Environmental Research 195 (2021) 110884

Contents lists available at ScienceDirect

Environmental Research
journal homepage: www.elsevier.com/locate/envres

Thermal plasma activation and UV/H2O2 oxidative degradation of

pharmaceutical residues
Martien H.F. Graumans a, *, Wilfred F.L.M. Hoeben b, Maurice F.P. van Dael a, Rob B.M. Anzion a,
Frans G.M. Russel c, Paul T.J. Scheepers a
a
Department for Health Evidence, Radboud Institute for Health Sciences, Radboudumc, Nijmegen, the Netherlands
b
Department of Electrical Energy Systems, Eindhoven University of Technology, Eindhoven, the Netherlands
c
Department of Pharmacology and Toxicology, Radboud Institute for Molecular Life Sciences, Radboudumc, Nijmegen, the Netherlands

A R T I C L E I N F O A B S T R A C T

Keywords: The aquatic environment becomes increasingly contaminated by anthropogenic pollutants such as pharmaceu­
Advanced oxidation processes tical residues. Due to poor biodegradation and continuous discharge of persistent compounds in sewage water
Medicine residues samples, pharmaceutical residues might end up in surface waters when not removed. To minimize this pollution,
Complex water matrices
onsite wastewater treatment techniques might complement conventional waste water treatment plants (WWTPs).
Hospital sewage water
Advanced oxidation processes are useful techniques, since reactive oxygen species (ROS) are used for the
degradation of unwanted medicine residues. In this paper we have studied the advanced oxidation in a controlled
laboratory setting using thermal plasma and UV/H2O2 treatment. Five different matrices, Milli-Q water, tap
water, synthetic urine, diluted urine and synthetic sewage water were spiked with 14 pharmaceuticals with a
concentration of 5 μg/L. All compounds were reduced or completely decomposed by both 150 W thermal plasma
and UV/H2O2 treatment. Additionally, also hospital sewage water was tested. First the concentrations of 10
pharmaceutical residues were determined by liquid chromatography mass spectrometry (LC-MS/MS). The
pharmaceutical concentration ranged from 0.08 up to 2400 μg/L. With the application of 150 W thermal plasma
or UV/H2O2, it was found that overall pharmaceutical degradation in hospital sewage water were nearly
equivalent to the results obtained in the synthetic sewage water. However, based on the chemical abatement
kinetics it was demonstrated that the degree of degradation decreases with increasing matrix complexity. Since
reactive oxygen and nitrogen species (RONS) are continuously produced, thermal plasma treatment has the
advantage over UV/H2O2 treatment.

1. Introduction other wastewater ingredients such as soap or other detergents (Beier

et al., 2010; Ferrando-Climent et al., 2014; Peake et al., 2016; Ajo et al.,
Aquatic ecosystems are increasingly polluted with anthropogenic 2018). Excreted pharmaceuticals may be metabolized to more bioactive
compounds such as industrial waste, personal care products, micro­ compounds than the administered parent drug. Conventional waste­
plastics, pesticides and pharmaceutical residues. Pharmaceutical pollu­ water treatment plants (WWTPs) cannot completely remove all phar­
tion has received more attention in recent years, due to their effect on maceutical residues. Due to continuous discharge and various
the aquatic ecosystem and potential risk for the drinking water quality wastewater treatment technologies, the degradation efficiency for each
(Peake et al., 2016; Petrie et al., 2016; Banaschik et al., 2018). Improper individual compound will differ from region to region (Peake et al.,
disposal, runoff from manure and the continuous excretion from users 2016). Certain WWTPs do convert glucuronide metabolites back into
cause the introduction of pharmaceuticals into the hydrosphere their pristine structure by enzymatic processes (Ajo et al., 2018; Li et al.,
(Magureanu et al., 2015; Souza and Feris 2016; Li et al., 2019). Com­ 2019). Antiepileptics, contrast agents, analgesics, antibiotics and cyto­
plete or partially metabolized pharmaceuticals are discharged via do­ static drugs are pharmaceutical classes often identified in raw hospital
mestic and hospital effluents. These effluents are often complex effluents (Beier et al., 2010; Ferrando-Climent et al., 2014; Ajo et al.,
chemical mixtures where pharmaceuticals and metabolites mix with 2018). To complement and ease conventional WWTPs, onsite

* Corresponding author. P.O. Box 9101, 6500 HB, Nijmegen, (133) Geert Grooteplein-Noord 21, the Netherlands.
E-mail address: [email protected] (M.H.F. Graumans).

https://doi.org/10.1016/j.envres.2021.110884
Received 23 December 2020; Received in revised form 8 February 2021; Accepted 10 February 2021
Available online 22 February 2021
0013-9351/© 2021 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
M.H.F. Graumans et al. Environmental Research 195 (2021) 110884

wastewater treatment techniques are proposed (Gerrity et al., 2010; previous studies high degradation efficiencies were observed for phar­
Banaschik et al., 2015; Ajo et al., 2018). Different techniques such as maceuticals treated with non-thermal plasma (Banaschik et al., 2018) or
coagulation-flocculation (Suarez et al., 2009), reverse osmosis (Beier the combination of non-thermal plasma with iron as a catalyst
et al., 2010) or advanced oxidation processes (AOPs) (Kohler et al., (Marković et al., 2015). Plasma-driven activation of water is a new AOP
2012; Ajo et al., 2018) are studied on pilot-scale to treat hospital that uses a hot arc (thermal plasma) in air over water to increase both
wastewater. However, further development of a fully operational on-site reactive oxygen and nitrogen species (RONS) in water (Hoeben et al.,
treatment technique is often economically too expensive for hospitals to 2019; Graumans et al., 2020). In a previous optimization
realize (Banaschik et al., 2015; Ajo et al., 2018). AOPs covers the prin­ laboratory-scale experiment we demonstrated that a thermal arc dis­
ciples of using ROS for the degradation of contaminants by ozone, charged in air over water is capable of the degradation of cyclophos­
hydrogen peroxide, peroxone (O3/H2O2) Fenton reaction (Fe2+/H2O2), phamide in tap water (Graumans et al., 2020). The aim of the present
UV, photocatalytic, supercritical and non-thermal gas discharges based study was to determine the applicability of thermal plasma to complex
oxidative degradation technology (Banaschik et al., 2015; Magureanu matrices. Pharmaceuticals were added to different aquatic matrices, see
et al., 2015; Marković et al., 2015). According to Ajo et al., (2018), Fig. 1. Subsequently a generic extraction technique was developed to
plasma treatment is the most efficient technique to produce ROS. In determine 14 selected compounds. This method was used to analyze

Fig. 1. The 14 selected pharmaceutical compounds, including their classifications and physicochemical properties.

2
M.H.F. Graumans et al. Environmental Research 195 (2021) 110884

pharmaceutical residues in synthetic and real wastewater matrices such container with an appropriate aperture (12 cm, wide) was placed over
as hospital and domestic sewage water. After analytical determination the glass beaker. The UV-C irradiance at the aqueous matrix surface was
and characterization of the pharmaceutical residues the hospital effluent extended and derived from our previous oxidation study (Graumans
was also plasma-treated, to determine the efficacy on real wastewater et al., 2020), (SI paragraph S1).
samples. Due to the wide variety in physicochemical properties of the
selected pharmaceuticals, in addition to solid phase extraction, a novel 2.3. Simulated aqueous matrices
liquid-liquid extraction method for the polar component metformin was
developed. Thermal plasma and UV-C/H2O2 oxidation was applied in Milli-Q,
tap water, synthetic urine, diluted urine and synthetic sewage water
2. Materials and methods matrix. All tested aqueous solutions were freshly prepared at the start of
the experiment. Milli-Q water was individually tested but also used for
2.1. Chemicals and reagents the preparation of synthetic urine and synthetic sewage water. A healthy
male volunteer consented to provide urine. This urine was diluted to a
Pharmaceutical standards, iopamidol (IOP), diatrizoic acid (DIA), 20% v/v ratio, in Milli-Q water, to simulate sewer dilution of flushed
fluoxetine (FLU), diclofenac (DF), metoprolol (MET), carbamazepine urine. Physicochemical characteristics of the tap water matrix were
(CB), terbutaline (TER), phenazone (PHE), acetaminophen (APAP), adopted from the freely available January–December 2019 quality
ciprofloxacin (CIP), doxycycline (DOX) and metformin (MF), all with standard report (Vitens, Arnhem). An average pH of 7.73 was reported,
>98% purity were retrieved from either Sigma Aldrich/Fluka Analytical with a total organic carbon (TOC) concentration of <0.5 mg/L and
(Zwijndrecht, the Netherlands) or Merck Group (Darmstadt, Germany). turbidity of 0.24 NTU, detailed information is given in the supplemental
Endoxan and Iomeron were obtained from Baxter (Utrecht, the data (Table S3). To standardize the advanced oxidation processes, both
Netherlands) and Bracco (Konstanz, Germany) respectively used to techniques were used in 500 mL aqueous matrix and during treatment
prepare a stock solution for cyclophosphamide (CP) and iomeprol magnetically stirred. The initial pharmaceutical concentration (n = 14)
(IOM). Mobile phase and buffer solutions were freshly prepared in ul­ was 5 μg/L. The oxidative pharmaceutical degradation was determined
trapure water using a Milli-Q Academic A10 system with a resistivity of by collecting samples at time intervals of 5, 10, 15, 30, 45, 60, 90 and
18 MΩ cm (Millipore, Amsterdam the Netherlands). Pharmaceutical 120 min. Prior to and after oxidative treatment the temperature and pH
stock solutions, with a concentration of 1.0 mg/mL were prepared in were measured. Direct LC-MS/MS injection was applied on the matrices,
50:50 (v/v) methanol water. Single aliquots of 10.0 μg/mL were sub­ Milli-Q water, tap water and synthetic urine. Sample pre-treatment was
sequently prepared in Milli-Q water and further diluted together to have applied on diluted urine, synthetic sewage water and actual wastewater
a pharmaceutical mixture (n = 14) with a concentration of 100.0 μg/L. samples prior to LC-MS/MS analysis. Centrifugal filtration was used for
Prior to analysis and solid phase extraction calibration standards were the urine samples, where solid phase extraction and liquid-liquid
freshly prepared, ranging from 0.5 up to 100.0 μg/L (n = 6). Deuterated extraction were used to pretreat wastewater matrixes.
internal standards (dIS) were all obtained from Toronto Research
Chemicals (North York, Canada) and prepared as single stock solutions 2.3.1. Synthetic urine
(1.0 mg/mL) in methanol (SI paragraph S4). Compounds needed for the For the preparation of synthetic urine, all needed analytes were
preparation of buffers and simulation matrices included NH4HCO2, dissolved separately in 100 mL Milli-Q, to minimize the formation of
NaCl, Na2SO4, KCl, KH2PO4, CaCl2 ⋅ 2H2O, NH4Cl, K2HPO4, MgSO4, precipitation or undissolved constituents. The final product comprises;
sodium dodecyl sulfate, creatine, urea, peptone and meat extract. These 25.0 g urea, 2.9 g NaCl, 2.3 g Na2SO4, 1.6 g KCl, 1.4 g KH2PO4, 1.1 g
chemicals were either acquired from Sigma Aldrich/Fluka Analytical CaCl2 ⋅ 2H2O, 1.1 g creatinine and 1.0 g NH4Cl in 1000 mL Milli-Q,
(Zwijndrecht, the Netherlands) or Merck Group (Darmstadt, Germany). simulating human urine composition (Giannakis et al., 2018).
UPLC-MS analytical grade chemicals formic acid (FA), methanol
(MeOH) and acetonitrile (ACN) were purchased at Merck Group 2.3.2. Synthetic sewage
(Darmstadt, Germany) and Boom B.V. (Meppel, the Netherlands) To simulate the composition of sewage water, a standardized pro­
respectively. tocol from the Organization for Economic Co-operation and Develop­
ment (OECD) was used. Similar to the preparation of artificial urine, all
2.2. Advanced oxidation processes analytes were prepared separately in 100 mL of Milli-Q. The final syn­
thetic sewage water solution contained; 160.0 mg Peptone, 110.0 mg
2.2.1. Plasma-driven water activation Meat Extract, 30.0 mg Urea, 28.0 mg K2HPO4, 7.0 mg NaCl, 3.0 mg
Thermal plasma discharge was performed using a laboratory-scale CaCl2 ⋅ 2H2O and 0.5 mg MgSO4 in 1000 mL Milli-Q water (OECD,
plasma activation unit from VitalFluid (Veldhoven, the Netherlands), 2001).
comprising a 150 W, 1 MHz dual resonant power modulator with 8 kV
arc operation voltage, creating a hot arc in air over water (Pemen et al., 2.4. Sewage water sample collection
2017; Hoeben et al., 2019). A thermal electric discharge with voltages
up to 40 kV was generated in air over water. Thermal arc operation was End-of-pipe hospital sewage water (HSW) samples were periodically
performed in ambient air with a power output of 150 Watt (W). To collected for one week in September 2019 (composite sample). The HSW
maintain the temperature of the water during activation between 20 and samples were collected in two separate jerry cans to obtain a sample
35 ◦ C, the plasma unit was equipped with an active Peltier cooler. including the weekend (Thursday – Monday 12–Sep 16, 2019) compared
to working days sample (Monday – Thursday 16–Sep 19, 2019). In
2.2.2. UV-C/H2O2 addition to compare the pharmaceutical concentration present in HSW,
Stationary UV-C irradiation was performed by using a modified also domestic sewage water (DSW) was sampled on Oct 31, 2019 (as a
Aetaire UV air disinfection unit. To irradiate UV-C (Philips PL-L 60 W/4 grab sample). These samples were taken from a residential area with no
P HO UV-C lamp, Ekkersrijt, the Netherlands) at 253.7 nm, an opening supply of HSW. Upon arrival in the laboratory the wastewater samples
of 28 cm × 7 cm was used. The distance between the lamp and aqueous were directly centrifuged and vacuum-filtered using a Whatman Filter
matrix surface was approximately 20 cm. At the start of the advanced grade 1 (Maidstone, United Kingdom). The filtered wastewater samples
oxidation process, 0.22 mM (~10 mg/L) hydrogen peroxide (H2O2) were distributed in separate clean plastic aliquots, labelled and stored in
from J.T. Baker, Avantor Performance Materials (Deventer, the the fridge (4 ◦ C) or freezer (− 20 ◦ C). Samples stored in the fridge were
Netherlands) was added. To minimize UV scattering and reflection, a analyzed within 72 h to maintain the approximate pharmaceutical shelf

3
M.H.F. Graumans et al. Environmental Research 195 (2021) 110884

life. used as collision gas during MS/MS. The voltages selected for the
capillary and cone were 2 kV and 20 V, respectively. Multiple reaction
2.5. Centrifugal filtration monitoring (MRM) transitions for all pharmaceuticals were determined
by direct infusion to the MS/MS. These transitions are presented as a
To analyze the pharmaceutical oxidative degradation in diluted precursor ion [M+H]+ combined with the corresponding fragmented
urine matrix, 0.5 mL urine sample was placed into a Spin-X centrifugal daughter ion. The compounds were all individually infused via the flu­
tube from Costar (Salt Lake City, USA). This Spin-X tube is equipped idics interface by using a flow rate of 10 μL/min. The optimized results
with a 0.22 μm cellulose acetate filter and placed on top of a 2.0 mL tube. are given in Table S4 (SI paragraph S4).
To filter the urine sample, an Eppendorf MiniSpin (Nijmegen, the The LC unit consists of a solvent degasser, quaternary solvent man­
Netherlands) centrifuge is used for 5 min at maximum speed (134000 ager, seal wash pump, column oven and auto-sampler. Gradient LC
rpm). After spinning, a clear urine sample was transferred to a LC-MS/ conditions were applied at a flow of 0.5 mL/min, using 2.5 μL of in­
MS vial for analysis. jection volume and a column oven temperature set at 40 ◦ C. Eluent (A)
consisted of 0.1% v/v formic acid (FA, 98–100%) in Milli-Q water in
2.6. Solid phase extraction combination with eluent (B) 100% v/v acetonitrile (ACN, UHPLC-MS).
An Acquity UPLC BEH C18 (2.1 mm × 100 mm, 1.7 μm) reversed
Acidic buffer solution (ABS) was prepared by dissolving 10 mM of phase column was used for the chromatographic separation. The LC
ammonium formate (NH4HCO2, ≥99% HPLC grade salt) and 3.25 mL of gradient program was started with 100% eluent A until 0.20 min, and
formic acid (FA, 98–100%) in Milli-Q to obtain a pH of ~2.2. Oasis HLB quickly changed to 100% B within 2.0 min. This 100% B was maintained
3 cc (cm3) solid phase extraction (SPE) cartridges with 60 mg of solid until 4.3 min and subsequently ramped to 100% A at 4.5 min. The
sorbent from Waters Corporation (Milford, USA), were selected for the gradient program was stopped at 8 min.
extraction. SPE cartridges were conditioned with 2.0 mL of methanol
and 2.0 mL of ABS. Prior to analysis, either 1.0 mL of analyte mixture, 2.9. Data analysis
including 100 μL dIS, HSW or DSW was diluted in 4 mL of ABS. Vacuum
extraction was performed using an SPE manifold from Analytichem in­ For all pharmaceutical compounds the method selectivity was ob­
ternational (Bay Meadows Ln, USA). An aliquot of 5.0 mL aqueous tained by using two MRM transitions per analyte. For each deuterated
pharmaceutical matrix was loaded on the SPE cartridge and subse­ internal standard only one MRM transition was selected. Linear curve
quently washed with 1.0 mL ABS to remove impurities. After sample fitting (y = ax + b) was applied using TargetLynx LC-MS/MS data
loading, the retained pharmaceuticals were extracted using 5.0 mL of acquisition software (Waters Corporation Milford, USA). Data quantifi­
methanol. Next, the extracts were vaporized under a gentle stream of cation for metformin (MF) was done by using a quadratic curve fitting (y
nitrogen using a sample concentrator (Stuart, Stone, United Kingdom). = ax2 + bx + c). To minimize the y-value error of the low analyte
at 40 ◦ C. The vaporized samples were reconstituted in mobile phase A concentrations, all calibration points were equally weighted by using a
(0.1% formic acid) prior to LC-MS/MS analysis. To develop an extrac­ weighting factor 1/X (Almeida et al., 2002; Gu et al., 2014). The phar­
tion method for pharmaceuticals in a complex matrix in a broad con­ maceutical oxidative degradation kinetics were plotted using GraphPad
centration range, the recovery of the selected pharmaceuticals was Prism 5.03 statistical software (Graphpad Inc., CA, USA) and the con­
determined in Milli-Q and synthetic sewage water by using three version level was determined using eq. (1) (Graumans et al., 2020). SPE
controlled concentrations, 1.0, 10.0 and 50.0 μg/L. recovery data was determined by dividing the extract concentration
through the original concentration, see eq. (2). A two-sided t-test was
2.7. Liquid-liquid extraction of metformin applied to test for statistical significance of degradation over time,
assuming first-order kinetics Eq. (3) (Miller and Miller 2010).
Due to the high hydrophilicity of metformin, no satisfactory recovery
R = (1 − Ct / C0 )⋅100 % (1)
was obtained with SPE. To improve the recovery for metformin, a liquid-
liquid extraction with acetonitrile at subzero temperatures was per­
formed according to a previously described method (Yoshida and Akane R‾ = Conversion level (%) (n = 4)
1999). An aliquot of 0.5 mL of a solution with a known metformin Ct = Concentration of the test substance at the end of treatment (μg/
concentration was extracted by adding 0.1 mL of 2.0 mM sodium L)
dodecyl sulfate (SDS) including 0.5 mL of acetonitrile. Sample aliquots C0 = Initial compound concentration (μg/L)
were vortexed and centrifuged at maximum speed for 5 min. By placing
Concentration (Extract)
the tubes for 30 min. at − 20 ◦ C, the aqueous phase was separated from Recovery (S%) = ⋅100 % (2)
Concentration (original)
the acetonitrile phase. Compounds with affinity for the organic layer
were retrieved from the acetonitrile phase. The acetonitrile layer was √̅̅̅̅̅̅̅̅̅̅̅
r⋅ n − 2
directly transferred to an LC-MS/MS vial for analysis. For the extraction t = √̅̅̅̅̅̅̅̅̅̅̅̅̅ (3)
from actual sewage water, the dimensions were increased by a tenfold, 1 − r2
using the ratio 5:1:5 mL (Sewage Water: SDS: Acetonitrile).
t = t-calculated
2.8. Liquid chromatography tandem mass spectrometry analysis r2 = Correlation coefficient of the log-normal transformed line

An Acquity UPLC system from Waters Corporation (Milford, USA) 3. Results and discussion
was used for the analytical separation of pharmaceuticals and deuter­
ated internal standards. The LC system was online coupled to a Xevo TQ- 3.1. Pharmaceutical degradation using thermal plasma activation
S micro quadrupole mass spectrometer, with electro spray ionization
(ESI), operated in positive ion mode for all selected compounds. Nitro­ The oxidative conversion levels of 14 analytes, spiked in five
gen gas was applied at 50 L/h in the cone and at 1100 L/h for des­ different matrices, were evaluated in a controlled setting. Selected
olvation. The desolvation temperature was set at 600 ◦ C, and argon was pharmaceuticals at given initial concentrations were degraded by

4
M.H.F. Graumans et al. Environmental Research 195 (2021) 110884

Table 1
Pharmaceutical removal by thermal plasma treatment in different matrices
Comp. Milli-Q Water Tap Water Synthetic Urine Urine Synthetic Sewage
a - - - -
R‾ k (min t1/2 R‾ k(min t1/2 R‾ k(min t1/2 R‾ k(min t1/2 R‾ k(min- t1/2
1 1 1 1 1
(%) ) (min) (%) ) (min) (%) ) (min) (%) ) (min) (%) ) (min)
Ct(min) Ct(min) Ct(min) Ct(min) Ct(min)

IOM 66.9 0.009 77.0 62.2 0.007 99.0 43.1 0.005 138.6 25.1 0.002 346.6 56.3 0.008 86.6
120 0.96 r2 120 0.84 r2 120 0.93 r2 120 0.73 r2 120 0.93 r2
IOP 63.0 0.008 86.6 47.1 0.005 138.6 0.0 - - 35.5 0.003 231.1 33.1 0.007 99.0
120 0.99 r2 120 0.89 r2 - - 120 0.48 r2b 120 0.75 r2
DIA 41.3 0.004 173.3 31.8 0.003 231.1 30.7 0.003 231.1 5.9 <0.001 >693.2 15.3 0.002 346.6
120 0.97 r2 120 0.89 r2 120 0.91 r2 120 0.02 r2b 120 0.51 r2
CIP 99.4 0.030 23.1 72.2 0.006 115.5 68.6 0.006 115.2 81.6 0.015 46.2 69.6 0.013 53.3
120 0.67 r2 120 0.29 120 0.64 r2 120 0.97 r2 120 0.97 r2
r2b
DOX 100.0 0.032 21.7 100.0 0.030 23.1 100.0 0.041 16.9 81.9 0.013 53.3 100.0 0.059 11.8
90 0.85 r2 120 0.83 r2 90 0.83 r2 120 0.84 r2 90 0.96 r2
FLU 82.8 0.017 40.8 81.2 0.014 49.5 70.2 0.010 69.3 67.6 0.010 69.3 82.5 0.016 43.3
120 0.98 r2 120 0.97 r2 120 0.99 r2 120 0.98 r2 120 0.96 r2
DF 100.0 0.520 1.3 100.0 0.295 2.4 100.0 0.108 6.4 100.0 0.097 7.1 99.1 0.026 26.7
15 1.00 r2 15 1.00 r2 15 1.00 r2 45 0.96 r2 120 0.58 r2
MET 72.3 0.011 63.0 61.7 0.008 86.6 49.6 0.005 138.6 9.6 <0.001 >693.2 64.7 0.009 77.0
120 0.99 r2 120 0.98 r2 120 0.98 r2 120 0.11 r2 120 0.96 r2
CP 100.0 0.160 4.3 100.0 0.185 3.8 100.0 0.078 8.9 100.0 0.049 14.2 100.0 0.121 5.7
120 0.98 r2 60 0.91 r2 60 0.97 r2 90 0.82 r2 90 0.97 r2
CB 90.5 0.019 36.5 84.4 0.019 36.5 62.3 0.008 86.6 37.9 0.004 173.3 79.4 0.014 49.5
120 0.99 r2 120 0.99 r2 120 0.98 r2 120 0.93 r2 120 0.97 r2
TER 92.7 0.021 33.0 95.2 0.046 15.1 89.5 0.018 38.5 91.9 0.021 33.0 85.3 0.016 43.3
120 0.98 r2 120 0.95 r2 120 0.95 r2 120 0.99 r2 120 0.98 r2
PHE 100.0 0.149 4.7 100.0 0.192 3.6 100.0 0.207 3.4 100.0 0.059 11.8 100.0 0.263 2.6
30 0.70 r2 30 0.75 r2 30 0.98 r2 90 0.91 r2 45 0.97 r2
APAP 100.0 0.774 0.9 100.0 0.115 6.0 100.0 0.271 2.6 100.0 0.065 10.7 100.0 0.739 0.94
15 1.00 r2 30 1.00 r2 45 0.97 r2 60 0.98 r2 30 1.00 r2
MF 17.6 0.001 693.2 42.8 0.004 179.3 0.0 - - 29.3 0.003 231.1 72.9 0.011 63.0
120 0.94 r2 120 0.74 r2 - - 120 0.82 r2 120 0.71 r2
a
First-order reaction rate equation: -d[A]/dt. k, the first-order rate constant, is the slope of the line. Where [A] is the drug concentration and t(1/2) ¼ ln 2/k
b
No significant change due to moderate to low correlation coefficient in linear modelling of assumed first-order kinetics (r2 < 0.50).

generated RONS. A time-based decomposition plot demonstrates that

the chemical degradation kinetics are distinct for all analytes tested, see
Fig. 2. DOX, DF, CP, PHE and APAP were all rapidly 100% decomposed,
with similar abatement kinetics. The compounds CIP, FLU, CB and TER
were converted for at least >90% within 120 min. The slowest degra­
dation was observed for MET (72.3%), IOM (66.9%), IOP (63.0%), DIA
(41.3%) and MF (17.6%) after 120 min treatment. Although the chem­
ical degradation kinetics differ between compounds and matrices, dur­
ing thermal plasma treatment the analytes were observed to be
subjected to exponential decay. Log-normal transformation of this con­
centration data resulted in linear time-based decomposition plots, see
Fig. S5 (SI paragraph S5). The slope of this linear decay represents the
first-order rate constant (k) where r2 > 0.50 demonstrates significant
correlation for first-order degradation kinetics (SI paragraph S5).
Comparison of the pharmaceutical degradation rate constants in
Milli-Q with other matrices (Table 1 and Fig. S3 and 4 (SI paragraph
Fig. 2. The time-based decomposition plot for the pharmaceuticals IOM ( ), S5)), it is demonstrated that the k value decreases with increasing matrix
IOP ( ), DIA ( ), CIP( ), DOX ( ), FLU( ), DF ( ), MET ( ), CP( ), CB( ), complexity. To demonstrate the influence of co-existing substances
TER ( ), PHE( ), APAP ( ) and MF( ), treated in Milli-Q water with 150 W
present in these treated matrices, the overall rate constant (k‾) was used
thermal plasma.
to calculate the inhibition ratio, see eq. (4) (Tokumura et al., 2016). A
ratio higher than 1.000 suggests inhibition by constituents and organic
matter in the sample matrix indicating that the aqueous composition is
of major influence on plasma oxidation efficiency. The matrix
oxidation to an energy density and compound dependent extent using complexity and interference of co-existing substances was consecutively
150 W thermal plasma treatment, (Table 1). Milli-Q was used as ultra- demonstrated to be:
pure water matrix where all molecules are solely oxidized by plasma-

1.000(Milli− Q) < 1.344(Synthetic ​ Sewage) < 1.893(Tap ​ water) < 2.119(Synthetic ​ urine) < 5.000(Urine) .

5
M.H.F. Graumans et al. Environmental Research 195 (2021) 110884

and CB (75%) within 70 min. We observed decomposition levels of

41.3%, 100% and 90.5% within 120 min for DIA, DF and CB, respec­
k(Milli− Q)
Inhibition ​ ratio = (4) tively. Although the oxidative degradation of these pharmaceuticals
k(Matrix)
shows similarities, it must be noted that the plasma degradation in our
It is demonstrated that the aqueous composition is of great influence study is not solely based on hydroxyl radical interaction. With plasma
on the plasma oxidation efficiency. Despite the overall inhibitory effect discharge in ambient air and striking over water, also reactive nitrogen
of co-existing substances in various tested media, synthetic sewage and species are transferred in the liquid matrix (Hoeben et al., 2019; Grau­
tap water, compared to Milli-Q, showed improved degradation rate mans et al., 2020), such as nitric oxide (NO), nitrogen dioxide (NO2),
constants for DOX, CP, TER, PHE and MF. The relatively small nitrous acid (HNO2), nitric acid (HNO3) and peroxynitrous acid
improvement in degradation for these compounds is attributed to matrix (ONOOH). In addition to oxidative degradation, these RNS can also
conductivity. Enhanced conductivity will increase the ability of facili­ induce nitration and nitrosation reactions (Magureanu et al., 2018;
tating electrical current through an aqueous medium, suggesting that Hoeben et al., 2019). Looking at the currently tabulated conversion
the pre-existence of ions in tap water and synthetic sewage accelerate levels (R‾) and corresponding half-lives (t1/2), (Table 1), it is clear that
the transfer of specific RONS (Brisset et al., 2011; Thirumdas et al., certain molecular structures are more suspectable to plasma produced
2018; Hoeben et al., 2019; Shimizu et al., 2020). Increased solution RONS than others. Aromatic ring systems, unsaturated double bonds (C
conductivity also increases the current of the thermal plasma arc, pro­ = C) and functional groups with electron donating properties have
moting the chemical activity of the plasma discharge in the gas phase. A increased reactivity towards electrophilic oxidizing agents (HO• or NO+ 2)
similar effect would have been expected in synthetic and diluted urine (Brown, 2007; Joshi and Thagard 2013; Banaschik et al., 2018;
matrices, however, an inhibitory effect was observed. This effect is Magureanu et al., 2018). The reactivity of aromatic ring systems is
probably explained by the high concentration of the co-existing minerals however, not self-evident since an electrophile, such as a hydroxyl
and organic constituents present in g/L ranges compared to mg/L range radical, is needed to interact (Brown, 2007).
in tap water and synthetic sewage water (OECD, 2001; Giannakis et al., A technology such as plasma-induced water activation is used to
2018). Shih and Locke (2011) demonstrated in their study an optimal generate electrophilic compounds, demonstrating that HO• radicals
conductivity of 150 μS/cm for the production of reactive species. readily initiate an addition reaction by binding to an aromatic ring
Further increase of conductivity (up to 500 μS/cm) resulted in a decline system, Fig. 3A. (Banaschik et al., 2018). A similar reaction pathway is
of radicals and molecular species formed. Due to the presence of expected during our thermal plasma treatment process, but are com­
co-existing substances there is less direct contact with water molecules plemented by nitration processes (Magureanu et al., 2018; Hoeben et al.,
to produce hydroxyl radicals (R1). On the other hand minerals and 2019; Graumans et al., 2020). With prolonged thermal plasma activa­
organic constituents may cause quenching or the production of less tion, the aqueous matrix pH is decreased by RONS (Table 2). Nitration
reactive radicals (Shih and Locke 2011; Giannakis et al., 2018; Shimizu reactions usually occur at acidic pH (<6), where a nitronium ion (NO+ 2)
et al., 2020). Slower and no gradual abatement kinetics was therefore is substituted or added as an electrophile at the aromatic ring system
also observed in urine diltued with water, showing a poor correlation of (Squadrito and Pryor 1998; Brown, 2007). Chiron et al., (2010), iden­
r2 < 0.50 for the degradation of IOP (0.48 r2), DIA (0.02 r2) and MET tified 3-nitro-APAP as the nitration transformation product of APAP in
(0.11 r2). WWTPs (Chiron et al., 2010), see Fig. 3B. Nitration reactions in WWTP
are initiated by nitrifying bacteria who produce nitric oxide (NO•). NO•
is a precursor for nitrating agents such as peroxynitrite (ONOO− ). In
3.2. Plasma oxidation chemistry acidic samples ONOOH is very unstable and will rapidly decompose into
HO•, NO2 and HNO3 (Chiron et al., 2010; Jewell et al., 2014; Hoeben
Many factors can influence thermal plasma degradation efficiency. In et al., 2019). Additionally, ONOO− can also react with CO2 yielding
our previous study we observed that electrical power input and initial NO2• and CO3•- radicals (Chiron et al., 2010). For thermal plasma acti­
analyte concentration were important parameters (Graumans et al., vation it is expected that similar RONS are produced (R3-6). However, it
2020). Additionally, it was demonstrated that by continued activation of is expected that the combined attack of hydroxyl radical and molecular
a relatively large volume of tap water (500 mL) and low pharmaceutical oxygen is the main degradation pathway in this thermal plasma study
concentration (4 μg/L), a logistic oxidative degradation pattern for CP and that nitration oxidation will complement the oxidative process
elapsed. Comparing the previously non-cooled plasma results with the under favorable circumstances (Chiron et al., 2010; Jewell et al., 2014;
current data, no logistic oxidative degradation pattern is found for CP Thirumdas et al., 2018; Magureanu et al., 2018; Hoeben et al., 2019).
and other treated analytes, see Fig. S6 (SI paragraph S6). This
remarkable dissimilarity is attributed to the temperature of the plasma
treated solution. Active cooling has a positive effect on the production of
thermally labile species such as ozone, hydrogen peroxide and transient
reactive nitrogen species, while high temperatures benefit nitric acid
formation (Hoeben et al., 2019). Hydroxyl radicals (HO•) and their
coupling product hydrogen peroxide (H2O2) are formed by
plasma-water contact, either in the gas phase or at the plasma-water
interface (R1) (Joshi and Thagard 2013). Two HO• radicals combine
to form measurable hydrogen peroxide concentrations (R2) (Magureanu
et al., 2018). On the other hand, in presence of organic molecules, such
as pharmaceuticals HO• radicals will react by initiating oxidative
degradation (Joshi and Thagard 2013; Magureanu et al., 2018).
Fig. 3. During thermal plasma treatment distinct chemical processes will occur
(R1) according to the produced RONS. Several RONS will initiate chemical processes
(R2) on reactive molecular structures containing aromatic rings, unsaturated bonds
or electron donating function groups. HO• radical attack is the favored degra­
Pharmaceutical decomposition initiated by HO• radicals produced dation pathway, reacting for example rapidly with the aromatic ring of DF by
with immersed non-thermal pulsed corona plasma, was demonstrated forming 4-hydroxy-DF (A). Additionally, nitration processes will also occur
before (Banaschik et al., 2018). Oxidative conversion levels for three where for example NO2• radicals can add to aromatic ring of APAP by pro­
identical compounds in Milli-Q were found for DIA (40%), DF (100%) ducing 3-nitro-APAP (B).

6
M.H.F. Graumans et al. Environmental Research 195 (2021) 110884

Table 2 molecules is relatively low (Bourin et al., 1997). DIA is weak acidic
Average pH values of n ¼ 2 thermal plasma treated matrices. compound, where IOM and IOP are both nonionized over a wide pH
150W Milli-Q Tap Synthetic Urine Synthetic range, see Table S5 (SI paragraph S7).
Plasma Water Water Urine pH x‾ Sewage The fully substituted aromatic ring with iodine atoms is character­
pH x‾ pH x‾ pH x‾ pH x‾ istic for contrast agents, since these structures facilitate X-ray absorption
Start 0 min 5.0 7.7 6.5 7.4 6.9 during clinical diagnosis (Bourin et al., 1997; Borowska et al., 2015).
End120 1.6 1.6 2.2 2.5 1.7 That these compounds are difficult to decompose during plasma
min oxidation is explained by the fully substituted aromatic ring, where the
iodine functional groups hinder the availability of the unsaturated C-
Our previous study on degradation data of CP without active cooling atoms (Rosati 1994; Bourin et al., 1997; Zhao et al., 2014; Banaschik
indicated pharmaceutical dissolution after prolonged activation (Grau­ et al., 2018). Assuming that hydroxyl radical attack is the major
mans et al., 2019). This suggests that minimal HO• radicals were formed, pathway during thermal plasma degradation study, it confirms that slow
but that molecular dissolution was catalyzed by acidic pH (<6) and molecular invasion can occur (Banaschik et al., 2018). On the other
increasing water temperature, both acting as stimulants for the nitration hand it is noteworthy, that the hydroxyl radical can attack on the ipso
reaction rate (Chiron et al., 2010). positions of the aromatic ring, as observed with e.g. DIA, IOP and IOM
(Jeong et al., 2010). Despite their slow chemical reactivity towards HO•
(R3) radicals, iodinated compounds are light sensitive Photolysis of these
(R4) molecules causes deiodination (Eloy et al., 1991; Wols et al., 2013).
Allard et al., (2016) demonstrated rapidly photolytically degraded IOP
(R5) and DIA up to >90% within 2 min in deionized water by immersed UV
irradiation at 254 nm (Allard et al., 2016). Although we use UV/H2O2
(R6)
treatment, similar fast complete conversion levels were found for IOM,
IOP and DIA within 15 min, demonstrating that the used UV-C source
3.3. UV-C/H2O2 oxidative chemistry (254 nm) is the most important technique for contrast agent degrada­
tion. In accordance with the study of Wols, we found that the majority of
UV-C/H2O2 advanced oxidation process is another useful technique pharmaceuticals do not photolytically degrade at 254 nm UV irradia­
for the degradation of micropollutants such as pharmaceuticals (Wols tion. Based on their data it was seen that DF, PHE and FLU can
et al., 2013). During the application with UV-C/H2O2 treatment, certain moderately decompose photolytically in Milli-Q, where MET, CP, CB,
molecules are sensitive for both interactions and are rapidly converted APAP and MF are scarcely sensitive to UV-C irradiation (Wols et al.,
by photolysis and hydroxyl radical attack (Wols et al., 2013). Hydrogen 2013). The poorest conversion levels were observed for APAP, CP and
peroxide in the gas phase is rapidly dissociated by UV photons, pro­ MF. Although UV-C/H2O2 treatment is an interplay between photo­
ducing highly reactive hydroxyl radicals (•OH) that can interact with chemical degradation accompanied with HO• radicals, we found that
nearly all organic compounds (von Sonntag 2008; Zhao et al., 2014).•OH UV-C irradiation has a large impact on the molecular degradation.
radicals will bind to C = C or C = N double bonds, initiate hydrogen (H)
abstraction, or trigger electron transfer reactions (von Sonntag 2008; 3.4. Effect of complex matrices on UV-C/H2O2 treatment
Banaschik et al., 2018). According to the degradation results (Fig. 4.), it
is seen that UV-C/H2O2 attacks nearly all pristine molecular structures Also during UV-C/H2O2 oxidative treatment the matrix complexity
completely in Milli-Q water, Table 3. Pseudo first-order reaction kinetics has a significant effect on the oxidative degradation of molecules. Using
ranged between 0.007 up to 0.591 min− 1, including corresponding again the inhibition factor (Eq. (4)), it has been observed that the
half-lives (t1/2) between 0.4 and 99.0 min. In contrast to thermal plasma coexisting substances in the aqueous matrix have even a greater influ­
treatment, the complete conversion of IOM, IOP and DIA with ence on UV-C/H2O2 oxidation than during plasma treatment. The
UV-C/H2O2 is most remarkable, since the molecular reactivity of these pharmaceutical degradation kinetics (Table 3) decrease when the matrix
complexity increases:

1.000(Milli− Q) : ​ 0.544(Tap ​ water) < 3.115(Synthetic Sewage) < 15.283(Synthetic urine)

< 54.00(Urine).

The rapid conversion levels in Milli-Q are presumably caused by the

absence of minerals and organic impurities. Retardation of the chemical
degradation kinetics in complex matrices like synthetic urine, diluted
urine and synthetic sewage are attributed to minerals and the natural
organic matter. Organic impurities, such as urea, meat extract and
peptone will absorb UV light at the expense of a less effective photolytic
process. Indirectly the production of HO• radicals is also diminished and
minerals such as phosphate and chlorine will induce the production of
less reactive secondary radicals (R7-8) (Borowska et al., 2015; Gian­
nakis et al., 2018).
(R7)

Fig. 4. UV-C/H2O2 oxidative degradation pattern in Milli-Q for the compounds

(R8)
IOM ( ), IOP ( ), DIA ( ), CIP( ), DOX ( ), FLU( ), DF ( ), MET ( ), CP( ), On the other hand, it was observed that the MF oxidation is increased
CB( ), TER ( 2), PHE( ), APAP ( ) and MF( ). in tap water. Chemical degradation kinetics are improved in tap water
matrix (0.028 min− 1) compared to Milli-Q (0.007 min− 1), showing
97.4% versus 52.9% degradation within 120 min, respectively. This
effect is attributed to photoreactive NO−3 (Lin et al., 2020), and in this
study also improved degradation of MF was demonstrated in water

7
M.H.F. Graumans et al. Environmental Research 195 (2021) 110884

Table 3
Pharmaceutical removal by UV-C/H2O2 treatment in different matrices.
Milli-Q Water Tap Water Synthetic Urine Urine Synthetic Sewage

R‾ k t1/2 R‾ k t1/2 R‾ k t1/2 R‾ k t1/2 R‾ k t1/2

(%) (min− 1) (min) (%) (min− 1) (min) (%) (min− 1) (min) (%) (min− 1) (min) (%) (min− 1) (min)
Ct(min) Ct(min) Ct(min) Ct Ct(min)
(min)

IOM 100.0 0.477 1.6 100.0 0.464 1.5 80.7 0.014 49.0 43.7 0.005 138.6 100.0 0.124 5.6
15 1.00 r2 15 1.00 r2 120 0.97 r2 120 0.58 r2 30 0.99 r2
IOP 100.0 0.587 1.2 100.0 0.361 1.9 85.3 0.015 46.2 53.2 0.007 99.0 100.0 0.101 6.9
15 1.00 r2 15 1.00 r2 120 0.98 r2 120 0.84 r2 30 0.99 r2
DIA 100.0 0.591 1.2 100.0 0.848 0.8 86.3 0.018 38.5 39.8 0.005 138.6 100.0 0.126 5.5
15 1.00 r2 15 1.00 r2 120 0.97 r2 120 0.88 r2 30 0.98 r2
CIP 100.0 n/aa – 100.0 0.292 2.4 86.2 0.015 46.2 8.6 0.001 693.2 100.0 0.159 4.4
15 - 30 0.99 r2 120 0.95 r2 120 0.44 r2 * 60 0.99 r2
DOX 100.0 0.067 10.4 n.d. – – 60.5 0.007 99.0 23.8 0.003 231.1 83.0 0.014 49.5
45 0.97 r2 - – 120 0.86 r2 120 0.44 r2 * 120 0.92 r2
FLU 87.9 0.016 43.3 100.0 0.109 6.4 62.5 0.008 86.6 18.5 0.002 346.6 94.8 0.027 25.7
120 0.95 r2 60 0.86 r2 120 0.99 r2 120 0.85 r2 120 0.84 r2
DF 100.0 n/aa – 100.0 1.853 0.4 100.0 0.046 15.1 69.7 0.010 69.3 99.3 0.036 19.3
15 - 15 1.00 r2 120 0.97 r2 120 0.99 r2 120 0.56 r2
MET 100.0 0.068 10.2 97.3 0.043 16.1 12.1 0.001 693.2 3.7 0.002 346.6 46.1 0.005 138.6
60 0.99 r2 120 0.99 r2 120 0.85 r2 120 0.59 r2 120 0.90 r2
CP 87.9 0.018 38.5 53.2 0.017 40.8 4.6 n.d. – 6.6 <0.001 >693.2 17.8 0.001 693.2
120 0.99 r2 120 0.99 r2 120 - 120 0.39 r2 * 120 0.54 r2
CB 100.0 0.075 9.2 88.5 0.019 99.0 8.1 0.001 693.2 N.C. – – 27.9 0.003 231.1
120 0.99 r2 120 0.99 r2 120 0.36 r2 * – – 120 0.81 r2
TER 100.0 0.079 8.8 100.0 n.d. – 53.1 0.005 138.6 4.1 <0.001 >693.2 70.7 0.009 77.0
90 0.99 r2 15 - 120 0.92 r2 120 0.05 r2 * 120 0.98 r2
PHE 100.0 0.249 2.8 100.0 0.013 1.6 78.5 0.013 53.3 35.0 0.004 173.3 100.0 0.110 63.0
30 0.93 r2 45 0.99 r2 120 0.99 r2 120 0.92 r2 120 0.99 r2
APAP 97.8 0.031 22.4 100.0 0.127 5.5 24.1 0.002 346.6 N.C. – – 70.1 0.009 77.0
120 0.75 r2 30 0.91 r2 120 0.95 r2 – – 90 0.95 r2
MF 52.9 0.007 99.0 97.4 0.028 24.8 40.7 0.003 231.1 11.3 0.001 693.2 9.6 0.002 346.6
120 0.92 r2 120 0.73 r2 120 0.52 r2 120 0.68 r2 120 0.47 r2 *

*: No significant change due to moderate to low correlation coefficient in linear modelling of assumed first-order kinetics (r2 leg < 0.50).
a
No reaction rate could be determined, by the rapid interaction of the pharmaceutical with either H2O2 or UV-C irradiation.

samples with amplified environmental factors, compared to pure water. 2020). Improved degradation results were found in the presence of high
Their experimental design was based on simulating environmental concentrations NO−3 (0.60–0.62 mg/L) and to a lesser extent Cl− (up to
photolysis by using 6 h of UV-A irradiation at 40 cm distance. During <4 mg/L), which both promote the photo-reactivity by producing HO•
this process, reactive oxygen species (ROS) are formed confirming that radicals (R9-R11) (Wang et al., 2017; Lin et al., 2020) As seen in the
HO• radical attack is the favored MF degradation pathway (Lin et al., suggested chemical reactions, it is expected that primarily NO2 and NO3

Fig. 5. SPE recovery percentages obtained in spiked synthetic sewage water (A). The extraction of known concentrations 1.0, 10.0 and 50.0 μg/L is compared to
100%( 6) recovery for the compounds; IOM ( ), IOP ( ), DIA ( ), CIP( ), DOX ( ), FLU( ), DF ( ), MET ( ), CP( ), CB( 5), TER ( ), PHE( ), APAP ( ) and MF
( ). Liquid-liquid extraction compared to SPE extraction (B). Recovery was determined for 50 μg/L MF( 9), without ( ) and with SDS ( ).

8
M.H.F. Graumans et al. Environmental Research 195 (2021) 110884

will play an improved role in UV-C/H2O2 tap water. As demonstrated in LC-MS/MS system contamination or have any recovery loss of other
Table S3 (SI paragraph S3), it is seen that both Cl2 and NO3 are present pharmaceuticals, an additional liquid-liquid extraction at subzero tem­
in concentrations exceeded the threshold of <4.0 mg/L (Lin et al., peratures from Yoshida et al. (1999) was modified especially for MF
2020). (Yoshida and Akane 1999). SDS was added to the sample to initiate a
cation-anion interaction between MF and SDS. ACN was added to extract
(R9)
UV
NO−3 →NO−3 ∗ the SDS-MF complex from the aqueous matrix. As presented in Fig. 5B, it
is seen that the LLE method can extract MF now with much better re­
(R10)
covery, ranging from 64.9 up to ~100%.
(R11)
3.6. Pharmaceutical detection and mitigation
3.5. Pharmaceutical extraction using solid phase extraction and liquid-
liquid extraction The optimized SPE and LLE extraction methods were used on real
wastewater samples taken from a hospital and residential area (Fig. 6).
A versatile SPE methodology was developed in synthetic sewage The concentrations (μg/L) of the 10 identified pharmaceuticals were
water for the detection of 14 different pharmaceuticals in raw hospital comparable with other studies of Vieno et al. (2006); Yin et al. (2010),
and residential wastewater samples. The extraction method was devel­ Petrie et al. (2016) and Ajo et al. (2018). Results reported for IOM and
oped and optimized for synthetic sewage water to achieve an optimum APAP are considered semi-quantitative because these values were
recovery in a sterile matrix with comparable properties as raw waste­ extrapolated in a range beyond the highest standard concentration (100
water. Due to the wide range of physicochemical properties of our μg/L). Even with this limitation of our analytical method, nearly all 14
selected pharmaceuticals (n = 14), the current developed SPE and LLE selected pharmaceuticals were detected in the hospital effluent
methodology are a compromise to achieve the best overall extraction (Table 4). The presence of IOM, DIA and CP, illustrates the difference
efficiency, see supporting information paragraph S8 for the detailed between hospital (HSW) and domestic sewage water (DSW), since
results. The currently developed SPE methodology has recovery per­ contrast agents and certain anticancer drugs are hospital administered
centages (S%) ranging from 27 up to > 100.0% for 13 pharmaceuticals only pharmaceuticals. Absence of IOP, is clarified by the fact that IOM is
in synthetic sewage water, see Fig. 5. the mainly used contrast agent for radiographic examinations at the
The extraction of MF, was insufficient, since recovery percentages Radboudumc. Despite the frequent usage of antibiotics, DOX has not
using Oasis HLB SPE cartridges were very low (S% >10.0). To minimize been detected in HSW and DSW, due to its limited stability and affinity

Fig. 6. Wastewater sampling points in the city of Nijmegen (the Netherlands). Sewage water was taken from a well at the Radboud University Medical Center (A) and
in a residential area (B). The sewer system of the end of pipe hospital effluent has no connection to the domestic sewage water well.

9
M.H.F. Graumans et al. Environmental Research 195 (2021) 110884

Table 4
Pharmaceutical detection in hospital sewage water compared to domestic sewage water in the city of Nijmegen.
Hospital sewage water (A)
Thursday – Monday (12 - 16 Sep 2019, including a weekend)
x‾ Concentration (μg/L)
(±sd) n = 4

IOMb IOP DIA DOX CIP FLU DF MET CP CB TER PHE APAP MF

2013.3 n.d. 8.2 n.d. 15.8 0.16 1.9 0.97 0.08 0.15 n.d. n.d. 0.43a 30.1
(±121.2) (±1.3) (±4.6) (±0.05) (±1.7) (±0.14) (±0.03) (±0.09) (±13.6)

Hospital sewage water (A)

Monday – Thursday (16 - 19 Sep 2019, working days)
x‾ Concentration (μg/L)
(±sd) n = 4

IOM IOP DIA DOX CIP FLU DF MET CP CB TER PHE APAP MF

2387.3 n.d. 10.1 n.d. 14.5 0.15 3.1 1.3 (±0.3) 0.34 0.08 n.d. n.d. 82.1 (±81.9) 36.8
(±190.9) (±1.9) (±4.9) (±0.04) (±3.6) (±0.10) (±0.23) (±14.9)

Hospital sewage water (Pooled)

x‾ Concentration (μg/L)
(±sd) n = 2

IOM IOP DIA DOX CIP FLU DF MET CP CB TER PHE APAP MF

2070.5 n.d. 7.8 n.d. 11.6 0.20 0.3 1.2 0.22 0.17 n.d. n.d. 0.07a (±0.16) 19.0
(±99.8) (±0.4) (±1.6) (±0.02) (±0.05) (±0.08) (±0.01) (±0.10) (±1.6)

Domestic sewage water (B)

Thursday (31 Oct 2019)
x‾ Concentration (μg/L)
(±sd) n = 4

IOM IOP DIA DOX CIP FLU DF MET CP CB TER PHE APAP MF
b
n.d. n.d. n.d. n.d. 1.1 0.5 (±0.4) 5.6 2.5 (±0.7) n.d. 1.9 (±0.7) n.d. n.d. 482.9 76.4
(±0.5) (±4.7) (±286.2) (±30.0)
a
It is recommended to analyse wastewater within 24h, to minimize biodegradation of APAP.
b
Determined concentrations exceeding the calibration range were quantified by linear extrapolation using the corresponding slope and intercept prepared in
synthetic sewage water.

for charged cations Ca2+ and Mg2+ (Soeborg et al., 2004; Loftsson (~10 μg/L), CIP (~13 μg/L), APAP (~300 μg/L) and MF (~35 μg/L)
2014). Tetracycline antibiotics, such as DOX, are stable under acidic were all much higher in HSW than in the spiked (5 μg/L) synthetic
conditions but undergo chemical reactions at neutral and basic pH, like sewage water matrix, see Table S8 (SI paragraph S10). According to the
epimerization where hydroxyl and hydrogen group substituents ex­ controlled experimental results, it has become clear, that the efficiency
change within the molecule by forming iso- or anhydrotetracyclines of pharmaceutical degradation mainly depends on pH, matrix
(Loftsson 2014). Tetracyclines predominantly absorb to sediment par­ complexity and oxidative technique used. A high pharmaceutical con­
ticles making it difficult to determine detectable amounts in wastewater centration is also of influence, but primarily slows down the rate of
samples (Thornton, 2001). That TER and PHE have not been found in the chemical degradation (Graumans et al., 2019). That the high concen­
sewage water samples might be attributed to fact that other pharma­ tration of IOM with UV-C/H2O2 treatment decreases so rapidly, is
ceuticals are used with the same therapeutic effect. Alternatives for TER attributed to its sensitivity to UV irradiation. To demonstrate effective­
and PHE are salbutamol and propyphenazone, molecules with very ness of both oxidative treatment techniques in hospital effluent, the total
similar chemical properties and therapeutic effects (Zuehlke et al., mass of the revealed pharmaceuticals was calculated according to a
2007). Additionally, the route of discharge is also of importance, since previously described method (Ajo et al., 2018). The overall conversion
concentrations of pharmaceuticals can vary per region. It is expected level was calculated with and without the contrast agents since the
that the presence of TER is limited in urban sewage water samples, since mass-based occurrence of IOM and DIA was relatively high compared to
the compound is mainly used as food additive in the veterinary sector to the other pharmaceuticals (Table 5).
increase the amounts of lean meat of livestock (Zhou et al., 2017). The According to the results provided in Table 5 and S8 (SI paragraph
concentrations of pharmaceuticals detected in both HSW and DSW, S10), it is suggested that thermal plasma treatment has the advantage
provides qualitative information about their consumption. In the HSW it over UV/H2O2 treatment, since in addition to ROS, also RNS are
is worth mentioning the difference between the weekday and weekend continuously produced in plasma activated water. With the exception of
sample, where the DSW sample reflects the continuous emission of IOM and DIA, thermal plasma oxidative degradation proves its effec­
prescription and over the counter medicines (Table 4). tiveness in very complex matrices. The less gradual chemical degrada­
tion observed, during oxidative treatment in HSW is attributed to the
3.6.1. Oxidative degradation chemistry in hospital sewage water matrix complexity. In our study only 10 compounds were detected in the
Previous pharmaceutical oxidative degradation results were ob­ HSW, but it is important to mention that many more analytes are to be
tained in a controlled setting. To evaluate the efficiency of plasma- expected (Ajo et al., 2018). Besides the pharmaceuticals that were not
induced water activation and UV-C/H2O2 on untreated hospital determined in this study, it is anticipated that molecules such as soaps,
effluent, pharmaceutical decomposition of the 10 identified compounds detergents, biotransformation products and hormones are also present
was determined, see Fig. 7. The overall conversion levels for IOM, DIA, in HSW. All these compounds will have an effect on the overall degra­
CIP, FLU, DF, MET, CP, CB, APAP and MF are all near equivalent to the dation efficiency e.g. competitive inhibition. The observed increase in
results obtained in the synthetic sewage water matrix. This is remark­ sample concentration overtime for the analytes CP and CB is caused by
able since the determined concentrations for IOM (~2400 μg/L), DIA their metabolites. During oxidative treatment, their non-detected

10
M.H.F. Graumans et al. Environmental Research 195 (2021) 110884

Fig. 7. Thermal plasma oxidation (A) was applied on hospital sewage water (HSW, n = 3). 10 compounds were identified, showing the conversion levels for; IOM ( ,
32.5%), DIA ( , 23.8%), CIP( , 69.7%), FLU( , 11.8%), DF ( , 93.6%), MET ( , 43.3%), CP( , 100.0%), CB( , 36.4%), APAP ( , 100.0%), MF( , 77.7%). UV-C/
H2O2 oxidative degradation (B) demonstrated; IOM ( , 99.9%), DIA ( , 100.0%), CIP( , 98.6%), FLU( , 4.0%), DF ( , 100.0%), MET ( , 36.5%), CP( , 82.3%), CB
( , N.C), APAP ( , 31.6%), MF( , 13.9%). For both techniques the pharmaceutical decomposition in HSW ( ) was compared to the controlled degradation results in
synthetic sewage water ( ). It is illustrated that both plasma oxidation (C) and UV-C/H2O2 (D) oxidative treatment in HSW initiate equivalent abatement as in the
simulated matrix.

Table 5
Total identified pharmaceutical concentration in hospital sewage water, demonstrating the effectiveness of plasma oxidation and UV/H2O2 treatment with and without
contrast agents.
AOP Total Conc. pharmaceuticals (μg/L) Removal (%) Total Conc. w/o IOM and DIA (μg/L) Removal (%) w/o IOM and DIA

Plasma 150WStart 2782.1 41.6 344.1 96.7

End120 min 1624.7 11.2
UV-C/H2O2 2881.9 91.4 371.8 33.2
End120 min 248.5 248.3

biological metabolites are reformed into the pristine structures. This will matrix complexity and an unexpected broad concentration range in real
cause an increased peak value relative to the measured initial concen­ life wastewater our analytical approach for synthetic sewage water
tration. Continuation of the oxidative treatment will cause further showed some limitations for quantification of pharmaceuticals like IOM
abatement and make it possible to degrade the compound completely and APAP. Additionally, our lowest calibration point (0.5 μg/L) did not
(Aziz et al., 2018; Ajo et al., 2018; Graumans et al., 2020). optimally serve quantification of FLU, DF, CP and CB, since their
determined concentrations ranged between the LOD and lowest cali­
3.7. Strengths and limitations of the study bration point in synthetic sewage water. Nonetheless, this laboratory
study represents a useful first step in the experimental validation of
We developed an experimental set-up to provide insight into the plasma activation technology to complement onsite wastewater
degradation of 14 pharmaceuticals in complex matrices by use of AOP. treatment.
The application of plasma treatment compared to UV-C/H2O2 on
simulated matrices, indicates dependency of the degradation efficiency 4. Conclusion
on matrix complexity and also varies with the method of oxidative
treatment and the structural properties of the pharmaceuticals tested. With this study we have extensively studied the degradation of
The simulation matrices were compared with end-of-pipe hospital pharmaceuticals in multiple complex matrices using thermal plasma and
sewage water, sampled and processed by advanced oxidation. Due to UV/H2O2 oxidative treatment. Despite the matrix complexity has each

11
M.H.F. Graumans et al. Environmental Research 195 (2021) 110884

technique its own added value towards specific pharmaceutical classes. Ferrando-Climent, L., Rodriguez-Mozaz, S., Barcelo, D., 2014. Incidence of anticancer
drugs in an aquatic urban system: from hospital effluents through urban wastewater
Single dose hydrogen peroxide in stationary UV-C/H2O2 treatment has a
to natural environment. Environ. Pollut. 193, 216–223. https://doi.org/10.1016/j.
temporary effect, and involved photons cannot produce RNS, whereas envpol.2014.07.002.
ongoing plasma activation can produce both ROS and RNS continuously. Gerrity, D., Stanford, B.D., Trenholm, R.A., Snyder, S.A., 2010. An evaluation of a pilot-
The continuous production of reactive species makes plasma activation a scale nonthermal plasma advanced oxidation process for trace organic compound
degradation. Water Res. 44 (2), 493–504. https://doi.org/10.1016/j.
useful technique for the application in complex matrices such as hospital watres.2009.09.029.
effluent. The pharmaceutical degradation by temperature-controlled Giannakis, S., Androulaki, B., Comninellis, C., Pulgarin, C., 2018. Wastewater and urine
thermal plasma treatment is complemented by nitrification under treatment by UVC-based advanced oxidation processes: implications from the
interactions of bacteria, viruses, and chemical contaminants. Chem. Eng. J. 343,
acidic conditions. Further research using cell-based assays should reveal 270–282. https://doi.org/10.1016/j.cej.2018.03.019.
whether the toxicity of contaminated matrices is decreased. Graumans, M.H.F., Hoeben, W., Russel, F.G.M., Scheepers, P.T.J., 2020. Oxidative
degradation of cyclophosphamide using thermal plasma activation and UV/H2O2
treatment in tap water. Environ. Res. 182, 109046. https://doi.org/10.1016/j.
Declaration of competing interest envres.2019.109046.
Gu, H., Liu, G., Wang, J., Aubry, A.F., Arnold, M.E., 2014. Selecting the correct weighting
factors for linear and quadratic calibration curves with least-squares regression
The authors declare that they have no known competing financial algorithm in bioanalytical LC-MS/MS assays and impacts of using incorrect
interests or personal relationships that could have appeared to influence weighting factors on curve stability, data quality, and assay performance. Anal.
Chem. 86 (18), 8959–8966. https://doi.org/10.1021/ac5018265.
the work reported in this paper. Hoeben, W.F.L.M., van Ooij, P.P., Schram, D.C., Huiskamp, T., Pemen, A.J.M., Lukes, P.,
2019. On the possibilities of straightforward characterization of plasma activated
water. Plasma Chem. Plasma Process. 39 (3), 597–626. https://doi.org/10.1007/
Acknowledgements
s11090-019-09976-7.
Jeong, J., Jung, J., Cooper, W.J., Song, W., 2010. Degradation mechanisms and kinetic
Financial support was made possible by the INTERREG Deutschland- studies for the treatment of X-ray contrast media compounds by advanced oxidation/
Nederland program [grant number 142118]. The authors would like to reduction processes. Water Res. 44 (15), 4391–4398. https://doi.org/10.1016/j.
watres.2010.05.054.
thank VitalFluid who provided the PAW lab unit and UV-C source. Jewell, K.S., Wick, A., Ternes, T.A., 2014. Comparisons between abiotic nitration and
biotransformation reactions of phenolic micropollutants in activated sludge. Water
Res. 48, 478–489. https://doi.org/10.1016/j.watres.2013.10.010.
Appendix A. Supplementary data
Joshi, R.P., Thagard, S.M., 2013. Streamer-like electrical discharges in water: Part II.
Environmental applications. Plasma Chem. Plasma Process. 33 (1), 17–49. https://
Supplementary data to this article can be found online at https://doi. doi.org/10.1007/s11090-013-9436-x.
Kohler, C., Venditti, S., Igos, E., Klepiszewski, K., Benetto, E., Cornelissen, A., 2012.
org/10.1016/j.envres.2021.110884.
Elimination of pharmaceutical residues in biologically pre-treated hospital
wastewater using advanced UV irradiation technology: a comparative assessment.
References J. Hazard Mater. 239–240, 70–77. https://doi.org/10.1016/j.jhazmat.2012.06.006.
Li, Y., Niu, X.M., Yao, C., Yang, W., Lu, G.H., 2019. Distribution, removal, and risk
assessment of pharmaceuticals and their metabolites in five sewage plants. Int. J.
Ajo, P., Preis, S., Vornamo, T., Manttari, M., Kallioinen, M., Louhi-Kultanen, M., 2018.
Environ. Res. Publ. Health 16 (23). ARTN 472910.3390/ijerph16234729.
Hospital wastewater treatment with pilot-scale pulsed corona discharge for removal
Lin, W.T., Zhang, X.H., Li, P., Tan, Y.Z., Ren, Y., 2020. Ultraviolet photolysis of
of pharmaceutical residues (vol 6, pg 1569, 2018). J Environ Chem Eng 6 (2).
metformin: mechanisms of environmental factors, identification of intermediates,
https://doi.org/10.1016/j.jece.2018.06.013. A1-A1.
and density functional theory calculations. Environ. Sci. Pollut. Res. 27 (14),
Allard, S., Criquet, J., Prunier, A., et al., 2016. Photodecomposition of iodinated contrast
17043–17053. https://doi.org/10.1007/s11356-020-08255-9.
media and subsequent formation of toxic iodinated moieties during final disinfection
Loftsson, T., 2014. Tetracyclines Drug Stability for Pharmaceutical Scientists.
with chlorinated oxidants. Water Res. 103, 453–461. https://doi.org/10.1016/j.
Magureanu, M., Mandache, N.B., Parvulescu, V.I., 2015. Degradation of pharmaceutical
watres.2016.07.050.
compounds in water by non-thermal plasma treatment. Water Res 81, 124–136.
Almeida, A.M., Castel-Branco, M.M., Falcão, A.C., 2002. Linear regression for calibration
https://doi.org/10.1016/j.watres.2015.05.037.
lines revisited: weighting schemes for bioanalytical methods. J. Chromatogr. B 774,
Magureanu, M., Bradu, C., Parvulescu, V.I., 2018. Plasma processes for the treatment of
215–222. https://doi.org/10.1016/S1570-0232(02)00244-1.
water contaminated with harmful organic compounds. J. Phys. D Appl. Phys. 51 (31)
Aziz, K., Miessner, H., Mueller, S., et al., 2018. Comparative study on 2,4-dichlorophe­
https://doi.org/10.1088/1361-6463/aacd9c. ARTN 313002.
noxyacetic acid and 2,4-dichlorophenol removal from aqueous solutions via
Markovic, M., Jovic, M., Stankovic, D., Kovacevic, V., Roglic, G., Gojgic-Cvijovic, G.,
ozonation, photocatalysis and non-thermal plasma using a planar falling film
Manojlovic, D., 2015. Application of non-thermal plasma reactor and Fenton
reactor. J. Hazard Mater. 343, 107–115. https://doi.org/10.1016/j.
reaction for degradation of ibuprofen. Sci. Total Environ. 505, 1148–1155. https://
jhazmat.2017.09.025.
doi.org/10.1016/j.scitotenv.2014.11.017.
Banaschik, R., Lukes, P., Jablonowski, H., Hammer, M.U., Weltmann, K.D., Kolb, J.F.,
Miller, J.N., Miller, J.C., 2010. Statistics and Chemometrics for Analytical Chemistry.
2015. Potential of pulsed corona discharges generated in water for the degradation
Pearson, pp. 110–153.
of persistent pharmaceutical residues. Water Res. 84, 127–135. https://doi.org/
OECD, 2001. OECD GUIDELINE for the TESTING of CHEMICALS Simulation Test -
10.1016/j.watres.2015.07.018.
Aerobic Sewage Treatment: 303 A: Activated Sludge Units - 303 B: Biofilms.
Banaschik, R., Jablonowski, H., Bednarski, P.J., Kolb, J.F., 2018. Degradation and
https://www.oecd-ilibrary.org/.
intermediates of diclofenac as instructive example for decomposition of recalcitrant
Peake, B.M., Braund, R., Tong, A.Y.C., Tremblay, L.A., 2016. The Life-Cycle of
pharmaceuticals by hydroxyl radicals generated with pulsed corona plasma in water.
Pharmaceuticals in the Environment. Woodhead Publishing, Cambridge, United
J. Hazard Mater. 342, 651–660. https://doi.org/10.1016/j.jhazmat.2017.08.058.
Kingdom.
Beier, S., Koster, S., Veltmann, K., Schroder, H., Pinnekamp, J., 2010. Treatment of
Pemen, A.J.M., van Ooij, P.P., Beckers, F.J.C.M., et al., 2017. Power modulator for high-
hospital wastewater effluent by nanofiltration and reverse osmosis. Water Sci.
yield production of plasma-activated water. Ieee T Plasma Sci 45 (10), 2725–2733.
Technol. 61 (7), 1691–1698. https://doi.org/10.2166/wst.2010.119.
https://doi.org/10.1109/Tps.2017.2739484.
Borowska, E., Felis, E., Zabczynski, S., 2015. Degradation of iodinated contrast media in
Petrie, B., Youdan, J., Barden, R., Kasprzyk-Hordern, B., 2016. Multi-residue analysis of
aquatic environment by means of UV, UV/TiO2 process, and by activated sludge.
90 emerging contaminants in liquid and solid environmental matrices by ultra-high-
Water, Air, Soil Pollut. 226 (5). ARTN 15110.1007/s11270-015-2383-9.
performance liquid chromatography tandem mass spectrometry. J. Chromatogr. A
Bourin, M., Jolliet, P., Ballereau, F., 1997. An overview of the clinical pharmacokinetics
1431, 64–78. https://doi.org/10.1016/j.chroma.2015.12.036.
of x-ray contrast media. Clin. Pharmacokinet. 32 (3), 180–193. https://doi.org/
Rosati, G., 1994. Clinical-pharmacology of iomeprol. Eur. J. Radiol. 18, S51–S60.
10.2165/00003088-199732030-00002.
https://doi.org/10.1016/0720-048x(94)90094-9.
Brisset, J.L., Benstaali, B., Moussa, D., Fanmoe, J., Njoyim-Tamungang, E., 2011. Acidity
Shih, K.Y., Locke, B.R., 2011. Optical and electrical diagnostics of the effects of
control of plasma-chemical oxidation: applications to dye removal, urban waste
conductivity on liquid phase electrical discharge. Ieee T Plasma Sci 39 (3), 883–892.
abatement and microbial inactivation. Plasma Sources Sci. Technol. 20 (3) https://
https://doi.org/10.1109/Tps.2010.2098052.
doi.org/10.1088/0963-0252/20/3/034021. Artn 034021.
Shimizu, T., Kishimoto, N., Sato, T., 2020. Effect of electrical conductivity of water on
Brown, T.L., Bursten, B.E., Langford, S., Sagatys, D., Duffy, 2007. Alkenes, Alkynes and
plasma-driven gas flow by needle-water discharge at atmospheric pressure.
Arenes Chemistry. The central science a broad perspectivePearson Education
J. Electrost. 104. ARTN 10342210.1016/j.elstat.2020.103422.
Australia, French Forest, pp. 888–937.
Soeborg, T., Ingerslev, F., Halling-Sorensen, B., 2004. Chemical stability of
Chiron, S., Gomez, E., Fenet, H., 2010. Nitration processes of acetaminophen in nitrifying
chlortetracycline and chlortetracycline degradation products and epimers in soil
activated sludge. Environ. Sci. Technol. 44 (1), 284–289. https://doi.org/10.1021/
interstitial water. Chemosphere 57 (10), 1515–1524. https://doi.org/10.1016/j.
es902129c.
chemosphere.2004.09.020.
Eloy, R., Corot, C., Belleville, J., 1991. Contrast media for angiography: physicochemical
properties, pharmacokinetics and biocompatibility. Clin. Mater. 7 (2), 89–197.
https://doi.org/10.1016/0267-6605(91)90045-h.

12
M.H.F. Graumans et al. Environmental Research 195 (2021) 110884

Souza, F.S., Feris, L.A., 2016. Hospital and municipal wastewater: identification of von Sonntag, C., 2008. Advanced oxidation processes: mechanistic aspects. Water Sci.
relevant pharmaceutical compounds. Water Environ. Res. 88 (9), 871–877. https:// Technol. 58 (5), 1015–1021. https://doi.org/10.2166/wst.2008.467.
doi.org/10.2175/106143016X14609975747603. Wang, Y.F., Roddick, F.A., Fan, L.H., 2017. Direct and indirect photolysis of seven
Squadrito, G.L., Pryor, W.A., 1998. Oxidative chemistry of nitric oxide: the roles of micropollutants in secondary effluent from a wastewater lagoon. Chemosphere 185,
superoxide, peroxynitrite, and carbon dioxide. Free Radic. Biol. Med. 25 (4–5), 297–308. https://doi.org/10.1016/j.chemosphere.2017.06.122.
392–403. https://doi.org/10.1016/s0891-5849(98)00095-1. Wols, B.A., Hofman-Caris, C.H., Harmsen, D.J., Beerendonk, E.F., 2013. Degradation of
Suarez, S., Lema, J.M., Omil, F., 2009. Pre-treatment of hospital wastewater by 40 selected pharmaceuticals by UV/H2O2. Water Res. 47 (15), 5876–5888. https://
coagulation-flocculation and flotation. Bioresour. Technol. 100 (7), 2138–2146. doi.org/10.1016/j.watres.2013.07.008.
https://doi.org/10.1016/j.biortech.2008.11.015. Yin, J., Yang, Y., Li, K., Zhang, J., Shao, B., 2010. Analysis of anticancer drugs in sewage
Thirumdas, R., Kothakota, A., Annapure, U., et al., 2018. Plasma activated water (PAW): water by selective SPE and UPLC-ESI-MS-MS. J. Chromatogr. Sci. 48 (10), 781–789.
chemistry, physico-chemical properties, applications in food and agriculture. Trends https://doi.org/10.1093/chromsci/48.10.781.
Food Sci. Technol. 77, 21–31. https://doi.org/10.1016/j.tifs.2018.05.007. Yoshida, M., Akane, A., 1999. Subzero-temperature liquid-liquid extraction of
Thornton, L., Butler, D., Docx, P., Hession, M., Makropoulos, C., McMullen, M., benzodiazepines for high performance liquid chromatography. Anal. Chem. 71 (9),
Nieuwenhuijsen, M., Pitman, A., Rautiu, R., Sawyer, R., Smith, S., White, D., 1918–1921. https://doi.org/10.1021/ac981276g.
Wilderer, P., Paris, S., Marani, D., Braguglia, C., Palerm, J., 2001. Pollutants in Zhao, C., Arroyo-Mora, L.E., DeCaprio, A.P., Sharma, V.K., Dionysiou, D.D., O’Shea, K.E.,
Urban Waste Water and Sewage Sludge. European Commission, Luxembourg. 2014. Reductive and oxidative degradation of iopamidol, iodinated X-ray contrast
Tokumura, M., Sugawara, A., Raknuzzaman, M., Habibullah-Al-Mamun, M., media, by Fe(III)-oxalate under UV and visible light treatment. Water Res. 67,
Masunaga, S., 2016. Comprehensive study on effects of water matrices on removal of 144–153. https://doi.org/10.1016/j.watres.2014.09.009.
pharmaceuticals by three different kinds of advanced oxidation processes. Zhou, L., Sleiman, M., Ferronato, C., Chovelon, J.M., de Sainte-Claire, P., Richard, C.,
Chemosphere 159, 317–325. https://doi.org/10.1016/j.chemosphere.2016.06.019. 2017. Sulfate radical induced degradation of beta 2-adrenoceptor agonists
Vieno, N.M., Tuhkanen, T., Kronberg, L., 2006. Analysis of neutral and basic salbutamol and terbutaline: phenoxyl radical dependent mechanisms. Water Res.
pharmaceuticals in sewage treatment plants and in recipient rivers using solid phase 123, 715–723. https://doi.org/10.1016/j.watres.2017.07.025.
extraction and liquid chromatography-tandem mass spectrometry detection. Zuehlke, S., Duennbier, U., Heberer, T., 2007. Investigation of the behavior and
J. Chromatogr. A 1134 (1–2), 101–111. https://doi.org/10.1016/j. metabolism of pharmaceutical residues during purification of contaminated ground
chroma.2006.08.077. water used for drinking water supply. Chemosphere 69 (11), 1673–1680. https://
doi.org/10.1016/j.chemosphere.2007.06.020.

13