Enhanced anaerobic digestion of food waste by thermal and ozonation pretreatment methods

lable at ScienceDirect

Journal of Environmental Management 146 (2014) 142e149

Contents lists avai

Journal of Environmental Management

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

Enhanced anaerobic digestion of food waste by thermal and ozonationpretreatment methods

Javkhlan Ariunbaatar a, d, *, Antonio Panico b, Luigi Frunzo c, Giovanni Esposito a,Piet N.L. Lens d, Francesco Pirozzi e

a Department of Civil and Mechanical Engineering, University of Cassino and the Southern Lazio, Via Di Biasio, 43, 03043 Cassino, FR, Italyb Telematic University Pegaso, Piazza Trieste e Trento, 48, 80132 Naples, Italyc Department of Mathematics and Applications Renato Caccioppoli, University of Naples Federico II, Via Claudio, 21, 80125 Naples, Italyd UNESCO-IHE Institute for Water Education, Westvest 7, 2611 AX Delft, The Netherlandse Department of Civil, Architectural and Environmental Engineering, University of Naples Federico II, Via Claudio, 21, 80125 Naples, Italy

a r t i c l e i n f o

Article history:Received 20 March 2014Received in revised form26 June 2014Accepted 30 July 2014Available online

Keywords:Anaerobic digestionThermal pretreatmentOzonation pretreatmentOrganic solid wasteEnergy requirement

* Corresponding author. Department of Civil and Mversity of Cassino and the Southern Lazio, Via Di Biasio

E-mail addresses: [email protected], j.ariunbaatar@u(J. Ariunbaatar).

http://dx.doi.org/10.1016/j.jenvman.2014.07.0420301-4797/© 2014 Elsevier Ltd. All rights reserved.

a b s t r a c t

Treatment of food waste by anaerobic digestion can lead to an energy production coupled to a reductionof the volume and greenhouse gas emissions from this waste type. According to EU Regulation EC1774/2002, food waste should be pasteurized/sterilized before or after anaerobic digestion. With respect tothis regulation and also considering the slow kinetics of the anaerobic digestion process, thermal andchemical pretreatments of food waste prior to mesophilic anaerobic digestion were studied. A series ofbatch experiments to determine the biomethane potential of untreated as well as pretreated food wastewas carried out. All tested conditions of both thermal and ozonation pretreatments resulted in anenhanced biomethane production. The kinetics of the anaerobic digestion process were, however,accelerated by thermal pretreatment at lower temperatures (<120 �C) only. The best result of647.5 ± 10.6 mlCH4/gVS, which is approximately 52% higher as compared to the specific biomethaneproduction of untreated food waste, was obtained with thermal pretreatment at 80 �C for 1.5 h. On thebasis of net energy calculations, the enhanced biomethane production could cover the energy require-ment of the thermal pretreatment. In contrast, the enhanced biomethane production with ozonationpretreatment is insufficient to supply the required energy for the ozonator.

© 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Food waste (FW) is the largest fraction of municipal solid waste(MSW). A study by the Food and Agricultural Organization (FAO,2011) suggests that one-third of the food produced for humanconsumption is lost or wasted globally, which amounts to about 1.3billion tons per year (FAO, 2011). The generation of MSW and FWare predicted to increase with 51 and 44%, respectively, by 2025;and if the current integrated solid waste management is practised,the global methane production from landfilled FW will increasefrom 3 to 48 Gkg by 2025, contributing to global warming (Adhikariand Barrington, 2006). While it is important to reduce the amountof FW generated, it is also necessary to develop sustainable

echanical Engineering, Uni-, 43, 03043 Cassino, FR, Italy.nesco-ihe.org, [email protected]

treatment and management schemes (Carlsson et al., 2012; Zaman,2013). Hence, these have become an interesting research field inthe scientific community.

As FW has a high moisture content and is readily biodegradable,it serves as a perfect substrate for anaerobic digestion (AD)(Kirchmayr et al., 2003; Zhang et al., 2007). The AD process ischaracterized by a series of biochemical transformations broughtabout by microbial consortia, which convert complex macromole-cules into low molecular weight compounds such as biomethane,carbon dioxide, water and ammonia (Mudhoo and Kumar, 2013).Treating FW with AD produces renewable energy and yields areduction of the amount of waste and greenhouse gas (GHG)emissions. Curry and Pillay (2012) estimated the potential energyrecovery from FW based on the FAO studies, and suggested that 1.3billion ton of waste can produce 894 TWh/year, which is approxi-mately 5% of the total global electrical energy utilization. Never-theless, the long retention time of the AD process is a majorconcern. Therefore, to accelerate the process and to enhance thebiomethane production, methods for pretreating FW prior to the

Table 1Composition of synthetic FW used for the experiment.

% Wet weightfraction

Average fromliteraturereview (%)a

Distribution ofmiscellaneousfraction overthe knowna

fraction (%)

Final concentrationapplied in the BMPtest (%)

Fruits and vegetables 58.4 20.2 78.6e79.0Pasta/rice/flour/cereals 3.6 1.3 4.9e5.0Bread and bakery 4.7 1.7 6.4e6.0Meat and fish 6.1 2.1 8.2e8.0Dairy products 1.4 0.5 1.9e2.0Miscellaneous 25.8 e 0

Total 100 25.8 100

a MTT Agrifood Research Finland (2010).

J. Ariunbaatar et al. / Journal of Environmental Management 146 (2014) 142e149 143

AD process have been developed (Carlsson et al., 2012; Espositoet al., 2011a, 2011b; Mata-Alvarez et al., 2000; Carrere et al., 2010).

Various mechanical, biological, chemical, thermal pretreatmentmethods or a combination of them can be applied for FW. The ef-fects of various pretreatment methods are highly differentdepending on the characteristics of the substrates and the pre-treatment type (Ariunbaatar et al., 2014). Although according to EUregulation EC1774/2002, FW is categorized as a catering waste, andit should be pasteurized or sterilized prior to or after AD (Kirchmayret al., 2003). Taking this regulation into account, a thermal or achemical pretreatment of FW could be more effective. These pre-treatments could cause the degradation of complex molecules aswell as the solubilization of recalcitrant particles, making thesubstrate more available for the anaerobes.

Thermal pretreatment is one of the easiest and most studiedpretreatment methods and has already been applied at a full-scale(Carlsson et al., 2012; Carrere et al., 2010). Among various chemicalmethods, ozonation is an attractive method, as it does not increasethe salt concentration in the reactor and does not have oxidantresidues in the organic waste (Carrere et al., 2010). However, pre-vious research on thermal and ozonation pretreatment methodshave been conducted mostly on wastewater sludge, and only a fewstudies were conducted on the organic fraction of municipal solidwaste (OFMSW) such as FW. Ma et al. (2011) obtained a 24% in-crease of biomethane production from FW with a thermal pre-treatment at 120 �C, whereas Liu et al. (2012) obtained a 7.9%decrease of the biomethane production from FW with thermalpretreatment at 170 �C. Cesaro and Belgiorno (2013) obtained anegligible increase with ozonation pretreatment of source-separated OFMSW (SS-OFMSW).

To the best of our knowledge, no study has been conducted onthe comparison of thermal and ozonation pretreatment to enhancethe AD of FW. Therefore, this research aims at investigating theeffects of thermal and ozonation pretreatments. A series of batchbiomethane potential (BMP) tests were conducted to investigatethe effect of temperature and treatment time of thermal andozonation pretreatments. Moreover, the net energy productionfrom applying these pretreatment methods, which could be usedfor a generation of electricity and heat, was estimated.

2. Materials and methods

2.1. Substrate and inoculum

MSW is the most complex solid waste stream, as opposed tomore homogenous waste streams resulting from industrial oragricultural activities (Sim and Wu, 2010). The generation rate andcomposition of FW depends on many factors such as the region,season, culture, economic income and demographics. To reduceexperimental bias due to the different compositions of collectedFW, the substrate used for this researchwas synthetically generatedbased on an average compositional analysis of FW in some Euro-pean countries, including UK, Finland, Portugal and Italy (Table 1)(MTT Agrifood Research Finland, 2010).

Table 1 shows the fractions of synthetic FW used in this exper-iment as well as the results from the study on mixed FW compo-sition in selected European countries (MTT Agrifood ResearchFinland, 2010). In order to make the substrate preparationsimpler, an assumption was made to eliminate the mixed meals,drinks and snacks fraction. The calculationwasmade assuming thatthe miscellaneous fraction of FW (25.8%) contains the same 58.4%fruits/vegetables, 3.6% pasta/rice, 4.7% bread/bakery, 6.1% meat/fish, 1.4% dairy products ratio, thus resulting in the additional dis-tribution of the miscellaneous fraction over these known fractions.Based on the final concentration of the FW composition shown in

Table 1, different types of uncooked food were mixed and blendedin order to obtain a homogenized synthetic FW that represents thetypical FW of the above-mentioned EU countries.

2.2. Pretreatment of FW

EU Regulation EC1774/2002 dictates that catering waste shouldbe pasteurized at >70 �C for at least an hour, or at >133 �C for20e30 min. With respect to this regulation, pretreatment at70e140 �C for an hour and at 140e150 �C for 30minwas conductedto investigate their potential to enhance the AD of FW. Moreover, aset of experiments was subsequently conducted if a longer pre-treatment time could result in a further enhancement of the bio-methane production. Pretreatment times of 1.5, 4 and 8 h wereinvestigated at the selected temperature.

A simple oven (WTC Binder) was used for the thermal pre-treatment. The FW was directly put in a 1 L glass bottle GL 45(Schott Duran), and then placed inside the oven. After the pre-treatment, the bottle was cooled until room temperature and it wasdirectly used for the BMP tests.

There are no regulations for ozonation pretreatment of FW priorto AD. An UV generator (model-Fischer) was used for the ozonationpretreatment. It produces 0.6 mmol O3 with a flow rate of 35 L/hourusing ambient air. The FW was placed in a vessel with inlet andoutlet tubes. The ozone was introduced from the bottom for10e60 min, and forced to flow out from the top, which generated0.168e1.008 gO3. Four concentrations (0.034 gO3/gTS, 0.068 gO3/gTS, 0.101 gO3/gTS, 0.202 gO3/gTS) of ozone doses were applied atroom temperature prior to the BMP test. To reduce the potentialozone inhibition that can have an immediate killing effect onanaerobic microbes, the vessel was flushed with nitrogen gas afterozonation.

2.3. Biomethane potential test

As there is no standard protocol for BMP tests (Raposo et al.,2011; Esposito et al., 2012a), the most common reported methodwas applied (Raposo et al., 2011; Esposito et al., 2012a, 2012b,2012c; Browne and Murphy, 2013). BMP tests were conducted ina 1 L glass bottle at mesophilic (32e34 �C) conditions. All thebottles were in duplicates and were placed on a magnetic stirrer(model-VELP) to provide continuous mixing. The substrate toinoculum (S/I) ratio was 0.5 gVS/gVS. The inoculum used for theBMP tests was from a full-scale AD plant located in Capaccio-Salerno (Italy). The plant treats the buffalo dung together withthe milk whey and sewage sludge generated from the mozzarellaproducing industry. The expected microbial consortia responsiblefor the AD process would be the typical methanogens mostcommonly found in rumen, i.e. Methanobrevibacter,

J. Ariunbaatar et al. / Journal of Environmental Management 146 (2014) 142e149144

Methanomicrobium, Methanobacterium, and Methanosarcina(Daquiado et al., 2014).

Biomethane was measured once a day by a volumetric methodas described by Esposito et al. (2012c) (Browne and Murphy, 2013).Each BMP test bottle was connected to an inverted 1 L glass bottlecontaining an alkaline solution (120 gNaOH/L) to absorb the carbondioxide. The cumulative biomethane production (CBP) wasnormalized to standard temperature and pressure (STP).

2.4. Analytical methods

Total Solids (TS), Volatile Solids (VS) and Total Kjeldahl Nitrogen(TKN) of both the synthetic FW and the inoculum were analysedaccording to the APHA standard methods (APHA). Total proteinswere calculated based on TKN, using a correction coefficient of 6.25(Codex Guidelines on Nutrition Labelling CAC/GL 2-1985 (Rev.1 e,1993). Total carbohydrates were determined with the phenol-sulphuric method and measured spectrophotometrically (TUVSR03210002) using glucose as standard solution (Codex Guidelineson Nutrition Labelling CAC/GL 2-1985 (Rev.1e, 1993)). Total lipidswere extractedwith amixture of chloroformandmethanol (1:2 byv/v), dried and weighted (Phillips et al., 1997).

2.5. Net energy production

The net energy production was calculated based on the extraenergy produced (E Produced) and the required energy for operatingthe pretreatments. The extra energy from the enhanced bio-methane production can be calculated as follows (Ma et al., 2011):

EProduced ¼ EBiomethane*VBiomethane*h (1)

where:EBiomethane ¼ energy content of biomethane (6.5 kWh/m3);VBiomethane ¼ extra biomethane produced due to pretreatment

(m3);h ¼ conversion factor (0.85 for thermal energy);The total required energy for the thermal pretreatment is the

sum of the required energy (EThermal) to obtain the desired pre-treatment temperature and the energy of the pretreatmentchamber (EChamber) to maintain the heat (Ma et al., 2011):

EThermal ¼ CFW*MFW* DT þ CWater*MWater*DT (2)

where:CFW ¼ heat capacity of dry food waste (1.92 kJ kg�1 �C�1);MFW ¼ dry mass of food waste and/or TS (kg/ton FW);CWater ¼ heat capacity of water (4.18 kJ kg�1 �C�1);MWater ¼ mass of water in FW (kg/ton FW);DT ¼ temperature increase from room temperature to desired

temperature (�C)

and EChamber ¼ DT*A*ðk=sÞ*t (3)

where:A ¼ total surface area of the pretreatment chamber (m2);

s ¼ thickness of the pretreatment chamber wall (m); k ¼ heatconductivity of material used of pretreatment chamber (W/m, �C);t ¼ pretreatment time (hours).

The density of FW ranges between 0.3 and 1 ton/m3 dependingon its characteristics and compaction (Tchobanoglous et al., 1993).For simplicity, 1 ton/m3 was considered for this research. Hence, asmall pretreatment chamber with 1.1 m-height and 0.55 m-radiuswidth, made of polyurethane (k ¼ 0.022 W/m, �C) was consideredfor the thermal pretreatment of 1 ton FW. Since EChamber depends

on the outdoor temperature, various scenarios of ambient airtemperature (�10 to 20 �C) were considered.

The total energy required forozonationdepends on theozonationmethod and the characteristics of the ozonator. Ozone generationfrom air with the lowest energy efficiency of 2e3% requires 40 kWh/kgO3 energy, whereas a high-energy efficiency of 30% requires2.5 kWh/kgO3 energy (Fridman, 2008). The average (21.3 kWh/kgO3)of reported values was used to estimate the required energy forozonation pretreatment. The calculation of the net energy produc-tion could not be compared with any other research, as so far noliterature was found specifically referring to FW.

3. Results

3.1. Characteristics of substrate and inoculum

The results of the chemical and physical characterization of boththe synthetic FW and the inoculum are shown in Table 2. Thesynthetic FW contains a high percentage (76.5 ± 0.7% VS) of car-bohydrates, making it a suitable substrate for the AD process (Neveset al., 2008). Values shown in Table 2 are the averages of the threesets of experiments and standard deviations are calculated basedon the values of triplicate experiments of each set. The inoculumcontains a higher amount of protein and lipids than carbohydrates.This suggests that the TS are mainly contained in the microbialbiomass and very little FW substrate is available in the inoculum.

3.2. Cumulative biomethane production

3.2.1. Thermal pretreatment: effect of pretreatment temperatureThe first set of experiments was conducted to investigate the

effect of temperature (70e140 �C) to pretreat FW for an hour.Biomethane production of FW reached its maximum amount after154 days, though the experiment was kept running for another2 months to make sure the maximumwas attained. The CBP curvesare shown in Fig. 1A.

FW pretreated with the thermal method produced more bio-methane than the untreated FW (Fig. 1). The CBP of pretreated FWwas enhanced by 22.2 ± 1.3, 18.9 ± 4.1, 9.9 ± 0.6, 7.5 ± 0.9, 3.8 ± 1.2%at pretreatment temperatures of 80, 100, 70, 120 and 140 �C,respectively.

The next set of BMP tests was carried out with FW pretreated at140e150 �C for 30 min (Fig. 1B). FW pretreated at higher temper-atures produced less methane than the untreated FW during theinitial 16e18 days. At the end of the experiment, the CBP of pre-treated substrates were nevertheless increased by 6.9 ± 0.3 and4.5 ± 0.8% at 140 and 150 �C, respectively. After the thermal pre-treatment at 120, 140 and 150 �C for both 1 h and for 30 min at 140and 150 �C, the substrate turned brown.

The effect of the thermal pretreatment on the AD process isparticularly clear when comparing the specific biomethane pro-duction (SBP) of the initial 20 days of biomethanation (Fig. 2).Most ofthe organic matter (80e85%) is converted into biomethane in theinitial 20 days. Fig. 2 shows that all the thermally pretreated FWsubstrates have a higher SBP than the untreated FW(426.0 ± 8.5 mlCH4/gVS). The highest SBP of 539.8 ± 8.7 mlCH4/gVSwas achievedwith a pretreatment at 80 �C, followed by 516.1± 7.1 at100 �C, 492.1 ± 16.3 at 120 �C and 479.3 ± 7.9 at 70 �C. The energyrequirement for a thermal pretreatment higher than 100 �C ismostlyutilized for evaporating thewater, thus high temperatures (>100 �C)were not suitable for the pretreatment of FW due to a higher energyrequirement and lower enhancement of the SBP. The BMP tests onthe effect of treatment time were carried out with temperatures at70 and 80 �C. Although for comparison reason, the net energy pro-duction was estimated for 120 �C.

Table 2Characterization of FW and inoculum used in this experiment.

TS (%) VS (%) VS/TS (%) TKN (mg/L) Protein (%VS) Lipid (%VS) Carbohydrates (% VS)

FW 22.2 ± 0.2 21.1 ± 0.2 89.9 ± 1.9 4.7 ± 0.6 14.3 ± 1.8 9.2 ± 1.1 76.5 ± 0.7Inoculum 2.7 ± 0.2 1.5 ± 0.1 57.0 ± 1.8 0.8 ± 0.1 59.3 ± 5.2 38.7 ± 5.3 2.1 ± 0.1

Fig. 1. CBP curves of FW pretreated at various temperatures for (A) 1 h and (B) 30 min.

J. Ariunbaatar et al. / Journal of Environmental Management 146 (2014) 142e149 145

3.2.2. Thermal pretreatment: effect of pretreatment timeFig. 3A and B show that all pretreatment conditions applied

resulted in a higher CBP when compared to the production of un-treated FW. As shown in Fig. 4A and B the highest SBP achieved waswith 1.5 h of pretreatment and amounted to 647.5 ± 10.6 and

Fig. 2. Effect of thermal pretreatment on the specific biomethane production duringthe initial 20 days of the BMP test.

510.6 ± 11.9 mlCH4/gVS at 80 and 70 �C, respectively. It is inter-esting to note that after 14 days of biomethanation, the substratetreated at 80 �C for 1.5 h showed a sudden increase in biomethaneproduction, making up an additional increase to the CBP curve.Longer pretreatment times of 4 and 8 h resulted in a higher SBP ascompared to the untreated FW, though the accumulated increase isless when compared to the SBP of 1 h pretreated FW at the sametemperature. It is worthwhile to note that the FW pretreated at70 and 80 �C for 4 and 8 h turned light brownish.

3.2.3. Ozonation pretreatmentCMP curves of the untreated and ozonated FW are shown in

Fig. 5. The net methane yield of untreated FW was440.3 ± 2.6 mlCH4/gVS, which is consistent with the first set ofexperiments and comparable with previous research. The BMPtests were kept on running for almost 220 days until the bio-methanation was ceased. All ozonated FW produced less bio-methane as compared to the untreated substrate during the initial15 days. However, thereafter all the ozonated FW started producinghigher amounts of biomethane than the untreated FW. At the endof the experiment, ozonation pretreatment resulted in 35.2 ± 1.5,46.4 ± 2.8, 32.9 ± 1.8, 22.2 ± 1.3% higher CMP at ozone doses of0.034 gO3/gTS, 0.068 gO3/gTS, 0.101 gO3/gTS, 0.202 gO3/gTS,respectively. Similar to the thermal pretreatment at 80 �C for 1.5 h,the ozonation pretreatment also caused an additional increase inthe CBP curves (Fig. 5A) after 18 and 36 days of biomethanation.

The SBP of the 20 days biomethanation (Fig. 5B) shows that thenet SBP of the untreated substrate was 420.9 ± 9.5 mlCH4/gVS,which is consistent with the results from the first set (Fig. 2). Thehighest SBP of 9.2 ± 0.7% was achieved with an ozone dose of0.068 gO3/gTS, followed by an increase of 7.8 ± 0.1%with 0.034 gO3/gTS. Therefore, the required energy estimation was carried out forthese conditions.

3.3. Net energy production

On the basis of the BMP experimental results, thermal pre-treatment at 80 �C for 1.5 h, gave the highest enhancement of SBP.Its net energy production was calculated for different scenarios(Fig. 6). Each scenario resulted in positive net energy production(Fig. 6); thus the extra biomethane produced due to pretreatment issufficient to generate the energy to apply the pretreatment. Incontrast, the pretreatment at 120 �C resulted in a negative netenergy production (Fig. 6), suggesting the higher temperatures arenot suitable for pretreating FW. Table 3 shows the calculation of thenet energy production for ozonation pretreatment. Each conditionresulted in a negative energy balance, which means the requiredenergy for ozonation pretreatment exceeds the energy that can begenerated from the extra biomethane produced.

4. Discussion

4.1. Effects of pretreatment methods

The CBP curves (Figs. 1A, 1B, 3A, 3B, 5A) of the untreated FWsuggest a typical AD of a substrate rich in carbohydrates (Neveset al., 2008; Appels et al., 2010), which agrees with the chemical

Fig. 3. CBP curves of FW pretreated at (A) 70 �C; and (B) 80 �C for various treatmenttimes.

Fig. 4. Effect of (A) 70 �C; and (B) 80 �C thermal treatment time on the biomethaneproduction during the initial 20 days of the BMP test.

J. Ariunbaatar et al. / Journal of Environmental Management 146 (2014) 142e149146

analyses of the FW (Table 2). AD of lipids and proteins are relativelyslow as compared to carbohydrates (Vavillin et al., 2008), andBreure et al. (1986) suggested that a complete degradation ofproteins cannot be achieved in the presence of high carbohydrateconcentrations (Breure et al., 1986). Hence, the entire potentialbiomethane source cannot be recovered from a normal unstimu-lated biomethanation of complex substrates (such as FW), whichcontains both easily biodegradable (carbohydrates) and recalcitrantorganic matter (lipids and proteins). This study, however, showedthat pretreatment of FW with thermal and ozonation methodsprior to AD can enhance the CBP (Fig. 1A, B, 3A, B, 5 A). The resultssuggest that the recalcitrant organic matter was degraded to lesscomplex substrates that are easily available for the anaerobic mi-crobes. In this regard, focussing only on the favourable C/N ratio,which is reported to be in the range of 14.7e36.4 (Zhang et al.,2007; Esposito et al., 2011a) for the AD of FW is not suitable, asFW contains considerable amounts of recalcitrant complex sub-stances. Moreover, thermal and ozonation pretreatments disinfectthe substrates, which contribute to a hospitable environment forthe methanogenic consortia in the anaerobic digesters. Conse-quently, the more specialised microbial community could convertmore organic matter to biomethane. Nevertheless, the effects ofpretreament methods were different depending on the conditionsapplied.

4.1.1. Effect of thermal pretreatmentThermal pretreatment resulted in an enhanced CBP at all the

tested conditions (Figs. 2, 3 and 5), which agrees with the previous

research (Mata-Alvarez et al., 2000; Ariunbaatar et al., 2014; Neveset al., 2008; Eskicioglu et al., 2006). These results indicate that thethermal pretreatment caused a deflocculation of macromolecules(Eskicioglu et al., 2006; Protot et al., 2011), which increases thesurface area of the substrates as proposed by previous research.Esposito et al. (2011b) confirmed that the increased surface arearesults in a better contact between the substrate and the microbialpopulation, thus more organic matter is converted into biomethane(Esposito et al., 2011b).

In addition to the well-known enhancement of the CBP, thisstudy showed the various effects of pretreatment temperature andtime that was not very well explained specifically for FW by pre-vious research. The effects of temperature and treatment time onthe CBP and SBP were not linear, but parabolic (Figs. 2 and 4). Thissuggests that the thermal pretreatment also caused the degrada-tion of complex substances and/or increased the soluble organicmatter (Valo et al., 2004), resulting in the Maillard reaction, i.e. areaction between amino acids and sugars. The product from theMaillard reaction, melanoidins, is difficult to degrade anaerobically(Carrere et al., 2010; Vavillin et al., 2008). Depending on the type ofcarbohydrates and proteins in the substrates, the temperaturerange to cause Maillard reaction differs, though the colour devel-opment is an important indication of the reaction (Ma et al., 2011;Appels et al., 2010). The FW pretreated at higher (>120 �C) tem-peratures indeed turned brownish. Liu et al. (2012) obtained asimilar conclusionwith a study on the thermal pre-treatment of FWand fruit and vegetablewaste at 175 �C, which resulted in a 7.9% and11.7% decrease of the CBP, respectively, due to the formation ofmelanoidins (Liu et al., 2012). Moreover, the FW pretreated at lower(70 and 80 �C) temperatures for longer times (4 and 8 h) turnedlight brownish, suggesting an incomplete or mild Maillard reaction

Fig. 5. Effect of ozone on A) CMP curves; and B) SBP of FW during the initial 20 days.

Table 3Cost benefit analysis of ozonation pretreatment.

0.034 gO3/gTS 0.068 gO3/gTS

Extra biomethane (V, m3/ton FW) 7.4 ± 0.1 8.7 ± 0.9Extra energy (kWh/ton FW) 40.8 ± 0.5 48.2 ± 4.8Required ozone (kgO3/ton FW) 7.55 15.10Required extra energy (kWh/ton FW) 160.77 321.54Net energy production (kWh/ton FW) �120.0 ± 0.5 �273.4 ± 4.5

J. Ariunbaatar et al. / Journal of Environmental Management 146 (2014) 142e149 147

had occurred. Bougrier et al. (2008) proposed that the thermalpretreatment could also cause a reaction between the soluble car-bohydrates and soluble proteins, forming amadori like compounds(Protot et al., 2011). These amadori compounds are the by-productsof melanoidins (Appels et al., 2010; Vavillin et al., 2008; Breureet al., 1986; Eskicioglu et al., 2006; Protot et al., 2011); and theformation of such compounds might have also yielded a lowerenhancement of the SBP at these pretreatment conditions.

Further to the Maillard reaction, which is a confirmation ofincreased degradation of proteins and carbohydrates, degradationof lipid compounds was also induced by the thermal pretreatment.

Fig. 6. Effect of ambient air temperature on the net energy production from thermalpretreatment at 80 �C for 1.5 h.

As suggested by Cirne et al. (2007), the major obstacle of bio-methane production from lipid compounds are the long chain fattyacids (LCFA), which yields a long (6e10 days) lag phase (Cirne et al.,2007). However, this inhibition due to LCFA is not permanent and ittakes time for the LCFA consuming anaerobic microbes to grow.Therefore, when organic substrates contain lipid compounds, theCBP curves usually illustrate a sudden increase. Figs. 1A, 1B, 3A, 3Band 6 exhibited such a sudden increase in the CBP curves afterapproximately 2 weeks.

Besides the melanoidins and the LCFA inhibition due toincreased degradation of the organic matter, the lower biomethaneproduction of the thermally pretreated FW during the initial days(Figs. 1 and 3) can be explained by volatilization of short chainorganics, which are a potential biomethane source (Eskicioglu et al.,2006). Since FW contains also the easily biodegradable, highlyvolatile carbohydrates, higher pretreatment temperatures(>140 �C) and longer treatment times (>4 h) can result in a loss ofthese fermentable sugars. Therefore, to obtain the highest amountof potential biomethane production from FW and to prevent apossible inhibition as well as a loss of potential biomethane, it isimportant to have a balance between the degradation of carbohy-drate, lipid and protein substrates (Vavillin et al., 2008; Breure et al.,1986).

4.1.2. Effect of ozonation pretreatmentOzonation pretreatment yielded 22e46% enhancement of CBP

(Fig. 5A), which is comparable with the previous results by Cesaroand Belgiorno (2013), who reported a 37% increase of CMP fromozonated source-separated OFMSW. Even though the CBPenhancement is comparable, the ozone dose for such enhancementis much lower (0.068 gO3/gTS as compared to 0.16 gO3/gTS) in thisresearch, and the CBP curves illustrate different trends. Fig. 6 showsthat all the ozonated FW produced less biomethane as compared tountreated FW during the initial 18 days.

Ozone is a strong oxidant, which decomposes itself into radicalsthat react with organic substrates in two ways: directly and indi-rectly (Carballa et al., 2007). The direct reaction based on the rad-icals of ozone can destroy the easily fermentable sugar, thusresulting in a loss of biomethane production. This effect is com-parable with the more extreme thermal pretreatment conditions,e.g. higher temperatures and longer treatment times (Section 4.1.1).The indirect reaction of ozone, which depends on the hydroxyl ion,causes the degradation of complex organic compounds such aslipids and proteins in FW, thus yielding a sudden increase in thebiomethane production (Fig. 4A). However, a previous study on theAD of ozonated SS-OFMSW produced a higher biomethane yieldfrom the beginning of the AD process (Cesaro and Belgiorno, 2013),probably the SS-OFMSW used for their experiment contained ahigher level of lipids and proteins. Unfortunately, the authors didnot analyse the chemical content of the substrate. Based on theresults obtained in this study (Table 3), ozonation was found to bean inefficient method to enhance the AD of FW. Even thoughozonation resulted in a higher CBP at all concentrations, consid-ering the initial 20 days of the AD process a high ozone dose of0.202 gO3/gTS found to be an inhibitory condition. It can be

J. Ariunbaatar et al. / Journal of Environmental Management 146 (2014) 142e149148

explained by a higher loss of fermentable sugars at a higher con-centration of ozone, as FW contains mostly carbohydrates (Table 2).Ozonation could be an attractive method for a substrate with a highcontent of more complex and recalcitrant organics.

4.2. Net energy production

On the basis of the net energy estimation, the enhanced bio-methane production could cover the required energy for the ther-mal pretreatment at 80 �C for 1.5 h, regardless of the ambient airtemperature; whereas the pretreatment at 120 �C gave a negativenet energy production (Fig. 6). Due to lack of existing literature onthe subject, no comparison on the net energy production by othersubstrates or systems could be carried out. In order to comparewith previous results, which reported a profit of 8.5e9.1 V/tonFWat 120 �C (Ma et al., 2011), the net energy productionwas convertedto a net profit when considering a thermal energy cost of 0.07 V/kWh (Ma et al., 2011). The net energy produced after thermalpretreatment could yield a profit of 7.65e13.45 V/tonFW at 80 �Cfor 1.5 h, depending on the ambient air temperature of the plantlocation; whereas a pretreatment at 120 �C could have a profit of0.41 V/tonFW only if the ambient temperature is 20 �C or higher.However, this research considered not only the required energy toreach the desired temperature, but also the energy to maintain theheat, with respect to the ambient air temperature, resulting in alower profit as compared to other research.

5. Conclusions

This research investigated the thermal and ozonation pretreat-ment methods to enhance the biomethanation of a synthetic FW,which was prepared mimicking a typical FW in selected Europeancountries. Based on a series of batch experiments, a thermal pre-treatment at 80 �C for 1.5 h yielded the highest enhancement (52%),amounting to 647.5 ± 10.6 mlCH4/gVS. The enhanced biomethaneproduction was enough to supply the required energy for thethermal pretreatment. Thermal pretreatment at a higher temper-atures (>120 �C) and a longer time (>4 h) caused the formation ofmore complex substrates, such as melanoidins, which are difficultfor anaerobes to digest. Pretreatment with a high dose of ozone(0.034e0.202 gO3/gTS) resulted in a loss of fermentable sugars.Therefore, such aggressive pretreatment methods found out to beineffective for the enhancement of AD treating FW.

Acknowledgements

This research was financially supported by Erasmus MundusJoint Doctoral Program on ETECOS3 (Environmental Technologiesfor Contaminated Solids, Soils and Sediments) under the EU grantagreement FPA n� 2010-0009 as well as the project on “Integratedsystem to treat buffalo slurry, aimed to recover water and safeenergy e STABULUM”, which is funded by the Decision of the Eu-ropean Commission No.C (2010) 1261 in agreement with theAgriculture Department of the Campania Region in the context ofthe Programme of Rural Development 2007e2013, Measure 124Cooperation for development of new products, processes andtechnologies in the agriculture and food sectors. Moreover, theauthors would like to thank Prof. Andreozzi, Robert and theDepartment of Chemical Engineering, University of Naples, Feder-ico II for their help in conducting the ozonation pretreatment.

References

Adhikari, B.K., Barrington, S., 2006. Predicted growth of world urban food waste andmethane production. Waste Manage. Res. 24, 421e433.

APHA, Standard Methods for the Examination of Water and Wastewater twenty-first ed. ISBN: 0875530478.

Appels, L., Degreve, J., Van der Bruggen, B., Van Impe, J., Dewil, R., 2010. Influence oflow temperature thermal pre-treatment on sludge solubilization, heavy metalrelease and anaerobic digestion. Bioresour. Technol. 101, 5743e5748.

Ariunbaatar, J., Panico, A., Esposito, G., Pirozzi, F., Lens, P.N.L., 2014. Pretreatmentmethods to enhance anaerobic digestion of organic solid waste. Appl. Energy15, 143e156.

Bougrier, C., Delgenes, J.P., Carrere, H., 2008. Effects of thermal treatments on fivedifferent waste activated sludge samples solubilisation, physical properties andanaerobic digestion. Chem. Eng. J. 139, 236e244.

Breure, A.M., Mooijman, K.A., van Andel, J.G., 1986. Protein degradation in anaerobicdigestion: influence of volatile fatty acids and carbohydrates on hydrolysis andacidogenic fermentation of gelatin. Appl. Microbiol. Biotechnol. 24, 426e431.

Browne, J.D., Murphy, J.D., 2013. Assessment of the resource associated with bio-methane from food waste. Appl. Energy 104, 170e177.

Carballa, M., Manterola, G., Larrea, L., Ternes, T., Omil, F., Lema, J.M., 2007. Influenceof ozone pretreatment on sludge anaerobic digestion: removal of pharmaceu-tical and personal care products. Chemosphere 67, 1444e1452.

Carlsson, M., Lagerkvist, A., Morgan-Sagastume, F., 2012. The effects of substratepretreatment on anaerobic digestion: a review. Waste Manage. 32, 1634e1650.

Carrere, H., Dumas, C., Battimeli, A., Batstone, D.J., Delgenes, J.P., Steyer, J.P., Ferrer, I.,2010. Pretreatment methods to improve sludge anaerobic degradability: a re-view. Hazard. Mater. 183, 1e15.

Cesaro, A., Belgiorno, V., 2013. Sonolysis and ozonation as pretreatment foranaerobic digestion of solid organic waste. Ultrason. Sonochem. 20, 931e936.

Cirne, D.G., Paloumet, X., Bjornsson, L., Alves, M.M., Mattiasson, B., 2007. Anaerobicdigestion of lipid-rich waste e effects of lipid concentration. Renew. Energy 32,965e975.

Codex Guidelines on Nutrition Labelling CAC/GL 2-1985 (Rev1 e 1993).Curry, N., Pillay, P., 2012. Biomethane prediction and design of a food waste to

energy system for the urban environment. Renew. Energy 41, 200e209.Daquiado, A.R., Cho, K.M., Kim, T.Y., Kim, S.C., Chang, H.H., Lee, Y.B., 2014. Meth-

anogenic archaea diversity in Hanwoo (Bos Taurus coreanae) rumen fluid, rectaldung, and barn floor manure using a culture-independent method based onmcrA gene sequences. Anaerobe 27, 77e81.

Eskicioglu, C., Kennedy, K.J., Droste, R.L., 2006. Characterization of soluble organicmatter of waste activated sludge before and after thermal pretreatment. WaterRes. 40, 3725e3736.

Esposito, G., Frunzo, L., Panico, A., Pirozzi, F., 2011a. Model calibration and validationfor OFMSW and sewage sludge co-digestion reactors. Waste Manag. 31,2527e2535.

Esposito, G., Frunzo, L., Panico, A., Pirozzi, F., 2011b. Modelling the effect of the OLRand OFMSW particle size on the performances of an anaerobic co-digestionreactor. Process Biochem. 46, 557e565.

Esposito, G., Frunzo, L., Panico, A., Pirozzi, F., 2012a. Enhanced bio-methane pro-duction by co-digestion of different organic wastes. Environ. Technol. 33,2733e2740.

Esposito, G., Frunzo, L., Giordano, A., Liotta, F., Pirozzi, F., 2012b. Anaerobic co-digestion of organic wastes. Rev. Environ. Sci. Biotechnol. 11 (4), 325e341.http://dx.doi.org/10.1007/s11157-012-9277-8.

Esposito, G., Frunzo, L., Liotta, F., Panico, A., Pirozzi, F., 2012c. Bio-methane potentialtests to measure the biomethane production from the digestion and co-digestion of complex organic substrates. Open. Environ. Eng. J. 5, 1e8.

FAO, 2011. Swedish Institute for Food and Biotechnology, Global Food Losses andFood Waste e Study Conducted for the International Congress.

Fridman, A., 2008. Plasma Chemistry. Cambridge University Press, 978-0-521-84735-3-Frontmatter.

Kirchmayr, R., Scherzer, R., Baggesen, D.L., Braun, R., Wellinger, A., 2003. Animal By-products and Anaerobic Digestion, Requirements of the European Regulation(EC) No. 1774/2002., IEA Bioenergy Task 37, Energy from Biomethane andLandfill Gas in Cooperation with Bioexell e Biomethane Center of Excellence.

Liu, X., Wang, W., Gao, X., Zhou, Y., Shen, R., 2012. Effect of thermal pretreatment onthe physical and chemical properties of municipal biomass waste. WasteManag. 32, 249e255.

Ma, J., Duong, T.H., Smits, M., Vestraete, W., Carballa, M., 2011. Enhanced bio-methanation of kitchen waste by different pretreatments. Bioresour. Technol.102, 592e599.

Mata-Alvarez, J., Mace, S., Llabres, P., 2000. Anaerobic digestion of organic solidwastes. An overview of research achievements and perspectives. Bioresour.Technol. 74, 3e16.

MTT Agrifood Research Finland, 2010. Valorisation of Food Waste to Biomethane,Deliverable Report to 7th Framework Programme., Theme Energy.

Mudhoo, A., Kumar, S., 2013. Effect of heavy metals as stress factors on anaerobicdigestion processes and biogas production from biomass. Int. J. Environ. Sci.Technol. 10 (6), 1383e1398.

Neves, L., Goncalo, E., Oliveira, R., Alves, M.M., 2008. Influence of composition onthe biomethanation potential of restaurant waste at mesophilic temperatures.Waste Manag. 28, 965e972.

Phillips, K.M., Tarrag�o-Trani, M.T., Grove, T.M., Grün, I., Lugogo, R., Harris, R.F.,Stewart, K.K., 1997. Simplified gravimetric determination of total fat in food com-posites after chloroformemethanol extraction. J. Am. Oil Chem. 74 (2), 137e142.

Protot, A., Julien, L., Christophe, D., Partick, L., 2011. Sludge disintegration duringheat treatment at low temperature: a better understanding of involvedmechanisms with a multi-parametric approach. Biochem. Eng. J. 54, 178e184.

J. Ariunbaatar et al. / Journal of Environmental Management 146 (2014) 142e149 149

Raposo, F., De la Rubia, M.A., Fernandez-Cegri, V., Borja, R., 2011. Anaerobic diges-tion of solid organic substrates in batch mode: an overview relating to methaneyields and experimental procedures. Renew. Sustain. Energy Rev. 16, 861e877.

Sim, E.Y.S., Wu, T.Y., 2010. The potential reuse of biodegradable municipal solidwaste (MSW) as feedstocks in vermicomposting. J. Sci. Food Agric. 90 (13),2153e2162.

Tchobanoglous, G., Theisen, H., Vigil, S.A., 1993. Integrated Solid Waste Manage-ment. ISBN 0071128654, 9780071128650.

Valo, A., Carrere, H., Delgene, J.P., 2004. Thermal, chemical and thermo-chemicalpre-treatment of waste activated sludge for anaerobic digestion. Chem. Tech-nol. Biotechnol. 79, 1197e1203. http://dx.doi.org/10.1002/jctb.1106.

Vavillin, V.A., Fernandez, B., Palatsi, J., Flotats, X., 2008. Hydrolysis kinetics inanaerobic degradation or particulate organic material: an overview. WasteManag. 28, 939e951.

Zaman, A.U., 2013. Identification of waste management development drivers andpotential emerging wast treatment technologies. Int. J. Environ. Sci. Technol. 10(3), 455e464.

Zhang, R., El-Mashad, H.M., Hartman, K., Liu, G., Choate, C., Gamble, P., 2007.Characterization of food waste as feedstock for anaerobic digestion. Bioresour.Technol. 98, 929e935.