626 35.64 ± 1.02 64.36 ± 1.875 4.4 14 18.5 0.210 ± 0.015 0.01135 ± 0.0014 1706 ± 14 671 1035 0.607 29.82 ± 2.01 70.18 ± 0.966 2.9 10 28.2 0.542 ± 0.019 0.01922 ± 0.0004 1863 ± 21 705 1158 0.622 28.64 ± 0.52 71.36 ± 0.527 1.9 6 45.3 1.046 ± 0.011 0.02309 ± 0.0003 1890 ± 13 721 1169 0.619 28.61 ± 1.85 71.39 ± 1.01

8 1.9 6 44.4 C 1.077 ± 0.015 0.02426 ± 0.0007 1936 ± 22 735 1201 0.620 29.40 ± 0.85 70.60 ± 1.969 1.5 6 57.6 0.874 ± 0.017 0.01517 ± 0.0010 1350 ± 10 545 805 0.596 28.12 ± 1.15 71.88 ± 1.21

a DAC has been calculated from VFA through Eq. (2).b ASC has been calculated through Eq. (1).

446 M.A. Romero Aguilar et al. / Chemical Engineering Journal 219 (2013) 443–449

where DOC, it is the dissolved organic carbon measured by organiccarbon analyzer; AiH, it represents the concentration of each indi-vidual VFA measured by gas chromatography; ni, it is the numberof carbons of each AiH; MWi, it is the molecular weight of each AiH.

According to this, DAC represents an average of carbon consid-ering the ‘‘carbon/molecular weight’’ ratios of each VFA indepen-dently measured by gas chromatography.

3. Results and discussion

In Table 3, the evolutions of the hydrogen production (HP), spe-cific hydrogen production (SHP), dissolved organic carbon (DOC),dissolved acid carbon (DAC), acidogenic substrate as carbon(ASC) and biogas composition are summarized for the differentHRTs used. The data presented are the average values of the lastperiod of each HRT.

3.1. Effect of the HRT on the biogas composition and hydrogenproduction

As can be seen in Table 3, only hydrogen and carbon dioxidewere present in the biogas composition along the experiment;methane was not generated since the methanogenic archaeas(and others hydrogen-consumers microorganisms) were inhibitedas a consequence of the imposed pH control (for 15, 10 and6.6 days) and subsequent low HRTs (under 6.6 days) which areoptimal for the acidogenic bacteria [16].

Fig. 2 shows the composition of biogas throughout the nineexperimental stages. As can be seen, the values of the hydrogenpercentage in biogas were in the range 25–50% whereas the stablevalues corresponding to the different HRTs used ranged from 27 to36% (Table 3), which is analogous to the percentages reported by

0 15 30 45 60 75 90 105 120 135 15020

30

40

50

60

70

80 1.5C1.9C

1.9B

2.9B4.4B6.6B6.6A10A

H2

CO2

Bio

gas

com

posi

tion

(%

)

Time (days)

15A

Fig. 2. Evolution of biogas composition.

others authors [12,22] in their experiments about hydrogen pro-duction by dark fermentation from food waste or municipal solidwastes and slaughterhouse.

On the other hand, as it is shown in Table 3, as the HRT is de-creased from 15 days to 1.9 days, the HP is gradually increasedfrom 0.034 L H2/Lreactor day (stage 1) to the maximum value of1.077 L H2/Lreactor day (stage 8). This inverse relationship betweenthe HRT and the daily hydrogen production in acidogenic anaero-bic digester has been corroborated by a lot of authors [23–26]. Thisfact may be due to the H2-producing bacteria (Clostridium sp.) areselected among the different populations of microorganisms in-volved on the anaerobic digestion at low HRTs [27]. In addition,it must be highlighted that the maximum HP obtained in this work(1.077 L H2/Lreactor day at 1.9-day HRT) is higher to the maximumHP reported by Kim et al. [28] for similar OLR (44 g VS/L day) intheir studies of optimization of hydrogen generation from foodwaste as a function of the HRT. These authors obtained a maximumHP of 0.79 L H2/Lreactor day at 1-day HRT.

On the contrary, when the HRT is decreased from 1.9 days to1.5 days between the stages 8 and 9, the HP decreases from1.077 L H2/Lreactor day to 0.874 L H2/Lreactor day, i.e. it decreasesabout 19% versus the maximum value. This fact indicates that adestabilization episode between the hydrolytic and acidogenicpopulations of microorganisms could be occurring on the reactor.This hypothesis is discussed later in the text through the ASCparameter (Section 3.2).

About the SHP, its evolution is analogous to the HP. As the HRTis decreased, the SHP is gradually increased from 5.7 mL H2/gVSadded (stage 1) to the maximum of 25.4 mL H2/g VSadded (stage8). Similarly, in the stage 9 the SHP decreases until 15.2 mL H2/gVSadded, i.e. its value decreases 40.2% versus the maximum valueof the previous stage.

According to Hawkes et al. [24], the theoretical maximum SHPreported from the acidogenesis of glucose is obtained between 8-h HRT and 12-h HRT. Therefore, the decreasing of the SHP betweenthe stages 8 and 9, in which the HRT is decreased from 1.9 days to1.5 days, may be indicative of unbalance between the hydrolyticand acidogenic phases by working with particulate wastes as ithas been stated above.

Data from literature about the specific hydrogen productivityfor food waste are higher than the obtained in this study forOFMSW. Thus, Shin et al. [12] have obtained SHP of 28.4–46.3 mL H2/g VSadded in batch tests performed for 2 g VS/L foodwaste at thermophilic temperature (55 �C) while the SHP was onlyof 1.3–5 mL H2/g VSadded operating at mesophilic temperature(35 �C). However, the authors found that SHP was depending onthe VS content of the test, reaching the maximum of 91.5 mL H2/g VSadded for 6 g VS/L at 55 �C. Moreover, a SHP of 43 mL H2/gVSadded was obtained by Liu et al. [8] in the acidogenesis of house-

0 15 30 45 60 75 90 105 120 135 150

250

200

150

100

1.5C1.9C

1.9B

2.9B4.4B6.6B6.6A10A

L H

2

Time (days)

15A

50

Fig. 3. Accumulated hydrogen production.

0 15 30 45 60 75 90 105 120 135 1500

500

1000

1500

2000

2500DOC ASC DAC 1.

5C1.9C

1.9B

2.9B4.4B6.6B6.6A10A

Con

cent

rati

on (

mg/

L)

Time (days)

15A

Fig. 4. Evolution of parameters related to organic matter solubilization.

0.40.50.60.70.80.91.0 1.

5C1.9C

1.9B

2.9B4.4B6.6B6.6A10A

tio

ASC

/DO

C

15A

M.A. Romero Aguilar et al. / Chemical Engineering Journal 219 (2013) 443–449 447

hold solid waste operating in a CSTR at mesophilic temperature(37 �C) and 2-days HRT.

Finally it should be noted that the evolution of the SHP has beentaken into account to determine when the feeding regime shouldbe changed. As it can be seen in Table 3, when the HRT is decreasedfrom 15 to 10 days (OLR raising of 36.7%) between stages 1 and 2,the SHP is notably increased from 5.7 to 8.7 mL H2/g VSadded

(increasing of 52.6%). This improvement is related to the OLR in-crease. However, when the HRT is decreased from 10 to 6.6 days(OLR raising of 47.6%) between stages 2 and 3, the SHP is decreasedfrom 8.7 to 5.7 mL H2/g VSadded (decreasing of 34.5%). This effect onthe hydrogen yield when the HRT is reduced may occurs becausemicrobial washing-out is taking place due to operational procedureused for semicontinuous feeding (in type-A feeding, one dose each2 days was used and hence, volume withdrawn instantaneouslyfrom the reactor in the feeding process is twice that required forthe imposed HRT) and, therefore, the feeding regime was increasedfrom type A to B between the stages 3 and 4.

As can be seen in Table 3, when type-B feeding regime was im-posed at 6.6-day HRT (stage 4), the SHP was increased from 5.7 to7.9 mL H2/g VSadded (increasing of 38.6%) without a significant OLRchange (�0.8%). Subsequent changes between stages 4 and 7 arerelated to OLR increases (54.2%, 52.4% and 60.6% respectively)but, for the last change (between stage 6 and 7), SHP increase islower than expected (20.3%). This slowing down in SHP increasewith the OLR increase may be related with the washing-out effectabove commented.

However, when the feeding frequency is increased betweenstages 7 and 8 (1.9-day HRT) without significant OLR change(2%), the SHP is increased slightly to 24.3 mL H2/g VSadded (increas-ing of 5.2%). In this case, the increase is less pronounced and it isnot related to OLR or feeding type change. As can be seen at thenext subsection, process limitation or inhibition could be thereason.

The above mentioned behavior of the process can be observedin Fig. 3, showing the accumulated hydrogen production duringall the stages. As can be seen, the slope of the curve increasesslightly for the HRTs between 15 and 6.6 days despite the increasein the OLR. However, from 4.4 to 1.9 days HRT a very higher slopewas obtained as a consequence of the OLR increasing. Finally, for1.5-day HRT the slope of the curve diminishes. This behavior indi-cates that the hydrogen yield of the process is optimum for therange from 4.4 to 1.9-day HRT.

0 15 30 45 60 75 90 105 120 135 1500.00.10.20.3R

a

Time (days)

Fig. 5. Evolution of ASC/DOC ratio.

3.2. Effect of the HRT on the solubilized organic matter

As it has been stated above, the ASC is an indirect parameterclosely related with the hydrolytic and acidogenic phases and thehydrogen production. Its value is the difference between two clas-

sical parameters: DOC, whose value is representative of theamount of solubilized carbon during the hydrolysis (which hasnot yet been converted to VFA by the acidogenic bacteria) andthe DAC, whose value is representative of the amount of carbonin acid form coming from the VFA generated during the acidogen-esis [21]. Thus, its value represents the amount of solubilized car-bon that could be transformed to VFA and, hence, in hydrogen. ASCcould be considered as a reservoir of VFA and hydrogen.

Fig. 4 shows the evolution of the above mentioned parameters(DOC, DAC and ASC) versus time for all the stages. After of eachchange of HRT a transitory evolution of the parameters can be ob-served. However, the system evolves, in all the cases, towards thestable value corresponding to pseudo-stationary state, in except of1.5-day HRT.

Reactors in stationary state show that hydrolysis and acidogen-esis phases are well coupled. Therefore, all solubilized substrate(measured as DOC) is converted into VFA (except a fraction resis-tant to acidogenesis) and, hence, the ASC/DOC ratio should beapproximately constant. However, when any operational conditionis changed, during a transitory period both phases could be decou-pled and ASC/DOC ratio could show variations in its value.

If both phases are decoupled due to acidogenic stage limitationsor inhibitions, the ASC/DOC ratio should increase significantly (or-ganic matter susceptible of acidogenesis is solubilized but it is notconverted to VFA). On the contrary, if the limiting step is thehydrolysis, as usual in solid wastes, the ASC/DOC value could main-tain or even go down if hydrolysis is extremely slow due to masstransfer limitations and, therefore, ASC is rapidly transformed toVFA by the acetogenic microorganisms.

In Fig. 5 the values of the ratio ASC/DOC have been plotted ver-sus time. The evolution of the parameter shows clearly the decou-pling of the hydrolysis and acidogenesis stages at the beginning ofeach HRT and the subsequent stabilization of the parameter. Be-sides, it can be observed from 6.6B forward that the ratio ASC/

0.000

0.005

0.010

0.015

0.020

0.025

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

1.5C1.9B 1.9C2.9B4.4B6.6B6.6A10A

SHP

(L

H2/g

VS ad

ded )

Rat

io A

SC/D

OC

HRT (days)

Ratio ASC/DOC

15A

SHP

Fig. 6. Comparative between ASC/DOC ratio and SHP.

448 M.A. Romero Aguilar et al. / Chemical Engineering Journal 219 (2013) 443–449

DOC became much more stable, indicating the effective coupling ofthe hydrolysis and acidogenesis.

Fig. 6 shows the average ASC/DOC ratio and SHP obtained asstabilized values for each HRT. Taking into account the previousstatements, an increment in the hydrogen production should haveassociated a parallel increment on the VFA concentration and,therefore, in the ASC/DOC decrease. As can be seen in Fig. 6, aboutthe three initial stages ASC/DOC is getting down, probably reachingits lower stable limit. The semicontinuous type-A feeding, withmassive lack of microbial populations during every feeding, mustto induce an abnormal behavior probably related to waste coloni-zation limitations.

Also, it may be seen how in the type-B periods (stages 4–7), theASC/DOC values are approximately constant (0.618 ± 0.008) andthe hydrogen production is clearly increased in one order of mag-nitude as the HRT decreases. This fact suggests that the hydrolysisand acidogenesis stages are well coupled and hydrolysis is the lim-iting step indeed. The transition to type-C feeding has not influenceon ASC/DOC ratio since the OLR is similar and high.

Finally, between the stages 8 and 9, there is not a categoricaldecreasing in ASC/DOC ratio but all the parameters related withthe solubilized organic matter concentration (DOC, DAC and ASC)and the hydrogen production show a substantial decrease versusthe previous stage. Concretely, the SHP decreases 37.4% and theASC, DAC and DOC drop 33.0%, 25.9% and 30.3% respectively (seeTable 3). Therefore, according to the previous statements, at 1.5-day HRT (stage 9) the process could be failing due to limitations re-lated to colonization of substrate particles.

4. Conclusions

Based on the results and discussion, the main conclusions de-rived from this research are as follows:

1. Acidogenic anaerobic digestion of a real OFMSW (average parti-cle size of 15 mm) coming from a full-scale MBT plant at ther-mophilic-dry conditions in a laboratory-scale CSTR is feasibleand balanced at least until 1.9-day HRT and 44.4 g VS/Lreactor -day with a semicontinuous feeding regime of twice a day(type-C). The higher hydrogen production has been found atthese conditions. According to the specific criterion selectedby authors, 1.9-days HRT may be considered as the optimumfor this work. Further improvement would be expected to per-form up-scale process without semicontinuous feeding regimelimitations of lab-scale.

2. Under HRTs of 6.6 and 1.9 days, the feeding regime was deter-minant for maintain enough microbial population into reactorand to avoid washing-out phenomena with decreasing of SHP.

3. HRT lower than 1.9 days at type-C feeding regime (1.5-day HRTin this research) induces a failure. It is not possible to identifyunequivocally the reason (hydrolysis failure, imbalance of theprocess, organic loading rate shock, limitations to particles col-onization of the substrates, etc.). The authors suggest limita-tions related to particle colonization.

4. A novel interpretation of the process based on the analysis ofthe ASC/DOC ratio at different stages has provided a betterunderstanding of the hydrolytic and acidogenic phases behaviorin the dark fermentation of OFMSW in stationary and transitorystates.

Acknowledgments

This work was supported by the ‘‘Ministerio de Ciencia e Innova-ción’’ of Spain (Projects CTM2007-62164 – associated grant refer-ence BES-2008-004390 – and CTM2010-17654), the ‘‘Consejeríade Innovación, Ciencia y Empresa’’ of the ‘‘Junta de Andalucía, Spain’’(Project P07-TEP-02472) and the European Regional DevelopmentFund (ERDF).

References

[1] J. Llopis, Types of systems for collecting municipal solid waste in Europeaccording to levels of development, final master thesis in sustainability, June2011.

[2] A. Bedoya, J.C. Castrillón, J.E. Ramírez, J.E. Vásquez, M. Zabala, Biologicalproduction of hydrogen: a literature survey, Dyna 154 (2007) 137–157.

[3] X.M. Guo, E. Trably, E. Latrille, H. Carrere, J.P. Steyer, Hydrogen production fromagricultural waste by dark fermentation: a review, Int. J. Hydrogen Energy 35(2010) 10660–10673.

[4] J.J. Lay, Y.J. Lee, T. Noike, Feasibility of biological hydrogen production fromorganic fraction of municipal solid waste, Water Resour. 33 (1999) 2579–2586.

[5] S.E. Oh, S. Van Ginkel, B.E. Logan, The relative effectiveness of pH control andheat treatment for enhancing biohydrogen gas production, Environ. Sci.Technol. 37 (2003) 5186–5190.

[6] D. Das, T. Veziroglu, Hydrogen production by biological processes: a survey ofliterature, Int. J. Hydrogen Energy 26 (2001) 13–28.

[7] Hydrogen is on the way, sustainable energy, magazine on European. Researchand Innovation, 42 (2004).

[8] D. Liu, R.J. Zeng, I. Angelidaki, Hydrogen and methane production fromhousehold solid waste in the two-stage fermentation process, Water Resour.40 (2006) 2230–2236.

[9] F. Fantozzi, C. Buratti, Biogas production from different substrates in anexperimental continuously stirred tank reactor anaerobic digester, Bioresour.Technol. 100 (2009) 5783–5789.

[10] B.Y. Xiao, J.X. Liu, Biological hydrogen production from sterilized sewagesludge by anaerobic self-fermentation, J. Hazard. Mater. 168 (2009) 163–167.

[11] D. Liu, B. Min, I. Angelidaki, Biohydrogen production from household solidwaste (HSW) at extreme-thermophilic temperature (70 �C) e influence of pHand acetate concentration, Int. J. Hydrogen Energy 33 (2008) 6985–6992.

[12] H.-S. Shin, J.-H. Youn, S.-H. Kim, Hydrogen production from food waste inanerobic mesophilic and thermophilic acidogenesis, Int. J. Hydrogen Energy 29(2004) 1355–1363.

[13] W.J. Jewell, S. Dell’Orto, K. Fanfoni, S. Fast, E. Gottung, D. Jackson, R. Kabrick,Dry Fermentation of Agricultural Residue. Cornell University Final Report forU.S. Department of Energy-Solar Energy Research Institute, SERI/TR-09038-7,Golden, Colorado, 1982.

[14] B.K. Richards, R.J. Cummings, W.J. Jewell, F.G. Herndon, High solids anaerobicmethane fermentation of sorghum and cellulose, Biomass Bioenergy 1 (1991)47–53.

[15] W.J. Jewell, R.J. Cummings, B.K. Richards, Methane fermentation of energycrops: maximum conversion kinetics and in situ biogas purification, BiomassBioenergy 5 (1993) 261–278.

[16] R.J. Zoetemeyer, J.C. Heuvel, A. Cohen, pH influence on acidogenic dissimilationof glucose in an anaerobic digestor, Water Res. 16 (1982) 303–311.

[17] H. Yang, P. Shao, T. Lu, J. Shen, D. Wang, X. Yuan, Continuous bio-hydrogenproduction from citric acid wastewater via facultative anaerobic bacteria, Int. J.Hydrogen Energy 31 (2006) 1306–1313.

[18] J. Fernández, M. Pérez, L.I. Romero, Effect of substrate concentration on drymesophilic anaerobic digestion of organic fraction of municipal solid waste(OFMSW), Bioresour. Technol. 99 (2008) 6075–6080.

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

[20] L.A. Fdez-Güelfo, C. Álvarez-Gallego, D. Sales Márquez, L.I. Romero García, Theeffect of different pretreatments on biomethanation kinetics of industrialorganic fraction of municipal solid waste (OFMSW), Chem. Eng. J. 171 (2011)411–417.

M.A. Romero Aguilar et al. / Chemical Engineering Journal 219 (2013) 443–449 449

[21] L.A. Fdez-Güelfo, C. Álvarez-Gallego, D. Sales Márquez, L.I. Romero García, Newindirect parameters for interpreting a destabilization episode in an anaerobicreactor, Chem. Eng. J. 180 (2012) 32–38.

[22] X. Gómez, A. Morán, M.J. Cuetos, M.E. Sánchez, The production of hydrogen bydark fermentation of municipal solid wastes and slaughterhouse waste: a two-phase process, J. Power Sources 157 (2006) 727–732.

[23] H.-S. Shin, J.-H. Youn, Conversion of food waste into hydrogen bythermophilicacidogenesis, Biodegradation 40 (2006) 2230–2236.

[24] F.R. Hawkes, R. Dinsdale, D.L. Hawkes, I. Hussy, Sustainable fermentativehydrogen production: challenges for process optimization, Int. J. HydrogenEnergy 27 (2002) 1339–1347.

[25] K.-W. Jung, D.-H. Kim, S.-H. Kim, S.-H. Shin, Bioreactor design for continuousdark fermentative hydrogen production, Bioresour. Technol. 102 (2011) 8612–8620.

[26] Y. Ueno, M. Tatara, H. Fukui, T. Makiuchi, M. Goto, K. Sode, Production ofhydrogen and methane from organic solid wastes by phase-separation ofanaerobic process, Bioresour. Technol. 98 (2007) 1861–1865.

[27] D. Jones, D. Woods, Acetone–butanol fermentation revisited, Microbiol. Rev. 50(1986) 484–524.

[28] S.-H. Kim, S.-K. Han, H.-S. Shin, Optimization of continuous hydrogenfermentation of food waste as a function of solids retention timeindependent of hydraulic retention time, Process Biochem. 43 (2008) 213–218.