w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 2 5 3 1 – 2 5 3 9

Avai lab le a t www.sc iencedi rec t .com

journa l homepage : www.e lsev ie r . com/ loca te /wat res

Application of high rate, high temperature anaerobicdigestion to fungal thermozyme hydrolysates fromcarbohydrate wastes

C. Forbesa,*,1, C. O’Reillya, L. McLaughlinb, G. Gilleranb, M. Tuohyb, E. Collerana

aEnvironmental Microbiology Research Unit, Department of Microbiology, National University of Ireland, Galway, IrelandbMolecular Glycobiology Group, Department of Biochemistry, National University of Ireland, Galway, Ireland

a r t i c l e i n f o

Article history:

Received 11 November 2008

Received in revised form

18 January 2009

Accepted 1 March 2009

Published online 19 March 2009

Keywords:

Municipal solid waste

UAHR

Thermophilic anaerobic digestion

Fungal thermozymes

Biogas

* Corresponding author: Tel.: þ353 214 90197E-mail address: c.forbes@ucc.ie (C. Forbe

1 Present address: Environmental Research0043-1354/$ – see front matter ª 2009 Elsevidoi:10.1016/j.watres.2009.03.014

a b s t r a c t

The objective of this study was to examine the feasibility of using a two-step, fully bio-

logical and sustainable strategy for the treatment of carbohydrate rich wastes. The primary

step in this strategy involves the application of thermostable enzymes produced by the

thermophilic, aerobic fungus, Talaromyces emersonii, to carbohydrate wastes producing

a liquid hydrolysate discharged at elevated temperatures. To assess the potential of ther-

mophilic treatment of this hydrolysate, a comparative study of thermophilic and meso-

philic digestion of four sugar rich thermozyme hydrolysate waste streams was conducted

by operating two high rate upflow anaerobic hybrid reactors (UAHR) at 37 �C (R1) and 55 �C

(R2). The operational performance of both reactors was monitored from start-up by

assessing COD removal efficiencies, volatile fatty acid (VFA) discharge and % methane of

the biogas produced. Rapid start-up of both R1 and R2 was achieved on an influent

composed of the typical sugar components of the organic fraction of municipal solid waste

(OFMSW). Both reactors were subsequently challenged in terms of volumetric loading rate

(VLR) and it was found that a VLR of 9 gCOD l�1 d�1 at a hydraulic retention time (HRT) of 1

day severely affected the thermophilic reactor with instability characterised by a build up

of volatile fatty acid (VFA) intermediates in the effluent. The influent to both reactors was

changed to a simple glucose and sucrose-based influent supplied at a VLR of 4.5 gCOD l�1 d�1 and HRT of 2 days prior to the introduction of thermozyme hydrolysates. Four unique

thermozyme hydrolysates were subsequently supplied to the reactors, each for a period of

10 HRTs. The applied hydrolysates were derived from apple pulp, bread, carob powder and

cardboard, all of which were successfully and comparably converted by both reactors. The

% total carbohydrate removal by both reactors was monitored during the application of

the sugar rich thermozyme hydrolysates. This approach offers a sustainable technology for

the treatment of carbohydrate rich wastes and highlights the potential of these wastes as

substrates for the generation of second-generation biofuels.

ª 2009 Elsevier Ltd. All rights reserved.

5; fax: þ353 214 901932.s).

Institute, University College Cork, Lee Road, Ireland.er Ltd. All rights reserved.

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 2 5 3 1 – 2 5 3 92532

1. Introduction

Carbohydrate based wastes constitute a significant proportion

of the organic fraction of municipal solid waste (OFMSW), as

an abundant by-product of the agri-feed/market, gardening,

food-processing and service industries. Stringent legislation

at a European level (European Council Directive 1999/31/EC)

stipulates that EU member states must reduce the quantities

of biodegradable municipal waste consigned to landfill. In

order to comply with imposed targets, member states have

been obliged to set up national strategies in order to imple-

ment the reduction of biodegradable waste landfilled. In

recent years, a number of novel approaches to the recycling of

carbohydrate wastes have been proposed including ethanol

and biodiesel production as well as the production of biode-

gradable plastic from food wastes and carbon dioxide (Call-

aghan et al., 1999). Arguably, the most valuable approach, both

environmentally and economically, is the anaerobic treat-

ment of the OFMSW which simultaneously reduces the

pollution potential of the waste while producing renewable

energy in the form of a biofuel, methane gas.

The focus of research and application of anaerobic digestion

technology to date has centred on sewage sludge, animal

manures and industrial wastewaters from the food-processing,

fermentation and pharmaceutical/fine-chemical industries.

Recent years, however, have seen increasing research into the

development of appropriate digester designs for bio-

methanation of primary biomass sources, such as waste food

materials, food-processing residues, yard-wastes and local

authority green wastes for the production of second generation

biogas (Mata-Alvarez et al., 2000; Ahn etal., 2001; Kimetal., 2002;

Nishio and Nakashimada, 2007; Demirel and Scherer, 2008;

Forster-Carneiro et al., 2008; Stabnikova et al., 2008). A variety of

anaerobic digestion designs for the treatment of food wastes

have been investigated within a variety of temperature ranges

(Bernal et al., 1992; Rintala and Lepisto, 1997; Solera et al., 2002;

Poirrier and Lema, 2004). The rate-limiting step in the anaerobic

treatment of carbohydrate wastes has been identified as the

hydrolysis of high molecular weight polysaccharides (Noike

et al., 1985), resulting in solid retention times in the region of 10–

30 days in single stage Continuously Stirred Tank Reactors

(CSTRs) (Gunaseelan, 1997; Callaghan et al., 1999). In light of this,

pre-hydrolysis of the carbohydrate rich fraction of OFMSW prior

to anaerobic digestion is an attractive option by which these

long chain polymers can be converted to readily degradable

subunits prior to anaerobic treatment.

By applying fungal enzymes to carbohydrate rich wastes as

a pre-treatment strategy, the complex enzymatic systems that

some aerobic fungi possess, including those which mineralise

recalcitrant polymers such as lignin, can be fully exploited. This

permits more complete anaerobic digestion of the resultant

hydrolysate, with concomitant higher methane yields. This

approach has been investigated previously by other research

groups using substrates such as orange processing wastes (Sri-

latha et al., 1995), agro-wastes (Okeke and Obi, 1995) and in

combination with steam pre-treatment to improve the anaer-

obic biodegradability of beech wood-meal (Sawada et al., 1995).

The pre-treatment strategy utilised in the current study

involved treatment of carbohydrate rich wastes by

thermostableenzymesproducedusingsolid-state fermentation

technology (SSF) at elevated temperatures with the non-sporu-

lating, GRAS (generally regarded as safe) fungus, Talaromyces

emersonii. SSFona variety of carbohydrate richwastes, including

food wastes, has previously been used routinely for medium to

large-scale enzyme production (Considine et al., 1988; Tuohy

and Coughlan, 1992; Tuohy et al., 1994). The application of

thermozymes to carbohydrate treatment has the potential to

maximise the extent and rate of hydrolysis of the poly-

saccharides of selected food wastes and of the organic fraction

of municipal solid waste (OFMSW) prior to anaerobic digestion.

Due to the high temperatures involved, the hydrolysate can be

applied at highloading ratesand short hydraulic retentiontimes

(HRT) to high-rate thermophilic anaerobic digesters, such as the

UASB and UAHR, resulting in rapid methanogenesis of the

simple sugars contained in the hydrolysate. The study pre-

sented in this paper offers a comparison between thermophilic

AD and the more commonly utilised mesophilic process for the

treatment of thermozyme hydrolysates. The aim was to assess

which process proved most efficient and robust in the treatment

of carbohydrate rich hydrolysates.

Four unique carbohydrate rich wastes were selected for

investigation using this two-step strategy. These wastes, apple

pomace, bread, cardboard and carob powder, were used as

substrate for the degradative activity of thermozymes isolated

from T. emersonii prior to anaerobic treatment. Apple pomace was

chosen as a representative fruit and vegetable waste as it is

composed of the common polymers associated with this type of

waste; cellulose, hemicelluloses and pectin, as well as soluble

sugars (Grohmann and Bothast, 1994). The second waste stream

for investigation was the hydrolysate product of bread. Bread is

a refined product of cereals, rich in carbohydrates, and thus

a suitable substrate for investigation as it contains large amounts

of carbohydrate. Bread is a waste worthy of investigation, if only

for its ubiquity. The readily anaerobic biodegradable nature of

wholemeal bread has previously been demonstrated in meso-

philic batch tests in a study conducted by Veeken and Hamelers

(1999). Cardboard was selected for investigation as a representa-

tive of the ligno-cellulosic component of the organic fraction of

MSW. The anaerobic digestion of ligno-cellulosic wastes has been

extensively investigated previously, with mainly newsprint used

assubstrateforawidevarietyofanaerobictreatmentapproaches.

Various designs and pre-treatment strategies have been investi-

gated, although these strategies are not utilised widely in the full-

scale treatment of these wastes (Thompson et al., 2001; Ahn and

Forster, 2002). The final waste investigated was carob powder.

This plant derivative is a chocolate and sugar replacement which

is becoming more widespread in breads and pastry products

due to increasing levels of diabetes, obesity and allergies. Carob

pods are comprised of about 76% carbohydrate. Carob powder

was used as a substrate in this trial due to the variety of sugars of

which it is composed (Table 3).

2. Materials and methods

2.1. Source of biomass

A mesophilic granular sludge was obtained from a full sca-

le, anaerobic upflow internal circulation (IC) reactor treating

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 2 5 3 1 – 2 5 3 9 2533

dairy wastewater at Carbery Milk Products, Ballineen, CO,

Cork.

2.2. Fungal pre-treatment

Pre-treatment of the wastes was carried out by the Molecular

Glycobiology group, N.U.I., Galway, as previously described

(Tuohy et al., 1990).

2.3. Reactor design and operation

Two laboratory scale UAHRs, both with an active volume of 3.9 l

were inoculated with mesophilic granular sludge to give an

initial volatile suspended solids (VSS) concentration of 11.5 g l�1

reactor liquid volume. Mesophilic conditions were maintained

in reactor 1 (R1) operated at 37 �C while reactor 2 (R2) was

operated at 55 �C from inoculation onwards, to encourage

adaptation of the biomass to thermophilic conditions (Dinsdale

et al., 1997). The remaining reactor active volume was filled

with a mixture of sucrose and acetate at a 1:1 ratio on a COD

basis, to a total of 1 gCOD l�1 until initiation of biogas production

at a concentration greater than 40% CH4 was achieved. Once this

methaneconcentrationwasrecorded, thetrial commenced(day

1) and the reactors were supplied with a synthetic influent

comprised of a mixture of equal concentrations (on a COD basis)

of nine sugars found to be most prevalent in canteen waste

hydrolysate (Hutnan et al., 2000). These were L-Arabinose,

fucose, galactose, galacturonic acid, glucose, mannose, rham-

nose, sucrose and D-xylose to a final concentration of 9 g COD�1.

The COD:N:P ratio was maintained at 1000:5:0.5 by supple-

mentation with NH4Cl and KH2PO4 to the required concentra-

tions. The pH of the influent mixture was buffered by addition of

NaHCO3 at a concentration of 12 g l�1 and supplemented with

micronutrients (1 ml l�1), as recommended by Shelton and

Tiedje (1984). Influent was initially supplied at a volumetric

loading rate (VLR) of 1.125 gCOD l�1 d�1 and an HRT of 8 days.

The VLR was increased stepwise from 1.125 gCOD l�1 d�1 to

2.5 gCOD l�1 d�1 and to 4.5 gCOD l�1 d�1 by reducing the HRT to 4

days and 2 days on days 10 and 23, respectively. On day 128, the

HRT was further reduced to 24 h with a resultant VLR of

9 gCOD l�1 d�1. This destabilised both the reactors and so the

HRT of both was raised to 2 days on day 171 to encourage

recovery of reactor performances. From day 171 until the end of

the trial the reactors were operated at a 2-day HRT, with a VLR of

4.5 gCOD l�1 d�1. On day 171, the influent to both reactors was

changed tosucroseand glucose, of an equal ratioona COD basis,

to a VLR of 4.5 gCOD l�1 d�1. This change was effected for

economic reasons as the influent used up until this point in the

start-up of the reactors incurred unsustainably high operational

costs. Application of fungal hydrolysates commenced on day

486 with the feeding of apple pulp hydrolysate. Each of the four

unique hydrolysates supplied as influent to the reactors was

supplied at a concentration of 9 gCOD l�1 so that the reactor

biomass could be challenged in terms of hydrolysate composi-

tion rather than loading rate. Each feeding period lasted for at

least 10 HRTs so that acclimatisation to new influent streams

could be monitored. The treatment trial was divided into 10

operational periods characterised by changes in either the VLR

or influent composition. The operational parameters for each

period of the trial as well as average results obtained are sum-

marised in Table 1 (R1) and Table 2 (R2).

2.4. Analytical techniques

Samples of reactor influent/effluent were routinely analysed

for COD (g l�1), VFAs (mg l�1) and pH (APHA, 2001). The total

carbohydrate content of samples was quantitatively deter-

mined according to the phenol–sulphuric acid method of

Dubois et al. (1956). Biogas was sampled and analysed for the

percentage methane content of the biogas produced by R1 and

R2 using gas chromatography. The concentration of key

carbohydrates in the thermozyme hydrolysates were analysed

by running cellobiose, glucose, galactose, D-xylose, mannose, 5-

hydroxymethylfurfural (HMF) and furfural as standards in

HPLC analysis (Table 3). HPLC analysis was carried out by the

Molecular Glycobiology group, N.U.I., Galway.

3. Results

3.1. Reactor start-up

A rapid start-up was achieved for the reactor operated at 37 �C

(R1), with COD removal efficiency of 96% recorded after 9 days of

operation (Fig. 1). Throughout the start-up period, R1 effluent

VFA concentrations were low or below the range of detection.

This was observed at all VLRs applied throughout periods 1–3,

with the exception of slight fluctuations. Start-up of R2 at 55 �C

was also highly successful, although improvements in opera-

tional performance were more gradual than those observed for

the mesophilic reactor (Fig. 2). The more lengthy start-up time

was reflected in both the COD removal efficiency of R2 and the

VFA content of the effluent produced (Figs. 1 and 2). The

methane content of the biogas produced by the mesophilic

reactor was, in general, higher than that from the thermophilic

reactor during the start-up period, although fluctuations in the

concentration were noted for both reactors (Fig. 3; Tables 1 and

2). Following initial start-up, the stable operation of both reac-

tors was compromised by two changes in reactor operation.

Firstly, the upflow velocity of the influent waste stream was

increased from 2 m h�1 to 5 m h�1 on day 90 in an attempt to

promote good mixing of the sludge bed. In consequence,

washout of biomass from both the mesophilic and thermophilic

reactors occurred on days 96 and 106, respectively. This proved

immediately detrimental to the performance of both reactors,

with decreased COD removal efficiencies and increased VFA

discharge noted until day 113. Period 5 of the trial period was

characterised by an increase in VLR to 9 gCOD l�1 d�1 and

reduction of HRT to 1 day, and this had a more severe impact on

the performance of both reactors (Figs. 1–3). In order to prevent

failure of the thermophilic reactor on day 171, the influent to

both reactors was changed to an equal mixture, on a COD basis,

of glucose and sucrose to 9 g l�1 and the HRT was increased to 2

days. The COD removal efficiency of both reactors improved

immediately, with COD removal efficiencies increasing to>97%

from day 201 onwards (Fig. 1). Effluent VFA concentrations from

R1 decreased rapidly to below the limits of detection by day 218

(Fig. 2A) while the decrease in R2 effluent VFA concentrations

was more gradual (Fig. 2B) until day 248. Although the %

Table 1 – Operation parameters and average performance data for R1 (37 8C) UAHR.

Period 1 2 3 4 5 6 7 8 9 10

Days 0–10 10–23 23–128 128–171 171–486 486–513 513–533 533–620 620–646 646–670

Influent type Synthetic influentof 9 sugars

G/Sf Apple Bread G/S Carob C/boardg

Influent parameters

CODa (mg l�1) 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000

HRTb (days) 8 4 2 1 2 2 2 2 2 2

VLRc gCOD l�1 d�1 1.125 2.25 4.5 9 4.5 4.5 4.5 4.5 4.5 4.5

Effluent parameters

(period mean)

% COD removal 81.4 97.7 96.5 78.2 96.4 88.0 88.4 95.72 92.0 87.9

% Carbd removal NR NR NR NR NR 99.0 99.5 92.7 92.0

pH 7.67 7.70 7.77 7.62 7.86 7.66 7.75 7.59 7.85 7.76

CH4 (%) 61.0 59.2 73.0 38.7 47 50.1 59.6 43.2 48.2 52.0

VFAe (mg l�1)

Acetate 179.9 10.4 89.4 271.6 6.6 16.6 6.6 150.1 154.1 204.3

Propionate 92.2 5.0 31.3 81.4 1.6 2.2 0.3 140.2 41.1 37.2

Butyrate 0 0 0 0 0.3 0.9 0 0 0 0

NR, not recorded.

a Chemical oxygen demand.

b Hydraulic retention time.

c Volumetric loading rate.

d Carbohydrate.

e Volatile fatty acid.

f Glucose/sucrose.

g Cardboard.

Table 2 – Operation parameters and average performance data for R2 (55 8C) UAHR.

Period 1 2 3 4 5 6 7 8 9 10

Days 0–10 10–23 23–128 128–171 171–486 486–513 513–533 533–620 620–-646 646–670

Influent type Synthetic influentof 9 sugars

G/Sf Apple Bread G/S Carob C/boardg

Influent parameters

CODa (mg l�1) 9000 9000 9000 9000 9000 9000 9000 9000 9000 9000

HRTb (days) 8 4 2 1 2 2 2 2 2 2

VLRc gCOD l�1 d�1 1.125 2.25 4.5 9 4.5 4.5 4.5 4.5 4.5 4.5

Effluent parameters

(period mean)

% COD removal 54.7 73.5 88.5 66.0 96.6 89.7 88.2 94.7 82.4 88.2

% Carbd removal NR NR NR NR NR 98.0 98.9 89.6 95.1

pH 7.59 7.73 7.89 7.89 8.00 7.71 7.82 7.50 7.78 7.99

CH4 (%) 45.0 51.0 46.9 35.2 48.3 47.5 53.1 41.6 45.5 47.9

VFAe (mg l�1)

Acetate 1309.0 437.3 485.6 792.4 72.5 29.4 112.2 221.5 193.2 17.5

Propionate 357.3 389.8 196.9 479.3 73.2 3.5 66.1 141.6 238.1 16.6

Butyrate 108.5 7.6 8.0 46.4 6.5 0 1.6 1.4 0 0

NR, not recorded.

a Chemical oxygen demand.

b Hydraulic retention time.

c Volumetric loading rate.

d Carbohydrate.

e Volatile fatty acid.

f Glucose/sucrose.

g Cardboard.

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 2 5 3 1 – 2 5 3 92534

Table 3 – Concentration of key carbohydrates in each of the fungal thermozyme hydrolysates used throughout the trial.

Concentration of analyte (g l�1)

Cellobiose Glucose Galactose D-xylose Mannose HMFa Fulfural

Apple pulp hydrolysate – 0.5 1.42 – 0.26 – –

Bread hydrolysate – 3.01 – 0.63 0.04 – –

Carob hydrolysate 1.87 1.09 – 1.69 1.47 0.21 –

Cardboard hydrolysate – 0.77 – 0.36 – – –

a 5-Hydroxymethylfurfural.

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 2 5 3 1 – 2 5 3 9 2535

methane content of the biogas fluctuated initially for both

reactors, by day 228 the values recorded were similar to those

obtained under stable operation prior to day 128.

3.2. Application of thermozyme hydrolysates

Apple pulp hydrolysate was supplied to the bioreactors on day

486 of the trial (Period 6). The COD removal efficiency averaged

89.7% and 88.0% during this period, for the thermophilic and

mesophilic reactors, respectively (Tables 1 and 2). The average

% methane composition of the R2 biogas was the same as that

from the period before while the average % methane produced

by R1 increased slightly from 47% to 50% (Tables 1 and 2). Both

reactors achieved >98% carbohydrate removal from the

influent throughout this period (Tables 1 and 2). Bread

hydrolysate was supplied to the reactors for 10 HRTs

throughout Period 7 of the reactor trial. While COD removal

efficiency by both reactors during this period was comparable

to the performance exhibited throughout feeding of apple

pulp hydrolysate, this feed period was characterised by

a marked rise in the methane content of the biogas produced

by both reactors (Fig. 3). The overall average % methane

increased to 59.6% of the biogas produced by R1 – a rise of

almost 10% from the previous feeding period, and to 53.1% by

the thermophilic reactor (Tables 1 and 2). Total carbohydrate

Fig. 1 – % Soluble COD removal by R1 & R2 during the trial. (

operational periods are denoted by arrows (Tables 1 and 2).

removal remained at almost 100% for both reactors, indicating

that effluent COD was comprised of intermediate VFAs. The

reactors were sustained on glucose and sucrose as influent

during Period 8 of the trial. Carob powder hydrolysate was

subsequently supplied on day 620 until day 646 (Period 9). An

immediate decrease in COD removal efficiency by both reac-

tors was noted upon addition of the new influent, which was

unsurprising given the relative complexity of the carob

powder hydrolysate (Table 3). After an initial period of adap-

tation of 2.5 HRTs (Fig. 1), reactor performance improved until

the end of the trial on day 646, with high COD removal effi-

ciencies of 98% by R1, and 96% by R2. This period was char-

acterised by better performance by the mesophilic reactor

than its thermophilic counterpart, especially in terms of COD

removal and low effluent propionate levels (Figs. 1 and 2).

Period 10 of the reactor trial proceeded between days 646 and

670, during which a hydrolysate produced from cardboard

waste was supplied as influent to the bioreactors. This period

was of note due to the superior performance of the thermo-

philic reactor in the treatment of this waste stream. On

commencement of this period, the COD removal efficiencies

of both reactors decreased, but in a reverse of the trend

observed during application of carob powder, this decrease

was more marked for the mesophilic reactor. Although the

performance of both reactors subsequently improved, the

, mesophilic; , thermophilic); VLR ( ). Different

Fig. 2 – Effluent VFA levels (mg lL1) from the mesophilic reactor, R1 (A), and the thermophilic reactor, R2 (B) throughout the

trial period. Different operational periods are denoted by arrows (Tables 1 and 2).

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 2 5 3 1 – 2 5 3 92536

mesophilic reactor was prone to unstable operation

throughout this period. By the end of the trial period, the COD

removal efficiency of both reactors had reached 97%.

4. Discussion

The focus of this study was to examine the feasibility of ther-

mophilic anaerobic treatment of hydrolysates generated from

the action of thermozymes produced by T. emersonii. In light of

this, one of the major objectives was to develop a thermophilic

adapted sludge. This was achieved very successfully by oper-

ating R2 at 55 �C from the commencement of the trial, while

maintaining a control reactor, R1, at 37 �C. During the start-up

period, the mesophilic reactor exhibited superior performance

with regard to key parameters of reactor performance such as

COD removal efficiency, methane content of biogas, and VFA

content of the reactor effluent. The short start-up time observed

for both R1 and R2 can be attributed to the use of established

anaerobic granules as inoculum and the ability of the hybrid

design to retain high biomass content within the lower, sludge

bed section of the reactor. Stable operation was observed after 9

days of mesophilic reactor operation, and 78 days of

thermophilic operation. Rapid start-up of thermophilic reactors

using mesophilic sludge has been reported by a number of

authors. Fang and Lau (1996), for example, demonstrated stable

thermophilic operation within 75 days. Other authors have

observed similar fast start-up of thermophilic reactors by using

a strategy of single increase of temperature (Wiegant and Let-

tinga, 1985; van Lier et al., 1994; Ohtsuki et al., 1992; Dinsdale

et al., 1997; Syutsubo et al., 1997; Philpott, 2000). During the

subsequent destabilisation of both reactors upon the applica-

tion of ahigher loadingrate, thethermophilicbiomasswasmore

susceptible with more marked decreases in COD removal effi-

ciency, percentage methane production and increasing levels of

effluent VFAs, especially propionate. While the mesophilic

reactor gradually adapted and recovered from this shock

loading, the loading rate had to be lowered to prevent total

failure of the thermophilic reactor due to build up of high

concentrations of VFAs.

During the application of apple and bread hydrolysates as

influent, the performances of both reactors were highly

comparable with equal capacities for COD reduction: 88% and

88.4% by R1, and, 89.7% and 88.2% by R2 on apple and bread

hydrolysates, respectively. The VFA content of the R2 effluent

in these trials remained higher than that of the R1 effluent,

Fig. 3 – % Methane content of the biogas produced by R1 & R2 ( , mesophilic , thermophilic) throughout the trial

period. Different operational periods are denoted by arrows (Tables 1 and 2).

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 2 5 3 1 – 2 5 3 9 2537

with acetate, followed by propionate as the main contributors

to effluent COD. Lata et al. (2002) also reported that acetate

and propionate were the initial and main contributors to

effluent COD concentration during the mesophilic treatment

of vegetable market waste. The highly comparable state of

operation in both reactors continued until the influent was

changed to carob powder hydrolysate. During treatment of

this feedstock, the mesophilic reactor was superior in all

respects. This situation reversed, however, upon feeding

cardboard hydrolysate to the reactors during which period the

thermophilic reactor out-performed the mesophilic reactor in

all respects, including lower concentrations of VFA in the

reactor effluent. In a comparative study of thermophilic and

mesophilic upflow filters treating paper pulp liquor at influent

organic loading rates of up to 3.87 g l�1 d�1, Ahn and Forster

(2002) reported that the thermophilic reactor achieved a more

stable level of performance, in terms of COD removal and

methane content of the biogas, than that of the mesophilic

reactor. Analogous results were obtained by other authors,

who reported higher treatment efficiencies of thermophilic

over mesophilic treatment (Harris and Dague, 1993; Borja

et al., 1995; Mendez et al., 1995). Many other studies have

revealed that thermophilic high-rate processes are charac-

terised by high concentrations of effluent VFAs, particularly

under high loading conditions. Propionate is often the first

and main VFA which accumulates (Rudd et al., 1985; Wiegant

and Lettinga, 1985; Wiegant and de Man, 1986; Harris and

Dague, 1993; Duran and Speece, 1997). Accumulation of

propionate in the effluent from the thermophilic reactor was

not observed to be significant in this study.

5. Conclusions

Rapid reactor start-up and subsequent successful mesophilic

and thermophilic anaerobic treatment of the thermozyme

hydrolysates of apple pomace, bread, carob powder and

cardboard was achieved by using UAHR reactor technology

and a granular source of inoculum. A successful one-step

transition to 55 �C and subsequent high-rate thermophilic

treatment was realised. Treatment efficiencies of >98%

soluble COD removal were attained by both reactors

throughout the trial at a VLR of 4.5 gCOD l�1 d�1 and an HRT of

48 h. Overall, the performance of the mesophilic reactor was

observed to be less susceptible than its thermophilic coun-

terpart to perturbations in influent composition and loading

rate. However, the thermophilic reactor performed compa-

rably to the mesophilic reactor when exhibiting stable oper-

ating conditions. Both reactors degraded a variety of

hydrolysates readily, with apple pomace, bread, and card-

board hydrolysates mineralised comparably. Ahn and Forster

(2002) acknowledged that, in assessing the potential of ther-

mophilic digesters, the main technological concern is not

whether a thermophilic system will operate but whether it

has real advantages over a mesophilic system for any specific

wastewater. It is clear that in the context of this study, the

comparable performance at 55 �C to that of the control reactor

at 37 �C strongly suggests that thermophilic treatment is

a positive option. The fact that the thermozyme hydrolysates

are produced at high temperature (60–80 �C) is a further

impetus to the application of thermophilic AD, without the

need for cooling the hydrolysate, in the 2-stage system

proposed in this study.

Acknowledgements

The receipt of financial support from the Irish Environmental

Protection Agency is gratefully acknowledged.

r e f e r e n c e s

Ahn, J.H., Forster, C.F., 2002. A comparison of mesophilic andthermophilic anaerobic upflow filters treating paper-pulp-liquors. Process Biochemistry 38 (2), 257–262.

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 2 5 3 1 – 2 5 3 92538

Ahn, Y.H., Min, K.S., Speece, R.E., 2001. Full scale UASB reactorperformance in the brewery industry. EnvironmentalTechnology 22 (4), 463–476.

APHA (2001). Standard Methods for the Examination of Water andWastewaters. Washington DC 20005: American Public HeathAssociation, American Waterworks Association and WaterEnvironment Federation.

Bernal, O., Llabres, P., Cecchi, F., Mata-Alvarez, J., 1992. Acomparative study of the thermophilic biomethanization ofputrescible organic wastes. Odpadnı vody/Wastewaters 1 (1),197–206.

Borja, R., Martin, A., Banks, C.J., Alonso, V., Chica, A., 1995. Akinetic study of anaerobic digestion of olive mill wastewater atmesophilic and thermophilic temperatures. EnvironmentalPollution 88 (1), 13–18.

Callaghan, F.J., Wase, D.A.J., Thayanithy, K., Forster, C.F., 1999.Co-digestion of waste organic solids: batch studies.Bioresource Technology 67 (2), 117–122.

Considine, P.J., Ororke, A., Hackett, T.J., Coughlan, M.P., 1988.Hydrolysis of beet pulp polysaccharides by extracts of solid-state cultures of Penicillium Capsulatum. Biotechnology andBioengineering 31 (5), 433–438.

Demirel, B., Scherer, P., 2008. Production of methane from sugar beetsilage without manure addition by a single-stage anaerobicdigestion process. Biomass and Bioenergy 32 (3), 203–209.

Dinsdale, R.M., Hawkes, F.R., Hawkes, D.L., 1997. Comparison ofmesophilic and thermophilic upflow anaerobic sludge blanketreactors treating instant coffee production wastewater. WaterResearch 31 (1), 163–169.

Dubois, M., Gilles, K.A., Hamilton, J., Rebers, P.A., Smith, F., 1956.Colorimetric method for determination of sugars and relatedsubstances. Analytical Chemistry 28, 350–356.

Duran, M., Speece, R.E., 1997. Temperature-staged anaerobicprocesses. Environmental Technology 18 (7), 747–753.

Fang, H.H.P., Lau, I.W.C., 1996. Startup of thermophilic (55 degreesC) UASB reactors using different mesophilic seed sludges.Water Science and Technology 34 (5–6), 445–452.

Forster-Carneiro, T., Perez, M., Romero, L.I., 2008. Influence oftotal solid and inoculum contents on performance ofanaerobic reactors treating food waste. BioresourceTechnology 99 (15), 6994–7002.

Grohmann, K., Bothast, R.J., 1994. Pectin-rich residues generatedby processing of citrus fruits, apples, and sugar-beets –enzymatic hydrolysis and biological conversion to value-added products. Enzymatic Conversion of Biomass for FuelsProduction 566, 372–390.

Gunaseelan, V.N., 1997. Anaerobic digestion of biomass formethane production: a review. Biomass & Bioenergy 13 (1–2),83–114.

Harris, W.L., Dague, R.R., 1993. Comparative performance ofanaerobic filters at mesophilic and thermophilictemperatures. Water Environment Research 65 (6),764–771.

Hutnan, M., Drtil, M., Mrafkova, L., 2000. Anaerobicbiodegradation of sugar beet pulp. Biodegradation 11 (4),203–211.

Kim, J.K., Cho, J.H., Lee, J.S., Hahm, K.S., Park, D.H., Kim, S.W.,2002. Mass production of methane from food wastes withconcomitant wastewater treatment. Applied Biochemistryand Biotechnology 98, 753–764.

Lata, K., Rajeshwari, K.V., Pant, D.C., Kishore, V.V.N., 2002.Volatile fatty acid production during anaerobic mesophilicdigestion of tea and vegetable market wastes. World Journal ofMicrobiology & Biotechnology 18 (6), 589–592.

Mata-Alvarez, J., Mace, S., Llabres, P., 2000. Anaerobic digestion oforganic solid wastes. An overview of research achievementsand perspectives. Bioresource Technology 74 (1), 3–16.

Mendez, R., Lema, J.M., Soto, M., 1995. Treatment of seafood-processing wastewaters in mesophilic and thermophilicanaerobic filters. Water Environment Research 67 (1), 33–45.

Nishio, N., Nakashimada, Y., 2007. Recent development ofanaerobic digestion processes for energy recovery fromwastes. Journal of Bioscience and Bioengineering 103 (2),105–112.

Noike, T., Endo, G., Chang, J.E., Yaguchi, J.I., Matsumoto, J.I., 1985.Characteristics of carbohydrate degradation and the rate-limiting step in anaerobic digestion. Biotechnology andBioengineering 27 (10), 1482–1489.

Ohtsuki, T., Watanabe, M., Miyaji, Y., 1992. Start up ofa thermophilic UASB (Upflow Anaerobic Sludge Blanket)reactor using micro-carrier and mesophilic granular sludge.Water Science and Technology 26 (3–4), 877–886.

Okeke, B.C., Obi, S.K.C., 1995. Saccharification of agro-wastematerials by fungal cellulases and hemicellulases. BioresourceTechnology 51 (1), 23–27.

Philpott, U., 2000. Mesophilic and thermophilic UASB treatment ofsulphate-containing wastewater. In: Department ofMicrobiology, vol. Ph.D., National University of Ireland, Galway.

Poirrier, P., Lema, J.M., 2004. Anaerobic hydrolysis andacidogenesis of carbohydrate polymers at psychrophilicconditions in a membrane reactor. In: 10th World Congress onAnaerobic Digestion, Montreal, Canada.

Rintala, J.A., Lepisto, S.S., 1997. Pilot-scale thermophilic anaerobictreatment of wastewaters from seasonal vegetable processingindustry. Water Science and Technology 36 (2–3), 279–285.

Rudd, T., Hicks, S.J., Lester, J.N., 1985. Comparison of thetreatment of a synthetic meat waste by mesophilic andthermophilic anaerobic fluidized-bed reactors. EnvironmentalTechnology Letters 6 (5), 209–224.

Sawada, T., Nakamura, Y., Kobayashi, F., Kuwahara, M.,Watanabe, T., 1995. Effects of fungal pretreatment and steamexplosion pretreatment on enzymatic saccharification of plantbiomass. Biotechnology and Bioengineering 48 (6), 719–724.

Shelton, D.R., Tiedje, J.M., 1984. General method for determininganaerobic biodegradation potential. Applied andEnvironmental Microbiology 47 (4), 850–857.

Solera, R., Romero, L.I., Sales, D., 2002. The evolution of biomassin a two-phase anaerobic treatment process during start-up.Chemical and Biochemical Engineering Quarterly 16 (1), 25–29.

Srilatha, H.R., Nand, K., Babu, K.S., Madhukara, K., 1995. Fungalpretreatment of orange processing waste by solid-statefermentation for improved production of methane. ProcessBiochemistry 30 (4), 327–331.

Stabnikova, O., Liu, X.-Y., Wang, J.-Y., 2008. Anaerobic digestionof food waste in a hybrid anaerobic solid-liquid system withleachate recirculation in an acidogenic reactor. BiochemicalEngineering Journal 41 (2), 198–201.

Syutsubo, K., Harada, H., Ohashi, A., Suzuki, H., 1997. An effectivestart-up of thermophilic UASB reactor by seedingmesophilically-grown granular sludge. Water Science andTechnology 36 (6–7), 391–398.

Thompson, G., Swain, J., Kay, M., Forster, C.F., 2001. Thetreatment of pulp and paper mill effluent: a review.Bioresource Technology 77 (3), 275–286.

Tuohy, M.G., Coughlan, M.P., 1992. Production of thermostablexylan-degrading enzymes by Talaromyces Emersonii.Bioresource Technology 39 (2), 131–137.

Tuohy, M.G., Coughlan, T.L., Coughlan, M.F., 1990. Solid stateversus liquid cultivation of Talaromyces emersonii on strawsand pulps: enzyme productivity. Advances in BiologicalTreatments of Lignocellulosic Materials, 153–175.

Tuohy, M.G., Laffey, C.D., Coughlan, M.P., 1994. Characterization ofthe individual components of the xylanolytic enzyme system ofTalaromyces emersonii. Bioresource Technology 50 (1), 37.

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 2 5 3 1 – 2 5 3 9 2539

van Lier, J.B., Boersma, F., Debets, M.M.W.H., Lettinga, G., 1994.High rate thermophilic anaerobic wastewater treatment incompartmentalized upflow reactors. Water Science andTechnology 30 (12), 251–261.

Veeken, A., Hamelers, B., 1999. Effect of temperature onhydrolysis rates of selected biowaste components.Bioresource Technology 69 (3), 249–254.

Wiegant, W.M., de Man, A.W.A., 1986. Granulation of biomass inthermophilic upflow anaerobic sludge blanket reactorstreating acidified wastewaters. Biotechnology andBioengineering 28 (5), 718–727.

Wiegant, W.M., Lettinga, G., 1985. Thermophilic anaerobicdigestion of sugars in upflow anaerobic sludge blanketreactors. Biotechnology and Bioengineering 27 (11), 1603–1607.