Substituent effects on the electrocatalytic reduction of oxygen on quinone-modified glassy carbon electrodes

Article abstract of DOI:10.1039/b315963a

The reduction of oxygen catalysed by two sulfur-containing derivatives of anthraquinone has been investigated. The quinones were chemically grafted to a glassy carbon electrode surface and the oxygen reduction reaction in alkaline solution followed a two-

Full text of DOI:10.1039/b315963a

View Online / Journal Homepage / Table of Contents for this issue
R E S E A R C H P A P E R
Substituent effects on the electrocatalytic reduction of oxygen on
quinone-modified glassy carbon electrodes
a
Fakhradin Mirkhalaf, Kaido Tammeveski and David J. Schiffrin*
b
a
a
Chemistry Department, University of Liverpool, Liverpool, UK L69 3BX.
E-mail: d.j.schiffrin@liv.ac.uk
Institute of Physical Chemistry, University of Tartu, Jakobi 2, 51014 Tartu, Estonia
b
Received 9th December 2003, Accepted 30th January 2004
F|rst published as an AdvanceArticle on the web 23rd February 2004
The reduction of oxygen catalysed by two sulfur-containing derivatives of anthraquinone has been investigated.
The quinones were chemically grafted to a glassy carbon electrode surface and the oxygen reduction reaction
in alkaline solution followed a two-electron pathway yielding hydrogen peroxide. The mechanism proposed
corresponds to an electrochemical–chemical (EC) reaction where the semiquinone radical anion
electrochemically formed reacts chemically with oxygen. The influence on reactivity of electron withdrawing
groups present in the quinone molecules has been studied and it is shown that there is an unexpected
small dependence between reactivity and standard potential of the grafted quinones. It is shown that
steric effects appear important in determining their electrocatalytic properties.
would be expected to shift the half wave potential to more
positive values, but their influence on the rate of reaction of
the semiquinone intermediate with molecular oxygen is still
not known.
1
Introduction
The reduction of oxygen on quinone modified carbon surfaces
has been extensively investigated. These studies include
1
,2
1
,4-naphthoquinone, 9,10-anthraquinone-2-sulfonate
3
and
,10-phenanthrenequinone adsorbed on highly oriented pyro-
The purpose of the present work was to compare the kinetics
of oxygen reduction on glassy carbon (GC) electrodes modified
with thiophenyl and benzenesulfonyl derivatives, with that for
the unsubstituted anthraquinone. The influence of the chemi-
cal functionalities on their electrochemical properties was also
studied. These compounds were chosen for two reasons, to
include groups close to both carbonyl functions and to
provide substituents with widely different electron withdrawing
properties. The concentration of KOH was chosen the same as
in previous studies to allow a comparison of the kinetic
9
lytic graphite (HOPG), and anthraquinone derivatives
4
–7
adsorbed on glassy carbon (GC);
1,4-naphthoquinone and
8
,9
3,8–12
its derivatives; anthraquinone (AQ) and its derivatives
and phenanthrenequinone (PQ) covalently attached to GC.
Carbon paste electrodes modified with 1,4-naphthoquinone
derivatives have also been employed for this purpose. A large
increase in the electrocatalytic effect of AQ modified GC on
the rate of oxygen reduction by employing ultrasonic waves
1
3
1
4
1
5
has recently been observed. It has also been shown that the
reduction behaviour is significantly altered when the
10,12,13
parameters calculated.
O
2
Grafting of the quinones was carried out by the electro-
chemical reduction of the corresponding diazonium salts
according to:
1
6
anodically pre-treated GC electrode is modified with AQ.
Quinone grafting by the electrochemical reduction of diazo-
nium salts of anthraquinone
on glassy carbon leads to a considerable improvement of the
rate of oxygen reduction to hydrogen peroxide. The reduction
rate is proportional to the surface concentration of the semi-
quinone radical formed by the electrochemical reduction of
the quinone groups.
oxygen to yield superoxide (O
17
1
0,12
13
and phenanthrenequinone
RN2 þ eꢁ ! R þ N2
þ
ꢀ
ðIVÞ
ðVÞ
ꢀ
R þ C ! R C
C represents here the carbon surface. This method was
previously employed for the attachment of AQ and PQ to a
1
0,12,13
The radical reacts with molecular
) which either disproportion-
ꢀ
ꢁ
10,13
18
2
GC electrode
and to carbon powder.
ates or is further reduced to form the hydroperoxide anion in
alkaline solution. The reaction follows an EC mechanism:
Thiophenyl and benzenesulfonyl derivatives of AQ
(compounds AQ1 and AQ2, Scheme 1) were synthesized and
used for the electrochemical investigations in organic media
Q þ eꢁ ! Qꢀꢁ
ðIÞ
ðIIÞ
(
Scheme 1). Diazonium tetrafluoroborate salts of the above
kc
compounds (compounds AQ3 and AQ4) were synthesized
for the modification of GC surfaces. Compounds AQ3 and
AQ4 had two points of attachment to the carbon surface.
The structure of the derivatives employed was dictated by syn-
thetic requirements, which made difficult the preparation of a
mono-derivative. Although this may give rise to some hetero-
geneity in the organisation of the grafted quinones, the sub-
strate itself is so highly heterogeneous that the possibility of
multiple orientations of the quinone is not important. For qui-
ꢀ
ꢁ
ꢀꢁ
Q
þ O2 ƒƒƒ! O þ Q
2
ꢁ
e
;H2O
ꢀ
ꢁ
ꢁ
2
O
ƒƒƒƒ! HO
ðIIIÞ
2
where Q is the surface quinone. Reaction (II) is assumed to be
the rate-determining step. Thus, carbon surfaces modified with
AQ derivatives are good electrocatalytic candidates for the
production of hydrogen peroxide. Different quinones have
been employed for the modification of carbon electrodes, but
the influence of the attached chemical functionalities on the
rate of oxygen reduction is still not understood. The presence
þ
none molecules attached by one end only, the –N2 residue
will decompose to yield a –H termination. The electrochemical
parameters obtained for the kinetics of oxygen reduction on
2 2
of electron-withdrawing groups (e.g. –CN, –SO , –NO )
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 1 3 2 1 – 1 3 2 7
1321

View Online
Scheme 1 Chemical structures of the quinone compounds used.
these electrodes were compared with previous measurements
with unsubstituted AQ.
grafted on GC from a 10 mM solution in DMF containing
0.1 M TBABF . The electrode was polished with 1 and 0.3
1
0,12
4
mm alumina and the subsequent grafting procedure was similar
to that previously described. This involved cycling the poten-
1
0
ꢁ1
2
Experimental
tial four times between þ0.75 and ꢁ0.45 V vs. SCE at 0.1 V s
and finally keeping the electrode at ꢁ0.2 V in the same solution
for 2 min to ensure completion of the reaction. The modified
electrodes were rinsed with DMF, acetone and water to
remove physically adsorbed compounds and then voltammo-
grams were recorded in 0.1 M KOH to check the surface
coverage obtained. Electrode modification with physically
adsorbed compounds was performed by placing a drop of 1
mM solution of the corresponding compound in acetonitrile
on the electrode surface and leaving it to evaporate in air. A
saturated calomel electrode (SCE) was used as the reference
and all potentials are referred to this electrode. For compari-
son with other reference electrodes, the half wave potential
Compounds AQ1 and AQ2 were synthesised in the same way
as AQ3 and AQ4 (see below) except that the starting material
was thiophenol instead of p-aminothiophenol.
Compound AQ3 was synthesized from bis-1,5(4,4 -
aminothiophenylanthraquinone). The latter was prepared with
0
90% yield from the reaction of 20 mmol of 1,5-dichloro-AQ
and 40 mmol of p-aminothiophenol in dimethylformamide
(
1
DMF) in the presence of potassium carbonate at 110–
ꢂ
30 C and the product recrystallised from DMF. Elemental
analysis gave C: calc. 69.00%, found 68.48%, H: calc. 3.56%,
found 3.69%, N calc. 6.19%, found 5.90%, S calc. 14.17%,
found 14.32%. This diamino compound was diazotised in
acetic acid containing fluoroboric acid by the addition of
þ
of the ferrocene/ferrocenium couple (Fc/Fc ) was measured
ꢂ
and a value of 0.415 V with respect to the SCE used was found.
All experiments were carried out in a three-compartment glass
cell at room temperature. The solutions were deoxygenated
with pure argon. Both argon and high purity oxygen were
supplied by BOC gases, UK.
sodium nitrite at 10 C. For the synthesis of AQ4, the amino
groups were protected by acetylation in acetic acid and acetic
anhydride before diazotation. The thioether groups were then
oxidised in excess of 30% hydrogen peroxide in boiling acetic
acid and recrystallised from acetic acid. Elemental analysis
0
Non-linear regression analysis (NLR) fits were performed
1
for the product, bis-1,5(4,4 -N-acetylamino)benzenesulfonyl-
using Origin software (version 6, Microcal Inc.).
anthraquinone was as follows: C: calc 59.79% found 58.22%,
H: calc 3.67, found 3.84, N: calc. 4.64% found 4.45%. The
acetylated compound was then hydrolysed to the amine by
reacting with KOH and diazotised into diazonium fluoro-
borate. Dication main peaks for AQ3 and AQ4 were observed
at 239 and 271 Da, respectively, with electrospray mass
spectrometry in accordance with their molecular weights. All
the chemicals were used with the highest purity available.
A commercial GC disk–Au ring electrode (Pine Instru-
ment Company, USA) with a disk geometrical area of 0.164
3 Results and discussion
3
.1 Compounds AQ1 and AQ2 in DMF
Fig. 1 shows voltammograms of AQ1 at different sweep rates.
A large peak separation (ca. 100 mV) that increases with sweep
rate is observed. These results show only a one-electron
transfer reaction with an average half-wave potential extra-
polated to zero sweep rate of ꢁ0.778 ꢃ 0.005 V. A second
reduction wave could only be observed at potentials more
2
cm was employed for the rotating disk electrode (RDE)
and rotating ring disk electrode (RRDE) measurements. The
electrode collection efficiency (N) was determined with ferro-
cene in DMF as 0.22. An AFMSRX rotator and MSRX speed
controller (Pine Instrument Company, USA) were employed.
The potential was controlled with a PGSTAT20 (Eco-Chemie
B.V., The Netherlands) potentiostat/galvanostat running
under the General Purpose Electrochemical System (GPES)
software.
p
c
1/2
negative than ꢁ1.4 V. The current function I /n
and the
ratio of anodic to cathodic currents (I /I ) were independent
c
a
p
p
of sweep rate. I and Ia are the cathodic and anodic peak
c
p
p
currents and n is the sweep rate. Water has a strong influence
on the electrochemistry of quinones and for this reason,
voltammograms were also recorded for quinone solutions to
which activated alumina (heated to 230 C overnight) had been
added to eliminate residual water. No difference in the current
and potential of the first wave was observed. Voltammograms
ꢂ
The electrochemical properties of compounds AQ1 and
AQ2 were studied in 0.5–5 mM solutions in DMF containing
0
base electrolyte. TBABF (99.9%, Aldrich) was vacuum dried
.1 M tetrabutylammonium tetrafluoroborate (TBABF
4
) as
4
2
were also recorded with the addition of 10% H O to the DMF.
and DMF (Aldrich, anhydrous 99.8%) was dehydrated over
molecular sieves before use. Compounds AQ3 and AQ4 were
Although a positive shift (ca. þ200 mV) was observed for the
second peak, the quinone/semiquinone reduction wave was
1
322
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 1 3 2 1 – 1 3 2 7
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4

View Online
Fig. 1 Cyclic voltammograms of compound AQ1 in Ar-saturated
DMF containing 0.1 M TBABF and 5 mM AQ1 at different sweep
rates: 20, 40, 80, 150, 250, 400 mV s
Fig. 3 Koutecky–Levich plots of compound AQ1 in DMF at
different potentials (data from Fig. 2(a)).
4
ꢁ1
.
unaffected confirming that water does not influence the electro-
chemical behaviour studied.
Fig. 2 shows rotating ring-disk polarisation curves for AQ1
in DMF and some of the corresponding Koutecky–Levich
observed at ꢁ0.55 and ꢁ0.85 V. A further reduction process
is also observed for potentials more negative than ꢁ1.2 V,
1
9
probably due to the reduction of the benzenesulfonyl groups.
The first peak corresponds to a reversible one-electron transfer
process (Fig. 5) to yield the corresponding semiquinone. The
radical formed is stable in the time scale of the voltammetric
(
K.–L.) plots at different potentials are shown in Fig. 3
showing a characteristic quasi-reversible behaviour. The
diffusion coefficient was calculated from the limiting slopes,
assuming a one-electron reaction (Table 1).
Fig. 4 shows a cyclic voltammogram of compound AQ2 in
DMF. In this case, two well-separated reduction peaks are
a
p
c
p
experiments, as indicated by a current ratio I /I close to unity
and independent of sweep rate. The current efficiency for the
ring is close to 100% and independent of rotation rate,
confirming that the reduction to the semiquinone is a simple
reversible one electron-transfer reaction. Fig. 6 shows RRDE
polarisation curves for this compound and the corresponding
Koutecky–Levich plots are presented in Fig. 7. From these
results and the sweep rate dependence of the cathodic peak
potentials (plot not shown), values of n and D were calculated
ꢁ
6
2
ꢁ1
as 0.98 and 6.1 ꢄ 10 cm s , respectively.
The kinetics parameters were calculated using a non-linear
regression analysis (NLR) by fitting the experimental data
to the equation for a quasi-reversible reaction:
1
0
2
0
0
1
1
1 þ bꢁ1e
f ðEꢁE Þ
jD
¼
ꢁ
þ
ꢁaf ðEꢁE Þ
FC k e
ð1Þ
0
ꢅ
0
j
2
)
/3
where b ¼ (D
R
/DOx
and f ¼ F/RT and the other terms
have their usual meaning. The diffusion coefficient for the
reduced and oxidised species was considered to be equal.
Averages of the results obtained from different rotation rates
are given in Table 1. In order to ascertain the validity of the
parameters obtained from NLR, polarisation curves for ferro-
cene in DMF as a standard reversible redox compound were
also measured on the same electrode. This is particularly
important since the behaviour of GC surfaces is strongly
affected by its variable chemical composition determined by
the experimental conditions and/or the method of surface pre-
2
1
paration, as already described by Kneten and McCreery. The
electrochemical standard rate constant however, could not be
determined by the NLR method employed since no difference
in the calculated fit and measured current could be found for
Table 1 Reduction parameters obtained from NLR analysis of the
quinones studied in DMF solution
3
10 k /
0
D/
2
cm s
ꢁ1
ꢁ1
Compound
E1/2/V
a
n
cm s
ꢁ
ꢁ
ꢁ6
6
6
Fig. 2 RRDE current–potential curves for compound AQ1 in
Ar-saturated DMF containing 0.1 M TBABF and 5 mM AQ1.
4
AQ1
AQ2
AQ
ꢁ0.778 ꢃ 0.005 4.6 ꢄ 10
ꢁ0.511 ꢃ 0.001 6.1 ꢄ 10
ꢁ0.825 ꢃ 0.006 8.9 ꢄ 10
0.66 ꢃ 0.08
0.55 ꢃ 0.02
0.47 ꢃ 0.03
1
1
1
1.0 ꢃ 0.3
16 ꢃ 1
Rotation rates: (1) 200, (2) 400, (3) 800, (4) 1500, (5) 2500 and (6)
.
6.7 ꢃ 0.9
3
500 rpm; (a) disk, (b) ring. Sweep rate ¼ 10 mV sꢁ1
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 1 3 2 1 – 1 3 2 7
1323

View Online
Fig. 4 Cyclic voltammogram of compound AQ2 in Ar-saturated
DMF containing 0.1 M TBABF
4
and 0.58 mM AQ2; sweep rate ¼ 20
ꢁ
1
.
mV s
0
k > 0.1. Thus, the rate constants measured in DMF are not
dependent on the electrode surface treatment observed by
McCreery and co-workers for aqueous redox couples.
2
1,22
For comparison purposes, the same measurements were car-
ried out with unsubstituted anthraquinone (Table 1). The rate
of the electron transfer reaction for AQ1 is lower than AQ2
and AQ. The addition of thioether functionalities to AQ shifts
the potential 47 mV to a more positive value. This shift is much
larger (i.e. 314 mV) for the benzenesulfonyl derivative (AQ2)
due to the stronger electron withdrawing properties of
this group, which results in the stabilisation of the radical
intermediate.
Fig. 6 RRDE current–potential curves for AQ2 in Ar-saturated
DMF containing 0.1 M TBABF and 0.58 mM AQ2; rotation
rate ¼ (1) 100, (2) 200, (3) 400, (4) 800, (5) 1100, (6) 1400, (7) 1800
3
.2 Compounds AQ3 and AQ4 grafted to GC
4
Fig. 8 shows the first and second sweeps for the electro-
chemical surface grafting of the quinones by reduction of the
corresponding diazonium salt. A large cathodic wave is
observed for the first sweep, but this is absent in the second
sweep, clearly indicating an almost complete blocking of the
surface by the grafted quinone.
The cyclic voltammograms of the quinone-modified GC
electrodes in Ar-saturated 0.1 M KOH (Fig. 9) present a large
peak-to-peak separation for AQ3. The peak currents show a
linear relationship with sweep rate, a clear indication of a
surface process (results not shown). The charge densities of
the cathodic peaks correspond to surface coverages of
ꢁ1
and (8) 2100 rpm; (a) disk, (b) ring currents, sweep rate ¼ 10 mV s
.
respectively. These values can be compared with that for
ꢁ10
23
ꢁ2
,
9
,10-anthraquinone-2-monosulfonate, 1.4 ꢄ 10
mol cm
obtained on the basal plane of pyrolytic graphite. The differ-
ence between these surfaces is due to the surface roughness of
GC and the orientation of the grafted molecules. It can be con-
cluded that the surface coverages obtained with GC were close
to a monolayer.
The large peak-to-peak separation obtained is unexpected
for compounds attached to a surface and indicates that the
surface reactions proceed by different routes for reduction
.0 ꢄ 10ꢁ and 4.3 ꢄ 10
10
ꢁ10
ꢁ2
mol cm for AQ3 and AQ4,
4
Fig. 5 Cyclic voltammograms of compound AQ2 in Ar-saturated
DMF containing 0.1 M TBABF
and 0.58 mM AQ2: sweep rate ¼ 10,
0, 40, 80, 120, 160, 250, 400 mV s
4
Fig. 7 Koutecky–Levich plot obtained from the results shown in
Fig. 6. Potential: (1) ꢁ0.53, (2) ꢁ0.55, (3) ꢁ0.6, (4) ꢁ0.65 and (5) ꢁ0.7 V.
ꢁ1
2
.
1
324
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 1 3 2 1 – 1 3 2 7
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4

View Online
Table 2 Half-wave potentials of AQ derivatives attached to a GC
ꢁ
1
.
electrode in 0.1 M KOH, sweep rate ¼ 40 mV s
Electrode
E
1/2/V (vs. SCE)
1
0
10
GC/AQ
GC/AQ3
ꢁ0.845
ꢁ0.873 ꢃ 0.005
ꢁ0.694 ꢃ 0.001
ꢁ0.699 ꢃ 0.005
GC/AQ4
AQ2 physically adsorbed on GC
3 2 6 5 2
CF SO and C H SO groups have the largest electron-
withdrawing property on the ionisation of benzoic acid with
s values of 0.93 and 0.7, respectively. The value of s for
CF S, however, is only 0.5 indicating a small electron affinity
3
of thioether groups. A correlation between Hammett substitu-
ent parameters and the shift of the reduction potentials of
Fig. 8 First (a) and second (b) sweeps for the grafting of compound
AQ3 on GC in Ar-saturated DMF containing 0.1 M TBABF and 10
25
benzoquinone (BQ) has been demonstrated. It was shown
4
ꢁ1
that by increasing the s value of the substituent the positive
shift of the BQ reduction potential is increased. Therefore, it
can be concluded that the thioether substituent in AQ1 does
not have a large effect on the reduction potential due to the
weak electron-withdrawing ability of the –SAr group. In
mM AQ3; sweep rate ¼ 100 mV s
.
and oxidation, as previously observed for other quinones.
These reactions follow electron and proton transfer steps. In
the present case, it is proposed that the reduction sequence
follows electron transfer to yield the corresponding semiqui-
2
contrast, the compound containing –SO Ar groups (AQ2)
shows a large positive shift with respect to AQ both in DMF
and in alkaline solution when chemically attached to GC
(Tables 1 and 2).
2
4
none:
AQ þ eꢁ ! AQꢀꢁ
ðVIÞ
followed by either
3
.3 Electrocatalysis of O reduction by grafted quinones
2
AQ þ eꢁ ! AQ2ꢁ þ Hþ ! HAQꢁ
ꢀ
ꢁ
ðVIIÞ
2
Fig. 10 shows RDE results for O reduction on a GC/AQ3
electrode at different rotation rates; the corresponding K.–L.
plots are given in Fig. 11. As can be seen, the attached quinone
considerably enhances the rate of O reduction in comparison
or
AQ þ Hþ ! HAQ þ eꢁ ! HAQꢁ
ꢀ
ꢁ
ꢀ
ðVIIIÞ
2
to a bare GC electrode. The value of n calculated from these
results using the K.–L. plots in Fig. 11 is 1.95 ꢃ 0.05 showing
This mechanism involves two consecutive reductions. Either of
the reactions leads to the formation of the corresponding
quinol anion according to the overall reaction:
2
that the O reduction sequence stops at the peroxide stage.
Fig. 12 compares O reduction results for GC/AQ4 and for
2
1
0
bare GC electrodes. In contrast to previous work for GC/
AQ, the effect of the native quinone groups of GC is not
apparent at low rotation rates. This reflects the closeness of
the values of the redox potentials of the native and attached
quinone groups (see later). Fig. 13 shows K–L plots of these
results at different potentials. The number of electrons
involved in the reduction is almost independent of potential
and is close to two (Fig. 13), i.e., oxygen reduction also stops
at the peroxide stage for this compound. The percentage of
AQ þ H þ 2eꢁ ! HAQꢁ
þ
ðIXÞ
The average value of the half-wave potential for AQ3 was
ꢁ
0.873 ꢃ 0.005 V, which is 28 mV more negative than that
1
0
of grafted anthraquinone (Table 2). The value of E1/2 for
the attached benzenesulfonyl derivative (AQ4) is 151 mV more
positive than that of anthraquinone. A similar large positive
shift is also observed for AQ2 physically adsorbed on GC
(
Table 2). The effect of electron-withdrawing groups on the
Hammett substituent parameters (s) and thus on the oxidation
potential, has already been described. It is known that
2
5
Fig. 10 RDE polarisation curves for oxygen reduction on a GC/
AQ3 electrode in O
2
-saturated 0.1 M KOH. Rotation rates ¼ (1)
3
60, (2) 610, (3) 900, (4) 1500, (5) 2500, (6) 3800 and (7) 4600 rpm.
Fig. 9 Cyclic voltammograms of GC grafted with AQ3 and AQ4 in
Ar-saturated 0.1 M KOH solution; sweep rate ¼ 100 mV s
For comparison (dashed line), the polarisation curve for a bare GC
electrode in the same solution at 960 rpm is shown.
ꢁ
1
.
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 1 3 2 1 – 1 3 2 7
1325

View Online
2
Fig. 11 Koutecky–Levich plot for GC/AQ3 in O -saturated KOH at
potentials: ꢁ0.7, ꢁ0.8, ꢁ1 and ꢁ1.1 V (from top to bottom).
2
Fig. 13 Koutecky–Levich plot of GC/AQ4 in O -saturated 0.1 M
KOH at different potentials: ꢁ0.65, ꢁ0.8, ꢁ1.0 and ꢁ1.15 V, respec-
tively. The inset shows the dependence of the number of electrons
transferred on potential.
peroxide formation (F) was calculated using eqn. (2) as
previously proposed:
2
6
2
constant for the chemical reaction of O with the semiquinone
2
00 ꢄ IR=N
F ¼
ð2Þ
radical (reaction (II)) for the surface quinone groups and G is
i
ID þ IR=N
the quinone surface concentration. Table 3 compares the
kinetics parameters of oxygen reduction electrocatalysed by
the different grafted quinones.
The standard potential of the native GC quinone groups,
where ID and IR are disk and ring currents and N is the elec-
trode collection efficiency. An average of peroxide production
yield of 89 ꢃ 2% and 97 ꢃ 2% was calculated for the electrode
modified with compounds AQ3 and AQ4, respectively.
0
E , falls within ꢃ50 mV for the different quinones studied.
1
This quantity is not easy to measure, in particular for AQ4, due
to the potential overlap of the reduction range for both native
and grafted quinones. In spite of this problem, the standard
potential of the native quinones is not strongly affected by
the grafted material. The effect of attachment appears to be
Attempts were made to use the electrodes modified with
physically adsorbed AQ1 and AQ2 for comparing their
kinetics parameters with those of quinones chemically attached
to the electrode surface. However, the physisorbed compounds
were rapidly desorbed from the surface on cycling in alkaline
solution. Similar behaviour was previously observed with
0
more noticeable for k . This quantity refers to unit geometric
area and the trend observed, AQ > AQ3 > AQ4, probably
reflects the greater coverage of active free carbon surface by
1
0
unsubstituted AQ physically adsorbed on GC.
The kinetics parameters for oxygen reduction were calcu-
lated from the RDE measurements using the non-linear regres-
the grafted quinones. The component k G does not change
1
1
1
2,13
much for the different quinones studied. This term corresponds
to the reactivity of the semiquinone formed by reduction of the
native quinone groups with oxygen. Since the thermodynamic
driving force for the radical–molecular oxygen reaction is not
very strongly dependent on the grafted quinone (see entries in
sion analysis previously described.
are the same as those in previous publications.
The symbols employed
Briefly,
1
0,12,13
both the native GC quinone groups and the grafted quinones
are considered in the analysis (subscripts 1 and 2, respectively).
0
0
E is the standard potential of quinone i, k is the electroche-
i
mical rate constant for oxygen reduction on the free glassy car-
column 1 of Table 3), it is not surprising that k G is unaffected
1 1
by the properties of the attached quinones.
Perhaps the most interesting observation is the comparison
of the chemical rate constant for reaction (II) for the different
i
bon surface, a is the transfer coefficient, DE is the difference
between the standard potentials of the Q /Q and Q/Q
i
couples for the different surface quinone species, k is the rate
ꢀ
ꢁ
2ꢁ
ꢀ
ꢁ
9
8
quinones. The values of k calculated were 1.2 ꢄ 10 , 3 ꢄ 10
2
8
3
ꢁ1 ꢁ1
and 1.6 ꢄ 10 cm mol
s
for AQ, AQ3 and AQ4, respec-
tively. The last column in Table 3 reveals that there is a
systematic decrease in its value as the standard potential of
the quinone is made more positive and the thermodynamic
driving force for electron transfer is increased, i.e. for more
negative values of the Gibbs free energy of reaction. This
effect, however, is relatively small. The trend is what would
be expected for a simple activated chemical reaction, with an
increase in the activation energy as the different substituents
increase the thermodynamic stability of the intermediate
radical. It is somewhat surprising though, that such small
2
variations in k are observed considering that the change in
standard potential going from AQ to AQ4 is 151 mV. This
indicates that other parameters must be considered to rationa-
lise the values of the chemical rate constant measured, such as
surface orientation of the electrocatalytic species and conse-
quent steric effects of the intermediates of the reaction between
the semiquinone radical and oxygen.
Fig. 12 RDE polarisation curves for oxygen reduction on a GC/
AQ4 electrode in O
(
(
In relation to the above, it is interesting to note that the rate
constants for the reduction of the quinones in solution do not
follow the expected Marcus normal behaviour. This would
2
-saturated 0.1 M KOH; rotation rates ¼ (1) 200,
2) 300, (3) 600, (4) 900, (5) 1500 and (6) 2500 rpm. For comparison
dashed line), the polarisation curve for a bare GC electrode in the
2
0
0
same solution at 960 rpm is shown.
require the values of k to follow the order AQ < AQ1 < AQ2,
1
326
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 1 3 2 1 – 1 3 2 7
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4

View Online
Table 3 Summary of non-linear regression parameters for oxygen reduction on GC electrodes grafted with AQ derivatives
0
0
4
10 k /cm s
0
ꢁ1
ꢁ1
ꢁ1
Electrode
E /V
E /V
a
DE
1
/V
DE
2
/V
k
1
G
1
/cm s
2 2
k G /cm s
1
2
1
0
GC/AQ
GC/AQ3
GC/AQ4
ꢁ0.56 ꢃ 0.01
ꢁ0.50 ꢃ 0.02
ꢁ0.64 ꢃ 0.03
ꢁ0.9 ꢃ 0.02
ꢁ0.76 ꢃ 0.03
ꢁ0.67 ꢃ 0.01
3.2 ꢃ 1.5
2.5 ꢃ 1
0.23 ꢃ 0.01
0.36 ꢃ 0.03
0.30 ꢃ 0.04
ꢁ0.25 ꢃ 0.02
ꢁ0.10 ꢃ 0.01
ꢁ0.39 ꢃ 0.08
ꢁ0.3 ꢃ 0.1
ꢁ0.22 ꢃ 0.01
ꢁ0.50 ꢃ 0.03
0.019 ꢃ 0.002
0.017 ꢃ 0.005
0.026 ꢃ 0.002
0.3 ꢃ 0.15
0.12 ꢃ 0.06
0.07 ꢃ 0.04
0.6 ꢃ 0.1
which is not the order observed. The likely reason for this is
the nature of the substituents that determines the degree of
proximity of the redox centre to the surface. This seems
to be the overriding effect, determining the value of the electro-
chemical rate constant.
References
1
2
E. Yeager, Electrochim. Acta, 1984, 29, 1527.
Z. W. Zhang, D. A. Tryk and E. B. Yeager, in Proc. Workshop on
the Electrochemistry of Carbon, ed. S. Sarangapani, J.R. Akridge
and B. Schumm, The Electrochemical Society, Pennington, NJ,
1
984, p. 158.
3
4
5
6
7
8
9
0
1
M. S. Hossain, D. Tryk and E. Yeager, Electrochim. Acta, 1989,
34, 1733.
A. Salimi, H. Eshghi, H. Sharghi, S. M. Golabi and M.
Shamsipur, Electroanalysis, 1999, 11, 114.
A. Salimi, M. F. Mousavi, H. Eshghi, H. Sharghi and M.
Shamsipur, Bull. Chem. Soc. Jpn., 1999, 72, 2121.
T. Nagaoka, T. Sakai, K. Ogura and T. Yoshino, Anal. Chem.,
4
Conclusions
A large positive shift of the redox potential of compound AQ2
in DMF solution compared to AQ1 and AQ was observed.
The rate constants for the one-electron reduction of AQ1
and AQ2 to the corresponding semiquinone radical in DMF
were determined using RDE data and non-linear regression
analysis. The electron transfer reaction for AQ2 is 16 times fas-
ter than for AQ1 in DMF. This could be due to the delocalisa-
tion of electrons caused by the sulfone groups, thus stabilising
the radical formed. A positive potential shift for covalently
attached AQ4 compared with grafted AQ is also observed in
alkaline solution, which is in agreement with that expected
from inductive effects of the benzenesulfonyl group on radical
formation as also observed in DMF. Therefore, reaction (I) is
thermodynamically more feasible when electron-withdrawing
functionalities are incorporated in the AQ molecule. However,
1
986, 58, 1953.
T. Nagaoka, T. Sakai, K. Ogura and T. Yoshino, J. Chem. Soc.,
Faraday Trans. 1, 1987, 83, 1823.
G. S. Calabrese, R. M. Buchanan and M. S. Wrighton,
J. Am. Chem. Soc., 1982, 104, 5786.
G. S. Calabrese, R. M. Buchanan and M. S. Wrighton,
J. Am. Chem. Soc., 1983, 105, 5594.
K. Tammeveski, K. Kontturi, R. J. Nichols, R. J. Potter and
D. J. Schiffrin, J. Electroanal. Chem., 2001, 515, 101.
V. A. Bogdanovskaya, M. R. Tarasevich, M. L. Hidekel, G. I.
Kozub and S. B. Orlov, Elektrokhim., 1984, 20, 164.
1
1
12 A. Sarapuu, K. Vaik, D. J. Schiffrin and K. Tammeveski,
J. Electroanal. Chem., 2003, 541, 23.
1
1
1
3
4
5
K. Vaik, A. Sarapuu, K. Tammeveski, F. Mirkhalaf and D. J.
Schiffrin, J. Electroanal. Chem., 2004, 564, 159.
S. M. Golabi and J. B. Raoof, J. Electroanal. Chem., 1996,
the chemical rate parameter (k
2 2
G ) of oxygen reduction for the
surface bound molecules is lower for compounds AQ3 and
AQ4 in comparison with that of AQ. It is proposed that these
functionalities reduce the rate of the reaction (II) due to both
the stabilisation of the semiquinone radical and by steric
hindrance by these groups. Since the rate determining step is
reaction (II), a suitable functionality to improve reaction (I)
does not necessarily enhance the rate of the chemical reaction
between the semiquinone radical anion and molecular oxygen.
This is an indication of kinetic control of reaction (II), which is
not influenced by the thermodynamic feasibility of the electro-
chemical step. The number of electrons transferred per O2
molecule for all cases was close to two indicating that the
reduction of oxygen stops at the peroxide stage.
4
16, 75.
A. Salimi, C. A. Banks and R. G. Compton, Phys. Chem. Chem.
Phys., 2003, 5, 3988.
16 K. Vaik, D. J. Schiffrin and K. Tammeveski, Electrochem.
Commun., 2004, 6, 1.
1
7
P. Allongue, M. Delamar, B. De sˆ bat, O. Fagebaume, R. Hitmi,
J. Pinson and J.-M. Sav e´ ant, J. Am. Chem. Soc., 1997,
119, 201.
G. G. Wildgoose, M. Pandurangappa, N. S. Lawrence, L. Jiang,
1
8
T. G. J. Jones and R. G. Compton, Talanta, 2003, 60, 887.
19 L. Horner and H. Neumann, Chem. Ber., 1965, 98, 1715.
20
21
22
23
24
A. J. Bard and L. Faulkner, Electrochemical Methods, Wiley,
New York, 2nd edn., 2000.
K. R. Kneten and R. L. McCreery, Anal. Chem., 1992,
6
4, 2518.
M. T. McDermott, K. Kneten and R. McCreery, J. Phys. Chem.,
1992, 96, 3124.
A. P. Brown and F. C. Anson, Anal. Chem., 1977,
49, 1589.
G. J. Gordillo and D. J. Schiffrin, Faraday Discuss., 2000,
Acknowledgements
Support from the European Union Framework V Growth
programme, CLETEPEG project Contract No. G5RD-CT-
1
16, 89.
2001-00463, is gratefully acknowledged. This work was partly
2
2
5
6
R. M. Scribner, J. Org. Chem., 1966, 31, 3671.
M. Bron, P. Bogdanoff, S. Fiechter, I. Dorbandt, M. Hilgendorff,
H. Schulenburg and H. Tributsch, J. Electroanal. Chem., 2001,
500, 510.
supported by the Estonian Science Foundation. F. M. is
grateful to Kashan University, Iran for the award of a study
leave at Liverpool University.
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 1 3 2 1 – 1 3 2 7
1327