Journal of Environmental Management 98 (2012) 168e174

Contents lists available at SciVerse ScienceDirect

Journal of Environmental Management

journal homepage: www.elsevier.com/locate/jenvman

Degradation, mineralization and antibiotic inactivation of amoxicillin by

UV-A/TiO2 photocatalysis
Despina Dimitrakopoulou a, Irene Rethemiotaki b, Zacharias Frontistis a, Nikolaos P. Xekoukoulotakis a,
Danae Venieri a, Dionissios Mantzavinos a, *
a
Department of Environmental Engineering, Technical University of Crete, Polytechneioupolis, GR-73100 Chania, Greece
b
School of Science and Technology, The Hellenic Open University, 18 Parodos Aristotelous St., GR-26335 Patras, Greece

a r t i c l e i n f o a b s t r a c t

Article history: The UV-A/TiO2 photocatalytic decomposition of amoxicillin (AMX) in aqueous suspensions was inves-
Received 18 July 2011 tigated. Experiments were performed at antibiotic concentrations between 2.5 and 30 mg/L, eight
Received in revised form commercially available TiO2 catalysts at loadings between 100 and 750 mg/L, acidic or near-neutral
19 December 2011
conditions (pH 5 or 7.5) and two different matrices (ultrapure water and secondary treated effluent)
Accepted 8 January 2012
Available online 24 January 2012
at a photon flux of 8
104 E/(L min). Of the various catalysts tested, Degussa P25 was highly active, i.e.
complete AMX degradation and 93% mineralization could be achieved after 25 and 90 min of reaction,
respectively at 10 mg/L AMX and 250 mg/L titania. In general, mineralization was slower than degra-
Keywords:
Activity
dation due to the formation of stable transformation by-products. For the range of concentrations
Kinetics studied, initial degradation rates can be approached by a LangmuireHinshelwood kinetic model, while
Pharmaceuticals the reaction order with respect to AMX shifts from first to zeroth as initial concentration increases from
Removal 2.5 to 5 mg/L to higher values. Degradation in treated effluent was partly impeded compared to pure
Semiconductor water due to the inherent presence of organic and inorganic constituents that compete for hydroxyl
Water radicals. Although increasing solution pH from 5 to 7.5 had no effect on degradation, it retarded
mineralization. The antibiotic activity of AMX prior to and after photocatalytic degradation was tested to
three reference bacterial strains, namely Escherichia coli (ATCC 23716), Klebsiella pneumoniae (NCTC
5056) and Enterococcus faecalis (ATCC 14506). The first two were found to be highly resistant at AMX
concentrations up to 25 mg/L, while the latter could partly be inactivated at lower AMX concentrations
(i.e. 10 mg/L) and/or in the presence of photocatalytic by-products.
Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction 2011). Antibiotics are used extensively in human and veterinary

Pharmaceuticals constitute an emerging class of micro-
pollutants, whose occurrence in the environment is due to point
sources, like production effluents and waste disposal, as well as
diffuse sources, like run-off from fields and anthropogenic effluents
(Kümmerer, 2009a). A variety of pharmaceuticals have been
detected in many environmental samples worldwide including
wastewater treatment plant effluents, surface water, seawater,
groundwater, soils and sediments and this is closely related to their
bioresistant nature (Fatta-Kassinos et al., 2010).
Among the various pharmaceutical compounds, antibiotics have
been paid particular attention because of their potential role in the
development of antibiotic-resistant bacteria (Homem and Santos,
* Corresponding author. Tel.: þ302821037797; fax: þ302821035852.
E-mail address: mantzavi@mred.tuc.gr (D. Mantzavinos).
medicine, as well as in aquaculture to prevent or treat microbial
infections. Most antibiotics tested to date are known to be bio-
recalcitrant under aerobic conditions (Kümmerer, 2009b), thus
escaping intact from conventional wastewater treatment plants. In
this light, non-biological methods have been employed to treat
antibiotics (and other pharmaceuticals too), such as advanced
oxidation processes, membrane separation, adsorption, coagula-
tion, as well as various combinations of them; two recent review
articles (Homem and Santos, 2011; Klavarioti et al., 2009)
summarize the more representative studies in this field.
Amoxicillin (AMX) is a semi-synthetic b-lactam antibiotic used
in human medicine to treat several diseases and in veterinary
practice as growth promoter. As it is excreted by the body almost
unmetabolized (e.g. excretion levels as much as 86  8% in human
urine have been reported (Garcia-Reiriz et al., 2007)), AMX itself
rather than a metabolite is likely to be found in environmental
samples (Morse and Jackson, 2004). AMX has been detected at the

0301-4797/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jenvman.2012.01.010
 D. Dimitrakopoulou et al. / Journal of Environmental Management 98 (2012) 168e174 169

mg/L level in environmental samples, like secondary treated Table 1

effluents and surface water (Kasprzyk-Hordern et al., 2008; TiO2 catalysts used in this study. A: anatase; R: rutile.

Schreiber and Szewzyk, 2008), while its presence in antibiotic Catalyst Crystal BET area, Particle Supplier
manufacturing effluents may reach the mg/L level (Arslan-Alaton form m2/g size, nm
et al., 2004). Degussa P25 75%A:25%R 50 21 Degussa AG
Advanced oxidation processes to degrade AMX in water Hombikat UV100 A > 99% <250 5 Sachtleben Chemie
GmbH
include ozonation (Andreozzi et al., 2005), dark Fenton oxidation
Millennium PC50 A > 97% 45e55 20e30 Millennium Inorganic
(Ay and Kargi, 2010), photo-Fenton oxidation (Mavronikola et al., Chemicals
2009; Trovo et al., 2008) and semiconductor photocatalysis using Millennium PC100 A > 95% 80e100 15e25 Millennium Inorganic
undoped and doped titania samples under UV-A or solar irradia- Chemicals
tion (Klauson et al., 2010). In other studies performed by a certain Millennium PC105 A > 95% 75e95 15e25 Millennium Inorganic
Chemicals
research group, the treatment of a synthetic antibiotic formulation Millennium PC500 A > 75% <300 5e10 Millennium Inorganic
wastewater containing amoxicillin, cloxacillin and ampicillin Chemicals
at a total concentration of about 300 mg/L was studied by means Tronox AK1 A > 97% 90 20 Kerr McGee Chemicals
of UV-A/ZnO photocatalysis (Elmolla and Chaudhuri, 2010a), Aldrich Anatase AA A > 99% 190e290 15 Aldrich
UV-A/TiO2 photocatalysis (Elmolla and Chaudhuri, 2010b) and
Fenton oxidation with and without UV-A irradiation (Elmolla and
Chaudhuri, 2010c).
the potassium ferrioxalate method and was found 8
104 E/
Since the literature concerning the heterogeneous photo-
(L min). Experiments were conducted in an immersion well, batch
catalytic treatment of AMX in environmentally relevant systems
type, laboratory scale photoreactor, purchased from Ace Glass
and/or concentrations is scarce, we decided to study, in a system-
(Vineland, NJ, USA) and described in detail elsewhere (Fotiadis
atic way, the effect of various process parameters such as TiO2 type
et al., 2007).
and loading, initial antibiotic concentration, solution pH and the
In a typical photocatalytic run, 0.35 L of the aqueous solution
water matrix (i.e. ultrapure water and treated wastewater) on the
containing the desired concentration of AMX in the range
kinetics of AMX degradation and mineralization, as well as on its
2.5e30 mg/L were loaded in the reaction vessel. These concentra-
antibiotic activity.
tions, although greater than those typically found in environmental
samples, were chosen to allow (i) the assessment of process effi-
2. Materials and methods ciency within a measurable time scale, and (ii) the accurate deter-
mination of residual AMX and organic carbon with the analytical
2.1. Materials techniques employed in this work. The solution was slurried with
the appropriate amount of catalyst and magnetically stirred for
AMX (C16H19N3O5S), whose structure is shown in Fig. 1, was 30 min in the dark to ensure complete equilibration of adsorption/
purchased from Fluka (CAS 26787-78-0) and used as received. Most desorption of AMX onto the TiO2 surface. After that period, the UV-
experiments were performed in ultrapure water taken from a water A lamp was turned on, while pure O2 was continuously sparged in
purification system (EASYpureRF-Barnstead/Thermolyne, USA); in the reaction mixture under stirring. During photocatalytic experi-
an additional experiment, the secondary effluent of a municipal ments, temperature was maintained at 25  C with a temperature
wastewater treatment plant was spiked with AMX to assess the control unit. Most experiments were performed at inherent solu-
effect of water matrix on degradation. Eight commercially available tion pH which was left uncontrolled during the reaction; the
TiO2 samples, whose main properties are summarized in Table 1, inherent solution pH was about 5 (the exact value was dependent
were employed for slurry photocatalytic experiments. of AMX initial concentration and the type of catalyst) in experi-
ments with ultrapure water and 7.5 in the experiment with the
2.2. Experimental procedure secondary effluent. For those runs with ultrapure water where the
initial pH had to be adjusted to 7.5, this was done adding the
UV-A irradiation was provided by a 9 W lamp (Radium Ralutec, appropriate amount of 0.5 N NaOH.
9W/78) emitting predominantly at 350e400 nm. The photon flux All of the experiments were run in duplicate and mean values
emitted from the lamp was determined actinometrically using are quoted as results with the standard error not exceeding 5%. The
only exception was the run in secondary effluent which was per-
formed once.
0.18

0.15 pKa2=7.5 2.3. Analytical techniques

NH2
H H
N S Samples periodically taken from the reactor were centrifuged to
Absorbance, AU

0.12 CH3
O N CH3 remove catalyst particles and then analyzed for (i) total organic
HO O
0.09 COOH carbon (TOC) content on a Shimadzu 5050A TOC analyzer, whose
pKa3=9.6 operation is based on the catalytic combustion/non-dispersive
pKa1=2.7
0.06
infrared gas analysis, to assess the extent of mineralization that
had occurred, and (ii) residual AMX concentration. The latter
was followed by high performance liquid chromatography
0.03
(Waters 2690) equipped with a Luna C18 (2) column (5 mm,
4.6 mm
250 mm) and a security guard column (4 mm
3 mm)
0
200 215 230 245 260 275 290 305
both purchased from Phenomenex, and a diode array detector
(Waters 996) set at 275 nm. The mobile phase consisted of 70%
Wavelength, nm
0.025 M KH2PO4 buffer solution in ultrapure water and 30%
Fig. 1. The absorbance spectrum and molecular structure of AMX. acetonitrile at a flow rate of 0.5 mL/min and inherent temperature.
170 D. Dimitrakopoulou et al. / Journal of Environmental Management 98 (2012) 168e174

2.4. Measurement of antibiotic activity
Antimicrobial susceptibility test was performed by determining
the minimum inhibitory concentration (MIC) of AMX using
Escherichia coli (ATCC 23716), Klebsiella pneumoniae (NCTC 5056)
and Enterococcus faecalis (ATCC 14506) as reference strains. MICs
were determined by the broth dilution method (microdilution),
labeling 96-well sterile microtitre trays with the appropriate anti-
biotic dilutions (Andrews, 2001; NCCLS, 2002). The range of the
tested antibiotic concentrations was between 5 and 30 mg/L.
Bacterial inocula were prepared using Mueller-Hinton broth
(Merck) adjusted to 2
104e105 CFU/mL, which is the appropriate
inoculum size for standard MIC. Microtitre trays were incubated for
18e24 h at 37  C, followed by optical density measurement at
630 nm, using a microplate reader (Labtech LT-4000 Plate Reader).
The same procedure was followed to test the antibiotic activity of
the treated AMX.
3. Results and discussion
3.1. Catalyst screening and activity
To assess the relative catalytic activity of various TiO2 samples,
preliminary experiments were conducted at 10 mg/L AMX initial
concentration, 250 mg/L catalyst loading and inherent solution pH.
Fig. 2 shows that Degussa P25 is highly active leading to complete
AMX disappearance after 25 min and over 90% mineralization after
90 min of reaction. To achieve similar AMX conversions with other
TiO2 samples, reaction times of at least 45 min are required, while
mineralization does not exceed 75% after 90 min. It should be noted
here that Degussa P25 is more active than the rest despite its lower
surface area (although the relationship between surface area and
photoactivity is debatable (Uyguner and Bekbolet, 2004)).
The superiority of Degussa P25 is attributed (Emilio et al., 2006;
Martin et al., 1994) to the slower electron/hole recombination
taking place on the catalyst surface compared to other TiO2 pho-
tocatalysts; another explanation ascribes the higher activity of
Degussa P25 to its structure which is a mixture of anatase and
rutile; this mixture is more active than the individual pure crys-
talline phases (Bickley et al., 1991). The results of Fig. 2 are not
surprising since Degussa P25 has been reported to be an effective
photocatalyst in numerous water/wastewater treatment applica-
tions (Malato et al., 2009). According to the above findings, all
subsequent photocatalytic experiments were performed with
Degussa P25 TiO2.
To confirm that AMX degradation is due to the synergy between
photonic energy and the catalyst surface rather than direct
photolysis and/or photooxidation, an additional experiment was
performed in the absence of catalyst leading to just 4% AMX
conversion after 90 min (Fig. 2a); this is consistent with the fact
that the absorbance spectrum of AMX (Fig. 1) fades out at about
290 nm, i.e. well below the emission spectrum of the irradiation
source.

a 3.2. Effect of catalyst loading

100
The effect of catalyst loading on 10 mg/L AMX degradation
80
and mineralization was studied in the range 100e750 mg/L and
inherent solution pH and the results are shown in Fig. 3. As seen,
AMX reduction, %

both AMX degradation and mineralization increase considerably

60 with increasing catalyst concentration from 100 to 250 mg/L
P25 UV100
beyond which the effect is less pronounced. Nevertheless,
AK1 PC50
PC100 PC105 mineralization appears to be more sensitive to catalyst concen-
40 PC500 AA tration changes than AMX degradation. For instance, the 20 min-
No catalyst TOC reduction is 43%, 60%, 65% and 70% at 100, 250, 500 and
750 mg/L catalyst, respectively; on the other hand, the 20 min-
20
AMX conversion is 65% at 100 mg/L catalyst and complete at any
other concentration. This is so because mineralization comprises
0 a sequence of reactions involving the oxidation of AMX photo-
0 15 30 45 60 75 90 catalytic transformation by-products to CO2 and water, which is
Time, min more difficult to occur than the partial AMX oxidation. Elmolla
b and Chaudhuri (2010b) reported a 42% AMX conversion after
100 300 min of UV-A irradiation at 500 mg/L pure anatase TiO2, which
P25 UV100
AK1 PC50 increased to 57  2% at 1500  500 mg/L catalyst. Their findings
PC100 PC105 are not directly comparable to ours since different experimental
80 PC500 AA conditions were employed including the type of catalyst (pure
anatase instead of Degussa P25), the photon flux (not explicitly
TOC reduction, %

60
given), the composition and concentration of the solution
(mixture of AMX with two other antibiotics at 300 mg/L cumu-
lative concentration instead of 10 mg/L AMX) and the catalyst to
40 substrate concentration ratio (in the range 1.7e6.7 in their work
and 10e75 in ours).
The increase in photocatalytic activity with increasing catalyst
20 loading indicates a heterogeneous catalytic regime since the
fraction of incident light absorbed by the semiconductor
progressively increases in suspensions containing higher amounts
0
of TiO2 (Abellan et al., 2009). The catalyst concentration above
0 15 30 45 60 75 90
which conversion levels off depends on several factors (e.g.
Time, min
reactor geometry, substrate concentration, wavelength and
Fig. 2. Relative activity of various TiO2 samples (250 mg/L) for 10 mg/L AMX intensity of light source) and corresponds to the point where all
(a) degradation and (b) mineralization at inherent pH ¼ 5. catalyst particles, i.e. all the surface exposed, are fully illuminated
 D. Dimitrakopoulou et al. / Journal of Environmental Management 98 (2012) 168e174 171

a 30 2.5
100
2
25

AMX concentration, mg/L

1/r , (L.min)/mg
1.5
80
20 1
AMX reduction, %

60 15
0.5

0
0 0.1 0.2 0.3 0.4

40 10 1/C , L/mg

100mg/L
250mg/L 5
20 500mg/L
750mg/L 0
0 0 5 10 15 20 25 30 35 40 45
0 15 30 45 60 75 90 Time, min
Time, min
Fig. 4. Change of AMX concentration as a function of its initial concentration at
b 500 mg/L Degussa P25 and inherent pH ¼ 5. Inset graph: the LeH kinetic model for
100
AMX degradation.

80
for the lower and higher concentration, respectively), the experi-
mental data fit the LeH model well (inset Fig. 4). From the intercept
TOC reduction, %

60 and the slope of the resulting straight line, the values for kr and K
are 2.28 mg/(L min) and 0.09 L/mg, respectively.
The ability of semiconductor photocatalysis to oxidize organic
40 species is due to the valence band holes photogenerated onto the
100mg/L titania surface, as well as the hydroxyl radicals produced from the
250mg/L reaction of holes with surface-bound water and hydroxyl anions
20 500mg/L (Turchi and Ollis, 1990). Hydroxyl radicals are considered as the
750mg/L primary oxidizing agent acting not only on the surface but also in
0
the solution phase; in this respect, the LeH model (and the asso-
0 15 30 45 60 75 90
ciated kinetic constants) should be viewed cautiously since it may
not take into account all possible photocatalytic mechanisms
Time, min
(Turchi and Ollis, 1989).
Fig. 3. Effect of Degussa P25 loading on 10 mg/L AMX (a) degradation and When the extent of adsorption and/or substrate concentration is
(b) mineralization at inherent pH ¼ 5. small, i.e. KCeq << 1, eqn. (1) is reduced to a first-order kinetic
expression as follows:
 
(Herrmann, 1999). At higher concentrations, a screening effect of dC Co
excess particles may occur, thus masking part of the photosensi-  ¼ kapp C5ln ¼ kapp t5ln 1  X ¼ kapp t (2)
dt C
tive surface and consequently hindering light penetration.
At the conditions employed in this work, P25 TiO2 in the range where kapp is an apparent rate constant and X is the conversion.
250e750 mg/L appears to have no significant effect on conversion, Fig. 5 shows the data of Fig. 4 plotted in the form of eqn. (2), where
i.e. it can provide a sufficient amount of catalyst particles without kapp values can be computed from the slopes of the resulting
impeding light penetration; therefore, all subsequent runs were
performed at 500 mg/L catalyst.
3

3.3. Effect of AMX concentration

2.5
The effect of varying AMX initial concentration (Co) was studied
in the range 2.5e30 mg/L at 500 mg/L Degussa P25 loading and the 2
results are shown in Fig. 4; regardless the initial concentration,
Ln(Co/C)

complete degradation is always achieved in less than 45 min of 1.5

reaction. Heterogeneous photocatalytic reactions can be described 30mg/L
by a LangmuireHinshelwood (LeH) kinetic model, i.e.: 1 20mg/L
10mg/L
KCeq 1 1 1 1
ro ¼ kr 5 ¼ þ (1) 0.5 5mg/L
1 þ KCeq ro kr K Ceq kr
2.5mg/L
where ro is the initial reaction rate, Ceq is the equilibrium concen- 0
tration, K is the adsorption constant onto the catalyst surface and kr 0 5 10 15 20 25 30
is the intrinsic reaction rate constant. Computing ro from the data of
Time, min
Fig. 4 for the first 4 min of reaction and taking Ceq z Co since the
extent of dark adsorption is quite low (i.e. between 3.2% and 13.7% Fig. 5. Plot of eqn. (2) for the AMX concentrationetime profiles of Fig. 4.
172 D. Dimitrakopoulou et al. / Journal of Environmental Management 98 (2012) 168e174
straight lines. These values are given in Fig. 6 alongside AMX
conversion after 4 min of reaction. (It should be noted that data
fitting to eqn. (2) was applied for AMX conversions up to 90% and
the coefficient of linear regression (r2) was always greater than
98.5%). The kinetics of AMX degradation in the range 2.5e5 mg/L
are, indeed, first order since the rate constant is independent of
the initial concentration (and so is conversion). However, at higher
concentrations both the rate constant and the conversion decrease,
thus implying deviation from first-order kinetics (although fitting
to eqn. (2) is still excellent). For example, if the reaction were zeroth
order with respect to AMX concentration, conversion would
decrease proportionately with increasing concentration, i.e.:
dC kapp
 ¼ kapp 5Co  C ¼ kapp t5X ¼ t (3)
dt Co
This is clearly demonstrated at the two higher concentrations,
where a 50% increase (from 20 to 30 mg/L) results in an almost
equal conversion decrease. For concentrations in the range
5e20 mg/L, the decrease in conversion is proportionately lower
than the respective increase in concentration, thus implying
kinetics between zeroth and first order.
This transition in kinetics can be explained as follows: the rate of
photocatalytic formation of oxidizing species should be a function
of the photon flux and the catalyst loading and, therefore, constant
at fixed operating conditions. In this respect, an increase in AMX
concentration would increase the probability of hydroxyl radicals
attack on AMX and, consequently, the reaction rate. This is partic-
ularly true during the early stages of reaction when the concen-
tration of competing transformation by-products is low. Fig. 7
shows their evolution depicted from the carbon balance between
residual AMX and measured TOC. This increase is linear in the range
2.5e5 mg/L, e.g. ro is 0.41 and 0.81 mg/(L min) at 2.5 and 5 mg/L,
respectively denoting first-order kinetics. At higher concentrations,
hydroxyl radicals become the limiting reactant and, consequently,
lower order kinetics would occur. For the range of initial concen-
trations examined, the photonic efficiency, F, takes values between
0.14% and 0.57%. The efficiency is computed as follows:
ro V
F ¼
100 (4)
I
where the initial reaction rate is expressed in M/min, I is the photon
flux and V is the liquid volume.
Klauson et al. (2010), who studied AMX degradation in the range
1e100 mg/L with 1000 mg/L Degussa P25 under UV-A radiation at
0.5 mW/cm2 and pH ¼ 6, reported about 90% conversion after
Fig. 6. Effect of initial AMX concentration on conversion (B) and the apparent rate
constant (C) of eqn. (2) at 500 mg/L Degussa P25 and inherent pH ¼ 5.
7
Concentration of by-products, mg C/L
6
5
4
3
30mg/L
2 20mg/L
5mg/L
1 2.5mg/L
0
0 5 10 15 20 25 30 35 40 45
Time, min
Fig. 7. Accumulation of liquid-phase transformation by-products during AMX degra-
dation at various initial concentrations, 500 mg/L Degussa P25 and inherent pH ¼ 5.
360 min at 10 mg/L AMX concentration; this value consistently
decreased increasing AMX concentration and became about 30% at
100 mg/L. This was explained considering a transition from the
kinetic control regime at lower concentrations to mass transfer
limitations at higher concentrations. In the present study, complete
conversion of 10 mg/L AMX was achieved after 25 min of reaction,
i.e. an order of magnitude faster than that reported by Klauson et al.
(2010); this may be due to a 10-fold greater irradiance employed in
this study since other experimental conditions were comparable.
Based on the kinetic analysis of this work, the photocatalytic
degradation of AMX (and alike pharmaceuticals) in environmen-
tally relevant matrices like surface water, groundwater, run-offs etc
is likely to follow first-order kinetics since its concentration will be
several orders of magnitude lower than in this study; conversely,
manufacturing effluents typically contain antibiotics at the mg/L
level and, in this case, lower order kinetics may, indeed, apply.
3.4. Effects of initial solution pH and the water matrix
Fig. 8 shows the effect of initial pH in ultrapure water on 10 mg/L
AMX degradation and mineralization. As clearly seen, increasing
pH from acidic to near-neutral conditions has no effect on AMX
degradation and this may be associated with the ionization states of
both the substrate and the catalyst. As seen in Fig. 1, AMX has three
pKa values at 2.7, 7.5 and 9.6 (Andreozzi et al., 2005). AMX is
positively charged at pH ¼ 5 (Elmolla and Chaudhuri, 2010a) and so
is the surface of Degussa P25, whose the point of zero charge is 6.7
(Fernandez-Ibanez et al., 2003). On the other hand, they are both
negatively charged at near-neutral/slightly alkaline conditions.
Therefore, neither pH appears to favor the electrostatic attraction
between AMX and the catalyst surface (the degree of dark
adsorption was measured 8  2% at either pH) and this may explain
the nearly identical degradation rates. These results are in good
agreement with those reported by Elmolla and Chaudhuri (2010b),
who found that a pH change from 5 to 8 had practically no influence
on AMX degradation (e.g. the 300 min-conversion was 57  2%).
Similarly, no striking differences were observed by Klauson et al.
(2010) regarding the effect of pH in the range 6e9 on AMX
degradation.
On the other hand, mineralization is favored at acidic conditions
with 93% TOC reduction occurring after 90 min, while the respec-
tive value at pH ¼ 7.5 is only 75%. If certain transformation by-
products were negatively charged at pH ¼ 5, their binding to the
 D. Dimitrakopoulou et al. / Journal of Environmental Management 98 (2012) 168e174
100
80
AMX or TOC reduction, %
60
40
AMX/pH=5
AMX/pH=7.5
TOC/pH=5
20
TOC/pH=7.5
AMX/pH=7.5 (SE)
0 0
0 15 30 45 60 75
Time, min
Fig. 8. Effect of initial solution pH and water matrix on 10 mg/L AMX degradation and
mineralization at 500 mg/L Degussa P25.
catalyst surface would be enhanced and this could increase
mineralization; nevertheless, this is speculative since identification
of by-products was not possible with the analytical tools employed
in this study. In addition, it is well known that TiO2 particles tend
to form agglomerates when dispersed in aqueous media, and
such agglomeration strongly depends on parameters such as ionic
strength and the pH of the suspension (French et al., 2009; Li et al.,
2010). It has been reported (Li et al., 2010) that agglomeration
of Degussa P25 particles in water decreases at acidic conditions
compared to neutral conditions, thus increasing the effective
surface area of the catalyst. This increase can also explain, in part,
the increased photocatalytic mineralization observed at acidic
conditions. If this is true, the fact that the transformation of AMX
to by-products is not accelerated at acidic conditions may be
explained by the fact that partial oxidation reactions are less
difficult to happen than total oxidation (mineralization).
To assess the role of water matrix, 10 mg/L AMX were spiked in
secondary effluent (SE) and subject to photocatalytic treatment at
inherent pH ¼ 7.5. As seen in Fig. 8, degradation in ultrapure water
is considerably faster than in SE, i.e. to achieve complete AMX
removal, the respective treatment times are 25 and 60 min. The SE
inherently contains about 8 mg/L of organic matter (this is 1.3 times
the organic carbon contained in 10 mg/L AMX) that would compete
with AMX for hydroxyl radicals and other oxidizing species, thus
reducing conversion. Furthermore, radical scavenging induced by
the presence of about 180 mg/L bicarbonates and 220 mg/L chlo-
rides in SE may also be responsible for decreased efficiency. These
results pinpoint the role of water matrix, which is likely to be
more crucial in real applications considering that far lower AMX
concentrations would typically exist in municipal wastewater
influents and discharges.
3.5. Antibiotic activity
Fig. 9 shows the antimicrobial activity of AMX against E. faecalis
as a function of its concentration (filled symbols). Partial activity is
developed in the range 10e30 mg/L AMX, with the extent of
inactivation not exceeding 42% at the higher concentration tested.
Similar tests with E. coli and K. pneumoniae revealed that both were
resistant up to 25 mg/L AMX, while only at 30 mg/L was antimi-
crobial activity partly developed, i.e. 52  1% inactivation (data not
shown for brevity). MIC values obtained in this study are rather
expected considering that all experiments have been conducted
with bacterial reference strains. According to Andrews (2001) the
173
40
Growth inhibition, %
30
20
10
0
5 10 15 20 25 30
90
AMX concentration, mg/L
Fig. 9. Growth inhibition of E. faecalis as a function of (C) original AMX concen-
tration in untreated solutions and (B) residual AMX concentration during oxidation
of 30 mg/L AMX at 500 mg/L Degussa P25 and inherent pH ¼ 5.
suggested ranges for MIC determinations regarding AMX are
0.25e128 mg/L for Enterobacteriaceae (E. coli and K. pneumoniae)
and 0.12e128 mg/L for Enterococci (E. faecalis). However, these
ranges may be different when dealing with environmental bacterial
isolates, as they may develop resistance producing chromosomal
beta-lactamases and inactivating the corresponding antibiotic
compounds (Lateef et al., 2005).
Fig. 9 also shows the effect of oxidation on the antimicrobial
properties of 30 mg/L AMX, where activity is plotted against the
residual, unreacted AMX concentration (open symbols) rather than
the respective treatment time. It is evident that photocatalytic
transformation by-products possess some activity to E. faecalis and
this is more pronounced at <5 mg/L AMX concentrations, where
activity seems to be exclusively due to by-products.
The respective tests (not shown) with E. coli and K. pneumoniae
resulted in complete removal of antibiotic activity as soon as the
residual AMX concentration dropped below 25 mg/L (i.e. after
5 min); transformation by-products accumulating in the reaction
mixture (i.e. 6.4 mg C/L after 30 min) appeared to have no antibiotic
and/or toxic effect.
4. Conclusions
The conclusions drawn from the present study can be summa-
rized as follows:
1) Semiconductor photocatalysis based on Degussa P25 TiO2 and
UV-A irradiation is an efficient method for the degradation of
AMX in water. Nonetheless, mineralization is a slower process
associated with the formation of persistent transformation by-
products. Process performance is affected by several factors,
namely irradiation time, photocatalyst type and loading, solu-
tion pH and the water matrix.
2) The rate of AMX degradation depends on the substrate to
oxidizing species concentration ratio. The order of reaction
regarding the substrate is first at low ratios and it decreases
towards zeroth at higher ratios. This is consistent with the
decreased conversions recorded in experiments with treated
effluent compared to runs in pure water since reactive oxygen
species are wasted in reactions with the effluent organic
matter.
3) The antibiotic activity of AMX depends on its concentration and
the test bacteria in question. In the case of E. coli and
K. pneumoniae, inactivation is exclusively due to the presence of
174 D. Dimitrakopoulou et al. / Journal of Environmental Management 98 (2012) 168e174
AMX at relatively high concentrations, i.e. over 25 mg/L, since
below this value no antimicrobial activity occurs regardless
the presence of transformation by-products in the reaction
mixture. Conversely, E. faecalis is less resistant to AMX and it is
also partially killed by its reaction by-products.
References
Abellan, M.N., Gimenez, J., Esplugas, S., 2009. Photocatalytic degradation of anti-
biotics: the case of sulfamethoxazole and trimethoprim. Catal. Today 144,
131e136.
Andreozzi, R., Canterino, M., Marotta, R., Paxeus, N., 2005. Antibiotic removal from
wastewaters: the ozonation of amoxicillin. J. Hazard. Mater. 122, 243e250.
Andrews, J.M., 2001. Determination of minimum inhibitory concentrations.
J. Antimicrob. Chemother. 48, 5e16.
Arslan-Alaton, I., Dogruel, S., Baykal, E., Gerone, G., 2004. Combined chemical and
biological oxidation of penicillin formulation effluent. J. Environ. Manage. 73,
155e163.
Ay, F., Kargi, F., 2010. Advanced oxidation of amoxicillin by Fenton’s reagent treat-
ment. J. Hazard. Mater. 179, 622e627.
Bickley, R.I., Gonzalez-Carreno, T., Lees, J.S., Palmisano, L., Tilley, R.J.D., 1991.
A structural investigation of titanium dioxide photocatalysts. J. Solid State
Chem. 92, 178e190.
Elmolla, E.S., Chaudhuri, M., 2010a. Degradation of amoxicillin, ampicillin and
cloxacillin antibiotics in aqueous solution by the UV/ZnO photocatalytic
process. J. Hazard. Mater. 173, 445e449.
Elmolla, E.S., Chaudhuri, M., 2010b. Photocatalytic degradation of amoxicillin,
ampicillin and cloxacillin antibiotics in aqueous solution using UV/TiO2 and UV/
H2O2/TiO2 photocatalysis. Desalination 252, 46e52.
Elmolla, E.S., Chaudhuri, M., 2010c. Comparison of different advanced oxidation
processes for treatment of antibiotic aqueous solution. Desalination 256,
43e47.
Emilio, C.A., Litter, M.I., Marinus, K., Bouchard, M., Colbeau-Justin, C., 2006. Phenol
degradation on platinized-TiO2 photocatalysts related to charge-carrier
dynamics. Langmuir 22, 3606e3613.
Fatta-Kassinos, D., Bester, K., Kümmerer, K., 2010. Xenobiotics in the Urban Water
Cycle: Mass Flows, Environmental Processes, Mitigation and Treatment Strat-
egies. In: Environmental Pollution, vol. 16. Springer.
Fernandez-Ibanez, P., Blanco, J., Malato, S., de las Nieves, F.J., 2003. Application of
the colloidal stability of TiO2 particles for recovery and reuse in solar photo-
catalysis. Water Res. 37, 3180e3188.
Fotiadis, C., Xekoukoulotakis, N.P., Mantzavinos, D., 2007. Photocatalytic treatment
of wastewater from cottonseed processing: effect of operating conditions,
aerobic biodegradability and ecotoxicity. Catal. Today 124, 247e253.
French, R.A., Jacobson, A.R., Kim, B., Isley, S.L., Lee Penn, R., Baveye, P.C., 2009.
Influence of ionic strength, pH, and cation valence on aggregation kinetics of
titanium dioxide nanoparticles. Environ. Sci. Technol. 43, 1354e1359.
Garcia-Reiriz, A., Damiani, P.C., Olivieri, A.C., 2007. Different strategies for the direct
determination of amoxicillin in human urine by second-order multivariate
analysis of kinetic-spectrophotometric data. Talanta 71, 806e815.
Herrmann, J.M., 1999. Heterogeneous photocatalysis: fundamentals and applica-
tions to the removal of various types of aqueous pollutants. Catal. Today 53,
115e129.
Homem, V., Santos, L., 2011. Degradation and removal methods of antibiotics from
aqueous matrices e a review. J. Environ. Manage. 92, 2304e2347.
Kasprzyk-Hordern, B., Dinsdale, R.M., Guwy, A.J., 2008. The occurrence of phar-
maceuticals, personal care products, endocrine disruptors and illicit drugs in
surface water in South Wales, UK. Water Res. 42, 3498e3518.
Klauson, D., Babkina, J., Stepanova, K., Krichevskaya, M., Preis, S., 2010. Aqueous
photocatalytic oxidation of amoxicillin. Catal. Today 151, 39e45.
Klavarioti, M., Mantzavinos, D., Kassinos, D., 2009. Removal of residual pharma-
ceuticals from aqueous systems by advanced oxidation processes. Environ. Int.
35, 402e417.
Kümmerer, K., 2009a. Antibiotics in the aquatic environment e a review e part I.
Chemosphere 75, 417e434.
Kümmerer, K., 2009b. The presence of pharmaceuticals in the environment due to
human use e present knowledge and future challenges. J. Environ. Manage. 90,
2354e2366.
Lateef, A., Oloke, J.K., Gueguimkana, E.B., 2005. The prevalence of bacterial resis-
tance in clinical, food, water and some environmental samples in Southwest
Nigeria. Environ. Monit. Assess. 100, 59e69.
Li, G., Lv, L., Fan, H., Ma, J., Li, Y., Wan, Y., Zhao, X.S., 2010. Effect of the agglomeration
of TiO2 nanoparticles on their photocatalytic performance in the aqueous
phase. J. Colloid Interface Sci. 348, 342e347.
Malato, S., Fernandez-Ibanez, P., Maldonado, M.I., Blanco, J., Gernjak, W., 2009.
Decontamination and disinfection of water by solar photocatalysis: recent
overview and trends. Catal. Today 147, 1e59.
Martin, S.T., Herrmann, H., Choi, W., Hoffmann, M.R., 1994. Time-resolved micro-
wave conductivity. Part 1. TiO2 photoreactivity and size quantization. J. Chem.
Soc. Faraday Trans. 21, 3315e3322.
Mavronikola, C., Demetriou, M., Hapeshi, E., Partassides, D., Michael, C.,
Mantzavinos, D., Kassinos, D., 2009. Mineralisation of the antibiotic amoxicillin
in pure and surface waters by artificial UVA- and sunlight-induced Fenton
oxidation. J. Chem. Technol. Biotechnol. 84, 1211e1217.
Morse, A., Jackson, A., 2004. Fate of amoxicillin in two water reclamation systems.
Water Air Soil Poll 157, 117e132.
National Committee for Clinical Laboratory Standards, 2002. Performance Stan-
dards for Antimicrobial Disk and Dilution Susceptibility Tests for Bacteria Iso-
lated from Animals. Approved standard M31eA2. National Committee for
Clinical Laboratory Standards, Wayne, PA.
Schreiber, F., Szewzyk, U., 2008. Environmentally relevant concentrations of phar-
maceuticals influence the initial adhesion of bacteria. Aquatic Toxicol. 87,
227e233.
Trovo, A.G., Santos Melo, S.A., Pupo Nogueira, R.F., 2008. Photodegradation of the
pharmaceuticals amoxicillin, bezafibrate and paracetamol by the photo-Fenton
process e application to sewage treatment plant effluent. J. Photochem. Pho-
tobiol. A 198, 215e220.
Turchi, C.S., Ollis, D.F., 1989. Mixed reactant photocatalysis: intermediates and
mutual rate inhibition. J. Catal. 119, 483e496.
Turchi, C.S., Ollis, D.F., 1990. Photocatalytic degradation of organic water contami-
nants: mechanisms involving hydroxyl radical attack. J. Catal. 122, 178e192.
Uyguner, C.S., Bekbolet, M., 2004. Photocatalytic degradation of natural organic
matter: kinetic considerations and light intensity dependence. Int. J. Photo-
energy 6, 73e80.