traugott 2006

17
Mobility of pharmaceuticals carbamazepine, diclofenac, ibuprofen, and propyphenazone in miscible-displacement experiments Traugott J. Scheytt a, * , Petra Mersmann b , Thomas Heberer c a Institute of Applied Geosciences, Technical University of Berlin, Ackerstr. 71-76, 13355 Berlin, Germany b Staatliches Umweltamt Krefeld, St. Toeniser Str. 60, 47803 Krefeld, Germany c Federal Institute for Risk Assessment, FG 702, Thielallee 88-92, 14195 Berlin, Germany Received 11 May 2004; received in revised form 1 November 2005; accepted 3 November 2005 Available online 15 December 2005 Abstract Many pharmaceuticals pass the unsaturated zone before reaching an aquifer. Therefore, laboratory sand column transport experiments were conducted to study the transport behavior of carbamazepine, diclofenac, ibuprofen, and propyphenazone under unsaturated conditions. The test water was artificial sewage effluent to simulate the infiltration of reused wastewater. The test water was spiked with the pharmaceutically active compounds and the tracer LiCl. Afterwards it was passed through laboratory sand columns, one experiment for each pharmaceutical. The physical and chemical parameters were recorded and general ions measured. Pharmaceuticals were measured using solid phase extraction, derivatization, and detection with GC-MS. The column experiments indicate a significant elimination of ibuprofen (54%), propyphenazone (55%), and diclofenac (35%), whereas carbamazepine was not eliminated. Retardation factors varied between 1.84 for carbamazepine, 2.51 for propyphenazone, 3.00 for ibuprofen, and 4.80 for diclofenac. These results show that mobility and elimination of diclofenac, ibuprofen, and propyphenazone is about in the same range as for experiments under saturated conditions whereas carbamazepine had a significantly lower sorption and elimination under unsaturated conditions. D 2005 Elsevier B.V. All rights reserved. Keywords: Sand column; Drugs; Sorption; Degradation; Unsaturated zone 0169-7722/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.jconhyd.2005.11.002 * Corresponding author. Tel.: +49 30 314 72417; fax: +49 30 314 25674. E-mail address: [email protected] (T.J. Scheytt). Journal of Contaminant Hydrology 83 (2006) 53 – 69 www.elsevier.com/locate/jconhyd

Upload: amir-keshvari

Post on 10-Mar-2015

37 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Traugott 2006

Journal of Contaminant Hydrology 83 (2006) 53–69

www.elsevier.com/locate/jconhyd

Mobility of pharmaceuticals carbamazepine,

diclofenac, ibuprofen, and propyphenazone in

miscible-displacement experiments

Traugott J. Scheytt a,*, Petra Mersmann b, Thomas Heberer c

a Institute of Applied Geosciences, Technical University of Berlin, Ackerstr. 71-76, 13355 Berlin, Germanyb Staatliches Umweltamt Krefeld, St. Toeniser Str. 60, 47803 Krefeld, Germany

c Federal Institute for Risk Assessment, FG 702, Thielallee 88-92, 14195 Berlin, Germany

Received 11 May 2004; received in revised form 1 November 2005; accepted 3 November 2005

Available online 15 December 2005

Abstract

Many pharmaceuticals pass the unsaturated zone before reaching an aquifer. Therefore, laboratory sand

column transport experiments were conducted to study the transport behavior of carbamazepine, diclofenac,

ibuprofen, and propyphenazone under unsaturated conditions. The test water was artificial sewage effluent

to simulate the infiltration of reused wastewater. The test water was spiked with the pharmaceutically active

compounds and the tracer LiCl. Afterwards it was passed through laboratory sand columns, one experiment

for each pharmaceutical. The physical and chemical parameters were recorded and general ions measured.

Pharmaceuticals were measured using solid phase extraction, derivatization, and detection with GC-MS.

The column experiments indicate a significant elimination of ibuprofen (54%), propyphenazone (55%),

and diclofenac (35%), whereas carbamazepine was not eliminated. Retardation factors varied between 1.84

for carbamazepine, 2.51 for propyphenazone, 3.00 for ibuprofen, and 4.80 for diclofenac. These results

show that mobility and elimination of diclofenac, ibuprofen, and propyphenazone is about in the same

range as for experiments under saturated conditions whereas carbamazepine had a significantly lower

sorption and elimination under unsaturated conditions.

D 2005 Elsevier B.V. All rights reserved.

Keywords: Sand column; Drugs; Sorption; Degradation; Unsaturated zone

0169-7722/$ -

doi:10.1016/j.j

* Correspond

E-mail add

see front matter D 2005 Elsevier B.V. All rights reserved.

conhyd.2005.11.002

ing author. Tel.: +49 30 314 72417; fax: +49 30 314 25674.

ress: [email protected] (T.J. Scheytt).

Page 2: Traugott 2006

T.J. Scheytt et al. / Journal of Contaminant Hydrology 83 (2006) 53–6954

1. Introduction

Pharmaceutically active substances (PhACs) from human medical care have been detected in

sewage water, in surface water and even in groundwater (Heberer, 2002; Scheytt, 2002; Tixier

et al., 2003). After application, many drugs are excreted as the parent substances or water-

soluble metabolites (Mutschler, 1991). These substances are only partially eliminated in

wastewater treatment plants and subsequently can enter the surface waters (Ternes, 1998).

Pharmaceutically active substances in wastewater and surface water can reach groundwater by

several ways including river bank filtration, artificial groundwater recharge, naturally occurring

influent groundwater flow conditions, and leaky sewage systems. In the case of wastewater

reuse, sewage water irrigation, and application of sewage sludge on agricultural land the

pharmaceuticals pass through the unsaturated zone before reaching an aquifer.

Not only excretion but also disposal of pharmaceuticals leads to occurrences in the aquatic

system, because pharmaceuticals are either disposed into the toilet or with the garbage.

According to Stiftung Warentest (2000), the amount of pharmaceuticals in Germany that is

prescribed but not used and therefore disposed sums up to approximately 4500 t annually. Other

sources (Glaeske, 1998) estimate even higher amounts of up to 16,000 t/year of disposed

pharmaceuticals from human medical care for Germany only. Based on evaluations from

pharmacies 60% to 80% of those disposed drugs are either flushed down the toilet or disposed

with normal household waste (Heeke and Gunther, 1993). With the exception of cytostatics,

which are considered hazardous waste, all other pharmaceuticals are disposed as normal

household disposal (Kummerer et al., 1998).

Eckel et al. (1993) reported that a landfill in Florida that received wastes in 1968 and 1969

has contaminated nearby shallow groundwater with pharmaceuticals. Analysis of a sample of

contaminated groundwater collected in 1984 from a well approximately 300 m from the landfill

tentatively indicated the presence of the addictive sedative pentobarbital, among many other

compounds. Awell at the same location was resampled in 1991, and pentobarbital was positively

identified using gas chromatography/tandem mass spectrometry at a concentration of 1000 ng

L�1. The persistence over 21 years of a supposedly unstable drug in anoxic groundwater raises

the possibility that other pharmacologically active compounds may be found in plumes coming

from landfills that have accepted medical wastes.

Another example of pharmaceuticals percolating into groundwater from landfills is presented

by Holm et al. (1995), who found down gradient of the landfill Grindsted (Denmark)

pharmaceuticals in groundwater. This landfill was not only used for household waste but also for

disposal of waste from pharmaceuticals production in the period between 1962 and 1975.

Maximum concentrations measured in groundwater were 10,440 Ag L�1 of sulfonamide and

4000 Ag L�1 of propyphenazone (Holm et al., 1995). Ahel et al. (1998) report the occurrence of

pharmaceuticals down gradient of a landfill near Zagreb (Croatia). This landfill was used not

only for household disposal but also for disposal of industrial waste. The authors found

propyphenazone not only in groundwater but also in the sediments beneath the landfill (Ahel

et al., 1998; Ahel and Jelicic, 2000). Concentrations reached up to 0.1 mg propyphenazone per

kg solids.

Finally, production sites are potential sources of pharmaceuticals. Data on input of

pharmaceuticals from production sites are rare. Skanavis (1999) reports an occurrence of

groundwater contamination by pharmaceuticals in Puerto Rico; El-Gohary et al. (1995) describe

the reuse of wastewater from pharmaceutical production sites north of Cairo in Egypt. Although

no systematic reports on occurrences of pharmaceutically active substances beneath production

Page 3: Traugott 2006

Fig. 1. Molecular structures of carbamazepine, diclofenac, ibuprofen, and propyphenazone.

T.J. Scheytt et al. / Journal of Contaminant Hydrology 83 (2006) 53–69 55

sites or storage facilities are available, in all cases pharmaceutical compounds have most likely

passed the unsaturated zone before reaching an aquifer.

The substances tested in the present experiments have been used extensively in human

medical care and have been observed repeatedly in surface water and in groundwater worldwide

(Heberer, 2002; Tixier et al., 2003). The PhACs tested are prescribed in high amounts and

belong to the compounds most widely prescribed (Scheytt, 2002). In Germany, in 1999 the

prescription volume reached up to 120 t for carbamazepine, up to 250 t for diclofenac, up to

140 t for ibuprofen, and up to 7.6 t for propyphenazone (Scheytt, 2002). The total amounts of

the pharmaceuticals administered to the patients might have been even higher because these

Table 1

Physical and chemical properties of the test compounds

Compound Vapor pressure [mm Hg] Solubility [mg L�1] log KOW pKa (20 8C)

Carbamazepine 1.4 d 10�7a 17.66b 2.25c 14.0

Diclofenac 6.14 d 10�8a 2.37d 4.51e 4.16f

4.02c

Ibuprofen 1.86 d 10�4a 21g 3.5h 4.52f

3.97e

Propyphenazone 5.2 d 10�6a 2400i 2.32j

Temperature is 25 8C, if not stated otherwise; 1 Pa=7.500617 d 10�3 mm Hg.a Calculated value; from: Neely and Blau (1985).b Calculated value; from: Meylan et al. (1996).c Syracuse Science Center (2002).d Fini et al. (1993).e Avdeef et al. (1998).f Rafols et al. (1997).g Yalkowsky and Dannenfelser (1992).h Stuer-Lauridsen et al. (2000).i Merck Index (2001).j Holm et al. (1995).

Page 4: Traugott 2006

T.J. Scheytt et al. / Journal of Contaminant Hydrology 83 (2006) 53–6956

numbers do not include hospital applications and the amounts purchased without prescription as

so-called over-the-counter drugs. Due to the amounts used and due to the nature of their

application, PhACs belong to the environmentally relevant compounds. Because they are

produced and administered with the aim of causing a biological effect, their occurrence in

groundwater is not only of scientific but also of public interest. To date, many of the possible

actions and biochemical ramifications on non-target organisms (aquatic biota) are not fully

understood, and many may be completely unknown (Daughton and Ternes, 1999).

The molecular structures of the pharmaceutically active compounds investigated are shown in

Fig. 1. The physical and chemical properties are compiled in Table 1. Carbamazepine is an anti-

epileptical drug used worldwide. This drug has been detected frequently in surface water and

occasionally in groundwater (Scheytt, 2002). Diclofenac is used in human medical care as an

analgesic, antiarthritic, antirheumatic compound. It belongs to the group of nonsteroidal anti-

inflammatory drugs (Mutschler, 1991). Ibuprofen is a nonsteroidal anti-inflammatory, analgesic,

and antipyretic drug. It is an important nonprescription drug and used widely (Buser et al.,

1999). Propyphenazone is a mild analgesic (pain-killer) pharmaceutical that is normally used in

combination drugs. Today, it is produced in Japan and Eastern Europe (Holm et al., 1995).

The aim of this study was to determine the mobility and transport behavior of the PhACs

carbamazepine, diclofenac, ibuprofen, and propyphenazone under water unsaturated conditions

in laboratory sand columns. Four sand column experiments were conducted with one experiment

for every compound. Only very few investigations on the mobility of pharmaceuticals under

unsaturated conditions have been published so far (Preuß et al., 2001; Rabølle and Spliid, 2000;

Stuer-Lauridsen et al., 2000). Mersmann et al. (2002) and Scheytt et al. (2004) showed great

differences in the mobility behavior of those pharmaceuticals in sand column experiments under

water-saturated conditions. Therefore, the experiments presented here should also reveal whether

there are significant differences of transport and mobility under unsaturated conditions compared

to saturated conditions.

2. Material and methods

2.1. Test sediment

One type of sediment was used in all the experiments, though each experiment was conducted

with fresh sediment. The sediment was obtained from well drillings of Berlin Water Works

(Berliner Wasserbetriebe) from the unsaturated zone and represents a typical sediment from the

unsaturated zone from the Berlin area. This sediment was sampled from a depth of 1.0–2.5 m

below ground surface at a location near water works Stolper Heide (Brandenburg, Germany),

north of Berlin (Germany). This sediment is typically found within the unsaturated zone and

several PhACs contaminations are reported from sediments very similar to the sediment chosen

here (Scheytt et al., 2001). The sediment consists of fine-grained alluvial sands with a low

content of organic carbon (fraction of organic carbon=0.13%). All sediment samples were air-

dried, gently crushed, sieved through a 2-mm sieve, and stored at 4 8C. Organic carbon was

analyzed using a Shimadzu Total Organic Carbon Analyzer (TOC-5050a). The cation exchange

capacity (CEC) was measured for both 135 sediments. The potential CEC (CECpot) was tested

by adding a buffered BaCl solution (pH 8.1) to the air-dried sediment (b2 mm). Another

empirical measure of the CEC was made by assessing the maximum concentrations of weakly

bound cations such as ammonium that can be sorbed. This so-called effective CEC (CECeff) was

measured by applying NH4Cl with no adjustment of the pH. Units of CEC vary, but the unit used

Page 5: Traugott 2006

Table 2

Sediment characteristics for sediment bStolper HeideQ

Parameters Texture [%]

pH [CaCl2] 4.8 Clay (b2 Am) 0

Total carbon [g kg�1] 1.3 Silt (2–63 Am) 0.7

Fraction of organic carbon, fOC [kg kg�1] 0.0013 Fine sand (63–200 Am) 42.05

CECeff [mmolc kg�1] 4 Medium sand (200–630 Am) 56.46

CECpot [mmolc kg�1] 15 Coarse sand (200–2000 Am) 0.8

Hydraulic conductivity using Beyer method [m s�1] 1 d 10�4 Gravel (N2 mm) 0.2

T.J. Scheytt et al. / Journal of Contaminant Hydrology 83 (2006) 53–69 57

here, [mmolc kg�1] with bcQ referring to the charge are commonly used in soil sciences. The

uniformity coefficient Cu which is a measure of the particle size range was obtained from the

grain size distribution curve. Characteristics of the sediment are listed in Table 2.

2.2. Water

The sand column experiments under unsaturated conditions were planned to simulate

pharmaceutical transport during wastewater reuse operations. Therefore, column experiments

were conducted using artificially synthesized water, representing an effluent from sewage

treatment. The composition of this sewage water was synthesized following data on average

treated sewage water compositions in Germany (Eiswirth et al., 1995; Matthess, 1994). The

chemical compositions of the water and the components to create this water are presented in

Tables 3 and 4.

2.3. Chemical analysis

For the analysis of the pharmaceutical compounds the water samples were adjusted to a pH

value of 2 and then extracted by solid-phase extraction (SPE) using a non-endcapped reversed

phase adsorbent (RP-C18 Bakerbond Polar Plus). Before extraction, the samples were spiked

with 100 ng of 4-chlorophenoxy-butyric acid (100 AL of a 1 ng AL�1 solution in methanol) or

dihydrocarbamazepine used as surrogate standards for analytical quality control (Reddersen and

Heberer, 2003). The acidic analytes diclofenac, ibuprofen, and propyphenazone and the

surrogate standard were derivatized with pentafluorobenzyl bromide (2% in toluene),

carbamazepine was derivatized with N-tert-butyldimethylsilyl-N-methyltrifluoroacetamide

(MTBSTFA) making them amendable to gas chromatographic separation using 5%-phenyl-

methylpolysiloxane column (HP5MS, 30 m�0.25 mm inner diameter, 0.25 Am film thickness)

from Agilent Technologies (Waldbronn, Germany) (Reddersen and Heberer, 2003). Of the

sample extracts (100 AL for each sample), 2 AL was analyzed by capillary gas chromatography-

Table 3

Concentrations of cations and anions and physical and chemical properties of the synthesized treated sewage water

Cations Anions Physical and chemical parameter

Na+ [mg L�1] 81.3 Cl� [mg L�1] 208.0 SC [AS cm�1] 1068

K+ [mg L�1] 16.7 SO42� [mg L�1] 100.5 pH 8.0

Mg2+ [mg L�1] 12.2 NO3� [mg L�1] 38.1 EH [mV] 508

Ca2+ [mg L�1] 46.5 PO43� [mg L�1] 37.6 O2 [mg L�1] 9.3

Page 6: Traugott 2006

Table 4

Substances used to synthesize the test water

KH2PO4 66.13 [mg L�1] CaCl2d 2H2O 183.40 [mg L�1]

Na2SO4 68.02 [mg L�1] NH4NO3 26.62 [mg L�1]

NaCl 140.12 [mg L�1]

MgSO4d 7H2O 131.83 [mg L�1] NaOH Adjustment of pH=8.0

T.J. Scheytt et al. / Journal of Contaminant Hydrology 83 (2006) 53–6958

mass spectrometry (GC-MS) with selected ion monitoring (SIM) using an HP 5890 gas

chromatograph and an HP 5970B quadrupol mass spectrometer from Agilent Technologies

(Waldbronn, Germany). Depending on the sample volume (100 to 200 mL) and the matrix, the

limits of determination were between 1 and 10 ng L�1, and the limits of quantitation were

between 5 and 25 ng L�1. The analytical recoveries ranged from 80% to 120%. For further

analytical details refer to Heberer et al. (1998) or Reddersen and Heberer (2003).

All water samples from the column experiment were frozen immediately after sampling. Each

water sample chosen for analysis of general ions was filtered through a 0.45-Am filter (Schull &

Schleicher) and analyzed afterwards. Analyses for calcium, magnesium and lithium were carried

out on using a Thermo Jarrel Ash ICP-AES; sodium, potassium, iron and manganese were

analyzed using flame AAS Philips PU 9400. Analysis of anions was performed on a Dionex ion

chromatograph DX 120.

2.4. Setup column experiments

Air-dried, homogenized sediment was manually packed into the stainless steel column

measuring 0.35�0.136 m (inner diameter). The sediment was added in approximately 3-cm

increments and tapped lightly with a plunger. Care was taken to avoid obvious layering of the

material or segregation of sediment by particle size. A gauze net and 1–2 mm diameter glass

globes were placed at the bottom of the column to prevent leaching of sediment particles; at the

top only the globes where used without a net. The bulk density ranges between 1.73 and

1.87 g cm�3. Other column characteristics and experimental conditions are provided in Table 5.

Table 5

Characteristics and experimental conditions for column runs

Carbamazepine Diclofenac Ibuprofen Propyphenazone

Q m3 h�1 1.31 d 10�4 1.33 d 10�4 1.30 d 10�4 1.33 d 10�4

(m day�1) 0.216 0.221 0.216 0.221

vp,avg m day�1 0.77 0.71 0.84 0.88

h vol.% 28 28 29 26

qb g cm�3 1.80 1.86 1.87 1.73

PV mL 1190 1155 1235 1125

DL m2 day�1 7.2 d 10�3 8.7 d 10�3 9.9 d 10�3 1.1 d 10�2

h m 0.29 0.28 0.29 0.30

d m 0.136 0.136 0.136 0.136

Vs m3 4.21 d 10�3 4.06 d 10�3 4.21 d 10�3 4.35 d 10�3

Transport parameter based on chloride and specific conductance breakthrough curves.

Q =volumetric flow rate; vp,avg=average pore water velocity; h =moisture content; qb=bulk density; PV= pore volume;

DL=longitudinal dispersion coefficient; h =column height filled with sediment; d =column diameter Vs=volume of the

column filled with sediment (total volume).

Page 7: Traugott 2006

Fig. 2. Diagram of the sand column setup for unsaturated flow.

T.J. Scheytt et al. / Journal of Contaminant Hydrology 83 (2006) 53–69 59

The stainless steel irrigation head was fixed on top of the column (Fig. 2). The irrigation head

consists of 69 single needles through which the water was applied. The artificial sewage effluent

was irrigated to obtain the desired water flux of 130 mL h�1 on average, equivalent to

approximately 0.22 m day�1 of dirrigationT. This amount of water lead to pore water velocities

vp,avg between 0.71 and 0.88 m day�1 with volumetric water contents between 26% and 29%.

The water was drained by gravity through the column (free drainage), the column was open to

atmospheric pressures. The eluted liquid was collected in fractions of approximately 25 mL

with a fraction collector and frozen in a refrigerator. From all these samples collected, an

appropriate subset of samples was chosen for laboratory analyses. After the experiment the sand

column was segmented. The samples were stored without any further analysis done on the

samples. The column was irrigated with the artificial sewage effluent for a period of 165 h,

equivalent to approximately 18 exchanged pore volumes. The artificial sewage effluent was

spiked with lithium chloride (LiCl) as tracer and the pharmaceuticals (one compound for each

experiment) and then passed through the column for approximately 70 h. After this first

phase, the second phase took another 95 h with the column being flushed with artificial sewage

effluent but without lithium chloride or pharmaceuticals. Beside the tracer and the

pharmaceutical compounds, all other parameters (e.g. flow rate) were kept the same during

the experiment and during the study. The experiments took place at a room/ambient temperature

of approximately 21 8C. The eluted liquid was analyzed for contents of anions, cations, and

pharmaceutical chemicals and lithium.

The concentration of the tracer lithium chloride was 61 mg L�1 and the pharmaceuticals had

a concentration of 1 Ag L�1 in the spiked water. These concentrations of pharmaceuticals used

for the experiments have been detected in the unsaturated zone at the sewage irrigation farms

south of Berlin (Scheytt et al., 2001) and also in groundwater in the Berlin area (Scheytt, 2002)

and other places worldwide (Heberer, 2002). Physico-chemical parameters (EH, pH, temperature,

oxygen saturation, specific conductance) were measured every 10 min using respective

Page 8: Traugott 2006

T.J. Scheytt et al. / Journal of Contaminant Hydrology 83 (2006) 53–6960

electrodes coupled to a data logger. Lithium chloride was chosen as a tracer because the

background concentration of lithium was below detection limit in all sediment and water

samples used for the experiments. Lithium also shows a high mobility and can be analyzed in a

rapid and cost-effective manner.

The column was dismantled immediately after the end of each experiment and the water

content was determined at different lengths within the column. The results show that no

saturated zones could be identified within the column. However, the possibility of a water-

saturated zone at the bottom of the column cannot be ruled out completely. Water contents within

the column varied between 21% in the uppermost 6 to 8 cm and 30% to 36% at the bottom of the

column, indicating predominantly unsaturated conditions throughout the column.

2.5. Theoretical considerations and sand column transport characteristics

The one-dimensional convective–dispersive solute transport through a homogeneous

medium, with linear and reversible equilibrium sorption without degradation of the solute,

under steady-state flow can be described by the Convective–Dispersive Equation (CDE):

Rf

BC

Bt¼ D

B2C

Bx2� vp;avg

BC

Bxð1Þ

where Rf is the retardation factor [–], C the normalized concentration (C/C0) in the fluid phase at

distance x [m] and time t [s], D is the hydrodynamic dispersion coefficient [m2 s�1] and vp,avgthe average pore water velocity [m s�super1]. Molecular diffusion was neglected because of the

high pore water velocity. Hydromechanic dispersion can be described as

DL ¼ aLvbp;avg ð2Þ

where DL is the longitudinal hydromechanic dispersion coefficient [m2 s�1] and aL the

longitudinal dispersivity [m]; b has values between 0.9 and 1.2. Studies indicated that for most

purposes b could generally be taken as unity for sandy sediments (Freeze and Cherry, 1979).

Studies on packed sand columns have found dispersivity to be higher under unsaturated

conditions (Jin et al., 2000). Other studies found a decrease in dispersivity with desaturation in

packed sand columns (James and Rubin, 1986; Wierenga and van Genuchten, 1989). Many

factors contribute to changes in the hydrodynamic properties of sediment under unsaturated

conditions, including the existence of regions of immobile water (two-region immobile water

model), a wider variety in pore water velocities when media is unsaturated or an increase in air-

filled pore space that increases the tortuosity of the solute flow path with desaturation (De Smedt

and Wierenga, 1984).

CXTFIT code of Parker and van Genuchten (1984) in its updated and modified version by

Toride et al. (1995) was used to perform the simulations. To calculate the pore water velocity for

unsaturated media (average pore water velocity vp,avg), the dispersion coefficient, and other

transport characteristics (Table 5), the breakthrough curves of chloride were fitted for each

experiment using the deterministic equilibrium CDE and the inverse mode to fit Eq. (1). With

these data, the breakthrough curves of lithium were fitted for each experiment allowing for

sorption of lithium. Lithium transport and the subsequent fitting were used to obtain information

about flow conditions during our experiments. In all cases, fitting of chloride and lithium was

performed using the deterministic equilibrium CDE and the inverse mode. Except for the

carbamazepine breakthrough curve, fitting of the breakthrough curves for the pharmaceuticals

Page 9: Traugott 2006

T.J. Scheytt et al. / Journal of Contaminant Hydrology 83 (2006) 53–69 61

was performed using deterministic nonequilibrium CDE allowing for retardation and

degradation. In these cases, the two-region physical nonequilibrium model was chosen with

the solute and adsorbed phase degradation rates being independent.

An adequate analytical solution of Eq. (1) is presented by van Genuchten and Parker (1984).

This analytical solution was used to obtain the retardation factors Rf.

Rf ¼vp;avg

vsp;avg¼ 1þ qb

hKd ð3Þ

where qb is the bulk density [kg L�1 or: g cm�3 if water density=1.000 g cm�3], h the

volumetric water content [–], and Kd the sorption distribution coefficient [L kg�1]. Retardation

factors observed in column studies often differ from values determined by batch equilibration

experiments (Porro et al., 2000; Gamerdinger et al., 2001). Increasing retardation factors could

be due to decreasing water content, as expected from Eq. (3), or due to limitations in availability

of sorption sites in batch vs. column experiments, or due to rate-limited geochemical reactions

and physical nonequilibrium present under flowing (column) vs. static (batch) conditions

(Hutchison et al., 2003).

The sorptive exchange of substances between the water phase and the solid phase may be

described by sorption coefficients, that are defined as the ratio of the concentration of the

substance in the solid phase csorb (mg kg�1) and in water cw (mg L�1) at equilibrium. The

simplest case of an isotherm occurs with a linear regression and a distribution coefficient Kd

(L kg�1).

csorb ¼ Kdcw ð4Þ

The distribution coefficient can also be derived from column tests using Eq. (3). This

distribution coefficient is different than that derived from batch experiments (Scheytt et al.,

2005) and will be termed the transport distribution coefficient Kd, Trans.

The mass balance for the pharmaceuticals was calculated by subtracting input mass and

output mass. Output mass was obtained by graphical integration of the area marked by the

breakthrough curves. For all breakthrough curves of the pharmaceuticals, the maximum C/C0

value reached and the average normalized concentration are reported to indicate the amount

eliminated through biological and chemical degradation and mineralization. To quantify the

elimination, first-order decay coefficients for the different pharmaceuticals from curve fitting

simulations are presented. Although no methods were applied to identify metabolites in the

effluent this degradation is most likely biodegradation.

3. Results

3.1. General ions and parameters

3.1.1. Physical and chemical parameters

Physical and chemical parameters were logged continuously. With the exception of specific

conductance, all other parameters did not show significant variations during the experiments.

Due to the application of LiCl as tracer the breakthrough could be monitored using the specific

conductance. Depending on the experiment, specific conductance increased between 0.6 and

1.0 exchanged pore volumes and reached a constant value of approximately 1000 AS cm�1 after

1.3 and 3.0 pore volumes. When the column was flushed with pure water specific conductance

decreased within 0.8 to 1.0 pore volumes.

Page 10: Traugott 2006

Fig. 3. Concentrations of cations and anions of column experiment carbamazepine. End of spiking is indicated by dashed

vertical line.

T.J. Scheytt et al. / Journal of Contaminant Hydrology 83 (2006) 53–6962

Redox potential was measured at around +500 mV at the outflow of the column, oxygen

constant ranged between 4.0 and 9.0 mg L�1. These values indicate that the prevailing

conditions in the sand column were aerobic. The pH value of the test water (pH 8.0) decreased

when passing the column to about 5.5, which is close to the value of the sediment (pH=4.8).

3.1.2. General ions

The outflow of the column was sampled every hour. About 20 samples were chosen to

measure general ions. Concentrations of those ions did show no great concentration variations

with the exception of lithium and chloride, which were added to the test water (Fig. 3).

Potassium concentrations reached steady state conditions after less than three pore water

exchanges, with values between 16 and 17.5 mg L�1. Sodium concentrations reached 75 to 90

mg L�1 after no more than two pore volume exchanges. Calcium concentration declined

whereas magnesium values increased within the first pore volume exchanges. Magnesium

reached 11 to 13 mg L�1 and calcium 43 to 50 mg L�1.

Among the anions the concentrations of nitrate decrease after one pore volume to 48 to 50 mg

L�1, concentrations of sulfate are stable through the whole experiment with 93 to 105 mg L�1.

Concentrations of chloride increased after the start of the experiment to approximately 215 mg

L�1 and decreased after cessation of spiking of the test water to 172 to 175 mg L�1. In general,

steady state conditions were reached after no more than 3 exchanged pore volumes for all ions.

Page 11: Traugott 2006

Fig. 4. Breakthrough curve of carbamazepine and the tracer lithium. End of spiking is indicated by dashed vertical line.

T.J. Scheytt et al. / Journal of Contaminant Hydrology 83 (2006) 53–69 63

3.2. Breakthrough curves of PhACs

The test water was spiked with the tracer LiCl and the PhACs carbamazepine, diclofenac,

ibuprofen, and propyphenazone at the beginning of each experiment. The breakthrough curves

are plotted as normalized concentrations (C/C0) and exchanged pore volumes. Curve fitting for

lithium breakthrough curves (Figs. 4–7) reveals that lithium is retarded with a retardation factor

between 1.33 (ibuprofen experiment) and 1.44 (propyphenazone experiment). The results of the

curve fitting for lithium are presented as solid lines in Figs. 4–7. The transport behavior indicates

that preferential flow and macropores do not seem to play a significant role during mass

transport.

3.2.1. Carbamazepine

The increase of concentrations of carbamazepine is retarded compared to the increase of

lithium concentration (Fig. 4). The retardation factor was found to be 1.84 and the average

Fig. 5. Breakthrough curve of diclofenac and the tracer lithium. End of spiking is indicated by dashed vertical line.

Page 12: Traugott 2006

Fig. 6. Breakthrough curve of ibuprofen and the tracer lithium. End of spiking is indicated by dashed vertical line.

T.J. Scheytt et al. / Journal of Contaminant Hydrology 83 (2006) 53–6964

normalized concentration was found to be 93% (C/C0=0.93). Therefore, degradation of

carbamazepine is very unlikely under the prevailing conditions. After the end of spiking, the

concentrations decreased slightly slower than lithium concentration and could not be detected in

the outflow after 5 exchanged pore volumes. The recovery of carbamazepine was found to be

102%, indicating in agreement with the other results that carbamazepine was not degraded

during the experiment and also, carbamazepine was desorbed completely.

3.2.2. Diclofenac

The transport behavior of diclofenac differs significantly from the behavior of carbamazepine.

The increase of concentration of diclofenac is clearly retarded and highest normalized

concentration reached was a C/C0 ratio of 0.83 (Fig. 5). Average normalized concentration

was 0.55. After the end of spiking the concentration decreased later than lithium. Recovery of

diclofenac was 63%, which shows an elimination of approximately 40% of the inflow mass.

Retardation factor was calculated to be 4.80.

Fig. 7. Breakthrough curve of propyphenazone and the tracer lithium. End of spiking is indicated by dashed vertical line.

Page 13: Traugott 2006

Table 6

Results of sand column experiments under unsaturated conditions

Compound Average C/C0 Recovery [%] Rf l [h�1] Kd, Trans [L kg�1]

Carbamazepine 0.93 102 1.84 0.006 0.131

Diclofenac 0.55 63 4.80 0.37 0.572

Ibuprofen 0.45 46 3.00 0.95 0.310

Propyphenazone 0.40 45 2.51 0.86 0.227

Kd, Trans value based on retardation factor, l =first-order decay coefficient from curve fitting.

T.J. Scheytt et al. / Journal of Contaminant Hydrology 83 (2006) 53–69 65

3.2.3. Ibuprofen

Ibuprofen was measured at the outflow of the column immediately after the start of the

experiment, but concentration decreased again and a sustained increase in concentration started

after approximately two exchanged pore volumes (Fig. 6). Based on the concentration increase,

the retardation factor was calculated to be 3.00 with the highest normalized concentration of 0.7.

However, concentrations decreased afterwards even before the end of spiking of the water.

Recovery was 46% that coincides very well with the average normalized concentration of C/C0

of 0.45. These numbers indicate that ibuprofen is not only significantly retarded under the

prevailing aerobic unsaturated conditions but also eliminated during soil passage.

3.2.4. Propyphenazone

The transport pattern of propyphenazone under unsaturated conditions is presented in Fig. 7.

In the beginning, no retardation was observed. Highest normalized concentration found was

0.5 after two exchanged pore volumes. The retardation factor came out to be 2.51. The average

C/C0 ratio was 0.4 and the amount recovered was 45%. The low C/C0 ratio and the small

amount recovered indicate significant elimination.

4. Discussion and conclusion

The objective of these solute transport experiments was to evaluate the transport behavior of

PhACs under unsaturated conditions. The water flux of the unsaturated column experiments was

in the range of the amount of water irrigated at the sewage irrigation farms south of Berlin

(Scheytt et al., 2001). These experiments were set up to simulate the field conditions

encountered at the irrigation farms south of Berlin. At those sewage irrigation farms drainage

channels were used to collect the water after percolating through the unsaturated zone. The

experiments should also be capable of describing the transport of pharmaceuticals through the

unsaturated zone in humid areas and during wastewater reuse in similar cases.

The sand column characteristics compared very well among the different single experiments.

However, due to the higher velocities of the unsaturated experiments presented in this

publication the dispersion coefficients as well as dispersivities were higher under unsaturated

conditions than under saturated conditions.

The sorption coefficients Kd, Trans calculated on the basis of Eq. (3) are presented in Table 6.

Retardation factors obtained through an analytical solution of Eq. (1) are based on the

assumption that the sorption isotherm is linear and that there is physical and chemical

equilibrium. If pore water velocity is high there may be a possibility that equilibrium conditions

are not reached. In this case, sorption would increase with decreasing pore water velocity in the

column. However, the experiments were designed to simulate the transport in the unsaturated

zone under wastewater reuse conditions with pore water velocities occurring under these

Page 14: Traugott 2006

Table 7

Comparison of results: retardation, recovery, and sorption for the pharmaceutically active substances

Carbamazepine Diclofenac Ibuprofen Propyphenazone

Sata Rf 2.8–3.3 2.0–2.6 4.0b 1.6–2.0

Recovery 93–105% 97–106% 9–46% 88–104%

Unsatc Rf 1.84 4.80 3.00 2.51

Recovery 102% 63% 46% 45%

log KOCd Columnc 2.00 2.64 2.38 2.24

log KOCd Batche 2.00–2.21 2.43–3.87 2.94–3.13 1.89–2.80

a Sat=column experiments under water saturated conditions; DOC=0.2%, pH=6.7, medium sand (Scheytt et al., 2004;

Mersmann et al., 2002).b Extrapolated value.c Unsat=column experiments under unsaturated conditions (this publication).d log KOC=organic carbon normalized sorption coefficient; log KOC=log(Kd/fOC).e Batch=results from batch experiments utilizing the same sediment as in the unsaturated sand column experiments

(Scheytt et al., 2005).

T.J. Scheytt et al. / Journal of Contaminant Hydrology 83 (2006) 53–6966

conditions. In order to compare those results with the results from previously conducted batch

experiments with the same sediment the values were normalized to the fraction of organic carbon

in the sediment. The results from the column experiments from this publication fit very well for

diclofenac and propyphenazone, whereas carbamazepine and ibuprofen yielded a lower sorption

at the column experiments compared to the batch tests. As sorption is expected to increase with

decreasing water content this finding is quite particular. However, several authors (Begin et al.,

2003; Hutchison et al., 2003) mention that retardation factors often indicate little trend with

water content.

Compared to the results from previously conducted sand column experiments under water-

saturated conditions (Mersmann et al., 2002; Scheytt et al., 2004), the transport behavior

deviated under unsaturated conditions significantly from saturated conditions.

Carbamazepine had a significantly lower retardation coefficient than under water-saturated

conditions and elimination did not occur (Table 7). This may explain why carbamazepine is

found quite regularly in groundwater samples, whereas other pharmaceuticals are barely

detected. For diclofenac, retardation and elimination were much higher under unsaturated

conditions than under saturated conditions, whereas for ibuprofen these parameters were about

in the same range as in the saturated conditions. The experiments presented here were intended

to simulate irrigation of sewage effluent. No precautions were taken to shield the container or

other parts of the experiment against light, but initial concentration was measured regularly.

Elimination of diclofenac under unsaturated conditions might be due in part to photochemical

biodegradation as has been reported by Buser et al. (1998) and has also been observed in own

degradation experiments not presented here. However, this photodegradation was not rapid

enough to prevent an increase in concentration up to a normalized concentration of 0.83. High

biodegradation rates of ibuprofen have already been reported several times (Buser et al., 1999;

Preuß et al., 2001; Winkler et al., 2001). In our experiment, it seems as if the elimination has a

lag of approximately 5 exchanged pore volumes. Concentrations of ibuprofen decrease after

5 pore volumes, even before the end of spiking of the test water. This could be an indication of

biodegradation in the sand column with an adaptation time for microorganisms.

For propyphenazone, the experiments revealed a high retardation factor and relatively high

elimination. This is in contradiction to previously conducted sand column experiments under

water-saturated conditions (Scheytt et al., 2004; Mersmann et al., 2002) and to findings by

Page 15: Traugott 2006

T.J. Scheytt et al. / Journal of Contaminant Hydrology 83 (2006) 53–69 67

Mohle et al. (1999). However, degradation has been reported already by Holm et al. (1995)

down gradient of a landfill site in Denmark and propyphenazone may also be prone to

irreversible sorption. Altogether, the pharmaceuticals of this study showed higher elimination

and lower mobility under unsaturated conditions than during saturated transport. However, they

were mobile and persistent enough to be expected in groundwater.

Acknowledgement

This investigation was funded by the German Research Foundation (Deutsche Forschungs-

gemeinschaft). The authors would like to thank G. Fricke and F. Ilmaz, for their help in the

laboratory and with the analysis of the pharmaceuticals. We thank Hongbin Zhan, Peter

Grathwohl and one anonymous reviewer for their constructive criticism of a previous version of

this paper.

References

Ahel, M., Jelicic, I., 2000. Occurrence of phenazone analgesics in landfill-leachate polluted groundwater. In: Keith, L.H.,

Needham, L.L., Jones-Lepp, T.L. (Eds.), Issues in the Analysis of Environmental Endocrine Disruptors, Proc. ACS

Symp. vol. 40 (1), pp. 109–111. Washington, DC.

Ahel, M., Mikac, N., Cosovic, B., Prohic, E., Soukup, V., 1998. The impact of contamination from a municipal solid

waste landfill (Zagreb, Croatia) on underlying soil. Water Sci. Technol. 37 (8), 203–210.

Avdeef, A., Box, K.J., Comer, J.E.A., Hibbert, C., Tam, K.Y., 1998. pH-metric logP 10. Determination of liposomal

membrane–water partitioning coefficients of ionizable drugs. Pharm. Res. 15 (2), 209–215.

Begin, L., Fortin, J., Caron, J., 2003. Evaluation of fluoride retardation factor in unsaturated and undisturbed soil

columns. Soil Sci. Soc. Am. J. 67, 1635–1646.

Buser, H.-R., Poiger, T., Muller, M.D., 1998. Occurrence and fate of the pharmaceutical drug diclofenac in surface

waters: rapid photodegradation in a lake. Environ. Sci. Technol. 32, 3449–3456.

Buser, H.-R., Poiger, T., Muller, M.D., 1999. Occurrence and environmental behaviour of the chiral pharmaceutical drug

ibuprofen in surface waters and in wastewater. Environ. Sci. Technol. 33, 2529–2535.

Daughton, C.G., Ternes, T.A., 1999. Pharmaceuticals and personal care products in the environment: agents of subtle

change? Environ. Health Perspect., Suppl. 107, 907–937.

De Smedt, F., Wierenga, P.J., 1984. Solute transport through columns of glass beads. Water Resour. Res. 20, 225–232.

Eckel, W.P., Ross, B., Isensee, R.K., 1993. Pentobarbital found in ground water. Ground Water 5 (31), 801–804.

Eiswirth, M., Hotzl, H., Lazar, C., Merkler, G.-P., Kramp, J., 1995. Qualifizierende Leckagedetektion bei

Abwasserkanalen. Proceed. 12th No-Dig Conf., Dresden, pp. 179–194.

El-Gohary, F.A., Abou-Elela, S.I., Aly, H.I., 1995. Evaluation of biological technologies for wastewater treatment in the

pharmaceutical industry. Water Sci. Technol. 32 (11), 13–20.

Fini, A., Fazio, G., Rabasco, A.M., 1993. 1-Octanol/water partitioning of diclofenac salts. Acta Technol. Legis Medicam.

4, 33–34.

Freeze, R.A., Cherry, J.A., 1979. Groundwater. Prentice-Hall, Englewood Cliffs, NJ.

Gamerdinger, A.P., Kaplan, D.I., Wellmann, D.M., Serne, R.J., 2001. Two-region flow model and decreased sorption of

uranium (VI) during transport in Hanford groundwater and unsaturated sands. Water Resour. Res. 37, 3155–3162.

Glaeske, G., 1998. Arzneimittel in Gewassern—Risiko fur Mensch, Tier und Umwelt?—Konsequenzen unter

Berucksichtigung des Arzneimittelverbrauchs. In: Hessische Landesanstalt fur Umwelt (Eds.), Arzneimittel in

Gewassern—Risiko fur Mensch, Tier 468 und Umwelt? Wiesbaden, pp. 105–109.

Heberer, Th., 2002. Occurrence, fate, and removal of pharmaceutical residues in the aquatic environment: a review of

recent research data. Toxicol. Lett. 131, 5–17.

Heberer, Th., Schmidt-Baumler, K., Stan, H.-J., 1998. Occurrence and distribution of organic contaminants in the aquatic

system in Berlin: part I. Drug residues and other polar contaminants in Berlin surface and groundwater. Acta

Hydrochim. Hydrobiol. 26 (5), 272–278.

Heeke, A., Gunther, J., 1993. Arzneimittel im Mull—Eine Studie der AOK Essen uber nicht verbrauchte Arzneimittel.

Dtsch. Apoth.-Ztg. 133 (46), 15–21.

Page 16: Traugott 2006

T.J. Scheytt et al. / Journal of Contaminant Hydrology 83 (2006) 53–6968

Holm, J.V., Rugge, K., Bjerg, P.L., Christensen, T.H., 1995. Occurrence and distribution of pharmaceutical compounds in

the groundwater downgradient of a landfill (Grindsted, Denmark). Environ. Sci. Technol. 29, 1415–1420.

Hutchison, J.M., Seaman, J.C., Aburime, S.A., Radcliffe, D.E., 2003. Chromate transport and retention in variably

saturated soil columns. Vadose Zone J. 2, 702–714.

James, R.V., Rubin, J., 1986. Transport of chloride ion in a water-unsaturated soil exhibiting anion exclusion. Soil Sci.

Soc. Am. J. 50, 177–181.

Jin, Y., Chu, Y., Li, Y., 2000. Virus removal and transport in saturated and unsaturated sand columns. J. Contam. Hydrol.

43, 111–128.

Kummerer, K., Erbe, T., Gartiser, S., Brinker, L., 1998. AOX-emissions from hospitals into municipal waste water.

Chemosphere 36 (11), 2437–2445.

Matthess, G., 1994. Die Beschaffenheit des Grundwassers. Borntrager, Berlin. 499 pp.

Merck Index, 2001. An Encyclopedia of Chemicals, Drugs, and Biologicals, 13th edition. Merck Research Laboratories,

Whitehouse Station, NJ.

Mersmann, P., Scheytt, T., Heberer, Th., 2002. Column experiments on the transport behavior of pharmaceutically active

substances in the saturated zone. Acta Hydrochim. Hydrobiol. 30, 275–284.

Meylan, W.M., Howard, P.H., Boethling, R.S., 1996. Improved method for estimating water solubility from octanol water

partition coefficient. Environ. Toxicol. Chem. 15 (2), 100–106.

Mohle, E., Horvath, S., Merz, W., Metzger, J.W., 1999. Bestimmung von schwer abbaubaren organischen Verbindungen

im Abwasser—Identifizierung von Arzneimittelruckstanden (Detection of persistent organic compounds in sewage

effluent—identification of pharmaceutical residues). Vom Wasser 92, 207–223.

Mutschler, E., 1991. Arzneimittelwirkungen—Lehrbuch der Pharmakologie und Toxikologie (Textbook of Pharmacology

and Toxicology), 6th edition. Wissenschaftliche Verlagsgesellschaft, Stuggart, 879 pp.

Neely, W.B., Blau, G.E., 1985. Environmental Exposure from Chemicals, 2nd edition. CRC Press, Boca Raton, FL.

Parker, J.C., van Genuchten, M.Th., 1984. Determining transport parameters from laboratory and field tracer

experiments. Va. Agric. Exp. Stn., 83–84.

Porro, I., Newman, M.E., Dunnivant, F.M., 2000. Comparison of batch and column methods for determining strontium

distribution coefficients for unsaturated transport in basalt. Environ. Sci. Technol. 34, 1679–1686.

Preuß, G., Willme, U., Zullei-Seibert, N., 2001. Behaviour of some pharmaceuticals during artificial groundwater

recharge—elimination of effects on microbiology. Acta Hydrochim. Hydrobiol. 29 (5), 269–277.

Rabølle, M., Spliid, N.H., 2000. Sorption and mobility of metronidazole, olaquindox, oxytetracycline and tylosine in soil.

Chemosphere 40, 715–722.

Rafols, C., Roses, M., Bosch, E., 1997. A comparison between different approaches to estimate the aqueous pKa of

several non-steroidal anti-inflammatory drugs. Anal. Chim. Acta 338, 127–134.

Reddersen, K., Heberer, Th., 2003. Multi-methods for the trace-level determination of pharmaceutical residues in sewage,

surface and ground water samples applying GC-MS. J. Sep. Sci. 26, 1443–1450.

Scheytt, T., 2002. Pharmaceuticals in groundwater—input, degradation, and transport (in German). Habilitation thesis,

Technical University Berlin, 148 pp.

Scheytt, T., Grams, S., Rejman-Rasinski, E., Heberer, Th., Stan, H.-J., 2001. Pharmaceuticals in groundwater: clofibric

acid beneath sewage farms south of Berlin, Germany. In: Daughton, C.G., Jones-Lepp, T.L. (Eds.), Pharmaceuticals

and Personal Care Products in the Environment—Scientific and Regulatory Issues, ACS Symposium Series, vol. 791.

ACS/Oxford University Press, pp. 84–99.

Scheytt, T., Mersmann, P., Leidig, M., Pekdeger, A., Heberer, Th., 2004. Transport of pharmaceutically active

compounds in saturated soil columns. Ground Water 42 (5), 767–773.

Scheytt, T., Mersmann, P., Lindstadt, R., Heberer, Th., 2005. Determination of sorption coefficients of pharmaceutically

active substances carbamazepine, diclofenac, and ibuprofen in sandy sediments. Chemosphere 60 (2), 245–253.

Skanavis, C., 1999. Groundwater disaster in Puerto Rico—the need for environmental education. J. Environ. Health 62

(2), 29–35.

Stiftung Warentest (Eds.), 2000. Handbuch Medikamente—Uber 5000 Arzneimittel fur Sie bewertet. Stiftung Warentest,

Berlin, 767 pp.

Stuer-Lauridsen, F., Birkved, M., Hansen, L.P., Holten Lutzhøft, H.-C., Halling-Sørensen, B., 2000. Environmental risk

assessment of human pharmaceuticals in Denmark after normal therapeutic use. Chemosphere 40, 783–793.

Syracuse Science Center, 2002. Database of Experimental Octanol–Water Partition Coefficients (Log P)—Homepage

http://esc-plaza.syrres.com/interkow/kowdemo.htm.

Ternes, T.A., 1998. Occurrence of drugs in German sewage treatment plants and rivers. Water Res. 32 (11), 3245–3260.

Tixier, C., Singer, H.P., Oellers, S., Muller, S.R., 2003. Occurrence and fate of carbamazepine, clofibric acid, diclofenac,

ibuprofen, ketoprofen, and naproxen in surface waters. Environ. Sci. Technol. 37 (6), 1061–1068.

Page 17: Traugott 2006

T.J. Scheytt et al. / Journal of Contaminant Hydrology 83 (2006) 53–69 69

Toride, N., Leij, F.J., van Genuchten, M.Th., 1995. The CXTFIT code for estimating transport parameters from laboratory

or field tracer experiments, version 2.1. Research Report, vol. 137. US Salinity Laboratory. 121 pp.

van Genuchten, M.Th., Parker, J.C., 1984. Boundary conditions for displacement experiments through short laboratory

soil columns. Soil Sci. Soc. Am. J. 48, 703–708.

Wierenga, P.J., van Genuchten, M.Th., 1989. Solute transport through small and large unsaturated soil columns. Ground

Water 27, 35–42.

Winkler, M., Lawrence, J.R., Neu, T.R., 2001. Selective degradation of ibuprofen and clofibric acid in two model river

biofilm systems. Water Res. 35, 3197–3205.

Yalkowsky, S.H., Dannenfelser, R.M., 1992. Aquasol Database of Aqueous Solubility—Version 5. College of Pharmacy,

University of Arizona, Tucson, AZ.