Mendez Arriaga si colab. 2008

Available at www.sciencedirect.com

else

o

solar irradiation

Fabiola Mendez-Arriaga, San

Chemical Engineering Department, University

a r t i c l e i n f o

drives to a satisfactory biodegradability index BOD5/COD (between 0.16 and 0.42) and, only

rved.

Recent studies have demonstrated the presence of pharmaceu- the consumed Top 200 drugs are ingested orally, and approxi-

ARTICLE IN PRESS

WAT E R R E S E A R C H 42 ( 2008 ) 585 594Corresponding author. Tel.: +34 93 4021293; fax: +34 93 402 1291.tical compounds in rivers, lakes and superficial freshwater

(Halling-Sorensen et al., 1998; Ternes, 1998; Daughton and

Ternes, 1999; Carballa et al., 2004). Non-steroidal anti-inflam-

matory drugs (NSAIDs) are a special group of pharmaceuticals

mately 5% of them correspond to NSAIDs (Takagi et al., 2006).

Some important examples of this family of medicines are

ibuprofen, naproxen, diclofenac and ketoprofen, although there

are more than 50 different types available commercially.

0043-1354/$ - see front matter & 2007 Elsevier Ltd. All rights reserved.doi:10.1016/j.watres.2007.08.002

E-mail addresses: [email protected] (F. Mendez-Arriaga), [email protected] (J. Gimenez).1. Introduction that are often found as a persistent toxic waste and are one ofthe most widely available drugs in the world. In Spain, 55% ofin this case, a post-biological treatment could be suggested.

& 2007 Elsevier Ltd. All rights resewere the most important residual compounds after the photocatalytic treatment of IBP. The

inhibition percentage of bioluminescence from Vibro fischerias a toxicity parameter

increased during the irradiation time due to the residual concentration of the hydroxyl

metabolites generated. However, after 120min, in experiments with 40mg/L of dissolved

oxygen, a decrease of the % inhibition was observed. Only photocatalytic treatment of IBPArticle history:

Received 4 January 2007

Received in revised form

29 June 2007

Accepted 1 August 2007

Available online 6 August 2007

Keywords:

Diclofenac

Ibuprofen

Naproxen

NSAID

Photocatalysis

Solar

TiO2tiago Esplugas, Jaime Gimenez

of Barcelona, Mart i Franque`s 1, 08028 Barcelona, Spain

a b s t r a c t

The aim of this work is to evaluate and compare the degradation achieved for three non-

steroidal anti-inflammatory drugs (NSAIDs) by heterogeneous TiO2 photocatalytic means in

aqueous solution at laboratory scale. The selected pharmaceutical compounds were

diclofenac (DCF), naproxen (NPX) and ibuprofen (IBP). These compounds were used in their

sodium salt chemical form.

Previous experiments (adsorption, photolysis and thermodegradation) were developed

to evaluate non-catalytic degradation for each NSAID. Photocatalytic experiments were

carried out in a Xe-lamp reactor in order to study the influences of different operational

conditions (catalyst load, temperature and dissolved oxygen concentration). These results

showed that the optimum amount of TiO2, to achieve maximum degradation, of IBP was

1 g/L. In contrast, the maximum degradation for DCF or NPX was observed at a TiO2 loading

of 0.1 g/L. Temperature had a significant effect only for NPX degradation, achieving almost

99% phototransformation. No significant differences were observed for DCF and IBP at 20,

30 and 40 1C. Dissolved oxygen concentration was an important parameter to increase the

degradation for NPX and IBP. However, it was observed that its rate of mineralization did

not increase. Intermediate metabolites were detected in all cases. Hydroxyl metabolitesanti-inflammatory drugs with TiO2 and simulatedjournal homepage: www.

Photocatalytic degradation of nvier.com/locate/watres

n-steroidal

The detection levels rarely exceed the mg/L levels (Stumpf

(Durapore PVDF Millipore 0.22mm) in a high-performance

ARTICLE IN PRESS

2 (et al., 1998; Ternes, 1998; Tixier et al., 2003); however, their

presence at low or high concentrations could bring about

harmful toxicological consequences (Fent et al., 2006; Prakash

Reddy et al., 2006).

Frequently, NSAIDs are prescribed for skeletalmuscle pain

and inflammatory rheumatic disorders; however, they also

present analgesic and antipyretic effects. The main common

characteristic in the NSAID group is the carboxylic aryl acid

moiety that provides their acidic properties.

Investigations on physical and chemical remediation tech-

nologies had recently reported the removal of some NSAIDs.

Filtration, adsorption, coagulationflocculation and flotation

(Carballa et al., 2005) are some examples. On the other hand,

advanced oxidation processes (AOPs) appear to be an alter-

native for the degradation of hazardous pollutants (Bauer

et al., 1999) and are usually clean technologies with an

economic advantage when AOPs are used with a natural light

source (Malato et al., 2001). AOPs are a potential alternative

for the degradation of NSAIDs by non-selective hydroxyl

radical attack on organic molecules and, sometimes, com-

plete mineralization is achieved (Ravina et al., 2002). Actually,

AOPs have been employed to remove NSAIDs. Special

development did show the ozonation process (Huber, 2003;

Ternes, 2003; Zweiner, 2000) and recently the photo-Fenton

solar process (Perez-Estrada et al., 2005a, b).

Studies with TiO2 heterogeneous catalysts have also been

reported. Oxidation of organic compounds by means of TiO2was achieved by hydroxyl radical generation through the

e/h+ pair generated when the semiconductor is exposed to

UV radiation. In a recent study, Calza et al. (2006) present an

exhaustive study on the heterogeneous photocatalytic degra-

dation of DCF over TiO2 aqueous suspension at low DCF

concentrations (0.079.24ppm).

In our study, relatively high initial concentrations of NSAID

were tested to obtain the most plausible amount of photo-

products and their subsequent characterization in order to

measure the final toxicity and biodegradability achieved. The

aim of this research is to describe the capability of TiO2 to

remove pollutants in aqueous solutions with high NSAID

concentrations, as well as to compare the degradation degree

of three separate NSAIDs, at different operation conditions in

a tubular laboratory reactor with a solar-simulated light

source.

2. Experimental section

2.1. Materials

Synthetic amorphous titanium dioxide (Degussa P-25) was

used as received. The NSAIDs DCF (2-[2-[(2,6-dichloropheny-In the environment, NSAIDs have been detected in hospital

wastewaters (Kummerer, 2001), in and out STP (sewage

treatment plant) effluents (Carballa et al., 2004), in surface

water such as rivers and lakes (Boyd et al., 2003), in marine

waters (Weigel et al., 2004) as well as in soil matrices (Scheytt

et al., 2006).

WAT E R R E S E A R CH 4586l)amino]phenyl]acetic acid), NPX ((S)-6 methoxy-a-methyl 2naphthaleneacetic acid)) and IBP (2-[3-(2-methylpropyl)phe-liquid chromatograph (HPLC) from Waters using a C18 RP

Trace Extrasil OD52-5 Micromet 250.46 Teknockroma col-umn, and a Waters 996 photodiode array detector using the

Empower Pro software 2002 Water Co. Conditions of each

HPLC analysis are shown in Table 1.

Total organic carbon (TOC) was measured in a Shimadzu

TOC-V CNS instrument. Spectrophotometric measurements

to obtain the absorption spectrum were carried out in a

Perkin-Elmer UV/VIS Lambda 20 (220700nm range) spectro-

photometer. BOD5 determinations were carried out according

to the Standard Methods (5120) by a respirometric single

measuring system and OxiTop procedure. COD determination

was carried out according to Norm. France NFT 90101.

Multi-temperaturemeasurements (Crison 621) and pH (Crison

GLP 22) were monitored online. Air and pure oxygen sparged

through the system were purchased at Air Liquide, both were

used to vary the dissolved oxygen concentration. Dissolved

oxygen concentration was measured online by Crison Oxi

330i WTWOxi Cal-SL sensor. Microtox tests were employed to

obtain the inhibition of bioluminescence of Vibrio fischeri at

15min of incubation time. Millipore water quality (18MO) was

used during all experiments. For the identification of by-

products, the final sample mixture was analyzed by electro-

spray ionization/mass spectrometry using a PerSeptive, TOF

Mariner Jasco AS-2050 plus IS mass spectrometer.

2.3. Photoreactor

A stirred tank (1.5 L) was filled with the NSAIDTiO2 solution.

The aqueous suspension was continuously pumped (Hei-

dolph Pumpdrive 5130 120300 rpm) to the Solarbox (Co.fo.-

me.gra 220V 50Hz) and recirculated to the batch tank. In the

Solarbox, the Duran tubular photoreactor (0.078L) was

irradiated by the Xe-OP lamp (Phillips 1kW) with a photon

flux equal to 6.9 mEinstein/s (290400nm). The jacket tem-

perature of the stirred tank was controlled with an ultra-

thermostat bath (Selecta; Frigiterm-10). All connections and

pipes employed were made of Teflon and/or glass material to

avoid losses by adsorption. The different types of equipment

and instruments employed in our installation are depicted in

Fig. 2.

3. Results and discussion

3.1. Previous experiments: photolysis, thermo-degradation and adsorption

To avoid confusion between photocatalytic and other degra-

dation phenomena, it was necessary to evaluate the non-nyl]propanoic acid), all of them in their sodium salt form,

were purchased from Sigma-Aldrich and used without pre-

treatment (Fig. 1).

2.2. Analytical procedures

NSAID concentrations were monitored after sample filtration

2008 ) 585 594photocatalytic influence of photolysis, thermodegradation

and adsorption of each NSAID onto the TiO2 catalyst.

ARTICLE IN PRESS

2 (O

NH

Cl

Cl O

DCF NPX

Na+O-

WAT E R R E S E A R C H 4Photolysis experiments were carried out with 1.5 L volume

solutions (200ppm initial concentration) placed in the stirred

tank and pumped to the Solarbox without the catalyst. The Xe

lamp was switched on and aliquots were taken to evaluate

the NSAID concentrations and TOC at different irradiation

times. The temperature was maintained constant at 30 1C.

Calza et al. (2006) confirmed that no significant influence by

photolysis at short-time experiments on DCF was detected;

however, in our experience, DCF and NPX concentrations

decreased until 75% when the DCF or NPX solution was

continually exposed to the Xe lamp during 2h (see Fig. 4a).

This degradation by the photolysis phenomenon has an

Fig. 1 Non-steroidal anti-infl

Table 1 Parameters of HPLC analysis employed

NSAID Mobile phase composition Flow-rat(mL/min

DCF Acetonitrileb10mM ammonium

formiatec (5050)

1.25

NPX Acetonitrileb0.05M ammonium

phosphate dihydrogend (8020)

1.00

IBP Acetonitroleb0.25 acetic acide (7525) 1.75

a All in isocratic mode.b Panreac 99.8%.c Fluka 99% HPLC grade.d Aldrich 98% ACS.e Fluka grade LCMS.

Fig. 2 Experimental equipmeO

O-Na+

O

O-Na+

IBP

2008 ) 585 594 587important influence on DCF and NPX degradation due to

their absorption spectra, which overleap on the Xe-lamp

emission spectrum, specifically above 290nm (see Fig. 3). The

results of these experiments are in agreement with Packer

et al. (2003). On the other hand, the small absorbance in the

UV region for IBP (see also Fig. 3) helps to explain why any

photolysis degradation was reached.

TOC decrease was only significant for DCF photolysis

(25%) (see Fig. 4b) and it could be possible due to the

photolysis contribution and also because of the sub-products

generated. The initial transparent solution of DCF changed

during photolysis to a light-brown color at the end of the

ammatory drugs selected.

e)a

Injectionvolume (mL)

l Detection(nm)

Temperature(1C)

20 280 25

20 254 25

20 254 25

nt and instrument design.

ARTICLE IN PRESS

2 (DCF

NPX

IBP

Emission spectrum

of Xe Lamp 1 kW

0

1

2

3

4

290 300 310 320 330 340 350 360 370 380 390 400

Abs (A

U)

0

0.5

1

1.5

2

2.5

Irra

dia

nce (W

m2)

WAT E R R E S E A R CH 4588experiments. The pH values during the photolysis experiment

decreased from 6 to 3, but no precipitation was observed. The

acidic decrease allows us to conclude that there is an

important influence of the photolysis effect on the degrada-

tion of DCF.

To evaluate the possible degradation due to temperature, 1 L

of NSAID solution (each one separated) with a 200ppm initial

concentration was placed in the stirred tank and heated at 20,

40, 60 and 80 1C. Once the set point temperature was reached,

an aliquot for HPLC and another one for TOC analysis were

taken. No thermo-degradation and TOC decreases were

observed for any NSAID tested in the 2080 1C range.

Concentration profiles in all cases were always constant and

0

50

100

150

200

250

0 30 60 90 120 150

Irradiation time (min)

NS

AID

(m

g/L

)

DCF NPX IBP

0

50

100

150

200

250

0 30 60 90 120 150


TO

C (

mg

/L)

DCF NPX IBP

Fig. 4 (a) Photolysis experiments. [NSAID] vs. irradiation

time for each NSAID. (b) Photolysis experiments. [TOC] vs.

irradiation time for each NSAID.

Wavelength (nm)

Fig. 3 Absorption spectra of DCF, NPX and IBP in Millipore

H2O and emission spectrum of Xe lamp (Allen et al., 2000).close to the initial concentration. This result was also seen in

measured TOC values.

On the other hand, adsorption isotherms were obtained for

each NSAID. To evaluate the decrease of the pollutant due to

adsorption on the catalyst, different compositions of NSAID

TiO2 solution were prepared in 25mL hermetic closed flasks

and submerged in a thermostatic bath at constant tempera-

ture (30 1C). The equilibrium concentration of each adsorbed

NSAID was determined by the difference after measuring the

supernatant concentration at 24h darkness. In Fig. 5, the

NSAID adsorption on TiO2 is depicted varying TiO2 loading at

an initial NSAID concentration of 500ppm.

Low percentages of adsorption were reached in all cases:

4.8470.39, 8.8371.29 and 2.1171.3 for DCF, NPX and IBP,

respectively. From a practical point of view, the adsorption

stage between any NSAID and the catalyst TiO2 does not

represent a significant loss in reduction on the concentration

of the substrate during the main photocatalytic process.

3.2. Photocatalytic experiments

0.0

1.0

2.0

3.0

0 0.2 0.4 0.6 0.8 1

TiO2 g/L

NS

AID

mM

liquid equilibrium phase

adsorbed phase

Fig. 5 DCF (&), NPX (n) and IBP (J) supernatant

concentrations and their respective adsorbed

concentrations on different loadings of TiO2 at an NSAID

concentration of 500ppm (filled points).

2008 ) 585 594Conditions, technical parameters, operational process values

and results of NSAID, TOC conversions, biodegradability

index (BOD5/COD) and toxicity for photocatalytic experiments

are shown in Table 2.

3.2.1. Process conditions: influence of temperature and flowrateAny temperature effects on DCF and IBP mineralization were

observed in the tested range (20, 30 and 40 1C). In the case of

NPX, an increase in degradation and a slight rise in TOC

conversion were reached for experiments carried out at 40 1C.

At the end of the experiments, a brown color was observed,

which could be explained by a polymerization reaction.

Flow rates (0.1, 0.2 and 0.4mL/min) have not influenced the

degradation rate in any NSAID tested in accordance with the

operational configuration of our system, because the ratio

between the irradiated volume and the total volume is 0.052.

The range of flow rates tested does not produce an important

difference in the crossing irradiated volume per minute and

no effect is observed.

ARTICLE IN PRESS

OD

te

2 (Table 2 Experimental photocatalytic setup, XNSAID, XTOC, B

NSAID TiO2 Dissolved O2 Flow-ra

WAT E R R E S E A R C H 43.2.2. Influence of TiO2 loadingIn heterogeneous photocatalysis, the rate of degradation is

not always proportional to the catalyst load. An optimal point

exists where TiO2 loaded shows a maximum degradation rate.

Over this value, scattering phenomena can appear and active

(g/L) (mg/L) recirculatio(L/min)

DCF D1a 0 82b 0.2

D2 0.07 82 0.2

D3 0.1 82 0.2

D4 0.5 82 0.2

D5 1 82 0.2

D6 2 82 0.2

D7 0.1 82 0.1

D8 0.1 82 0.4

D9 0.1 82 0.2

D10 0.1 82 0.2

D11 0.1 40 0.2

NPX N1a 0 82 0.2

N2 0.1 82 0.2

N3 0.5 82 0.2

N4 1 82 0.2

N5 0.1 82 0.1

N6 0.1 82 0.4

N7 0.1 82 0.4

N8 0.1 82 0.4

N9 0.1 40 0.4

I1a 0 82 0.2

I2 0.1 82 0.2

I3 0.2 82 0.2

I4 0.5 82 0.2

I5 0.75 82 0.2

I6 1 82 0.2

I7 1 82 0.1

IBP I8 1 82 0.4

I9 1 82 0.1

I10 1 82 0.1

I11 1 40 0.1

I11 (0h) 1 40 0.1

I11 (0.5h) 1 40 0.1

I11 (1.0h) 1 40 0.1

I11 (2.0h) 1 40 0.1

I11 (3.0h) 1 40 0.1

I11 (6.0h) 1 40 0.1

I11 (12.0h) 1 40 0.1

I12 1 8 0.1

I12 (0.5h) 1 8 0.1

I12 (1.0h) 1 8 0.1

I12 (2.0h) 1 8 0.1

I12 (3.0h) 1 8 0.1

I13 (25 ppm) 1 40 0.1

I14 (50 ppm) 1 40 0.1

I15 (100 ppm) 1 40 0.1

Irradiation time: 4h excluding the experiments indicated.

Initial concentration: 200ppm excluding the experiments indicated.a Photolysis experiment.b 82 expression indicates a free [O2] consumption, from an initial conc5/COD and/or % Vibrio fisheri (V.f.) inhibition

Temperature X X BOD5/COD and/or

2008 ) 585 594 589sites on the surface of the catalyst do not bring more e/h+

generation and therefore increase in degradation rate does

not occur.

For DCF and NPX, the addition of TiO2 does not improve the

degradation rate in comparison with photolysis experiments;

n (1C) NSAID TOC (% V.f. inhibition)

30 0.95 0.25 0.02

30 0.94 0.52 o0.0230 0.90 0.59 o0.0230 0.89 0.50 o0.0230 0.75 0.38 o0.0230 0.69 0.29 o0.0230 0.90 0.68 o0.0230 0.94 0.68 0.02

20 0.87 0.68 o0.0240 0.91 0.69 0.02

30 0.98 0.70 0.02 (14)

30 0.90 0.07 0.02

30 0.44 0.19 o0.0230 0.28 0.11 o0.0230 0.28 0.12 o0.0230 0.48 0.18 o0.0230 0.60 0.20 o0.0220 0.62 0.20 o0.0240 0.99 0.26 0.03

40 0.99 0.26 0.05 (97)

30 0.0 0.0 0.02

30 0.14 0.05 0.02

30 0.26 0.10 0.02

30 0.38 0.14 0.02

30 0.55 0.11 0.02

30 0.70 0.13 0.02

30 0.61 0.13 0.02

30 0.71 0.18 0.02

20 0.65 0.13 0.02

40 0.79 0.15 0.02

30 0.99 0.14 0.16 (87)

30 0 0 0 (77)

30 0.19 0.05 (79)

30 0.21 0.11 0.03 (82)

30 0.59 0.14 0.10 (91)

30 0.70 0.15 (88)

30 0.99 0.16 0.20

30 0.99 0.17 0.29

30 0.68 0.14 0.04 (79)

30 0.20 0.03 0.04 (89)

30 0.28 0.06 0.04 (87)

30 0.33 0.10 0.04 (78)

30 0.49 0.12 0.04 (84)

30 0.99 0.80 0.42

30 0.99 0.74 0.30

30 0.99 0.50 0.19

entration of 8mg/L to a final concentration of 2mg/L.

however, for TOC rate degradation, an enhancement was

shown when TiO2 was employed. Only for IBP the TiO2 loaded

has an important rise in the IBP conversion (00.75) but a

slight increase in the mineralization rate was achieved

(0.030.21). Within the TiO2 load range tested, 1 g/L was found

to be the optimum value to achieve a maximum conversion of

IBP and TOC. Fig. 6 shows the highest conversion obtained for

each NSAID tested.

An approach to describe the apparent kinetic order of the

photodegradation reaction was achieved through the simple

first-order kinetics model. The kinetics constant (K), shown in

Table 3, highlights the fact that the reaction rate of NSAID

degradation is not so different for each compounds tested.

3.2.3. Ibuprofen caseinitial concentration varying: influenceon biodegradability indexFrom a practical point of view, biodegradability and toxicity

tests were carried out for the conditions where higher NSAID

and TOC conversions were reached. Even though good

conversions are obtained, the by-products generated can be

less biodegradable or more toxic than the original ones, and

then the photocatalytic process lacks practical application to

a next biological process.

Under free oxygen consumption (28g/L) or constant (8 g/L)

ARTICLE IN PRESS

0.2

N

WAT E R R E S E A R CH 42 (5900

0 50 100 150 200 250 300


Fig. 6 Maximum degradation ratio for each NSAID tested.

DCF at 0.1g/L TiO2; NPX at 0.1 g/L TiO2; and IBP at 1g/L TiO2.

Table 3 Kinetic constant values for each NSAID

At 30 1C and 0.2mL/min TiO2 (g/L) K (min1)

DCF 0.1 9.6 103NPX 0.1 7.0 103

3In this sense, pure NSAID solutions (with initial concentra-

tions ranging from 2 to 500ppm) were tested and the

biodegradability index obtained in all cases was zero.

Biodegradability indexes reached had no significant improve-

ment for DCF or NPX still at saturated O2 concentrations

under any combination of operational conditions at initial

high concentrations (see last column in Table 2). This leads to

the conclusion that the final dissolution at the end of the

photocatalytic treatment would not be able to be biologically

treated, although acceptable conversions were observed.

0.4

0.6

0.8

1

SA

ID/N

SA

IDo

DCF NPX IBPIBP 1.0 9.1 10of dissolved oxygen, the reaction rate is remarkably stopped

and the maximum conversion of IBP obtained is almost 70%,

probably because of the subsequent exhaust or insufficient

concentration of O2 in dissolution. A high recombination of

e/h+ pair is reached and no more hydroxyl radicals are

available to react. Therefore, transfer of the hydroxyl radical

to the solution limits the photocatalytic reaction. Under

excess of oxygen in dissolution, the conversion of IBP is

almost complete possibly being the most efficient production

of hydroxyl radicals due to the less recombination of e/h+

species in the bulk of the catalyst. On the other hand, the

sparge got a major exposition or contact between the TiO2, O2and the pollutant, raising the probability of collisions among

the IBP and the reactive species. In contrast, the slight

improvement on the TOC removals could be explained due

to the strong ability of IBP to scavenge the hydroxyl radical

generated. If the acceptor specie (O2) is present in excess, the

production of hydroxyl radicals increases. IBP then acts more

efficiently as a scavenger of those hydroxyl radicals and

diminishes almost totally its concentration by the hydroxyla-

tion process. Mineralization is not higher because the primary

process present is the generation of hydroxyl metabolites.

The probable generation of acid compounds (propionic,Only the case of IBP showed an enhancement in the BOD5/

COD index when the solution had an excess of dissolved

oxygen. Next sections will focus on IBP results due to the

improvement in its biodegradability index.

In Figs. 7a, b and 8 are shown IBP removal, TOC conversion

and biodegradability index, respectively, as a function of the

initial IBP concentration (200, 100, 50 and 25ppm) at 1 g/L of

TiO2 and 40mg/L of dissolved oxygen. At the lowest initial

concentration of IBP, it is possible to obtain the maximum

degradation in approximately 60min, with a mineralization

rate of almost 50%. The biodegradability index reached the

highest BOD5/COD relationship at the lowest initial substrate

concentration (Fig. 8). The photocatalytic process showed a

higher effectiveness in degradation at low initial concentra-

tions of IBP. Its higher biodegradability index can be related to

the high oxidation level reached by the by-products gener-

ated.

Removal at the end of reaction andmineralization are lower

at high initial concentrations of IBP; however, in all cases the

photocatalytic treatment shows an enhancement in the

biodegradability index in comparison to IBP (high or low

concentrations) without treatment (from 0 to 0.160.42).

3.2.4. Dissolved O2 concentration varying: influence onbiodegradability indexDissolved O2 concentration plays an important role on

photocatalytic processes (Blanco and Malato, 2003). In aerobic

photocatalytic heterogeneous systems the absorbed oxygen is

the principal acceptor species. An important enhancement in

the IBP degradation was observed at saturated dissolved O2concentration (40mg/L) and a reduction of irradiation time

was also observed. Mineralization shows, however, a slight

enhancement, approximately only 15% (see Fig. 9).

2008 ) 585 594formic and hydroxypropionic acid) from the loss of the

carboxyl moiety of IBP is not the main reaction.

ARTICLE IN PRESS

2 (150

200

25 ppm

50 ppm

WAT E R R E S E A R C H 4Moreover, at long irradiation exposure times, hydroxylation

also appears to be the main reaction because the conversion

of IBP is completed and mineralization did not increase. (see

Section 3.2.5).

The improvement of the biodegradability ratio of IBP due to

high dissolved oxygen concentration is the most important

aspect to remark under an excess of dissolved oxygen in the

0

50

100

0 50 100 150 200 250

Irradiation Time (min)

IBP

ppm

100 ppm200 ppm

X T

OC

Fig. 7 (a) Initial IBP concentration effect on IBP rem

0

0.1

0.2

0.3

0.4

0.5

25 ppm 50 ppm 100 ppm 200 ppm

IBP initial concentration

BO

D5/C

OD

Fig. 8 Initial IBP concentration effect on the BOD5/COD after

4h of photocatalytic treatment.

0

0.25

0.5

0.75

1

0 60 120 180 240


X I

BP

0

0.1

0.2

0.3

0.4

0.5

X T

OC

Fig. 9 O2 effect on the XTOC (empty points) and XIBP(filled points): (B) 82g/L O2 consumed; (&) 8 g/L O2; and

(n) 40g/L O2.0.50

0.75

1.0025 ppm

50 ppm

100 ppm

200 ppm

2008 ) 585 594 591dissolution and it is also observed at long irradiation times. In

Figs. 10 and 11, the irradiation time effect and concentration

of dissolved oxygen on BOD5/COD are shown.

3.2.5. Intermediates generated and their toxicityNo differences were observed in the type of by-products

identified among experiments carried out at saturated or

free O2 concentrations. However, the intensity obtained in

experiments where O2 was not in excess was two times

higher than those where O2 was maintained at saturated

concentrations.

0.00

0.25

0 50 100 150 200 250


oval. (b) Initial IBP concentration effect on XTOC.

IBP biodegradability ratio

0

0.1

0.2

0.3

0.4

1 hr 2 hr 4 hr 6 hr 12 hr

Irradiation Time

BO

D5/C

OD

Fig. 10 Irradiation time effect on BOD5/COD. [IBP]o 200ppm, 1g/L TiO2, 40mg/L O2, 30 1C.

0

0.1

0.2

0.3

0.4

8-2 (free

consume)

8 g/L (O2 eq

AIR)

40 g/L (O2 SAT)

Oxygen concentration

BO

D5/C

OD

Fig. 11 Dissolved oxygen concentration effect on BOD5/

COD. All cases [IBP]o 200ppm, 1g/L TiO2, 30 1C, 4h.

ARTICLE IN PRESS

O-O

HO

O

O

HO

OH

OH

[221]

[238]

5]

[162]

[133]

C2H5

O-

C

O

C

O-

OH

O

III

OH

OH

H3C

H3C

OH

O

O

2 (O-O

HO

O-OHO

OH

OH

[221]

[221]

[20

[178]

HO

HO

OH

[178]

C O-

O

C O-

O

I

II

IV

WAT E R R E S E A R CH 4592Fig. 12 shows the by-products detected by LCMS from the

photocatalytic treatment (in marked circle) and possible

transformation pathways followed by ibuprofen.

The most probable degradation (or transformation) of

IBP is due to a dominant photochemical scavenger of the

OHdmediated process by the inclusion of the radical in

the methylpropyl phenyl positions I (1-hydroxyIBP) and

II (2-hydroxyIBP)), and in the arylcarboxylic moiety (III

and IV). Direct demethylation or decarboxylation was not

observed. However, complementary results obtained from

the degradation of IBP by means of the photo-Fenton reagent

(not yet published) helped to assure that the hydroxylation

process can be the first step of degradation, followed by

a second step of demethylation or decarboxylation with

other different by-products with smaller m/z (178, 162, 133)

and also propionic, formic or hydroxypropionic acid or

sodium salts. Only at low initial concentrations of IBP did

the biodegradability increase (Fig. 8) due to the photocatalytic

reaction. Then, the highest concentration of hydroxylated

by-products is produced. From these, the formation of acidic

or salt compounds could continue with the degradation

process. In contrast, at high initial concentrations of IBP

(200ppm), the only step of the hydroxylation reaction is seen

and a slight formation of acidic compounds is observed

(Fig. 12).

During the photocatalytic reaction, the generation of by-

products increases until a maximum concentration is

reached. In our case, it can be related to the toxicity results

Fig. 12 Intermediates observed at the end of photocatalytic tr

irradiation time.O-O

C O-

O

O-

OH

HO

H

O

2008 ) 585 594obtained. The toxicity of the solution at different reaction

times was measured by a Microtox assay and can be observed

in Fig. 13. The increase in the inhibition of the bacteria test

goes hand in hand with the increase in toxic by-products,

which corresponds to hydroxyl forms of ibuprofen, generated

during the photocatalytic reaction. The most important

difference (expressed in inhibition percentages), between

the experiments at 40 and 28mg/L free oxygen consumption,

is the production rate of the toxic by-products at oxygen

eatment of IBP. [IBP]o 200ppm, 1g/L TiO2, 40mg/L O2, 4h

0

5

10

15

20

0 30 60 120 180 240


EC

50 (

%v/v

)

65

70

75

80

85

90

95

% inhib

ition V

ib.fis

ch.

IBP 8-2 free O2 consume

IBP 40 mg/L O2

Fig. 13 Changes of EC50% v/v (in bars) and % inhibition

(linked symbols) of Vibrio fisheri bioluminescence as a

function of irradiation time.

conversion of the initial IBP concentration. Biological treat-

treatment plants. Water Res. 38 (12), 29182926.Carballa, M., Omill, F., Lema, J., 2005. Removal of cosmetics

ARTICLE IN PRESS

2 (ment seems to be a good post-alternative after a photo-

catalytic treatment, because reports on biofilms or batch

biological reactors have shown the ability to reduce 90% of the

initial IBP and its metabolites (Zweiner et al., 2002). In this

way, the photocatalytic by-products, generated and promoted

at excess O2 concentration, increased the biodegradability

index at high concentrations of IBP.

4. Conclusions

Although NSAIDs are a group of pollutants with special

characteristics for specific functions like drugs in human use,

no general pattern is possible to assign in the photochemical

reactions for their degradation by catalytic means. No general

tendencies were found for the photolytic degradation, ad-

sorption or thermodegradation experiments. However, a

strong influence by photolysis is observed in the case of

DCF and NPX.

For all cases, an improvement is reached on the TOC

reduction when increasing the TiO2 loading. For DCF and NPX

cases, the by-products generated were not recommended for

a post-biological treatment due to the low biodegradability

index reached (sometimes equal to the original one).

Only IBP was an important increase in its removal and

biodegradability index reached after a photocatalytic process.

In IBP degradation, the main photocatalytic process is

hydroxylation and it depends strongly on aerated or non-

aerated conditions. The photocatalytic degradation of IBP is

only an early transformation process going through hydro-

xylated by-products, but complete mineralization is not

observed. High oxidation levels can be assured by direct

decarboxylation attack on the hydroxyl radical avoiding thesaturation conditions. It is possible to observe a maximum

value of % inhibition at 120min and, after that, a decrease

appears due to the diminishing concentration of toxic by-

products. In the case of non-saturated conditions, the

production of toxic metabolites is still increasing and it could

be possible for them to rise until a maximum value of

approximately 92% for times longer than 240min. Analog

results are observed for the EC50 (% v/v) parameter. At 40mg/L

of O2, EC50 seems to give a minimum value and then it

increases in accordance with the maximum concentration of

hydroxyl metabolites generated.

Moreover, Packer et al. (2003) assured that the half-life time

of IBP in the environment can be function of the OHd

concentration, i.e. [1015], [1016] and [1017]M, corresponding

to 29, 296 and 2960h, respectively. IBP degradation by OHd

scavenger reaction is a slow process in a normal environ-

ment. In contrast, the application of a photocatalytic treat-

ment on the degradation of IBP increases the concentration

rate of the hydroxyl radical, thus promoting the generation of

faster hydroxyl metabolites. The metabolites generated dur-

ing the photocatalysis process are also reported as a product

of the biochemical pathway (Zweiner et al., 2002) but no more

than 10% of the original amount of IBP. By photocatalytic

means, it is possible to increase the percentage up to 100%

WAT E R R E S E A R C H 4hydroxylation step. By photocatalytic means, a direct attack

on the carboxylic moiety is not observed. Moreover, theingredients and pharmaceuticals in sewage primary treat-ment. Water Res. 30, 47904796.

Daughton, C.G., Ternes, T., 1999. Pharmaceuticals and personalcare products in the environment: agents of subtle change?Environ. Health Perspect. 107 (6), 907938.

Fent, K., Weston, A., Caminada, D., 2006. Ecotoxicology of humanpharmaceuticals. Aquat. Toxicol. 76 (2), 122159.

Halling-Sorensen, B., Nors, S., Lanzky, P., Ingerslev, F., Holten, H.,Jorgensen, S., 1998. Occurrence, fate end effects of pharma-ceutical substances in the environmenta review. Chemo-sphere 36 (2), 357393.

Huber, M., 2003. Oxidation of pharmaceuticals during ozonationand advanced oxidation processes. Environ. Sci. Technol. 37,10161024.hydroxylation rate is directly related to the concentration of

dissolved oxygen due to higher hydroxyl radical concentra-

tions that are able to react.

The hydroxyl metabolites generated are in accordance with

the toxicological results obtained. The inhibition percentage

decreases if concentrations of the toxic metabolites decrease

at long irradiation times, but also if the hydroxyl radical rate

increases with an excess in the dissolved oxygen concentra-

tion. IBP, as a pollutant in water at high concentrations, can

be degraded in a biological treatment after pre-oxidation by

photocatalysis treatment with TiO2.

Acknowledgments

The authors would like to thank the financial support from

the Spanish Ministry of Education and Science (Projects

CTQ2005-00446/PPQ and CTQ2004-02311/PPQ), the Barcelona

University for the predoctoral grant, Marc Esplugas for his

help in the management of the research and to Dr. Marta

Vilaseca Casas and Dr. Mara Reixach Riba for their help and

technical support in the HPLC and LCMS analyses.

R E F E R E N C E S

Allen, J., Allen, S., Baertischi, S., 2000. 2-Nitrobenzaldehyde: aconvenient UV-A and UV-B chemical actinometer for drugphotoestability testing. J Pharm. Biomed. Anal. 24, 167178.

Bauer, R., Waldner, G., Fallmann, H., Hager, S., Klare, M., Krutzler,T., Malato, S., Maletzky, P., 1999. The photo Fenton reactionand the TiO2/UV process for waste water treatmentnoveldevelopments. Catal. Today 53, 131141.

Blanco, J., Malato, S., 2003. Solar Detoxification. Renewable EnergySeries. UNESCO Publishing, France.

Boyd, G., Reemtsma, H., Grimm, A., Mitra, S., 2003. Pharmaceu-ticals and personal care products (PPCPs) in surface andtreated waters of Louisiana, USA and Ontario, Canada. Sci.Total Environ. 311, 135149.

Calza, P., Sakkas, V., Medana, C., Baiocchi, C., Dimou, A., Pelizzeti,E., Albanis, T., 2006. Photocatalytic degradation study ofdiclofenac over aqueous TiO2 suspensions. Appl. Catal.:Environ. 67, 197205.

Carballa, M., Omil, F., Lema, J., Llompart, M., Garca-Jares, C.,Rodrguez, I., Gomez, M., Ternes, T., 2004. Behavior ofpharmaceuticals, cosmetics and hormones in a sewage

2008 ) 585 594 593Kummerer, K., 2001. Drugs in the environment: emission of drugs,diagnostic aids and disinfectants into wastewater by hospitals

in relation in relation to other sourcesa review. Chemo-sphere 45, 957969.

Malato, S., Blanco, J., Vidal, A., Richter, C., 2001. Photocatalysiswhit solar energy at a pilot-plant scale: an overview. Appl.Catal. B: Environ. 37, 115.

Packer, J., Werner, J., Latch, D., McNeil, K., Arnold, W., 2003.Photochemical fate of pharmaceuticals in the environment:naproxen, diclofenac, clofibric acid and ibuprofen. Aquat. Sci.65, 343351.

Perez-Estrada, L., Malato, S., Gernjak, W., Aguera, A., Thurman, M.,Ferrer, I., Fernandez-Alba, A., 2005a. Photo Fenton degradation ofdiclofenac: identification of main intermediates and degradationpathway. Environ. Sci. Technol. 39 (21), 83008306.

Perez-Estrada, L., Maldonado, M., Gernjak, W., Aguera, A.,Fernandez-Alba, F., Ballesteros, M., Malato, S., 2005b. Decom-position of diclofenac by solar driven photocatalysis at pilotplant scale. Catal. Today 101, 219226.

Prakash Reddy, N.C., Anjaneyulu, Y., Sivasankari, B., Ananda, K.,2006. Comparative toxic studies in birds using nimesulide anddiclofenac sodium. Environ. Toxicol. Pharmacol. 22 (2), 142147.

Ravina, M., Campanella, L., Kiwi, J., 2002. Accelerated mineraliza-tion of the drug diclofenac via Fenton reactions in a concentricphotoreactor. Water Res. 36, 35533560.

Scheytt, T., Mersmann, P., Heberer, T., 2006. Mobility of pharma-ceuticals carbamazepine, diclofenac, ibuprofen, and propy-phenazone in miscible-displacement experiments. J. Contam.Hydrol. 83, 5369.

Stumpf, M., Ternes, T., Harbere, K., Baumann, W., 1998. Isolierungvon Ibuprofen-Metaboliten und deren Bedeutung als Konta-minanten der aquatischen Umwelt. Vom Wasser 91,291303.

Takagi, T., Ramachandran, C., Bermejo, M., Yamashita, S., Yu, L.,Amidon, G., 2006. A provisional biopharmaceutical classifica-tion of the top 200 drug products in the United States, GreatBritain, Spain and Japan. Mol. Pharm. 3 (6), 631643.

Ternes, T., 1998. Occurrence of drugs in German sewagetreatments plants and rivers. Water Res. 32 (11), 32453260.

Ternes, T., 2003. Ozonation: a tool for removal of pharmaceuticals,contrast media and fragrances from wastewater? Water Res.37, 19761982.

Tixier, C., Singer, H., Oellers, S., Muller, S., 2003. Ocurrence andfate of carbamazepine, ibuprofen, ketoprofen and naproxen insurface waters. Environ. Sci. Technol. 37 (6), 10611068.

Weigel, S., Berger, U., Jensen, E., Kallenborn, R., Thoresen, H.,Huhnerfuss, H., 2004. Determination of selected pharmaceu-ticals and caffeine in sewage and seawater from Tromso/Norway whit emphasis on ibuprofen and its metabolites.Chemosphere 56, 583592.

Zweiner, C., 2000. Oxidative treatment of pharmaceuticals inwater. Water Res. 34, 18811885.

Zweiner, C., Seeger, S., Glauner, T., Frimmel, F., 2002. Metabolitesfrom the biodegradation of pharmaceuticals residues ofibuprofen in biofilm reactors and batch experiments. Anal.Bioanal. Chem. 372, 569575.

ARTICLE IN PRESS

WAT E R R E S E A R CH 42 ( 2008 ) 585 594594

Photocatalytic degradation of non-steroidal anti-inflammatory drugs with TiO2 and simulated solar irradiationIntroductionExperimental sectionMaterialsAnalytical proceduresPhotoreactor

Results and discussionPrevious experiments: photolysis, thermo-degradation and adsorptionPhotocatalytic experimentsProcess conditions: influence of temperature and flow rateInfluence of TiO2 loadingIbuprofen case--initial concentration varying: influence on biodegradability indexDissolved O2 concentration varying: influence on biodegradability indexIntermediates generated and their toxicity

ConclusionsAcknowledgmentsReferences

Documents

Mendez Arriaga si colab. 2008