w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 1 7 8 5 – 1 7 9 6

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

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

Aerated treatment pond technology with biofilmpromoting mats for the bioremediation of benzene,MTBE and ammonium contaminated groundwater

Sven Jechalke a,*, Carsten Vogt a, Nils Reiche b, Alessandro G. Franchini c, Helko Borsdorf b,Thomas R. Neu d, Hans H. Richnow a

a Department of Isotope Biogeochemistry, Helmholtz Centre for Environmental Research – UFZ, Permoserstr. 15, D-04318 Leipzig, Germanyb Department of Monitoring- and Exploration Technologies, Helmholtz Centre for Environmental Research – UFZ, Permoserstr. 15,

D-04318 Leipzig, Germanyc Department of Environmental Biotechnology, Helmholtz Centre for Environmental Research – UFZ, Permoserstr. 15,

D-04318 Leipzig, Germanyd Department of River Ecology, Helmholtz Centre for Environmental Research – UFZ, Permoserstr. 15, D-04318 Leipzig, Germany

a r t i c l e i n f o

Article history:

Received 4 August 2009

Received in revised form

1 December 2009

Accepted 1 December 2009

Available online 13 January 2010

Keywords:

Groundwater remediation

Geotextiles

MTBE and benzene degradation

Compartment transfer

* Corresponding author. Tel.: þ493412351360E-mail address: sven.jechalke@ufz.de (S.

0043-1354/$ – see front matter ª 2009 Elsevidoi:10.1016/j.watres.2009.12.002

a b s t r a c t

A novel aerated treatment pond for enhanced biodegradation of groundwater contaminants

was tested under field conditions. Coconut fibre and polypropylene textiles were used to

encourage the development of contaminant-degrading biofilms. Groundwater contami-

nants targeted for removal were benzene, methyl tert-butyl ether (MTBE) and ammonium.

Here, we present data from the first 14 months of operation and compare contaminant

removal rates, volatilization losses, and biofilm development in one pond equipped with

coconut fibre to another pond with polypropylene textiles. Oxygen concentrations were

constantly monitored and adjusted by automated aeration modules. A natural transition

from anoxic to oxic zones was simulated to minimize the volatilization rate of volatile

organic contaminants. Both ponds showed constant reductions in benzene concentrations

from 20 mg/L at the inflow to about 1 mg/L at the outflow of the system. A dynamic air

chamber (DAC) measurement revealed that only 1% of benzene loss was due to volatiliza-

tion, and suggests that benzene loss was predominantly due to aerobic mineralization.

MTBE concentration was reduced from around 4 mg/L at the inflow to 3.4–2.4 mg/L in the

system effluent during the first 8 months of operation, and was further reduced to 1.2 mg/L

during the subsequent 6 months of operation. Ammonium concentrations decreased only

slightly from around 59 mg/L at the inflow to 56 mg/L in the outflow, indicating no signif-

icant nitrification during the first 14 months of continuous operation. Confocal laser scan-

ning microscopy (CLSM) demonstrated that microorganisms rapidly colonized both the

coconut fibre and polypropylene textiles. Microbial community structure analysis per-

formed using denaturing gradient gel electrophoresis (DGGE) revealed little similarity

between patterns from water and textile samples. Coconut textiles were shown to be more

effective than polypropylene fibre textiles for promoting the recruitment and development

of MTBE-degrading biofilms. Biofilms of both textiles contained high numbers of benzene

metabolizing bacteria suggesting that these materials provide favourable growth conditions

for benzene degrading microorganisms.

ª 2009 Elsevier Ltd. All rights reserved.

; fax: þ493412351443.Jechalke).er Ltd. All rights reserved.

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 1 7 8 5 – 1 7 9 61786

1. Introduction

Table 1 – Composition of inflowing and outflowing groundwater from both basins. Values representaverages from 14 months of operation with standarddeviations (±SD) for 33–40 measurements.

Parameter[mg/L]

Inflow Outflow basin1 coconutmaterial

Outflow basin2 polypropylene

material

Ammonium 59 � 5 56 � 5 56 � 4

Nitrate <0.1 <0.1 <0.1

Sulphate 6 � 4 5 � 4 5 � 3

Iron 5.8 � 0.7 3 � 2 3 � 1

Iron(II) 6 � 1 1 � 1 1.7 � 0.8

Manganese 1.6 � 0.1 1.4 � 0.3 1.4 � 0.2

Magnesia 60 � 2 60 � 2 60 � 2

Sodium 139 � 7 138 � 7 139 � 6

Chloride 119 � 7 120 � 7 121 � 6

Phosphate 1.4 � 0.7 0.7 � 0.3 0.5 � 0.2

Phosphorus 0.9 � 0.1 0.5 � 0.2 0.5 � 0.2

Calcium 210 � 10 195 � 17 197 � 13

Potassium 13.6 � 0.6 13.7 � 0.6 13.7 � 0.6

Benzene 20 � 2 0.002 � 0.001 0.001 � 0.001

MTBE 3.9 � 0.5 2.5 � 0.8 2.5 � 0.6

Fuel, fuel additives and ammonium are frequently detected

water pollutants worldwide (Christensen et al., 2001; Squillace

et al., 1996). Benzene, toluene, ethylbenzene, the three xylene

isomers (m-, o-, and p-xylene, BTEX compounds) and methyl

tertiary-butyl ether (MTBE) are highly soluble and therefore

extremely mobile in groundwater systems (Squillace et al.,

1996). Therefore, these compounds are of environmental

concern and represent suitable organic contaminants for

testing groundwater remediation system effectiveness.

Benzene, the most toxic and water soluble BTEX compound,

can be degraded by many microorganisms under oxic (Agteren

et al., 1998) and even hypoxic conditions (Yerushalmi et al.,

2002). MTBE biodegradation is slow and difficult due to steric

effects within the molecule, but has been shown by several

studies to be degraded under aerobic conditions (Ferreira et al.,

2006; Schmidt et al., 2004). Notably, growth rates and biomass

yields for aerobic MTBE degraders have been shown to be much

lower than aerobic benzene degraders (Fortin et al., 2001). Both

MTBE and benzene are highly recalcitrant under anoxic

conditions (Foght, 2008; Haggblomet al., 2007). Ammonium can

be aerobically oxidized to nitrate by slow growing microor-

ganisms in a two-step process known as nitrification.

Groundwater fuel contamination is generally character-

ized by a large chemical oxygen demand (COD). Therefore,

increased oxygen concentrations are positively correlated

with biodegradation rates. However, it is a challenge to supply

sufficient amounts of oxygen into contaminated aquifers due

to its low solubility. Enhanced biodegradation of BTEX (Borden

et al., 1997) and MTBE (Salanitro et al., 2000; Wilson et al., 2002)

has been achieved by active measures such as direct oxygen

injection, or passive measures like introduction of oxygen-

releasing compounds into the system. However, the long-

term efficiency of reactive barriers may be impacted by biofilm

clogging or precipitate formation (Scherer et al., 2000).

Aerobic ponds have found worldwide application in

municipal wastewater treatment; however, degradation of

fuel related contaminants in combination with ammonium is

poorly investigated in these systems (Thorneby et al., 2006).

Photosynthetic algae and bacteria as well physical aeration

are often used to support oxic degradation processes.

Retaining biomass in the system is a prerequisite for efficient

oxidation of contaminants and has been shown to enhance

biodegradation potential in benchtop scale experiments

(Korkut et al., 2006). Direct implementation of aerated

trenches in contaminated shallow aquifers could represent

a cost effective treatment technique, and has not yet been

described to our knowledge.

The aim of this study was to evaluate the effectiveness of

aerated trench systems for the reduction of BTEX and MTBE at

a field scale. Our model system simulates a scenario where

contaminated groundwater is passed through a treatment

facility at a constant recharge rate, directly from the aquifer,

before being released into the environment. The treatment

aims to reduce COD and contaminant levels in the effluent by

promoting aerobic biodegradation carried out by organisms

contained in contaminant-degrading biofilms. We test the

effectiveness of two different geotextiles, a polypropylene

fleece and a natural coconut fibre. Polypropylene fibres are

relatively inert with regard to extreme pH, salinity and

temperature conditions in comparison to polyester (Mathur

et al., 1994) and have been shown to support biofilm formation

of nitrifying bacteria (Korkut et al., 2006; McLean et al., 2000;

Takamizawa et al., 1993), whereas coconut textile is natural

and cost efficient. Here, the implementation of two aerated

trench systems after fourteen months of continuous opera-

tion is discussed in terms of contaminant degradation rates

and performance of the different textiles in the system.

2. Materials and methods

2.1. Site location and groundwater composition

The model treatment facility was set up next to a refinery

plant in Leuna, Germany. Due to spills, improper handling,

and war damage, the groundwater in this area is heavily

contaminated with high concentrations of ammonium, the

fuel additive MTBE, benzene, and considerable amounts of

iron (Table 1). Groundwater for processing was obtained from

a well located downstream from the refinery.

2.2. Setup of the aerobic pond system

The system consists of two parallel basins (basin 1 and basin 2),

each 5 m long, 1.15 m wide and 2.2 m deep (Fig. 1). The inflow

and effluent groundwaters pass through a gravel layer of high

porosity before entering or leaving the basin, simulating infil-

tration and exfiltration into or out of an engineered open

surface water body. The groundwater flow is regulated via

tubes located at the in and outflow of the basins at a depth of

2.15 m. Each system is separated into different compartments,

two permeable segments of 45 cm at the sides of the inflow and

outflow, filled with coarse gravel (8–16 mm) and an open water

surface area of 4.1 m in length. Five barriers of geotextiles direct

Fig. 1 – Cross section of one of the two parallel geotextile supported aerated treatment pond systems. Solid black circles

represent the location of oxygen sensors, white circles in black boxes indicate aeration modules, and dark lines in water

body show positions of geotextile barriers.

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 1 7 8 5 – 1 7 9 6 1787

the water flow path through the basin area with 30 cm high

openings at the bottom or top in alternating alignment. The

systems differ only in the geotextile materials used. System 1

contains coconut fibre mats (Angermunder Matten, Anger-

munde, Germany) as a natural material with a thickness of

1–2 cm and a density of 960 g/m2. System 2 contains synthetic

polypropylene fleece (Ludwig Kunststoffe GmbH, Berg,

Germany) with a thickness of 0.5 cm and a density of 320 g/m2.

Aeration modules (porous ceramic filter modules, Pall Schu-

macher GmbH, Germany) are located on the bottom of each

pond (before the 1st, 3rd and 5th barrier) and oxygen sensors

(TriOxmatic� 700 IQ sensors, WTW Wissenschaftlich-Techni-

sche Werkstatten GmbH, Weilheim, Germany) are suspended

in the water column at 60, 120 and 180 cm from the water

surface after the same barriers. One temperature sensor

(RM-Typ RL-5060-7, Rossel Messtechnik GmbH, Dresden,

Germany) per basin is positioned on the bottom of the tank, at

the midpoint of the water surface area. Two additional

geotextile barriers, consisting of 9 squares (20 � 20 cm) of

geotextiles positioned along the entire depth of the water

column, were installed for biofilm sampling and deployed in

front of the 2nd geotextile barrier of both trenches.

2.3. Process parameters

The experiment was initiated on 1st November 2007. Anoxic

groundwater was pumped into the systems at a constant rate

of 2.0 � 0.4 m3/day, resulting in a hydraulic retention time of

6.3 days. Outflow rates for both basins were 1.9 � 0.4 m3/day

and suggest that there are negligible water losses due to

evaporation.

Both basins were aerated with ambient air. Each aeration

module was operated separately and was regulated by taking

the mean oxygen content measured by the following three

oxygen sensors described in the previous paragraphs. The

oxygen content was measured at 15 min intervals. The

minimum oxygen concentration was held at 1 mg/L until

the 29th of November 2007. Afterwards, a gradient was

created ranging from 0 mg/L at the inflow area (1.5 m from the

inflow, aeration module switched off), 0.5 mg/L in the middle

of the basins (2.8 m from the inflow), to 1 mg/L in the rear

portion of the basins (4.2 m from the inflow).

2.4. Volatilization measurement

The volatilization losses for benzene and MTBE were

measured using a specially designed dynamic air chamber

(DAC) installed for one week on top of the basin with coconut

mats during two sampling campaigns in July and September

2008. The DAC was constructed from a steel frame covered

with a 200 mm ETFE foil. It is 1.8 m in height and its volume

accounts for around 10 m3. An adjustable air blower was

operated at 400 � 20 m3/h. Determination of benzene and

MTBE air concentrations was performed using active

sampling onto sorbent tubes (150 mg Tenax TA� plus 100 mg

Chromosorb106�) according to EPA Method TO-17 (U.S. Envi-

ronmental Protection Agency, 1999) at the inlet and outlet of

the DAC. Duplicate samples were taken hourly with

a sampling rate of 25 ml/min. The inflow and outflow

groundwater contaminant concentrations were monitored

simultaneously. More detailed information regarding design,

setup and function of the dynamic air chamber will be pre-

sented in a separate manuscript (Reiche et al., submitted for

publication).

2.5. Sampling procedure

Groundwater at the inflow and effluent from the two systems

was monitored periodically (2–4 times a month). After dis-

carding 500 ml of water from the sampling port, samples were

taken without headspace in amber glass bottles for organic

carbon, inorganic carbon and COD analysis, and in clear glass

bottles for biological oxygen demand (BOD) analysis. Samples

for cation measurement (Fe(II), K, Ca, Na, Mg, Mn, and P) were

acidified directly after sampling with HNO3 to a pH of 1–2. For

measurement of anions (Cl�, SO42�, NO2

�, NO3�, PO4

3�) and NH4þ,

samples were taken without headspace in 50 ml polyethylene

bottles (VWR International GmbH, Darmstadt, Germany). All

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 1 7 8 5 – 1 7 9 61788

samples were stored at 4 �C until measurement. Samples for

benzene and MTBE concentration profiles along distinct water

depths were taken in 1 L Duran� bottles (Schott, Germany),

closed with screw caps, and equipped with two tube connec-

tions. One was connected to an N72 gas pump (KNF Neuberger

GmbH, Freiburg, Germany) that flushed air into the bottle until

it reached the sampling depth. After the pump was discon-

nected, the bottle was filled through the vacant tube due to the

surrounding water pressure.

2.6. Analytical methods

The water temperatures of the inflowing groundwater, and at

the bottom central portion of each system, were measured at

15 min intervals (RM-Typ RL-5060-7, Rossel Messtechnik

GmbH, Dresden, Germany). The pH of the inflow and outflow

was monitored by a SensoLyt� 700 IQ system (WTW Wissen-

schaftlich-Technische Werkstatten GmbH, Weilheim,

Germany).

Cations were analysed following the DIN 38406-E5 protocol

(German Institute for Standardisation, 1983). Ammonium

quantification was performed photometrically at a detection

limit of 0.01 mg/L (EPOS Analyser 5060, Eppendorf AG,

Hamburg, Germany). The elements Ca, Fe, K, Mg, Mn, Na, and

P were analysed by inductively coupled plasma atomic emis-

sion spectroscopy (ICP-AES) using the a Spectro Ciros CCD

(SPECTRO Analytical Instruments GmbH, Kleve, Germany)

with detection limits of 0.03, 0.03, 0.2, 0.01, 0.02, 0.4, and

0.1 mg/L, respectively.

Anions were analysed following the DIN 38405 (German

Institute for Standardisation, 1979–2006) and the EN ISO

10304-2 (International Organization for Standardization, 1995)

protocols. Chloride, sulphate and nitrate were analysed using

the ion chromatograph DX500 (Dionex GmbH, Idstein,

Germany). The concentration of ortho-phosphate was ana-

lysed with a detection limit of 0.07 mg/L using an EPOS Ana-

lyser 5060 (Eppendorf AG, Hamburg, Germany).

The total organic carbon (TOC) was analysed according to

the DIN EN 1484 protocol (German Institute for Stand-

ardisation, 1997), in which TOC is the sum of purgeable

organic carbon (POC) and non-purgeable organic carbon

(NPOC). The BOD was analysed according to the DIN 38409 T52

protocol. The COD was analysed using the LCK 314 cuvette test

system (HACH LANGE GmbH, Dusseldorf, Germany) according

to the manufacturers instructions. Benzene concentrations

were determined by taking a 1 ml aliquot from a sample with

a sterile syringe, and added to 9 ml water (adjusted to pH 2.5

with sulphuric acid) in 20 ml GC-vials. Samples were analysed

on a gas-chromatograph equipped with a flame ionization

detector (GC-FID) (CP-3800 GC, Varian Inc., USA) described

elsewhere (Fischer et al., 2008).

The concentration of MTBE was determined by another

GC-FID system (HP 6890 GC, Agilent Technologies, Waldbronn,

Germany) described elsewhere (Rosell et al., 2010).

2.7. Determination of viable cell counts

Most probable number (MPN) counts for aerobic benzene and

MTBE metabolizing bacteria were determined using 10 repli-

cates and a 1:10 dilution carried out to 10�8 in 96 well plates

(8 � 12 wells, Nunc, Apogent, Denmark). Sample volumes of

500 ml from in and outflow water or 500 ml of extracted cell

solution from biofilms were inoculated into 4.5 ml of R2A

medium (Reasoner and Geldreich, 1985) for growth of total

aerobic bacteria and Brunner mineral medium supplemented

with benzene (Vogt et al., 2002) or into mineral salt medium

(Rohwerder et al., 2006) including vitamins, cobalt and MTBE

for growth of specific degraders, respectively. For contami-

nant specific degraders, plates were incubated in gas-tight jars

(Oxoid, United Kingdom) and 300 ml of pure benzene or MTBE

was added to the gas phase. The first and last columns of the

plates were filled with sterile media as controls, covered with

sterile lids, and incubated at 27 �C. Bacterial growth was

visually characterized in 96-well plates until no further

changes were observed.

2.8. Biofilm analysis

2.8.1. Confocal laser scanning microscopy (CLSM)Textile samples for CLSM-analysis of biofilms were taken

from the additional textile carriers. Samples of around 3 cm2

were transported while avoiding shaking at 4 �C in sterile

50 ml Teflon screwcap Duran bottles (Schott, Germany) filled

with 30 ml of 0.2 mm filtered water from the appropriate pond.

Textile sampling was performed in January, April, July, and

October 2008 at depths of 30, 120 and 180 cm. Both textiles

were examined for possible autofluorescence in order to select

the appropriate fluorochromes, mounted in 5 cm Petri dishes

and stained with SYTO� 9. The nucleic acid signal, the auto-

fluorescence and reflection were recorded. Images were

collected with a TCS-SP1, connected to an upright microscope

and controlled using confocal software, version 2.61 (Leica

Microsystems GmbH, Wetzlar, Germany). Samples were

examined using water immersible lenses (20� NA 0.5, 63� NA

0.9). Excitation was at 488 nm (reflection and Syto 9) and

633 nm (chlorophyll A autofluorescence). Emission signals

were detected at 480–495 nm (reflection), 500–550 nm (Syto 9)

and 650–750 nm (chlorophyll A). Raw image data sets are

presented as maximum intensity projections without any

further adjustments.

2.8.2. Extraction of cells from textilesThe cells were extracted from the textiles using a modified soil

extraction method (Riis et al., 1998). Briefly, 20 ml of a sterile

0.2% solution of Na4P2O7 (pH 8.5, adjusted with HCl) was added

to 0.5 g textile (dry weight) in 50 ml centrifugation tubes (VWR

International GmbH, Darmstadt, Germany). After vortexing for

2 min at maximum speed (Vortex Genie�2, Scientific Indus-

tries, inc., Bohemia, New York, USA), the sample was shaken

horizontally at 200 rpm for 30 min at room temperature and

homogenized by ice-cooled sonication (10 min, 50 W, 0.3 s

impulses). The residual textile was dried at 60 �C and weighed.

2.8.3. Degradation experimentsMicrocosm experiments with textile samples were performed

to evaluate the capacity of the biofilms to degrade benzene

and MTBE. Textile samples were taken from the additional

textile carriers located at a depth of 170 cm and transported at

4 �C in 50 ml centrifugation tubes. For benzene degradation,

1 cm2 textile samples were inoculated in duplicate into 240 ml

Table 2 – Average oxygen content [mg/L] during the test period from 1st of December 2007 until the 31st of December 2008.

Depth Distance to inflow [m]

1.5 2.8 4.2

Average SD Average SD Average SD

Basin 1 Coconut 60 cm 0.3 1.0 0.5 0.8 1.5 1.2

120 cm 0.3 0.7 0.5 0.6 1.0 0.7

180 cm 0.5 0.9 0.5 0.7 0.8 0.7

Basin 2 Polypropylene 60 cm 0.6 1.4 0.5 0.7 1.5 1.3

120 cm 0.5 0.9 0.4 0.5 1.2 0.9

180 cm 0.2 0.5 0.5 0.6 0.7 0.7

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 1 7 8 5 – 1 7 9 6 1789

cultivation bottles filled with 55 ml mineral salt medium

spiked with benzene to a final concentration of 50 mg/L. The

bottles were sealed with Teflon caps and incubated on

a shaker (85 rpm, 25 �C).

For the MTBE degradation experiments, 2 cm2 textile

samples were inoculated in triplicate into 240 ml cultivation

bottles filled with 60 ml of mineral salt medium (Rohwerder

et al., 2006) including vitamins, cobalt, and around 160 mg/L

MTBE. The bottles were sealed with Teflon caps and incubated

on a shaker (130 rpm, 30 �C). Concentrations of MTBE were

analysed periodically in duplicate.

2.9. Microbial community structure analysis

Denaturing gradient gel electrophoresis (DGGE) was carried out

using the DCode Universal Mutation Detection System (Biorad,

Munich, Germany) and a modified protocol described else-

where (Imfeld et al., 2008; Nikolausz et al., 2008). For DNA

extraction, 240 ml water samples were filtered onto cellulose

acetate filters (pore size 0.2 mm, diameter 45 mm, Sartorius AG,

Gottingen, Germany) and DNA was isolated from the filters or

directly from 0.2 g of textile using the FastDNA� Kit (MP

Biomedicals, Irvine CA) and the FastPrep� Instrument (Savant

Instruments, Inc. Holbrook, New York. Bacterial specific

primers27F (Lane, 1991) and1378R (Heuer etal., 1997) wereused

for PCR. Nested PCR was carried out using the universal primers

GC-968-GC (Nubel et al., 1996) and Univ1378R. 16S PCR products

(500 ng) weredirectly applied to 8% (wt/vol) polyacrylamide gels

containing linear denaturing gradients of between 30 and 60%

urea/formamide (7 M urea and 40% formamide (vol/vol) as 100%

denaturants) andseparated by electrophoresisat 60 V for16 h at

constant temperature (60 �C) in a 1� TAE buffer. The gel was

silver-stained according to Bassam et al. (1991) and scanned

with an UMAX Astra 2400s (Umax Systems GmbH, Willich,

Germany) flatbed scanner. Similarity matrices, based on band

abundances, were produced using the GelComparII software

(Applied Maths, Kortrijk, Belgium). Dendrograms were gener-

ated using the unweighted pair group method using arithmetic

means (UPGMA) algorithm.

3. Results

3.1. Physico-chemical conditions within the pond

Dissolved oxygen levels at 1.5 m from the inflow reached

concentrations up to 0.6 mg/L (Table 2), even when the

aeration was switched off in this zone of the basin. Addi-

tionally, diurnal fluctuations in oxygen levels (not shown)

indicate oxygen production by algae during the day and

consumption over night. This diurnal pattern was also

observed on the textiles (see below). The pH of the inflow

groundwater was constant at 7.1 � 0.9, measured for the

duration of the experiment at 15 min intervals, and 7.5� 0.5 in

both ponds. The average annual temperature of the inflowing

groundwater was 12� 3 �C. In the two basins, the temperature

was correlated with the ambient air temperature (R2 of 0.7,

data not shown) and ranged from 3 �C in the winter period to

25 �C in the summer period (see Supplementary information).

3.2. Contaminant removal

The monthly average inflow concentrations of benzene were

between 16.6 mg/L and 23.3 mg/L. Benzene outflow concen-

trations varied but were predominantly found to be around

1 mg/L, a value below the German guidelines for drinking

water quality (DVGW, 2001), and consistently remained

under the guideline value of 10 mg/L set by the World Health

Organization (WHO, 2003) for drinking water quality (see

supplementary information). The following calculations are

based on a loading rate of 2 m3/day. Each system removed

33.2–46.6 g benzene per day or 5.8–8.1 g/m2/day. Horizontal

concentration profiles throughout the system showed that

the largest concentration decrease (for benzene and also for

MTBE) took place within the front part of the system, within

a distance of up to 2.8 m from the inflow (Fig. 2). In micro-

cosm experiments with fresh textile samples, benzene was

immediately degraded without lag-phase (Fig. 3). Microcosm

experiments showed that sorption of benzene and MTBE to

the geotextiles in the system reached equilibrium after a few

days (data not shown), and can therefore be considered to be

negligible.

During the first year of operation, the residual MTBE

concentration in the effluent from both systems was consis-

tently 60–80% of the inflow concentration. Hence, each basin

removed 1.6–3.1 g MTBE per day or 0.3–0.5 g/m2/day. Beginning

in September ’08, after around 317 days of operation, both

systems started to develop differently. In the basin with

coconut fibre material, MTBE was increasingly degraded

(Fig. 4). This observation was confirmed by microcosm

experiments with coconut fibre collected in August 08, where

MTBE was degraded while no degradation was seen in

microcosms with polypropylene textile samples (Fig. 5).

However, the overall reduction in MTBE concentrations were

Inflow

1.5 m

- 60 c

m

1.5 m

- 120

cm

1.5 m

- 180

cm

2.8 m

- 120

cm

4.2 m

- 120 c

m

Outflow

mg/

L

0

5

10

15

20

25

Benzene concentration - coconutBenzene concentration - polypropylene MTBE concentration - coconut MTBE concentration - polypropylene

Fig. 2 – Benzene and MTBE concentration gradients

analysed in April 08. Samples were taken from the inflow,

from different sections along the water flow path of each

basin and from the outflow. In x axis, first number

represents the distance to the inflow and second number

the sampling depth from the water surface.

Novem

ber 07

Decem

ber 0

7

Janu

ary 08

Februa

ry08

March 08

April 0

8

May08

June

08

July

08

Augus

t 08

Septem

ber 0

8

Octobe

r 08

Novem

ber 08

Decem

ber 08

MTB

E co

ncen

tratio

n %

of i

nflo

w a

vera

ge

0

20

40

60

80

100

outflow coconutoutflow polypropylene

Fig. 4 – Percentage of residual MTBE in basin outflows,

compared with the average inflow concentration of each

month. Error bars represent standard deviations for 2–4

samplings.

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 1 7 8 5 – 1 7 9 61790

not sufficient to reach the 20–40 mg/L EPA drinking water

advisory standard (US Environmental Protection Agency,

1997).

A slight removal of 1.7–10.7% ammonium was observed in

the outflow of both basins (see supplementary information)

over the entire course of the experiment, while the largest

removal occurred in February and March 2008. An average

removal rate of 2.0–12.6 g ammonium per day, or 0.4–2.2 g/m2/

day, was observed for each basin.

time [h]0 5 10 15 20 25

resi

dual

ben

zene

[%]

0

20

40

60

80

100

120

coconut textile polypropylene fleece control without textile

Fig. 3 – Microcosm experiment with textile pieces sampled

in June 08. Textiles were incubated as duplicates in

mineral salt medium under oxic conditions with benzene

added as carbon source and incubated at 25 8C. Benzene

amended medium was used as a control.

3.3. Volatilization of contaminants

The volatilization rate of benzene and MTBE for the basin

containing the coconut fibre mats, analysed in July 2008, was

254 mg/min and 1590 mg/min, respectively. This implies that

about 1% and 53% of the benzene and MTBE losses from the

system were due to volatilization, respectively. During the

second campaign in September 2008, the percentage of mass

loss to volatilization amounted to 0.4% for benzene and 48%

for MTBE (Table 3).

3.4. Microbiological observations

Average bacterial numbers at the inflow and outflow from the

system were 1� 103 MPN/ml and 1� 106 MPN/ml, respectively.

time [days]0 20 40 60 80 100

resi

dual

MTB

E [%

]

0

20

40

60

80

100

120

coconut textilepolypropylene fleececontrol without textile

Fig. 5 – Microcosm experiment with textile pieces sampled

in August 08. Textiles were incubated as triplicates in

mineral salt medium under aerobic conditions with MTBE

added as a carbon source and incubated at 25 8C. Medium

with added MTBE was used as control.

Table 3 – Volatilization rates and calculated percentage of mass loss due to volatilization for the basin with coconut fibrematerial. Values are from two sampling campaigns in July and September 2008, standard deviation values (SD) areprovided.

Date of samplingcampaign

Benzene MTBE Temperatures

Volatilizationrate (mg/min)

SD(mg/min)

Volatilizationpart of mass

loss (%)

SD(%)

Volatilizationrate (mg/min)

SD(mg/min)

Volatilizationpart of

mass loss (%)

SD(%)

Water(�C)

SD(�C)

DACair (�C)

SD(�C)

July (n ¼ 3) 254 115 1 0.5 1590 200 53 13 18.5 0.17 33.5 9.4

September (n ¼ 5) 120 57 0.4 0.2 1480 175 48 7 14.5 0.65 17.5 4.6

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 1 7 8 5 – 1 7 9 6 1791

MPNs for aerobic benzene metabolizing bacteria ranged from

3.3 � 101 to 9.2 � 102 MPN/ml in the inflow, and 2.9 � 102 to

1.1 � 104 MPN/ml (Fig. 6 a) in the effluent. Biofilms were

observed by CLSM on polypropylene and coconut textiles,

during the entire sampling period, beginning with the first

sampling in January (see Supplementary data). Algae were

predominantly found in the uppermost, sunlight-exposed

zones but were present along the entire surface of the textile

in lower concentrations. The uppermost region of the poly-

propylene textile developed a thick layer of algae and cyano-

bacteria, but this was not observed on the coconut fibres.

Precipitates were found predominantly on textiles sampled

from deep zones. On both textiles, high numbers of cultivable

aerobic bacteria were detected. The total amount of cells per

cm2 or gram textile is difficult to compare to the amount of

cells per cm3 in the water. However, by comparing the MPN

results for cultivable aerobic and benzene metabolizing

bacteria, we conclude that the percentage of benzene

degrading bacteria on textiles is higher than the percentage in

the outflow from the trenches (Fig. 6b). The cells counts from

mineral salt media supplemented with MTBE as a carbon

source could not be clearly distinguished. Only minimal

growth was observed, and might have been caused only by

februa

rymarc

hap

rilmay jun

e july

augu

st

septe

mber

octob

er

nove

mber

dece

mber

MPN

/ml o

r MPN

/cm

² tex

tile

1e+1

1e+2

1e+3

1e+4

1e+5

1e+6

1e+7

1e+8

1e+9

Textile coconut Textile polypropylene Inflow Outflow coconut Outflow polypropylene

a

Fig. 6 – (a) MPN of benzene metabolizing aerobic bacteria in yea

compared to total aerobic bacteria.

metabolization of MTBE or by growth on residual benzene

from the inoculum. Microbial community compositions from

water and textile samples were compared using DGGE, and

analysed by clustering using Pearson correlation and UPGMA

methods. Complex community patterns were observed for

basin water and textile samples and were significantly altered

compared to inflow groundwater (Fig. 7). Water samples from

the basins were very similar, whereas textile samples differed

from each other and also from all water samples.

3.5. Total organic carbon (TOC), chemical oxygendemand (COD), and biological oxygen demand (BOD)

The system showed overall reductions of 61% for COD, 57–59%

for TOC, and 71–73% for BOD. These parameters together with

the calculated electron equivalents for the COD values and the

summarized average MTBE, benzene, iron(II) and ammonium

concentrations are shown (Table 4). To facilitate better

comparison, the organic carbon content based on the MTBE

and benzene concentrations was calculated and individual

POC/NPOC values are provided for inflow and outflow

groundwaters. The TOC fractions in the inflow were 29.5% and

70.5% for NPOC and POC, respectively. The NPOC showed no

februa

rymarc

hap

rilmay jun

e july

augu

st

septe

mber

octob

er

nove

mber

dece

mber

% b

enze

ne d

egra

ders

0.00.20.40.60.8

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

100.0

b

r 2008. (b) Percentage of benzene metabolizing bacteria

Fig. 7 – Cluster analysis of DGGE profiles of water and textile samples, collected in August 08; Lanes were clustered by

Pearson correlation and UPGMA.

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 1 7 8 5 – 1 7 9 61792

significant decrease whereas the POC decreased up to 92%, and

equals the calculated sum of MTBE and benzene in mg C/L. In

fact, this reduction is mainly due to benzene removal. By

comparing calculated electron equivalents, we can also

conclude that the differences in COD, are mainly due to

reductions in benzene concentrations. The removal of iron(II)

accounts for 2% of the difference in electron equivalents. The

oxygen demand for ammonium oxidation was not assessed by

the COD measurement and thus did not contribute to this

parameter. Calculated electron equivalents necessary for

oxidation of the ammonium to nitrate in the inflow ground-

water (26 � 2 mM) are nearly double the amount of oxygen

compared to the COD. BOD 5 and BOD 10 values were

comparable to the results obtained from the TOC and COD

measurements. Regarding total values, the BOD measured was

50–58% and 34–37% of the COD at the inflow and outflows,

respectively.

Table 4 – Total organic carbon (TOC), chemical oxygendemand (COD) and biological oxygen demand (BOD)values of the inflowing and outflowing groundwater ofthe two basins. Numbers represent averages from 14months of operation with standard deviations (±SD) for36–39 measurements. For comparison, electronequivalents for MTBE, benzene, ammonium and iron areshown as well.

Units Inflow Outflowcoconut

Outflowpolyprop.

TOC (POC þ NPOC) mg C/L 42 � 3 17 � 4 18 � 3

NPOC mg C/L 12 � 2 15 � 2 16 � 3

POC mg C/L 29 � 2 3 � 3 2 � 2

MTBE þ benzene mg C/L 21 � 3 1.7 � 0.6 1.7 � 0.4

COD mg/L 119 � 7 46 � 13 47 � 9

COD eeq 14.9 � 0.8 6 � 2 6 � 1

MTBE þ benzene eeq 9 � 1 0.9 � 0.3 0.9 � 0.2

Iron(II) eeq 0.29 � 0.07 0.07 � 0.05 0.09 � 0.04

Ammonium eeq 26 � 2 25 � 2 25 � 2

BOD 5 mg/L 60 � 10 17 � 10 16 � 8

BOD 10 mg/L 69 � 9 20 � 11 19 � 9

#eeq: electron equivalents [mM].

#POC: purgeable organic carbon.

#NPOC: non-purgeable organic carbon.

4. Discussion

4.1. Contaminant removal

Biodegradation of contaminants in aquifers is often rate-

limited by the availability of oxygen. In this study, we inves-

tigated the direct implementation of an aerated treatment

system in a shallow aquifer to stimulate aerobic degradation

processes. A transition zone between an anoxic compartment

in the inflow and an oxic compartment in the outflow area

was simulated by adjusting an oxygen gradient from 0 mg/L in

the inflow, 0.5 mg/L in the middle, to 1 mg/L in the outflow

areas of the systems. Low oxygen concentrations were

selected to minimize volatilization losses and maximize

biodegradation rates for the contaminants. Our system was

designed to minimize clogging and may be used for ground-

waters unsuitable for conventional reactive barrier treat-

ments. After treatment using our methods, water may be

amended with additional electron acceptors and released into

an aquifer or discharged into a traditional drainage system.

4.1.1. Organic carbon and oxygen demand reductionBoth systems showed nearly complete and constant benzene

removal down to drinking water limits. This indicates that the

system is stable and operates independent from water

temperature fluctuations (from 3 to 25 �C). The reduction in

TOC and BOD was caused by the removal of purgeable organic

carbon whereas the amount of non-purgeable organic carbon

remained stable. The removed purgeable organic carbon

content originated from MTBE and benzene, as confirmed by

calculations of the organic carbon amounts and the respective

electron equivalents (Table 4). The organic carbon responsible

for residual TOC and COD was not studied in detail but likely

originates from humic substances, as well as allochthonous

and autochthonous organic carbon like proteins, peptides,

polysaccharides, and pedogenic or aquagenic refractory

compounds (Oliveira et al., 2006). However, it is important to

note that benzene may be oxidized incompletely in the COD

measurement due to the low boiling point temperature

(Dedkov et al., 2000). This can lead to a slight underestimation

of the COD in the inflow. Other volatile organic contaminants

such as toluene, xylene, trimethylbenzene and naphthalene

were considered negligible due to low average inflow

concentrations of up to 0.4 mg/L and complete removal in the

outflow (data not shown). Iron(II) was not completely absent

in the outflows. Inflowing iron(II) was probably not completely

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 1 7 8 5 – 1 7 9 6 1793

oxidized in the basins due to the low oxygen concentrations,

or iron(II) might also be produced by iron reduction in anoxic

zones of the system, e.g. deeper biofilms layers. Regarding the

calculated electron equivalents, the removal of iron(II) is of

minor importance for the reduction of the COD in this system

(Table 4).

4.1.2. BenzeneEffective benzene biodegradation was expected, since this

pollutant has been degraded in environmental systems even

under hypoxic conditions (Agteren et al., 1998; Yerushalmi

et al., 2002) and treatment efficiencies for aerobic bioreactors

up to 100% have been described. However, most of the studies

did not take into account the role of stripping of VOC due to

aeration (Farhadian et al., 2008). In our systems, benzene was

almost completely degraded in the first third of the basin,

a zone which had not been aerated. Presumably, the oxygen

needed for benzene breakdown was likely provided by the

photosynthetic activity of algae as well as from oxygen

enriched water circulating back from the next aeration

modules. In addition, diurnal temperature fluctuations of the

ambient air may have triggered convective flow and transport

of oxygen to the bottom of the system, as reported by Schmid

et al. (2005). Microcosm experiments showed that sorption of

benzene and MTBE to the textile material in this system was

negligible (data not shown). Thus, benzene was rapidly bio-

degraded in our system, and only small amounts were vola-

tilized to the atmosphere. Additionally, this indicates that the

maximum treatment capacity for this system was not reached

with the actual groundwater load of benzene. System perfor-

mance for benzene was similar for both geotextiles used.

4.1.3. MTBEMTBE biodegradation was by far not as effective as benzene

biodegradation. MTBE is more resistant to enzymatic attacks

due to its tertiary carbon atom and the ether bond (Davidson

and Creek, 2000). However, an MTBE-degrading b-proteo-

bacterium strain L108, capable of growing on fuel oxygenate

ethers as the sole source of carbon and energy, was previously

isolated from the Leuna site (Rohwerder et al., 2006) and

demonstrates a natural attenuation potential for MTBE at this

location. Volatilization analyses showed that 48–53% of the

removed MTBE was stripped. The residual amount might be

biodegraded, although biodegradation could not be demon-

strated in vitro during the first 6 months of system operation.

However, from August ‘08, the coconut textile basin showed

increasing MTBE removal rates of up to 67% in December ‘08

(Fig. 4). Since the Henry coefficient is influenced by tempera-

ture, the higher amount of removal in the late summer period

could be related to the higher water temperature, and there-

fore to a higher volatilization rate of MTBE. However, in this

case, both systems should have been influenced in the same

manner, a result not observed. Indeed, MTBE biodegradation

was confirmed by laboratory microcosm experiments per-

formed with coconut textile samples in August ‘08, around 8

months after operation was initiated, while no biodegradation

was observed for polypropylene after 100 days of incubation

(Fig. 5). Maximum growth rates (mmax) for aerobic MTBE

degraders are an order of magnitude lower than reported for

BTEX and nitrifying organisms (Waul et al., 2007), and this

may have led to extensive lag-phases for MTBE degraders or

even MTBE degraders being out-competed in some sections of

the system. Further research is needed to determine if auxil-

iary substances provided by the coconut material which might

consist of cellulose, hemicellulose, lignin, pectin, waxes, and

water soluble substances (Bledzki and Gassan, 1999) enhance

or stimulate the growth of aerobic MTBE degraders.

4.1.4. AmmoniumAmmonium concentrations decreased only slightly from an

average concentration of 59 mg/L in the inflow to 56 mg/L in

the outflows (Table 1), indicating inhibited or absent nitrifi-

cation during the first 14 months of operation. Nitrification is

a chemolithoautotrophic process that is performed in two

sequential oxygen demanding steps. Using a representative

measurement of yield and oxygen consumption by Nitro-

somonas and Nitrobacter, (i) the oxidation of ammonium to

nitrite requires 3.16 mg O2/mg NH4–N or 1.38 mol O2/mol NH4þ

and (ii) the oxidation of nitrite to nitrate requires 1.11 mg O2/

mg NO2–N or 0.49 mol O2/mol NO2� (Ahn, 2006), resulting in

a demand of 6.1 mM O2 or 24.3 mM electron equivalents

required for complete oxidation of 59 mg/L ammonium. This

value is in agreement with our calculation of 26� 2 mM for the

theoretical requirement of electron equivalents (Table 4). In

laboratory experiments, it was shown that oxygen concen-

trations higher than 0.1–0.2 mg/L support growth of ammo-

nium oxidizers. Even at lower oxygen concentrations down to

0.05 mg/L, nitrification was observed for a certain time (Abe-

liovich, 1987). Hence, the oxygen concentrations in the basins

(0.5–1 mg/L) should have been sufficient for ammonium

oxidation, as well as the observed pH range of 7.1 in the

inflowing groundwater and 7.5 in the effluent from the basins

(Ahn, 2006). However, nitrifying bacteria such as Nitrosomonas

and Nitrobacter are characterized by low growth yields of

0.15 mg cells/mg NH4–N and 0.02 mg cells/mg NO2–N oxidized,

respectively (Ahn, 2006). It was described that a high organic

loading can result in decreased nitrification, probably due to

faster growing heterotrophic bacteria dominating the surface

of the biofilm, and leads to oxygen limitations for the nitri-

fying bacteria growing deeper inside the biofilm (Elenter et al.,

2007). We speculate that low growth rates for nitrifiers in

combination with competition by heterotrophic bacteria,

suboptimal growth conditions, and possibly grazing of nitri-

fying bacteria by higher trophic level organisms (e.g. protozoa)

might be responsible for the low nitrification rate observed in

the system thus far.

4.1.5. PhosphateAround 1.4 mg/L phosphate and 0.9 mg/L phosphorous

entered the system on average. Soil microbial biomass ratios

were suggested to be related to Redfield ratios of 60:7:1 for

C:N:P (Cleveland and Liptzin, 2007). Based on these ratios and

considering reported growth yields for benzene (Reardon

et al., 2000), MTBE (Muller et al., 2007), ammonium, and nitrite

(Ahn, 2006) of 1.2 g/g, 0.869 g/g, 0.15 g/g and 0.02 g/g, respec-

tively, the system was not considered to be limited by phos-

phorous (see supplementary information). The theoretical

phosphorous uptake of 0.9 mg P/L matches the inflow

concentration of phosphorous and may represent an over-

estimation because other elements, e.g. oxygen, hydrogen,

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 1 7 8 5 – 1 7 9 61794

and sulphur are not considered in the biomass. Consequently,

phosphorous was always detected in the outflow of the

system at concentrations of 0.5 � 0.2 mg/L. The removal of up

to 0.4 mg/L phosphorous in the system might be a result of

biomass formation. However, the calculated amount of

phosphorus necessary for complete degradation of benzene,

MTBE and ammonium is 1.35 mg/L (see supplementary

information) and indicates that phosphorus might become

a growth-limiting factor during higher overall degradation

rates.

4.2. Investigation of biofilms and microbial community

The textiles in the aerated pond system enhanced degradation

rates for organic contaminates by supporting the develop-

ment of contaminant-degrading biofilms.

Significant biofilms were formed on both textile types, as

observed by repeated CLSM-analysis of textile samples. Algae

and bacteria were found along all sampling depths throughout

the experimental period. However the upper portion was

dominated by algae and cyanobacteria, and the middle and

bottom portions by bacteria (see supplementary data). Freshly

sampled textiles showed no lag-phase for aerobic benzene

degradation in laboratory microcosm experiments, demon-

strating the presence of active benzene degrading microor-

ganisms in textile biofilms within the system (Fig. 3). Up to

76% of the total MPNs from the system were accounted for in

the selective benzene enrichments, indicating the presence

and retention of a selectively-enriched benzene degrading

community. The two geotextiles showed similar benzene

degrading performance, suggesting that it is not the nature of

the material but rather the surface area for biofilm formation

that is important for the performance of the system.

Furthermore, we can extrapolate that larger areas of textiles

for biofilm growth would generally enhance the performance

of the system with regard to benzene degradation rates (Fig. 2).

The inflowing anoxic groundwater contained very low

numbers of cultivable aerobic bacteria (around 103 MPN/mL).

However, the number of aerobic benzene metabolizing

bacteria at the inflow was only tenfold lower (Fig. 6), demon-

strating that the inflow is a continuous source of aerobic

benzene degraders for the trench system but also that an

enrichment of aerobic benzene degraders had already taken

place in the groundwater.

The microbial communities from the inflowing ground-

water, the basin water, and the textile samples changed due to

the different environmental parameters within each

compartment (Fig. 7). The inflowing groundwater is anoxic,

and major parts of the basins showed minimum oxygen

concentrations between 0.0 and 1.0 mg/L. Thus, the ground-

water community may be comprised of facultative and strict

anaerobes, whereas the basin communities may be domi-

nated by facultative or obligate aerobes. Corresponding to this

hypothesis, the groundwater community was significantly

different to the basin community. Furthermore, the commu-

nities colonizing the textiles showed little similarity to the

water communities. This confirms the results from the MPN

experiments showing that the textiles support the formation

of a specialized community for contaminant degradation and

therefore can enhance the biodegradation capacity of the

community. Community profiles from the mats also differed

and suggest that specific surface properties of coconut and

polypropylene support the growth on distinct biofilms. For

example, the density of the polypropylene fleece compared to

the coconut fibre material could lead to isolated anaerobic

pockets within the material. Further bacterial community

analysis is needed to confirm this hypothesis.

5. Conclusion

The following conclusions can be drawn from the current

study:

1. Biofilms developed on both textile materials within two

months after initiating system operation. Results from

microbial community analysis and laboratory microcosm

experiments indicate that the development of a distinct

microbial community, adapted to contaminant degrada-

tion, on the surfaces of both tested geotextile materials was

achieved.

2. Benzene was effectively biodegraded from 20 mg/L inflow

concentration to less than 2 mg/L (99.9%) at the outflow of

both pond variants, indicating that surface provided for

growth of biofilms is a major factor for improving biodeg-

radation rates.

3. During the first year, inflow concentration reductions from

4 mg/L to 2.5 mg/L (on average 38% removal) were observed

for MTBE and were not sufficient to reach drinking water

quality standards; MTBE also showed high stripping rates

(48–53%). After 8 months of system operation, the basin

equipped with coconut fibre textile started to increasingly

degrade MTBE, an observation confirmed in laboratory

microcosm experiments.

4. Low degradation rates were observed for ammonium,

indicating a limitation or inhibition by system conditions or

outcompeting of nitrifyiers by heterotrophic organisms.

5. The COD was significantly depleted in the system from

119 mg/L in the inflow to 46–47 mg/L in the outflows (61%

removal) showing the potential of the trench to improve

water quality.

6. The textile materials, coconut fibre mats and polypropylene

fleece, are both suitable for support and development of

contaminant-degrading bacterial biofilms and could be an

option for low-cost enhancement of degradation capacities

of contaminant treatment ponds. Studies on the long-term

durability as well as the long-term performance of both

materials are currently in progress.

Acknowledgements

We thank Ute Kuhlicke for operating the CLSM, Krista Ver-

steeg for technical assistance in the MPN series, Marcell

Nikolausz for help and advice with DGGE analysis, Peter

Mosig, Stefan Kukla and Francesca Loper for technical assis-

tance in the system operation and sampling, and the depart-

ment of Analytical Chemistry as well as Grit Weichert of the

Centre for Environmental Biotechnology for analytical

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 1 7 8 5 – 1 7 9 6 1795

support. We acknowledge the suggestions and recommenda-

tions of two unknown reviewers, Brandon E. L. Morris and

Monica Rosell for greatly improving the quality of the manu-

script. This work was supported by the Helmholtz Centre for

Environmental Research – UFZ in the scope of the SAFIRA II

Research Programme (Revitalization of Contaminated Land

and Groundwater at Megasites, subproject ‘‘Compartment

Transfer - CoTra’’).

Appendix.Supplementary data

The supplementary materials can be viewed at doi:10.1016/j.

watres.2009.12.002.

r e f e r e n c e s

Abeliovich, A., 1987. Nitrifying bacteria in wastewater reservoirs.Applied and Environmental Microbiology 53 (4), 754–760.

Agteren, M.H.v., Keuning, S., Janssen, D.B., 1998. Handbook onBiodegradation and Biological Treatment of HazardousOrganic Compounds. Kluwer Academic Publishers, TheNetherlands.

Ahn, Y.H., 2006. Sustainable nitrogen eliminationbiotechnologies: a review. Process Biochemistry 41 (8),1709–1721.

Bassam, B.J., Caetanoanolles, G., Gresshoff, P.M., 1991. Fast andsensitive silver staining of DNA in polyacrylamide gels.Analytical Biochemistry 196 (1), 80–83.

Bledzki, A.K., Gassan, J., 1999. Composites reinforced withcellulose based fibres. Progress in Polymer Science 24 (2),221–274.

Borden, R.C., Goin, R.T., Kao, C.-M., 1997. Control of BTEXmigration using a biologically enhanced permeable barrier.Ground Water Monitoring & Remediation 17 (1), 70–80.

Christensen, T.H., Kjeldsen, P., Bjerg, P.L., Jensen, D.L.,Christensen, J.B., Baun, A., Albrechtsen, H.-J., Heron, G., 2001.Biogeochemistry of landfill leachate plumes. AppliedGeochemistry 16 (7–8), 659.

Cleveland, C.C., Liptzin, D., 2007. C:N:P stoichiometry in soil: isthere a ‘‘Redfield ratio’’ for the microbial biomass?Biogeochemistry 85 (3), 235–252.

Davidson, J.M., Creek, D.N., 2000. Using the gasoline additiveMTBE in forensic environmental investigations.Environmental Forensics 1 (1), 31–36.

Dedkov, Y.M., Elizarova, O.V., Kel’ina, S.Y., 2000. Dichromatemethod for the determination of chemical oxygen demand.Journal of Analytical Chemistry 55 (8), 777–781.

DVGW, D.V.d.G.-u.W.e.V., 2001. Verordnung zur Novellierung derTrinkwasserverordnung vom 21. Mai 2001, Bonn.

Elenter, D., Milferstedt, K., Zhang, W., Hausner, M., Morgenroth, E.,2007. Influence of detachment on substrate removal andmicrobial ecology in a heterotrophic/autotrophic biofilm. WaterResearch 41 (20), 4657–4671.

Farhadian, M., Duchez, D., Vachelard, C., Larroche, C., 2008.Monoaromatics removal from polluted water throughbioreactors – a review. Water Research 42 (6–7), 1325–1341.

Ferreira, N.L., Malandain, C., Fayolle-Guichard, F., 2006. Enzymesand genes involved in the aerobic biodegradation of methyltert-butyl ether (MTBE). Applied Microbiology andBiotechnology 72 (2), 252–262.

Fischer, A., Herklotz, I., Herrmann, S., Thullner, M., Weelink, S.A.B.,Stams, A.J.M., Schlomann, M., Richnow, H.H., Vogt, C., 2008.Combined carbon and hydrogen isotope fractionationinvestigationsfor elucidatingbenzenebiodegradationpathways.Environmental Science & Technology 42 (12), 4356–4363.

Foght, J., 2008. Anaerobic biodegradation of aromatichydrocarbons: pathways and prospects. Journal of MolecularMicrobiology and Biotechnology 15 (2–3), 93–120.

Fortin, N.Y., Morales, M., Nakagawa, Y., Focht, D.D.,Deshusses, M.A., 2001. Methyl tert-butyl ether (MTBE)degradation by a microbial consortium. EnvironmentalMicrobiology 3 (6), 407–416.

German Institute for Standardisation, 1979–2006. DeutscheEinheitsverfahren zur Wasser-, Abwasser- undSchlammuntersuchung – Anionen. (German StandardMethods for Analyzing Water, Wastewater and Sludge –Anions). Deutsches Institut fur Normung e.V. (DIN), Berlin,Germany.

German Institute for Standardisation, 1983. DeutscheEinheitsverfahren zur Wasser-, Abwasser- undSchlammuntersuchung – Kationen. (German StandardMethods for Analyzing Water, Wastewater and Sludge –Cations). Deutsches Institut fur Normung e.V. (DIN), Berlin,Germany.

German Institute for Standardisation, 1997. Wasseranalytik –Anleitungen zur Bestimmung des gesamten organischenKohlenstoffs (TOC) und des gelosten organischen Kohlenstoffs(DOC). (German Standard Method for Measurement of TotalOrganic Carbon and Dissolved Organic Carbon). DeutschesInstitut fur Normung e.V. (DIN), Berlin, Germany.

Haggblom, M.M., Youngster, L.K.G., Somsamak, P., Richnow, H.H.,Laskin, A.I., Sariaslani, S., Gadd, G.M., 2007. Advances inApplied Microbiology. Academic Press, p. 1.

Heuer, H., Krsek, M., Baker, P., Smalla, K., Wellington, E.M.H.,1997. Analysis of actinomycete communities by specificamplification of genes encoding 16S rRNA and gel-electrophoretic separation in denaturing gradients. Appliedand Environmental Microbiology 63 (8), 3233–3241.

Imfeld, G., Nijenhuis, I., Nikolausz, M., Zeiger, S., Paschke, H.,Drangmeister, J., Grossmann, J., Richnow, H.H., Weber, S.,2008. Assessment of in situ degradation of chlorinatedethenes and bacterial community structure in a complexcontaminated groundwater system. Water Research42 (4–5), 871.

International Organization for Standardization, 1995. WaterQuality – Determination of Dissolved Anions by LiquidChromatography of Ions – Part 2: Determination of Bromide,Chloride, Nitrate, Nitrite, Orthophosphate and Sulfate inWaste Water. ISO, Geneva, Switzerland.

Korkut, E.N., Martin, J.P., Yaman, C., 2006. Wastewater treatmentwith biomass attached to porous geotextile baffles. Journal ofEnvironmental Engineering-Asce 132 (2), 284–288.

Lane, D.J., 1991. 16S/23S rRNA Sequencing, vol. 6. Wiley&Sons Ltd,Chichester, England.

Mathur, A., Netravali, A.N., Orourke, T.D., 1994. Chemical agingeffects on the physio-mechanical properties of polyester andpolypropylene geotextiles. Geotextiles and Geomembranes 13(9), 591–626.

McLean, B.M., Baskaran, K., Connor, M.A., 2000. The use of algal-bacterial biofilms to enhance nitrification rates in lagoons:experience under laboratory and pilot-scale conditions. WaterScience and Technology 42 (10–11), 187–194.

Muller, R.H., Rohwerder, T., Harms, H., 2007. Carbon conversionefficiency and limits of productive bacterial degradation ofmethyl tert-butyl ether and related compounds. Applied andEnvironmental Microbiology 73 (6), 1783–1791.

Nikolausz, M., Kappelmeyer, U., Szekely, A., Rusznyak, A.,Marialigeti, K., Kastner, M., 2008. Diurnal redox fluctuation

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 1 7 8 5 – 1 7 9 61796

and microbial activity in the rhizosphere of wetland plants.European Journal of Soil Biology 44 (3), 324–333.

Nubel, U., Engelen, B., Felske, A., Snaidr, J., Wieshuber, A.,Amann, R.I., Ludwig, W., Backhaus, H., 1996. Sequenceheterogeneities of genes encoding 16S rRNAs in Paenibacilluspolymyxa detected by temperature gradient gelelectrophoresis. Journal of Bacteriology 178 (19), 5636–5643.

Oliveira, J.L., Boroski, M., Azevedo, J.C.R., Nozaki, J., 2006.Spectroscopic investigation of humic substances in a tropicallake during a complete hydrological cycle. Acta HydrochimicaEt Hydrobiologica 34 (6), 608–617.

Reardon, K.F., Mosteller, D.C., Rogers, J.D.B., 2000. Biodegradationkinetics of benzene, toluene, and phenol as single and mixedsubstrates for Pseudomonas putida F1. Biotechnology andBioengineering 69 (4), 385–400.

Reasoner, D.J., Geldreich, E.E., 1985. A new medium for theenumeration and subculture of bacteria from potable water.Applied and Environmental Microbiology 49 (1), 1–7.

Reiche, N., Lorenz, W., Borsdorf, H. Development and applicationof dynamic air chambers for measurement of volatilizationfluxes of benzene and MTBE from constructed wetlandsplanted with common reed. Chemosphere, submitted forpublication.

Riis, V., Lorbeer, H., Babel, W., 1998. Extraction of microorganismsfrom soil: evaluation of the efficiency by counting methodsand activity measurements. Soil Biology and Biochemistry 30(12), 1573.

Rohwerder, T., Breuer, U., Benndorf, D., Lechner, U., Muller, R.H.,2006. The alkyl tert-butyl ether intermediate 2-hydroxyisobutyrate is degraded via a novel cobalamin-dependent mutase pathway. Applied and EnvironmentalMicrobiology 72 (6), 4128–4135.

Rosell, M., Finsterbusch, S., Jechalke, S., Hubschmann, T., Vogt, C.,Richnow, H.H., 2010. Evaluation of the effects of low oxygenconcentration on stable isotope fractionation during aerobicMTBE biodegradation. Environmental Science & Technology 44(1), 309.

Salanitro, J.P., Johnson, P.C., Spinnler, G.E., Maner, P.M.,Wisniewski, H.L., Bruce, C., 2000. Field scale demonstration ofenhanced MTBE bioremediation through aquiferbioaugmentation and oxygenation. Environmental Science &Technology 34 (19), 4152–4162.

Scherer, M.M., Richter, S., Valentine, R.L., Alvarez, P.J.J., 2000.Chemistry and microbiology of permeable reactive barriers forin situ groundwater clean up. Critical Reviews inEnvironmental Science and Technology 30 (3), 363–411.

Schmid, B.H., Koskiaho, J., Puustinen, M., 2005. Convectiveoxygen transport in a constructed wetland pond: mechanism,

measurements and modelling by multilayer perceptrons.Journal of Environmental Science and Health Part A-Toxic/Hazardous Substances & Environmental Engineering 40 (6–7),1281–1292.

Schmidt, T.C., Schirmer, M., Weiss, H., Haderlein, S.B., 2004.Microbial degradation of methyl tert-butyl ether and tert-butylalcohol in the subsurface. Journal of Contaminant Hydrology70 (3–4), 173–203.

Squillace, P.J., Zogorski, J.S., Wilber, W.G., Price, C.V., 1996.Preliminary assessment of the occurrence and possiblesources of MTBE in groundwater in the United States,1993�1994. Environmental Science & Technology 30 (5),1721–1730.

Takamizawa, K., Fukunaga, I., Inoue, Z., 1993. Promotion ofnitrification and denitrification by recirculation of effluent andbiofilm process. Environmental Technology 14 (10), 981–987.

Thorneby, L., Mathiasson, L., Martensson, L., Hogland, W., 2006.The performance of a natural treatment system for landfillleachate with special emphasis on the fate of organicpollutants. Waste Management & Research 24 (2), 183–194.

U.S. Environmental Protection Agency, 1999. CompendiumMethod TO-17, Determination of Volatile Organic Compoundsin Ambient Air Using Active Sampling onto Sorbent Tubes.Center for Environmental Research Information, Office ofResearch and Development, Cincinnati, OH.

US Environmental Protection Agency, 1997. Drinking WaterAdvisory: Consumer Acceptability Advice and Health EffectsAnalysis on Methyl Tertiary-Butyl Ether (MTBE). Office ofWater, Washington, DC, USA, p. 48.

Vogt, C., Alfreider, A., Lorbeer, H., Ahlheim, J., Feist, B.,Boehme, O., Weiss, H., Babel, W., Wuensche, L., 2002. Twopilot plant reactors designed for the in situ bioremediation ofchlorobenzene-contaminated ground water: hydrogeologicaland chemical characteristics and bacterial consortia. WaterAir and Soil Pollution: Focus 2, 161–170.

Waul, C.K., Arvin, E., Schmidt, J.E., 2007. In: Barcelo, D. (Ed.), FuelOxygenates. Springer-Verlag, Berlin Heidelberg, Germany, pp.213–248.

WHO, 2003. Benzene in Drinking-water. Background Documentfor Preparation of WHO Guidelines for Drinking-water Quality.World Health Organization, Geneva.

Wilson, R.D., Mackay, D.M., Scow, K.M., 2002. In situ MTBEbiodegradation supported by diffusive oxygen release.Environmental Science & Technology 36 (2), 190–199.

Yerushalmi, L., Lascourreges, J.F., Guiot, S.R., 2002. Kinetics ofbenzene biotransformation under microaerophilic andoxygen-limited conditions. Biotechnology and Bioengineering79 (3), 347–355.