Pyrolyzed photoresist carbon electrodes in aprotic solvent: Bilirubin electrochemistry and interaction with electrogenerated superoxide

Article abstract of DOI:10.1016/j.electacta.2014.09.046

Pyrolyzed photoresist carbon electrodes (PPCEs) are fabricated by the photopatterning of a negative tone epoxy-based photoresist, SU-8, through optimized standard UV photolithography. The electrochemical characteristics of PPCEs are investigated in dimethyl sulfoxide (DMSO), observing a wider accessible potential window and a smaller capacitance with respect to glassy carbon electrodes. PPCEs are used to study the cyclic voltammetric behavior of bilirubin (BR) in DMSO. Detailed information is obtained on the multiple steps involved both in the electrochemical oxidation and reduction of BR. Interesting points concerning the electrochemical oxidation of BR to biliverdin are clarified, identifying the formation of an intermediate whose fate depends on the time scale of the electrochemical experiment. PPCEs are also used to electrogenerate the superoxide anion O₂^-? in DMSO for studying possible reactions between BR and O₂^-?. The results obtained demonstrate that BR is an efficient superoxide scavenger and that a concentration 2 mM of BR is high enough to consume all the O₂^-? generated by oxygen reduction at the PPCE/DMSO interface.

Full text of DOI:10.1016/j.electacta.2014.09.046

Electrochimica Acta 147 (2014) 401–407
Contents lists available at ScienceDirect
Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta
Pyrolyzed Photoresist Carbon Electrodes in Aprotic Solvent: Bilirubin
Electrochemistry and Interaction with Electrogenerated Superoxide
M. Silvestrini , A. Mardegan , R. Kamath , M. Madou c,d,e, L.M. Moretto a,
a
b
c
f
b
a,
S. Passamonti , P. Scopece , P. Ugo *
a
Department of Molecular Sciences and Nanosystems, University Ca’Foscari of Venice, Santa Marta 2137, 30123 Venice, Italy
Veneto Nanotech, via delle Industrie 5, 30175 Venice-Marghera, Italy
Department of Biomedical Engineering, University of CA, Irvine, CA 92697, USA
Materials and Manufacturing Technology, University of CA, Irvine, CA 92697, USA
Department of Mechanical and Aerospace Engineering, University of CA, Irvine, CA 92697, USA
Department of Life Sciences, University of Trieste, 34127 Trieste, Italy
b
c
d
e
f
A R T I C L E I N F O
A B S T R A C T
Article history:
Received 9 July 2014
Received in revised form 25 August 2014
Accepted 9 September 2014
Available online 28 September 2014
Pyrolyzed photoresist carbon electrodes (PPCEs) are fabricated by the photopatterning of a negative tone
epoxy-based photoresist, SU-8, through optimized standard UV photolithography. The electrochemical
characteristics of PPCEs are investigated in dimethyl sulfoxide (DMSO), observing a wider accessible
potential window and a smaller capacitance with respect to glassy carbon electrodes. PPCEs are used to
study the cyclic voltammetric behavior of bilirubin (BR) in DMSO. Detailed information is obtained on the
multiple steps involved both in the electrochemical oxidation and reduction of BR. Interesting points
concerning the electrochemical oxidation of BR to biliverdin are clarified, identifying the formation of an
intermediate whose fate depends on the time scale of the electrochemical experiment.
Keywords:
Pyrolyzed photoresist carbon
Bilirubin
Superoxide
Dimethyl sulfoxide
Voltammetry
ꢀ
ꢁ
PPCEs are also used to electrogenerate the superoxide anion O
2
in DMSO for studying possible
ꢀꢁ
2
reactions between BR and O
. The results obtained demonstrate that BR is an efficient superoxide
ꢀ
ꢁ
scavenger and that a concentration 2 mM of BR is high enough to consume all the O
oxygen reduction at the PPCE/DMSO interface.
2
generated by
ã 2014 Published by Elsevier Ltd.
1
. Introduction
Recently, a new procedure to fabricate carbon electrodes based
on the controlled pyrolysis of polymeric photoresist has been
introduced, obtaining so-called pyrolyzed photoresist carbon
electrodes (PPCEs) [10]. This is a very promising approach since,
by using simple UV photo-lithography it is possible to pattern
high-performance carbon electrodes with complex geometries,
such as interdigitated arrays of microelectrodes [11,12]. The
optimization of the pyrolysis conditions allows one to obtain
PPCEs in which the carbon is mostly amorphous GC [13]. Studies
performed in aqueous media demonstrated that the electrochem-
ical behavior of PPCEs compares with that of classical electrodes
prepared from bulk glassy carbon rods, with the advantage of being
cheaper and with the possibility to customize their design [13]. In
this paper we report, for the first time, the application of PPCEs for
voltammetric measurements in an aprotic solvent, namely
dimethyl sulfoxide (DMSO), used here as medium to dissolve
bilirubin (BR) and to study its electrochemical behavior together
with its reactivity towards electrogenerated superoxide radical
anions.
Carbon electrodes possess many advantages including low
fabrication cost, wide accessible potential window, chemical and
electrochemical stability [1] so that, in many cases, these electro-
des are superior with respect to noble metal electrodes [2]. Among
common carbon electrode materials, such as pyrolytic graphite,
carbon fibers or carbon paste [3–6], glassy carbon (GC) is the most
widely used [7,8]. Glassy carbon is impermeable to gases and
liquids, has a very small porosity and can be polished to a mirror-
like finish an almost infinitive number of times. Moreover, it is very
resistant both to high oxidizing and reducing potentials thereby
allowing one to perform electrochemistry in a 4–5 V wide potential
window [9], when suitable electrolytes are used.
*
Corresponding author.
E-mail address: ugo@unive.it (P. Ugo).
http://dx.doi.org/10.1016/j.electacta.2014.09.046
013-4686/ã 2014 Published by Elsevier Ltd.
0

4
02
M. Silvestrini et al. / Electrochimica Acta 147 (2014) 401–407
Scheme 1. Structural formulas of bilirubin and its oxidation products, namely biliverdin, purpurin and choletelin.
2ꢀ
ꢁ
Bilirubin (BR) is the yellow-orange bile pigment found in blood,
electrochemical generation of O
(by reduction of dissolved
mostly bound to the plasma protein albumin. It is a linear
tetrapyrrole (see scheme 1), insoluble in water at neutral pH, but
very soluble in organic solvents. BR is an important serum
biomarker used in clinical medicine for assessing hemolysis,
hepatic function and cardiovascular risk [14]. This molecule
originates from the degradation of the heme moiety in hemoglo-
bin, other hemoproteins, such as cytochromes, catalase, peroxidase
and tryptophan pyrrolase, and free heme [15]. In human body BR is
present mainly as conjugated and unconjugated BR; the first forms
a complex with gluconic acid, which makes it water soluble.
Because of the important role played by this redox compound in
several physiological processes and diseases [16], BR has been the
subject of several previous electrochemical studies performed
both in aqueous [17–21] and non aqueous media [22–29]. The
results of these studies indicate that the electrochemical oxidation
of BR is a multistep process in which, depending on the applied
potential, BR is electrochemically oxidized to biliverdin (BV),
purpurin or choletelin (see Scheme 1) [22].
oxygen) and to study the reactivity of BR towards this radical as
a proof of the radical scavenging capabilities of BR.
2. Experimental section
2.1. Electrochemical apparatus
All voltammetric measurements were carried out at room
ꢃ
temperature (22 ꢂ1 C) with a CHI1222A potentiostat controlled
via a personal computer with its own software. A three-electrode
single-compartment cell made of dark glass and equipped with a
2
PPCE (geometric area = 0.031 cm ) or a glassy carbon working
2
electrode (GCE; geometric area = 0.071 cm ), a platinum counter
electrode and a platinum pseudo - reference electrode (Pt-pseudo).
All potential values are finally referred to the E1/2 of the ferrocene/
+
ferricinium (Fc/Fc ) redox couple recorded in the same experi-
mental conditions as the data of interest [31,32]. For the latter aim,
1 mM Fc was added to the electrolyte and the relevant E1/2 was
Because of the poor water solubility of unconjugated BR, the
use of non aqueous media is to be preferred for gathering
fundamental information on its electrochemical behavior. This
notwithstanding, even for measurements performed in aprotic
media like DMF [22–27] and DMSO [28,29], in the literature
there is no agreement on the oxidation mechanisms or even on
the number of electrons involved in the oxidation steps, since
different authors report different results. For instance, con-
cerning the first oxidation process, coulometric data support a
two-electrons oxidation [28,29]. However, voltammetric data
gathered at the corresponding oxidation peak do not agree with
evaluated as E1/2 = (Ep
and backward peak potentials, respectively; in all these measure-
ments the peak-to-peak separation (DEp= Ep - Ep ) for Fc/Fc was
f b
f b f b
+ Ep )/2 where Ep and Ep are the forward
+
60 ꢂ 5 mV and E1/2 = +0.37 V vs Pt-pseudo. PPCEs were used as
obtained after the pyrolysis procedure. Before use, GCEs were
mirror polished with a 0.3 and a 0.05 mm alumina slurry on
microcloth pads, rinsed with water and dried.
Coulometric measurements were carried out at room temper-
ature with an AMEL 552 potentiostat with the associated AMEL
integrator 731. In order to avoid the generation of products
derived from higher oxidation potential process (see below), the
potentials applied during electrolysis were kept approximately
30 mV below the peak potential recorded voltammetrically for
the process of interest. When required, electrochemical measure-
ments were performed under a nitrogen atmosphere, after
purging the electrolyte solution for 15 minutes. Nitrogen was
dried by bubbling through concentrated sulfuric acid and pre-
a two-electron process; for instance, |E
p
- Ep/2| values are
significantly larger than the 28.5 mV value expected for a
reversible two-electron oxidation [30]. Moreover, Ribo et al.
[
26] reported voltammograms in which the first oxidation peak
of bilirubin presented the same peak current as the one-
electron oxidation of similar molecules. Similarly, for the BR
reduction, it is not clear which peaks are ascribed to the direct
reduction of BR and which are related to the electroactivity of
BV electrogenerated in the anodic portion of the cyclic
voltammetric experiments. These controversial results
prompted us to undertake the present study with the goal of
studying the electrochemical performance of PPCEs in DMSO,
and then revisiting the electrochemical behavior of BR in this
aprotic solvent. The findings enabled us to use the PPCE for the
2
saturated with DMSO vapors. Measurements involving the O /
ꢀꢁ
O
2
couple were performed in solutions equilibrated with
atmospheric oxygen.
The supporting electrolyte tetrabutylammonium tetrafluoro-
ꢃ
4
borate (TBABF ) was dried overnight in a vacuum oven at 40 C
and stored in a desiccator before use. All laboratory glassware was
carefully dried with a heat-gun before use in order to remove
water.

M. Silvestrini et al. / Electrochimica Acta 147 (2014) 401–407
403
2.2. UV–vis Spectroscopy and AFM
The spectra were recorded using an UV–vis Perkin Elmer
Lambda 2 spectrophotometer equipped with quartz cuvettes with
a light path of 1 cm.
AFM measurements were performed in air on samples dried at
room conditions, using a Veeco NS IV Dimension 3100 Scanning
TM
Probe Microscope. Images were collected in Tapping mode using
standard silicon cantilevers with typical resonant frequency
150 kHz and elastic constant 5 N/m. The surface roughness (Rq)
was calculated as the root mean square average of the roughness
profile ordinates.
2.3. Materials
BR (Frontier Scientific, ꢄ 98%), DMSO (Sigma-Aldrich, ꢄ 99.5%)
and TBABF
4
(Fluka) were used without further purification.
Fig. 1. Cyclic voltammograms recorded in DMSO, 0.1 M TBABF
4
at PPCE (full line)
Nitrogen was purchased from SIAD (grade 5.0, purity 99.999%).
SU-8 was obtained from MicroChem Inc. MA, USA. Silicon wafers
were obtained from Noel Technologies, CA, USA. Purified water was
produced using a Milli-Ro plus Milli-Q (Millipore) water purifica-
tion system. All other chemicals were reagent grade.
ꢀ1
and GCE (dashed line). Scan rate 100 mV s
.
From the blank voltammograms recorded at different scan rates
data not shown) it was possible to calculate the double-layer
(
capacitance (Cdl) values of both electrode materials, by using the
equation [30]:
2.4. PPCE preparation
I
C
= Cdl A v
where v is the voltammetric scan rate, A is the electrode area
and I is the capacitive current calculated as half of the difference
between the forward and backward currents, at ꢀ0.7 V vs Fc/Fc . By
A negative tone epoxy-based photoresist, SU-8, was photo-
patterned employing standard UV photolithography, using previ-
ously described procedures [13]. Briefly, the SU-8 photoresist was
C
+
ꢀ
2
this method, values of Cdl equal to 93.5 and 191.4 mF cm were
spin-coated on SiO
2
coated wafers and pre-dried for 30 min at
obtained for PPCE and GCE, respectively. Note that the capacitance
value measured here for GC in DMSO is rather close to the value
ꢃ
1
20 C. After deposition, the casting solvent was evaporated from
the photoresist in a soft bake or pre-bake step. During this step,
performed at 95 C, the use of a hot plate was recommended to
ꢀ
2
measured for GC in acetonitrile, namely 187
mF cm [33]. The
ꢃ
lower I observed for the PPCE indicates a lower surface roughness
C
avoid skin effects. Indeed, baking the photoresist layer in a
convection oven would evaporate the solvent present on the top
surface of the photoresist first, thus slowing down the evaporation
of the solvent from the bulk of the polymer and thus resulting in a
resist skin. The resist was then exposed to UV light (360 nm) which
generates a low concentration of the catalyst for the cross-linking
reaction. After a post-exposure bake, which was needed to fully
polymerize the SU-8, the patterned features were developed. In
this step, the non-polymerized SU-8 was dissolved upon immer-
sion of the wafer in a developing agent provided by MicroChem Inc.
The patterned structures were carbonized in a three-step
pyrolysis process in an open ended quartz-tube furnace. First, the
with respect to conventional GCE. This was confirmed by AFM
observations (Fig. S1, Supplementary data) where a surface
roughness, Rq = 0.8 nm was observed for PPCE vs a Rq = 3.6 nm
for GCE. These results indicate that the PPCE surface is even
smoother than the surface of a highly polished GCE.
3
.2. Voltammetric characterization of BR in DMSO
Fig. 2, full line shows the CV pattern recorded with a PPCE in
1
4
mM BR and 0.1 M TBABF , in DMSO. In the first segment (anodic
scan direction) two well defined oxidation peaks, namely A1 at
+
+
0.25 V, A2 at +0.43 V vs Fc/Fc followed by a broad shoulder near
samples were heated under N
2
flow (flow rate: 2000 sccm) at
the anodic limit of the voltammogram are observed. In the second
ꢃ
3
00 C for 60 min. The temperature was then ramped up at a rate of
ꢃ
ꢃ
10 C/min to 900 C. The samples were kept at this final pyrolysis
temperature for 1 h before cooling.
Before carrying out the electrochemical measurements, the
TM
electrodes were insulated with a Kapton1 (DuPont ) tape with a
circular hole punched in it to expose a defined geometric area, of
2
0
3
3
.031 cm to the solution only.
. Results and discussion
.1. Characterization of PPCE in DMSO
Fig. 1 reports the blank CVs recorded in 0.1 M supporting
ꢀ1
4
electrolyte (TBABF ) at 100 mV s with a PPCE (full line) and a
conventional GCE (dashed line) in the absence of dissolved oxygen.
The current values in these voltammograms are normalized with
respect to the geometrical area of the electrodes and reported as
current densities (J). The comparison shows a similar cathodic limit
for the accessible potential window on both electrode materials.
On the anodic side, a significantly higher background current is
observed at the GCE when approaching +0.7 V.
Fig. 2. Cyclic voltammograms recorded at a PPCE in 1 mM BR (full line) and in 1 mM
Fc (dashed line). Supporting electrolyte 0.1 M TBABF in DMSO, scan rate 20 mV s .
4
ꢀ
1

4
04
M. Silvestrini et al. / Electrochimica Acta 147 (2014) 401–407
Table 1
Voltammetric parameters from data in Fig. 2.
Peak
E
p
(V) vs Fc/Fc+
|E
p
- Ep/2| (mV)
Ip (mA)
A1
C1”
C2
Fc
+0.25
ꢀ0.75
ꢀ1.79
+0.03
67
59
45
57
2.86
0.79
0.99
2.79
segment (return scan), four peaks (C1', C1”, C2 and C3) are detected
at negative potential values. The main features of this CV overlap
with the voltammogram recorded with a GC disk electrode in the
same experimental conditions [34]. The dashed line CV was
recorded on the same electrode and same supporting electrolyte
but with 1 mM Fc added to the solution. It is worth stressing that
the peak current A1 is the same of that recorded for an equimolar
concentration of Fc which undergoes an one-electron oxidation
[31,32]; such an evidence supports that A1 is an one-electron peak.
Table 1 compares some of the characteristic parameters
extractable with satisfactory accuracy from the CVs in Fig. 2.
These data show that |E - Ep/2| values (where E and Ep/2 are the
p
p
peak potential and the potential at the half-peak height,
respectively) do not substantially differ from the value of
56.5 mV expected for an one-electron reversible process, apart
from peak C2.
It is worth pointing out here that, in general, the features of a CV
of BR on a PPCE resemble those presented in the literature for
measurements of BR solutions on Hg electrodes in DMSO [28] and
DMF [26], and on Pt in DMF [22–24,27], however, in these previous
works peak A1 was always considered as a two-electron peak.
Fig. 3 shows the CVs recorded at a PPCE in the same
experimental conditions as in Fig. 2, but at different scan rates.
The plot in the inset shows that peak current A1 scales linearly
with the square root of the scan rate (v1 ), indicating a diffusion
controlled process. On the other hand, peak A2 decreases while
increasing the scan rate up to the point where it disappears at a
/2
ꢀ1
scan rate of 100 mV s . Such a dependence suggests that peak
A2 is related to the oxidation of a product generated after peak A1
via a relatively slow chemical reaction. Peaks C1', C2 and C3,
4
Fig. 4. Cyclic voltammograms recorded in 1 mM BR, 0.1 M TBABF , DMSO (a) by
reversing the scan potential immediately after peak A1 (thick full line), A2 (thin full
line) and the anodic limit (dotted line); (b) initiating the potential sweep in the
1
/2
increase linearly with v
.
ꢀ1
positive (thin full line) or negative (thick full line) direction. Scan rate 50 mV s
.
In Fig. 4a we illustrate the changes in a CV upon reversing the
scan potential immediately after peak A1 (thick full line), A2 (thin
full line) and the end of the anodic segment (dotted line). The
comparison between these voltammograms makes it clear that
peaks C1' and C1” are associated with peak A1. When reversing the
scan direction after peak A1, the peaks C1' and C1” indeed do
increase (see also inset in Fig. 4a) as a consequence of the longer
oxidation time. On the contrary, peaks C2 and C3 do not undergo
any substantial change when changing the anodic limit of the scan.
Fig. 4b shows that when the scan is started in the negative
direction (thick full line), C1' and C1” are not detected while peaks
C2 and C3 are present. This indicates that peaks C2 and C3 are
related to the direct electrochemical reduction of BR, as previously
reported for BR reduction in DMSO at Hg and Au electrodes by
Slifstein and Ariel [28]. These authors attributed the two peaks to
two independent irreversible reduction processes generating
unknown dimerization or polymerization products.
It can be noted that in the positive part of the voltammogram a
new peak appears at ꢀ0.03 V, this peak being associated to the re-
oxidation of the reduction products of BR.
UV-Vis spectra (Fig. S2, Supplementary data) recorded on the
products after exhaustive electrolysis at a potential close to peak
A1 are characterized by the typical biliverdin (BV) features,
displaying two absorption maxima at 389 and 685 nm (Fig. S2b,
Supplementary data). Note that, at the same time, the absorbance
maximum at 456 nm (Fig. S2a, Supplementary data), typical for
BR, disappeared and the solution color turned from orange to
green. Coulometric integration of the current indicated that for
4
Fig. 3. Cyclic voltammograms recorded in 1 mM BR, 0.1 M TBABF , DMSO at
ꢀ
1
1/2
different scan rates (10, 20, 50 and 100 mV s ). Inset: plot of Ip (A1) vs v
.

M. Silvestrini et al. / Electrochimica Acta 147 (2014) 401–407
405
Table 2
completing the electrolysis two equivalents of electrons per mole
of BR are required.
Voltammetric data relevant to the CVs reported in Fig. 5.
v (mV sꢀ1)
Ip
(
m
A)
D
Ep (mV)
E1/2 (V)
The apparent incongruence of the observation of an one-electron
oxidation peak A1 on the voltammetric time-scale while a two-
electron oxidation is observed during the slower exhaustive
electrolysis process matches with the occurrence of the mechanism
c
2
0
ꢀ2.77
ꢀ4.27
ꢀ5.85
176
196
220
ꢀ1.23
ꢀ1.23
ꢀ1.23
50
00
1
(
1)–(4) inwhich BR is oxidized at first to the radical BR+ꢁ (reaction 1).
This radical can then dimerize (reaction 2) to finally generates BV by
losing hydrogen ions in reaction 3. On the voltammetric time-scale,
the first oxidation is the rate determining step, so that the peak
displays the features typical of a one-electron process.
nitrogen. The reduction peak current scales linearly with the
square root of the scan rate (not shown), while all the other
relevant voltammetric parameters, listed in Table 2, agree with the
occurrence of a quasi-reversible one-electron diffusion-controlled
reduction. This evidence together with a comparison with previous
literature data [35], indicate that the process involved is the one-
electron reduction of dissolved oxygen to generate the superoxide
radical anion, according to reaction (5):
+ꢁ
ꢀ
2
BR ! 2BR + 2e
(1)
(2)
(3)
(4)
BR+ ! (BR)
ꢁ
2+
2
2
+
! 2 BV + 4H + 2 eꢀ
+
ꢀ
2ꢀꢁ
BR)
2
O
2
+ e ! O
(5)
At variance with what has been observed at a GCE [36,37], the
reduction of O at the PPCE is not fully reversible. The quinone
2
+
ꢀ
BR ! BV + 2H + 2e
On the (longer) coulometric time scale, the whole sequence
expressed by the overall reaction (4) is operative so that the two-
electrons–two-protons oxidation of BR to BV is completed. Such
observation agrees also with voltammetric data that show that
peak A2, attributed to the oxidation of BV to purpurin, is detected
only at scan rates < 50 mV/s. Concerning peak A2, it is worth
stressing that electrolysis experiments performed at potential
values close to this peak, produced indeed a purple solution
(
8 conjugated double bonds) whose UV–vis spectrum was
characterized by the appearance of three new absorption peaks
at 336, 518 and 546 nm (Fig. S2c, Supplementary data), typical for
purpurin generation [24,34]. Finally, it can be noted that a broad
shoulder is detected just before the positive limit of the accessible
potential window; its location close to this limit prevented us to
perform any detailed analysis however, on the basis of previous
studies [24], it can be attributed to the formation of choletelin.
5.1. Reactivity of BR with electrogenerated superoxide anion
Fig. 5 shows the voltammograms recorded at a PPCE in a 0.1 M
TBABF , DMSO solution equilibrated with ambient air. A single
4
catho-anodic peak system centered at approximately -1.23 V vs Fc/
+
Fc is detected, which disappears upon prolonged bubbling with
Fig. 6. (a) Cyclic voltammograms recorded in 0.1 M TBABF
presence of atmospheric oxygen at different BR concentrations, namely: 0 (full line),
.8 (dashed line) and 4 mM (dotted line); (b) CVs recorded before (full line) and after
4
, DMSO at a PPCE in the
0
Fig. 5. Cyclic voltammograms recorded in 0.1 M TBABF
4
, DMSO in the presence of
the addition of 2 mM (dashed line) and 4 mM (dotted line) BR, in the presence of
ꢀ
1
dissolved oxygen at a PPCE at different scan rates, namely: 20 mV s (dashed line),
atmospheric oxygen dissolved, and after deoxygenation of the solution (dashed-
ꢀ
1
ꢀ1
ꢀ1
5
0 mV s (dotted line) and 100 mV s (full line).
dotted line). Scan rate 50 mV s
.

4
06
M. Silvestrini et al. / Electrochimica Acta 147 (2014) 401–407
groups, typically present on the GC surface, act as catalysts for
oxygen reduction, which displays as a reversible process; the lack
of reversibility at the PPCE is probably due to the fact that quinone
groups are not produced during the pyrolysis of SU-8 since this
higher than at a PPCE with the same geometric area. These features
constitute an additional advantage to the previously demonstrated
possibility to pattern PPCEs by photolitography in almost any
shape and dimension [10–13].
process is performed operating under an inert N
Fig. 6a shows the effect of the addition of increasing amounts of
BR on the oxygen voltammetric reduction: the height of the
2
atmosphere [10].
The application of PPCE to the study of the electrochemical
behaviour of BR allowed us to clarify some mechanistic points. In
particular, it was demonstrated that the oxidation of BR to BV
involves the formation of a relatively stable intermediate whose
fate depends on the timescale of the experiment. It was also shown
that PPCE can be applied also for the electrochemical generation of
cathodic peak progressively increases while
a concomitant
decrease, up to the total disappearance, of the anodic peak current
is observed. Upon increasing the voltammetric scan rate (data not
shown) only the reduction peak continues to be detected in the
voltammogram.
ꢀꢁ
superoxide and to the study of reactions between O
2
and BR. In
ꢀꢁ
2
living systems, O
can arise from oxygen-consuming reactions
In Fig. 6b we report the CVs recorded in the presence of oxygen,
over a wider cathodic potential window in, before (full line) and
after the addition of 2 (dashed line) and 4 (dotted line) mM of BR. In
these CVs the BR reduction peaks C2 and C3 are detectable, as
shown by the dashed-dotted line CV in Fig. 6b, recorded in de-
aerated DMSO in the presence of 4 mM BR. The dashed line CV in
the same figure shows that, when 2 mM BR is added to DMSO
solution in equilibrium with atmospheric oxygen, only the oxygen
reduction peak is detected, while no superoxide reoxidation peak
is observed in the return scan. At the same time, the peak C2,
relevant to BR reduction, disappears. These evidences indicate that
and its concentration is tightly regulated by numbers of enzymatic
or non-enzymatic reactions [39], including reaction with BR. The
data shown above, obtained in an aprotic solvent, enable to
ꢀꢁ
conclude that bilirubin acts as O
2
scavenger. Since the solvent
DMSO loosely simulates lipid environments where bilirubin
distributes (e.g. phospholipid bilayers, [40]), these data suggest
ꢀꢁ
that BR may be an effective
O
2
scavenger in biological
membranes, protecting lipids from peroxidation and subsequent
loss of barrier integrity [41].
Acknowledgments
ꢀ
ꢁ
electrogenerated O
2
effectively reacts and consumes all the BR
present at the interface, transforming it into a species which is not
reducible within the investigated potential window. When the
concentration of BR is increased to 4 mM (see Fig. 6b, dotted line),
peak C2 is now detected, however with a peak current significantly
smaller than that recorded in the same solution, after deaeration
This work was supported financially by the Cross-Border
Cooperation Italy-Slovenia Program 2007-2013 - Strategic Project
TRANS2CARE and, partially, by MIUR (Rome), project PRIN
2010AXENJ8, and Veneto Region. We thank Ermelinda Fioriniello
for preliminary measurements with GCE and Dr. Veronica Vascotto
for the AFM measurements.
(
Fig. 6b, dashed-dotted line). This indicates that 4 mM BR still
ꢀ
ꢁ
reacts with O
2
, but an excess of BR remains unreacted at the
ꢀꢁ
interface. All the above evidences support a significant O
2
Appendix A. Supplementary data
scavenging capability by BR as well as the occurrence of a
stoichiometric reaction between the tetrapyrrole and the radical,
so that, in the experimental conditions used here, a 2 mM
concentration of BR is high enough to consume all the superoxide
electrogenerated at the PPCE/DMSO interface. Only one previous
report performed by Hu and coworkers in dimethylformammide
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.elec-
tacta.2014.09.046.
References
[
38], described a possible reaction between superoxide and BR. It is
[
1] K. Kinoshita, Carbon Electrochemical and Physico-chemical Properties, John
Wiley & Sons, New York, 1988.
worth stressing that these previous authors focused their study
only on observing changes in the O
ꢀꢁ
2
/O
2
CV pattern without
[2] R.L. McCreery, Chem. Rev. 108 (2008) 2646–2687.
demonstrating the simultaneous consumption of BR, as done in the
present work.
[3] R.N. Adams, Electrochemistry at Solid Electrodes, Marcel Dekker, New York,
969.
4] R.L. McCreery, K.K. Cline, Carbon Electrodes, in: P.T. Kissinger, W.R. Heineman
Eds.), Laboratory Techniques Electroanalytical Chemistry, Marcel Dekker,
1
[
(
New York, 1996, pp. 293–332.
6
. Conclusions
[
[
5] A.L. Beilby, W. Brooks, G.L. Lawrence, Anal. Chem. 36 (1964) 22–26.
6] I. Švancara, K. Kalcher, A. Walcarius, K. Vytras, Electroanalysis with Carbon
Paste Electrodes, CRC Press, Boca Raton (FL USA), 2012.
7] R.L. McCreery, in A.J. Bard, Ed., Electroanalytical Chemistry, vol 18, Marcel
Dekker, 1991, New York.
[8] W.E. Van der Linden, J.W. Dieker, Anal Chim. Acta 119 (1980) 1–24.
9] K. Izutsu, Electrochemistry in Nonaqueous Solutions, 2nd edition, Wiley-VCH
Verlag GmbH & Co. KGaA, Weinheim, 2009 Chapter 8.
10] S. Ranganathan, R.L. McCreery, S.M. Majji, M. Madou, J. Electrochem. Soc. 147
The results obtained in this study indicate that PPCEs can be
[
successfully employed for performing reliable voltammetric
measurements in aprotic solvents. The comparison of the
electrochemical behaviour of PPCE vs. GCE puts is evidence some
significant differences and advantages for the former. First of all,
PPCEs present a very small double-layer charging current with a
capacitance which is less than half that of GCEs. This is mainly a
consequence of the very smooth surface of the PPCE, as detected by
AFM measurements. The identity between geometrical and “real”
surface make PPCEs the electrodes of choice for performing
electroanalytical measurements with very low background current
[
[
(2000) 277–282.
[
11] J.I. Heo, D.S. Shim, G. Turon Teixidor, S. Oh, H. Shin, M.J. Madou, J. Electrochem.
Soc. 158 (2011) J76–J80.
12] R.R. Kamath, M.J. Madou, Anal. Chem. 86 (2014) 2963–2971.
[
[13] A. Mardegan, R. Kamath, S. Sharma, P. Scopece, P. Ugo, M. Madou, J.
Electrochem. Soc. 160 (2013) B132–B137.
[
[
14] X. Wang, J.R. Chowdhury, N.R. Chowdhury, Current Paediatrics 16 (2006)
0–74.
7
(
that is improved S/N ratio) as well as for the accurate evaluation of
15] N.R. Chowdhury, Y. Lu, J.R. Chowdhury, in The Textbook of Hepatology: From
Basic Science to Clinical Practice, in: J. Rodes, J.-P. Benhamou, A. Blei, J. Reichen,
M. Rizzetto (Eds.), J. Wiley-Blackwell, New York, 2007 Chapter 2.3.5.
16] J. Fevery, Liver Int. 28 (2008) 592–605.
voltammetrically relevant parameters (e.g. diffusion coefficients)
whose determination is influenced by the exact knowledge of the
electrode area. As far as electrochemistry in aprotic solvents is
concerned, it can be noted that PPCEs are characterized by a
slightly wider accessible potential window than GCEs, particularly
[
[
17] I. Taurino, V. Van Hoof, G. De Micheli, S. Carrara, Thin Solid Films 548 (2013)
546–550.
[
18] M. Rahman, K. Lee, D. Park, M. Won, Y. Shim, Biosens. Bioelectron. 23 (2008)
8
57–864.
19] J. Klemm, M. Prodromidis, M. Karayannis, Electroanalysis 12 (2000) 292–295.
[20] A. Fortuney, G. Guilbault, Electroanalysis 8 (1996) 229–232.
+
on the anodic side; for instance, at +0.75 V vs. Fc/Fc the
background current in DMSO, 0.1 M TBABF at a GCE is 3-times
4
[

M. Silvestrini et al. / Electrochimica Acta 147 (2014) 401–407
407
[
21] J. Wang, D. Luo, P. Farias, J. Electronal. Chem. Interfacial Electrochem. 185
1985) 61–71.
[31] S. Daniele, P. Ugo, G.A. Mazzocchin, G. Bontempelli, Anal Chim. Acta 173 (1985)
141–148.
(
[
[
22] J.D. Van Norman, Anal. Chem. 45 (1973) 173–175.
23] P. Longhi, P. Manitto, D. Monti, T. Mussini, S. Rondinini, Electrochim Acta 26
[32] G. Gritzner, J. Kuta. Pure Appl. Chem. 56 (1984) 461–466.
[33] I.-T. Kim, M. Egashira, N. Yoshimoto, M. Morita, Electrochim Acta 55 (2010)
6632–6638.
(
1981) 541–543.
[
24] J.R. Pradko, S. Pons, S. Bandyopadhyay, J.F. McAleer, A.S. Hinman, Bioelec-
[34] E. Fioriniello, Thesis, University Ca’ Foscari of Venice, 2012.
[35] D.T. Sawyer, J.L. Roberts Jr., J. Electroanal. Chem. 12 (1966) 90–101.
[36] D. Vasudevan, H. Wendt, J. Electroanal. Chem. 192 (1995) 69–74.
[37] D.T. Sawyer, G. Chlericato Jr., C.T. Angells, E.J. Nanni Jr., T.T. Suchlya, Anal Chem.
54 (1982) 1720–1724.
trochem. Bioener. 13 (1984) 267–285.
[25] S. Dong, Y. Yang, G. Cheng, J. Electroanal. Chem. 324 (1992) 219–228.
26] J.M. Ribo, J.-A. Ferrera, J. Claret, K. Grubmayr, Bioelectrochem. Bioener. 29
[
(
1992) 1–17.
[
27] J. Niu, S. Dong, Electroanal. 6 (1994) 437–443.
[38] Y. Wei, X. Ji, X. Dang, S. Hu, Bioelectrochem. 61 (2003) 51–56.
[39] S.I. Liochev, Free Radic. Biol. Med. 60 (2013) 1–4.
[40] S.M. Bornsek, L. Ziberna, T. Polak, A. Vanzo, N.P. Ulrih, V. Abram, F. Tramer,
S. Passamonti, Food Chem. 134 (2012) 1878–1884.
[
[
[
28] M. Ch. Slifstein, J. Ariel, Electroanal, Chem. Interfacial Electrochem. 48 (1973)
47–463.
29] F. Moussa, G. Kanoute, C. Herrenknecht, P. Levillain, F. Trivin, Anal. Chem. 60
1988) 1179–1185.
30] A.J. Bard, L.R. Faulkner, Electrochemical Methods. Fundamentals and
Applications, John Wiley & Sons Inc., New York, 2001.
4
(
[41] A.C. Bulmer, H.J. Verkade, K.H. Wagner Prog, Lipid Res. 52 (2013) 193–205.