Science of the Total Environment 407 (2009) 2557–2564

Contents lists available at ScienceDirect

Science of the Total Environment

j ourna l homepage: www.e lsev ie r.com/ locate /sc i totenv

Treating landfill leachate using passive aeration trickling filters; effects of leachatecharacteristics and temperature on rates and process dynamics

Richard Matthews ⁎, Michael Winson, John ScullionInstitute of Biological, Environmental and Rural Sciences, Aberystwyth University, Ceredigion, Wales, SY23 3DA, UK

⁎ Corresponding author. Tel.: +44 1970 622304; fax: +E-mail address: rwm@aber.ac.uk (R. Matthews).

0048-9697/$ – see front matter © 2009 Elsevier B.V. Aldoi:10.1016/j.scitotenv.2009.01.034

a b s t r a c t

a r t i c l e i n f o

Article history:

Biological ammoniacal-nitro Received 29 July 2008Received in revised form 12 January 2009Accepted 14 January 2009Available online 13 February 2009

Keywords:Landfill leachateTreatmentTrickling filterTemperatureNH4

+-NTONTOC

gen (NH4+-N) and organic carbon (TOC) treatment was investigated in replicated

mesoscale attached microbial film trickling filters, treating strong and weak strength landfill leachates in batchmodeat temperaturesof 3,10,15 and30°C. Comparing leachates, ratesofNH4

+-N reduction (0.126–0.159gm−2 d−1)were predominantly unaffected by leachate characteristics; therewere significant differences inTOC rates (0.072–0.194 g m−2 d−1) but no trend relating to leachate strength. Rates of total oxidised nitrogen (TON) accumulation(0.012–0.144 g m−2 d−1) were slower for strong leachates. Comparing temperatures, treatment rates variedbetween0.029–0.319gNH4

+-Nm−2 d−1 and0.033–0.251gCm−2 d−1 generally increasingwith rising temperatures;rates at 3 °C were 9 and 13% of those at 30 °C for NH4

+-N and TOC respectively. For the weak leachates (NH4+-

Nb140 mg l−1) complete oxidation of NH4+-N was achieved. For the strong leachates (NH4

+-N 883–1150 mg l−1) abiphasic treatment response resulted inNH4

+-N removal efficiencies of between68and88%and for one leachatenodirect transformation of NH4

+-N to TON in bulk leachate. The temporal decoupling of NH4+-N oxidation and TON

accumulation in this leachate couldnot be fully explainedbydenitrification, volatilisation or anammox, suggestingtemporary storage of Nwithin the treatment system. This study demonstrates that passive aeration tricklingfilterscan treat well-buffered high NH4

+-N strength landfill leachates under a range of temperatures and that leachatestrength hasno effect on initial NH4

+-N treatment rates.Whether this approach is a practicable option depends on arange of site specific factors.

© 2009 Elsevier B.V. All rights reserved.

1. Introduction

Water percolating through putrescible wastes emplaced in alandfill can mobilise organic and inorganic components and decom-position products to generate a potentially polluting liquid leachate.Leachate characteristics vary between and within landfill sitesdependent on climate, site management practices and the composi-tion and degradation stage of the emplaced wastes. A series ofinterrelated physical, chemical and biological processes determinewaste progression through progressive degradation stages withimplications for the quality, quantity, and composition of any leachateproduced. Young acetogenic leachates from operational sites containhigh concentrations of ammoniacal-nitrogen (NH4

+-N), Total OrganicCarbon (TOC), Chemical Oxygen Demand (COD) and Biological OxygenDemand (BOD) (Renou et al., 2007). High concentrations of NH4

+-N andTOC persist even in agedmethanogenic leachates and these pollutantsare particularly harmful if released in high concentrations into aquaticsystems (Haarstad and Maehlum, 1999). For these reasons leachatetreatment is necessary throughout the operation, restoration andlengthy aftercare period of a landfill.

44 1970 622350.

l rights reserved.

Leachate treatment should be simple, universal and adaptable(Renou et al., 2007). Many treatment schemes utilise aerobic biologicalnitrification, the coupled two step oxidation of ammonia to nitrite(nitritation) by ammonia oxidising bacteria (AOB) and nitrite to nitrate(nitration) by nitrite oxidising bacteria (NOB) for the treatment of NH4

+-N and heterotrophic bacteria for organic carbon treatment in variousreactor configurations (Knox, 1985; Ilies and Mavinic, 2001; Kim et al.,2006). Whilst the coexistence of nitrifiers and heterotrophs can allowfor simultaneous ammonia and carbon oxidation, interspecies compe-tition for substrates can affect mass transfer reactions and conse-quently the stability and performance of a treatment system (Fzd-Polanco et al., 2000; Rostron et al., 2001; Carrera et al., 2004).

Biological processes have been shown to be effective in removingnitrogenous and organic matter from acetogenic leachates (Renouet al., 2007) but bio-refractory organic compounds (e.g. humicsubstances or surfactants) can limit process effectiveness (Wisz-niowski et al., 2006). Di Iaconi et al. (2006) reported very lowammonia removal efficiencies treating a mature landfill leachate andexponential decreases in nitrification rate with increasing COD:Nratios have been observed by Carrera et al. (2004) treating highstrength ammonium wastewaters. Consequently, air-stripping orchemical precipitation pre-treatment of strong leachates has beenrecommended (Li and Zhao, 1999) in order to provide or maintainfavourable conditions for nitrifiers and ensure effective nitrification.

Fig. 1. Schematic of a trickling filter unit.

2558 R. Matthews et al. / Science of the Total Environment 407 (2009) 2557–2564

In activated sludge processes elevated concentrations of NH4+-N

have been shown to reduce microfauna abundance (Puigagut et al.,2005), inhibit microbial activity (Li and Zhao, 1999) and, in a biofilmairlift reactor to inhibit nitrogen transformations (Kim et al., 2006).Whilst increasing temperatures can increase cellular metabolism(Fdz-Polanco et al., 1994) they can also selectively inhibit nitrogentransformations in strong leachates by increasing levels of freeammonia (FA). FA inhibition of AOB is less predictable than that ofNOB; Anthonisen et al. (1976) reported nitration inhibition by FA at0.1–1.0 mg l−1 NH3-N and nitritation inhibition at 10–150 mg l−1 NH3-N, resulting respectively in nitrite and ammonia accumulation.However Villaverde et al. (2000) found nitrifying microbial filmacclimation to FA, suggesting that tolerance to FA may be treatmentsystem or microbial community dependent.

Temperature perturbations can also alter microbial communitycomposition and function, affecting species diversity and hencetreatment system stability. Fdz-Polanco et al. (1994) found tempera-ture to be the decisive parameter in the nitrifying capacity of abiological aerated filter (BAF) whilst Ilies and Mavinic (2001)attributed nitrification failure, during high strength landfill leachatetreatment, primarily to decreased temperature. However whilstdecreasing temperatures may reduce specific microbial growth ornitrification rates, increased oxygen saturation may allow improvedsubstrate penetration and potentially overcome diffusion limitationproblems often associated with immobilised biomass (Parker et al.,1995; Rostron et al., 2001).

Biomass immobilised in trickling filter systems can offer treatmentadvantages in terms of simplicity of use, reduced commissioning,maintenance, energy and operating costs and improved processstability (Eding et al., 2006). Recent research on passively aerated,attachedmicrobial film trickling filter systems has been confined, witha notable exception (Knox, 1985), to tertiary nitrifying applications forlow strength wastewaters such as those encountered in domesticsewage (Wik, 1999), aquaculture (Eding et al., 2006) or potable water(Tekerlekopoulou and Vayenas, 2008) treatment. There has been littleconsideration given to their capability or potential on-site applicationeither as a stand-alone treatment technology or as part of anintegrated treatment train for more polluted landfill leachates.

The aims of this study were: (1) to determine treatment rates forstrong leachates relative to those achieved for weak leachates, (2) toinvestigate potential inhibition of N transformation processes instrong leachates and (3) to assess the effect of variations intemperature on these processes and rates.

2. Materials and methods

The treatment system investigated was based on one currently inuse on several rural, closed and partially restored landfill sites inmid-Wales, UK. Mesoscale batch treatment systems modelled on thesesite systems were used in a series of experiments with a variety oflandfill leachates under controlled temperature regimes. This ap-proach allowed performance to be evaluated at scales incorporatingthe inherent heterogeneity of full-scale treatment systems whilstallowing for the replication necessary to validate experimental out-comes statistically.

2.1. Experimental system

Sixteen trickling filter units (Fig. 1) were constructed. Each unitconsisted of a cylindrical aeration tower (1.0 m by 0.3 m dia., internalvolume 70 l) over a 50 litre capacity recycle sump. Each towerwas filled with randomly packed polypropylene plastic media(Mass Transfer Cascade Filterpak™ Stoke, UK; specific surface area220 m2 m−3) to give an average surface area for microbial colonisationof 15.6 m2 per unit with an average void ratio of 92%.

Leachate was continuously re-circulated (50 l per hour) around theunit, via a submersible pump located in the sump, to a dispersal platewhich ensured homogenous distribution and continuous mediawetting. Leachate trickled through the tower before returning to thesump. A flow splitter located on the submersible pump ensured thatleachate within the sump was continuously mixed.

Aeration, necessary for effective nitrification, was achieved pas-sively by the passage of leachate through the voids in the towermedia.Effluent sampled from the base of towers was approximately oxygensaturated (mean DO 6.5 mg l−1 at 30 °C; 9.9 mg l−1 at 3 °C) and, whenoperated in batchmode, DO levels in the sumpprogressively increased,after an initial lag period (1–6 h) to N6mg l−1. The large void ratio of themedia combined with hydraulic scour from the leachate ensured thatdetrimental clogging or blinding of the plastic media did not occur.

For each leachate–temperature combination investigated, fourreplicate units were used to address issues of reproducibility inbiofilm systems (Lewandowski et al., 2004). Units were conditioned toleachates by pre-treatment runs prior to each experiment commen-cing. Before the current period of experimentation, all units had beenoperated concurrently under controlled conditions for over two years,treating a variety of landfill leachates sourced from operational andclosed landfill sites. Results from previous experiments with thesesystems (Matthews et al., 2006), suggested that an effective microbialfilm had developed and been maintained on the plastic media.

2.2. Leachate

Landfill leachate was collected from four landfill sites to provide arange of leachate characteristics (Table 1). NYCC leachate was collectedfrom a closed and fully restored landfill and is typical of an agedmethanogenic leachate with low NH4

+-N, TOC, BOD and COD concentra-tions. BRTH leachate was collected from a closed and recently restoredlandfill, the NH4

+-N, TOC, BOD and COD concentrations are typical of amethanogenic leachate derived from aging wastes. TREC leachate wascollected from an operational landfill accepting municipal solid (MSW)and commercial and industrialwastes; this acetogenic leachatehadhighNH4

+-N and TOC concentrations and a particularly high BOD:COD ratio.NYCO, the strongest NH4

+-N leachate, was collected from an operationallandfill acceptingMSW; the high TOC, BOD and COD concentrations aretypical of a young acetogenic leachate originating from freshly emplacedputrescible wastes.

Leachates from the closed sites (NYCC, BRTH) were collected fromsubterranean storage sumps, whilst on operational sites (TREC, NYCO)

Table 1Leachate characteristics in this study

Site Status pH EC NH4+-N NO3

−-N NO2−-N TOC COD BOD

µS cm−1 mg l−1

NYCC Closed 7.73 (0.06) 937 (47) 39.2 (3.2) 6.43 (4.79) 0.84 (1.10) 59.2 (8.6) 131 (12) 22 (4)BRTH Closed 7.02 (0.06) 1869 (187) 132.8 (17.8) 2.07 (1.54) nd nd 308 (33) 137 (11) 42 (6)TREC Operational 7.86 (0.15) 14,062 (910) 883 (59) 175.8 (30.1) nd nd 959 (42) 1575 (33) 124 (9)NYCO Operational 8.24 (0.05) 23,562 (1123) 1150 (76) 29.4 (15.8) 4.45 (0.17) 1350 (56) 5315 (143) 639 (14)

Mean (standard deviation); nd = not detected.

2559R. Matthews et al. / Science of the Total Environment 407 (2009) 2557–2564

they were collected directly from leachate wells located in active cells.Leachates were transported by tanker and used within 24 h minimis-ing any potential changes in characteristics arising from storage.Individual leachates were used in four separate experimental runswithout dilution, pH manipulation or nutrient additions.

2.3. Temperature

A range of temperatures was selected in order to investigatetreatment system response; 3 °C as a likely lower limit, 30 °C as close tothe optimumfor nitrification (Fdz-Polanco et al.,1994) and10 and15 °Cto reflect common operating conditions encountered in temperateclimates such as the UK (Knox, 1985).

Temperature treatmentswere randomlyassigned to individual units.Four replicate units were operated at 3, 10, 15 and 30 °C±0.2 °C intemperature controlled rooms. Thermal shockwas avoided by graduallyadjusting rooms to target temperature over a period of oneweekwhilsttreating the leachate to be used in the subsequent experimental phase.Treatment rates were monitored during this period to ensure that eachunit was operating consistently. At the end of this conditioning period,unitswere emptied and refilledwith freshly collected leachate. Leachatecirculation (reported as experimental time ‘0’) commenced when bulkleachate temperature stabilised to the designated room temperature.The random reassignment of units to temperature treatments wasrepeated for each leachate tested.

2.4. Sampling and analyses

Leachate samples were collected from the continuously mixedrecirculation sumps of individual units. Frequency of sample collectionvaried between leachate dependent on strength, ranging from every2 h for NYCC, to every 12 h for BRTH, to daily for TREC and NYCO.Additional sampling was undertaken in the early stages of eachexperiment, complementing the existing sampling regime, to providegreater detail on start up conditions. Complete oxidation or stabilisa-tion of NH4

+-N concentrations determined the end of each experiment.Bulk leachate pH and redox (Eh) (Hanna 301 meters) and dissolved

oxygen (Jenway 970 DO meter) were recorded in situ. Other analyseswere completed within 4 h of sampling. NH4

+-Nwas determined by thephenol–hypochlorite method (Harwood and Kühn, 1970) using a CecilSeries 2 digital spectrophotometer measuring absorbance at 630 nm.Nitrite (NO2

−) and nitrate (NO3−) concentrations (TON) were measured

by ion exchange chromatography (Metrohm Ion Chromatograph.Eluent: 0.2 M Na2CO3 and 0.75 M NaHCO3; Regenerant: 0.025 MH2SO4). Samples were filtered (Whatman 0.2 µm membrane filter)prior to injection. Organic carbon (TOC) was measured using aShimadzu 5050A TOC analyser (Shimadzu, Japan). Variations inconcentrations of FA (Eq. (1)) as a function of total NH4

+-N concentra-tion, pH and temperature were calculated from measured parametersaccording to Anthonisen et al. (1976).

FA free NH3½ � =NH +

4 QN� �

d 10pH� �

Kb=Kwð Þ + 10pH� � ð1Þ

where NH4+-N is the total ammoniacal-nitrogen concentration and Kb

and Kw are the ionisation constants for ammonia and waterrespectively. Additional sampling for nitrous oxide (syringe samplingof tower headspace and analysis by gas chromatography — Pye-Unicam Series 104) and volatilised ammonia (trapping from head-space air pumped through 2% boric acid) was undertaken on selectedoccasions during the main experimental programme to determineprimarily whether denitrification and ammonia volatilisation wereoccurring, rather than to quantify process rates. The experimental set-up made accurate rate quantification impracticable.

2.5. Preliminary experiments

Preliminary experiments across the range of temperatures used inthemain experimental programmewere undertaken to determine theextent of any abiotic losses of N by NH4

+-N volatilisation and mediaadsorption. Scaled-down versions of the main passive aeration,trickling filter systemwere set up in incubators using non-inoculatedmedia and synthetic ammonium nitrate solutions (pH and NH4

+-Nconcentrations as for stronger leachates). Changes in solution NH4

+-Nand TON were recorded with volatilised NH4

+-N trapped in boric acid;at the end of trial runs media were extracted with 2 M KCl andmineral-N concentrations determined by steam distillation–titrationto detect any adsorbed N species. Results from these experiments,even under themost favourable conditions for volatilisation (30 °C, pH8.5, NH4

+-N 1684 mg l−1), showed reductions in NH4+-N concentrations

and total ammonia trapped over 4 days to be b2.5% of the solutionNH4

+-N content. Mineral-N species in media extracts were consistentlybelow detectable limits, effectively discounting N adsorption onto themedia as an important process.

2.6. Statistical analysis and data handling

Treatment rates for individual experimental units were deter-mined by fitting linear regressions to variations in measuredparameters over time (Weon et al., 2004; Scullion et al., 2007). R-squared values for all linear regressions were greater than 0.92 andregularly exceeded 0.97 (pb0.001). Since biofilm viability and cover-age were not measured directly, quoted rates (g m2 biofilm d−1)assume complete biomass coverage of the available surface area of theplastic media. For the strong leachates (TREC, NYCO) a biphasictreatment was apparent; after rapid removal of 70–80% of NH4

+-N andTOC, treatment rates declined. Treatment rates reported for theseleachates are for the initial, non-limited treatment phase. Reasons forthis decline are considered in the Discussion.

Treatment responses (leachate and temperature) and their inter-actions were evaluated by two-way ANOVA, without data transforma-tion, after appropriate tests for normality and equality of variance;least significant difference (LSD) analysis at 95% confidence was usedto differentiate individual means. The random assignment of treat-ments to units, standardised pre-experimental procedures and therobust environmental controls imposed were considered sufficient toallow data from the four experimental leachate runs to be analysed asa single experiment.

Fig. 2. Leachate–temperature treatment rate (g m2 media surface area) interactioneffects for a) TON accumulation; b) TOC reduction for □ NYCC, ○ BRTH, ▲ TREC and■ NYCO.

Table 3Average NH4

+-N loss and concomitant TON gain (mg) in bulk leachate at temperatures of3, 10, 15 and 30 °C over the entire experimental period

2560 R. Matthews et al. / Science of the Total Environment 407 (2009) 2557–2564

3. Results

Significant leachate and temperature effects were obtained for allof the main treatment rate characteristics measured (Table 2).Comparing leachates, NYCO had the highest (pb0.01) treatmentrates for NH4

+-N but the lowest (pb0.001) TON accumulation ratealthough these treatment rate differences were less than 20%. Therewas no significant difference in TON accumulation rates betweenBRTH and TREC but NYCC leachate had a higher (pb0.01) rate. Allleachates differed significantly in their TOC reduction rates ranking inthe order BRTHNNYCONTRECNNYCC. The BRTH rate was almost threetimes that of the NYCC leachate and more than 50% higher than thoseof the two operational site leachates. Higher temperatures causedprogressive increases in treatment rates, but for 10 and 15 °Cdifferences were significant for NH4

+-N only; all other temperatureeffects were statistically significant (pb0.05). Comparing 10 with30 °C, rates were slightly less than four times greater at the highertemperature for NH4

+-N and less than three times greater for TOC. At3 °C NH4

+-N, TON and TOC rates were respectively 9, 5 and 13% ofcorresponding values for the 30 °C optimum.

Highly significant leachate–temperature interaction effects (Fig. 2)were found for TON (pb0.001) and TOC (pb0.001). TON interactioneffects can be attributed mainly to the very limited response toincreased temperatures in the NYCO leachate and the marked BRTHleachate response between 15 and 30 °C; there was no correspondinginteraction for NH4

+-N. The TOC interaction can be attributed to theelevated NYCO leachate response at 15 °C and the reduced response at30 °C; BRTH leachate again exhibited a marked rate response between15 and 30 °C.

When losses of NH4+-N over the entire experimental period were

compared with gains in TON (Table 3) significant differences wereapparent. For the two closed site leachates NYCC and BRTH and forTREC at the higher temperatures most, and in some cases more, of theNH4

+-N lost was accounted for as TON. For TREC at 3 °C and NYCO at 3and 10 °C very little of the reduction in NH4

+-N was accounted for asTON, even when measurements were continued over a prolongedperiod after the initial period of rapid NH4

+-N loss. For the strong NYCOleachate at 15 and 30 °C about 50% of the NH4

+-N loss was recovered asTON by the end of the experiment.

The NH4+-N rate response to increasing temperature was similar for

all leachates but the NYCO TON response differed from that of otherleachates as already noted. For this reason changes in concentrationsof NH4

+-N and TON over time, for the two strong operational siteleachates are presented in more detail. At temperatures of 10, 15 and30 °C, TREC leachate (Fig. 3) TON concentrations increased more orless in linewith reductions in NH4

+-N, following initially rapid and thensubsequently reduced treatment phases. TREC leachate at 3 °C andNYCO leachate (Fig. 4) at 3 and 10 °C produced very little TON despitesubstantial losses of NH4

+-N. At 15 and 30 °C, substantial amounts of

Table 2Average treatment rates (g m2 media surface area d−1) for measured parameters fora) leachate and b) temperature over the main period of leachate amelioration

a) Leachate

NYCC BRTH TREC NYCO

NH4-N 0.126b 0.136b 0.139b 0.159aTON 0.144c 0.122b 0.110b 0.012aTOC 0.072d 0.194c 0.102b 0.128a

b) Temperature

3 10 15 30

NH4-N 0.029d 0.085c 0.127b 0.319aTON 0.012c 0.064b 0.079b 0.233aTOC 0.033c 0.096b 0.115b 0.251a

Within a rowmeans with a common letter suffix do not differ at a 5% level of probability(least significant difference).

TON were accumulated in the NYCO leachates but only after a delay of100–250 h after the main period of NH4

+-N loss. For individualreplicates of these last two treatment combinations, TON accumula-tion had either stabilised at levels equivalent to N80% of NH4

+-N loss orexhibited on-going rapid TON accumulation up to the final sampling.

These same leachate–temperature combinations also differed inrelation to accumulation of NO2

−–N. Whereas large concentrations ofNO2

−–N were detected (Fig. 5) during the intermediate phases of TRECleachate treatment at the higher temperatures, little nitrite wasdetected at any stage of the treatment process at 3 °C. For the strongNYCO and the weak BRTH and NYCC leachates there were very lowconcentrations (b5 mg l−1) of NO2

−–N in the initial treatment phases

Leachate Temperature NH4+-N loss TON gain Significance

NYCC 3 1847 (460) 1365 (582) ns10 2155 (199) 3199 (163) ⁎⁎

15 1843 (148) 3387 (177) ⁎⁎⁎

30 1824 (47) 3332 (104) ⁎⁎⁎

BRTH 3 6182 (209) 5568 (1162) ns10 6506 (176) 7445 (2045) ns15 6848 (367) 6416 (238) ns30 7129 (219) 8379 (551) ⁎

TREC 3 11,017 (3076) 2356 (2274) ⁎⁎

10 21,715 (2963) 25,664 (11,872) ns15 29,892 (6169) 24,933 (8916) ns30 25,089 (4535) 35,202 (4491) ns

NYCO 3 24,953 (5695) 1452 (767) ⁎⁎

10 44,728 (4312) 3598 (1725) ⁎⁎⁎

15 49,065 (9960) 24,539 (1296) ns30 56,895 (8313) 26,718 (18,731) ns

Mean (standard deviation); paired t-test, ns — not significant, ⁎pN0.05, ⁎⁎pN0.01,⁎⁎⁎pN0.001.

Fig. 4. NYCO leachate — mean concentrations (mg l−1) of a) NH4+-N and b) TON in bulk

leachate over experimental period at ○ 3; ▲ 10; □ 15 and ♦ 30 °C.

Fig. 5. Mean NO2−N concentrations in bulk leachate over experimental period for TREC

leachate at ○ 3; ▲ 10; □ 15 and ♦ 30 °C.

Fig. 3. TREC leachate — mean concentrations (mg l−1) of a) NH4+-N and b) TON in bulk

leachate over experimental period at ○ 3; ▲ 10; □ 15 and ♦ 30 °C.

2561R. Matthews et al. / Science of the Total Environment 407 (2009) 2557–2564

and no apparent accumulation. After 200 h no NO2−–Nwas detected in

the bulk NYCO leachate or in any of the weak, closed site leachates.Temporal trends in pH and Eh for TREC and NYCO are presented

in Fig. 6 because of their known influence on N transformations.Whereas pH fell to about 6.5 following an initial lag period in thehigher temperature TREC and NYCO leachates, at the lower tempera-tures leachate pH remained high or increased. TREC leachates alsoshowed markedly lower redox potentials than those for NYCO, withthe 3 °C treatment exceptional in maintaining a negative Eh over mostof the experimental period despite high (8.7 mg l− 1) DOconcentrations.

4. Discussion

The findings reported here are the first to make a direct replicatedcomparison in passive aeration trickling filters between strongand weak strength landfill leachate treatment under a range oftemperatures. Although there were some statistically significantdifferences between leachates and some leachate–temperatureinteractions, reductions in NH4

+-N occurred at broadly similar ratesacross awide range (40–1150mg l−1) of initial NH4

+-N concentrations.Significant differences in TOC treatment rates were not related toinitial TOC concentrations. For different leachate–temperaturecombinations N transformation processes and their dynamics varied,with TON rates tending to decrease with increasing initial NH4

+-Nconcentrations.

Limited control over operating temperatures is often considered aweakness in trickling filter systems. Average rates of NH4

+-N reductionfor all leachates at temperatures of 3, 10, 15 and 30 °C were 0.029,0.085, 0.127 and 0.319 g m−2 d−1 respectively and compare favourablywith those reported by Knox (1985). For TOC treatment rates variedbetween 0.033 at 3 °C and 0.251 g m−2 d−1 at 30 °C and exhibited abiphasic pattern characterised by initially rapid and subsequentlyreduced rates. Mean residual TOC concentrations in the weak NYCCand BRTH leachates were 32±4 and 46±7 mg l−1 respectively withremoval efficiencies of 46 and 85%. For the strong TREC and NYCOleachates TOC removal efficiencies of 50 and 68% were achieved withresidual concentrations of 476±35 and 416±36 mg l−1. All treatedleachates had however very low residual BOD (6–27 mg l−1 equivalentto 80–96% removal efficiencies) suggesting that much of the TOCremaining, when experiments terminated, was not readily biodegrad-able. Knox (1985) attributed comparable, poor leachate TOC removalsin a trickling filter to similarly recalcitrant, poorly biodegradable TOC.

All process rates showed a more or less linear response toincreasing temperatures rather than the exponential increase pre-dicted by Arrhenius kinetics, suggesting mass transfer rather thankinetic controls. Overall, temperature was a key factor affectingoperational efficiency. Nevertheless, treatment of leachate at 3 °Ccontinued at rates equivalent to 9 and 13% of optimum for NH4

+-N and

Fig. 6. Temporal trends in pH and Eh for TREC leachate (up to 624 h a and b) and NYCO leachate (up to 912 h c and d) at ○ 3; ▲ 10; □ 15 and ♦ 30 °C.

2562 R. Matthews et al. / Science of the Total Environment 407 (2009) 2557–2564

TOC, suggesting that the major nitrification inhibition encountered byIlies and Mavinic (2001) at temperatures of 10 °C, in an activatedsludge process may not be as pronounced for our systems. Knox(1985) and Parker et al. (1995) also found that the percentage declinein reaction rates in nitrifying trickling filters, with decreasingtemperature, was less than that for activated sludge processes.

The exponentially declining rates of nitrification with increasinginfluent COD:N ratios reported by Carrera et al. (2004) treating highstrength ammoniumwastewaters in an activated sludge process, werenot evident in our studies. Our rates of NH4

+-N reduction were broadlysimilar despite initial leachate COD:N ratios varying between 1.03 and4.62. The leachate with the highest COD:N ratio had the highest rate ofNH4

+-N reduction, although ammonia volatilisation may partiallyexplain this finding. Elevated COD:N ratios have also been shown toaffect the spatial distribution and activity of nitrifying populations inan up-flow BAF treating synthetic wastewaters (Fzd-Polanco et al.,2000). Results presented here and those of extensive treatability trials(Matthews et al., 2006; Scullion et al., 2007) suggest that potentialcompetition between faster growing heterotrophic bacteria andslower growing nitrifiers, was not an important factor limitingtreatment rates in these very well established, as opposed to newly

colonising, microbial systems. Lydmark et al. (2006) have reportedvariably distributed microbial communities in full scale nitrifyingtrickling filters and some stratification of microbial metabolicprocesses, which may allow for effective activity of various bacterialcommunities within microenvironment niches.

Data reported here also suggest that NH4+-N treatment was not

itself inhibited by high (up to 1150 mg l−1) concentrations of NH4+-N.

There was however some evidence (Fig. 5) consistent with a transientinhibition of nitration resulting in nitrite accumulation in the earlierstages of TREC leachate treatment, possibly due to FA inhibition ofNOB (Kim et al., 2006); NH4

+-N oxidation is considered to be lesssensitive to FA inhibition than NO2

−–N oxidation (Anthonisen et al.,1976). Whereas NO2

−–N was not detected in either of the weak NYCCor BRTH leachates, nor did it accumulate in the NYCO leachate, markedaccumulations (up to 350±154 mg l−1 at 30 °C) occurred during theintermediate treatment phases of TREC leachate. The particularly lowredox values (Fig. 6) in bulk TREC leachate during the phase of rapidNO2

−N accumulation may have favoured this increase. For TRECleachate, calculated free nitrous acid (FNA) concentrations (max0.53 mg l−1 at 30 °C) increased to within the 0.22–2.8 mg l−1 FNAnitrification inhibitory range reported by Anthonisen et al. (1976) in

2563R. Matthews et al. / Science of the Total Environment 407 (2009) 2557–2564

activated sludge, possibly limiting further increases in nitriteconcentrations by inhibiting nitritation without affecting nitrationprocesses already inhibited by FA.

Although the NYCO leachate had the highest calculated mean FAconcentrations (80 mg l−1 and 450 mg l−1 at temperatures of 3 and30 °C respectively), its redox potentials weremarkedly higher, and lessfavourable to nitrite accumulation, than those for TREC leachates.Redox potential may therefore be a co-factor explaining in part theabsence of nitrite accumulation in the high NH4

+-N NYCO leachates.Variations inmicrobial acclimation to high concentrations of FA, foundto influence the degree of nitration inhibition encountered insubmerged nitrifying biofilters (Villaverde et al., 2000), may be afurther factor. Determining inhibitory concentrations of FA withinsystems remains difficult because of co-factors; variations in masstransfer controls between different treatment systems furthercomplicate generalisation of inhibition thresholds.

All leachates were used without pH manipulation or nutrientadditions. For the strong TREC and NYCO leachates, at all but thelowest temperature, NH4

+-N treatment followed a biphasic patterncharacterised by initially rapid and subsequently reduced rates. It iswidely acknowledged that nitrification may be subject to substrate ormetabolite inhibition or limitation; loss of buffering capacity, reducedconcentrations of NH3 at low pH and bicarbonate limitation and canall adversely affect nitrification of high strength nitrogenous waste-waters (Wett and Rauch, 2003). For TREC leachate at 30 °C inorganiccarbon (IC) concentrations (data not shown) declined from 308±42 mg l−1 to b10 mg l−1 within 168 h; over the same period pHdeclined from N8 to 6.6 (Fig. 6a) and amarked slow down in the rate ofNH4

+-N reduction (Fig. 3a) was apparent. For NYCO at 30 °C a similarpattern emerged, IC fell rapidly from 1061±46 mg l−1 to b2 mg l−1 buta pH N7 was maintained (Fig. 6c). Whilst pH may indicate a potentiallimitation, Tarre and Green (2004) have shown that it may notnecessarily itself be a limiting factor.

For the weak NYCC and BRTH leachates and for the strong TRECleachate at 15 and 30 °C, reductions in NH4

+-N can be attributed tonitrification as reflected by corresponding gains in TON. For TRECleachate (Fig. 3b) at 3 °C and for NYCO leachate (Fig. 4b) at alltemperatures there was a marked reduction of NH4

+-N over theexperimental period. For NYCO at the two higher temperatures rapidTON accumulation occurred only towards the final stages of experi-ments long after NH4

+N reductions had ceased; for individualreplicates of these treatments TON (predominantly NO3

−–N) concen-trations were equivalent to N80% of NH4

+-N losses. It remains possiblethat the other lower temperature leachates would have followed asimilar, delayed pattern of TON accumulation. Alternatively, sincethese leachates maintained high pH and NH4

+-N concentrations andlow Eh, some nitrogen may have been lost by concurrent nitrificationand denitrification within the immobilised microbial film (Wik, 1999)or ammonia volatilisation (Garzón-Zúñiga et al., 2007). We detectednitrous oxide and gaseous ammonia emissions in ambient air withintemperature controlled rooms, at least in the early stages of leachatetreatment, but we estimated these losses as ammonia to beapproximately 2% of total NH4

+-N reductions, in line with data frompreliminary trials. N2 emissions from the oxidation of ammoniumwith nitrite by anaerobic ammonia oxidising (anammox) bacteria inan attached microbial film, rotating biological contractor (RBC) havebeen reported by Egli et al. (2003) treating high strength landfillleachates. Anammox bacteria have also been detected in nitrifyingtrickling filters (Lydmark et al., 2006); the application of fluorescencein situ hybridisation (FISH) techniques could have determinedwhether there was a potential for loss by this pathway in ourexperiments.

Although there may have been some losses of N throughdenitrification, volatilisation or anammox TON accumulation gener-ally balanced for the most part NH4

+N reductions, albeit after aconsiderable delay in the NYCO leachate. For at least some leachates,

significant quantities of N were either temporarily removed from thebulk leachate or present in a form not detected by the analyticalmethods employed. An increase in biomass-N (Fzd-Polanco et al.,2000; Rostron et al., 2001) or temporary storage of inorganic-Nwithinthe microbial film (Wik, 1999) may have occurred over the treatmentperiod. The mechanism underlying this ‘storage’ phenomenon isworthy of further investigation since it has implications for treatmentsystems where input loadings are variable.

Experiments reported here were not designed to optimise processrates. In other experiments with the same experimental system,Matthews et al. (2006) reported three to four fold increases intreatment rates by manipulating leachate loading and contact timeduration between leachate and the aeration tower media. In practice,leachates would be treated in continuous flow, rather than the batchmode used in these experiments; this would allow for leachaterecirculation and the possibility of diluting stronger leachates so as toreduce any potential for FA inhibition. Weak leachates from the twoclosed sites were treated to ‘completion’ for NH4

+-N. For the strongleachates, although marked reductions in treatment rates occurred,nitrification inhibition due to pH or inorganic carbon substratelimitation could be avoided by appropriately managed inputs.Whether trickling filters are effective as a stand-alone treatmentoption will depend on a range of site specific, practical and economicfactors. Findings reported here indicate that with appropriatemanagement their use could be extended to stronger leachates inaddition to the well established ‘polishing’ role for previously treatedor weak leachates.

5. Conclusions

The trickling filter system reported here proved capable of treatinga variety of landfill leachates typical of operational and closed siteswith a broadly consistent performance. The treatment rates reportedare lower than those for some other treatment systems but they wereachieved without their additional operational complexity, monitoringrequirements and costs. For some strong leachates NH4

+-N andnitrogen oxidation processes were temporally decoupled. The extentto which this phenomenon is attributable to transient N storagewithin microbial films, to the presence of undetected N-forms withinbulk leachate, or a combination of these factors, requires furtherinvestigation.

Acknowledgements

This work was funded by Biffaward and Biffa Waste. Practicalsupport, helpful advice and comments from D. Broomfield, S. Mollard,S. Morgan, D. Robinson-Todd, S. Simmons and S. Beech and twoanonymous reviewers are gratefully acknowledged.

References

Anthonisen AC, Loehr RC, Prakasam TBS, Srinath EG. Inhibition of nitrification byammonia and nitrous-acid. J Water Pollut Cont Fed 1976;48(5):835–52.

Carrera J, Vicent T, Lafuente J. Effect of influent COD:N ratio on biological nitrogenremoval (BNR) from high-strength ammonium industrial wastewater. ProcessBiochem 2004;39(12):2035–41.

Di Iaconi C, Ramadori R, Lopez A. Combined biological and chemical degradation fortreating a mature municipal landfill leachate. Biochem Eng J 2006;31(2):118–24.

Eding EH, Kamstra A, Verreth JAJ, Huisman EA, Klapwijk A. Design and operation ofnitrifying trickling filters in recirculating aquaculture: a review. Aquacult Eng2006;34(3):234–60.

Egli K, Bosshard F, Werlen C, Lais P, Siegrist H, Zehnder AJB, et al. Microbial compositionand structure of a rotating biological contractor biofilm treating ammonium-richwastewater without organic carbon. Microb Ecol 2003;45(4):419–32.

Fdz-Polanco F, Villaverde S, García PA. Temperature effect on nitrifying bacteria activity inbiofilters: activation and free ammonia inhibition. Water Sci Technol 1994;30(11):121–30.

Fzd-Polanco F, Méndez E, Urueña MA, Villaverde S, García PA. Spatial distribution ofheterotrophs and nitrifiers in a submerged biofilter for nitrification. Water Res2000;34(16):4081–9.

2564 R. Matthews et al. / Science of the Total Environment 407 (2009) 2557–2564

Garzón-Zúñiga MA, Lessard P, Aubry G, Buelna G. Aeration effect on the efficiency ofswine manure treatment in a trickling filter packed with organic materials. WaterSci Technol 2007;55(10):135–43.

Haarstad K, Maehlum T. Important aspects of long-term production and treatment ofmunicipal solid waste leachate. Waste Manage Res 1999;17(6):470–7.

Harwood JE, Kühn AL. Colorimetric method for ammonia in natural waters. Water Res1970;4(12):805–12.

Ilies P, Mavinic DS. The effect of decreased ambient temperature on the biologicalnitrification and denitrification of a high ammonia landfill leachate. Water Res2001;35(8):2065–72.

Kim D-J, Lee D-I, Keller J. Effect of temperature and free ammonia on nitrification andnitrite accumulation in landfill leachate and analysis of its nitrifying bacterialcommunity by FISH. Bioresource Technol 2006;97(3):459–68.

KnoxK. Leachate treatmentwith nitrification of ammonia.WaterRes 1985;19(7):895–904.Lewandowski Z, BeyenalH, StookeyD.Reproducibilityof biofilmprocessesand themeaning

of steady state in biofilm reactors. Water Sci Technol 2004;49(11–12):359–64.Li XZ, Zhao QL. Inhibition of microbial activity of activated sludge by ammonia in

leachate. Environ Int 1999;25(8):961–8.Lydmark P, Lind M, Sörensson F, Hermansson M. Vertical distribution of nitrifying

populations in bacterial biofilms from a full scale nitrifying trickling filter. EnvironMicrobiol 2006;8(11):2036–49.

Matthews R, Winson M, Scullion J. Aerobic biological treatment of landfill leachate. WitTrans Ecol Envir 2006;92:437–46.

Parker D, Lutz M, Andersson B, Aspegren H. Effect of operating variables on nitrificationrates in trickling filters. Water Environ Res 1995;67(7):1111–8.

Puigagut J, Salvadó H, García J. Short-term harmful effects of ammonia on activatedsludge microfauna. Water Res 2005;39(18):4397–404.

Renou S, Givaudan JG, Poulain S, Dirassouyan F, Moulin P. Landfill leachate treatment:review and opportunity. J Hazard Mater 2007;150(3):468–93.

RostronWM, Stuckey DC, Young AA. Nitrification of high strength ammoniawastewaters:comparative study of immobilisation media. Water Res 2001;35(5):1169–78.

Scullion J, Winson M, Matthews R. Inhibition and recovery in a fixed microbial filmleachate treatment system subject to shock loading of copper and zinc. Water Res2007;41(18):4129–38.

Tarre S, Green M. High-rate nitrification at low pH in suspended- and attached-biomassreactors. Appl Environ Microb 2004;70(11):6481–7.

Tekerlekopoulou AG, Vayenas DV. Simultaneous biological removal of ammonia, iron andmanganese from potable water using a trickling filter. Biochem Eng J 2008;39(1):215–20.

Villaverde S, Fdz-Polanco F, García PA. Nitrifying biofilm acclimation to free ammonia insubmerged biofilters. Start-up influence. Water Res 2000;34(2):602–10.

Weon SY, Lee SI, Koopman B. Effect of temperature and dissolved oxygen on biologicalnitrification at high ammonia concentrations. Environ Technol 2004;25(1):1211–9.

Wett B, Rauch W. The role of inorganic carbon limitation in biological nitrogen removalof extremely ammonia concentrated wastewater. Water Res 2003;37(5):1100–10.

Wik T. Adsorption and denitrification in nitrifying tricklingfilters.Water Res 1999;33(6):1500–8.

Wiszniowski J, Robert D, Surmacz-Gorska J, Miksch K, Weber JV. Landfill leachatetreatment methods: a review. Environ Chem Lett 2006;4(1):51–61.