Electrochemical oxidation of amoxicillin on carbon nanotubes and carbon nanotube supported metal modified electrodes

Article abstract of DOI:10.1016/j.cattod.2019.06.039

The electrolysis of amoxicillin (AMX) was carried out on CNT, Pt/CNT and Ru/CNT modified electrodes based on Carbon Toray in 0.1 M NaOH, 0.1 M NaCl and 0.1 M Na₂CO₃/NaHCO₃ buffer (pH 10) media with the aim of studying the significance of two factors, electrode material and pH, on the oxidative degradation and mineralization of AMX. For this purpose, the electrolysis products were identified by HPLC-MS and GC–MS, and quantified by HPLC-UV-RID and IC. The highest carbon mineralization efficiency, corresponding to 30% of the oxidized AMX, was found for Pt/CNT modified electrode in carbonate buffer medium. Regarding to the AMX conversion, the results show that the effect of pH is higher than that of the electrode material. Principal component analysis allowed to determine the experimental parameters vs. product distribution relationship and to elucidate the oxidation pathways for the studied electrodes. The results show that the hydroxylation of the aromatic ring and the nitrogen atom play an important role on the efficient degradation of AMX.

Full text of DOI:10.1016/j.cattod.2019.06.039

Accepted Manuscript
Title: Electrochemical oxidation of amoxicillin on carbon
nanotubes and carbon nanotube supported metal modified
electrodes
Authors: Marta Ferreira, Iwona Kuzniarska-Biernacka,
´
´
Antonio M. Fonseca, Isabel C. Neves, Olıvia S.G.P. Soares,
´
Manuel F.R. Pereira, Jose L. Figueiredo, Pier Parpot
PII:
S0920-5861(19)30074-4
DOI:
Reference:
https://doi.org/10.1016/j.cattod.2019.06.039
CATTOD 12286
To appear in:
Catalysis Today
Received date:
Revised date:
Accepted date:
14 February 2019
30 May 2019
12 June 2019
Please cite this article as: Ferreira M, Kuzniarska-Biernacka I, Fonseca AM, Neves
IC, Soares OSGP, Pereira MFR, Figueiredo JL, Parpot P, Electrochemical oxidation
of amoxicillin on carbon nanotubes and carbon nanotube supported metal modified
electrodes, Catalysis Today (2019), https://doi.org/10.1016/j.cattod.2019.06.039
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Electrochemical oxidation of amoxicillin on carbon nanotubes and
carbon nanotube supported metal modified electrodes
Marta Ferreira,1 Iwona Kuzniarska-Biernacka,1 António M. Fonseca,1,2 Isabel C.
Neves,1,2 Olívia S.G.P. Soares,3 Manuel F.R. Pereira,3 José L. Figueiredo,3 and Pier
Parpot*1,2
1CQUM, Centro de Química, Escola de Ciências, Universidade do Minho, Campus
Gualtar, 4710-057, Braga, Portugal - parpot@quimica.uminho.pt
2CEB - Centre of Biological Engineering, Universidade do Minho, Campus Gualtar,
4710-057, Braga, Portugal
3Laboratório de Catálise e Materiais (LCM), Laboratório Associado LSRE-LCM,
Departamento de Engenharia Química, Faculdade de Engenharia, Universidade do
Porto, Rua Dr. Roberto Frias, 4200-465 Porto, Portugal
Graphical abstract
1

Eletrode material
Degradation
products
AMX
CNT
Pt
CNT
Chromatographic techniques




HPLC-MS
GC-MS
HPLC-UV-RID
IC
Ru
Electrolysis ofAMX
CNT
Highlights:
 Study of anode material and pH effect on the oxidation and mineralization of
AMX.
 Pattern recognition for electrode material/reaction intermediates relationship.
 Highest mineralization rate found for Pt/CNT anode in carbonate buffer medium.
 Influence of surface properties of electrocatalyst on AMX oxidation was
emphasized.
Abstract: The electrolysis of amoxicillin (AMX) was carried out on CNT, Pt/CNT and
Ru/CNT modified electrodes based on Carbon Toray in 0.1 M NaOH, 0.1 M NaCl and
0.1 M Na2CO3/NaHCO3 buffer (pH 10) media with the aim of studying the significance
of two factors, electrode material and pH, on the oxidative degradation and mineralization
of AMX. For this purpose, the electrolysis products were identified by HPLC-MS and
GC-MS, and quantified by HPLC-UV-RID and IC. The highest carbon mineralization
efficiency, corresponding to 30 % of the oxidized AMX, was found for Pt/CNT modified
electrode in carbonate buffer medium. Regarding to the AMX conversion, the results
show that the effect of pH is higher than that of the electrode material. Principal
component analysis allowed to determine the experimental parameters vs. product
distribution relationship and to elucidate the oxidation pathways for the studied
2

electrodes. The results show that the hydroxylation of the aromatic ring and the nitrogen
atom play an important role on the efficient degradation of AMX.
Keywords: Electro-oxidation, Amoxicillin, CNT and M/CNT modified electrodes,
Surface properties, Reaction pathways, Principal Component Analysis (PCA)
1. Introduction
The increasing amounts of pharmaceutical compounds and some of their degradation
products in water and wastewater is one of the most important threats for public health
and environment. Among these compounds, the class of antibiotics is the cause of special
concern, considering both their relative amount and their resistance to conventional
biological treatments. A significant number of research works in the field of the detection
and removal of pharmaceuticals, particularly antibiotics, was recently carried out [1-17]
having in mind this concern. Taking into account their significant use for human health
and veterinary treatments, a wide range of antibiotics and their degradation products can
be found in several environmental water bodies. The existing methods suffer drawbacks,
including the need for applying high temperatures and pressures, high cost of materials,
use of toxic or environmentally harmful reagents, problems of stability. Due to their
versatility, high efficiency, easy control, amenability to automation, environmental
compatibility and cost effectiveness, electrochemical methods can be considered an
effective tool for water remediation [18]. These are considered green methods [19] and
may constitute an alternative to the conventional approaches, overcoming existing
technological barriers. The inherent advantage is their environmental compatibility, since
the main reagent, the electron, is a “clean reagent”. Electrochemical processes generally
3

have lower temperature and pressure requirements than their equivalent non-
electrochemical counterparts. In this context, anodic oxidation may allow an efficient
oxidation in relatively mild conditions without harmful redox agents [20-22]. The applied
overpotential can decrease the energy barrier for oxidation, compared with the chemical
catalytic route, and consequently can change the rate of the electrode process.
Considering the catalytic route, the search for active, stable and environmentally
compatible catalysts is continuing. The peculiar characteristics of CNTs, namely high
surface area and high electrical conductivity, qualify them as adequate electrocatalysts or
as catalyst support materials for the oxidative degradation of organic pollutants in
wastewater [18].
Platinum, palladium and ruthenium are well known active and stable electrocatalysts for
organic compounds oxidation. Ru and RuO2 are also considered suitable for the
production of active chlorine compounds on the electrode surface, since they can provide
lower and higher over-potentials for chorine and oxygen production respectively, in
comparison with other metals or carbonaceous electrodes.
There are several contributions in the literature concerning the exhaustive oxidation of
organic compounds by electrochemical methods [23-36] but, to our knowledge, the
present work reports for the first time the electrochemical oxidation of AMX on metal
supported CNT modified electrodes. The electroreactivity of amoxicillin (AMX) on
different carbon nanotube supported metal catalysts was studied recently by our group,
having in mind a compromise between lower anodic potentials and higher current
densities [37]. This work includes the preliminary voltammetric study in order to
elucidate the electroreactivity of AMX on different M/CNT electrodes in different pH
media and to determine the kinetic parameters of the redox reactions. The choice of
electrode materials and pH media for the electrolysis of AMX was also discussed. The
4

present work aims to establish relationships between product distribution, electrode
material and pH. For this purpose, constant potential electrolyses of AMX were carried
out using CNT, Ru/CNT and Pt/CNT modified electrodes in several pH media, and the
corresponding analytical data were exploited using chemometric methods. One of the
critical issues in the field of pollutant oxidation is the exhaustive determination of
degradation products, since a decrease of total organic carbon (TOC) content alone could
not be sufficient, considering the possibility of producing some degradation products even
more harmful for human health and environment than the initial ones. Having in mind
this concern, an extensive analytical work was also carried out using several
chromatographic and spectroscopic methods.
2. Experimental
2.1. Preparation of carbon nanotube supported metal catalysts
The preparation of the carbon nanotubes (CNT, Nanocyl-3100) and carbon nanotube
supported metal (M/CNT) modified electrodes was performed according to established
procedures described in our previous works [18, 22, 23]. H2PtCl6 and RuCl3 (Alfa Aeser)
were used as metal precursors for preparing M/CNT by the incipient wetness method.
The amounts of Pt or Ru used for the preparation of the carbon nanotubes supported metal
catalysts were always 1 wt.%. During the thermal treatment, the monometallic catalysts
were first dried at 100 ºC for 24 h. In the next step, Pt/CNT was thermally treated under
nitrogen flow at 200 ºC for 1 h and reduced at 200 ºC under hydrogen flow for 3 h while
Ru/CNT was heat treated and reduced at 250 ºC (these temperatures were selected based
on previous temperature programmed reduction experiments). The resulting materials
were cooled to room temperature under a nitrogen flow. The details and the results of the
5

characterization of carbon nanotube supported metal catalysts are described elsewhere
[38].
For the preparation of M/CNT modified electrodes, a catalyst (loading of 0.5 mg cm-2)
prepared with ultra-pure water (Barnsted E-pure system, 18.2 MΩ cm at 20 °C) and
Nafion® suspension (5 wt.% Sigma-Aldrich®) was homogeneously deposited onto the
Carbon Toray paper (CT, geometrical area of 4.0 cm2, Quintech) and the solvent was then
evaporated at room temperature overnight.
2.2. Electrochemical Setup and product analysis
Cyclic voltammetry (CV) study was performed in a thermostated two-compartment glass
cell separated by ion exchange membranes (Nafion 117 and Ionac). The reference and
counter electrodes were a saturated calomel electrode and a platinum foil (99.95%)
electrode, respectively. Ultra-pure nitrogen (U Quality from Air Liquide) was used to
deaerate the solution before each experiment and to maintain an inert atmosphere over
the solution during the measurements. The cleanness of the surfaces was tested prior each
experiment by recording voltammograms in the supporting electrolyte alone.
The electrochemical oxidation of organic compounds, which involves oxygenated
species, is generally favored in alkaline medium, since an excess of OH- ions is expected.
Two alkaline media were studied in this work: carbonate buffer and 0.1 M NaOH. These
two different pH media provide different interactions between the electroactive substrate
and the electrode surface which could be partially or totally oxidized by adsorbed OH
species, depending on pH. The 0.1 M NaCl medium was chosen in order to study the
effect of introducing Cl– ions in the reactional medium since the presence of Cl- leads to
active chlorine species that are strong oxidants. The supporting electrolytes were prepared
using ultrapure water with sodium chloride (Panreac, 99.5 %), sodium carbonate
6

anhydrous (Panreac, 99.8 %), sodium bicarbonate (Panreac, 98 %) or sodium hydroxide
(Sigma-Aldrich).
The electrochemical instrumentation consists of a potentiostat/galvanostat from Amel
Instruments coupled to a computer by an AD/DA converter. The Labview software
(National Instruments) and a PCI-MIO-16E-4 I/O module were used for generating and
applying the potential program as well as acquiring data, such as current intensities.
Two chromatographic set-ups were used for the quantitative analysis of the reaction
products. One consisted of a high performance liquid chromatograph (HPLC) with an
isocratic pump and double on line detection, including an UV-Vis detector and a
refractometer, and the other was an ion chromatograph (IC, Dionex) with a conductivity
detector. The separation of reaction products was carried out using Aminex HPX-87 H
(Biorad) and Merck C18 analytical columns for HPLC-UV and an AS11-HC column
(Dionex Corp.) for IC. Some gaseous reaction products were detected by a dual cell
microvolume thermal conductivity detector (Vici, Valco Instruments Co. Inc.). The
product separation, in this case, was carried out by a packed GC column: chromosorb 104
(1.8 m × 1/8 in. × 2.1 mm). TOC analysis was performed by the NPOC (non-purgeable
organic carbon) method using a LTOC Total Organic Carbon Analyzer of Shimadzu
coupled to an ASI-L auto sampler of the same brand.
3. Results and Discussion
The electrochemical oxidation of amoxicillin (AMX) was carried out using carbon
nanotubes (CNT) and carbon nanotube supported metal (M/CNT) modified electrodes in
several pH media in order to determine the effect of the electrolysis conditions on product
distribution. The effect of different parameters like electrode material, pH and applied
7

potential on substrate conversion and selectivity was evaluated in order to achieve the
best conditions for amoxicillin removal.
3.1. Electrolysis of amoxicillin in NaOH medium
The electrolysis of AMX was carried out on CNT, Ru/CNT and Pt/CNT modified
electrodes at 1.5 V. vs. SCE, in 0.1 M NaOH medium. In this medium AMX is partially
transformed into amoxicillin penicilloic acid (APcA) and amoxicillin penilloic acid
(APIA) [8].
Cyclic voltammograms of different modified electrodes recorded at the beginning and at
the end of the electrolysis of 2 mM AMX in 0.1 M NaOH medium are displayed in Figure
1.
12
10
8
25
20
15
10
5
12
8
30
20
a
b
6
4
4
10
2
0
0
0
0
-5
-2
-4
-6
-8
-10
-20
-30
-4
-8
-10
-15
-20
-1.5
-1.0
-0.5
0.0
0.5
1.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
E / V vs. SCE
E / V vs. SCE
12
8
30
20
b
4
10
0
0
-10
-20
-30
-4
-8
-1.5
-1.0
-0.5
0.0
0.5
1.0
E / V vs. SCE
Figure 1. Cyclic voltammograms of a) CNT, b) Ru/CNT and c) Pt/CNT () in 0.1 M
NaOH, () at the beginning and () at the end of the electrolysis of 2 mM AMX.
8

A decrease of current intensities, especially in the range of -1.5 to 0 V vs. SCE, was
noticed at the end of the electrolysis in comparison with the cyclic voltammograms
recorded at the beginning of the electrolysis for the metal modified electrodes. For these
electrodes, especially for Pt/CNT, a shift of solvent reduction potential to more cathodic
values was noticed for the cyclic voltammograms corresponding to the end of the
electrolysis compared to those obtained at the beginning. These results show that
deactivation of the electrode surface by irreversibly adsorbed species could be important
for Pt/CNT in these conditions.
The consumption of AMX during the electrolysis was confirmed by UV/vis analysis,
where a significant decrease of the absorbance in the range of 230–270 nm was observed.
Figure 2 shows the UV spectra of the pure AMX solution and the electrolyzed solutions.
These data show also an increase of the absorbance in the range of 210-230 indicating the
formation of products with relatively few conjugated groups.
3.0
2.5
2.0
1.5
1.0
0.5
0.0
200
220
240
260
280
300
320
340
360
Wavelenght / nm
Figure 2. UV spectra of blank solution of AMX () and electrolyzed solutions of AMX
(t = 8 h of electrolysis) on () Pt/CNT and () Ru/CNT on modified electrodes in 0.1
M NaOH medium.
9

After 8 h of electrolysis 12, 43 and 58 % of AMX conversion was achieved, respectively
for CNT, Ru/CNT and Pt/CNT modified electrodes. A significant part of the bonded
-
nitrogen in the AMX molecule accumulates as NO3 and the chemically bonded sulfur is
2-
transformed partially into SO4 . The sulfur and nitrogen mineralization was determined
-
by ion chromatography (IC) as SO42- and NO3 percentages, while carbon mineralization
was determined by IC and TOC analysis and expressed as CO32- percentage. The best
2-
2-
mineralization efficiencies were obtained for Ru/CNT, with percentages of CO3 , SO4
-
and NO3 at the end of the electrolysis corresponding to respectively 27, 18 and 8 % of
the transformed AMX. For the Pt/CNT modified electrode, the percentages of these
compounds were respectively 19, 15 and 5 %. The decrease of the AMX concentration
and the increase of the carbonate concentration during the electrolysis after the correction
using blank solution are given in Figure 3. The blank data were obtained applying the
same potential to the electrode, in 0.1 M NaOH solution without AMX, during the same
time period corresponding to the electrolysis.
2
1.8
1.6
1.4
1.2
1
4
3
2
1
0
0.8
0
2
4
6
8
t / h
Figure 3. Evolution of AMX and carbonate concentrations during the electrolysis of
AMX at Ru/CNT with an applied potential of 1.5 V vs. SCE.
10

The presence of the low molecular weight carboxylic acids among the oxidation products
was confirmed by GC-MS analysis of their trimethylsilyl derivatives, TMS (see Figure
S1 in Supporting Informations). GC-MS analysis besides acetic, propanoic, glyceric, 2-
hydroxy butanoic, glycolic, oxalic, malonic, maleic, 2,3-hydroxypropanoic, malic, 2-
propenoic, tartaric, 4-hydroxybenzoic, mercapto-acetic (m-acetic) acids and glycine,
shows also the presence of indole-dione type compounds.
The distribution of the degradation products of AMX for the Pt/CNT and Ru/CNT
modified electrodes determined by IC and HPLC-UV-RID is given in Table 1. The
presence of some of these compounds was also confirmed by GC-MS using NIST and
Wiley spectral libraries. The amounts of low molecular weight carboxylic acids (C1-C6)
correspond respectively to 35.6 and 43 % of the oxidized AMX for Pt/CNT and Ru/CNT
electrodes, while the percentages of unidentified and/or unquantified products were 44
and 27.
The HPLC-MS chromatograms obtained with RP-18 column for AMX solution of control
experiment carried out in absence of applied potential and the electrolyzed solutions of
AMX in NaOH medium for Pt/CNT and Ru/CNT modified electrodes are presented in
Figure 4. A significant decrease of AMX, amoxicillin penicilloic acid (APcA) and
amoxicillin penilloic acid (APiA) peak areas was noticed after the electrolysis of AMX
at 1.5 V vs. SCE during 8 h for both metal carbon nanotube modified electrodes.
11

30000000
25000000
20000000
15000000
10000000
5000000
0
3500000
3000000
2500000
2000000
1500000
1000000
500000
0
a
338;382;418
b
374
340
97;179;195;261
384
148;165;260;282;287
219;263;364
99;122;186;200;
217;240;260
115:196;261
231;289
302;362
176;201
189;210;419
187;219;417
247
202;215;273;309;333
192
204; 275
354
245;267
233
0
5
10
15
20
0
5
10
15
20
t / min
t / min
Figure 4. HPLC-MS chromatograms in full scan mode for () AMX solution from
control experiment, (), oxidized solutions of AMX at Pt/CNT and () at Ru/CNT
modified electrodes in 0.1 M NaOH medium at the end of the electrolysis (t = 8 h) in a)
negative and b) positive ionization mode, (m/z values for each ion are associated to the
respective chromatographic peak).
Electrolysis products were identified using ESI-MS/MS in both negative and positive
ionization modes. Some of these products and the degradation pathways are presented in
Figure 5. Main products of the reactions provide ions with m/z 189, 261 and 231 and 167
for both Pt/CNT and Ru/CNT modified electrodes. Several degradation pathways can be
considered including hydroxylation, four-membered -lactam ring opening
(epimerization) and AMX diketopiperazine formation [9]. The hydroxylation of AMX
occurs at the positions more susceptible to an electrophilic attack like the aromatic ring
or the nitrogen atom. The ions with m/z 354 and 176 confirm the oxidation of the methyl
group in the thiazolidine ring. The ESI-MS-MS spectrum, which shows fragments at m/z
159 (C6H11N2OS) and 130 (C5H8NOS), is in agreement with this suggestion.
12

NH2
S
NH2
H
N
H
N
Epimerization
N
H
O
N
O
O
HO
O
HO
OH
AMX
APcA
m/z 384
m/z 366
NH2
O
H
N
H
N
N
H
N
H
H
N
O
NH2
O
OH
HO
O
S
O
HO
N
m/z 323
N
m/z 340
HO
O
NH2
O
COOH
H
N
m/z 211
OH
m/z 261
N
H
HO
OH
N
O
N
N
S
H
N
S
m/z 312
HO
HO
O
N
N
S
OH
H
N
HO
O
O
COOH
COOH
m/z 189
m/z 305
m/z 263
O
N
O
HO
OH
OH
NH
NH
m/z 383
O
H
N
HO
OH
HO
S
NH2
O
H
N
N
HO
HO
HO
m/z 197
N
O
N
O
OH
OH
O
N
NH
O
O
HO
OH
NH2
m/z 295
m/z 382
O
m/z 375 (352+23)
HO
OH
N
H O
m/z 181
O
S
H
OH
N
S
S
NO
O
HO
N
HO
OH
N
N
O
N
O
O
O
O
COOH
COOH
HO
HO
OH
m/z 233
m/z 277
m/z 167 (esi (+)-ms)
m/z 165 (esi (-)-ms)
m/z 428
COOH
S
S
S
H2N
O
N
O
H2N
HO
m/z 139 (esi (+)-ms)
m/z 137 (esi (-)-ms)
N
N
N
O
O
COOH
COOH
m/z 231
m/z 247
m/z 217
OH
S
OH
OH
S
HO
OH
S
S
O
COOH
S
HOOC
O
O
N
HN
N
O
O
COOH
COOH
m/z 123 (esi (+)-ms)
m/z 121 (esi (-)-ms)
COOH
COOH
COOH
m/z 197
m/z 178
m/z 190
m/z 160
m/z 130
m/z 227
m/z 176
H O
O
H O
H O
O
O
O
O
O
O
OH
HO
OH
O
H2N
H3C
HO
H
m/z 47
m/z 61
m/z 90
m/z 91
m/z 117
Figure 5. The main products of AMX electrolysis in 0.1 M NaOH medium.
The ESI-MS/MS spectra of some degradation products and the main fragments are given
in Figure S2 (Supporting Information).
It can be concluded that the presence of metal ions is crucial for both oxidation and
mineralization of AMX in 0.1 M NaOH medium. Higher conversion was found for
13

Pt/CNT modified electrode while higher mineralization rate was obtained with Ru/CNT
modified electrode. This result can be explained by the ability of the Ru electrode for an
exhaustive oxidation in comparison with Pt/CNT, which may need a promotor for the
adsorption of the oxygenated species in a higher extent on the electrode surface in order
to proceed with a complete oxidation.
The outcome data also indicate that the interaction between the carbon atom of the amide
group, localized between the phenolic and thiazolidine ring containing groups, and
surface OH species, could be important concerning the approach of the molecule. This is
one of the carbon atoms most susceptible for a nucleophilic attack from hydrous surface
species, since the carboxylic groups of APcA and the OH group of phenol, with negative
charges due to the deprotonation reactions in this pH medium, make more difficult an
approach of the molecule from these groups to the negatively charged electrode surface.
Most reaction products identified by HPLC-MS and ESI-MS-MS, containing thiazolidine
ring or phenol group separately are in agreement with this suggestion, showing that the
initial bond cleavage of the molecule most probably occurs between these two moieties.
For the Pt/CNT modified electrode higher amounts of oxidation products without a
significant degradation of the molecule were found in comparison with Ru/CNT modified
electrode, while the amount of molecular backbone chain cleavage products like phenol
hydroxypyrazine (BHP, Mw =188 g mol-1) were lower than what was found for Ru/CNT.
For the oxidative degradation initiated by the hydroxylation of the aromatic ring on Ru
modified electrode, the following reaction pathways can be envisaged, with the redox
couple RuOx+1/RuOx acting probably as mediator:-
14

R
R
Hydroxylation of
benzene ring
HO
HO
HO
(OH)n
(n= 1-> 4)
R
Mono and di-carboxylic acids (C1-C6) + CO2 + ROH
CO2 + H2O
(OH)n
For the Pt electrode, a surface reaction between adsorbed AMX and Pt-OH species can
be foreseen:
Pt + AMX  Pt_(AMX)ads
Pt_(AMX)ads + Pt-OH  products +2Pt
3.2 Electrolysis of AMX in 0.1 M NaCl medium
The electrochemical oxidation of AMX was carried out in 0.1 M NaCl medium on CNT,
Pt/CNT and Ru/CNT modified electrodes (Figure 6). In NaCl medium, the oxidation of
the chloride ions provides active chlorine species like hypochlorite ion which as strong
oxidants can provide further oxidation of organic compounds.
60
40
20
0
15
10
5
15
10
5
16
12
8
b
a
0
4
-5
0
0
-20
-40
-60
-4
-8
-12
-10
-15
-20
-5
-10
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
E / V vs. SCE
E / V vs. SCE
15

40
20
0
15
10
5
c
0
-20
-40
-5
-10
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
E / V vs. SCE
Figure 6. Cyclic voltammograms of a) CNT, b) Ru/CNT and c) Pt/CNT () in supporting
electrolyte alone, () at the beginning and () at the end of the electrolysis of 2 mM
AMX.
Cyclic voltammograms at the beginning and at the end of the electrolysis for CNT,
Pt/CNT and Ru/CNT show that the oxidation of AMX starts from -0.5 V vs SCE at CNT
and Ru/CNT and at -0.2 V vs. SCE at Pt/CNT, where the electrode surface becomes
partially covered by OH species. Higher current intensities for AMX oxidation were
noticed at the beginning of the electrolysis in comparison with NaOH medium.
Considering the point of zero charge of CNT, a positively charged electrode surface is
expected in this case contrarily to that found for 0.1 M NaOH medium. The approach of
the AMX molecule from the carboxylate anion is probably favored in this case.
For Pt/CNT modified electrode a decrease of the hydrogen adsorption/desorption region
was noticed in presence of AMX, since the adsorption of this compound competes with
hydrogen adsorption and consequently decreases the available sites for this purpose. The
shape of the cyclic voltammograms obtained after the electrolysis indicates an
electrode/electrolyte interface with high resistivity. These results show that the electrode
surface at the end of the electrolysis could be deactivated due to the formation of a passive
16

film. It should be emphasized that one of the reasons for the surface properties change
could be the pH drop noticed during the electrolysis due to the consumption of OH ions,
since the oxidation of water is one of the possible reactions occurring at these anodic
potentials. The decrease of pH, especially near to the electrode surface, may favor
dimerization reactions of some oxidation products containing the phenol group, and
consequently providing strongly adsorbed compounds that could prevent the adsorption
of oxygenated species on the electrode surface.
The products of the electrolysis were determined using coupling chromatographic
techniques (HPLC-MS, GC-MS) and quantified using HPLC-UV and IC. The distribution
of the products for different electrodes is given in Table S1 (Supporting Information).
The electrolyses of AMX carried out in 0.1 M NaCl provided 62, 64 and 54 % of
conversion for CNT, Pt/CNT and Ru/CNT modified electrodes, respectively; while the
mineralization efficiency for chemically bonded carbon, calculated by TOC analysis and
given as a percentage of the oxidized AMX were respectively 3, 22 and 19 %, for these
three modified electrodes. The selectivity towards sulphate, nitrate and ammonium ions
were respectively 10, 4 and 6 % for Ru/CNT; 12, 5.4 and 7 % for Pt/CNT and 8.6, 4.3
and 10 % for CNT. The percentages of low molecular weight carboxylic acids (C1-C6)
among the oxidation products were 17.5 and 12 % for Pt/CNT and Ru/CNT electrodes
respectively, while the percentages of unidentified and/or unquantified products were
60.5 and 65 % for these anodes, respectively. The higher amounts of these products in
NaCl medium can be explained by the formation of some chlorinated compounds. Higher
conversion of AMX on the CNT modified electrode can be explained by a stronger
interaction between AMX and the electrode surface than those expected for Pt/CNT and
Ru/CNT modified electrodes. The presence of metal favors the adsorption of oxygenated
species on electrode surface, especially in the case of Ru/CNT modified electrode. On the
17

other hand, the mineralization of AMX occurs at much lower extent on CNT electrode
than those found for M/CNT electrodes. These results show that for a more efficient
mineralization, adequate amounts of both organic species and oxygenated species on
electrode surface are necessary.
UV/vis spectroscopy study shows that at the end of the electrolysis, the decrease of the
absorption in the 220-260 nm region was accompanied by the appearance of a new
absorption band between 320-380 nm (see Figure S4 in the supporting information). This
new absorption band is attributed to the presence of highly conjugated compounds
including phenol hydroxypyrazine (PHp) [5]. The presence of PHp was also confirmed
by HPLC-MS and ESI-MS/MS. But the relative amount of this last product seems to be
lower in this case than that found in 0.1 M NaOH medium, indicating that other highly
conjugated products with similar structure should be responsible for the increase of
absorbance in the range between 320-380 nm.
The HPLC-MS chromatograms of the electrolyzed solutions of AMX in 0.1 M NaCl are
given in Figure S3 (Supporting Information). The molecular structures of products
identified by GC-MS, HPLC-MS and ESI-MS/MS are shown in Figure 7. These results
indicate larger amounts of higher molecular weight oxidation products, like ions with m/z
of 412, 419, 445, than those obtained in 0.1 M NaOH medium.
OH
S
NH
HO
N
N
O
O
HO
OH
m/z 412
18

H
N
NH2
O
S
HN
OH
O
HO
COOH
Cl
m/z 419
Figure 7. Some of the high molecular weight products of electrolysis of AMX in 0.1 M
NaCl medium, identified by HPLC-MS and ESI-MS/MS
3.3 Oxidation of AMX in 0.1 M carbonate buffer medium
The electrolysis of AMX in carbonate buffer medium was also carried out on CNT,
Pt/CNT and Ru/CNT electrodes, in order to evaluate the pH effect on AMX conversion
and product distribution. Figure 8 displays the cyclic voltammograms of CNT, Pt/CNT
and Ru/CNT at the beginning and at the end of the electrolysis of AMX, in carbonate
buffer medium.
30
20
10
0
80
60
40
20
0
60
40
20
0
b
a
-20
-40
-20
-40
-60
-10
-20
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
E V vs. SCE
E / V vs. SCE
19

40
20
0
c
-20
-40
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
E vs. SCE
Figure 8. Cyclic voltammograms of a) CNT, b) Ru/CNT and c) Pt/CNT modified
electrodes in NaHCO3/Na2CO3 carbonate buffer (), at the beginning () and at the end
() of electrolysis of 2 mM AMX.
The current intensities corresponding to the cyclic voltammograms at the end of the
electrolysis, especially for the potential range between -0.5 and 1 V vs. SCE, are higher
than those noticed at the beginning. This result could be due to two factors: (i) the
rearrangement of the electrode surface due to the oxidative treatment during the
electrolysis, which can increase the porosity of the carbon material and consequently the
active surface area, and, (ii) the production of species with higher electroreactivity
comparing to AMX.
The products of the electrolysis were identified by coupling techniques (GC-MS, HPLC-
MS) and quantified by IC and HPLC-UV-RID (Table S2, Supporting Information). The
electrolyzed solution corresponding to the oxidation in carbonate buffer was analyzed by
HPLC-(ESI-)-MS.
The product ion with m/z 97 obtained by negative ionization mode is in a much higher
extent than that found for the oxidation in NaOH medium. This ion is attributed to the
pentandienoic acid (CH2=CH-CH=CH-COOH) and the ESI-MS/MS spectrum of this ion
confirms this hypothesis. The best results concerning the mineralization of carbon content
20

of AMX were obtained with Pt/CNT modified electrode and correspond to 30 % of the
oxidized AMX. This value falls to 10 % for Ru/CNT modified electrode. Considering the
sulfur and nitrogen content of AMX, the selectivities towards sulphate, nitrate and
ammonium ions were 14, 3 and 9 % for Ru/CNT; 28, 6 and 11 % for Pt/CNT. The
degradation rate of AMX on Pt/CNT modified electrode is analogous to those obtained
with the widely used anode materials: Pt, DSA and BDD. The mineralization rate is
similar to that found for Pt (36 %) and DSA (41 %) and lower than that corresponding to
Ti4O7 (69 %) anodes [15]. BDD anode [15, 34] allows the highest mineralization rate (94
%). However, the high cost of BDD electrode and scarcity of suitable substrate limit its
large-scale application [32]. This mineralization rate, still modest, could be enhanced
through the modification of M/CNT ratio and/or CNT’s surface properties aiming at an
optimum coverage of adsorbed AMX and oxygenated species for an efficient
mineralization, since a lack of adequate coverage for these species probably leads to an
incomplete oxidation. The possibility of the modification of surface properties in order to
improve efficient oxidation and mineralization constitutes an advantage of an active
anode towards a passive one
In carbonate buffer medium, the amounts of low molecular weight carboxylic acids
correspond respectively to 42.6 and 31.8 % of the oxidized AMX for Pt/CNT and
Ru/CNT electrodes, while the percentages of unidentified and/or unquantified products
were 27.4 (Pt/CNT) and 58.2 (Ru/CNT).
The HPLC-ESI(-)-MS chromatograms of electrolyzed solution of AMX on Pt/CNT,
Ru/CNT and a blank solution of AMX are present in Figure S5 (Supporting Information).
A significant amount of low molecular weight compounds with very low retention times
was observed in carbonate buffer medium for both Pt/CNT and Ru/CNT modified
electrodes in comparison with the blank solution of AMX. On the other hand, a significant
21

decrease of the peak eluted at 11 min, attributed to AMX, was observed for electrolyzed
solutions.
Significant amount of low molecular weight compounds with low retention times was
formed in in carbonate buffer medium, in comparison with NaOH and NaCl media. The
HPLC-MS chromatograms in negative ionization mode obtained for Pt/CNT and Ru/CNT
modified electrodes in 0.1 M NaOH and in carbonate buffer (pH 10) are superposed in
Figure S5, in order to distinguish more easily the differences between these media.
It can be concluded from these chromatograms, that a higher amount of the low molecular
weight products with lower retention time were formed in carbonate buffer medium in
comparison with 0.1 M NaOH medium (see Figure S5). On the contrary, the amount of
products with higher molecular weight was decreased by a decrease in pH value from 13
to 10. These results indicate that the oxidative degradation of AMX takes place in a higher
extent in carbonate buffer medium than that observed in 0.1 M NaOH medium. This can
be explained by a much easier approach and subsequent adsorption of the electroactive
substrate on the electrode surface in this first pH medium, due to the decrease of the
concentration of species with multiple negative charges, in comparison with 0.1 M NaOH
medium, since the negatively charged electrode surface exerts less repulsion forces
towards the electroactive species in this last case. The mineralization rate of AMX at the
Pt/CNT electrode in carbonate buffer medium is slightly better than that found for
Ru/CNT electrode in 0.1 M NaOH medium. The highest conversion rates found in
carbonate buffer medium are not complemented by highest mineralization efficiencies
indicating that for the further oxidation of low molecular weight degradation products,
like carboxylic acids, the presence of a significant amount of oxygenated species on the
electrode surface could be also necessary. Probably a compromise between available sites
for the adsorption of the organic compound and adsorbed oxygenated species is required
22

for a complete oxidation, which leads to the mineralization of the molecule. It seems that
PHp is one of the products that influences the mineralization efficiency, since the relative
amount of this compound among the oxidation products is inversely proportional to the
decrease of TOC values.
3.4 Response surface methodology
Response surface methodology [37,39] was applied in order to determine the importance
of the electrode material (x1) and pH (x2) on the conversion (y). R software package and
Microcal Origin 7.5 Software were used for this purpose. The data used for the
determination of the parameter’s coefficients is given in Table 2. The codification of the
original values was carried out using levels of -1, 0 and 1 for CNT, Pt/CNT and Ru/CNT
modified electrodes and for pH values of 6, 10 and 13, respectively. The relationships
between the percentage of conversion and two factors: pH and electrode material, can be
defined by following equation:
2
2
Conversion (%) = 95.21+3.14 x1-11.46 x2-11.32 x1 -39.13 x2 + 8.71 x1x2
The R2 and Pr(>F) for regression (FREG) and for Lock of Fit (FLOF) were respectively 0.90,
0.016 and 0.16 indicating that the found polynomial model was suitable. It seems that the
effect of pH on the AMX conversion is higher than that was exerted by electrode material
and the effect of interaction between factors is not significant.
3.5 Principal Component Analysis
Principal Component Analysis (PCA) is one of the most powerful statistical tools of
multivariate data analysis which can be used to classify the samples through a small
number of new and non-correlated variables deduced from the raw data by dimension
reduction [40, 41]. R software package was used for this purpose. The original data matrix
23

was constituted from 6 samples including electrolyzed and blank solutions of AMX and
nearly 50 compounds determined by HPLC-MS analysis as variables. The first and the
second PCs represented 42 and 15 % of the total variance, respectively. The biplot graph
for the first and the second PCs is present in Figure 9. The scores of the first PC allow to
separate the solutions electrolyzed on Pt/CNT and Ru/CNT modified electrodes in 0.1 M
NaOH medium from the others including the blank solutions of AMX.
Figure 9. Biplot graph for a) PC2 vs. PC1 obtained from original matrix composed by
samples of electrolyzed solutions and main products determined by HPLC-MS analysis.
The degradation products which characterized these two solutions can be distinguished
from the others both by number and by relative amount, the former being higher for
Pt/CNT. The second PC allows to separate the electrolysis on Pt/CNT and Ru/CNT
modified electrodes in 0.1 M NaOH medium. The chemical compounds (loadings),
located near to the sample corresponding to the electrolysis on Ru/CNT were relatively
24

low molecular compounds obtained from the cleavage between phenyl hydroxypyrazine
(PHp) and thiazolidine (Thz) moieties, while a significant part of the compounds found
near to the Pt/CNT modified electrode correspond to the oxidation products with higher
molecular weights, which were probably provided by external groups cleavage rather than
the cleavage between these earlier moieties. These results show that the initial
degradation, which yields “PHp” and “Thz” like products is easier on Ru/CNT than that
found on Pt/CNT. The distribution of the loadings taking into account their proximities
to the scores indicates also that the mineralization on Ru/CNT electrode in 0.1 M NaOH
medium seems to be related to the hydroxylation of aromatic ring (m/z 383) and nitrogen
atom (m/z 382 and 159) and both on aromatic ring and nitrogen atom (m/z 413) [9].
PCA confirms also that the mineralization efficiency is related on the other hand to the
cleavage of carboxylic acid and ammonium groups of the APcA. In order to explain the
ions with m/z values of 340, 323, 295 and 277 situated near to the mineralization rate, the
following losses can be considered: 384 (C16H21N3O6S + H+)  340 (C16H21N3O6S + H+
- CO2)  323 (C16H21N3O6S + H+ - CO2 – NH3)  295 (C16H21N3O6S + H+ - CO2 – NH3
– CO2)  277 (C16H21N3O6S + H+ - CO2 – NH3 – CO2 – H2O).
The suggested structure for the other ion with m/z 245 related to the mineralization is
given in Figure 10.
O
H
N
+
N
H
O
Figure 10. The suggested structure for the other ion with m/z 245
The ESI-MS-M spectrum confirms this suggestion. It can be concluded from these results
that the hydroxylation of the aromatic ring and nitrogen atom and the degradation of the
25

molecule into phenolic and thiozalidine (Thz) groups are easier on Ru/CNT modified
electrode than that found for Pt/CNT modified electrode. The loading graph of PC2 vs.
PC1 shows also that the compounds with greater influence on the electrolyzed solution at
Ru/CNT in 0.1 M NaOH medium are inversely proportional with those related to the
blank solutions of AMX in the same medium, since the angle between them corresponds
to approximately 180º, leading to a correlation coefficient of -1. This fact shows that the
products of the electrolysis on Pt/CNT are related directly to AMX and its initial
hydrolysis products, characteristic of blank solutions. On the other hand, the compounds
corresponding to the electrolyzed solutions of AMX on Pt/CNT in the same medium are
less correlated to the initial AMX, ApcA and APiA found in the blank solution indicating
that some other parallel or successive transformations could be also important in this case.
The biplot graph shows also that electrolyzed solutions of AMX in carbonate buffer
medium show chemical compositions similar to the blank solutions and consequently
strengthen the hypothesis that these solutions, besides the remaining AMX and its
hydrolysis products, show small amount of intermediate oxidation products since beside
the mineralization products some unidentified low molecular weight products persist in
this case. It could also be noticed from these results that the electrolysis on Ru/CNT
electrode in 0.1 M NaCl privileges the formation of some high molecular weight products.
4. Conclusions
Among the different pH media studied in this work, the carbonate buffer medium
provides the higher AMX conversions concerning all the modified electrodes used in this
study. These encouraging results show that a conversion of 83 % was achieved on the
carbon nanotube supported platinum electrode in this medium and 30 % of the oxidized
26

AMX was transformed into CO2. These results can be explained by the fact that hydrous
platinum (PtOH) species can act as efficient catalyst for the oxidation of AMX in this
medium since the approach of AMX molecule could be easier considering that the relative
amount of AMX with multiple negative charges is lower that found in 0.1 M NaOH. The
results also show that the change of pH from 10 to 13 has a more important effect on the
mineralization of AMX for the Ru/CNT modified electrode than that observed for
Pt/CNT. This can be explained by the fact that, in alkaline media, the efficient
mineralization of AMX on Ru/CNT needs a higher coverage of RuOX on the electrode
surface, leading to the transfer of active oxygen between these species and the AMX
molecule, while the surface reaction between AMX and adsorbed oxygenated species can
be envisaged in the case of the Pt/CNT electrode. In NaCl medium, besides a high
conversion of the initial substrate into oxidation products, low mineralization values were
observed. The presence of halogenated and high molecular weight compounds was also
noticed in this case.
Acknowledgements
The authors thank FCT (Fundaꢀꢁo para a Ciꢂncia e a Tecnologia) for the PhD grant of
Marta Ferreira. This work has been developed under the scope of the projects:
BioTecNorte (operation NORTE-01-0145-FEDER-000004) and “AIProcMat@N2020-
Advanced Industrial Processes and Materials for a Sustainable Northern Region of
Portugal 2020”, NORTE-01-0145-FEDER-000006, supported by the Northern Portugal
Regional Operational Programme (NORTE 2020), under the Portugal 2020 Partnership
Agreement, through the European Regional Development Fund (ERDF). This work also
has been funded by national funds (FCT/MCTES (PIDDAC)) through the projects:
27

PTDC/AAGTEC/5269/2014, Centre of Chemistry (UID/QUI/00686/2013 and
UID/QUI/0686/2016) and Associate Laboratory LSRE-LCM - UID/EQU/50020/2019.
References:
[1]. J.C.G. Sousa, A.R. Ribeiro, M.O. Barbosa, M.F.R. Pereira, A.M.T. Silva, J. Hazard.
Mater. 344 (2018) 146-162.
[2]. D. Shahidi, R. Roy, A. Azzouz, Appl. Catal. B: Environm. 174 (2015) 277-292.
[3]. B. Silva, F. Costa, I.C. Neves, T. Tavares, Chap. 1- Introduction. In Psychiatric
Pharmaceuticals as Emerging Contaminants in Wastewater, Springer Book.
SpringerBriefs in Green Chemistry for Sustainability, 2015, Vol. III, pp. 1-18. 4.
[4]. D. Fatta-Kassinos, S. Meric, A. Nikolaou, Anal. Bioanal. Chem. 399 (2011) 251-275.
[5]. I. Gozlan, A Rotstein, D. Avisar, Chemosphere. 91 (2013) 985.
[6]. A.F. Martins, F. Mayer, E.C. Confortin, C.D. Frank, Wastewater. Clean. 37 (2009)
365–371.
[7]. D. Klauson, J. Babkina, K. Stepanova, M. Krichevskaya, S. Preis, Catal. Today. 151
(2010) 39-45.
[8]. D. Kanakaraju, J. Kockler, C.A. Motti, B.D. Glass, M. Oelgemoller, Appl. Catal. B:
Environm. 166-167 (2015) 45-55.
[9]. A.G. Trovó, R.F.P. Nogueira, A. Agüera, A.R. Fernandez-Alba, S. Malato, Water.
Res. 45 (2011) 1394-1402.
[10]. E.S. Elmolla. M. Chaudhuri, J. Hazard. Mater. 172 (2009) 1476-1481.
[11]. F.J. Benitez, J.L. Acero, F.J. Real, G. Roldán, Chemosphere 77 (2009) 53-59.
[12]. Y.J. Jung, W.G. Kim, Y. Yoon, J.-W. Kang, Y.M. Hong, H.W. Kim, Sci. Total
Environ. 420 (2012) 160-167.
[13]. M. Magureanu, D. Piroi, N.B. Mandache, V. David, A. Medvedovici, C. Bradu, V.I.
Parvulescu, Water Res. 45 (2011) 3407-3416.
28

[14]. G. Chen. Sep. Purif. Technol. 38 (2004) 11-41.
[15]. S. Ganiyu, N. Oturan, S. Raffy, M. Cretin, R. Esmilaire, E. Van. Hullebusch, G.
Esposito, M.A. Oturan, Water Res. 106 (2016) 171-182.
[16]. A.S. Kumar, S. Sornambikai, L. Deepika, J.-M. Zen, J. Matter. Chem. 20 (2010)
10152-10158.
[17]. Z. Frontistis, M. Antonopoulou, D. Venieri, I. Konstantinou, D. Mantzavinos, J.
Environ. Manage. 195 (2017) 100-109.
[18]. M. Ferreira, M.F. Pinto, O.S.G.P. Soares, M.F.R. Pereira, J.J.M. Órfão, J.L.
Figueiredo, I.C. Neves, A.M. Fonseca, P. Parpot, Electrochim. Acta 60 (2012) 278-
286.
[19]. C.H. Lam, A.J. Bloomfield, P.T. Anastas, Green Chem. 19 (2017) 1958-1968.
[20] C.A. Martínez-Huitle, S. Ferro, Chem. Soc. Rev. 35 (2006) 1324-1340.
[21]. S. Trasatti, Int. J. Hyd. Ener. 20 (1995) 835-844.
[22]. M. Ferreira, M.F. Pinto, I.C. Neves, A.M. Fonseca, O.S.G.P. Soares, J.J.M. Órfão,
M.F.R. Pereira, J.L. Figueiredo, P. Parpot. Chem. Eng. J. 260 (2015) 309-315.
[23]. M. Ferreira, M.F. Pinto, O.S.G.P. Soares, M.F.R. Pereira, J.J.M. Órfão, J.L.
Figueiredo, I.C. Neves, A.M. Fonseca, P. Parpot, Chem. Eng. J. 228 (2013) 374-380.
[24]. O. Scialdone, S. Randazzo, A. Galia, G. Silvestri, Water Res. 43 (2009) 2260-2272.
[25]. O. Scialdone, Electrochim. Acta. 54 (2009) 6140-6147.
[26]. C. Comninellis, Electrochim. Acta. 39 (1994) 1857-1862.
[27]. C. A. Martínez-Huitle, M. A. Rodrigo, I. Sires, O. Scialdone, Chem. Rev. 115 (2015)
13362-13407.
[28]. A.N.S. Rao, V. T. Venkatarangaiah, Environ. Sci. Pollut. Res. 21 (2014) 3197-3217.
[29]. I. Sirés, E. Brillas, M. A. Oturan M. A. Rodrigo, M. Panizza, Environ. Sci. Pollut.
Res. 21 (2014) 8336-8367.
29