Available online at www.sciencedirect.com

www.elsevier.com/locate/wasman

Waste Management 28 (2008) 1654–1659

Digestion of frozen/thawed food waste in the hybrid anaerobicsolid–liquid system

O. Stabnikova *, X.Y. Liu, J.Y. Wang

School of Civil and Environmental Engineering, Nanyang Technological University, 50 Nanyang Avenue, Singapore 639798, Singapore

Accepted 15 May 2007Available online 10 September 2007

Abstract

The hybrid anaerobic solid–liquid (HASL) system, which is a modified two-phase anaerobic digester, is to be used in an industrialscale operation to minimize disposal of food waste at incineration plants in Singapore. The aim of the present research was to evaluatefreezing/thawing of food waste as a pre-treatment for its anaerobic digestion in the HASL system. The hydrolytic and fermentation pro-cesses in the acidogenic reactor were enhanced when food waste was frozen for 24 h at �20 �C and then thawed for 12 h at 25 �C (exper-iment) in comparison with fresh food waste (control). The highest dissolved COD concentrations in the leachate from the acidogenicreactors were 16.9 g/l on day 3 in the control and 18.9 g/l on day 1 in the experiment. The highest VFA concentrations in the leachatefrom the acidogenic reactors were 11.7 g/l on day 3 in the control and 17.0 g/l on day 1 in the experiment. The same volume of methanewas produced during 12 days in the control and 7 days in the experiment. It gave the opportunity to diminish operational time of batchprocess by 42%. The effect of freezing/thawing of food waste as pre-treatment for its anaerobic digestion in the HASL system was com-parable with that of thermal pre-treatment of food waste at 150 �C for 1 h. However, estimation of energy required either to heat thesuspended food waste to 150 �C or to freeze the same quantity of food waste to �20 �C showed that freezing pre-treatment consumesabout 3 times less energy than thermal pre-treatment.� 2007 Elsevier Ltd. All rights reserved.

1. Introduction

The hybrid anaerobic solid–liquid (HASL) system, whichis a modified two-phase anaerobic digester, is to be used in anindustrial scale operation to minimize disposal of food wasteat incineration plants in Singapore (Wang et al., 2005). Sim-ilar to a conventional two-phase anaerobic digester, the firstphase, including hydrolysis and acidification of food waste,is performed in the acidogenic reactor. The second phase,including acetogenesis and methanogenesis, is carried outin the methanogenic reactor. As liquefaction (solubilization)and hydrolysis of solid wastes are rate-limiting steps inanaerobic digestion (Shin et al., 2001), an increase of theseprocessing rates should lead to better performance of anaer-obic treatment of food waste in the HASL system. To

0956-053X/$ - see front matter � 2007 Elsevier Ltd. All rights reserved.

doi:10.1016/j.wasman.2007.05.021

* Corresponding author. Tel.: +65 6790 4740; fax: +65 6791 0676.E-mail address: costab@ntu.edu.sg (O. Stabnikova).

promote solubilization of organic matter and release ofintracellular material from plant cells, cell disruption canbe used (Mata-Alvarez et al., 2000).

Cell disruption is used to improve anaerobic digestion ofactivated sludge. As a result of cell disruption, intracellularand cell wall polymers, including polysaccharides, proteins,lipids, and other macromolecules, are released into sur-rounding medium. Consequently, cell disruption enhancedanaerobic fermentation of activated sludge and increasedbiogas generation (Bien et al., 2004). For microbial or plantcell disruption, different disintegration methods such asmechanical (e.g., sonication), chemical (e.g., alkali treat-ment), osmotic (e.g., NaCl treatment), thermal (e.g., heattreatment), and biological (e.g., treatment by enzymes)were proposed (Erdinclerm and Vesilind, 2000; Mata-Alva-rez et al., 2000). Cell disruption can be achieved also duringfreezing of material at low temperature due to intracellularice crystals formation causing damage of cell membranes(Thomashow, 1998; Webb et al., 1996).

3

2

4

1

1, acidogenic reactor,

Ra; 2, methanogenic reactor, Rm; 3, peristaltic pump; 4, wet gas meter.

Fig. 1. 1, acidogenic reactor, Ra; 2, methanogenic reactor, Rm;3, peristaltic pump; 4, wet gas meter.

O. Stabnikova et al. / Waste Management 28 (2008) 1654–1659 1655

The aim of the present research was to evaluate freezing/thawing of food waste as a pre-treatment for its anaerobicdigestion in the HASL system.

2. Materials and methods

2.1. Food waste, anaerobic microbial sludge and microbial

granules

Food waste was collected from a canteen of the univer-sity. Waste was shredded into particles with an average sizeof 6.0 mm in a Robot–Coupe Shredder (CL50 Ultra,Hobart, France). The composition of food waste used inthe experiments was as follows (% of wet weight): vegetableroots, 50, orange peels, 20, rice, 15, and noodles, 15. Thetotal solids (TS) contents were 18.6 ± 1.7% and17.9 ± 0.7%, and the volatile solids (VS) contents were92.9 ± 1.2% and 94.6 ± 2.5% of TS in the mixed freshand frozen/thawed mixed food waste, respectively.

Anaerobic microbial sludge, collected from an anaerobicdigester of a local wastewater treatment plant, was used asinoculum for the acidogenic reactor. The concentrations ofTS and VS in the sludge were 3.8 ± 0.1 and 2.7 ± 0.1 g/l,respectively. Sludge granules collected from a 5 l UASBreactor, which had been operated for more than 6 monthswith a synthetic wastewater (SWW), were used as inoculumfor the methanogenic reactor. SWW had the followingcomposition (mg/l): peptone, 500; glucose, 1700; meatextract, 350; NaHCO3, 1500; NH4Cl, 400; K2HPO4, 100;CaCl2 Æ 6H2O, 24; MgSO4 Æ 7H2O, 27; FeSO4 Æ 7H2O, 20;MnSO4 Æ H2O, 0.05; (NH4)6Mo7O4 Æ 24H2O, 0.05; AlCl3,0.05; CoCl2 Æ 6H2O, 0.05; NiCl2, 0.05; ZnCl2, 0.05; CuCl2,0.05; H3BO3, 0.05; distilled water was added to 1 l. Theinfluent substrate concentration, in terms of COD, was2500 mg/l.

2.2. Characteristics and operation of the HASL system

The HASL system included an acidogenic columnreactor to treat solid food waste and an upflow anaero-bic sludge blanket (UASB) methanogenic reactor to treatliquid leachate, collected from the bottom of the acido-genic reactor (Fig. 1). Part of the effluent from the meth-anogenic reactor was used for the dilution of the acidleachate from the acidogenic reactor to maintain optimalpH for methanogenesis, and the rest of the effluent fromthe methanogenic reactor was recycled into the acido-genic reactor to avoid addition of water for food wastehydrolysis.

The experiment was carried out simultaneously in twoidentical HASL systems operated at 35 �C for 12 days.Food waste (800 g), anaerobic microbial sludge as inocu-lum (1 l), and distilled water (1 l) were placed in each acido-genic reactor. Untreated food waste was used in thecontrol. Food wastes (800 g), frozen for 24 h at �20 �Cand then thawed for 12 h at 25 �C, were used inexperiment.

2.3. Sampling and chemical analysis

Leachate from the acidogenic reactor and effluent fromthe methanogenic reactor were collected for analysis daily.The pH value was measured using a pH meter (CORNING145, Halstead, Essex, UK). Total suspended solids, volatilesolids, and chemical oxygen demand (COD) were deter-mined in the well-mixed samples in triplicate by standardmethods (Standard Methods for the Examination of Waterand Wastewater, 1998). For volatile fatty acids (VFA),determination samples were filtrated through Whatman0.2 lm nitrocellulose membrane filters and were then ana-lyzed using HPLC (Perkin Elmer, Series 200, Norwalk,CT, USA). The HPLC was equipped with a220 mm · 4.6 mm polypore H column and an UV 210 nmdetector. The mobile phase was 0.005 N H2SO4 with a flowrate of 0.15 ml/min.

Gas production was monitored by a wet gas meter(Ritter TG 05, Bochum, Germany), while gas composi-tion was analyzed by a Hewlett Packard GC HP5890A(HACH, Avondale, PA, USA) for methane, carbon diox-ide, and nitrogen. The GC was equipped with a thermalconductivity detector and a stainless-steel column packedwith Hayesep Q (80/100 mesh). The operational temper-atures of injector, detector and column were kept at100 �C, 200 �C and 50 �C, respectively. Helium was usedas a carrier gas at a flow rate of 40 ml/min. The pH,concentrations of dissolved COD and VFA, as well asbiogas production and methane content, were determineddaily.

All analytical determinations were performed at least intriplicate. Values of different parameters were expressed asmean ± standard deviation.

Acc

umul

ativ

e m

etha

ne p

rodu

ctio

n (l

)

0

5

10

15

20

25

30

35

40

Time (days)0 1 2 3 4 5 6 7 8 9 10 11 12 13

Time (days)0 1 2 3 4 5 6 7 8 9 10 11 12 13

CO

D (

mg/

l)

0

5000

10000

15000

20000

Leachate of Ra in CLeachate of Ra in E

Effluent of Rm in CEffluent of Rm in E

1656 O. Stabnikova et al. / Waste Management 28 (2008) 1654–1659

2.4. Comparison of structure of fresh and frozen/thawed

food waste

The structure of fresh and frozen/thawed food wastewas observed by a scanning electron microscopy (SEM)Stereoscan 420 (Leica Cambridge Instruments). Foodwaste were prepared for SEM by fixation for 1–4 h in 2%(v/v) glutaraldehyde, washed three times for 20 min with0.10 M sodium cacodylate buffer (pH 7.2), and were dehy-drated in a graded butyl alcohol series (50, 70, 85, 95, 100%v/v). Dehydrated food waste was dried using a freeze dryer(Bal-Tec CPD 030), sputter-coated with gold at 20 mA in ahigh vacuum (2.8 · 10�6 Torr) and low temperature(�170 �C) cryo-chamber for 90 s, and then viewed withSEM at 20 kV.

3. Results

The fresh and frozen/thawed food wastes were placed inthe acidogenic reactors of control and experimental HASLsystems, respectively. The results of anaerobic digestion offood waste in the HASL systems are shown in Figs. 2 and3. The lowest pH in the leachate from the acidogenic reac-tors in the control and the experiment were similar, 4.8 and5.0, respectively (Fig. 2a). However, the lowest pH in thecontrol was observed on day 3, meanwhile the same pHvalue was reached in the experiment on the first day ofthe process. Rapid drop of pH in the leachate from the aci-

VFA

(m

g/l)

0

4000

8000

12000

16000

Time (days)0 1 2 3 4 5 6 7 8 9 10 11 12 13

Time (days)0 1 2 3 4 5 6 7 8 9 10 11 12 13

pH

4

5

6

7

8

Leachate of Ra in CLeachate of Ra in E

Effluent of Rm in CEffluent of Rm in E

Fig. 2. Value of pH (a); concentration of VFA in the leachate from theacidogenic reactor (Ra) and in the effluent from the methanogenic reactor(Rm) (b) in control (C) and experiment (E).

Methane production in C Methane production in E

Fig. 3. Concentration of dissolved COD in the leachate from theacidogenic reactor (Ra) and in the effluent from the methanogenic reactor(Rm) (a); accumulative methane production in control (C) and experiment(E) (b).

dogenic reactor in the experiment demonstrated a higherrate of hydrolysis of frozen/thawed food waste. The initialdissolved COD concentration in the acidogenic reactorincreased almost twice in the experiment when frozen/thawed food waste was used (Fig. 3a). According toSEM images of fresh and frozen/thawed vegetable wastes(Fig. 4), the structure of food waste material after freez-ing/thawing became looser, which causes increase of dis-solved organics concentration.

High initial concentration of dissolved COD resulted inrapid rise of VFA concentration in the acidogenic reactorin experiment. The peaks of VFA and dissolved COD con-centrations were higher and were reached in a shorter timein the experiment than in the control (Figs. 2b, 3a). Thehighest VFA concentrations in the leachate from the acido-genic reactors were 11.7 g/l on day 3 in the control and17.0 g/l on day 1 in the experiment (Fig. 3b). The highestdissolved COD concentrations in the leachate from the aci-dogenic reactors were 16.9 g/l on day 3 in the control and18.9 g/l on day 1 in the experiment (Fig. 3a). The concen-trations of dissolved COD in the effluent from the metha-nogenic reactors were lower in the experiment than thatin the control (Fig. 3a). Maximum rates of COD, VFA

Fig. 4. SEM images of fresh (a) and frozen/thawed (b) vegetable waste.

O. Stabnikova et al. / Waste Management 28 (2008) 1654–1659 1657

and methane production were 6.3, 4.6 and 5.4 l/day,respectively, in the control and 7.3, 15.4 and 7.6 l/day,respectively, in the experiment. The maximum rates ofCOD, VFA and methane production increased by 20%,235% and 41%, respectively, in the experiment in compar-ison with the control. It indicated the intensification ofhydrolysis and acidogenesis in the acidogenic reactor andmethanogenesis in the methanogenic reactor in the experi-ment. However, the average value of all parameters for 12days of food waste anaerobic digestion in the control andthe experiment did not differ much. The average rates ofCOD, VFA and methane production increased by 4%,9% and 9% in the experiment in comparison with the con-trol. The total VS removal was 77.2% and 80.1% in thecontrol and the experiment. Yield of methane production,in terms of l/g VS removed, increased by 6.7% in the exper-iment in comparison with the control. The similar increaseby 7.1% was observed when food waste pre-treated at150 �C for 1 h was used for anaerobic digestion insteadof fresh food waste (Wang et al., 2006). Therefore, succes-sive freezing and thawing of food waste has not muchchanged the stoichiometric parameters in the HASL sys-tem, but significantly changed the maxima of the parame-ters, as well as kinetic parameters. Almost the samequantity of dissolved COD and VFA were produced inthe acidogenic reactor for 12 days in the control and for

7 days in the experiment (Table 1). The same volume ofmethane was produced in 12 days in the control and 7 daysin the experiment (Fig. 3b, Table 1). Meanwhile, the aver-age methane content in biogas was 66% in the control and69% in the experiment.

Use of frozen/thawed food waste instead of fresh onegave the opportunity to diminish operational time of thebatch process by 42%. The effect of freezing/thawing of foodwaste as a pre-treatment step for its anaerobic digestion inthe HASL system was comparable with thermal pre-treat-ment of food waste at 150 �C for 1 h, which helped to dimin-ish time needed to produce the same quantity of methanetwice in comparison with the anaerobic digestion of freshfood waste due to speed up of organic matter hydrolysis(Wang et al., 2006). Therefore, freezing and thawing of foodwaste can be considered as alternative method for thermalpre-treatment of food waste before its anaerobic digestion.The following additional considerations were taken intoaccount. To prevent the decay of collected waste, as wellas the production of the spores and toxins of micromycetesand formation of nuisance odor in place of waste collection,the food waste could be stored either outdoor in winter or inthe industrial refrigerator in summer in the countries withmoderate climate. It can be accumulated in the industrialrefrigerator in the tropical countries to avoid decay ofwaste. Freezing and thawing of food waste followed by cen-trifugation is considered a safe and energy-saving methodfor the treatment of food waste and is expected to be appli-cable widely for its recycling, transportation, and incinera-tion (Shimiya and Miyawaki, 2006). The freezing in arefrigerator, following with the thawing before anaerobicdigestion, is not only an essential procedure for food wastemanagement but also is a useful pre-treatment procedure asshown in the present study.

To calculate the energy required to increase temperatureof food waste from 25 �C to 150 �C, it was assumed thatthere was no heat loss or water evaporation. The energyfor heating of suspension of 800 g of food waste in1000 ml of water (Qt ), used for anaerobic digestion in pres-ent study, was calculated by equation:

Qt ¼ Corganic Morganic Dt þ Cwater ðMwater þ 1000ÞDt;

where Corganic is specific heat capacity of organic matter,0.89 J/g �C; Cwater is specific heat capacity of water,4.18 J/g �C; Morganic is mass of organic matter in 800 g offood waste with the content of TS 18.6%, 148.72 g; Mwater

is mass of water in 800 g of food waste, 651.28 g; 1000 ismass of water added to prepare food waste suspension;Dt is the difference between the final temperature, 150 �C,and the initial temperature, 25 �C.

Qt ¼ 0:89 J=g �C� 148:72 g� ð150� 25Þ �Cþ 4:18 J=g�C

� ð651:28þ 1000Þ g� ð150� 25Þ �C ¼ 879:33 kJ

If some portion of water will be evaporated during thermalpre-treatment, the calculated value will be even higherbecause of partial liquid–gas phase transition.

Table 1Comparison of anaerobic digestion of frozen/thawed food waste (experiment) with fresh food waste (control) in the HASL system

Parameters Process duration (days)

Control Experiment

7 12 7 12

Dissolved COD produced in the acidogenic reactor (g) 93.6 ± 3.7 115.0 ± 4.8 103 ± 4.4 119.6 ± 5.9VFA produced in the acidogenic reactor (g) 85.0 ± 2.6 102.7 ± 3.3 98.5 ± 4.7 112.8 ± 4.9Total methane production (l) 25.7 31.7 31.8 34.4Methane yield (l/gVS removed) 0.30 0.32Average methane content in biogas (%, v/v) 68 ± 2.6 65.8 ± 3.7 69.2 ± 2.5 66.1 ± 4.4

1658 O. Stabnikova et al. / Waste Management 28 (2008) 1654–1659

The energy for freezing of 800 g of food waste was cal-culated by following equation:

Qf ¼ Corganic Morganic Dt1 þ Cwater Mwater Dt1 þ Cice M ice Dt2

þ Lfwater Mwater;

where Cice is specific heat capacity of ice, 2.06 J/g �C; Mice

is mass of ice equal to mass of water in food waste; Lfwater

is latent heat of fusion, 334.4 J/g water; Dt1 is the differencebetween the initial temperature, 25 �C, and final tempera-ture, 0 �C; Dt2 is the difference between the initial tempera-ture, 0 �C, and the final temperature, �20 �C:

Qf ¼ 0:89 J=g �C� 148:72 g� ð25þ 20Þ �Cþ 4:18 J=g �C� 651:28 g� ð25� 0Þ �Cþ 2:06 J=g �C� 651:38 g� ð0þ 20Þ �Cþ 334:4 J=g� 651:28 g ¼ 318:64 kJ

The price of equipment for thermal treatment (industrialautoclave) is much higher than the cost of an industrialfreezer of the same volume because an autoclave is a vesselworking at elevated pressure, while a freezer works atatmospheric pressure. For example, the price of an indus-trial autoclave with a volume of 3 m3 is approximately$16,000, meanwhile the price of an industrial freezer witha volume of 3 m3 is approximately $4000.

Consequently, an estimation of energy required either toheat the suspended food waste to 150 �C or to freeze thesame quantity of food waste to �20 �C shows that freezingpre-treatment consumes about 3 times less energy thanheating pre-treatment. If the energy losses from heatingand refrigerating processes are not considered, the energyfor heating 800 g of food waste (TS = 18.6% w/w), sus-pended in 1000 ml of water, from 25 �C to 150 �C is879 kJ, and the energy for freezing 800 g of food wastefrom 25 �C to �20 �C is 319 kJ. The energy of methane,released during anaerobic digestion of 800 g of food waste,is 1273 kJ; therefore the balance of utilized and releasedenergy in food waste anaerobic digestion is more favorablefor freezing/thawing pre-treatment than for heating pre-treatment.

4. Conclusions

Freezing/thawing of food waste facilitated the hydro-lytic and fermentation processes in the acidogenic reactor

and ensured faster supply of nutrients in the methano-genic reactor. The same volume of methane was pro-duced in 12 days in control with fresh food waste and7 days in experiment with frozen/thawed food waste. Itgave an opportunity to diminish operational time ofbatch process by 42%. The effect of freezing/thawing offood waste as pre-treatment for its anaerobic digestionin the HASL system is comparable with that of thermalpre-treatment of food waste at 150 �C for 1 h. However,estimation of energy required either to heat the sus-pended food waste to 150 �C or to freeze the same quan-tity of food waste to �20 �C showed that freezing pre-treatment consumes about 3 times less energy than ther-mal pre-treatment.

Acknowledgement

The research was supported by a grant from the Na-tional Environment Agency of the Ministry of the Environ-ment and Water Resources of the Republic of Singapore.

References

Bien, J.B., Malina, G., Bien, J.D., Wolny, L., 2004. Enhancing

anaerobic fermentation of sewage sludge for increasing biogas

generation. Journal of Environmental Science and Health, Part A,

Toxic/hazardous Substances and Environmental Engineering 39,

939–949.

Erdinclerm, A., Vesilind, P.A., 2000. Effect of cell disruption on

compactibility of biological sludges. Water Science and Technology

42, 119–126.

Mata-Alvarez, J., Mace, S., Llabres, P., 2000. Anaerobic digestion of

organic solid wastes. An overview of research achievements and

perspectives. Bioresource Technology 74, 3–16.

Shimiya, Y., Miyawaki, O., 2006. A test apparatus for freezing-thawing

dehydration of food wastes. Japan Journal of Food Engineering 7, 31–

38.

Shin, H.S., Han, S.K., Song, Y.C., Lee, C.Y., 2001. Multi-step sequential

batch two-phase anaerobic composting of food waste. Environmental

Technology 22, 271–279.

Standard Methods for the Examination of Water and Wastewater, 1998.

20th ed. Washington, DC: American Public Health Association/Water

Environment Federation.

Thomashow, M.F., 1998. Role of cold-responsive genes in plant freezing

tolerance. Plant Physiology 118, 1–8.

Wang, J.Y., Zhang, H., Stabnikova, O., Tay, J.H., 2005. Comparison of

lab-scale and pilot-scale hybrid anaerobic solid–liquid systems oper-

ated in batch and semi-continuous modes. Process Biochemistry 40,

3580–3586.

O. Stabnikova et al. / Waste Management 28 (2008) 1654–1659 1659

Wang, J.Y., Liu, X.Y., Kao, J.C.M., Stabnikova, O., 2006. Digestion of

pre-treated food waste in hybrid anaerobic solid–liquid (HASL)

system. Journal of Chemical Technology and Biotechnology 83,

345–351.

Webb, M.S., Gilmour, S.J., Thomashow, M.F., Steponkus, P.L., 1996.

Effects of COR6.6 and COR15am polypeptides encoded by COR (cold-

regulated) genes of Arabidopsis thaliana on dehydration-induced phase

transitions of phospholipid membranes. Plant Physiology 111, 301–312.