A laboratory simulation of in situ leachate treatment in semi-aerobic bioreactor landfill

In this study, two laboratory-scale simulated landfill bioreactors were established, of which Reactor A was operated only with leachate recirculation and served as the control, and Reactor B was operated as semi-aerobic bioreactor landfill with leachate recirculation. In situ leachate treatment and accelerating organic decomposition in semi-aerobic bioreactor landfill was investigated. The results indicated that the introduction of air into the landfill was favourable for optimising the micro-organism growth environment and accelerating the degradation o� organic matter. It can be seen clearly �rom the results that NH� +-N can be removed in situ in the semi-aerobic bioreactor landfill with leachate recirculation. Moreover, semi-aerobic bioreactor landfill showed lower emissions for leachate than those in leachate from anaerobic landfill, with low concentrations of COD, VFA, NH� +-N and TKN, and which saved the disposing process o� the discharged leachate. The three-dimensional e�citationemission matrix fluorescence spectroscopy (EEMs) of dissolved organic matter (DOM) in Reactor B changed greatly, and fluorescence peak changed from protein-like fluorescence at Day 60 to humic-like fluorescence at Day 95 and 250, while in Reactor A, fluorescence peak of DOM was always protein-like fluorescence. The comparison of the EEMs indicated that the semi-aerobic landfill accelerated the organic decomposition.


Introduction
In 2��2, appro�imately ���% o� the municipal solid waste (MSW) generated in China was disposed of in landfills (Wang et al., 2006).However, anaerobic degradation of solid waste results in the production of leachate and landfill gas for a very long time in a conventional landfill.Leachate and landfill gas are the potential pollution sources �or the surrounding environment, and the long-term environmental impacts will last for several decades (Cossu et al., 2003;Bilgili et al., 2006).There�ore, there has been increased emphasis on the operation of landfills as bioreactors to enhance decomposition of solid waste, provide a reduction in landfill emissions over a relatively short time, and dispose leachate in situ (Price et al, 2003;Mehta et al., 2002;Pohland et al., 2000;Reinhart et al., 2002;Reinhart, 1996;Pohland et al., 1994;Townsend et al., 1996;Chan et al., 2002;Demir et al., 2004;Bilgili et al., 2004;Mark and Cristina;2006).However, ammonia nitrogen typically accumulates because ammonia is stable under anaerobic conditions.Thus, higher concentrations o� ammonia than those �ound in leachate from conventional landfills last long even after the organic fraction of the waste is stabilised (Burton and Watson-Craik, 1998;Onay and Pohland, 1998;Price et al., 2003).That is the reason why ammonia removal is an important aspect o� longterm landfill pollution control.
However, ex situ leachate treatment o� high concentration COD and NH � -N can be costly (Ferhan and Aktas 2000; Maree et al., 2004).Recently, in situ biological nitrogen removal for bioreactor landfills attracted more and more attention of the researchers.Air addition has recently been practised at a number of pilot-scale and field-scale landfills worldwide, where it was �ound that the organic �raction o� the waste decomposed �ar more rapidly under aerobic conditions than under anaerobic conditions, and ammonia was removed in situ by nitrification and denitrification in aerobic landfills (Berge et al., 2006;Onay et al., 2001;Reinhart et al., 2002;Read et al., 2001;Das et al., 2002;Themelis et al., 2001;Borglin et al., 2004;Boni et al., 1997).However, there are some disadvantages for the aerobic bioreactor landfills, such as needing �orced ventilation systems, comple� operation and management, and large energy consumption.Semi-aerobic landfills use natural ventilation instead o� mechanical ventilation �or o�ygen supply, and which create an aerobic region in the landfill.Semi-aerobic landfill system is propitious to the simultaneously occurring nitrification and denitrification and thus in situ disposes the leachate e��ectively and accelerates the stabilisation of the waste (Theng et al., 2005).In China, only a few researchers studied the semi-aerobic landfill technology, but no systematic and comprehensive studies have been conducted (Wang et al., 2006).
Moreover, fluorescence excitation-emission matrix spectroscopy (EEMs) provides much more detailed information about fluorescence properties of the organic matter that may reveal important in�ormation about its composition and biogeochemical cycling (Burdige, et al., 2004).Fluorescence spectroscopy has high sensitivity and specificity, and has already been used in natural water quality monitoring.Rapid analyses are now possible using fluorescence spectrophotometers such that the EEMs 134 can be generated in approximately 1min (Baker, 2002;Baker and Curry, 2004).Thus, fluorescence analyses may be providing a rapid method to characterise organic matter and to fingerprint organic pollutants in the leachate.
The main objective o� the research was to investigate in situ leachate treatment and accelerating organic decomposition and stabilisation of solid waste in semi-aerobic landfill by chemical analysis and fluorescence analyses.

Experimental equipments
Two simulated reactors, made o� polymethylmethacrylate (PMMA) of 1.5 cm thickness, with a diameter of 20 cm and of 1.5 cm thickness, with a diameter of 20 cm and of 1.5 cm thickness, with a diameter of 20 cm and height 1.2 m, providing an effective volume of 31.4 ℓ, were used in this experiment.Both of the reactors were wrapped with 1cm thick polyurethane polyfoam which acted as a heat-insulating material to prevent temperature redistribution between the reactors and the surrounding environment.Prior to filling, a 10 cm thickness of gravel was placed at the bottom of each landfill reactor to retain re�use and stop small particles �rom leaching out.Then about 17.3 kg of pretreated MSW mixture was filled into each landfill reactor.Finally, the waste mixture was covered with a �� cm depth o� pea gravel and a water distributor was installed at the upper part o� every reactor.Three sample collection ports o� solid wastes were set in the middle o� the reactors�� the leachate outlets, which were connected to the leachate recirculation tanks, were under the reactors.Besides, the vent-pipe was installed under the reactor B (Fig. 1).
The temperature in a Chinese landfill ranges from 22 to 52ºC, and is 27 to 45ºC over the first 2 yr after landfilling.The temperature of 35ºC is about the mean temperature over the temperature range noted in field (He et al., 2006).Thus, two simulated reactors were both placed in a thermostatic chamber and the temperature was kept at 33±2ºC.After filling, Reactor A was sealed with silicone caulk and monitored for leaks.Then, approximately 3 ℓ deionised water was added to the reactors in order to produce about 0.5 ℓ of leachate at the start of the experiment.

Experimental materials
Fresh waste was obtained from Asuwei Sanitary Landfill Site, located in Beijing, China.Plastic bags and large inorganic waste were removed by a first sorting step and further sorting was conducted in the laboratory.All of the waste was then shredded and mixed to avoid leachate preferential flow in simulated landfill reactors.The waste composition was as follows (by weight): Kitchen waste, 70.3±8.3%;paper, 10.2±2.1%;plastic, 8.1±1.9%;fabric, 1.2±0.6%;metal, 0.2±0.1%;and others, 1.0±0.2%.The relatively high proportion of kitchen waste or organic waste is considered to be a characteristic of MSW in China (He et al., 2005(He et al., , 2006a;;He and Shen, 2006;Wang et al., 2006), as well as several other developing countries (Sponza and Agdag, 2004;San and Onay, 2001).

Reactor operation
Reactor A served as a control; leachate was drained to a recirculation tank from Reactor A, and recycled by peristaltic pump every 2 d.Reactor B was operated similarly as Reactor A, but after Day 60, the valve of the vent-pipe was opened in order to simulate the semi-aerobic bioreactor landfill.Water was sprayed into the top of the reactors at about 300 mℓ/6d to increase the moisture content o� the waste.The daily load o� recycled leachate to reactors did not exceed 4 ℓ in order to avoid flooding of reactors.

Sampling and analytical methods
Leachate samples were collected from the recirculation tank every 2 d and analysed for chemical oxygen demand (COD), biological oxygen demand (BOD � ), pH, ammonia nitrogen (NH � + -N), NO - � -N, NO - 2 -N, Total Kjeldahl nitrogen(TKN) according to the Standard Methods (EPA of China, 1989).The pH value was measured by a Sartorius digital pH meter with PY-AS1 electrode (PB-10, Sartorius Inc., Germany).The ORP was measured by a Thermo Orion model 250A4 ORP meter with a 9179BN electrode (Thermo Orion Inc., USA).The concentration of the VFA was determined by gas chromatography (GC-6890N, Agilent Inc., USA).The gas content in biogas samples was analysed using a second gas chromatograph (HP5890, HP Inc., USA).
The leachate samples were centri�uged with a rotating speed o� �2 ��� r•min -� �or 2� min under �℃, and filtrated through glass-fibre membrane (0.45 μm) to remove the suspended matters that may react with DOM.The organic matter in the supernatant was dissolved organic matter (DOM).The total organic carbon (TOC) of the DOM was determined by TOC analyser (SHIMADZU TOC-5000).Methods described by Christensen (1998) and Ma et al. (2001) were adopted herein for dissolved organic matter (DOM) fractionation in leachate.The DOM in the leachate was fractionated into humic acid (HA), fulvic acid (FA) and hydrophilic (HyI) fractions.Amberlite XAD-8 resin was obtained from Rohm and Haas Co. and 732 cation exchange resins were obtained from Beijing Huideyi Co. Ltd. (Beijing, China).XAD-8 and 732 resins were cleaned by the method described in the literature (He et al., 2006b).The concentrations of HA, FA and HyI fractions were determined through measuring dissolved organic carbon (DOC) concentration of the fractionated sub-samples.
The EEMs was measured by fluorescence spectrometer (Perkin Elmer Luminescence Spectrometer LS50B).Each EEM was generated by scanning excitation wavelengths from 2�� to ��� nm at �� nm steps, and detecting the emitted fluorescence between 200 and 600 nm at 10 nm steps.Scan speed was � 2�� nm•min -� , permitting collection of a complete EEM in ~60 s.EEM spectra are illustrated as the elliptical shape of contours.The X-axis represents the emission spectra from 200 to 600 nm, whereas the Y-axis is the excitation wavelength from 200 to 480 nm.Contour lines, as the 3 rd dimension, are shown for each EEM spectra to represent the fluorescence intensity.

Results and discussions pH and ORP
Figure 2 shows the change o� pH over time.The pH values were in the range of 5.5 to 6.5 in the first 60 d of degradation in two reactors.After Day 60, pH values began to increase and reached 8 after Day 108 in semi-aerobic reactors.In the subsequent days, no considerable variation was observed �or pH in leachate �rom semi-aerobic reactor and pH values remained about 8. Different from semi-aerobic reactor, on Day 108, the pH was 6.6 in anaerobic reactor, and reached 7.5 after Day 180.These results indicated that the semi-aerobic reactor reaches optimal pH value much �aster than that the anaerobic reactor, showing the rapid degradation o� solid wastes in semi-aerobic condition.These results are in accordance with previous studies (Bilgili et al., 2006;Nakasaki et al., 1993;Cossu et al., 2003).
Oxidation-reduction potential (ORP) is used to monitor chemical reactions, to quanti�y ion activity, or to determine the o�idising or reducing properties o� a solution.It also provides in�ormation about the biological processes occurring under aerobic and anaerobic conditions.The ORP within a landfill deter-mines the mechanism of waste degradation.Bilgili et al. (2006) reported that high ORP (aerobic conditions) causes accelerated degradation of waste.The results of ORP are plotted in Fig. 3.In the beginning of the 60 d period, the ORP values were the same in the two reactors.After Day 60, the ORP values was increased in semi-aerobic reactor and reached 100 mV on Day 116.Oppositely, ORP values decreased below -100 mV in the anaerobic reactor a�ter ��� d, which showed that the degradation was changing �rom the acidogenic phase to the methanogenic phase a�ter the consumption o� the available o�ygen in the anaerobic reactor.Some researchers found that there is an optimum ORP requirement �or methanogenesis, which generally ranges �rom -100 to -300 mV (Bilgili et al., 2006).

Organics and gas composition
COD data for semi-aerobic and anaerobic reactors are presented in Fig. 4  5 600 and 360 mg/ℓ, respectively, and correspondingly, the ratio of BOD/COD was decreased from 0.78 and 0.76 at the beginning of operation to 0.36 and 0.16, respectively.The BOD/COD ratio indicates the amount o� biodegradable compounds in the leachate.The BOD/COD ratio of Reactor B was lower than that of Reactor B; this may be as a result of accelerating biodegradable organic degradation under semi-aerobic condition.These results are in accordance with previous studies (Bilgili et al., 2006;Cossu et al., 2003).Their studies show that the COD values o� leachate �rom aerobic reactors were lower than those in the leachate �rom the anaerobic reactor.
Volatile fatty acids (VFA) are the most important inter-mediates in the anaerobic digestion process.Moreover, since VFA accumulation may lead to process failure due to the pH drop they induce (Anderson and Yang, 1992), VFA concentrations have been monitored �or a long time as process per�ormance indicators.The variation of VFA in leachate from semi-aerobic and anaerobic reactors is presented in Fig. 6.As illustrated in the figure, VFA concentration accumulated during the beginning of 60 d in Reactor B and reached a maximum value of 31 900 on Day 60, and then decreased quickly.However, VFA concentration accumulated and increased to a ma�imum value o� �� ��� on Day 70 in Reactor A, and then decreased slowly.The variation of HA, FA and HyI in leachate samples is presented in Fig. 7.The results showed that the DOM in leachate was mainly composed of an HyI fraction in Reactors A and B in the first 60 d.After semi-aerobic operation for Reactor B, the HyI fraction considerably decreased; correspondingly, the HA and FA fractions increased quickly, while the variation of DOM compositions was not obvious in Reactor A. All of these results show that the residual organic comprised non-biodegradable large-molecule compounds in the leachate from Reactor B, such as humic acid, which was difficult to biodegrade (Zouboulis et al., 2003).
The methane appeared in the gas phase of Reactor A and Reactor B since Day 100 and 65, respectively (Fig. 8).The volume �raction o� methane increased slowly and reached ��% a�ter 300 d in Reactor A. Different from Reactor A in which methanogenesis was largely inhibited by VFA accumulation and low pH, methane started to appear since the first week in the Reactor B. However, the volume �raction o� methane never e�ceeded ��%, and the corresponding O 2 level was appro�imately 2%.
These results indicated that semi-aerobic landfill can optimise micro-organism growth environments, and have a positive e��ect on the balanced growth o� the acid-production phase and methane production (Dong et al., 2007), and accelerate organic decomposition.However, the activity o� methanogen was inhibited in the anaerobic landfill reactor due to the accumulation of organic acid during the acid-production phase, and organic matter degradation in the leachate was slowed down consequently.

Nitrogenous compounds
The data o� NH + � -N and TKN in semi-aerobic and anaerobic reactors are given in Figs.�� and ��, respectively.The NH + � -N concentrations in the two reactors increased greatly in the initial stage and were kept constant after 40 d.The maximum

137
NH + � -N and TKN concentrations were measured to be � ���� and 3 345 mg/ℓ for semi-aerobic reactor, and to be 2 950 and 3 340 mg/ℓfor anaerobic reactor, respectively.After 60 d the NH + � -N concentrations in leachate decreased quickly for Reactor B due to semi-aerobic operation.Moreover, the NO - � -N and NO - 2 -N concentrations sharply increased, and then decreased over the subsequent period (Fig. 11).The NO - � -N and NO - 2 -N was not detected in Reactor A. The final NH + � -N concentrations in Reactors A and B were 2 490 and 22 mg/ℓ on Day 370, respectively.Differences between semi-aerobic and anaerobic reactors indicate the nitri�ying e��ect.The results suggested that semi-aerobic operation is an e��ective way to enhance biological nitrification and remove NH + � -N.The reactor was developed to include aerobic, anoxic and anaerobic zone under semi-aerobic condition.The sources o� carbon and nitrate necessary �or denitrification could be supplied by utilising leachate recirculation to carry the residual C and N from the anaerobic zone into the aerobic zone, and subsequently to the anoxic zone at the top of the system.Accordingly, ammonia nitrogen could be removed in situ (Dong et al., 2007).
Most of the nitrogen in landfill bioreactors is NH + � -N and is produced �rom the degradation o� proteins and amino acids (Price et al., 2003).Some researchers reported that ammonia was the most significant long-term component of leachate (Christensen et al., 1998), because ammonia is stable under anaerobic conditions.In general, there should be no apparent increase or decrease in the concentration o� all nitrogen groups during the anaerobic degradation of solid waste (Bilgili et al., 2006).Thus, the same trend o� TKN concentrations was observed during the study.
As illustrated, it can be seen clearly that NH + � -N could be in situ removed by semi-aerobic operation with leachate recirculation, which could avoid the disposing process o� the discharged leachate.Moreover, during the in situ NH + � -N removal process, the biodegradable organic matters could be used as carbon source of denitrification in semi-aerobic reactor and thus avoided the inadequate o� carbon sources at the late stage o� the recirculated landfill.

EEMs of the leachate
In this study, we applied three-dimensional fluorescence EEMs spectra for characterising the leachate samples.Each EEM provided spectral in�ormation about the organic matter compositions o� leachate samples.Fluorescence spectra such as those we have obtained contains a number of distinct peaks that are generally ascribed to either humic-like or protein-like fluorescence, and at least six peaks were readily identified from EEM fluorescence spectra of leachate DOM (see Fig. 12).The first main peak was a distinctive and intense fluorescence peak at excitation and emission wavelengths (Ex/Em) of 210-230 nm/340-360 nm (peak A), which is identical in location to the diagnostic fluorescence centre observed previously (Baker and Curry, 2004).Moreover, there were three peaks that were identified at Ex/Em of 210-230/300-320 nm, 270-280/300-320 nm and 270-280/340-380 nm, marked by peak B, peak C and peak D, respectively.The four peaks have been ascribed as protein-like peaks, in which the fluorescence is associated with the tryptophan and tyrosine (Baker, 2001;Chen et al., 2003;Yamashita and Tanoue, 2003;Baker and Inverarity, 2004).A fifth peak was located around Ex/Em of 310-340/420-440 nm (Peak E), which is attributed to aromatic and aliphatic groups in the DOM fraction and commonly labeled as fulvic-like (Coble, 1996).In these leachates we also observed a �� th peak at Ex/Em of 220 to 240/420 to 440 nm (Peak F), a poorly understood fluorescent centre widely attributed to a component of the humic fraction (Burdige et al., 2004;Coble, 1996;Yan et al., 2000).
All of the fluorescence peaks of the leachate in the Reactor A on Day 60, 95 and 250 were protein-like fluorescence (some examples are shown as contour maps in Fig. 12a, b, c), which are attributed to protein tryptophan and tyrosine (Burdige et al., 2004;Chen et al., 2003;Lawrence et al., 1999; Wu and Tanoue,  12d) on Day 95 and 250, which are attributed to aromatic and aliphatic groups in the DOM fraction and commonly labelled as humiclike and fulvic-like, and related to the carbonyl and carboxyl in the humus (Wu and Tanoue, 2001).The locations of peaks E and F for the leachate on Day 250 were red shifted to longer wavelengths compared to those of the leachate on Day 95 and the intensity of the fluorescence peaks tend to increase.The change of peak locations and fluorescence intensity may be ascribed to an increase of molecular size, aromatic polycondensation, level of conjugated chromophores and humification degree of organic matter as the landfill time increases.The result was in accordance with previous studies (Senesi and Miano, 1991;Wei et al., 2007).Some reports presented that most of the organic matters were biodegradable in the leachate at the beginning of landfill, and then became nonbiodegradable (such as humic acid) at the late stage of landfill (Zouboulis et al., 2003).The EEMs of leachate indicated that organic matters were humic-like and fulvic-like in the leachate of Reactor B. Furthermore, the fluorescence results suggest a larger humification degree of leachate in Reactor B with respect to the corresponding leachate in Reactor A. This is also evidence to suggest that semi-aerobic operation stimulates the organic decomposition.

Conclusions
The results indicated that the semi-aerobic landfill could separate the acid �ormation phase and the methane �ermentation phase, optimise the micro-organism growth environment and accelerate the degradation of organic matter.Moreover, it showed lower emissions �or leachate than those in leachate �rom anaerobic landfill, with low concentrations of COD, BOD � , VFA, ammonia and TKN.In situ removal o� NH + � -N and organic matter in leachate could be carried out in a recycled semi-aerobic landfill and there�ore the disposal process o� the discharged leachate is avoided.Compared with traditional anaerobic and aerobic bioreactor landfill operations, the introduction of air into the landfills by natural ventilation induced a rapid and marked oxidation o� organic matter and nitrogen.The positive e��ect o� leachate recirculation was seen more clearly in a semi-aerobic bioreactor landfill operation than in the anaerobic bioreactor landfill.
Figure 1Schematic of simulated anaerobic and semi-aerobic reactors . The initial COD concentrations were around 40 000 mg/ℓ in leachate for Reactor A and B. COD concentrations increased rapidly a�ter the e�periment commenced, and reached the maximum values of 82 600 and 90 200 mg/ℓ for Reactor A and B after 65 and 55 days, respectively.After reaching to maximum values, COD concentrations began to decrease rapidly for Reactor B, and the decreasing rate of concentration from Reactor B was greatly faster than that from Reactor A. The final concentrations measured in Reactor A and B on Day 370 was 15 600 and 2 270 mg/ℓ, respectively.Changes of BOD � concentrations are presented in Fig. �.The BOD � results showed similar trend with COD.The final concentrations measured in Reactor A and B on Day 370 were

Figure 5 Figure 4 Figure 3 Figure 2
Figure 5Changes of BOD concentrations in semi-aerobic and anaerobic reactors

Figure 8 Figure 6 Figure 7
Figure 8Time evolution of gas compositions in semi-aerobic and anaerobic reactors Figure 10Changes of TKN concentrations in semi-aerobic and anaerobic reactors Figure 12Fluorescence EEM contour plots for DOM of the leachate in Reactors A and B