Electrochemical detection of nitric oxide and peroxynitrite anion in microchannels at highly sensitive platinum-black coated electrodes. application to ROS and RNS mixtures prior to biological investigations

Article abstract of DOI:10.1016/j.electacta.2014.08.046

The electrochemical detection of nitric oxide (NO) and peroxynitrite anion (ONOO^-) was investigated at Pt-black electrodes in microchannels. Owing to the high reactivity of these species under conditions close to physiological media, kinetic parameters were determined before evaluating the detection performances from synthetic solutions. Highly sensitive and stable Pt-black electrodes allowed detection limits down to 30 nM (NO) and 40 nM (ONOO^-) to be reached with very high sensitivities. As NO and ONOO^-are two relevant biological molecules involved in oxidative stress their simultaneous detections with other major molecules like hydrogen peroxide and nitrite were also performed and validated experimentally at pH 8.4. These results demonstrated that relative ROS/RNS contents in synthetic mixtures can be easily assessed at selected detection potentials. Beyond the interest of using small volumes in microfluidic channels, optimization of detection requires precise conditions easy to implement. These were delineated to lead in microdevices to high-performances detection of oxidative stress metabolites either from analytes or from the production of a few living cells like macrophages.

Full text of DOI:10.1016/j.electacta.2014.08.046

Electrochimica Acta 144 (2014) 111–118
Contents lists available at ScienceDirect
Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta
Electrochemical Detection of Nitric Oxide and Peroxynitrite Anion in
Microchannels at Highly Sensitive Platinum-Black Coated Electrodes.
Application to ROS and RNS Mixtures prior to Biological Investigations
Yun Li, Catherine Sella, Frédéric Lemaître, Manon Guille-Collignon, Laurent Thouin∗,
Christian Amatore∗
Ecole Normale Supérieure-PSL Research University, Département de Chimie, Sorbonne Universités - UPMC Univ Paris 06, CNRS UMR 8640 PASTEUR, 24, rue
Lhomond, 75005 Paris, France
a r t i c l e i n f o
a b s t r a c t
The electrochemical detection of nitric oxide (NO•) and peroxynitrite anion (ONOO−) was investigated at
Pt-black electrodes in microchannels. Owing to the high reactivity of these species under conditions close
to physiological media, kinetic parameters were determined before evaluating the detection perform-
ances from synthetic solutions. Highly sensitive and stable Pt-black electrodes allowed detection limits
down to 30 nM (NO•) and 40 nM (ONOO−) to be reached with very high sensitivities. As NO• and ONOO−
are two relevant biological molecules involved in oxidative stress their simultaneous detections with
other major molecules like hydrogen peroxide and nitrite were also performed and validated experi-
mentally at pH 8.4. These results demonstrated that relative ROS/RNS contents in synthetic mixtures
can be easily assessed at selected detection potentials. Beyond the interest of using small volumes in
microfluidic channels, optimization of detection requires precise conditions easy to implement. These
were delineated to lead in microdevices to high-performances detection of oxidative stress metabolites
either from analytes or from the production of a few living cells like macrophages.
Article history:
Received 16 May 2014
Received in revised form 2 August 2014
Accepted 5 August 2014
Available online 26 August 2014
Keywords:
RNS
ROS
electrochemical detection
microchannel electrode
Pt-black electrode
© 2014 Elsevier Ltd. All rights reserved.
superoxide anion O2 and NO• are produced in a restricted cell
•−
1. Introduction
environment, they couple at a rate close to the diffusion limit lead-
ing to the production of peroxynitrite anion ONOO− [5]. ONOO− is
reported as a powerful oxidant and cytotoxic agent. It is the source
of many oxidative, nitrating and nitrosating processes involving
several cell components [1,6].
Oxidative stress is a process involved in many physiological
and pathological situations (cancers, neurodegenerative diseases,
autoimmune pathologies). Over the last decades, broad researches
have been devoted to the study of the different phases of this pro-
cess. Recent results have demonstrated the ubiquitous imbrication
in metabolism of reactive oxygen species (ROS) with that of nitric
oxide (nitrogen monoxide, NO•) [1]. The exact balance between
these species and products controls many cell functions. They are
also meaningful risk markers in case of traumatic injuries [2]. How-
ever, the simultaneous production of superoxide and nitric oxide by
living cells leads to a series of complex and interconnected path-
ways in oxidative stress [3]. Effects of nitric oxide in organisms
depend on its local concentration but also on specific conditions
prevailing in vivo [4]. Nitric oxide is a weak oxidant by itself but it
reacts extremely fast with paramagnetic species. Therefore, when
The challenging development of analytical methods for moni-
toring the production of nitric oxide and peroxynitrite is therefore
highly warranted [7,8]. This is key to better understanding the
role played by these species in human health care, heavy dis-
eases and pathologies. Among existing methods, fluorescence and
chemiluminescence measurements allow extracellular high per-
formance detection of nitric oxide [9–15] and peroxynitrite [16,17].
However, in comparison, electroanalytical methods remain advan-
tageous due to their unparalleled spatial and temporal resolutions
[18]. Indeed, despite inherent difficulties, micro-electrochemistry
provides powerful tools to estimate concentration levels very close
to the location from which species are produced. Electroanalytical
techniques are also the only ones able to determine intracellu-
lar [19] and extracellular [3,20,21] minute fluxes of ROS/RNS in
real-time without disturbing the metabolism of living cells. They
are label-free since nitric oxide and peroxynitrite can be detected
∗
Corresponding authors.
E-mail addresses: laurent.thouin@ens.fr (L. Thouin), christian.amatore@ens.fr
(C. Amatore).
http://dx.doi.org/10.1016/j.electacta.2014.08.046
0013-4686/© 2014 Elsevier Ltd. All rights reserved.

112
Y. Li et al. / Electrochimica Acta 144 (2014) 111–118
electrochemically at specific potentials [8,22–29]. This contrasts
with the poor selectivity of many redox fluorescent markers [17].
Both species oxidize at relatively low potentials [30] even if their
detections are more demanding due to their short half-life under
physiological conditions[22,24,26,31,32]. Insomepreviousstudies,
we demonstrated that microelectrodes could be used to moni-
tor extracellular fluxes of nitric oxide and peroxynitrite released
from single activated fibroblasts [33–35] or macrophages [36–39].
Under these experimental conditions, the excellent sensitivity and
temporal resolution of the measurements resulted from the very
short distance separating the microelectrode surface from single
cells. However, the performances of such individual measurements
were in practice constrained by cellular variability. Repetitive series
of measurements were thus required to obtain reliable statisti-
cal information. In this respect, microelectrodes integrated into
microfluidic devices should facilitate the whole procedure if detec-
tion of extracellular fluxes of ROS and RNS is performed at once
on a collection of cells gathered in a microchamber [36,40,41]. This
approach would meet the two necessary requirements for detec-
tion of highly reactive species, i.e., small sample volumes and short
analysis times.
Detection of ROS and RNS in microsystems remains a challenge.
Under those circumstances, the pH employed, mixing dynamics
and composition of the media are important [42]. Electrochemi-
cal performances are also often limited by the ensuing poisoning
of metallic electrodes such as Pt electrodes [43]. Convective flow in
microsystems is expected thus to minimize these inhibitions and
increase at the same time the detection sensitivity. In this context,
we demonstrated recently that the simultaneous electrochemical
detection of hydrogen peroxide (H2O2) and nitrite (NO2−) could be
performed in microfluidic channels at highly sensitive platinum-
black coated platinum electrodes [44]. In the following work, we
wish to report an extension of this approach to the case of the
other two important reactive species, nitric oxide NO• and per-
oxynitrite anion ONOO−. One problem here is the kinetic stability
of these two species when the detection is not preformed closely
above a single cell (viz., at a few ␮m) but in a microfluidic channel.
Taking benefit of optimized Pt-black deposits, the electrochemical
detections were thus studied using calibrated synthetic solutions
of NO• and ONOO−. In association with H2O2 and NO2-, conditions
were further delineated to reach high-performances in microflu-
idic channels for detecting the four primary ROS and RNS in the
mixtures.
Fig. 1. (A) Voltammograms in a continuous flowing solution of 0.1 mM ONOO− at
bare-Pt (dashed line) and Pt-black electrodes (solid line) at pH 11.1. (B) Current
responses of ONOO− microinjections at bare-Pt electrode before (dashed line) and
after (solid line) electrode activation. (C) From top to bottom, current responses of
repetitive ONOO− microinjections at bare-Pt electrode after electrode activation. (D)
Current responses of ONOO− microinjections at Pt-black electrode before (dashed
line) and after (solid line) electrode activation. In (B-D) Microinjections of 0.1 mM
ONOO− solution at pH 11.1 and E = 0.45 V/REF. x-axis refers to time after solution
preparation.
aerated PBS buffer (pH 7.4) to provide DEA-NONOate tested
solutions. Hydrogen peroxide and nitrite solutions were freshly
prepared in PBS from stock solutions (H2O2: 35 wt %, Acros;
NO2-:1 M from 99.99% NaNO2, Aldrich) before experiments. Since
peroxynitrite and DEA-NONOate solutions decompose with time,
the duration t0 between the solution preparation and beginning of
experiments was recorded and reported systematically in kinetic
studies.
Pt-black deposition solutions were prepared from 1 mL
hydrogen hexachloroplatinate (IV) solution (8% wt in water; Sigma-
Aldrich) and 1.6 mg lead (II) acetate trihydrate (99.8%; Sigma)
added into 6.36 mL PBS. Lead ions were required as an impurity to
initiate and control the Pt-black growth during electrodeposition
[47].
2. Experimental section
2.1. Materials and solutions
2.2. UV-Vis spectra
Phosophate-buffered saline (PBS) was use as buffer for most
experiments. PBS was prepared from tablets (Life Technologies)
dissolved into pure water and composed of 10 mM Na2HPO4,
2.68 mM KCl and 140 mM NaCl (pH 7.4). Water was obtained from
a Milli-Q purification system (resistivity 18.2 Mꢀ cm at 25 ◦C; Mil-
lipore). Peroxynitrite solutions were prepared from dilution of
40 mM alkaline sodium peroxynitrite solution (NaONOO; Cayman
Chemicals) with a mixture of 0.01 M PBS and 0.1 M NaOH so as
to reach the required pH, between 11.1 and 8.4. This pH was
systematically controlled with a pH meter (Radiometer Analyti-
cal). Stock solutions of 50 mM DEA-NONOate [45,46] were initially
prepared by dissolving 10 mg of solid DEA-NONOate (Cayman
Chemicals) in 0.96 mL of 0.01 M NaOH. Solid DEA-NONOate was
previously stored under nitrogen at -80 ◦C to avoid any decomposi-
tion. Similarly, stock solutions were stored at 0 ◦C after preparation
for 24 h maximum before use. To initiate the release of NO•,
minute volumes of a stock solution were rapidly mixed with
Absorbance of peroxynitrite and DEA-NONOate solutions was
monitored at room temperature (22 ◦C) using a UV-Vis spectropho-
tometer (Lambda 45; Perkin Elmer) at ꢁ = 302 nm and 250 nm for
peroxynitrite and DEA-NONOate solutions respectively.
2.3. Microdevices
Design and fabrication of microdevices were as reported in our
previous work (see Fig. 1 in [44]). The upper layer comprised three
independent parallel channels made by casting polydimethylsilox-
ane (PDMS, RTV-615; Momentive Performance Materials) onto a
20 ␮m-thick SU-8 2015 photoresist (Microchem) patterned mold.
Reservoirs and injection channels were punched in the PDMS
layer after it was peeled off. The under layer was a glass sub-
strate on which several platinum microbands (working electrodes,

Y. Li et al. / Electrochimica Acta 144 (2014) 111–118
113
WEs; 197 ␮m width) were patterned by soft lithography and
deposited (Ti/Pt with 8 nm/25 nm thickness) using a sputtering
coater (K675XD; Emitech). Some WEs were further platinized upon
reduction of Pt (IV) from deposition solutions at a constant current
density (-8 mAcm−2). A silver microband was partly oxidized to be
used as a reference Ag/AgCl electrode (REF; 500 ␮m width). It was
fabricated by sputtering 50 nm Ag onto the underlying Pt surface
and oxidizing it by 5 mM FeCl3 (Sigma). The counter electrode was a
Pt electrode band (CE; 500 ␮m width) placed downstream the WEs.
The two channel components, PDMS layer and glass substrate, were
then assembled irreversibly by means of oxygen plasma (Harrick),
so that the microchannels were set perpendicular to the microband
electrodes (see Fig. 1 in ref [44]). REF and CE were located respec-
tively upstream and downstream the WEs according to the flow
direction. The effective lengths of all electrodes were limited by
microchannel width (200 ␮m) and the volume of solution above
the microbands was restricted by the microchannel height (20 ␮m).
During each experiment, only one WE was connected and only one
channel was filled with solution, the others remaining empty.
and anion are present at physiological pH. However, the acid is
unstable and rearranges into nitrate:
ONOOH → NO3− + H+
(1)
At higher pH, a decomposition pathway prevails with the pro-
duction of dioxygen and nitrite:
k
ONOOH + ONOO−−→ O2 + 2NO− + H+
(2)
2
This reaction proceeds via the formation of peroxynitric acid as
intermediate during the following sequence [24]:
ONOOH + ONOO− → NO2− + O2NOO− + H+ → 2NO2− + O2 + H+
(3)
However, in most cases of interest here, the overall decomposi-
tion of ONOO− in Eq. (2) is first order and decreases upon increasing
pH [51]. At physiological pH, the decay becomes more complex
and depends on the nature of the buffer [52]. The half-life time of
ONOO− is estimated about 1 s. This makes the ONOO− detection
particularly challenging for the analysis of its production by living
cells or its content in biological samples. Note that under other cir-
cumstances, ONOO− can also react rapidly with carbonyls [53] and
metals [54].
2.4. Electrochemical experiments
All electrochemical experiments were performed versus a home
made Ag/AgCl reference electrode (REF) at room temperature using
a potentiostat (Autolab PGSTAT 30; Eco Chemie) controlled by GPES
4.9 software. The flow was pressure-driven by a syringe pump
(Pump 11 Elite; Harvard Apparatus) at constant average flow rate
of 2 ␮Lmin−1. According to the device geometry, it corresponded to
a laminar flow regime [48]. Before some electrochemical measure-
ments, electrodes were activated by applying a series of alternating
potential pulses (+0.2 V, 1 s and -0.5 V, 1 s; 30 cycles) in flowing PBS.
Analytes were introduced in microchannels in two ways, either as
a continuous flow of solution or via microinjections of solution in a
continuous PBS flow through a peek sample loop (5 ␮L; Rheodyne)
to mimic oxidative bursts. In the first case, voltammograms were
recorded at a scan rate of 10 mVs−1. In both cases, chronoampero-
metric measurements were performed in order to monitor current
variations during the passage of analytes above electrodes or to
monitor current decay with time for kinetic evaluations. Calibration
curves were established from microinjections with various sample
concentrations down to 0.1 ␮M. Detection limits (LOD) were cal-
culated from independent tests (at least 10 tests) with samples in
the nanomolar concentration range. For this evaluation, three times
the standard deviation of the measurements and calibration slope
were considered.
The electrochemical oxidation of ONOO− is a one-electron reac-
tion, whatever the pH [30]:
ONOO− → ONOO• + e−
(4)
Conversely, ONOO− is sufficiently stable in alkaline conditions
with an acceptable extended lifetime [24,32,50] (e.g., 5 hours in
0.3 M NaOH at 25 ◦C). We then performed electrochemical experi-
ments at pH 11.1. Indeed, it was previously demonstrated that the
oxidation wave of ONOO− was independent of pH down to 8.5 [30].
This allows the electrochemical properties of ONOO− to be evalu-
ated from more stable solutions [7,30]. Note that these conditions
do not ensure at this stage any reliability of measurements in bio-
logical environments. Higher pH than 11.1 were avoided to prevent
degradation of Pt-black electrodes. Measurements were carried
out at bare-Pt and Pt-black electrodes fabricated according to the
procedure described in experimental section. Fig. 1A compares
voltammograms obtained in case of a continuous solution flow of
ONOO− at bare-Pt and Pt-black electrodes. In each case peroxyni-
trite started to be oxidized at potentials higher than 0.2 V/REF. At
Pt-black electrodes a steady-state diffusion mass transport regime
was reached above 0.45 V/REF as evidenced by a limiting current
(or current plateau) at high potentials. In contrast, a peak-shaped
voltammogram was observed under same conditions at bare-Pt
electrodes of same surface areas with current intensities two times
lower. This behavior demonstrated that oxidation of ONOO− on
bare-Pt was altered by a more complex pathway involving other
limiting steps and by some competitive poisoning of Pt surface.
This was confirmed by performing ONOO− detection at a constant
potential of 0.45 V/REF after several injections of ONOO− solutions
(Fig. 1B-D). Activation of electrodes, prior to detection, led to a
noticeable increase of current at bare-Pt electrodes (Fig. 1B). In
this case, a flow-removal of residues adsorbed on Pt surface as
well as a selective activation of Pt sites led to an enhancement of
the electrochemical oxidation. However, intensities reached after
activation still remained lower than those recorded at Pt-black
electrodes. Note also that the measurement repeatability was not
satisfactory since current responses for successive ONOO− microin-
jections decreased with time from one to the other (Fig. 1C). On the
contrary, currents recorded at Pt-black electrodes were more sta-
ble and no significant change was observed of current level before
and after activation (Fig. 1D). Such electrode treatment led to more
reproducible results with higher long-term performances as it was
2.5. Theoretical simulations
The differential equation associated to the kinetic of NO• release
(see Eq. (9)) was solved using COMSOL Multiphysics 4.2 software. A
mathematical model based on least squares analysis of experimen-
tal kinetics was used to ensure optimal fit between experimental
and calculated data. This model took into account predictions
resulting from previous simulations of mass transport regimes at
channel microband electrodes [48].
3. Results and discussion
3.1. Peroxynitrite detection
Since peroxynitrite decomposes rather rapidly at physiological
pH, we had to take into account its chemical stability during the fol-
lowing electrochemical investigation. The evaluation of the kinetics
was thus a necessary step before assessing the performances of
electrochemical detection achieved in microchannels. Since the pK
of peroxynitrous acid (ONOOH) is about 6.8 [5,49,50], both the acid

114
Y. Li et al. / Electrochimica Acta 144 (2014) 111–118
Fig. 3. (A) Current responses at Pt-black electrode of ONOO− microinjections with
various concentrations. From top to bottom: 4, 3, 2, 1, 0.5 and 0.1 mM, 50, 10, 5,
1, 0.5 and 0.1 ␮M. x-axis refers to time elapsed after solution preparation. (B) Cur-
rent response of 50 nM ONOO− microinjection. (C-D) Calibration curves of ONOO−
detection at Pt-black electrode. Currents monitored 300 s after solution preparation
and concentrations evaluated for k = 1.2 10−4 s−1. Linear curve fit with slope equal
to 1.04 10−3 AM−1. In (A-D) pH = 11.1 and E = 0.45 V/REF.
Fig. 2. (A) Current intensities monitored at Pt-black electrode in 0.1 mM
ONOO-continuous flowing solutions at several pH. E = 0.45 V/REF and t0 = 130 s. (B)
Repetitive UV-Vis spectra starting from 20 s after solution preparation with 51 s time
intervals of 0.1 mM ONOO− solution at pH 8.4. Absorbance decreases with time. (C)
Variations of UV-Vis peak intensities at several pH of 0.1 mM ONOO− solution at
ꢁ = 302 nm, t0 = 40 s. (D) Comparison between kinetic constants evaluated from elec-
trochemical detection (kEC) and UV-Vis spectra analysis (kUV) at times lower than
200 s after solution preparation. In (A,D) pH = 11.1 (open circle), 10.8 (solid triangle),
10.2 (solid circle), 9.4 (open square), 8.9 (open triangle) and 8.4 (solid square).
detection (LOD) assessed from calibration slope and repetitive
measurements in low concentration range was 40 nM. This limit
agreed with data reported elsewhere [7,32,55]. However, this
limit reflected other limitations than signal-to-noise ratio since
currents reached at 50 nM were still far larger than noise (Fig. 3B).
Note that such LOD does not relate to expected concentrations in
biological systems. It applies to any extracellular concentrations
emitted by a cells population. Their detected values depend on the
experimental conditions including device geometry. Evaluation
of concentrations in microfluidic channels is thus a prerequisite
for the determination of amounts of species released upstream.
The calibration curve allowed the diffusion coefficient of ONOO−
confirmed in case of tests performed with a large number of elec-
trodes (data not shown). From an electroanalytical point of view,
ONOO− detection was by far more efficient on optimized Pt-black
electrodes than on bare-Pt electrodes.
These first results were confirmed by investigating the kinetics
of ONOO− decomposition at lower pH. Electrochemical measure-
ments were achieved at Pt-black electrodes and compared to those
obtained under the same conditions from UV-Vis spectra (Fig. 2).
In both cases ONOO− decomposition was found to be first order
over a range of 200 s after the solution preparation (Fig. 2A and C).
At higher times, the kinetic profile was more complex as signif-
icant deviations from linear logarithmic plots occurred when pH
tended to 8.4. This trend was consistent with other mechanisms
proposed under similar conditions [5]. The evaluation of the first-
order kinetic constant k (Eq. (2)) from both types of measurements
led to very similar results (Fig. 2D). Accordingly, k increased by
almost two orders of magnitude from 10−4 to 10−2.5 s−1 when
pH decreased from 11.1 to 8.4. All these data were also in very good
agreement with other values reported in the literature [5,30,32,50].
These results demonstrated the feasibility and reliability of
ONOO− detection at Pt-black electrodes. Calibration curves were
established from injections of different ONOO− solutions at
pH 11.1 (Fig. 3A). For that purpose, a wide range of concentrations
was investigated after correcting them by taking into account
the kinetics of ONOO− decomposition at this pH. ONOO− con-
centrations were estimated after the time delay separating the
measurements from the solution preparation. Values on current
plateau were also subtracted from residual currents monitored in
the absence of ONOO−. Calibration curves thus obtained displayed
linear relationships with four decades linearity and a sensitivity
of 2.64 AM−1cm−2 (Fig. 3C-D). Conversely, linearity was no longer
observed at concentration below 100 nM (Fig. 3D). The limit of
to be estimated leading to DONOO- = (1.4 0.2) 10−5 cm2s−1
.
According to theoretical steady-state currents predicted under the
hydrodynamic regime of interest in the present microchannels
(see experimental section), this value was close to those usually
considered for RNS in similar conditions [56–59].
3.2. Nitric oxide detection
The performances of NO• detection were evaluated using
dynamic NO• concentrations generated through the decomposi-
tion of DEA-NONOate solutions. Preliminary studies evidenced that
the decomposition rate depended strongly on the concentration
of DEA-NONOate and pH. Therefore, NO• generation rates were
carefully investigated through kinetic evaluations performed under
the conditions used in this study. DEA-NONOate solutions chemi-
cally decompose into NO• (75% yield) when dissolved into aqueous
solutions at physiological pH according to [60]:
k1
Et2N − [N(O)NO]− Na+ + H+−→ 2 NO• + Et2NH + Na+
(5)
with a rate law of DEA-NONOate global decomposition given by:
V1 = k1[Et2N-[N(O)NO]−]
where k1 = 7.22 10−4 s−1 at 22 ◦C at pH 7.4 [61].
(6)

Y. Li et al. / Electrochimica Acta 144 (2014) 111–118
115
Fig. 5. (A) Analysis of UV-Vis peak intensities for three series of measurements
in 0.1 mM DEA-NONOate solution at pH 7.4. t0 = 40 s and ꢁ = 250 nm. Linear curve
fit with slope equal to k1 = 7.8 10−4 s−1. (B) Theoretical (solid lines) and experi-
mental (symbols) chronoamperometric currents at Pt-black electrodes for various
concentrations of DEA-NONOate flowing solutions. From top to bottom: 1.5 (solid
circle), 1 (open circle), 0.5 (solid square), 0.2 (open square), 0.1 (solid triangle) and
0.05 mM (open triangle). Curve fit with k1 = 7.8 10−4 s−1, k2 [O2] = 5.5 102 M−2s−1 and
i/[NO•] = 1.13 10−3 AM−1. (C-D) Calibration curves for NO• detection at Pt-black elec-
trode established from DEA-NONOate flowing solutions of various concentrations
ranging from 1500 to 0.1 ␮M, at pH 7.4. Corrected currents monitored 850 and 1200 s
after solution preparation. Linear curve with slope equal to 1.13 10−3 AM−1. In (B-D)
E = 0.62 V/REF and pH = 7.4.
Fig. 4. (A) Voltammograms in a continuously DEA-NONOate flowing solution at Pt-
black electrode. Voltammograms recorded 270 s (solid line) and 770 s (dashed line)
after solution preparation. (B) Background corrected voltammograms at bare-Pt
(dotted line) and Pt-black electrodes (solid line), 270 s after solution preparation. (C)
Corrected voltammograms at Pt-black electrodes, 230 (top) to 730 s (bottom) after
solution preparation. (D) UV-Vis spectra, 120 to 9000 s after solution preparation.
In (A-B) 0.5 mM DEA-NONOate solution, pH = 7.4. In (C-D) 0.1 mM DEA-NONOate
solution at pH = 7.4.
However, NO• concentration results from the competition
between the fast DEA-NONOate decomposition (Eq. (5)) and the
slower reaction of NO• with oxygen [62]:
electrochemical oxidation current for NO• required a correction
from background currents. So, voltammograms recorded after com-
plete NO• decomposition were subtracted from the initial response
(Fig. 4B). This was a necessary step since NO• reaction with oxygen
(Eq. (7)) produces nitrite anions that accumulate and start to oxi-
dize at slightly higher potentials [44]. It is also known that NO•
donors like DEA-NONOate generate potential interfering species
issued from own NO• reactivity or other decomposition products
[60]. After background currents subtraction a wave was obtained
with a maximum at E = 0.62 V/REF in contrast to a plateau expected
under steady-state regime of mass transport. This electrochemical
behavior could be related to the progressive electrode poisoning
by side-products of DEA-NONOate decomposition and products
released during the potential scan. This was also confirmed by the
progressive current amplitude decay in successive voltammograms
recorded as a function of time (Fig. 4C). Nevertheless, in com-
parison to bare-Pt electrodes (Fig. 4B), corrected voltammograms
displayed current magnitude almost 350 times larger on Pt-black
electrodes. In order to verify that the time-course and concentra-
tion level of NO• produced were adequately monitored, the kinetic
profile of NO• was also analyzed from series of UV-Vis spectra mon-
itored under the same conditions since their variations tracked the
rate of DEA-NONOate decomposition only (Fig. 4D). In this case,
an average value of k1 could be determined at ꢁ = 250 nm with
k1 = (7.8 0.5) 10−4 s−1 (Fig. 5A). By introducing this value in Eq.
(9), k2 was evaluated from the best fit of a series of chronoampe-
rometric curves monitored at 0.62 V/REF at several DEA-NONOate
concentrations. For that purpose, the diffusion coefficient of NO•
was considered to be equal to 1.5 10−5 cm2s−1 [44] and currents
were estimated on the basis of theoretical predictions accounting
k2
4 NO• + O2 + 2 H2O−→ 4 NO− + 4H+
(7)
2
which obeys a second order rate low given by:
V2 = 4k2[NO•]2[O2]
(8)
where k2 = (2.3 0.6) 106 M−2s−1 at 20-25 ◦C [46,62–67].
Since the efficiency of NO• release is about 75% in Eq. (5),1.5
mole of nitric oxide is generated per mole of DEA-NONOate [68].
Therefore, the kinetic profile of NO• release follows the balance
between V1 and V2:
d[NO•]/dt = 1.5k1exp(-k1t)[Et2N-[N(O)NO]−]0-4k2[NO•]2[O2]
(9)
where [Et2N-[N(O)NO]−]0 is the initial concentration of DEA-
NONOate. Since V1 is much larger than V2, NO• concentration
increases faster at very short times, reaches a maximum and then
decays slowly [45]. Note that the corresponding half-life time of
DEA-NONOate is 16 min at 22 ◦C in 0.1 M phosphate buffer at phys-
iological pH [61]. To ensure that the time-course and concentration
level of produced NO• were enough to complete adequate detec-
tions, we first performed measurements not later than 20 min (i.e.,
1200 s) after the solution preparation. Electrochemical detection
of NO• was performed through its one-electron oxidation process
according to:
NO• → NO+ + e−
(10)
Fig. 4A shows voltammograms obtained under these conditions
at Pt-black electrode at 270 and 770 s. However, extracting the

116
Y. Li et al. / Electrochimica Acta 144 (2014) 111–118
for the local hydrodynamic regime [48]. A good agreement was
observed between experimental and calculated currents for all ini-
tial DEA-NONOate concentration only for k2 [O2] = 5.5 102 M−1s−1
(Fig. 5B). Since O2 solubility at room temperature was about
0.27 mM, this afforded k2 = (2.2 0.4) 106 M−2s−1. All these results
were thus perfectly consistent with values of k1 and k2 reported in
the literature [46,61–66].
On the basis of these kinetic determinations, a calibration curve
was established over a submillimolar concentration range by evalu-
ating NO• concentrations at the time of each chronoamperometric
measurement. Accordingly, a linear variation was obtained over
two decades (Fig. 5C and D) showing that NO• detection on Pt-
black electrodes was very reliable over this concentration range.
Note that the upper limit corresponded to the initial concentration
of DEA-NONOate that could be used experimentally. Note also that
the linear curves reported in these figures did not represent the
result of curve fits but were the theoretical line predicted from the
previous determined values of k1 and k2. The LOD resulting from
repetitive measurements at low NO• concentrations (i.e., close to
50 nM) was about 30 nM. This limit was similar to those already
reported under various conditions with different geometries and
electrode dimensions [25,26,29,40,67–71]. However, the sensitiv-
ity reached here was 2.87 AM−1cm−2, being higher by one order of
magnitude than previously reported ones.
3.3. Detection of ROS and RNS mixtures
Fig. 6. (A) Normalized steady-state voltammograms obtained at Pt-black electrodes
from separate flowing solutions of H2O2, ONOO-, NO• and NO2−. pH = 7.4 for H2O2,
−
NO• and NO2 and pH = 11.1 for ONOO−. Dashed lines define the potentials in
Oxidation of ONOO− at Pt-black electrode occurs at potentials
close to H2O2 oxidation [44].−A similar situation may apply to NO•
oxidation in presence of NO2 [45]. In order to be able to perform
quantitative analysis of ROS and RNS mixtures at Pt-black elec-
trodes, preliminary tests were thus performed to ensure (i) that
no interferences or competitive reactions occurred during simul-
taneous oxidations and (ii) that the relative contributions of one
species to the global response could easily recovered at every detec-
tion potential. The choice of an optimal detection potential for one
species corresponded to a compromise between reaching a limiting
current (i.e., a current plateau) and minimizing the influence of the
others ROS/RNS on each individual current response. In the case
of simultaneous detections of two species, it was convenient to
select a potential related to significant contributions of the species
involved.
amperometry offering the best sensitivity and selectivity for the detection of each
species. From left to right: E = 0.30, 0.45, 0.62 and 0.85 V/REF. (B) Current responses of
0.1 mM H2O2 microinjections at pH = 7.4. E = 0.3 V/REF (solid line) and E = 0.45 V/REF
(dashed line). (C) Current responses of 0.1 mM ONOO− microinjections at pH = 11.1.
E = 0.3 V/REF (solid line) and E = 0.45 V/REF (dashed line).
current responses monitored with separate H2O2 and ONOO− solu-
tions and mixtures of identical concentrations. These experiments
were conducted at pH 8.4 (at pH slightly higher than physiologi-
cal conditions) in order to detect enough amount of ONOO− (see
decomposition kinetics in Fig. 2D). This pH value offered a good
compromise for testing the performance of the electrochemical
detections with synthetic mixtures since it provided the necessary
time delay between the solution preparation and measurements.
Fig. 7A and B illustrate the perfect agreement obtained between
currents in case of microinjections performed at this pH. The sum
of individual currents matched the current measured for the cor-
responding mixture at each potential.
−
Voltammetric profiles of H2O2 and NO2 were thus established
separately at pH 7.4 under identical conditions. They are reported
and compared in Fig. 6A to those monitored for NO• and ONOO−.
Note that ONOO− profile was recorded at pH 11.1 owing to its
chemical stability. It was checked that its electrochemical oxidation
(Eq. (4)) was pH independent for conditions approaching phys-
iological ones (data not shown). As observed, oxidation of H2O2
overlapped strongly with that of ONOO− while oxidation profiles
of NO• and NO2− were well separated from the others. Accordingly,
four optimal detection potentials were defined as reported by the
vertical dashed lines in Fig. 6A. Note that 0.45 and 0.62 V/REF were
potentials selected previously for ONOO− and NO• detections. Rel-
ative influences of H2O2 and ONOO− at 0.3 and 0.45 V/REF were
then analyzed from the voltammetric profiles and from current
responses obtained from a series of microinjections performed at
these potentials (Fig. 6B and C). Owing to the results in Fig. 6A, one
had:
The same approac−h was applied for detection of the two other
species NO• and NO2 at 0.62 and 0.85 V/REF. On the basis of the
voltammetric profiles reported in Fig. 6A one obtained within the
accuracy of the measurements:
i = (iH2O2 + iONOO-) + iNO
and:
at E = 0.62 V/REF
(13)
i = (iH2O2 + iONOO-) + iNO + iNO2-
at E = 0.85 V/REF
(14)
These relations were also confirmed by series of microinjec-
tions performed at these two potentials (Fig. 7C and D). Finally,
the good agreement observed between data demonstrated that it
was possible at the time scale of these measurements to evaluate
the contribution of all the four species and therefore to assess the
composition of unknown mixtures of H2O2, ONOO-, NO• and NO2
.
−
i = 0.85iH2O2 + 0.43iONOO- at E = 0.30 V/REF
and:
(11)
Indeed, provided that the mass transport regime in microchannel is
well controlled and predicted [48,72,73], the determination of mix-
ture composition from Eqs. (11) to (14) is straightforward. These
results clearly suggested that no side reaction occurred during
simultaneous oxidations of the four species. It gives the oppor-
tunity to detect, characterize and quantify small amounts of the
i = iH2O2 + iONOO- at E = 0.45 V/REF
where i is the global current measured at each potential. These
(12)
relations were tested and validated experimentally by comparing

Y. Li et al. / Electrochimica Acta 144 (2014) 111–118
117
particular to detect locally reactive species like ONOO− and NO•
very close from their production sites. Work is under progress to
extend this approach to the detection of oxidative stress metabo-
lites emitted by living cells.
Acknowledgments
This work has been supported in part by the CNRS (UMR8640),
Ecole Normale Supérieure, UPMC and French Ministry of Research.
Y. L. thanks the China Scholarship Council and the Chinese Embassy
for a PhD grant.
References
[1] P. Pacher, J.S. Beckman, L. Liaudet, Phys. Rev. 87 (2007) 315.
[2] V.C. Besson, N. Croci, R.G. Boulu, M. Plotkine, C. Marchand-Verrecchia, Brain.
Research. 989 (2003) 58.
[3] C. Amatore, S. Arbault, M. Guille, F. Lemaître, Chem. Rev. 108 (2008) 2585.
[4] J.C. Toledo Jr., O. Augusto, Chem. Res. Toxicol. 25 (2012) 975.
[5] R. Kissner, T. Nauser, P. Bugnon, P.G. Lye, W.H. Koppenol, Chem. Res. Toxicol.
10 (1997) 1285.
[6] D. Jourd’heuil, F.L. Jourd’heuil, P.S. Kutchukian, R.A. Musah, D.A. Wink, M.B.
Grisham, J. Biol. Chem. 276 (2001) 28799.
[7] M.K. Hulvey, C.N. Frankenfeld, S.M. Lunte, Anal. Chem. 82 (2010) 1608.
[8] D. Quinton, S. Griveau, F. Bedioui, Electrochem. Commun. 12 (2010) 1446.
[9] O.C. Zafiriou, M. McFarland, Anal. Chem. 52 (1980) 1662.
[10] J.N. Bates, Neuroprotocols 1 (1992) 141.
[11] D.C. Yao, A.G. Vlessidis, N.P. Evmiridis, A. Evangelou, S. Karkabounas, S. Tsam-
palas, Anal. Chim. Acta 458 (2002) 281.
[12] M.M. Tarpey, D.A. Wink, M.B. Grisham, Am. J. Physiol. Regul. Integr. Comp.
Physiol. 286 (2004) R431.
[13] P. Wardman, Free Radic. Biol. Med. 43 (2007) 995.
[14] Y.Y. Woldman, J. Sun, J.L. Zweier, V.V. Khramtsov, Free Radic. Biol. Med. 47
(2009) 1339.
Fig. 7. Current responses at Pt-black electrode as a function of the detection poten-
tial with microinjections of 0.1 mM separate solutions and mixtures of H2O2, ONOO-,
−
DEA-NONOate and NO2 at pH = 8.4. In dashed line, sum of individual current
responses obtained from separate solutions.
four ROS/RNS when released by populations of living cells [74].
No•t-e that in such case, the oxidation wave of superoxide anion
O2 is not expected to produce any interfere in these determina-
[15] T. Nagano, J. Clin. Biochem. Nutr. 45 (2009) 111.
[16] R. Radi, T.P. Cosgrove, J.S. Beckman, B.A. Freeman, Biochem. J. 290 (1993) 51.
[17] C.C. Winterbourn, Biochim. Biophys. Acta 1840 (2014) 730.
[18] R.M. Wightman, Science 311 (2006) 1570.
[19] Y. Wang, J.M. Noel, J. Velmurugan, W. Nogala, M.V. Mirkin, C. Lu, M. Guille
Collignon, F. Lemaitre, C. Amatore, Proc. Natl. Acad. Sci. USA 109 (2012) 11534.
[20] S. Borgmann, I. Radtke, T. Erichsen, A. Blochl, R. Heumann, W. Schuhmann,
ChemBioChem 7 (2006) 662.
•-
tions since O2 concentration will be much lower than H2O2 due
to its fast disproportionation reaction [56] (additionally catalyzed
by SOD) or fast reaction with NO• [5]. However, the accuracy of
measurements for NO• detection in biological media [42,75] is still
a relevant question. Finally, physiological pHs should be leading to
correct results whenever ROS/RNS transport between the cells and
the detecting electrodes is fast enough. Indeed, under our condi-
tions designed as not to stress the cells hydrodynamically (i.e., at
a realistic flow rate of 2 ␮Lmin−1 and with a microchannel cross
section of 20 × 200 ␮m), the distance between cells and electrodes
needs to be less than 1 cm to ensure a transit time below the sec-
ond timescale. This is fully consistent with the half life time of
peroxynitrite [52], which is the most reactive of the four species.
[21] S. Isik, J. Castillo, A. Blochl, E. Csoregi, W. Schuhmann, Bioelectrochemistry 70
(2007) 173.
[22] V.L. Lobachev, E.S. Rudakov, Uspekhi Khimii 75 (2006) 422.
[23] S. Borgmann, Anal. Bioanal. Chem. 394 (2009) 95.
[24] D. Gupta, B. Harish, R. Kissner, W.H. Koppenol, Dalton Trans. (2009) 5730.
[25] F. Bedioui, D. Quinton, S. Griveau, T. Nyokong, Phys. Chem. Chem. Phys. 12
(2010) 9976.
[26] B.J. Privett, J.H. Shin, M.H. Schoenfisch, Chem. Soc. Rev. 39 (2010) 1925.
[27] D.B. Gunasekara, M.K. Hulvey, S.M. Lunte, Electrophoresis 32 (2011) 832.
[28] D.B. Gunasekara, M.K. Hulvey, S.M. Lunte, J.A. da Silva, Anal. Bioanal. Chem. 403
(2012) 2377.
[29] R.A. Hunter, B.J. Privett, W.H. Henley, E.R. Breed, Z. Liang, R. Mittal, B.P. Yoseph,
J.E. McDunn, E.M. Burd, C.M. Coopersmith, J.M. Ramsey, M.H. Schoenfisch, Anal.
Chem. 85 (2013) 6066.
[30] C. Amatore, S. Arbault, D. Bruce, P. de Oliveira, M. Erard, M. Vuillaume, Chem.
Eur. J. 7 (2001) 4171.
4. Conclusion
[31] G. Ferrer-Sueta, R. Radi, ACS Chem. Bio. 4 (2009) 161.
[32] O. Adegoke, T. Nyokong, J. Lumin. 134 (2013) 448.
[33] S. Arbault, N. Sojic, D. Bruce, C. Amatore, A. Sarasin, M. Vuillaume, Carcinogene-
sis 25 (2004) 509.
[34] C. Amatore, S. Arbault, C. Bouton, K. Coffi, J.-C. Drapier, H. Ghandour, Y. Tong,
ChemBioChem 7 (2006) 653.
[35] C. Amatore, S. Arbault, M. Erard, Anal. Chem. 80 (2008) 9635.
[36] C. Amatore, S. Arbault, Y. Chen, C. Crozatier, I. Tapsoba, Lab Chip 7 (2007) 233.
[37] C. Amatore, S. Arbault, C. Bouton, J.-C. Drapier, H. Ghandour, A.C.W. Koh, Chem-
BioChem 9 (2008) 1472.
[38] M.R. Filipovic, A.C.W. Koh, S. Arbault, V. Niketic, A. Debus, U. Schleicher, C. Bog-
dan, M. Guille, F. Lemaitre, C. Amatore, I. Ivanovic-Burmazovic, Angew. Chem.
Int. Edit. 49 (2010) 4228.
[39] A.S. Bernard, C. Giroud, H.Y. Ching, A. Meunier, V. Ambike, C. Amatore, M. Guille
Collignon, F. Lemaitre, C. Policar, Dalton Trans. 41 (2012) 6399.
[40] W. Cha, Y.-C. Tung, M.E. Meyerhoff, S. Takayama, Anal. Chem. 82 (2010) 3300.
[41] E.C. Metto, K. Evans, P. Barney, A.H. Culbertson, D.B. Gunasekara, G. Caruso, M.K.
Hulvey, J.A. Fracassi da Silva, S.M. Lunte, C.T. Culbertson, Anal. Chem. 85 (2013)
10188.
[42] J.L. Harding, M.M. Reynolds, Anal. Chem. 86 (2014) 2025.
[43] M.W. Breiter, Electrochim. Acta 11 (1966) 905.
[44] Y. Li, C. Sella, F. Lemaitre, M.G. Collignon, L. Thouin, C. Amatore, Electroanalysis
25 (2013) 895.
As in the case of H2O2 and NO2-, electrochemical oxidations
of ONOO− and NO• at optimized Pt-black electrodes led to high
detection performances under microfluidic conditions. High reli-
ability and sensitivity of the measurements were achieved allowing
kinetic parameters to be determined accurately. Individual detec-
tions and quantifications were successfully performed in case of
mixtures of the four species. Indeed, upon taking into account
the steady-state mass transport regimes achieved at microchannel
electrodes, relative composition of solutions could be evaluated at
selected detection potentials.
These results validate the concept of simultaneous electrochem-
ical detections of the four important ROS and RNS in microfluidic
devices with the use of very small amount of analytes and living
biological materials. Temporal resolution below the second-time-
scale can be easily achieved at channel microelectrodes owing to
the geometry of the device and flow velocity. This study opens the
way to perform measurements under physiological conditions, in

118
Y. Li et al. / Electrochimica Acta 144 (2014) 111–118
[45] C. Amatore, S. Arbault, Y. Bouret, B. Cauli, M. Guille, A. Rancillac, J. Rossier,
ChemPhysChem 7 (2006) 181.
[60] L.K. Keefer, R.W. Nims, K.M. Davies, D.A. Wink, Methods Enzymol. 268 (1996)
281.
[46] S. Griveau, C. Dumezy, P. Goldner, F. Bedioui, Electrochem. Commun. 9 (2007)
2551.
[47] Y. Ikariyama, S. Yamauchi, T. Yukiashi, H. Ushioda, J. Electrochem. Soc. 136
(1989) 702.
[48] C. Amatore, N. Da Mota, C. Sella, L. Thouin, Anal. Chem. 79 (2007) 8502.
[49] W.H. Koppenol, Quimica Nova 21 (1998) 326.
[50] Y. Luo, C. Zhang, Y. She, R. Zhong, P. Wei, Reac. Kinet. Mech. Cat. 101 (2010)
291.
[61] J.A. Hrabie, J.R. Klose, D.A. Wink, L.K. Keefer, J. Org. Chem. 58 (1993) 1472.
[62] V.L. Pogrebnaya, A.P. Usov, A.V. Baranov, A.I. Nesterenko, P.I. Bezyazychnyi, J.
App. Chem. Ussr 48 (1975) 1004.
[63] V.G. Kahritonov, A.R. Sundquist, V.S. Sharma, J. Biol. Chem. 269 (1994) 5881.
[64] R.S. Lewis, W.M. Deen, Chem. Res. Toxicol. 7 (1994) 568.
[65] M. Pires, M. Rossi, D. Ross, Int. J. Chem. Kin. 26 (1994) 1207.
[66] S. Goldstein, G. Czapski, J. Am. Chem. Soc. 117 (1995) 12078.
[67] A. Cserey, M. Gratzl, Anal. Chem. 73 (2001) 3965.
[68] F. Bedioui, N. Villeneuve, Electroanalysis 15 (2003) 5.
[69] I.R. Davies, X. Zhang, Methods Enzymol. 436 (2008) 63.
[70] D. Quinton, A. Girard, L.T.T. Kim, V. Raimbault, L. Griscom, F. Razan, S. Griveau,
F. Bedioui, Lab Chip 11 (2011) 1342.
[71] S. Griveau, F. Bedioui, Anal. Bioanal. Chem. 405 (2013) 3475.
[72] C. Amatore, N. Da Mota, C. Sella, L. Thouin, Anal. Chem. 80 (2008) 4976.
[73] C. Amatore, C. Lemmer, P. Perrodin, C. Sella, L. Thouin, Electrochem. Commun.
13 (2011) 1459.
[51] S. Pfeiffer, A.C.F. Gorren, K. Schmidt, E.R. Werner, B. Hansert, D.S. Bohle, B.
Mayer, J. Biol. Chem. 272 (1997) 3465.
[52] C. Molina, R. Kissner, W.H. Koppenol, Dalton Trans. 42 (2013) 9898.
[53] S. Tibi, W.H. Koppenol, Helv. Chim. Acta 83 (2000) 2412.
[54] J.P. Crow, Free Radic. Biol. Med. 28 (2000) 1487.
[55] Z.J. Chen, W. Ren, Q.E. Wright, H.W. Ai, J. Am. Chem. Soc. 135 (2013) 14940.
[56] B.H.J. Bielski, D.E. Cabelli, R.L. Arudi, J. Phys. Chem. Ref. Data 14 (1985)
1041.
[57] R.B. Mikkelsen, P. Wardman, Oncogene 22 (2003) 5734.
[58] I.G. Zacharia, W.M. Deen, Ann. Biomed. Eng. 33 (2005) 214.
[59] X. Liu, J.L. Zweier, J. Electroanal. Chem. 688 (2013) 32.
[74] C. Amatore, S. Arbault, A.C.W. Koh, Anal. Chem. 82 (2010) 1411.
[75] R.A. Hunter, W.L. Storm, P.N. Coneski, M.H. Schoenfisch, Anal. Chem. 85 (2013)
1957.