w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 2 1 4 – 3 2 2 4

Avai lab le at www.sc iencedi rect .com

journa l homepage : www.e lsev i er . com/ loca te /wat res

Removal of Pharmaceutical and Personal Care Products(PPCPs) under nitrifying and denitrifying conditions

Sonia Suarez, Juan M. Lema, Francisco Omil*

School of Engineering, University of Santiago de Compostela, 15782 Santiago de Compostela, Galicia, Spain

a r t i c l e i n f o

Article history:

Received 21 September 2009

Received in revised form

17 February 2010

Accepted 27 February 2010

Available online 6 March 2010

Keywords:

Fragrances

Hormones

Pharmaceuticals

Redox conditions

Solid retention time

Temperature

* Corresponding author. Tel.: þ34 981 56 31 0E-mail addresses: sonia.suarez@usc.es (S

0043-1354/$ – see front matter ª 2010 Elsevidoi:10.1016/j.watres.2010.02.040

a b s t r a c t

The contribution of volatilization, sorption and transformation to the removal of 16

Pharmaceutical and Personal Care Products (PPCPs) in two lab-scale conventional activated

sludge reactors, working under nitrifying (aerobic) and denitrifying (anoxic) conditions for

more than 1.5 years, have been assessed. Pseudo-first order biological degradation rate

constants (kbiol) were calculated for the selected compounds in both reactors. Faster degra-

dation kinetics were measured in the nitrifying reactor compared to the denitrifying system

for the majority of PPCPs. Compounds could be classified according to their kbiol into very

highly (kbiol> 5 L gSS�1 d�1), highly (1< kbiol< 5 L gSS

�1 d�1), moderately (0.5< kbiol< 1 L gSS�1 d�1)

and hardly (kbiol< 0.5 L gSS�1 d�1) biodegradable.

Results indicated that fluoxetine (FLX), natural estrogens (E1þ E2) and musk fragrances

(HHCB, AHTN and ADBI) were transformed to a large extent under aerobic (>75%) and

anoxic (>65%) conditions, whereas naproxen (NPX), ethinylestradiol (EE2), roxithromycin

(ROX) and erythromycin (ERY) were only significantly transformed in the aerobic reactor

(>80%). The anti-depressant citalopram (CTL) was moderately biotransformed under both,

aerobic and anoxic conditions (>60% and >40%, respectively). Some compounds, as car-

bamazepine (CBZ), diazepam (DZP), sulfamethoxazole (SMX) and trimethoprim (TMP),

manifested high resistance to biological transformation.

Solids Retention Time (SRTaerobic >50 d and <50 d; SRTanoxic >20 d and <20 d) had

a slightly positive effect on the removal of FLX, NPX, CTL, EE2 and natural estrogens

(increase in removal efficiencies <10%). Removal of diclofenac (DCF) in the aerobic reactor

was positively affected by the development of nitrifying biomass and increased from 0% up

to 74%. Similarly, efficient anoxic transformation of ibuprofen (75%) was observed after an

adaptation period of 340 d. Temperature (16–26 �C) only had a slight effect on the removal

of CTL which increased in 4%.

ª 2010 Elsevier Ltd. All rights reserved.

1. Introduction conditions, which can be installed in different compartments

Conventional Sewage Treatment Plants (STPs) have been

designed to simultaneously eliminate organic matter and

nitrogen from urban wastewater by the implementation of

technologies that combine the existence of anoxic and aerobic

0; fax: þ34 981 52 80 50.. Suarez), juan.lema@uscer Ltd. All rights reserved

of the plant (e.g. activated sludge plants), or be sequentially

applied in one single reactor (e.g. sequential batch reactors).

Nowadays, removal mechanisms for conventional

contaminants under different redox conditions are under-

stood in detail and have been efficiently applied in most full-

.es (J.M. Lema), francisco.omil@usc.es (F. Omil)..

Table 1 – Concentration of PPCPs spiked to the reactorfeed (CFeed in mg LL1), CAS number, solid–waterdistribution coefficient (Kd in L kgL1) and Henry’s lawconstant (H in mg mL3 air/mg mL3 wastewater).

Compound CFeed CAS log Kd Ha

Anti-depressants

Fluoxetine (FLX) 20 054910-89-3 0.7b,c 3.6� 10�6

Citalopram (CTL) 20 059729-33-8 2.0b,d 1.1� 10�9

Estrogens

17b-Estradiol (E2) 10 50-28-2 2.8e 1.5� 10�9

17a-Ethinylestradiol (EE2) 10 57-63-6 2.5f 3.3� 10�10

Anti-inflammatories

Ibuprofen (IBP) 10 15687-21-1 0.9f 6.1� 10�6

Naproxen (NPX) 10 22204-53-1 1.1g 1.4� 10�8

Diclofenac (DCF) 10 15307-86-5 1.2f 1.9� 10�10

Anti-epileptic

Carbamazepine (CBZ) 20 298-46-4 0.1f 4.4� 10�9

Antibiotics

Trimethoprim (TMP) 20 738-70-5 2.3h 9.8� 10�13

Roxithromycin (ROX) 20 80214-83-1 2.2i 2.0� 10�29

Sulfamethoxazole (SMX) 20 723-46-6 2.4h 2.6� 10�11

Erythromycin (ERY) 20 114-07-8 2.2b 2.2� 10�27

Musks

Galaxolide (HHCB) 40 1222-05-5 3.3f 5.4� 10�3

Tonalide (AHTN) 40 1506-02-1 3.4f 5.1� 10�3

Celestolide (ADBI) 40 13171-00-1 3.9j 7.3� 10�1

Tranquilizer

Diazepam (DZP) 20 439-14-5 1.3f 1.5� 10�7

a Syracuse Research Corporation (SRC).

b Jones et al. (2002).

c Brooks et al. (2003).

d Stuer-Lauridsen et al. (2000).

e Clara et al. (2004a,b).

f Ternes et al. (2004).

g Urase and Kikuta (2005).

h Gobel et al. (2005).

i Joss et al. (2005).

j Kupper et al. (2006).

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 2 1 4 – 3 2 2 4 3215

scale STPs, although not the same can be said for micro-

pollutants. For the latter, the contribution of anoxic and

aerobic conditions to their overall removal was usually not

analyzed, since only influent and effluent concentrations of

micropollutants have been normally considered in the moni-

toring of STPs (Clara et al., 2005; Gobel et al., 2007; Jones et al.,

2007; Joss et al., 2005). There are some exceptions, as for

example kinetic experiments performed under aerobic,

anoxic and anaerobic redox conditions (Joss et al., 2004), as

well as individual samplings carried out in the denitrification

and nitrification tanks in full-scale STP (Andersen et al., 2003),

in order to analyze the behavior of estrogens (estrone: E1, 17b-

estradiol: E2 and 17a-ethinylestradiol: EE2) under different

redox conditions. It has also been shown that for EE2 the

enrichment of activated sludge in nitrifying bacteria could

enhance its transformation into metabolites devoid of estro-

genic activity (Vader et al., 2000). Additionally, biodegradation

of some other pharmaceuticals (ibuprofen: IBP, diclofenac:

DCF and clofibric acid) in oxic and anoxic biofilm reactors was

investigated by Zwiener et al. (2000).

Understanding the behavior of Pharmaceutical and

Personal Care Products (PPCPs) under different redox condi-

tions is not only essential for achieving deeper knowledge of

the whole wastewater treatment process, but also in order to

predict further pathways of those contaminants once released

into the environment (e.g. groundwater recharge, degradation

in surface water, etc.). In fact, different metabolites could be

formed under aerobic and anoxic conditions, as previously

reported for other pollutants such as nonylphenol ethoxylate

surfactants (Goel et al., 2003) or the pharmaceutical residue

phenazone (Greskowiak et al., 2006), indicating differences in

degradation pathways in anoxic and aerobic processes. This

could explain why some pollutants are better removed under

anoxic conditions despite the common evidence that higher

oxidation potentials in aerobic environments should favor

their degradation. In this way, Drewes et al. (2001) reported

negligible removal of tri-iodinated benzene derivatives (X-ray

contrast media) under aerobic redox conditions, which was

significantly enhanced in anoxic environments.

The objective of the present work was to evaluate the

potential of nitrifying and denitrifying conditions for the

elimination of a set of 16 Pharmaceutical and Personal Care

Products (PPCPs), including pharmaceuticals from different

therapeutic classes, hormones and musk ingredients (Table 1),

in order to better understand the overall removal process for

such micropollutants in full-scale STPs.

2. Materials and methods

2.1. Denitrifying and nitrifying reactors

Two 2-L continuous stirred-tank reactors coupled to a 1-L

settler were inoculated with biomass collected from

a conventional activated sludge pilot plant which was oper-

ated for more than 1.5 years in order to study the fate of the

same set of PPCPs in a different reactor configuration. One

reactor was operated under anoxic conditions, whereas in the

second plant nitrifying aerobic conditions were maintained.

Both reactors have been running at a Hydraulic Retention

Time (HRT) of 1 d.

The anoxic reactor was mechanically stirred and capped in

order to restrict the transfer of oxygen from air to the liquid

phase. Recirculation of biomass from the settler to the reactor

was carried out by means of a peristaltic pump in order to ach-

ieve an external recirculation ratio (Rext) of 30. Since the reactor

was aimed at promoting a denitrification process, a synthetic

feed containing an organic carbon source and nitrate was

continuously pumped into the system at a flow rate of

1.72� 0.35 L d�1 and with the following composition: 500 mg L�1

ofCOD (asNaCH3CO2),40 mg L�1 ofN-NO3,8 mg L�1 ofP-PO4 and

a solution of trace elements (FeCl3, H3BO3, CuSO4, KI, ZnSO4,

CoCl2, MnCl2 at concentrations in the range of 3–150 mg L�1).

In the aerobic plant oxygen was supplied through air

diffusers located at the bottom of the reactor and recirculation

of biomass was carried out by means of an air lift device in

order to maintain a Rext of 40. To achieve nitrifying condi-

tions, the reactor was fed at a flow rate of 1.94� 0.27 L d�1 with

an inorganic carbon source and ammonia at concentrations of

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 2 1 4 – 3 2 2 43216

1000 mg L�1 NaHCO3, 40 mg L�1 of N-NH4, 8 mg L�1 of P-PO4

and the same trace elements as those employed in the anoxic

reactor.

A start-up period, during which no PPCPs were added to the

reactors, was maintained for at least one Solid Retention Time

(SRT) in order to allow acclimation of bacterial population to

the new redox and feeding conditions. After this initial stage,

pharmaceuticals were incorporated to the feed at the

concentrations indicated in Table 1. PPCPs were directly

spiked into the feeding tank from stock solutions of individual

compounds dissolved in methanol (antibiotics and anti-

inflammatories) or acetone (musks, anti-depressants, CBZ

and DZP) at a concentration of 2000 mg L�1 which were stored

in the freezer. Both reactors have been continuously working

during 440 d at constant operation parameters, with two

exceptions. The first is temperature which varied seasonally

(16–26 �C) since it was not controlled (Fig. 1), while the second

is SRT for which two operational periods were distinguished:

SRT >50 d and <50 d; SRT >20 d and <20 d for the nitrifying

and anoxic system, respectively. The variability of SRT is due

to variable losses of biomass within the effluent derived from

the use of a settler as separation unit, but also due to the

fluctuations of biomass concentration inside the reactors

which was especially important in the aerobic plant (Fig. 1). In

the latter, biomass concentration decreased down to

0.5 g VSS L�1 until day 170, most probably due to the promo-

tion of endogenous respiration of the heterotrophic bacteria

time (d)0 100 200 300 400

T (º

C)

5

10

15

20

25

30

VSS

(g.L

-1)

0.0

0.5

1.0

1.5

2.0

2.5A

time (d)

0 100 200 300 400

T (º

C)

15

18

21

24

27

30

VSS

(g.L

-1)

0.0

0.5

1.0

1.5

2.0

2.5B

Fig. 1 – Temperature (>), VSS inside the reactor (:) and in

the effluent (C) of the nitrifying (A) and denitrifying (B)

reactor.

present in the inoculum, in the absence of an organic matter

source. After this initial stage, bacterial population stabilised

and initiated a growth period until day 300, at a rate of

7.8� 10�3 g VSS d�1, achieving a pseudo-steady sludge

concentration of 1.6 g VSS L�1. The yield constant of produced

biomass per amount of ammonia oxidized was 0.1 g VSS g�1

N-NH4, which is in accordance with the stoichiometric

parameter.

2.2. Analytical methods

Both reactors were weekly sampled, including the feed, reac-

tion medium and final effluent, in order to analyze conven-

tional parameters and PPCPs. Total and Volatile Suspended

Solids (TSS and VSS), nitrite and nitrate concentrations were

determined following Standard Methods (APHA, 1999),

ammoniacal nitrogen was measured with a spectrophotom-

eter (Shimadzu UV-1603, UV–visible) at 635 nm and Total,

Inorganic and Organic Carbon (TC, IC and TOC) were deter-

mined by a Shimadzu analyzer (TOC-5000). Additionally,

temperature, pH and dissolved oxygen have been regularly

followed in both reactors.

Analyses of PPCPs were performed after collecting 2-L

samples in glass or aluminum bottles, which were immediately

prefiltered (AP4004705, Millipore) and supplemented with

a pinch of sodium azide (w0.3 g L�1). For the analysis of PPCPs

sample extraction based on Solid Phase Extraction (SPE) or Solid

Phase MicroExtraction (SPME) was used as pre-concentration

technique prior to their quantitative determination. Liquid or

Gas Chromatography coupled to Mass Spectrometry (LC–MS or

GC–MS, respectively) was used for the final quantification.

Analysis of the soluble content of anti-inflammatory

compounds, carbamazepine (CBZ), diazepam (DZP) and

musks was performed following the methodology described

in Rodrıguez et al. (2003), which consists of adjusting the pH of

the samples to 2.5, adding meclofenamic acid and dihy-

drocarbamazepine as surrogate standards, SPE of 250 mL

samples using 60 mg OASIS HLB cartridges (Waters, Milford,

MA, USA) and elution from the cartridge using 3 mL of ethyl

acetate. This extract was divided into two fractions: one of

them was used for direct determination of CBZ, DZP, galax-

olide (HHCB), tonalide (AHTN) and celestolide (ADBI), while

the other one was employed for the analysis of anti-

inflammatories as their tertbutyldimethylsilyl derivatives.

Finally, GC/MS detection was carried out in a Varian CP 3900

chromatograph (Walnut Creek, CA, USA) equipped with

a split–splitless injector and connected to an ion-trap mass

spectrometer. Additionally, total concentrations of musks

were determined by Solid Phase Micro Extraction (SPME)

following the procedure developed by Garcıa-Jares et al. (2002).

Briefly, 10-mL samples were immersed in a bath at 100 �C for

5 min to equilibrate temperature. Then, a PDMS–DVB fiber

(65 mm polydimethylsiloxane–diviylbenzene, Supelco, USA)

was exposed to the headspace over the sample for 25 min.

Once the exposition finished, the fiber was immediately

inserted into the GC injector for chromatographic analysis.

The anti-depressants fluoxetine (FLX) and citalopram (CTL)

were analyzed by GC–MS (Varian Saturn 3) after SPME

according to Lamas et al. (2004). Briefly, water samples were

placed in 22-mL headspace vials. To improve the extraction

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 2 1 4 – 3 2 2 4 3217

a derivatization process was carried out with potassium

hydrogen carbonate and acetic anhydride (acetylation).

Afterwards SPME at 100 �C was carried out using a PDMS-DVB

fiber which was inserted into the GC injection port after

30 min in order to complete the analysis.

Estrogens were analyzed according to Quintana et al.

(2004). Accordingly, the pH of the samples was adjusted to 6

and methanol (1%) and the internal standard, 17b-estradiol-d4

(75 ng L�1), were added to the samples prior to its pre-

concentration by SPE (Oasis HLB 60 mg cartridges).

Cartridges were then dried and eluted with 3 mL of ethyl

acetate. This volume was reduced to approximately 0.3 mL

and further cleaned-up by passing it through a 500 mg Sep-

Pak silica cartridge. Analytes were then eluted with 10 mL of

ethyl acetate, and the extract reduced to 0.1 mL and derivat-

ized with N-methyl-N-(trimethylsilyl)trifluoroacetamide

(MSTFA) at 85 �C for 100 min. GC–MS–MS analysis was carried

out using a Varian CP 3800 gas chromatograph connected to

ion-trap mass spectrometer (Varian Saturn 2000).

For analysis of antibiotics, sulfadimidin-C13 and Caffeine-

C13 were used as internal standards and an acidic buffer

solution was added to the samples prior to sample extraction.

SPE was carried out with Isolute 101 cartridges and analytes

were eluted with methanol and acidic methanol. Final detec-

tion was performed in an LC–MS–MS in the positive ESI mode.

2.3. Mass balances

Mass balances were applied considering stripping, sorption

and degradation as possible removal processes for PPCPs.

Volatilization was only considered in the mass balances of

the aerobic reactor, since the anoxic plant was not aerated.

The relative fraction stripped to the gas phase was calculated

according to Eq. (1):

C�j;air

Cj;total¼ H$qair

1þ Kd;j$TSSþH$qair(1)

where C*j,air is the concentration of compound j that leaves the

reactor during aeration referred to the volume of wastewater

treated and Cj,total is the total concentration of compound j

(both in mg L�1), H the dimensionless Henry’s law constant, qair

the aeration applied per unit of wastewater treated

(Lair L�1wastewater), Kd,j the solid–water distribution coefficient

of compound j (L kg�1) and TSS the suspended solids

concentration inside the reactor (kg L�1).

According to Eq. (1) volatilization is not significant for any

compound (<5%, even for the relatively volatile compounds

galaxolide: HHCB and tonalide AHTN), except for celestolide

(58%), even under worst-case conditions (i.e. qair of

15 Lair L�1wastewater and TSS inside the aerobic plant of 0.8 g L�1).

For the latter, the mass flow that leaves the aerobic reactor due

to volatilization (Fj,Stripped in mg d�1) has been calculated

according to Eq. (2) and included in the mass balances:

Fj;Stripped ¼ C�j;air$Q (2)

being Q in Eq. (2) the feed flow rate in the aerobic reactor

(L d�1).

The mass flow of compound j leaving the reactors sorbed

onto solids contained in the final effluent (Fj,Sol in mg d�1) has

been estimated applying Eq. (3), where sorption equilibrium is

assumed:

Fj;Sol ¼ Q$Kd;j$TSSEff$Cj;Eff (3)

where, Cj,Eff is the dissolved concentration of compound j

(mg L�1) and TSSEff the suspended solids concentration (kg L�1),

both measured in the effluent.

Sorption coefficients (Kd,j) considered in the mass balances

have been taken from the literature (Table 1) with the excep-

tion of fragrances (HHCB, AHTN and celestolide: ADBI) for

which this parameter has been calculated from experimental

total and soluble concentrations measured in the effluent

(Ternes et al., 2004). Priority was always given to experimen-

tally determined values, although in the case of the two anti-

depressants (FLX and CTL) this parameter had to be estimated

from their KOW, following the procedure described in Jones

et al. (2002). It is worth to highlight that Kd values reported

by Gobel et al. (2005) for SMX vary up to 70%, being the reasons

for this not completely clear. Taking into account the physico-

chemical properties of this compound, sorption is not expec-

ted to be significant under the operational conditions

considered. On the one hand, lipophilic interactions with the

lipid fraction of the sludge (i.e. absorption) should not be

significant according to the log Kow of 0.9 reported for SMX

(Syracuse Research Corporation). Compounds can also sorb

onto sludge through the establishment of electrostatic inter-

actions between a positively charged compound and the

negatively charged surface of microorganisms (i.e. adsorp-

tion). According to Lin et al. (1997), SMX exhibits a positively

charged amino group at a very acidic pH (pKa1 1.8), while at pH

higher than 5.7 (pKa2) a loss of the sulfonamide proton yields

to its negatively charged conjugate, thus indicating that

interactions between SMX and sludge are not probable.

Total mass flow of compound j in the effluent (Fj,Eff in

mg L�1) has been calculated as the sum of the flow in the liquid

and the solid phase (Eq. (4)).

Fj;Eff ¼ Cj;Eff$Q$�1þ Kd;j$TSSEff

�(4)

In the case of influents, the mass flow (Fj,Feed in mg L�1) has

been calculated assuming that sorption is negligible, since the

synthetic feed did not contain any solid particles:

Fj;Feed ¼ Cj;Feed$Q (5)

where, Cj,Feed is the concentration of compound j in the feed

(mg L�1).

Assuming steady-state conditions for the reactors, biolog-

ical transformation can be calculated according to Eq. (6):

Ej;Anox ¼Fj;Feed � Fj;Eff

Fj;Feed$100

Ej;Aer ¼Fj;Feed �

�Fj;Eff þ Fj;Stripped

�

Fj;Feed$100

ð6Þ

where, Ej,Anox and Ej,Aer are the transformation efficiencies (%)

for compound j in the anoxic and aerobic reactors, respectively.

2.4. Degradation kinetics

Biological transformation of PPCPs can be described by

pseudo-first order kinetics (Joss et al., 2006):

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 2 1 4 – 3 2 2 43218

rS ¼ �dCj;total

dt¼ kbiol$VSS$Cj;dissolved (7)

where Cj,total and Cj,dissolved are the total and dissolved

concentrations of compound j (mg L�1), t the time (d), kbiol is the

reaction rate constant (L g�1 d�1) and VSS the volatile sus-

pended solids concentration (g L�1).

For calculating the pseudo-first order degradation rate

constant (kbiol) for the selected PPCPs the processes were

adjusted to a steady-state, continuous stirred-tank reactor

model:

Fj;Feed ��Fj;Eff þ Fj;Stripped

�� rS$V ¼ 0 (8)

where V is the reaction volume (L).

3. Results and discussion

3.1. Overall removal of PPCPs in the reactors

The efficiency of the aerobic reactor regarding nitrification

increased from w40% up to w90% during the first 30 d of

operation and remained afterwards stable around 98%. Simi-

larly, the anoxic reactor showed to be very efficient in the

removal of nitrate as nitrogen gas, reaching almost 100% after

the first 50 d of operation.

Mean concentrations of PPCPs in the influents and effluents

from both reactors together with their overall removal from the

liquid phase are summarized in Table 2. Data from the whole

operation period, without distinguishing between different

conditions of temperature and SRT, were considered in the

calculation of the mean, which led in some cases to significant

standard deviations. In all cases, removal was calculated from

measured influentconcentrations, rather thanfromthosespiked

to the feeding tank (Table 1). The differences between both could

be attributed to losses of the spiked compounds due to degra-

dation or sorption within the feeding system.

Table 2 – Mean concentrations of PPCPs and standard deviationand anoxic reactors (ppb or mg LL1). Removal from the liquid paveraged over the whole operational period. LOD: Limit of Det

PPCP LOD (ng L�1) Aerobic reactor

CFeed CEff

E1þ E2 0.7 6.6� 1.4 0.07� 0.04

IBP 27 8.1� 1.5 0.4� 0.3

HHCB 23 13.4� 8.6 0.4� 0.3

AHTN 23 15.5� 11.0 0.7� 0.3

ADBI 23 15.8� 7.0 0.4� 0.3

FLX 17 13.4� 3.7 1.0� 0.2

ROX 1.2 17.4� 5.9 1.6� 0.6

ERY 1.2 17.7� 2.2 2.0� 0.7

EE2 1.7 5.5� 2.6 0.8� 0.9

NPX 27 9.5� 0.9 1.3� 0.5

CTL 15 13.0� 7.4 4.5� 2.3

DCF 100 8.2� 1.9 6.2� 2.7

SMX 2.6 21.1� 1.6 16.4� 0.1

DZP 230 16.1� 4.1 13.3� 3.5

TMP 2.9 19.3� 1.0 16.4� 1.0

CBZ 470 19.0� 4.9 18.1� 5.9

n.a. not analyzed.

Ten of the considered PPCPs were removed to a high degree

(>85%) in the aerobic reactor, comprising hormones (E1þ E2

and EE2), the anti-inflammatory drugs IBP and naproxen

(NPX), the three musks (HHCB, AHTN, and ADBI), the anti-

depressant fluoxetine (FLX) and two antibiotics (roxi-

thromycin: ROX and erythromycin: ERY). These high removal

efficiencies occurring in aerobic treatment plants regarding

hormones, anti-inflammatory drugs and fragrances have

been already reported by several authors (Baronti et al., 2000;

Gomez et al., 2007; Joss et al., 2004; Kupper et al., 2006; Nakada

et al., 2006; Simonich et al., 2002), whereas eliminations

previously determined for FLX, ROX and ERY (Castiglioni et al.,

2006; Gobel et al., 2007; Joss et al., 2005; Vasskog et al., 2006)

were significantly lower than those measured in the present

work. The anoxic reactor has shown to be able to remove

fragrances, FLX and natural estrogens (E1þ E2) in an effective

way, although in a slightly lower degree (>70%) compared to

the aerobic reactor, whereas for the rest of these ten

compounds removal achieved was much less effective (<40%).

Anti-depressant citalopram (CTL) has been partially

removed in both reactors (60% and 44% in the aerobic and

anoxic reactor, respectively), similar to the overall removal

efficiencies determined by Vasskog et al. (2006) during one

sampling campaign performed at three STPs in Norway.

The rest of pharmaceuticals (DCF, sulfamethoxazole: SMX,

diazepam: DZP, trimethoprim: TMP and carbamazepine: CBZ)

have not been significantly transformed (<25%) by the biolog-

ical treatment with neither nitrifying nor denitrifying bacteria.

The high persistence of these compounds has also been

observed in different full-scale STP (Clara et al., 2004a,b; Gobel

et al., 2007; Lindberg et al., 2005; Lindqvist et al., 2005), with the

exception of SMX for which removal efficiencies reported

varied in a wide range. For example, eliminations of 0–84% and

(-138)–60% can be found in Castiglioni et al. (2006) and Gobel

et al. (2007), respectively, although this could be partially due

to the fact that real wastewaters, which have a more complex

s (n [ 10) in the feed (CFeed) and effluent (CEff) of the aerobichase (%) calculated for each time point, and afterwardsection of the analytical method.

Anoxic reactor

Removal CFeed CEff Removal

99� 0 6.4� 0.8 1.8� 0.3 72� 2

95� 4 8.0� 0.7 5.1� 2.0 37� 26

92� 12 10.2� 8.2 0.5� 0.2 86� 15

90� 13 11.7� 9.8 0.9� 0.1 82� 16

97� 2 10.9� 7.7 0.6� 0.2 88� 15

92� 3 14.6� 6.5 2.2� 0.8 84� 6

91� 0 18.8� 1.2 15.4� 2.5 15� 7

89� 2 23.9� 0.1 19.1� 2.3 20� 10

87� 11 5.8� 1.9 4.6� 1.4 20� 13

86� 5 9.0� 1.1 8.1� 0.4 9� 13

60� 17 16.0� 4.3 9.0� 3.0 44� 9

22� 28 6.4� 0.9 6.2� 0.6 2� 5

22� 5 n.a. n.a. –

17� 11 15.3� 5.8 12.2� 3.4 16� 17

14� 10 n.a. n.a. –

6� 12 17.9� 4.8 17.9� 5.6 1� 10

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 2 1 4 – 3 2 2 4 3219

matrix, were used in these works. For the two antibiotics (SMX

and TMP) the behavior in the anoxic reactor could not be

established, since none of these compounds have been detec-

ted in the feed or effluent of the reactor, although both of them

had been spiked to the synthetic feed. The reason for that is not

completely clear, although interferences with other

compounds existing in the feed might be a plausible reason.

3.2. Fate of PPCPs in the reactor. Application of massbalances

Mass balances were applied to each compound, according to

Eqs. (2)–(6) and results have been graphically represented,

including the contribution of both, biological transformation

(Ej,Anox and Ej,Aer) and sorption (Fj,sol/Fj,Feed� 100) to the overall

removal of PPCPs. Moreover, volatilization has also been

considered in the case of ADBI in the aerobic plant (Fj,Stripped/

Fj,Feed� 100). The residual fraction of each compound that

leaves the reactors with the effluent has been also included in

the plots (Cj,Eff�Q/Fj,Feed� 100). The whole operation period of

both reactors has been classified according to temperature

(low: 16–20 �C and high: 20–26 �C) and SRT. Additionally,

pseudo-first order degradation rate constants, kbiol, were

calculated (Table 3) according to Eqs. (7) and (8) once stable

operational conditions were achieved and specific nitrifying

and denitrifying biomass was developed inside the reactors.

Faster degradation kinetics were measured in the nitrifying

reactor compared to the denitrifying system for the majority

of PPCPs, with the exception of fragrances for which very high

kbiol were determined in both reactors. According to their kbiol

the selected PPCPs could be classified as very highly

(kbiol> 5 L gSS�1 d�1), highly (1< kbiol< 5 L gSS

�1 d�1), moderately

(0.5< kbiol< 1 L gSS�1 d�1) and hardly (kbiol< 0.5 L gSS

�1 d�1) biode-

gradable, which in turn led to removal efficiencies of >80%,

60–80%, 40–60% and <40%, respectively (Table 3).

Table 3 – Summary of PPCP transformations obtained in the norder reaction rate constants (kbiol in L gSS

L1 dL1) estimated uponinfluence of SRT, T and other sludge characteristics.

Compound Transformation kbi

Aerobic Anoxic Aerobic

HHCB þþ þ/þþ 170

E1þ E2 þþ þ 170

AHTN þþ þ/þþ 115

ADBI þþ þ/þþ 75

IBP þþ ��/þ 20

EE2 þþ ��/� 20

FLX þþ þþ 9

ROX þþ �� 9

ERY þþ ��/� 6

CTL þ �þ 3

NPX þþ �� 9

DCF ��/þ �� 1.2

SMX � n.a. 0.3

DZP �� �� <0.4

TMP � n.a. 0.15

CBZ �� �� <0.06

(��)<20%; (�) 20–40%; (�þ) 40–60%; (þ) 60–80%; (þþ)>80%; n.a. not analyz

sludge (Sludge) on the transformation degree is indicated as (yes) or (no)

Diclofenac showed to be very persistent during treatment in

the anoxic reactor (kbiol< 0.04 L gSS�1 d�1), but this was not the

case inside the aerobic plant. For the latter, the data trend in

Fig. 2B seems to indicate that there has been an initial adap-

tation period that coincides with the death and washout of

heterotrophic bacteria (w170 d) during which removal of DCF

increased from 0% to 25%. Afterwards, a correlation between

sludge concentration in the reactor and biological trans-

formation of DCF was observed, reaching maximum removals

of around 74%. The high deviations observed in Fig. 2A for

aerobic conditions, indicate that the behavior of DCF was

influenced to a higher extent by the biomass developed in the

system than by the operation temperature and SRT. The fate of

DCF under anoxic and oxic conditions has been investigated by

Zwiener et al. (2000) in biofilm reactors who found removal

efficiencies below 20% under both conditions. Taking into

account that those biofilm reactors had been inoculated with

municipal sewage sludge and that operation stopped after only

120 d, these results are comparable to those obtained in the

present research during the first months.

Naproxen and FLX were both biologically transformed to

a high degree in the aerobic reactor (Fig. 3A and B, respec-

tively) with less than 16 and 9% of residual mass flow in the

effluent, respectively. While FLX exhibited significant trans-

formation in the anoxic treatment (79–89%), that was not the

case of NPX. Fluoxetine concentrations have been measured

in the influent and effluent of three STPs in Norway (Vasskog

et al., 2006) being the removal efficiencies reported in the

range of 8–70% for FLX, which are below the values measured

in the nitrifying reactor of the present study. This could be due

to the higher proportion of nitrifiers in this plant, although no

definite conclusion can be made since data for comparison are

very scarce. A positive effect of increasing the SRT of the

reactor has been observed for FLX in the anoxic reactor (kbiol

w5 and w2.5 L gSS�1 d�1 at SRT >20 d and <20 d, respectively)

itrifying and denitrifying reactors, biological pseudo-firstachievement of stable operational conditions and observed

ol (L gSS�1 d�1) Influence

Anoxic SRT T Sludge

150 No No No

w3 Yes No No

60 No No No

75 No No No

w1.5 No No Yes

0.4 Yes No No

w5 Yes No No

0.2 n.a. n.a. No

0.15 n.a. n.a. No

0.5 Yes Yes No

<0.2 Yes No No

<0.04 No No Yes

n.a. n.a. n.a. No

<0.25 No No No

n.a. n.a. n.a. No

<0.03 No No No

ed. The influence of SRT, temperature (T ) and other characteristics of

.

0

20

40

60

80

100

0

0.5

1

1.5

2

2.5

0 100 200 300 400 500

IBP

rem

oval

(%)

VSS

(g/L

)

time (d)

-10

10

30

50

70

0 0.5 1 1.5 2 2.5

VSS (g/L)

DC

F re

mov

al (%

)

t<100d

t: 170-340d

t>400d

0%

20%

40%

60%

80%

100%

120%

<50d >50d >50d <20d >20d

DC

F fa

te

high high low high high

Aerobic Anoxic

T:

SRT:

0%

20%

40%

60%

80%

100%

120%

<50d >50d >50d <20d >20d <20d

IBP

fate

high high low high high low

Aerobic Anoxic

T:

SRT:

DC

A B

Fig. 2 – A, C) Fate of DCF and IBP in the aerobic and anoxic reactors, indicating the contribution of biological transformation

( ), sorption (-) and release within the effluent (,). B) Correlation between removal of DCF and biomass concentration in

the aerobic reactor for the different sampling dates (t). D) Correlation between removal of IBP in the anoxic reactor (C) and its

biomass concentration (>).

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 2 1 4 – 3 2 2 43220

and NPX in the aerobic plant, although for the latter the

increase in removal was very slight (3%). At the operational

conditions of these reactors it is not expected that an increase

in the SRT will lead to changes in the bacterial population,

although, as described in Omil et al. (2010) the enzymes

produces by them could act in a different way. For example,

Jones et al. (2007) propose that systems operating at high SRT

could promote the release of less specific enzymes due cell

lysis. Acclimation phenomena are also possible to occur,

during which microorganisms present in a given system are

able to degrade at a larger extent certain pollutants after

a period of time due to the broadening of their enzymatic

spectrum as has been shown for different xenobiotic (Layton

et al., 2000; Zwiener et al., 2000).

Ibuprofen and NPX exhibit both low sorption potential and

high aerobic biological degradation constants (kbiol IBP:

w20 L gSS�1 d�1; kbiol NPX: w9 L gSS

�1 d�1). As expected from their

kbiol, aerobic transformation of IBP was slightly better than for

NPX, between 93 and 96%. When the fate of IBP in the anoxic

process was analyzed according to operational conditions of

temperature and SRT, high deviations were observed (Fig. 2C),

thus, to better analyze its behavior, removal was plotted as

a function of time (Fig. 2D). The evolution of the biological

degradation rate constant with time (kbiol: 0.1 L gSS�1 d�1 between

day 0 and 200; 0.3 L gSS�1 d�1 from day 200 to 350 and 1.5 L gSS

�1 d�1

afterwards) indicates that adaptation of bacteria was respon-

sible for the wide range of transformation efficiencies

measured (16–75%). Adaptation of bacteria seems not to be

related to the presence of the pharmaceutical in the waste-

water, since biomass was taken from a reactor that had already

been fed with IBP for six months. Therefore, adaption was more

plausible due to changes in the characteristics of bacterial

population (e.g. its enzymatic spectrum) which developed

from a standard heterotrophic to a strict denitrifying biomass.

Zwiener et al. (2000) measured removals for IBP of more than

90% under oxic and of 15% under anoxic conditions, although,

as stated previously, they could have missed adaptation due to

a too early stop of the reactors (120 d).

Natural estrogens (E1 and E2) were highly transformed (99%)

under aerobic conditions and even in the anoxic reactor

transformation was significant (69–73%, Fig. 3C). Accordingly,

the calculated biological degradation constants were very high

in the aerobic and anoxic reactors, respectively (kbiol aerobic:

170 L gSS�1 d�1 and kbiol anoxic: 3 L gSS

�1 d�1). In the anoxic reactor,

a slight increase in the transformation degree was observed

when increasing the SRT of the plant (kbiol from 2.2 to

2.7 L gSS�1 d�1). Important eliminations of natural estrogens

under denitrifying conditions in full-scale STPs have also been

0%

20%

40%

60%

80%

100%

120%

<50d >50d >50d <20d >20d

NP

X fa

te

high high low high high

Aerobic Anoxic

T:

SRT:

0%

20%

40%

60%

80%

100%

120%

<50d >50d >50d <20d >20d <20d

FLX

fate

high high low high high low

Aerobic Anoxic

T:

SRT:

0%

20%

40%

60%

80%

100%

120%

<50d >50d >50d <20d >20d

EE

2 fa

te

high high low high high

Aerobic Anoxic

T:

SRT:

0%

20%

40%

60%

80%

100%

120%

<50d >50d >50d <20d >20d

E2+

E1

fate

high high low high high

Aerobic Anoxic

T:

SRT:

A B

C D

Fig. 3 – Fate of NPX, FLX, natural hormones (E1 D E2) and EE2 in the aerobic and anoxic reactors, indicating the contribution

of biological transformation ( ), sorption (-) and release within the effluent (,).

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 2 1 4 – 3 2 2 4 3221

reported by Andersen et al. (2003). The synthetic hormone

ethinylestradiol was only transformed appreciably in the

aerobic reactor (82–90%), whereas under anoxic conditions less

than 26% of the parent compound was degraded (Fig. 3D). This

behavior fits very well with the kinetic behavior of EE2 deter-

mined by Joss et al. (2004) for different redox conditions, where

it was shown that EE2 was removed at a significant rate only

under aerobic conditions. This was additionally observed in

combined anoxic/aerobic treatment plants (Andersen et al.,

2003). The role of ammonium monooxygenase (AMO), which

is the enzyme that catalyses the first step in nitrification, in the

degradation of EE2 was studied by different authors (Forrez

et al., 2008; Shi et al., 2004; Vader et al., 2000; Yi and Harper,

2007). Nitrifying sludge was reported to enhance trans-

formation of EE2, via hydroxylation that converts EE2 into

hydrophilic products devoid of estrogenic activity (Vader et al.,

2000), which was confirmed in the present work by comparing

kbiol obtained before the washout of heterotrophic bacteria

from the aerobic reactor and after the development of nitri-

fying sludge (kbiol: 9 and 20 L gSS�1 d�1, respectively). This was

additionally supported by Shi et al. (2004), who studied the role

of ammonia oxidizing bacteria on the degradation rate

constants of these estrogens, observing that when they are

inhibited, degradation of estrogens still occurs but at a slower

rate, being this specially significant for EE2. Yi and Harper

(2007) determined a linear relationship between nitrification

and EE2 removal in enriched nitrifying cultures. The biological

degradation rate constant estimated for heterotrophic sludge

fits well with that reported by Joss et al. (2004) concerning

aerobic batch experiments with sludge from a conventional

activated sludge treatment plant (kbiol: 8� 2 L gSS�1 d�1). The

transformation efficiency for EE2 in the aerobic reactor

increased around 8% when the plant was operated at higher

SRT.

In the case of antibiotics only two sampling campaigns

have been carried out, both at the same conditions of low

temperature and SRT below 20 d (Fig. 4). Roxithromycin and

erythromycin were transformed very efficiently (w90%) in the

aerobic reactor and to a larger extent than at similar SRTs in

full-scale STP (Gobel et al., 2007). This might point to a higher

affinity of nitrifying bacteria towards these compounds,

especially when comparing kbiol obtained in the present work

(kbiol ROX: 9 L gSS�1 d�1 and kbiol ERY: 6 L gSS

�1 d�1) with those

determined for activated sludge in Joss et al. (2006). On the

other hand, only slight transformations of these two antibi-

otics have been observed in the anoxic reactor (<27%). A slight

preference for a metabolism under oxic conditions for ROX

was also reported by Heberer et al. (2008) when studying the

influence of redox conditions on the elimination of antimi-

crobial residues during bank filtration. The other two antibi-

otics considered (SMX and TMP) have shown higher

persistence towards aerobic biological treatment, since the

maximum transformation observed was 26% and 21% for SMX

and TMP, respectively.

0%

20%

40%

60%

80%

100%

120%

ROX ERY-H2O

SMX TMP ROX ERY-H2O

Ant

ibio

tics

fate

Aerobic Anoxic

Fig. 4 – Fate of antibiotics (ROX, ERY, SMX and TMP) in the

aerobic and anoxic reactors, indicating the contribution of

biological transformation ( ), sorption (-) and release

within the effluent (,).

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 2 1 4 – 3 2 2 43222

Transformation of CTL in the aerobic reactor increased

from 62% up to 70% when SRT was raised to above 50 d and it

also improved a 4% when operating at higher temperatures.

When comparing this data with those included in Vasskog

et al. (2006), again an increase in removal efficiencies seems

to be achieved when enriching activated sludge in nitrifying

bacteria, although, as stated for Fluoxetine, this should be

confirmed with more data about CTL which are currently not

0%

20%

40%

60%

80%

100%

120%

<50d >50d >50d <20d >20d <20d

HH

CB

fate

high high low high high low

Aerobic Anoxic

T:

SRT:

0%

20%

40%

60%

80%

100%

120%

<50d >50d >50d

AD

BI f

ate

high high low

Aerobic

T:

SRT:

Fig. 5 – Fate of HHCB, AHTN and ADBI in the aerobic and anoxi

transformation ( ), sorption (-), release within the effluent (,

available in the literature. The efficiency of the anoxic plant

was somewhat lower, although still quite significant (41–46%).

For musk compound ADBI the three removal mechanisms,

i.e. volatilization, sorption and biodegradation, have to be

considered in the mass balance applied to the aerobic reactor

(Fig. 5), whereas for the other two fragrances (HHCB and

AHTN) volatilization could be neglected. Sorption coefficients

(Kd) applied in Eq. (4) have been experimentally determined

from measured soluble and total concentrations of fragrances.

The following results were obtained: 1.5� 103 and

4.9� 103 L kg�1 for HHCB, 2.0� 103 and 3.4� 103 L kg�1 for

AHTN and 5.2� 102 and 3.9� 103 L kg�1 for ADBI, in the

aerobic and anoxic reactor, respectively. These values are in

the range of 2� 103–1� 104 L kg�1 previously reported for

activated sludges (Joss et al., 2005; Kupper et al., 2006; Ternes

et al., 2004), except for ADBI in the aerobic plant where the

measured sorption coefficient was somewhat lower. For these

three compounds, sorption coefficients determined for the

anoxic reactor were somewhat higher than those obtained for

the aerobic unit, which is a factor commonly not considered

when applying mass balances in full-scale STP. Therefore, it is

strongly recommended to measure those coefficients for each

particular situation, especially in the case of highly lipophilic

compounds, at least until the significant discrepancies

between published Kd values and environmental factors that

can affect this partition coefficient are better understood.

Transformation of HHCB and AHTN during aerobic bio-

logical treatment reached 79–99% and 76–98%, respectively,

0%

20%

40%

60%

80%

100%

120%

<50d >50d >50d <20d >20d <20d

AH

TN fa

te

high high low high high low

Aerobic Anoxic

T:

SRT:

<20d >20d <20d high high low

Anoxic

c reactors, indicating the contribution of biological

) and for ADBI volatilization in the aerobic reactor ( ).

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 2 1 4 – 3 2 2 4 3223

whereas in the anoxic plant the efficiency was slightly lower,

in the range of 67–84% and 65–76%, respectively. The fraction

of these substances that left the reactors sorbed onto the

solids contained in the effluent was negligible in the aerobic

reactor and below 18% in the anoxic one (Fig. 5). Residual

concentration of ADBI in the aerobic effluent was below 4%,

being the two most significant removal pathways biological

transformation (80–96%) and volatilization (3–16%), since

contribution of sorption was negligible (Fig. 5). The anoxic

reactor also demonstrated high efficiency in the overall

removal of ADBI (78–93%), although in this case sorption (7–

10%) and biodegradation (69–89%) were the responsible

removal processes (Fig. 4). The very high biological degrada-

tion rate constants determined for fragrances (kbiol aerobic:

>75 L gSS�1 d�1 and kbiol anoxic: >60 L gSS

�1 d�1) does not corre-

spond with the behavior of fragrances described in previous

works, where their removal was associated to sorption rather

than to biological transformation (Bester, 2004; Joss et al.,

2005; Kupper et al., 2006). These differences could be a conse-

quence of working with higher SRT in the present reactors

compared to the mentioned works, where SRT below 25 d

were considered. This assumption is based on the consider-

ation that for lipophilic compounds, such as fragrances, the

retention time inside the reactor is determined by the SRT,

rather than by the HRT of the plant. Similar biotransformation

percentages to the measured in this study have been reported

by Clara et al. (2005) for AHTN and HHCB in a STP working at

SRT >52 d, which supports the previous discussion.

4. Conclusions

Considering the main removal mechanisms for PPCPs only

(bio)transformation was significant for the majority of

compounds. In the case of musk fragrances, a significant

fraction (7–18%) of compounds left the anoxic reactor sorbed

onto solids, whereas sorption was negligible in the case of the

aerobic plant, which was associated to the better settling

characteristics of the nitrifying biomass developed in that

reactor. Volatilisation was only significant for ADBI and

contributed between 3 and 16% to the removal of this

substance in the aerobic system.

The selected PPCPs could be classified according to their

aerobic and anoxic biodegradability as follows:

� Highly biodegradable under aerobic and anoxic conditions:

IBP, FLX, natural estrogens (E1þ E2) and musk fragrances

(HHCB, AHTN and ADBI).

� Highly biodegradable under aerobic conditions but persis-

tent in the anoxic reactor: DCF, NPX, EE2, ROX and ERY.

� Moderately biodegradable under aerobic and anoxic condi-

tions: CTL.

� Resistant to biological transformation: SMX, TMP, CBZ

and DZP.

The positive effect of increasing SRT has been demon-

strated for several compounds, with improvements in the

removal efficiency of w10% in the case of FLX, CTL and EE2,

whereas temperature only affected very slightly the removal

of CTL (4%). The characteristics of bacteria developed in the

reactors influenced very significantly the fate of DCF and IBP

in the plants. In fact, removal of those compounds was only

achieved after the growth of specific bacteria. The enrichment

of the activated sludge used as inoculums in nitrifying

bacteria has shown to enhance transformation of EE2, ROX

and ERY, which are compounds moderately transformed in

conventional activated sludge plants.

Acknowledgments

This work was supported by Spanish Ministry of Education

and Science (MICROFARM project: CTQ2007-66265/PPQ,

NOVEDAR_Consolider project: CSD2007-00055 and Research

Fellowship). The authors thank Oliver Gans (Austrian Federal

Environment Agency) for the analyses of antibiotics.

r e f e r e n c e s

Andersen, H., Siegrist, H., Halling-Sorensen, B., Ternes, T.A., 2003.Fate of estrogens in a municipal sewage treatment plant.Environmental Science & Technology 37, 4021–4026.

Baronti, C., Curini, R., D’Ascenzo, G., Di Corcia, A., Gentili, A.,Samperi, R., 2000. Monitoring natural and synthetic estrogens atactivated sludge sewage treatment plantsand in a receiving riverwater. Environmental Science & Technology 34 (24), 5059–5066.

Bester, K., 2004. Retention characteristics and balanceassessment for two polycyclic musk fragrances (HHCB andAHTN) in a typical German sewage treatment plant.Chemosphere 57 (8), 863–870.

Brooks, B.W., Foran, C.M., Richards, S.M., Weston, J., Turner, P.K.,Stanley, J.K., Solomon, K.R., Slattery, M., La Point, T.W., 2003.Aquatic ecotoxicology of fluoxetine. Toxicology Letters 142,169–183.

Castiglioni, S., Bagnati, R., Fanelli, R., Pomati, F., Calamari, D.,Zuccato, E., 2006. Removal of pharmaceuticals in sewagetreatment plants in Italy. Environmental Science &Technology 40 (1), 357–363.

Clara, M., Kreuzinger, N., Strenn, B., Gans, O., Kroiss, H., 2005. Thesolids retention time – a suitable design parameter to evaluatethe capacity of wastewater treatment plants to removemicropollutants. Water Research 39 (1), 97–106.

Clara, M., Strenn, B., Kreuzinger, N., 2004a. Carbamazepine asa possible anthropogenic marker in the aquatic environment:investigations on the behaviour of Carbamazepine inwastewater treatment and during groundwater infiltration.Water Research 38 (4), 947–954.

Clara, M., Strenn, B., Saracevic, E., Kreuzinger, N., 2004b.Adsorption of bisphenol-A, 17 beta-estradiole and 17 alpha-ethinylestradiole to sewage sludge. Chemosphere 56, 843–851.

Clesceri, L.S., Greenberg, A.E., Eaton, A.D., 1999. StandardMethods for the Examination of Water and Wastewater, 20thed. APHA-AWWA-WPCF, Washington, DC.

Drewes, J.E., Fox, P., Jekel, M., 2001. Occurrence of iodinated X-raycontrast media in domestic effluents and their fate duringindirect potable reuse. Journal of Environmental Science andHealth, Part A 36 (9), 1633–1645.

Forrez, I., Carballa, M., Boon, N., Verstraete, W., 2008. Biologicalremoval of 17a-ethinylestradiol (EE2) in an aerated nitrifyingfixed bed reactor during ammonium starvation. Journal ofChemical Technology & Biotechnology 84 (1), 119–125.

Garcıa-Jares, C., Llompart, M., Polo, M., Salgado, C., Macıas, S.,Cela, R., 2002. Optimisation of a solid-phase microextraction

w a t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 3 2 1 4 – 3 2 2 43224

method for synthetic musk compounds in water. Journal ofChromatography A. 963, 277–285.

Gobel, A., Thomsen, A., McArdell, C.S., Joss, A., Giger, W., 2005.Occurrence and sorption behavior of sulfonamides,macrolides, and trimethoprim in activated sludge.Environmental Science & Technology 39, 3981–3989.

Gobel, A., McArdell, C.S., Joss, A., Siegrist, H., Giger, W., 2007. Fateof sulfonamides, macrolides, and trimethoprim in differentwastewater treatment technologies. Science of the TotalEnvironment 372 (2-3), 361–371.

Goel, A., Muller, M.B., Sharma, M., Frimmel, F.H., 2003.Biodegradation of nonylphenol ethoxylate surfactants inbiofilm reactors. Acta Hydrochimica et Hydrobiologica 31 (2),108–119.

Gomez, M.J., Martinez Bueno, M.J., Lacorte, S., Fernandez-Alba, A.R.,Aguera, A., 2007. Pilot survey monitoring pharmaceuticals andrelated compounds in a sewage treatment plant located on theMediterranean coast. Chemosphere 66 (6), 993–1002.

Greskowiak, J., Prommer, H., Massmann, G., Nutzmann, G., 2006.Modeling seasonal redox dynamics and the correspondingfate of the pharmaceutical residue phenazone during artificialrecharge of groundwater. Environmental Science &Technology 40 (21), 6615–6621.

Heberer, T., Massmann, G., Fanck, B., Taute, T., Dunnbier, U.,2008. Behaviour and redox sensitivity of antimicrobialresidues during bank filtration. Chemosphere 73 (4), 451–460.

Jones, O.A.H., Voulvoulis, N., Lester, J.N., 2002. Aquaticenvironmental assessment of the top 25 English prescriptionpharmaceuticals. Water Research 36, 5013–5022.

Jones, O.A.H., Voulvoulis, N., Lester, J.N., 2007. The occurrenceand removal of selected pharmaceutical compounds ina sewage treatment works utilising activated sludgetreatment. Environmental Pollution 145, 738–744.

Joss, A., Zabczynski, S., Gobel, A., Hoffmann, B., Loffler, D.,McArdell, C.S., Ternes, T.A., Thomsen, A., Siegrist, H., 2006.Biological degradation of pharmaceuticals in municipalwastewater treatment: proposing a classification scheme.Water Research 40 (8), 1686–1696.

Joss, A., Andersen, H., Ternes, T., Richle, P.R., Siegrist, H., 2004.Removal of estrogens in municipal wastewater treatmentunder aerobic and anaerobic conditions: consequences forplant optimization. Environmental Science & Technology 38(11), 3047–3055.

Joss, A., Keller, E., Alder, A.C., Gobel, A., McArdell, C.S., Ternes, T.,Siegrist, H., 2005. Removal of pharmaceuticals and fragrances inbiological wastewater treatment. Water Research 39, 3139–3152.

Kupper, T., Plagellat, C., Braendli, R.C., de Alencastro, L.F.,Grandjean, D., Tarradellas, J., 2006. Fate and removal ofpolycyclic musks, UV filters and biocides during wastewatertreatment. Water Research 40 (14), 2603–2612.

Lamas, J.P., Salgado-Petinal, C., Garcıa-Jares, C., Llompart, M.,Cela, R., Gomez, M., 2004. Solid-phase microextraction–gaschromatography–mass spectrometry for the analysis ofselective serotonin reuptake inhibitors in environmentalwater. Journal of Chromatography A. 1046 (1–2), 241–247.

Layton, A.C., Gregory, B.W., Seward, J.R., Schultz, T.W., Sayler, G.S.,2000. Mineralization of steroidal hormones by biosolids inwastewater treatment systems in Tennessee USA.Environmental Science & Technology 34 (18), 3925–3931.

Lin, C.-E., Chang, C.-C., Lin, W.-C., 1997. Migration behavior andseparation of sulfonamides in capillary zone electrophoresisIII. Citrate buffer as a background electrolyte. Journal ofChromatography A 768 (1), 105–112.

Lindberg, R.H., Wennberg, P., Johansson, M.I., Tysklind, M.,Andersson, B.A.V., 2005. Screening of human antibioticsubstances and determination of weekly mass flows in fivesewage treatment plants in Sweden. Environmental Science &Technology 39 (10), 3421–3429.

Lindqvist, N., Tuhkanen, T., Kronberg, L., 2005. Occurrence ofacidic pharmaceuticals in raw and treated sewages and inreceiving waters. Water Research 39 (11), 2219–2228.

Nakada, N., Tanishima, T., Shinohara, H., Kiri, K., Takada, H.,2006. Pharmaceutical chemicals and endocrine disrupters inmunicipal wastewater in Tokyo and their removal duringactivated sludge treatment. Water Research 40 (17), 3297–3303.

Omil, F., Suarez, S., Carballa, M., Reif, R., Lema, J.M., 2010. Criteriafor designing sewage treatment plants for enhanced removalof organic micropollutants. In: Fatta-Kassinos, D., et al. (Eds.),Xenobiotics in the Urban Water Cycle. Mass Flows,Environmental Process, Mitigation and Treatment Strategies,Environmental Pollution, vol. 16. Springer ScienceþBusinessMedia B.V., p. 283 (Chapter 16).

Quintana, J.B., Carpinteiro, J., Rodriguez, I., Lorenzo, R.A.,Carro, A.M., Cela, R., 2004. Determination of natural andsynthetic estrogens in water by gas chromatography withmass spectrometric detection. Journal of Chromatography A.1024 (1–2), 177–185.

Rodrıguez, I., Quintana, J.B., Carpinteiro, J., Carro, A.M.,Lorenzo, R.A., Cela, R., 2003. Determination of acidic drugs insewage water by gas chromatography-mass spectrometry astert-butyldimethylsilyl derivates. Journal of ChromatographyA. 985, 265–274.

Shi, J., Fujisawa, S., Nakai, S., Hosomi, M., 2004. Biodegradation ofnatural and synthetic estrogens by nitrifying activated sludgeand ammonia-oxidizing bacterium Nitrosomonas europaea.Water Research 38 (9), 2323–2330.

Simonich, S.L., Federle, T.W., Eckhoff, W.S., Rottiers, A., Webb, S.,Sabaliunas, D., De Wolf, W., 2002. Removal of fragrancematerials during US and European wastewater treatment.Environmental Science & Technology 36 (13), 2839–2847.

Stuer-Lauridsen, F., Birkved, M., Hansen, L.P., Holten Lutzhoft, H.-C., Halling-Sorensen, B., 2000. Environmental risk assessmentof human pharmaceuticals in Denmark after normaltherapeutic use. Chemosphere 40, 783–793.

Ternes, T.A., Herrmann, N., Bonerz, M., Knacker, T., Siegrist, H.,Joss, A., 2004. A rapid method to measure the solid-waterdistribution coefficient (K-d) for pharmaceuticals and muskfragrances in sewage sludge. Water Research 38 (19), 4075–4084.

Urase, T., Kikuta, T., 2005. Separate estimation of adsorption anddegradation of pharmaceutical substances and estrogens inthe activated sludge process. Water Research 39, 1289–1300.

Vader, J.S., van Ginkel, C.G., Sperling, F.M.G.M., de Jong, J., deBoer, W., de Graaf, J.S., van der Most, M., Stokman, P.G.W.,2000. Degradation of ethinyl estradiol by nitrifying activatedsludge. Chemosphere 41 (8), 1239–1243.

Vasskog, T., Berger, U., Samuelsen, P.J., Kallenborn, R., Jensen, E.,2006. Selective serotonin reuptake inhibitors in sewageinfluents and effluents from Tromso, Norway. Journal ofChromatography A 1115 (1-2), 187–195.

Yi, T., Harper, W.F., 2007. The link between nitrification andbiotransformation of 17 alpha-ethinylestradiol.Environmental Science & Technology 41 (12), 4311–4316.

Zwiener, C., Glauner, T., Frimmel, F.H., 2000. Biodegradation ofpharmaceutical residues investigated by SPE-GC/ITD-MS andon-line derivatization. HRC Journal of High ResolutionChromatography 23 (7–8), 474–478.