Water Research 90 (2016) 301e308

Contents lists available at ScienceDirect

Water Research
journal homepage: www.elsevier.com/locate/watres

Kinetics and pathways of ibuprofen degradation by the UV/chlorine

advanced oxidation process
Yingying Xiang a, b, c, Jingyun Fang a, b, *, Chii Shang c, d, **
a
SYSU-HKUST Research Center for Innovative Environmental Technology (SHRCIET), School of Environmental Science and Engineering, Sun Yat-Sen
University, Guangzhou 510275, PR China
b
Guangdong Provincial Key Laboratory of Environmental Pollution Control and Remediation Technology (Sun Yat-Sen University), Guangzhou 510275, PR
China
c
Department of Civil and Environmental Engineering, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong
d
Hong Kong Branch of Chinese National Engineering Research Center for Control & Treatment of Heavy Metal Pollution, Hong Kong

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

Article history: The UV/chlorine advanced oxidation process (AOP), which forms reactive species such as hydroxyl
Received 29 August 2015 radicals (HO ) and reactive chlorine species (RCS) such as chlorine atoms (Cl ) and Cl2, is being

Received in revised form considered as an alternative to the UV/H2O2 AOP for the degradation of emerging contaminants. This
27 November 2015
study investigated the kinetics and pathways of the degradation of a recalcitrant pharmaceutical and
Accepted 28 November 2015
Available online 12 December 2015
personal care product (PPCP)dibuprofen (IBP)dby the UV/chlorine AOP. The degradation of IBP followed
the pseudo first-order kinetics. The first-order rate constant was 3.3 times higher in the UV/chlorine AOP
than in the UV/H2O2 AOP for a given chemical molar dosage at pH 6. The first-order rate constant
Keywords:
decreased from 3.1  103 s1 to 5.5  104 s1 with increasing pH from 6 to 9. Both HO and RCS

Pharmaceutical and personal care products

(PPCPs) contributed to the degradation, and the contribution of RCS increased from 22% to 30% with increasing



Ibuprofen pH from 6 to 9. The degradation was initiated by HO -induced hydroxylation and Cl -induced chlorine
UV/chlorine advanced oxidation process substitution, and sustained through decarboxylation, demethylation, chlorination and ring cleavage to
Hydroxyl radicals form more stable products. Significant amounts of chlorinated intermediates/byproducts were formed
Chlorine atoms from the UV/chlorine AOP, and four chlorinated products were newly identified. The yield of total organic
Water treatment chlorine (TOCl) was 31.6 mM after 90% degradation of 50 mM IBP under the experimental conditions. The
known disinfection by-products (DBPs) comprised 17.4% of the TOCl. The effects of water matrix in
filtered drinking water on the degradation were not significant, demonstrating the practicality of the UV/
chlorine AOP for the control of some refractory PPCPs. However, the toxicity of the chlorinated products
should be further assessed.
© 2015 Elsevier Ltd. All rights reserved.

1. Introduction Many PPCPs are resistant to conventional water and wastewater

treatment such as coagulation, sedimentation, filtration and bio-
Pharmaceutical and personal care products (PPCPs) have been logical processes (Vieno et al., 2007). Activated carbon adsorbs
discovered in wastewater effluent, seawater, surface water and some PPCPs mainly through hydrophobic interaction, but it has
drinking water around the world (Vieno et al., 2007). Many PPCPs difficulty removing compounds with low hydrophobicity (e.g.,
are toxic to aquatic organisms and their presence potentially poses naproxen, sotalol, and atenolol) (Vieno et al., 2007; Huerta-Fontela
a long-term threat to human health (Christen et al., 2010). et al., 2011). Ozone, chlorine and chlorine dioxide are selective
oxidants that can effectively degrade PPCPs with electron-donating
functional groups such as phenolic- and amine-containing com-
pounds (Lee and von Gunten, 2010). However, they work poorly
* Corresponding author. School of Environmental Science and Engineering, Sun against recalcitrant PPCPs such as ciprofloxacin, phenytoin and
Yat-Sen University, Guangzhou 510275, PR China.
atenolol (Vieno et al., 2007; Huerta-Fontela et al., 2011). UV
** Corresponding author. Department of Civil and Environmental Engineering, The
Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong photolysis is also selective but abatement of many PPCPs using this
Kong. method requires much higher UV fluences (~1500 mJ/cm2) than
E-mail addresses: fangjy3@mail.sysu.edu.cn (J. Fang), cechii@ust.hk (C. Shang).

http://dx.doi.org/10.1016/j.watres.2015.11.069
0043-1354/© 2015 Elsevier Ltd. All rights reserved.
302 Y. Xiang et al. / Water Research 90 (2016) 301e308

does disinfection (Rosenfeldt and Linden, 2004). purchased from SigmaeAldrich. Sodium sulfite (Na2SO3) was pur-
Advanced oxidation processes (AOPs) are effective for the chased from Fluka, and sodium hydroxyl (NaOH), sulfuric acid
degradation of recalcitrant PPCPs. The UV/peroxide (H2O2) AOP (H2SO4) and bicarbonate (NaHCO3) from VWR BDH Prolabo. Phos-


producing hydroxyl radicals (HO ) has been reported to unselec- phoric acid, methanol, acetonitrile and methyl tert-butylether
tively oxidize many PPCPs such as diatrizoic acid, metronidazole, (MtBE) of HPLC grade were purchased from Fisher Scientific.
phenazone, phenytoin (Wols et al., 2013; Yuan et al., 2009). How- The IBP stock solution (7 mM) was prepared by adding the pure
ever, the radical production is not efficient due to the low UV compound to deionized water containing 7 mM NaOH and stirring
absorbance of H2O2 (for example, εH2O2, 254 nm ¼ 19.6 M1 cm1), overnight. A free chlorine stock solution (about 2000 mg/L as Cl2)
which requires excessive H2O2 dosing (Barakat et al., 2005). In was diluted from the 4% sodium hypochlorite (NaClO) solution and
drinking water practices, only 5e10% of the added H2O2 is periodically standardized by DPD/FAS titration (USEPA, 1998). A
consumed during treatment (Rosenfeldt et al., 2013). The unreacted sodium sulfite solution (20 mM) was prepared weekly. A real
H2O2 requires quenching, which adds to the treatment cost. drinking water sample was collected after sand filtration from a
water treatment plant in Guangzhou, China and stored at 4  C. The


The UV/chlorine AOP producing HO and reactive chlorine spe-
cies (RCS) such as chlorine atoms (Cl ) (Eq. (1)) and Cl2, is recently



water quality parameters of the sample are shown in
being considered as an alternative to the UV/H2O2 AOP (Fang et al., Supplementary Information Table S1.
2014). The remaining chlorine can be used for residual protection in
drinking water distribution systems. The UV/chlorine AOP is found 2.2. UV irradiation
to be more efficient in producing reactive species at acidic and
neutral pH (Watts et al., 2007; Fang et al., 2014). Figure S1 shows the schematic diagram of the UV reactor used in
this study. A low pressure (LP) UV lamp (Heraeus, GPH 212T5L/4,
HOCl/OCl þ hv / HO /O  þ Cl




(1) 10W, 254 nm) in a quartz sleeve was placed in the centerline of a
700 mL cylindrical glass reactor with rapid mixing provided at the
The diverse reactive species, such as HO , Cl and Cl2, that




bottom of the reactor. A water circulating system kept the inner
coexist in the UV/chlorine system may complement each other in tank temperature at 25 ± 0.1  C. The photo flux (I0) was determined
to be 0.44 mE s1 for kinetic tests and 0.62 mE s1 for other tests


degrading a variety of contaminants. HO is a nonselective oxidant,
and it reacts with organic moieties with rate constants close to using iodide-iodate chemical actinometry (Bolton and Linden,
diffusion-controlled rates (Lee and von Gunten, 2010). Cl and Cl2



2003). The effective path length (L) was determined to be 3.5 cm
are strong oxidants with oxidation potentials of 2.47 V and 2.0 V, by measuring the photolysis rate of H2O2 (Garoma and Gurol,
2005). The average UV fluence rate (E0p) was determined to be


respectively (Hirakawa and Nosaka, 2002; Beitz et al., 1998). Cl is a
selective oxidant that reacts with electron-rich moieties through 1.05 mW cm2 for kinetic tests and 1.47 mW cm2 for other tests.
one-electron oxidation, H-abstraction, and addition to unsaturated


CeC bonds (Grebel et al., 2010). Moreover, Cl can react more 2.3. Experimental procedure


strongly with benzoic acid and phenol than can HO (Fang et al.,
2014). The UV/chlorine AOP is more efficient than the UV/H2O2 All experiments were carried out in the 700 mL photoreactor
AOP in the degradation of some emerging contaminants such as described above. The UV lamp and the water circulating system
benzotriazole, tolytriazole and iopamidole (Sichel et al., 2011) and were warmed up at least 30 min before experiments. Experiments
micropollutants such as trichloroethylene (Wang et al., 2012). were initiated by adding the NaClO stock solution to the UV reactor
However, the effectiveness of the UV/chlorine process in degrading containing the solution spiked with 10 mM IBP and 2 mM phosphate
buffer to give an initial chlorine concentration of 10e100 mM and



recalcitrant PPCPs is less known. The actions of radicals (HO , Cl and




Cl2 ), especially RCS (Cl and Cl2 ), on the transformation mecha- pH of 6, 7 or 9. Samples (1 mL) were collected at different time
nisms of recalcitrant PPCPs in the UV/chlorine AOP are currently intervals, quenched with 20 mM Na2SO3 (at a molar ratio of sulfite/
unclear. Moreover, whether or not the attacks by the reactive chlorine of 1.5) and stored at 4  C in the dark before HPLC analysis.
chlorine species (RCS) in the UV/chlorine process would cause Another test was conducted in the same manner with the initial
undesirable chlorinated byproducts to form is unknown and de- H2O2 concentration of 100 mM at pH 6. Control tests of IBP degra-
mands investigation. dation by UV direct photolysis and dark chlorination were also
Ibuprofen (IBP), a nonsteroidal anti-inflammatory drug, was conducted in a similar manner but in the absence of chlorine and
chosen as the target recalcitrant PPCP for its resistance to con- UV light, respectively.
ventional water treatment processes, its frequent occurrence in To identify the degradation intermediates/products and quan-
surface water, and its potential long-term adverse effects (Vieno tify the concentrations of total organic chlorine (TOCl) and disin-
et al., 2007; Ruggeri et al., 2013). The aims of this study were to fection by-products (DBPs) during IBP degradation by the UV/
(1) evaluate the degradation kinetics of IBP by the UV/chlorine AOP chlorine AOP, higher initial concentrations of IBP (50 mM) and
under varied conditions; (2) investigate the relative contributions chlorine (200 mM) were used. After a certain reaction time (1, 2, 5,
of the reactive species to the degradation; and (3) evaluate the 15 or 30 min) in each run, the reaction was quenched with 100 mM
formation of chlorinated byproducts and reveal the degradation Na2SO3 and a 400 mL sample was collected and subjected to the
pathways of IBP by the UV/chlorine AOP. pretreatment procedure (Text S1) developed by Zhai and Zhang
(2009) for UPLC/ESI-tqMS and UPLC/ESI-QTOF-MS analyses. Addi-
2. Materials and methods tional tests at pH 0 were conducted following the same procedure
but with the pH of the reaction mixture adjusted to zero using 95%
2.1. Chemicals H2SO4 prior to adding the NaClO stock solution to the reactor.

All solutions were prepared from reagent-grade chemicals and 2.4. Analytical methods
deionized water (18.2 U-cm) produced from a Milli-Q water puri-
fication system (Millipore, Reference). IBP, sodium hypochlorite Concentrations of IBP and NB were measured using a high
solution (NaClO), hydroxyl peroxide (H2O2), sodium sulfate performance liquid chromatography (HPLC) system (Agilent Tech-
(Na2SO4), sodium chloride (NaCl) and nitrobenzene (NB) were nologies 1260 Series) equipped with a Waters symmetry C18
 Y. Xiang et al. / Water Research 90 (2016) 301e308 303

column (4.6 mm  150 mm, 5 mm particle size) and a UV detector

set at 223 nm and 262 nm, respectively. The mobile phase for the 1.0
measurement of IBP and NB consisted of water (pH 2, adjusted by
phosphoric acid)/methanol (20:80, v/v%) and water/acetonitrile
(50:50, v/v%), respectively, at a flow rate of 1.0 mL/min. The con- 0.8
centration of free chlorine was determined using DPD/FAS titration
(USEPA, 1998). The concentration of H2O2 was determined by the I

[IBP] / [IBP]0
3
method (Klassen et al., 1994). The pH was measured with a pH 0.6 -9.35*10-4t
C/C0=e
meter (Thermo Scientific, Orion Star A111). 2
R =0.9993
The intermediates and products of IBP, after the aforementioned
pretreatment, were identified with a Waters UPLC system coupled 0.4
to the Waters Acquit ESI-tqMS device (UPLC/ESI-tqMS) and
equipped with a HSS T3 column (50  2.1 mm, 1.8 mm particle size, UV photolysis -3.06*10-3t
Waters). The mobile phase was composed of methanol and water at 0.2 C/C0=e
Dark chlorination
a flow rate of 0.4 mL/min with the column temperature of 35  C. UV/H2O2
2
R =0.9968
The composition of methanol/water (v/v%) changed linearly from UV/chlorine
5:95 to 90:10 in the first 8 min, returned to 10:90 in the next 0.0
0.1 min, and then changed linearly back to 5:95 in the following 0 200 400 600 800 1000 1200
Time (s)
2.9 min. The accurate m/z and the proposed chemical formula were
obtained from a Bruker maxis 4G UPLC/ESI-QTOF-MS system Fig. 1. Comparison of the time-dependent IBP degradation in pure water by UV direct
equipped with an Agilent SB-C18 column (50  2.1 mm, 1.8 mm photolysis, dark chlorination, the UV/chlorine AOP and the UV/H2O2 AOP. Conditions:
particle size). The mobile phase consisted of water and acetonitrile pH 6, [IBP]0 ¼ 10 mM, [H2O2]0 ¼ 100 mM, [chlorine]0 ¼ 100 mM, UV fluence
(ACN) at a flow rate of 0.4 mL/min. The composition of ACN/water rate ¼ 1.05 mW cm2.

(v/v%) changed linearly from 5:95 to 90:10 in the first 8 min,

remained at 90:10 for 10 min and returned to 10:90 in the next
3.2. Effects of chlorine dosages, pH, bicarbonate, chloride and
0.1 min, then changed linearly back to 5:95 in the following 2.9 min.
temperature
The values of MS parameters for the UPLC/ESI-tqMS and UPLC/ESI-
QTOF-MS are shown in Text S1. The TOCl was determined on a
Fig. 2a and b display the effect of chlorine dosages (0e100 mM)
precombustion station (AQF-100 Mitsubishi) with an ion chro-
and pH on the degradation of IBP by the UV/chlorine AOP. The IBP
matograph (IC) system (Dionex ICS 3000), following Standard
degradation rate increased monotonically with increasing chlorine
Method 5320B (USEPA, 1998). The DBPs were quantified using a gas
dosage. The pseudo first-order rate constants (k0 ) of the IBP
chromatograph coupled with an electronic capture detector (GC-
degradation increased from 6.1  104 s1 to 3.1  103 s1 with
ECD, Agilent Technologies) and equipped with a HP-5MS fused
increasing chlorine dosage from 10 mM to 100 mM, as shown in the
silica capillary column (30 m  0.25 mm, 0.25 mM, Agilent J&W).
inset of Fig. 2a. However, k0 increased more gradually with
The procedures and parameters for the measurements of volatile
increasing chlorine dosage at chlorine dosages above 70 mM. The
DBPs and haloacetic acids followed USEPA methods 551 and 552.3
effects of chlorine dosage on k0 of IBP were similar to that of benzoic
(USEPA, 1998), respectively, with details shown in Text S2.
acid (Fang et al., 2014), because of the scavenging of the reactive
species by excess free chlorine (Eqs (2)e(5)).

HO
þ HOCl / ClO
þ H2O k1 ¼ 2.0  109 M1s1 (2)

3. Results and discussion

HO
þ OCl / ClO
þ OH k2 ¼ 8.8  109 M1s1 (3)

3.1. Comparison of IBP degradation by the UV/chlorine and UV/

Cl
þ HOCl / Hþ þ Cl þ ClO
k3 ¼ 3.0  109 M1s1 (4)
H2O2 AOPs
Cl
þ OCl / Cl þ ClO
k4 ¼ 8.2  109 M1s1 (5)
Fig. 1 shows the time-dependent degradation of IBP by UV direct
photolysis, dark chlorination, the UV/chlorine AOP and the UV/
HO
þ IBP / products k5 ¼ 7.4(±1.2)  109 M1s1 (6)
H2O2 AOP at the same chemical molar dosage and pH 6. Only 10.2%
and 3.1% of IBP were degraded after 20 min by UV direct photolysis
k0 of IBP degradation by the UV/chlorine AOP decreased from
and dark chlorination, respectively, reconfirming the recalcitrance
3.1  103 s1 to 5.5  104 s1 with increasing pH from 6 to 9. The
of IBP to UV photolysis and chlorination (Yuan et al., 2009; Pinkston
effects of pH were similar to those of benzoic acid and trichloro-
and Sedlak, 2004). The IBP degradation by the UV/chlorine AOP and
ethylene (Fang et al., 2014; Wang et al., 2012) because HOCl has a
the UV/H2O2 AOP followed the pseudo first-order kinetics, with
higher quantum yield and smaller radical scavenging effect than
rate constants of 3.1  103 s1 and 9.4  104 s1, respectively. The OCl (Eqs. (2)e(5)) (Fang et al., 2014; Nowell and Hoigne
, 1992).
3.3 times higher rate constant of the former can be partially
Figures S2 and S3 show the effects of HCO 3 and chloride on the
explained by the higher radical production in the UV/chlorine AOP
degradation of IBP by the UV/chlorine AOP. k0 decreased from
than in the UV/H2O2 AOP at pH 6. The quantum yields of HOCl, OCl
9.23  104 s1 to 7.38  104 s1 with increasing HCO 3 concen-
and H2O2 under UV irradiation are 1.0e1.5, 0.87e1.3 and 0.5e1.0,
tration from 0.2 mM to 4 mM at pH 7. HCO 3 competed with IBP for
respectively (Jin et al., 2011; Wang et al., 2012; Crittenden et al.,
HO and Cl with the rate constants of 8.5  106 M1 s1 and

1999; Shu et al., 2014). The molar absorption coefficients of HOCl,

2.2  10 M1 s1, respectively (Fang et al., 2014). The reaction
8
OCl and H2O2 are 59, 66 and 19.6 M1 cm1, respectively (Barakat
between HCO 




3 and either HO or Cl generates CO3 , which is less
et al., 2005; Feng et al., 2007). The comparable or higher quantum
reactive to IBP (Vione et al., 2011). The degradation was not affected
yields and higher molar absorption coefficients of chlorine species
by chloride at 0.2e10.2 mM, consistent with the results from Fang
led to a higher production of radicals in the UV/chlorine AOP.
304 Y. Xiang et al. / Water Research 90 (2016) 301e308

1.1

1.0

0.9

0.8

[IBP] / [IBP]0
0.7

0.6

0.5

0.4

0.3

0.2
Ultrapure water
0.1 Real drinking water

0.0
0 200 400 600 800 1000 1200
Time (s)

Fig. 3. Time-dependent IBP degradation in real water by the UV/chlorine AOP. Con-
ditions: pH 7, [IBP]0 ¼ 10 mM, [chlorine]0 ¼ 100 mM, UV fluence rate ¼ 1.05 mW cm2.

chloride, bromide and nitrate in the real water were 0.67 mg/L,
0.98 mM, 12.3 mg/L, < 0.01 mg/L and 1.267 mg/L (as N), respectively
(Table S1). The effect of chloride at 12.3 mg/L on the IBP degrada-
tion was less than 3% (Figure S3). UV photolysis of nitrate produces



HO , but the formation of HO is negligible due to the low absorption
1 1
(3 M cm ) and quantum yield (0.09) of nitrate at 254 nm (Mack
and Bolton, 1999). DOM could act as UV254nm filter and radicals
scavenger to decrease the production of radicals and increase the
consumption of radicals, respectively (Fang et al., 2014). The rate
constant of DOM with HO was reported as 2.5  104 (mg/L)1 s1

(Katsoyiannis et al., 2011), and that with Cl was estimated to be
4 1 1
1.3  10 (mg/L) s (Fang et al., 2014). The presence of 1 mM
bicarbonate in deionized water depressed the pseudo first-order
rate constants of IBP degradation from 1.02  103 s1 to
Fig. 2. Time-dependent IBP degradation in pure water by the UV/chlorine AOP at 7.92  104 s1 (Figure S2). Therefore, the 9.2% lower IBP degra-
different chlorine dosages (a) and different pH (b). The inset shows the pseudo-first- dation rate in real water was mainly due to the competition of bi-
order rate constants as a function of chlorine dosages (a) and pH (b). Conditions:

carbonate and DOM with IBP for HO and Cl .

[IBP]0 ¼ 10 mM, [chlorine]0 ¼ 0e100 mM, pH ¼ 6e9, UV fluence rate ¼ 1.05 mW cm2.

3.4. Relative contributions of reactive species

et al. (2014).
Temperature did not have a significant effect on the degradation HO and reactive chlorine species (RCS) (e.g., Cl and Cl2) are

of IBP by the UV/chlorine AOP at pH 6 (Figure S4). k0 increased from likely the main reactive species in the UV/chlorine AOP, among
1.9  103 s1 to 3.1  103 s1 with increasing temperature from which Cl2 is generally less reactive (Grebel et al., 2010). As shown

5  C to 25  C. Based on the Arrhenius equation, the activation en- in Figure S3, the degradation rates of IBP were not affected by the
ergy Ea was estimated to be 18.02 kJ/mol and the Arrhenius con- increasing chloride concentration from 0.2 to 10.2 mM, which on
stant was 4.58 s1. Generally, reactions involving radicals such as the other hand increased the concentration of Cl2 by two orders of

HO and e


aq are not strongly affected by temperature due to the



magnitude. Meanwhile, the concentrations of HO and Cl changed
relatively small activation energies (6e30 kJ/mol) (Buxton et al., negligibly (Fang et al., 2014). Therefore, the constant degradation
1988). However, increasing temperature decreases the viscosity of rate of IBP suggests that the contribution of Cl2 to the degradation

water and may increase the quantum yield of HOCl/OCl photolysis

was negligible and that HO and Cl were the main reactive species
(Herrmann, 2007). It should be noted that dissolved oxygen was during the degradation.
present in the process, and its effect on the degradation of IBP could

The relative contributions of HO and RCS to IBP degradation by
be neglected (Figure S5). the UV/chlorine AOP at pH 6e9 were studied using NB as the probe
compound. NB only reacted with HO (kHO-NB ¼ 3.9  109 M1 s1)

and its reaction with Cl was negligible (Watts and Linden., 2007).
3.3. IBP degradation in real water

The relative contributions of HO and RCS were determined based
on the steady-state assumption of radical concentrations, as shown
Fig. 3 shows the IBP degradation in the real water sample within
in Eqs. (7) and (8).
the first 20 min. The degradation rate of IBP in the real water
sample was 9.2% lower than that in deionized water at 20 min. k0
kNB 0 ¼ kHO$NB ½HO$ss (7)
were 7.90  104 s1 and 1.02  103 s1, respectively, in real water
and deionized water. The concentrations of DOC, bicarbonate,
 Y. Xiang et al. / Water Research 90 (2016) 301e308 305

kIBP 0 ¼ kHO$IBP ½HO$ss þ kRCSIBP 0 þ kUVIBP 0 (8)

where kIBP0 and kNB0 are defined as the overall pseudo first-order


degradation rate constants of IBP and NB, respectively. kHO-IBP and



kHO-NB represent the second-order rate constants of HO reacting
0 0
with IBP and NB, respectively. kRCS-IBP and kUV-IBP represent the
pseudo first-order rate constants of IBP degradation by RCS
oxidation and UV photolysis, respectively. kUV-IBP0 was obtained
experimentally from the IBP degradation by UV direct photolysis.
[HO ]ss is defined as the steady-state concentration of HO . kNB0 was

experimentally obtained, and [HO ]ss was then calculated according


to Eq. (7) with kHO-NB obtained from the literature (Watts et al.,
2007). The only unknown in Eq. (8) was kRCS-IBP0, which was fitted



using Eq. (8). Then the relative contributions of HO and Cl to the
degradation can be expressed as kHO-IBP[HO ]ss/kIBP0 and kRCS-IBP0 /

kIBP0, respectively.
Fig. 4 shows the calculated pseudo first-order rate constants of


IBP degradation by HO and RCS oxidation and UV direct photolysis
in the UV/chlorine AOP at different pHs. At pH 6, the relative con-



tributions of HO and Cl to IBP degradation were 75% and 22%, Fig. 5. The formation of total organic chlorine (TOCl) and known disinfection
respectively, while UV direct photolysis contributed to 3% only. byproduct (DBP) from the degradation of IBP by the UV/chlorine AOPas a function of

time and by 30-min dark chlorination. Conditions: pH 6, [IBP]0 ¼ 50 mM,
With increasing pH from 6 to 9, the relative contribution of HO to [chlorine]0 ¼ 200 mM, UV fluence rate ¼ 1.47 mW cm2.
IBP degradation decreased to 53%, while the relative contributions


of Cl and UV direct photolysis increased to 30% and 17%,
respectively. remaining 82.6% TOCl is unknown. Meanwhile, the yield of the
known DBPs comprised 0.3% of the degraded IBP (as per mole of
3.5. The formation of TOCl and known DBPs during IBP degradation carbon) by the UV/chlorine process (Figure S6). The increases in the
by the UV/chlorine AOP concentrations of TOCl in the UV/chlorine AOP are consistent with
the results in Zhao et al. (2011), who used benzoic acid and NB as
Fig. 5 shows the formation of TOCl and known DBPs from the IBP the precursors. The large increases in TOCl were hypothesized to be
degradation by the UV/chlorine AOP and dark chlorination at the associated with the formation of intermediates that are more
initial IBP concentration of 50 mM and the chlorine dosage of reactive toward chlorine substitution. Therefore, the intermediates/
200 mM and pH 6. The UV/chlorine AOP generated significantly products and pathways of IBP degradation were also studied.
higher levels of TOCl and known DBPs than dark chlorination. The Besides the halogenated organic products, inorganic byproducts
TOCl formation was 31.6 mM from the UV/chlorine process but only including perchlorate (ClO  
4 ), chlorate (ClO3 ), and chlorite (ClO2 )
0.2 mM from dark chlorination after 30 min. The known DBPs, can also be formed in the UV/chlorine process (Rao et al., 2012). The
including trichloromethane (TCM), 1,1,1-trichloropropanone (1,1,1- formation of ClO  
4 , ClO3 and ClO2 was estimated to be < 0.1 mg/L,
TCP), chloral hydrate (CH), 1,1-dichloro-2-propanone (1,1-DCP), ~200 mg/L and <0.1 mg/L, respectively, under the experimental
dichloroacetic acid (DCAA) and trichloroacetic acid (TCAA), conditions (chlorine dosages ¼ 200 mM, UV fluence ¼ 1764 mJ/cm2),
comprised only 17.4% of TOCl, after adjustment of the number of based on their yields reported in Rao et al. (2012). Thus, the for-
chlorine atoms substituted in each product. The composition of the mation of inorganic byproducts should not be a problem for the UV/
chlorine process under oxidation conditions typically used in
drinking water treatment.
0.0030
UV photolysis
Reactive chlorine species 3.6. Intermediates/products and proposed pathways of IBP
Hydroxyl radicals degradation by the UV/chlorine AOP
0.0025

The mineralization of IBP by the UV/chlorine AOP is lower than

0.0020 7% under the experimental conditions (Figure S7), suggesting that
k0,IBP (s-1)

many carboxyl/chlorine-containing intermediates/products were

0.0015
formed and could not be efficiently mineralized during the degra-
dation of IBP. Fig. 6 shows the time-dependent evolution profile of
the intermediates/products identified by UPLC/ESI-tqMS. The
0.0010 relative abundance of the compounds C118 (m/z 117), C130 (m/z 129),
C172 (m/z 171), C142 (m/z 141) and C152 (m/z 151) increased mono-
tonically with increasing reaction time, and thus they were deemed
0.0005
to be the final products. Other compounds such as C222 (m/z 221),
C238 (m/z 237), C240 (m/z 239), C268 (m/z 267), C258 (m/z 257), C180
0.0000 (m/z 179) and C134 (m/z 133) were identified as intermediates, the
pH=6 pH=7 pH=9 abundance of which initially increased with time before decreasing.
The structures of the intermediates/products are proposed based
Fig. 4. The specific first-order rate constants of IBP degradation by HO , reactive

chlorine species and UV direct photolysis in the UV/chlorine AOP at pH 6e9. Condi-
on mass spectrograms obtained from UPLC/ESI-tqMS and further
tions: [IBP]0 ¼ 10 mM, [NB]0 ¼ 5 mM, [chlorine]0 ¼ 100 mM, UV fluence confirmed by UPLC/ESI-QTOF-MS (Table S2). Isomeric structures
rate ¼ 1.05 mW cm2. are postulated due to the peaks of different retention times with
306 Y. Xiang et al. / Water Research 90 (2016) 301e308

be non-chlorine-containing compounds. The corresponding abun-

m/z 221 dances of m/z 257 and m/z 259 with retention times of 4.16 min,
2.0 m/z 237
m/z 239
5.23 min, 7.15 min were 3:1, suggesting that C258-2 was a mono-
m/z 267 chloro compound with isomers.
m/z 257 The location of chlorine substitution can be estimated by
1.5
m/z 179 applying an optimum collision energy-dependent ESI-tqMS
m/z 133
Intensity ( x e )

method reported in Deng et al. (2014). Figure S10 shows that the
8

m/z 117
optimum collision energy of both C258-2 and C240 in the precursor
m/z 129
m/z 171 ion scan (PIS) of m/z 35 was 15 eV, indicating that the chlorine was
1.0 substituted to hydrogen in the benzene ring instead of in the
m/z 141
m/z 151 aliphatic structure. However, as C258-2 was a mixture of monochloro
compounds, it was possible that isomers existed with chlorine
0.5
substituted in the aliphatic part. The identified final products of IBP
degradation, including carboxylatic acids (C130, C118, C172) and
chlorine containing products (C142, C152), were aliphatic com-
pounds formed through the cleavage of the benzene ring. By using
0.0 the same method described above, C142 and C152 were identified as
0 5 10 15 20 25 30 35 40
Time (min) dichloro and monochloro products, respectively (Figure S11). The
accurate m/z and proposed structures of all the intermediates/
Fig. 6. Intensities of ESI-tqMS full scan of major intermediates/products during IBP products identified are shown in Table S2. TCM, 1,1,1-TCP, CH, 1,1-
degradation by the UV/chlorine AOP as a function of reaction time. Hollow symbols DCP, DCAA and TCAA were continuously formed throughout the
represent intermediates, solid symbols represent products. Conditions: pH 6, IBP degradation.
[IBP]0 ¼ 50 mM, [chlorine]0 ¼ 200 mM, UV fluence rate ¼ 1.47 mW cm2.
The proposed pathways of IBP degradation and the roles of
radicals attack can explain the significantly accelerated formation
the same mass-to-charge ratio in the UPLC/ESI-tqMS analysis. of chlorine-containing byproducts in the UV/chlorine AOP. Hy-

Based on the intermediates and products identified, the pathways droxylation of IBP by HO led to the formation of intermediates (e.g.
of IBP degradation by the UV/chlorine AOP are proposed in Scheme C222-1 and C238-1) that were more reactive toward electrophilic
1. Note that four newly identified chlorine-containing in- attack by chlorine species to form more chlorine-containing in-
termediates were marked with brackets. termediates/products, including C258-2, C142, C152, known DBPs and
Hydroxylation and chlorine substitution were observed at the other unknown TOCl, during IBP degradation by the UV/chlorine
beginning of the IBP degradation. The hydroxylated products, AOP. The chlorine-containing intermediates/products could be
including compounds C222 (m/z 221) and C238 (m/z 237), were parts of the unknown TOCl.
formed at higher levels than the monochloro compound C240 (m/z
239). The possible structures of C222-1,2,3,4 and C238-1,2 are proposed 4. Conclusions
according to the results of UPLC/ESI-QTOF-MS and UPLC/ESI-tqMS,


and the reported products of HO -induced IBP degradation (Li et al., The degradation of IBP by the UV/chlorine and UV/H2O2 AOPs
2014; Madhavan et al., 2010; Me
ndez-Arriaga et al., 2008). Further followed the pseudo first-order kinetics. The first-order rate con-
hydroxylation of C222 and C238 in the ring and on the side chains led stant was 3.3 time higher in the UV/chlorine AOP than in the UV/
to the formation of the multi-hydroxylated compound C268. H2O2 AOP for a given chemical molar dosage at pH 6. The degra-
Decarboxylation and demethylation on the side chains produced dation by the UV/chlorine AOP was sensitive to chlorine dosage,
C258-1 and aromatic intermediates with smaller m/z ratios of C180 alkalinity and solution pH, but not to temperature and the chloride
and C134. C258-1 and C134 have also been reported as degradation concentration tested. Increasing pH from 6 to 9 decreased the first-
intermediates of IBP in photolytic and photocatalytic systems (Da order rate constant from 3.1  103 s1 to 5.5  104 s1 and
Silva et al., 2014). increased the contribution of RCS from 22% to 30%.





To investigate whether HO and/or Cl were responsible for the The degradation was initiated by HO -induced hydroxylation
hydroxylation of IBP in the UV/chlorine AOP, a set of experiments and RCS-induced chlorine substitution, and sustained through


were conducted at pH 0, at which only Cl existed in the system decarboxylation, demethylation, chlorination and ring cleavage to
(Nowell and Hoigne
, 1992). Figure S8 shows the ESI-tqMS full scan form more stable products. The hydroxylation formed in-
spectra of the sample after 55% degradation of IBP by the UV/ termediates that were more reactive toward electrophilic attack by
chlorine AOP at pH 0. The main product was the monochloro chlorine species. The formation of chlorinated intermediates/
compound of C240 (m/z 239), whereas no hydroxylated products byproducts from the UV/chlorine AOP was significant. Four
were observed. This finding suggests that the formation of hy- chlorine-containing intermediates/products was identified. The


droxylated IBP was attributable to HO attack only. yield of total organic chlorine (TOCl) was 31.6 mM after 90%
By using the ESI-tqMS precursor ion scan (PIS) of m/z 35 and m/z degradation of 50 mM IBP under the experimental conditions. The
37 (Figure S9) and comparing the natural abundances of chlorine known DBPs comprised 17.4% of the TOCl.
isotopes (35Cl and 37Cl, ~3:1) in mass spectra (Zhai and Zhang, In summary, the effects of water matrix in filtered drinking
2009), the number of chlorine atoms substituted in a compound water on the degradation were not significant, demonstrating the
can be estimated. When a compound contained isomers, UPLC/ESI- practicality of the UV/chlorine AOP for the control of some re-
tqMS selected ion recording (SIR) and multiple reaction monitoring fractory PPCPs. On the other hand, the formation of DBPs, both
(MRM) modes were used to achieve separation and to estimate the known DBPs and unknown TOCl, was enhanced. The evaluation of
number of chlorine atoms in a compound. Taking C258 as an toxicity of the mixture produced after treating IBP with the UV/
example, by comparing the chromatograms of m/z 257 in the SIR chlorine AOP was out of the scope of this investigation, but the
mode and m/z 257 and m/z 35 in the MRM mode, compounds of m/z significant formation of chlorine-containing intermediates and
257 with retention times of 2.65 min and 3.26 min were deemed to final products was firstly recognized in the current study. Never-
theless, it is a very important information for evaluating the overall
 Y. Xiang et al. / Water Research 90 (2016) 301e308 307

OH
m/z 205 C13H18O2 Cl●
HO●
Main pathway
Hydroxylation

O O O O O
O O OH
HO OH OH OH OH OH
OH OH OH OH
OH OH Cl
OH OH
C222-1, m/z 221 C222-2, m/z 221 C222-3, m/z 221 C222-4, m/z 221 C238-1, m/z 237 C238-2, m/z 237 C240, m/z 239
C13H 18O3 C13H18O 3 C13H18O3 C13H18O3 C13H18O4 C13H18O4 C13H17O2Cl

HO● Hydroxylation
Decarboxylation
Cl● Demethylation
Chlorination
O
HO 2
O
HO OH OH OH
OH HO
OH O
OH OH 3 Cl OH
O OH O O
C268, m/z 267 C258-1, m/z 257 C258-2, m/z 257 C180, m/z 179 C134, m/z 133
C13H 16O6 C12H18O6 C12H15O 4Cl C 9 H 8 O4 C9H10O

HO● Ring opening

Cl● Chlorination

O O O O Cl Cl O
O
OH OH O OH HO Cl OH
OH HO
OH
C118, m/z 117 C130, m/z 129 C172-1, m/z 171 C142-1, m/z 141 C152-1, m/z 151
C5H10O3 C6H10O3 C8H12O4 C3H4O2Cl2 C5H9O3Cl

Cl Cl O O Cl
Cl Cl
Cl Cl HO Cl Cl HO
Cl Cl OH Cl
Cl
Cl OH O O
Cl Cl
TCM CH DCP TCP DCAA TCAA

Scheme 1. Proposed pathways of IBP degradation by the UV/chlorine AOP.

applicability of the UV/chlorine AOP in practice and should be References

further assessed. Limited information is available in the literature
on the overall CHO cell chronic cytotoxicity and CHO cell acute Barakat, M.A., Tseng, J.M., Huang, C.P., 2005. Hydrogen peroxide-assisted photo-
catalytic oxidation of phenolic compounds. Appl. Catal. B Environ. 59 (1e2),
genotoxicity of drinking water disinfected by the combination of 99e104.
UV (300 mJ/cm2) and chlorine (1.9 mg/L) (Plewa et al., 2012). Un- Beitz, T., Bechmann, W., Mitzner, R., 1998. Investigations of reactions of selected
expectedly, the chronic cytotoxicity and acute genotoxicity of water azaarenes with radicals in water. 2. Chlorine and bromine radicals. J. Phys.
Chem. A 102 (34), 6766e6771.
treated by chlorine followed by LPUV were slightly lower than that Bolton, J.R., Linden, K.G., 2003. Standardization of methods for fluence (UV Dose)
of water treated by chlorine alone. The toxicity assessment is determination in bench-scale UV experiments. J. Environ. Eng. 129 (3),
onging and will be reported in the near future. 209e215.


Buxton, G.V., Wood, N.D., Dyster, S., 1988. Ionisation constants of OH and HO2 in
aqueous solution up to 200  C. A pulse radiolysis study. J. Chem. Soc. Faraday
Acknowledgments Trans. 1 84 (4), 1113e1121.
Christen, V., Hickmann, S., Rechenberg, B., Fent, K., 2010. Highly active human
pharmaceuticals in aquatic systems: a concept for their identification based on
This research was supported by the Natural Science Foundation their mode of action. Aquat. Toxicol. 96 (3), 167e181.
of China (grant numbers 51108117, 51378515), the Hong Kong Crittenden, J.C., Hu, S., Hand, D.W., Green, S.A., 1999. A kinetic model for H2O2/UV
Research Grants Council (grant number 16208914), and the Science process in a completely mixed batch reactor. Water Res. 33 (10), 2315e2328.
Da Silva, J.C.C., Teodoro, J.A.R., Afonso, R.J.D.C., Aquino, S.F., Augusti, R., 2014.
& Research Development Program of Foshan (Contract No. Photolysis and photocatalysis of ibuprofen in aqueous medium: characteriza-
2012AA100091). tion of by-products via liquid chromatography coupled to high-resolution mass
spectrometry and assessment of their toxicities against Artemia Salina. J. Mass
Spectrom. 49 (2), 145e153.
Appendix A. Supplementary data Deng, Z., Yang, X., Shang, C., Zhang, X., 2014. Electrospray ionization-tandem mass
spectrometry method for differentiating chlorine substitution in disinfection
Supplementary data related to this article can be found at http:// byproduct formation. Environ. Sci. Technol. 48 (9), 4877e4884.
Fang, J., Fu, Y., Shang, C., 2014. The roles of reactive species in micropollutant
dx.doi.org/10.1016/j.watres.2015.11.069.
308 Y. Xiang et al. / Water Research 90 (2016) 301e308

degradation in the UV/free chlorine system. Environ. Sci. Technol. 48 (3), Differential toxicity of drinking water disinfected with combinations of ultra-
1859e1868. violet radiation and chlorine. Environ. Sci. Technol. 46 (14), 7811e7817.
Feng, Y., Smith, D.W., Bolton, J.R., 2007. Photolysis of aqueous free chlorine species Rao, B., Estrada, N., McGee, S., Mangold, J., Gu, B., Jackson, W.A., 2012. Perchlorate
(HOCl and OCl) with 254 nm ultraviolet light. J. Environ. Eng. Sci. 6 (3), production by photodecomposition of aqueous chlorine solutions. Environ. Sci.
277e284. Technol. 46 (21), 11635e11643.
Garoma, T., Gurol, M.D., 2005. Modeling aqueous ozone/UV process using oxalic Rosenfeldt, E.J., Linden, K.G., 2004. Degradation of endocrine disrupting chemicals
acid as probe chemical. Environ. Sci. Technol. 39 (20), 7964e7969. bisphenol A, ethinyl estradiol, and estradiol during UV photolysis and advanced
Grebel, J.E., Pignatello, J.J., Mitch, W.A., 2010. Effect of halide ions and carbonates on oxidation processes. Environ. Sci. Technol. 38 (20), 5476e5483.
organic contaminant degradation by hydroxyl radical-based advanced oxida- Rosenfeldt, E., Boal, A.K., Springer, J., Stanford, B., Rivera, S., Kashinkunti, R.D.,
tion processes in saline waters. Environ. Sci. Technol. 44 (17), 6822e6828. Metz, D.H., 2013. Comparison of UV-mediated advanced oxidation. J. Am. Water
Herrmann, H., 2007. On the photolysis of simple anions and neutral molecules as Works Assoc. 105 (7), 29e33.
sources of O/OH, SOx and Cl in aqueous solution. PCCP 9 (30), 3935e3964. Ruggeri, G., Ghigo, G., Maurino, V., Minero, C., Vione, D., 2013. Photochemical
Hirakawa, T., Nosaka, Y., 2002. Properties of O$ 2 and OH$ formed in TiO2 aqueous transformation of ibuprofen into harmful 4-isobutylacetophenone: pathways,
suspensions by photocatalytic reaction and the influence of H2O2 and some kinetics, and significance for surface waters. Water Res. 47 (16), 6109e6121.
ions. Langmuir 18 (8), 3247e3254. Shu, Z., Li, C., Belosevic, M., Bolton, J.R., El-Din, M.G., 2014. Application of a solar UV/
Huerta-Fontela, M., Galceran, M.T., Ventura, F., 2011. Occurrence and removal of chlorine advanced oxidation process to oil sands process-affected water
pharmaceuticals and hormones through drinking water treatment. Water Res. remediation. Environ. Sci. Technol. 48 (16), 9692e9701.
45 (3), 1432e1442. Sichel, C., Garcia, C., Andre, K., 2011. Feasibility studies: UV/chlorine advanced
Jin, J., El-Din, M.G., Bolton, J.R., 2011. Assessment of the UV/Chlorine process as an oxidation treatment for the removal of emerging contaminants. Water Res. 45
advanced oxidation process. Water Res. 45 (4), 1890e1896. (19), 6371e6380.
Katsoyiannis, I.A., Canonica, S., von Gunten, U., 2011. Efficiency and energy re- USEPA, 1998. Standard Methods for the Examination of Water and Wastewater,
quirements for the transformation of organic micropollutants by ozone, O3/ 20th ed. American Public Health Assoc, Washington. DC. USA.
H2O2 and UV/H2O2. Water Res. 45 (13), 3811e3822. Vione, D., Maddigapu, P.R., De Laurentiis, E., Minella, M., Pazzi, M., Maurino, V.,
Klassen, N.V., Marchington, D., McGowan, H.C.E., 1994. H2O2 determination by the Minero, C., Kouras, S., Richard, C., 2011. Modelling the photochemical fate of
I
3 method and by KMnO4 titration. Anal. Chem. 66 (18), 2921e2925. ibuprofen in surface waters. Water Res. 45 (20), 6725e6736.
Lee, Y., von Gunten, U., 2010. Oxidative transformation of micropollutants during Vieno, N.M., H€ arkki, H., Tuhkanen, T., Kronberg, L., 2007. Occurrence of pharma-
municipal wastewater treatment: comparison of kinetic aspects of selective ceuticals in river water and their elimination in a pilot-scale drinking water
(chlorine, chlorine dioxide, ferrateVI, and ozone) and non-selective oxidants treatment plant. Environ. Sci. Technol 41 (14), 5077e5084.
(hydroxyl radical). Water Res. 44 (2), 555e566. Wang, D., Bolton, J.R., Hofmann, R., 2012. Medium pressure UV combined with
Li, X., Wang, Y., Yuan, S., Li, Z., Wang, B., Huang, J., Deng, S., Yu, G., 2014. Degradation chlorine advanced oxidation for trichloroethylene destruction in a model water.
of the anti-inflammatory drug ibuprofen by electro-peroxone process. Water Water Res. 46 (15), 4677e4686.
Res. 63, 81e93. Watts, M.J., Linden, K.G., 2007. Chlorine photolysis and subsequent OH radical
Mack, J., Bolton, J.R., 1999. Photochemistry of nitrite and nitrate in aqueous solution: production during UV treatment of chlorinated water. Water Res 41 (13),
a review. J. Photochem. Photobiol. A Chem. 128 (1e3), 1e13. 2871e2878.
Madhavan, J., Grieser, F., Ashokkumar, M., 2010. Combined advanced oxidation Watts, M.J., Rosenfeldt, E.J., Linden, K.G., 2007. Comparative OH radical oxidation
processes for the synergistic degradation of ibuprofen in aqueous environ- using UV-Cl2 and UV-H2O2 processes. J. Water Supply Res. T 56 (8), 469e477.
ments. J. Hazard. Mater 178 (1e3), 202e208. Wols, B.A., Hofman-Caris, C.H.M., Harmsen, D.J.H., Beerendonk, E.F., 2013. Degra-
Me
ndez-Arriaga, F., Esplugas, S., Gime
nez, J., 2008. Photocatalytic degradation of dation of 40 selected pharmaceuticals by UV/H2O2. Water Res. 47 (15),
non-steroidal anti-inflammatory drugs with TiO2 and simulated solar irradia- 5876e5888.
tion. Water Res. 42 (3), 585e594. Yuan, F., Hu, C., Hu, X., Qu, J., Yang, M., 2009. Degradation of selected pharmaceu-
Nowell, L.H., Hoigne
, J., 1992. Photolysis of aqueous chlorine at sunlight and ul- ticals in aqueous solution with UV and UV/H2O2. Water Res. 43 (6), 1766e1774.
traviolet wavelengths-II. Hydroxyl radical production. Water Res. 26 (5), Zhai, H., Zhang, X., 2009. A new method for differentiating adducts of common
599e605. drinking water DBPs from higher molecular weight DBPs in electrospray
Pinkston, K.E., Sedlak, D.L., 2004. Transformation of aromatic ether- and amine- ionization-mass spectrometry analysis. Water Res. 43 (8), 2093e2100.
containing pharmaceuticals during chlorine disinfection. Environ. Sci. Tech- Zhao, Q., Shang, C., Zhang, X., Ding, G., Yang, X., 2011. Formation of halogenated
nol. 38 (14), 4019e4025. organic byproducts during medium-pressure UV and chlorine coexposure of
Plewa, M.J., Wagner, E.D., Metz, D.H., Kashinkunti, R., Jamriska, K.J., Meyer, M., 2012. model compounds, NOM and bromide. Water Res. 45 (19), 6545e6554.