Mechanism of glucose electrochemical oxidation on gold surface

Article abstract of DOI:10.1016/j.electacta.2010.04.069

The complex oxidation of glucose at the surface of gold electrodes was studied in detail in different conditions of pH, buffer and halide concentration. As observed in previous studies, an oxidative current peak occurs during the cathodic sweep showing a highly linear dependence on glucose concentration, when other electrolyte conditions are unchanged. The effect of the different conditions on the intensity of this peak has stressed the limitations of the previously proposed mechanisms. A mechanism able to explain the presence of this oxidative peak was proposed. The mechanism takes into account ion-sorption and electrochemical adsorption of OH^-, buffer species (K₂HPO₄/KH₂PO₄) and halides.

Full text of DOI:10.1016/j.electacta.2010.04.069

Electrochimica Acta 55 (2010) 5561–5568
Contents lists available at ScienceDirect
Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta
Mechanism of glucose electrochemical oxidation on gold surface
Mauro Pastaa,b, Fabio La Mantiab,1, Yi Cuib,∗
a
Dipartimento di Chimica Inorganica, Metallorganica e Analitica “Lamberto Malatesta”, Università degli Studi di Milano, Via Venezian 21, 20133 Milano, Italy
Department of Material Science and Engineering, Stanford University, 476 Lomita Mall, Stanford, CA 94305, United States
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 2 March 2010
Received in revised form 12 April 2010
Accepted 18 April 2010
Available online 24 April 2010
The complex oxidation of glucose at the surface of gold electrodes was studied in detail in different
conditions of pH, buffer and halide concentration. As observed in previous studies, an oxidative current
peak occurs during the cathodic sweep showing a highly linear dependence on glucose concentration,
when other electrolyte conditions are unchanged. The effect of the different conditions on the intensity
of this peak has stressed the limitations of the previously proposed mechanisms. A mechanism able to
explain the presence of this oxidative peak was proposed. The mechanism takes into account ion-sorption
Keywords:
Glucose electrochemical oxidation
Glucose sensing
−
and electrochemical adsorption of OH , buffer species (K2HPO4/KH2PO4) and halides.
©
2010 Published by Elsevier Ltd.
Gold electrochemical oxidation
Reaction mechanism
Electrochemical impedance spectroscopy
1
. Introduction
Gold is an attractive metal for the oxidation of glucose, because
its oxidation potential in neutral and alkaline medium is more neg-
ative compared to the other metals [13,14] and therefore has been
extensively examined [4,14–23].
Nikolaeva et al. proposed a mechanism for glucose electro-
oxidation at high potentials, in which a layer of gold oxide formed
on the surface of a gold electrode could have a great catalytic effect
on glucose oxidation [24]. The suggested mechanism was the fol-
lowing:
Electrochemical oxidation of glucose has generated much inter-
est over the years. It has been extensively studied for applications in
glucose–oxygen fuel cells [1] and, especially, in glucose sensors, [2]
whose optimization (in terms of response time, lifetime, sensitiv-
ity and selectivity) is required to improve the treatment of diabetes
mellitus, a chronic disease affecting millions of people around the
world [3].
or
The electrocatalytic oxidation of glucose in alkaline medium
was investigated using Cu, Ni, Fe, Pt and Au electrodes [4]. Among
them, platinum has been the most widely studied [5–9]; in partic-
ular Beden et al. applied a reflectance IR spectroscopic technique to
study the electro-oxidation process of d-glucose at platinum elec-
trodes in alkaline medium [10]. However, platinum also proved to
be extremely non-selective and susceptible to poisoning by various
components of blood and other physiological media over extended
use [9,11,12].
This reaction was followed by rapid electrochemical regenera-
tion of the surface oxide
−
−
AunO
+2OH ꢀ AunOm+H O + 2e
m−1
2
In a subsequent study, Makovos and Liu first identified a posi-
tive current peak during the cathodic sweep in cyclic voltammetry
and highlighted a highly linear dependence between the maxima of
current values and glucose concentration in a wide potential range
depending on the medium composition [25]. They also proposed
∗ Corresponding author. Tel.: +1 6507234613; fax: +1 6507254034.
E-mail address: yicui@stanford.edu (Y. Cui).
1
ISE member.
0
013-4686/$ – see front matter © 2010 Published by Elsevier Ltd.
doi:10.1016/j.electacta.2010.04.069

5
562
M. Pasta et al. / Electrochimica Acta 55 (2010) 5561–5568
a different mechanism for the peak generation, claiming that the
one provided by Nikolaeva et al. did not account for the fact that the
reaction proceeds only when the potential favors a partial reduction
of the gold oxide:
2. Experimental
d(+)-Glucose (dextrose) anhydrous, potassium chloride, sodium
fluoride, PBS (phosphate buffer solution: 100 mM NaCl, 10 mM
KCl, 10 mM KH PO4 and 10 mM K HPO ), potassium phosphate
2
2
4
−
+O2−
AunOm+2e ꢀ AunO
m−1
dibasic and potassium phosphate monobasic were purchased from
Followed by:
2
Sigma Aldrich. A gold pin electrode (surface area 0.0314 cm ) and
−
−
AunO
+2OH ꢀ AunOm+H O + 2e
a platinum counter electrode were purchased from Amel Electro-
chemistry.
m−1
2
Chlorides, amino acids, and human albumin were observed to
inhibit the reaction whereas urea and l-ascorbic acid contributed
a stabilizing effect to the performance of the electrode. They con-
cluded that gold oxide might be successfully utilized as catalyst for
glucose oxidation in a glucose sensor or a biological fuel cell, if a
proper separation of the inhibitors was achieved. In fact, this is the
basis of the pulsed amperometric detection technique (PAD) at gold
electrodes which, coupled with liquid chromatography (LC), gained
prominence for the analysis of complex mixtures of polyalcohols
and carbohydrates in alkaline medium [26].
Nevertheless the presence of the previously reported inhibitors
forbids the direct application of this system in blood, where it is
mainly required. Among these inhibiting species, chlorides are the
most problematic because they are present in a high concentra-
tion in the blood (about 0.1 M) and their separation from glucose is
difficult to achieve.
Electrochemical characterization was carried out using a Bio-
Logic VMP3 potentiostat–galvanostat multichannel equipped with
EIS board. A double junction Ag|AgCl|KCl (3.5 M) reference elec-
trode (RE) was used in the measurement. The double junction was
−
employed to prevent OH diffusion and reaction at the Ag/AgCl
interface; additionally, the RE potential was monitored after each
measurement to confirm that no change had taken place. Before
each measurement, the gold pin electrode surface was activated
−1
and stabilized in 0.1 M KOH by CV scans at 100 mV s between
−
0.7 and 0.8 V vs. RE until stable voltammograms were observed
about 20 scans). All the measurements were performed under inert
nitrogen) atmosphere and room temperature.
(
(
3. Results and discussion
Fig. 1 shows a typical cyclic voltammogram of glucose electro-
Makovos and Liu attempted to use a gold electrode in media
containing chloride ions, but discovered that chloride ions, even at
trace levels, strongly inhibit any response toward glucose.
oxidation in alkaline solution. The discussion of the overall
mechanism is not the aim of this work; a deep analysis has been
already discussed by Beden et al. at platinum electrodes and low
temperature [10]. The aim of the present work is to understand the
influence of different species in the electrolyte on the generation
of peak (*) in Fig. 1.
−
They suggested that the presence of Cl ions caused gold to
dissolve instead of forming an oxide layer:
Au + 4Cl−→ AuCl4− + 3e−
The Pourbaix diagram of Au in aqueous solutions (see Fig. 2)
stresses that in the pH range 7–12, gold is oxidized to Au(III)
hydroxide, the oxidation potential decreasing as the pH is
increased. Above pH 12.5, gold is oxidized to the water-soluble ion
Consequently, the unoxidized gold has no electrocatalytic activ-
ity with respect to the oxidation of glucose.
Many questions remain unanswered by the mechanisms pro-
posed by Nikoleava and Makovos:
HAuO32
−
.
According to the diagram, cyclic voltammetry measurements
performed at gold pin working electrode in 100 mM
a
a
(
(
1) If these mechanisms are correct, the charge under the cathodic
oxidative peak current should show a dependence on the oxide
thickness, not on the glucose concentration.
2) The last step proposed in both mechanisms is gold re-oxidation,
however it is difficult to imagine this re-oxidation taking place
at such a low potential (down to 0.0 V vs. RHE) when the same
reaction (initial formation of the oxide) in the anodic sweep
starts at 0.8 V vs. RHE.
(
(
3) The peak onset is very sharp, but there is also a long tail after
the main peak which is difficult to explain with the proposed
mechanism: instead it seems that the reduction of the oxide
enables the oxidation of the glucose.
4) The data does not conclusively suggest that gold oxide is really
what is formed in these conditions.
In the present study we investigated the effects of varying
chloride concentration, pH and buffer conditions on the electro-
oxidation of glucose at a gold electrode, proposing a mechanism
for the positive current peak formation during the cathodic sweep.
The comprehension of the mechanisms behind the formation of this
peak was necessary to the authors to correctly lay the groundwork
for upcoming work concerning glucose sensing.
Fig. 1. Cyclic voltammetry of a gold pin electrode in 0.1 M KOH without (red curve)
−
1
and with (green curve) 10 mM glucose; scan rate 100 mV s . (For interpretation
of the references to colors in this figure legend, the reader is referred to the web
version of this article.)

M. Pasta et al. / Electrochimica Acta 55 (2010) 5561–5568
5563
Fig. 4. Cyclic voltammetry at
a
gold pin electrode of 100 mM buffer
(
(
K2HPO4–KH2PO4) at different pH: 7.5 (green curve); 9.5 (red curve); 11.5
blue curve) in the presence of glucose (10 mM). Scan rate 20 mV s . (For interpre-
−
1
tation of the references to colors in this figure legend, the reader is referred to the
web version of this article.)
to Au(OH)3 thus avoiding the formation of the tetrachloroaurate
ion. This means that above pH 9, Au(OH) should be formed, even in
3
the presence of chlorides. To demonstrate this, CV measurements
at pH 11.5 (buffered, 100 mM) in the presence of 10 mM glucose
have been performed at different chloride concentrations (Fig. 6).
As expected, in these conditions the chloride ions do not completely
prevent the generation of peak (*), but the intensity of the peak does
◦
Fig. 2. Pourbaix diagram for the system Au–H2O at 25 C. The concentration of all
the soluble species is 10 M. Extract from Ref. [29].
−4
−
decrease as the concentration of Cl increases. The mechanism pro-
posed by Makovos and Liu [25] cannot explain this phenomenon.
We propose another mechanism: the chloride ions can strongly
adsorb to the gold active sites thus inhibiting the oxidative adsorp-
tion of glucose, a key step in the overall oxidation process. To
K HPO –KH PO solution without glucose at different pH values
2
4
2
4
(
see Fig. 3) show a peak relative to the formation of the gold
hydroxide (around 1 V vs. RHE), which is successively reduced in
the negative sweep (0.75 V vs. RHE). Two other peaks, one in the
oxidative (0.55 V vs. RHE) and the other in the reductive scan (0.5 V
vs. RHE) (both clearly visible at pH 11.5), have been attributed to
the chemisorption of the hydroxide ion to the Au surface [27].
When glucose (10 mM) is added to the electrolyte (Fig. 4) the
oxidative peak (*) appears. Its current density depends strongly on
the pH of the solution.
−
Fig. 5 shows the Pourbaix diagram of Au in a 2 M Cl aqueous
solution. The reason indicated by Makovos and Liu for the inac-
−
tivity of gold in the presence of Cl is the dissolution of Au to
−
AuCl4 which prevents the formation of the oxide. However the
Pourbaix diagram shows that above pH 9, gold is directly oxidized
Fig. 3. Cyclic voltammetry at
a
gold pin electrode of 100 mM buffer
(
(
K2HPO4–KH2PO4) at different pH: 7.5 (green curve); 9.5 (red curve); 11.5
blue curve). Scan rate 20 mV s . (For interpretation of the references to colors in
−1
−
◦
−2
−
Fig. 5. Pourbaix diagram for the system Au–H2O–Cl at 25 C. a = 10 M, [Cl ] = 2 M,
pO2 = pH2 = 1 atm. Extract from Ref. [30].
this figure legend, the reader is referred to the web version of this article.)

5
564
M. Pasta et al. / Electrochimica Acta 55 (2010) 5561–5568
Fig. 6. Cyclic voltammetry at a gold pin electrode of 10 mM glucose in 100 mM buffer
K2HPO4–KH2PO4) at pH 11.5 and different chloride concentrations: (blue) without
chloride ions, (red) 50 mM, (green) 100 mM. Scan rate 20 mV s . (For interpretation
of the references to colors in this figure legend, the reader is referred to the web
version of this article.)
Fig. 8. Cyclic voltammetry at a gold pin electrode of 10 mM glucose in 100 mM
(
buffer (K2HPO4–KH2PO4) at pH 11.5 and different fluoride concentrations: (red)
−
1
−1
without fluoride ions, (blue) 50 mM and (black) 100 mM. Scan rate 20 mV s . (For
interpretation of the references to colors in this figure legend, the reader is referred
to the web version of this article.)
prove the adsorption of the chlorides to the surface, electrochemi-
cal impedance spectroscopy (EIS) techniques has been employed at
fixed frequency and potential, while increasing the concentration
been confirmed with the EIS measurement at fixed frequency and
potential, while increasing the concentration of F− (Fig. 9). In this
case we can see a decrease of the Re(Z), while the −Im(Z) value is
−
of Cl . EIS at high frequency enables the gold/electrolyte interface
constant.
−
to be studied. An adsorption of Cl to the surface should increase
The very importance of the adsorbed species on the oxidation
of the glucose gives rise to questions about the possible limitations
generated by other species in solution, like the buffer. To address
this question cyclic voltammetry of glucose oxidation was per-
formed using only NaF (100 mM) as supporting electrolyte (Fig. 10).
The presence of a very low buffer (KH PO \K HPO ) concentration
the differential capacitance; as a consequence the negative imagi-
nary part of the impedance (−Im(Z)) should decrease. A preliminary
EIS was performed in the range from 100 kHz to 100 mHz to estab-
lish the optimal working frequency (5 kHz). The experiment was
performed recording the impedance every 10 s (Fig. 7). Obviously a
decrease of the Re(Z) is observed when increasing the KCl concen-
tration, due to the increase of the conductivity. According to the
results, after an initial period in which the value of −Im(Z) slightly
2
4
2
4
(1 mM) was necessary to obtain a reproducible cyclic voltamme-
try. The anodic limit was also shifted 150 mV toward more positive
potentials with respect to the 100 mM buffer measurement to allow
the Au(OH)3 oxidation peak formation.
−
increases, it decreases for concentrations higher than 40 mM Cl .
The suspected importance of the adsorbed ions on the overall
oxidation mechanism of glucose was confirmed by the replacement
Interesting conclusions can be drawn by comparing Figs. 4 and
10. At pH 7.5 without buffer it is possible to observe the presence of
the cathodic peak which is practically invisible with 100 mM buffer.
This effect stresses that the buffer is also adsorbed on the active
sites. The effect at higher pH is also very interesting. It is clearly
visible that with fluoride as the supporting electrolyte the peak
current density is about 5 times lower. This can be easily explained
by considering the real purpose of the buffer: to maintain a constant
pH near the gold surface [10]. This process is very slow when only
−
−
of Cl with F . It is expected that the interaction between a very
hard ion, such as fluoride, with a soft element, like gold, is hindered,
thus −Im(Z) should remain constant in response to increasing flu-
oride concentration. The cyclic voltammograms after successive
fluoride additions are reported in Fig. 8. As expected, the addition of
fluorides does not affect the features of the cyclic voltammogram,
−
meaning that the adsorption of F is negligible. This conclusion has
Fig. 7. Continuous EIS measurement at 5 kHz of Au pin electrode in 100 mM K2HPO4,
pH 11.5 with increasing chloride concentration. Re(Z) in red squares and −Im(Z) in
blue circles. (For interpretation of the references to colors in this figure legend, the
reader is referred to the web version of this article.)
Fig. 9. Continuous EIS measurement at 5 kHz of Au pin electrode in 100 mM K2HPO4,
at pH 11.5 with increasing fluoride concentration. Re(Z) in red squares and −Im(Z)
in blue circles. (For interpretation of the references to colors in this figure legend,
the reader is referred to the web version of this article.)

M. Pasta et al. / Electrochimica Acta 55 (2010) 5561–5568
5565
−
fluoride ions are present, being OH ions provided only through
transport. So the buffer has two contrasting effects: it adsorbs to
the surface, decreasing the active sites for the oxidation of glucose
and it also keeps constant the pH at the surface of the gold electrode.
This second effect is predominant at high pH values.
On the basis of the experimental data and considerations
reported in this section, a mechanism of the generation of the
oxidative peak (*) is proposed in the next section.
3.1. Model of glucose oxidation
Here the electrochemical oxidation of glucose to gluconate is
analyzed in mechanistic detail. The aim is to understand the origin
of the oxidative peak in the reductive sweep of the cyclic voltamme-
try, a peculiar feature already studied for its implication in glucose
sensing [25]. The mechanism proposed is based on previous stud-
ies on platinum [10] and gold [24] electrodes. Fig. 11 summarizes
the mechanism and the main species involved. At first, the glu-
cose molecule is electrochemically adsorbed at the surface of the
electrode by dehydrogenation (reaction (1)). The dehydrogenated
molecule can be transformed to gluconate by direct oxidation (reac-
Fig. 10. Cyclic voltammetry at
a gold pin electrode of 100 mM NaF, 1 mM
(
K2HPO4–KH2PO4) at different pH: 7.5 (green); 9.5 (red); 11.5 (blue) in the presence
−1
of glucose (10 mM). Scan rate 20 mV s
.
ꢀ
Fig. 11. Proposed mechanism at gold electrodes. G = d-glucose (␤-d-glucopyranose formed predominantly in water), G = dehydrogenated glucose (intermediate generated
ꢀ
ꢀ
ꢀꢀꢀ
by anomeric carbon dehydrogenation), G = d-gluconate, G = ı-gluconolactone.

5
566
M. Pasta et al. / Electrochimica Acta 55 (2010) 5561–5568
Fig. 12. (a) Potential distribution at the solid phase/electrolyte interface. (b) Current flow through the gold/electrolyte interface.
Table 1
Summary of fixed simulation parameters.
k1,k,o
10
k1,a,o
k2,k,o + k3,k,o
8 × 10−1
k2,o,o/k2,k,o
k3,o,o/k3,k,o
−
−7
1
0
10
1
1
NI/mol cm 3
−
ꢀ/cm
ꢀ2/cm
kad,1,k,o
kad,2,k,o
−
7
8 × 10−3
−1
4 × 10−3
2 × 10−3
1
0
10
tion (2)), which involves the production of a hydroxide ion and the
+
elimination of an H . An alternative path is the oxidization of the
dehydrogenated glucose to ␦-gluconolactone (reaction (3)). The ␦-
gluconolactone is then transformed to gluconate after reacting with
a hydroxide ion.
Fig. 13. Simulation of the electrochemical oxidation of glucose. No adsorption of
other species considered. Parameters are in Table 1. Green r = 1, red r = 0, blue r = ∞.
(For interpretation of the references to colors in this figure legend, the reader is
referred to the web version of this article.)
k
1
,R
ꢀ
+ H O + e ꢁ G + OH−
−
G
G
G
(1)
(2)
(3)
(
I)
2
(␧)
(␤)
(␧)
(␧)
k
1
,O
as:
k
2
,R
m
ꢀ
ꢀ−
+ H + e ꢁ G + OH−
+
−
ꢀ
r1 = k1,RNIꢁ − k1,Oa a (1 − ꢁ − ꢁ − ꢁꢀꢀꢀ)(aAu)
r2 = k2,RNIꢁ a − k2,ONIꢁ a␧
ꢀ
␧
␧
ꢀ
ꢀꢀ
(
I)
(␧)
(␤)
(I)
(␧)
G OH
k
2
,O
ꢀ
ꢀ
␧
ꢀ
(4)
H
OH
ꢀ
ꢀꢀ
␧
ꢀ
r3 = k3,RNIꢁ
a
− k3,ONIꢁ
k
H
3
,R
ꢀ
ꢀꢀ
+
(␧)
−
ꢀ
+ H + e ꢁ G
(
I)
(␤)
(I)
where NI is the number of sites available in the inner Helmholtz
k
3
,O
ꢀ
ꢀꢀ
ꢀꢀꢀ
layer, ꢁ , ꢁ , ꢁ are the fraction of sites occupied by the dehydro-
␧
genated glucose, gluconate and ␦-gluconolactone, respectively, a
G
Here the subscript I indicates the inner Helmholtz layer of the
␧
OH
is the activity of glucose molecules in the electrolyte, a the activ-
interface, ␧ the solution and ␤ the electrode. The electrochemi-
cal reactions are written following the electrochemical convention
of reduction as the direct reaction and oxidation as the inverse
ity of hydroxide ions in the electrolyte, a␧ the activity of protons
H
in the electrolyte, and aAu the activity of gold at the surface of the
gold electrode. The exponent m of aAu is dependent on the number
of gold atoms necessary to coordinate the oxidation of the glucose.
Empirically, it should be greater than 1. The kinetic constants of
the electrochemical reactions (1)–(3) are dependent on the electric
potential distribution at the interface:
ꢀ
ꢀꢀ
ꢀꢀꢀ
one. The molecules represented by the symbols G, G , G , and G
are reported in Fig. 11, and are glucose, dehydrogenated glucose,
gluconate, and ␦-gluconolactone, respectively. Our electrochemical
experiments confirm that the oxidation of glucose is blocked at the
⎡
⎣
⎡
⎣
⎡
⎣
ꢀ
ꢀ
ꢀ
ꢁ⎤
⎡ ꢀ
⎡ ꢀ
ꢁ⎤
F
F
F
˚0␤ − ˚H␧
2RT
F
˚0␤ − ˚H␧
2RT
k1,R
k1,O
⎦
⎣
⎦
=
=
=
exp
exp
exp
−
−
−
,
= exp
= exp
= exp
k1,R,0
k1,O,0
ꢁ⎤
⎦
ꢁ⎤
⎦
˚0␤ + ˚I − 2˚H␧
F
˚0␤ + ˚I − 2˚H␧
k2,R
k2,O
⎣
⎣
(5)
,
k2,R,0
2RT
ꢁ⎤
k2,O,0
2RT
ꢁ⎤
⎡ ꢀ
˚0␤ − ˚H␧
F
˚0␤ − ˚H␧
k3,R
k3,O
⎦
⎦
,
k3,R,0
2RT
k3,O,0
2RT
where F is the Faraday constant, R the universal gas constant, T
surface of the gold hydroxide, as previously pointed out by Xiang
et al. [28]. We described the reaction rates for the ith reaction, ri,
␤
the absolute temperature, ꢂ0 the electric potential at the surface

M. Pasta et al. / Electrochimica Acta 55 (2010) 5561–5568
5567
Fig. 14. Schematic of the formation of the oxidative peak in the cathodic scan. (a) As long as the potential of the electrode is below 1.05 V vs. RHE glucose can adsorb to the gold
surface, first step of the oxidation reaction. (b) At potential higher than 1.05 V vs. RHE gold surface is oxidized to gold hydroxide, inactive toward glucose electro-oxidation
(
c) during the cathodic scan gold is reduced at potentials around 0.6 V vs. RHE, glucose can adsorb again and get oxidized, generating the oxidative peak in the cathodic scan.
The different in potential for gold hydroxide formation and reduction is due to overpotential.
of the gold electrode, ꢂ␧ the electric potential of the solution at
except glucose is neglected (like in the fluorides). The simulations
H
the Helmholtz plane (dH in Fig. 12), and ˚I the electric potential in
the inner Helmholtz layer. The electric field in the inner Helmholtz
layer is not constant, due to the presence of the gluconate, which is
a charged species (see Fig. 12). In Fig. 12 the current flow through
the electrode is reported. The correlation between the fraction of
the occupied sites of the different compounds and the reaction rates
is given by:
should be compared with Fig. 10. The effect of the parameter r on
reaction (1) is negligible, while it is very important in the shape of
the cyclic voltammetry curve above 0.3 V vs. RHE. When r = ∞, most
of the dehydrogenated glucose is oxidized to ␦-gluconolactone near
0.6 V vs. RHE. The desorption of ␦-gluconolactone does not involve
current flow; a limiting current is reached. After the formation of
gold hydroxide (at 1.05 V vs. RHE) the oxidation of glucose stops,
and the surface desorbs all the adsorbed species. Decreasing the
potential, the reduction of gold hydroxide generates new gold sur-
face, active to the oxidation of glucose (Fig. 14). The peak (*) is
observed in the simulations. Its intensity is much higher than in
experimental data, due to the fact that diffusion has not been con-
sidered. Under these conditions, the limiting reactant is the number
of active sites at the gold surface, which is immediately saturated.
By decreasing the value of r, the intensity of the peak (*) decreases.
This is an effect of the gluconate. The gluconate is a negatively
charged species and therefore is attracted to the surface of the
electrode at high voltages. When the potential of the electrode is
increased, the gluconate desorbs much more slowly than the ␦-
gluconolactone, therefore less active sites are available when the
glucose oxidation starts again.
ꢀ
dꢁ
dt
dꢁ
ꢀNI
ꢀNI
ꢀNI
= −ꢀr + ꢀr + ꢀr
1
2
3
ꢀ
ꢀ
= −ꢀr − ꢀr
(6)
2
ad,1
dt
ꢀ
ꢀꢀ
dꢁ
dt
= −ꢀr − ꢀr
3
ad,2
where ꢀ is the distance between the surface of the electrode and
the Helmholtz plane, and rad,1 and rad,2 are the reaction rates of the
desorption of the gluconate and the ␦-gluconolactone, respectively.
The desorption reaction rates can be expressed by:
rad,1 = kad,1,RNIꢁ − kad,1,OaGꢀꢀ (1 − ꢁ − ꢁ − ꢁꢀꢀꢀ)(aAu
ꢀ
ꢀ
␧
ꢀ
ꢀꢀ
m
)
(7)
rad,2 = kad,2,RNIꢁ − kad,2,OaGꢀꢀꢀ (1 − ꢁ − ꢁ − ꢁꢀꢀꢀ)(aAu
ꢀ
ꢀꢀ
␧
ꢀ
ꢀꢀ
m
)
Fig. 15 shows the simulated results of the effect of adsorption
on the peak (*). The figure reports two curves, simulating the pres-
with the kinetic parameters equal to:
−
−
ꢂ
ꢃ
ꢂ
ꢃ
ence of F (no adsorption) and Cl (chemisorption) respectively.
F(˚ − ˚ ␧)
F(˚ − ˚ ␧)
kad,1,R
I
H
kad,1,O
I
H
−
As expected from the experimental data, the presence of Cl par-
=
exp
−
,
= exp
kad,1,R,0
2RT
kad,1,O,0
2RT
tially blocks the active sites, thus decreasing the intensity of the
−
k
k
peak (*). We want to stress that in the case of Cl adsorption, the
ad,2,R
ad,2,O
=
1,
= 1
kad,2,R,0
kad,2,O,0
reaction rate for each step in the oxidation of glucose decreases, as
observed in the experimental section (see Fig. 6). The simulations
support the mechanism we have proposed for the generation of the
peak (*) and the influence of the other species in the electrolyte on
its intensity.
(8)
We want to stress that the reaction of desorption of glu-
conate involves charged species, therefore passage of reductive
current, as reported in Fig. 12. On the other hand, desorption of
␦-gluconolactone does not involve passage of charged species.
To prove that the proposed mechanism can justify the presence
of the oxidative peak (*), simulations of the oxidation of glucose
were made using a program compiled with MATLAB 7.0. From an
intuitive point of view, when the activity of gold at the surface
decreases (due to the formation of gold hydroxide), the reaction
rates r , r , and r slow down, while reaction rad,1 and rad,2 become
1
2
3
more positive. The result is a decrease of the adsorbed species.
When the oxide is reduced, the naked gold surface is ready to oxi-
dize the glucose. We will demonstrate that the ratio of the reaction
rates of the two alternative mechanisms of formation of the glu-
conate (reactions (2) and (3)) is relevant in observing the oxidative
peak (*) (see Fig. 1). The simulations do not include transport phe-
nomena, nor the activity coefficients of the species, therefore it
is not expected to completely reproduce the experimental cyclic
voltammetry curve. Nevertheless, the main features (number of
peaks and the peak (*)) should be reproduced.
The fixed parameters of the simulations are given in Table 1. The
parameter r is the ratio between kd,2 and kd,3. In Fig. 13 the simula-
tions for different values of r are reported. Absorption of other ions
Fig. 15. Simulation of the electrochemical oxidation of glucose, with r = 0. Green no
adsorption, red adsorption (100 mM). (For interpretation of the references to colors
in this figure legend, the reader is referred to the web version of this article.)

5
568
M. Pasta et al. / Electrochimica Acta 55 (2010) 5561–5568
4
. Conclusions
[6] S. Ben Aoun, I. Taniguchi, Chem. Lett. 37 (2008) 936.
[
7] L.H.E. Yei, B. Beden, C. Lamy, J. Electroanal. Chem. Interfacial Electrochem. 246
1988) 349.
8] N.N. Nikolaeva, O.A. Khazova, Y.B. Vasil’ev, Elektrokhimiya 19 (1983) 1476.
(
This paper has proposed a mechanism to explain the oxidation
[
peak in the cathodic scan at gold electrodes in which the key step
is the competitive adsorption at the active sites of the ionic species
present in the solution (phosphate buffer, chlorides and OH ) and
[9] N.N. Nikolaeva, O.A. Khazova, Y.B. Vasil’ev, Elektrokhimiya 16 (1980) 1227.
10] B. Beden, F. Largeaud, K.B. Kokoh, C. Lamy, Electrochim. Acta 41 (1996) 701.
11] J.R. Rao, G.J. Richter, F. Von Sturm, E. Weidlich, Bioelectrochem. Bioenerg. 3
[
[
−
(
1976) 139.
the substrate (glucose). Simulations of the proposed mechanism
have supported the plausibility of the mechanism. The study rep-
resents the basis for upcoming work on glucose sensing at gold
electrodes.
[12] U. Gebhardt, G. Luft, G.J. Richter, F. Von Sturm, Bioelectrochem. Bioenerg. 5
(1978) 607.
[
13] M. Tominaga, T. Shimazoe, M. Nagashima, H. Kusuda, A. Kubo, Y. Kuwahara, I.
Taniguchi, J. Electroanal. Chem. 590 (2006) 37.
[14] M. Tominaga, T. Shimazoe, M. Nagashima, I. Taniguchi, Electrochem. Commun.
(2005) 189.
7
[
[
[
15] Y.-G. Zhou, S. Yang, Q.-Y. Qian, X.-H. Xia, Electrochem. Commun. 11 (2009) 216.
16] Y. Bai, W. Yang, Y. Sun, C. Sun, Sens. Actuators B B134 (2008) 471.
17] J.-J. Yu, S. Lu, J.-W. Li, F.-Q. Zhao, B.-Z. Zeng, J. Solid State Electrochem. 11 (2007)
1211.
18] C. Jin, I. Taniguchi, Chem. Eng. Technol. 30 (2007) 1298.
19] B.K. Jena, C.R. Raj, Chem. Eur. J. 12 (2006) 2702.
20] M. Tominaga, T. Shimazoe, M. Nagashima, I. Taniguchi, Chem. Lett. 34 (2005)
202.
21] M.W. Hsiao, R.R. Adzic, E.G. Yeager, J. Electrochem. Soc. 143 (1996) 759.
22] K.B. Kokoh, J.M. Leger, B. Beden, H. Huser, C. Lamy, Electrochim. Acta 37 (1992)
Acknowledgements
Y.C. acknowledges support from the King Abdullah University of
Science and Technology (KAUST) Investigator Award (No. KUS-l1-
[
[
[
001-12). The authors wish to thank Heather Deshazer for her help
in preparing the manuscript.
[
[
References
1909.
[
23] R.R. Adzic, M.W. Hsiao, E.B. Yeager, J. Electroanal. Chem. Interfacial Elec-
trochem. 260 (1989) 475.
[
[
[
1] S. Kerzenmacher, F J., R. Zengerle, F. von Stetten, J. Power Sources 182 (2008) 1.
2] A. Heller, B. Feldman, Chem. Rev. (Washington, DC, US), 108 (2008) 2482.
3] J.M. Ekoé, The Epidemiology of Diabetes Mellitus, Wiley-Blackwell/Hoboken,
Chichester, UK/NJ, 2008.
[24] N.N. Nikolaeva, O.A. Khazova, Y.B. Vasil’ev, Elektrokhimiya 19 (1983) 1042.
[25] E.B. Makovos, C.C. Liu, Bioelectrochem. Bioenerg. 15 (1986) 157.
[26] W.R. LaCourse, D.C. Johnson, Anal. Chem. 65 (1993) 50.
[27] D.W. Kirk, F.R. Foulkes, W.F. Graydon, J. Electrochem. Soc. 127 (1980) 1069.
[28] C. Xiang, Q. Xie, S. Yao, Electroanalysis 15 (2003) 987.
[29] M.J. Nicol, Gold Bull. 13 (1980) 105.
[
[
4] M. Tominaga, M. Nagashima, K. Nishiyama, I. Taniguchi, Electrochem. Commun.
9
(2007) 1892.
5] I.T. Bae, E. Yeager, X. Xiang, C.C. Liu, J. Electroanal. Chem. Interfacial Elec-
trochem. 309 (1991) 131.
[30] N.P. Finkelstein, R.D. Hancock, Gold Bull. 7 (1974) 72.