Unexpected complexity in the electro-oxidation of iodide on gold in the ionic liquid 1-ethyl-3-methylimidazolium bis(trifluoromethanesulfonyl)imide

Article abstract of DOI:10.1021/ac402150y

The electro-oxidation of iodide on a gold electrode in the room temperature ionic liquid 1-ethyl-3-methylimidazolium bis(trifluoromethanesulfonyl)imide has been investigated using transient cyclic voltammetry, linear-sweep semi-integral voltammetry, an electrochemical quartz crystal microbalance technique, and coulometry/electrogravimetry. Two oxidation processes are observed, with an electron stoichiometry of 1:1, compared with the well-known 2:1 electron stoichiometry observed on other commonly used electrode materials, such as platinum, glassy carbon, and boron-doped diamond, under identical conditions. Detailed mechanistic information, obtained in situ using an electrochemical quartz crystal microbalance, reveals that this unusual observation can be attributed to the dissolution of the gold electrode in the presence of iodide. Coulometric/electrogravimetric analysis suggests that the oxidation state of the soluble gold species is +1 and that diiodoaurate, [AuI₂]^-, is the likely intermediate. A proportionally smaller amount of triiodide intermediate is also detected by means of UV-vis spectroscopy. On this basis, it is proposed that iodide oxidation on gold occurs via two parallel pathways: predominantly via a diiodoaurate intermediate 2I^- + Au [AuI₂]^- + e^- and [AuI ₂]^- I₂ + Au + e^- and to a lesser extent via a triiodide intermediate 3I^- I₃^- + 2e^- and I₃^- 3/2I₂ + e^-. This proposed mechanism was further supported by voltammetric investigations with an authentic sample of the anionic [AuI₂]^- complex.

Full text of DOI:10.1021/ac402150y

Article
pubs.acs.org/ac
Unexpected Complexity in the Electro-Oxidation of Iodide on Gold in
the Ionic Liquid 1‑Ethyl-3-methylimidazolium
bis(trifluoromethanesulfonyl)imide
†
,‡
,†
,‡
§
,†
Cameron L. Bentley, Alan M. Bond,* Anthony F. Hollenkamp,* Peter J. Mahon, and Jie Zhang*
†
School of Chemistry, Monash University, Clayton, Vic 3800, Australia
‡
CSIRO Energy Technology, Box 312, Clayton South, Vic 3169, Australia
§
Faculty of Life and Social Sciences, Swinburne University of Technology, Hawthorn, Vic 3122, Australia
*
S Supporting Information
ABSTRACT: The electro-oxidation of iodide on a gold electrode in
the room temperature ionic liquid 1-ethyl-3-methylimidazolium
bis(trifluoromethanesulfonyl)imide has been investigated using
transient cyclic voltammetry, linear-sweep semi-integral voltammetry,
an electrochemical quartz crystal microbalance technique, and
coulometry/electrogravimetry. Two oxidation processes are ob-
served, with an electron stoichiometry of 1:1, compared with the
well-known 2:1 electron stoichiometry observed on other commonly
used electrode materials, such as platinum, glassy carbon, and boron-
doped diamond, under identical conditions. Detailed mechanistic
information, obtained in situ using an electrochemical quartz crystal
microbalance, reveals that this unusual observation can be attributed to the dissolution of the gold electrode in the presence of
iodide. Coulometric/electrogravimetric analysis suggests that the oxidation state of the soluble gold species is +1 and that
−
diiodoaurate, [AuI ] , is the likely intermediate. A proportionally smaller amount of triiodide intermediate is also detected by
2
means of UV−vis spectroscopy. On this basis, it is proposed that iodide oxidation on gold occurs via two parallel pathways:
−
−
−
−
−
predominantly via a diiodoaurate intermediate 2I + Au ⇌ [AuI ] + e and [AuI ] ⇌ I + Au + e and to a lesser extent via a
2
2
2
−
−
−
−
−
triiodide intermediate 3I ⇌ I + 2e and I ⇌ 3/2I + e . This proposed mechanism was further supported by voltammetric
investigations with an authentic sample of the anionic [AuI ] complex.
3
3
2
−
2
ince their introduction over two decades ago, air/water
stable nonhaloaluminate room temperature ionic liquids
which the oxidized and reduced forms of a suitable redox
couple are present so as to establish a redox shuttle system that
regenerates the reduced form of the dye. According to the
original work, solutions of tetra-alkyl ammonium salts of
S
(
RTILs) have been employed as replacements for volatile
1
organic solvents in a range of applications. RTILs are intrinsic
2
−
−
ionic conductors, and many possess favorable properties such
iodide/triiodide (I /I ) in acetonitrile meet the dual require-
3
as negligible vapor pressure and high chemical, thermal, and
electrochemical stability. Indeed, the distinctive physicochem-
ments for an appropriate electrode potential combined with6
sufficient transport properties of the redox-active species.
However, these electrolyte solutions impose severe restrictions
on device performance due to poor stability (high solvent
volatility and corrosivity of trace iodine) and limited safety
(solvent flammability) of DSSCs under light soaking
conditions. These issues can be averted by employing an
3
ical properties possessed by a particular RTIL are determined
by its constituent cation/anion and therefore can be tailored to
4
suit specific applications. The versatile nature of RTILs has
allowed them to be successfully employed as replacements for
“
conventional electrolytes” in a range of electrochemical
5
,6
7
5,6,11
devices, including dye-sensitized solar cells, lithium batteries,
fuel cells, and supercapacitors.
The dye-sensitized solar cell (DSSC), as reported by
electrolyte solution that is based on an appropriate RTIL.
Another issue that is embedded in DSSC technology is the
8
9
high cost of platinum, which is a significant driving force for the
6
10
development of alternative cathode materials. While obviously
̈
O’Regan and Gratzel, has been the subject of intensive
sensible, this endeavor runs counter to the well-known kinetic
research over the past two decades. Lower manufacturing costs
and great versatility in design (size, shape, flexibility, etc.) mean
that DSSCs are possible replacements for traditional (p-/n-)
silicon photovoltaics in a range of applications. The DSSC has
−
−
12
facility of the I /I redox process at noble-metal electrode
materials.
3
13−17
Platinum, in particular, has been a widely
three fundamental components: (1) a TiO semiconductor
Received: July 15, 2013
2
photoanode with an adhered photoactive dye; (2) a platinized
counter electrode (cathode); (3) an electrolyte solution in
Accepted: October 21, 2013
Published: October 21, 2013
©
2013 American Chemical Society
11319
dx.doi.org/10.1021/ac402150y | Anal. Chem. 2013, 85, 11319−11325

Analytical Chemistry
studied electrode material for the I /I /I redox processes in a
Article
−
−
material, it is far from “noble” in its behavior and actually
becomes intimately involved in the course of iodide electro-
oxidation. This in turn presents new insights into the operation
3
2
18
19
range of solvents, including water, acetonitrile, and ionic
1
7,20−24
liquids.
It is well-known that iodide is oxidized to
−
−
molecular iodine in an overall one electron per iodide ion
process:
of the I /I3 electrode in DSSCs.
−
−
EXPERIMENTAL SECTION
Reagents. The compound 1-ethyl-3-methylimidazolium
bis(trifluoromethanesulfonyl)imide ([C mim][NTf ], Io-li-tec
■
2
I ⇌ I + 2e
(1)
2
It has also been well-established that iodide and iodine combine
homogeneously to form the polyhalogen complex anion,
triiodide:
2
2
−
2
GmbH, 99.5%) was dried under high vacuum (≤10 mbar) at
4
5 °C for 48 h and stored under lithium. The residual water
was less than 50 ppm, as determined by Karl Fischer titration
Metrohm 831 KF Coulometer). Additionally, 1-ethyl-3-
−
−
I + I ⇌ I
(2)
2
3
(
The stability constant of this reaction is highly solvent
methylimidazolium iodide ([C mim]I, Io-li-tec GmbH,
2
3
−1
25
7
−1
dependent, ranging from ≈10 M in water to ≈10 M
>98%) was recrystallized twice from a 2:1 mixture of ethyl
acetate (Merck, EMSURE) and isopropanol (Merck, EM-
SURE) and then dried under high vacuum for 48 h. Care was
2
6,27
in acetonitrile,
and has a strong bearing on the observed
2
6
iodide oxidation mechanism. On platinum, the overall
oxidation process shown in eq 1 may occur in one step, giving
rise to a single voltammetric process (as observed in aqueous
taken during handling and storage of [C mim]I to avoid
2
exposure to light. Tetra-n-butylammonium diiodoaurate
2
6,28−30
−
media
) or in two steps (via an I3 intermediate), giving
([NBu ][AuI ]) was prepared by suspending gold foil (0.53
4
2
rise to two voltammetric processes (as observed in acetoni-
g, Sigma-Aldrich, ≥99.9%), iodine (0.39 g, Sigma-Aldrich,
2
6,28,31
17,20−24
trile
and ionic liquids
):
99.8%), and tetra-n-butylammonium iodide (1.08 g, [NBu ]I,
4
3
−
−
−
Sigma-Aldrich, ≥99%) in absolute ethanol (8 cm , Merck,
3
I ⇌ I + 2e
(3)
3
−
−
EMSURE) and refluxing for 72 h (2Au + I + I
⇌
3
−
35
2
[AuI ] ). The crude product, [NBu ][AuI ] (confirmed by
3
2
2
4
2
I ⇌ I + e−
−
3
2
electrospray ionization mass spectrometry) was recrystallized
(4)
36
twice from dry ethyl acetate, washed with diethyl ether
Analogous voltammetric behavior is observed on glassy carbon
(Merck, EMSURE), and dried under vacuum, for a final yield of
77%. To avoid excessive moisture pickup, this procedure was
and boron-doped diamond electrodes, albeit with significant
17
overpotentials when compared with the behavior at platinum.
carried out under an atmosphere of dry N using standard
2
In contrast, iodide electro-oxidation has not been studied
extensively on gold electrodes. Only a few studies have been
reported so far, and all were undertaken in conventional
Schlenk techniques. Tetra-n-butylammonium hexafluoro-
phosphate ([NBu ][PF ], Sigma-Aldrich) was dried under
4
6
vacuum at 80 °C for 24 h. Ferrocene (Fluka, >98%),
acetonitrile (Sigma-Aldrich, 99.8%), and acetone (Merck,
EMSURE) were used as supplied by the manufacturer;
[C mim][NTf ], [C mim]I, [NBu ]I, [NBu ][AuI ], [NBu ]-
32
organic/aqueous electrolyte media. Loo et al. investigated the
surface adsorbed complexes formed on a gold electrode during
oxidative polarization in aqueous halide media in situ using
surface-enhanced Raman spectroscopy; Au(I) and Au(III)
2
2
2
4
4
2
4
[PF ], and I were stored and handled under a dry argon
6
2
−
−
species of the type [AuX ] and [AuX ] , respectively, were
atmosphere in a glovebox.
2
4
detected at the electrode surface at increasingly positive
Electrochemical Systems and Procedures. Voltammet-
ric experiments were undertaken with a Gamry Reference 600
Potentiostat/Galvanostat/ZRA (Warminster, PA). The electro-
chemical quartz crystal microbalance experiments were carried
out using an impedance method (passive EQCM) on an
Agilent E5061A ENA Series Network Analyzer (Santa Clara,
CA) coupled to an Autolab PGSTAT302N potentiostat
(Metrohm, Herisau, SUI). All experiments were carried out
under benchtop conditions at ambient temperature (21 ± 1
−
−
−
potentials for all of the investigated halides (X = Cl , Br ,
−
33
and I ). A study by Qi et al. on the voltammetric response of
iodide on a gold electrode in aqueous media revealed a complex
reaction pathway, with a myriad of reactions being possible due
to significant overlap between the standard potentials for iodide
34
and gold oxidation. Guo et al. characterized the soluble
species formed upon oxidative polarization of a gold electrode
in acetonitrile containing dissolved iodide in situ using
electrochemical/electrospray mass spectrometry; two diffu-
sion-controlled voltammetric processes were observed, and
the products associated with each (at increasingly positive
°C). All electrolyte solutions were degassed with N prior to
2
experimentation, and a blanket of N was maintained over the
2
electrolytes while undertaking experiments. All voltammetric
experiments were carried out using a standard 3-electrode
arrangement with a platinum mesh auxiliary electrode and an
−
−
potential) were [AuI ] and I , respectively. Even this limited
2
3
number of studies suggest that the electrochemical behavior of
−
−
−
−
37
the I /I3 couple at gold electrodes is likely to be appreciably
more complicated than that observed at platinum and carbon
electrodes.
I /I reference electrode (Pt | 400 mM [C mim]I + 100
3
2
mM I2 in [C mim][NTf ]). UV−vis spectroscopy was
2
2
performed with a Varian Cary 50 Bio UV−vis spectropho-
In this paper, we report the electrochemical behavior of
iodide on a gold electrode in the room temperature ionic liquid
tometer (Agilent, Santa Clara, CA).
The platinum macrodisk electrode (nominal diameter = 1.6
mm) was purchased from BASi (Bioanalytical Systems, West
Lafayette, IN), and the gold macrodisk electrode (nominal
diameter = 3.0 mm) was purchased from Metrohm (Herisau,
Switzerland). Macroelectrodes were activated by polishing with
successively smaller (1, 0.3, and 0.05 μm) aqueous alumina
slurries (Kemet, Kent, U.K.) on a clean polishing cloth
(Buehler, Lake Bluff, IL). Adherent alumina were removed by
1
-ethyl-3-methylimidazolium bis(trifluoromethanesulfonyl)-
imide. A variety of techniques, including transient cyclic
voltammetry, linear sweep semi-integral voltammetry, EQCM,
UV−vis spectroscopy, and coulometry/electrogravimetry are
employed to elucidate the redox mechanism and characterize
the electrogenerated intermediate species. In the course of this
investigation, we confirm that when gold is the electrode
1
1320
dx.doi.org/10.1021/ac402150y | Anal. Chem. 2013, 85, 11319−11325

Analytical Chemistry
Article
sonication in deionized water (Milli-Q), before rinsing with
acetone. The EQCM working electrode was a 13 mm diameter
AT-cut α-quartz crystal (Bright Star Crystals, Rowville, AUS),
which had gold disks (5.0 mm diameter) vapor deposited on
each side; the electrode resonated in air at 10 ± 0.05 MHz. The
change in resonant frequency (Δf ) in the quartz crystal
0
electrode is related to the change in mass (Δm) by the
3
8,39
Sauerbrey equation:
⎛
⎞
f0 (m)
Ax ρ
⎜
⎜
⎝
⎟
⎟
Δf = f (m + Δm) − f (m) = −
Δm
0
0
0
q q ⎠
=
−SΔm
(5)
where x is the thickness of the quartz crystal, ρ is the density
q
q
−3
of the quartz crystal (2.645 g cm at 20 °C), A is the active
electrode surface area, and S is the Sauerbrey constant. Before
each experiment, the EQCM working electrode was cleaned
using concentrated nitric acid and rinsed successively with
distilled water and acetone. The gold foil electrode (Sigma-
Aldrich, ≥99.9%) used in the coulometry experiments was
cleaned in concentrated nitric acid and rinsed successively in
distilled water and acetone prior to use.
Figure 1. Cyclic voltammograms obtained from 10 mM [C mim]I in
2
The active area of the relevant electrodes was calibrated
voltammetrically using the oxidation of a ferrocene solution of
known concentration (1.7 mM in acetonitrile containing 0.1 M
[
C mim][NTf ] at a 3.0 mm diameter Au macroelectrode with scan
2 2
−1
rates (from top to bottom) of 1000, 500, 250, 100, and 50 mV s .
[
1
NBu ][PF ]) and adopting a diffusion coefficient of 2.24 ×
4
6
−5
2
−1
40
0
cm s as published under these conditions. Semi-
integration of the I−t data was performed using an algorithm
4
1
from de Levie, which was adapted from Oldham and
4
2
Myland. A Microsoft Excel macro that includes this algorithm
43
is available free online.
RESULTS AND DISCUSSION
■
−
Cyclic Voltammetry of I on Au. Initial investigations on
gold were carried out with a 3.0 mm diameter electrode at an
iodide concentration of 10 mM; cyclic voltammograms
obtained at scan rates ranging from 50 to 1000 mV s− are
shown in Figure 1. For comparison, experiments were also
undertaken using a Pt electrode (shown in Figure 2). At all scan
rates, there are two sets of oxidation/reduction peaks observed
on both Au and Pt electrodes. On the gold electrode, the
oxidation peak (Ep,ox) and reduction peak (Ep,red) potentials for
the first process both shift in a positive direction with scan rate,
1
though separation between the peaks (ΔE ) is constant with
p
scan rate (ΔE = 189.7 ± 0.1 mV). This is a stark contrast to
p
−
−
the first redox process on platinum (I /I ), whereby ΔE
changes significantly with scan rate as a result of sluggish
heterogeneous electron-transfer kinetics.
3
p
1
7,22,24
Figure 2. Cyclic voltammogram obtained from 10 mM [C mim]I in
The voltam-
2
[
C mim][NTf ] at a 1.6 mm diameter Pt macroelectrode at a scan rate
2 2
metric characteristics of the second process obtained on both
−1
of 250 mV s .
gold and platinum are similar; Ep,ox, Ep,red, and ΔE are all
p
relatively constant over the range of scan rates investigated,
−
−
with values of 533 ± 3, 393 ± 3 (both vs I /I ), and 139.9 ±
́
convolution procedure initially described by Saveant and co-
44
3
0
.1 mV respectively (on Au). Overall, iodide redox chemistry
workers, which, when applied under conditions whereby mass
appears to be more complex on a gold compared to platinum or
carbon electrodes, with noticeable broadening of the reduction
peaks for both of the processes.
transport occurs by semi-infinite planar diffusion, is known as
45
semi-integration, as described by Oldham. Semi-integration of
I−E data gives rise to a function, M(t), which is related to the
concentration of electroactive species at the electrode surface.
Semi-integration has previously been shown to be applicable in
−
Linear Sweep Semi-Integral Voltammetry of I on Au.
In order to identify the relative number of electrons involved in
each step, transient (peak-shaped) potential sweep voltam-
metric data (Figures 1 and 2) were transformed into forms
46
high-viscosity RTIL electrolytes; accordingly, the forward
sweep of the I−E data on gold and platinum electrodes were
semi-integrated, following background subtraction to remove
capacitive current (shown in Figure 3).
12
which closely resemble steady-state voltammetric curves.
Transforming the data is achieved using the general
1
1321
dx.doi.org/10.1021/ac402150y | Anal. Chem. 2013, 85, 11319−11325

Analytical Chemistry
Article
oxidation/reduction, respectively, with 1:1 electron stoichiom-
etry, as described by the following equations:
−
−
−
2
I + Au ⇌ [AuI ] + e
(7)
(8)
2
−
−
[
AuI ] ⇌ I + Au + e
2
2
This mechanism has been proposed on the basis of the
electrochemical behavior of gold in conventional media
3
2,33,52
containing iodide.
In addition, Compton and co-
53
workers have previously suggested that the formation of
gold−halide complex intermediates is feasible when investigat-
ing the oxidation of chloride on gold in a range of ionic liquids.
Interestingly, theoretical calculations (geometric optimization
and potential energy surface mapping) by Kloo and co-
54
−
−
workers indicated that [AuI ] and I are closely related and
2
3
should be more-or-less exchangeable in polyiodides and gold(I)
compounds containing the former. It is also worth noting that
direct participation (dissolution) of the gold electrode during
charge transfer parallels the well-known mechanism for
mercury oxidation in the presence of halides in nonaqueous
55
media:
−
−
−
2
Hg + 6X ⇌ 2[HgX ] + 4e
3
(9)
Figure 3. Comparison of the background corrected semi-integrals
obtained from 10 mM [C mim]I in [C mim][NTf ] at a scan rate of
−
−
2
2
2
Hg + 2[HgX ] ⇌ 3HgX + 2e
−1
(10)
250 mV s on gold () and platinum (---) macrodisk electrodes.
3
2
−
where X is chloride, bromide, or iodide. When investigated
using DC polarography, the iodide ion gives two well-defined
oxidation waves on a mercury electrode, the first being assigned
to the reversible oxidation of mercury to soluble triiodomercu-
rate and the second to the reversible oxidation of mercury to
The M−E curves resemble near steady-state voltammograms
obtained with the respective electrode materials (Au: Figure
S1a; Pt: Figure S1b). Intriguingly, the magnitude of the
plateaus relative to one another (proportional to the number of
electrons involved in each process) is different on gold and
platinum electrodes. On platinum, as is well-known, the
limiting current magnitudes exhibit a 2:1 ratio, corresponding
56
soluble mercuric iodide.
EQCM Studies on a Au Thin-Film Electrode. If the
above-proposed mechanism is valid, a change in mass of the
gold electrode can be expected during the voltammetric
experiment and thus may be detectable by EQCM. The cyclic
voltammogram of the first oxidation process and corresponding
−
−
−
to the I /I3 and I /I oxidation processes described by eqs 3
3
2
and 4, respectively. On gold, the relative limiting current
magnitudes exhibit a 1:1 ratio, indicative of a different iodide
oxidation mechanism. This observation may be linked to the
fact that gold is less inert in the presence of iodide compared
with platinum- or carbon-based materials. For example, gold is
known to be liable to chemical and/or electrochemical
Δf −E curve obtained with an Au quartz−crystal electrode at
0
an iodide concentration of 10 mM is shown in Figure 4. Indeed,
a decrease in electrode mass was detected during the forward
(oxidative) potential sweep, and the lost mass was partially
recovered during the reverse (reductive) potential sweep. This
observation is consistent with the mechanism described by eq
7. The experiment was repeated at an iodide concentration of
300 mM in order to improve the signal-to-noise ratio; a cyclic
4
7,48
49
dissolution in aqueous
contain a mixture of halide (complexant) and halogen
oxidant). Further, gold undergoes reversible dissolution/
and nonaqueous solutions that
(
5
0
−
−
deposition in solutions containing mixtures of I and I3 as
per the following reaction:
voltammogram and corresponding Δf −E curve is shown in
Figure 5. The upper potential limit of the potential sweep was
limited to 0 V versus I /I , as it was found that the thin film of
gold was completely stripped upon further scanning to more
positive potentials, exposing the underlying chromium adhesion
0
−
−
−
−
−
3
2
Au + I + I ⇌ 2[AuI ]
(6)
3
2
Indeed, iodide has been proposed as a substitute for highly
4
7,48,51
57
toxic cyanide in the dissolution phase of gold refining.
layer (shown in Figure S2). The EQCM/voltammetric
It is well-known that one electron is involved in the overall
electro-oxidation of each iodide ion on platinum. The
characteristics of the first process scale with concentration
because again the oxidation/reduction processes correspond to
a decrease/increase in electrode mass, respectively, which is
consistent with a reaction of the type described by eq 7.
Coulometric Studies with a Au Foil Electrode. To
confirm that the oxidation state of Au in the dissolved Au−
iodide complex is +1, electrogravimetric analysis (bulk
magnitude of the area-normalized M plateaus on both Au
L
and Pt are comparable (see Figure 3), suggesting that the
number of electrons involved during electro-oxidation of each
iodide ion on gold is also one. Therefore, the difference
between the iodide oxidation mechanism on gold and platinum
must arise from a disparity in the intermediate formed after the
initial oxidation process; on platinum it is well-established that
this intermediate is triiodide, as shown in the reaction described
by eq 3. In accordance with these results, the intermediate
formed on the gold electrode may: (a) be a soluble gold−
iodide compound/complex and (b) liberate iodine/iodide upon
−
−
electrolysis) at 0.2 V versus I /I3 was carried out. During
the bulk electrolysis experiment, the initially colorless electro-
−
lyte changed to a brown color characteristic of I , the presence
3
of which was confirmed using UV−vis spectroscopy. This
observation suggests that two oxidation processes are occurring
simultaneously at the investigated hold potential: the oxidation
1
1322
dx.doi.org/10.1021/ac402150y | Anal. Chem. 2013, 85, 11319−11325

Analytical Chemistry
Article
Figure 5. (a) Cyclic voltammogram and (b) corresponding Δf −E
0
curve obtained from 300 mM [C mim]I in [C mim][NTf ] at a Au
2
2
2
2
−1
quartz crystal electrode (A = 0.28 cm ) at a scan rate of 5 mV s .
Figure 4. (a) Cyclic voltammogram and (b) corresponding Δf −E
0
curve obtained from 10 mM [C mim]I in [C mim][NTf ] at a Au
The amount of electrogenerated triiodide (and hence Q −
2
2
2
)
2
−1
I3
quartz crystal electrode (A = 0.28 cm ) at a scan rate of 1 mV s .
was calculated using UV−vis spectroscopy after preparing a
standard curve (shown in Figure S3). The presence of the
soluble gold−iodide intermediate had no observable effect on
the UV−vis spectrum of triiodide, implying that the dissolved
of gold to diiodoaurate, as shown in the reaction described by
eq 7, and the oxidation of iodide to triiodide, as shown in the
reaction described by eq 3. The formation of two distinct
products during oxidation is consistent with the broad
reduction peaks observed during cyclic voltammetry (see
Figure 1). The quantity of electricity passed through the cell
−
species is colorless; the [AuI ] anion reportedly possesses no
2
48
UV−vis spectrophotometric activity. Thus, QAu was calcu-
lated using eq 11, and the oxidation state of gold was
determined using Faraday’s laws of electrolysis:
Q M
Au
(
Q ) is therefore consumed by the oxidation reactions
T
z =
mF
(12)
described by eqs 3 and 7, thus:
where m is the difference in the mass of the gold electrode
before and after electrolysis, M is the atomic mass of gold, F is
Q = Q + Q I
−
T
Au
(11)
3
1
1323
dx.doi.org/10.1021/ac402150y | Anal. Chem. 2013, 85, 11319−11325

Analytical Chemistry
Article
Faraday’s constant, and z is the valence number of the oxidized
gold species. By substituting the appropriate values of known
parameters into eq 12, z was determined to be 1.01, in
agreement with the reactions described by eqs 7 and 8. QAu
accounted for roughly 88% of Q , indicating that the reaction
T
described by eq 7 is the predominant oxidation pathway, which
explains why 1:1 electron stoichiometry is observed voltam-
metrically (as shown in Figure 3).
−
Voltammetric Studies of [AuI2] on Au. To further
support the proposed mechanism, [NBu ][AuI ] was synthe-
4
2
sized, and voltammetric experiments were undertaken in
C mim][NTf ] containing 2 mM of [NBu ][AuI ] on a gold
[
2
2
4
2
−1
electrode. A cyclic voltammogram recorded at 25 mV s is
shown in Figure 6a. As was to be expected based on the
reactions described by eqs 7 and 8, starting from open circuit
−
potential, [AuI ] gives rise to two waves of equal magnitude
2
(
(
one oxidative and one reductive). The midpoint potentials
Emid) defined as
Ep,ox + Ep,red
Emid
=
(13)
2
−
are comparable for both processes to those obtained with I
−
−
under identical conditions (see Figure 6b). Emid (all vs I /I )
3
on the second cycle is equal to 10 and 462 mV for the
processes occurring at lower and higher potentials, respectively,
−
−
for [AuI ] , compared to 15 and 455 mV for I . Similar
2
variation in Emid was observed when changing the direction of
the initial sweep and is thought to arise from the nature of the
redox mechanism (i.e., adsorption and dissolution/deposition
processes at the electrode surface give rise to slight
irreproducibility). On the basis of these and previous results,
we conclude that iodide oxidation to molecular iodine on a gold
electrode occurs predominantly via a diiodoaurate intermediate,
as shown in the reactions described by eqs 7 and 8, and to a
lesser extent, via a triiodide intermediate, analogous to what is
observed on platinum, as shown in the reactions described by
eqs 3 and 4.
CONCLUSIONS
■
The electro-oxidation of iodide on a gold electrode in the room
temperature ionic liquid 1-ethyl-3-methylimidazolium bis-
(trifluoromethanesulfonyl)imide has been investigated in detail
using a variety of techniques. Two oxidation processes with an
electron stoichiometry of 1:1 are observed on a gold electrode,
compared to the well-known 2:1 electron stoichiometry
observed on platinum under identical conditions. Dissolution
of the gold electrode, detected in situ using EQCM, was found
to coincide with the first iodide oxidation process, and the
oxidation state of the soluble gold species was found to be +1
using coulometric/electrogravimetric analysis. Quantitative
treatment of the coulometric data suggests that iodide oxidation
to molecular iodine on gold occurs predominantly via a
diiodoaurate intermediate, and to a lesser extent, via a triiodide
intermediate (detected using UV−vis spectroscopy). Diiodoau-
rate was further confirmed as the intermediate by investigating
the voltammetric response of this species at a gold electrode.
Uncovering a pathway for oxidation that is clearly distinct from
the established one involving triiodide, and with electron
transfer kinetics that appear to be superior, offers the prospect
Figure 6. Cyclic voltammogram obtained from (a) 2 mM [NBu ]-
4
[
AuI ] and (b) 4 mM [C mim]I in [C mim][NTf ] at a 3.0 mm
2 2 2 2
−1
diameter Au macroelectrode at a scan rate of 25 mV s . First (---) and
second () cycles are shown.
equivalents, there may be other conductive substrates where
analogous reactions and substrate interactions allow for fast,
reversible operation of the iodide−triiodide couple. In support
of the feasibility of this approach, it is worth recalling that the
rather moderate potential of the latter invites consideration of a
large range of possible electrode materials and composites.
ASSOCIATED CONTENT
Supporting Information
■
*
S
Near steady-state voltammograms obtained from the electro-
oxidation of iodide on gold and platinum (Figure S1), an
EQCM experiment demonstrating the complete stripping of
the gold thin film electrode during iodide electro-oxidation
(Figure S2), and the standard curve/UV−visible spectrum used
−
−
for enhancing the performance of the I /I redox couple in
3
dye-sensitized solar cells. While this is unlikely to involve the
replacement of platinum-coated electrodes with gold-coated
1
1324
dx.doi.org/10.1021/ac402150y | Anal. Chem. 2013, 85, 11319−11325

Analytical Chemistry
Article
to calculate the amount of triiodide formed during bulk
electrolysis (Figure S3). This material is available free of charge
via the Internet at http://pubs.acs.org.
(32) Loo, B. H. J. Phys. Chem. 1982, 86, 433−437.
(
33) Qi, P. H.; Hiskey, J. B. Hydrometallurgy 1993, 32, 161−179.
34) Zhao, S. F.; Lu, J. X.; Bond, A. M.; Zhang, J. Chem.Eur. J.
(
2
(
012, 18, 5290−5301.
35) Baker, L. J.; Bott, R. C.; Bowmaker, G. A.; Healy, P. C.; Skelton,
AUTHOR INFORMATION
Corresponding Authors
E-mail: alan.bond@monash.edu (A.M.B.).
E-mail: jie.zhang@monash.edu (J.Z.).
E-mail: tony.hollenkamp@csiro.au (A.F.H.).
■
B. W.; Schwerdtfeger, P.; White, A. H. J. Chem. Soc., Dalton Trans.
1
995, 1341−1347.
*
*
*
(36) Ohtsuka, I.; Nakayama, H.; Ishii, K. J. Raman Spectrosc. 1989,
20, 489−492.
̈
(37) Dias, A. P.; Papageorgiou, N.; Kalyanasundaram, K.; Gratzel, M.
Inorg. Chem. 1996, 35, 1168−1178.
Notes
(
(
38) Sauerbrey, G. Zeitschrift Fur Physik 1959, 155, 206−222.
The authors declare no competing financial interest.
39) Snook, G. A.; Bond, A. M.; Fletcher, S. J. Electroanal. Chem.
2
002, 526, 1−9.
ACKNOWLEDGMENTS
■
(40) Rogers, E. I.; Silvester, D. S.; Poole, D. L.; Aldous, L.; Hardacre,
We thank Dr. Thomas Ruether for assistance with synthesis of
the [NBu ][AuI ] complex.
C.; Compton, R. G. J. Phys. Chem. C 2008, 112, 2729−2735.
(
41) de Levie, R. How to Use Excel® in Analytical Chemistry and in
4
2
General Scientific Data Analysis; Cambridge University Press: Cam-
bridge, U.K., 2001.
REFERENCES
■
(
42) Oldham, K. B.; Myland, J. C. Fundamentals of Electrochemical
Science; Academic Press: San Diego, CA, 1994.
43) de Levie, R. Excellaneous. http://www.bowdoin.edu/∼rdelevie/
excellaneous/ (accessed September 19, 2013).
(
(
(
1) Wilkes, J. S. Green Chem. 2002, 4, 73−80.
2) MacFarlane, D. R.; Seddon, K. R. Aust. J. Chem. 2007, 60, 3−5.
3) Armand, M.; Endres, F.; MacFarlane, D. R.; Ohno, H.; Scrosati,
(
B. Nat. Mater. 2009, 8, 621−629.
4) Dean, P. M.; Pringle, J. M.; MacFarlane, D. R. Phys. Chem. Chem.
Phys. 2010, 12, 9144−9153.
5) Papageorgiou, N.; Athanassov, Y.; Armand, M.; Bonhote, P.;
Pettersson, H.; Azam, A.; Gratzel, M. J. Electrochem. Soc. 1996, 143,
099−3108.
6) Hagfeldt, A.; Boschloo, G.; Sun, L. C.; Kloo, L.; Pettersson, H.
Chem. Rev. 2010, 110, 6595−6663.
7) Howlett, P. C.; MacFarlane, D. R.; Hollenkamp, A. F.
Electrochem. Solid State Lett. 2004, 7, A97−A101.
8) de Souza, R. F.; Padilha, J. C.; Goncalves, R. S.; Dupont, J.
Electrochem. Commun. 2003, 5, 728−731.
9) Balducci, A.; Dugas, R.; Taberna, P. L.; Simon, P.; Plee, D.;
(
(
(
44) Saveant, J. M. J. Electroanal. Chem. 1973, 44, 169−187.
45) Grenness, M.; Oldham, K. B. Anal. Chem. 1972, 44, 1121−1129.
46) Bentley, C. L.; Bond, A. M.; Hollenkamp, A. F.; Mahon, P. J.;
(
(
Zhang, J. Anal. Chem. 2012, 85, 2239−2245.
47) Issaev, N.; Osteryoung, J. G. J. Electrochem. Soc. 1998, 145, 974−
81.
48) Issaev, N.; Osteryoung, J. G. J. Electrochem. Soc. 1996, 143,
477−3482.
̈
(
9
(
3
3
(
(
(
49) Nakao, Y. J. Chem. Soc., Chem. Commun. 1992, 426−427.
50) Nakao, Y.; Sone, K. Chem. Commun. 1996, 897−898.
51) Hu, Z.; Ritzdorf, T. J. Electrochem. Soc. 2007, 154, D543−D549.
52) Guo, T.; Li, L. D.; Cammarata, V.; Illies, A. J. Phys. Chem. B
(
(
(
(
(
2
(
005, 109, 7821−7825.
53) Villagran, C.; Banks, C. E.; Hardacre, C.; Compton, R. G. Anal.
Chem. 2004, 76, 1998−2003.
54) Svensson, P. H.; Rosdahl, J.; Kloo, L. Chem.Eur. J. 1999, 5,
05−311.
55) Bond, A. M. Modern Polarographic Methods in Analytical
Chemistry; Marcel Dekker: New York, 1980.
56) Matsui, Y.; Kurosaki, Y.; Date, Y. Bull. Chem. Soc. Jpn. 1970, 43,
707.
57) Deakin, M. R.; Buttry, D. A. Anal. Chem. 1989, 61, 1147A−
154A.
Mastragostino, M.; Passerini, S. J. Power Sources 2007, 165, 922−927.
(
(
10) O’Regan, B.; Grat
11) Wang, P.; Zakeeruddin, S. M.; Moser, J. E.; Humphry-Baker, R.;
zel, M. J. Am. Chem. Soc. 2004, 126, 7164−7165.
12) Bard, A. J.; Faulkner, L. R. Electrochemical Methods:
Fundamentals and Applications, 2nd ed.; Wiley: New York, 2001.
13) Swain, G. M. In Handbook of Electrochemistry, 1st ed.; Zoski, C.
̈
zel, M. Nature 1991, 353, 737−740.
(
3
(
Grat
(
̈
(
(
1
(
1
G., Ed.; Elsevier: Amsterdam, 2007; pp 111−153.
(
(
14) Hubbard, A. T. J. Phys. Chem. 1975, 79, 808−815.
15) Mebrahtu, T.; Rodriguez, J. F.; Bravo, B. G.; Soriaga, M. P. J.
Electroanal. Chem. 1987, 219, 327−333.
16) Bravo, B. G.; Mebrahtu, T.; Rodriguez, J. F.; Soriaga, M. P. J.
Electroanal. Chem. 1987, 221, 281−287.
17) Bentley, C. L.; Bond, A. M.; Hollenkamp, A. F.; Mahon, P. J.;
Zhang, J. Electrochim. Acta 2013, 109, 554−561.
18) Dane, L. M.; Janssen, L. J. J.; Hoogland, J. G. Electrochim. Acta
968, 13, 507−513.
19) Popov, A. I.; Geske, D. H. J. Am. Chem. Soc. 1958, 80, 1340−
352.
(
(
(
1
(
1
(
(
(
20) Kawano, R.; Watanabe, M. Chem. Commun. 2003, 330−331.
21) Kawano, R.; Watanabe, M. Chem. Commun. 2005, 2107−2109.
22) Ejigu, A.; Lovelock, K. R. J.; Licence, P.; Walsh, D. A.
Electrochim. Acta 2011, 56, 10313−10320.
(
(
23) Zhang, Y.; Zheng, J. B. Electrochim. Acta 2007, 52, 4082−4086.
24) Silvester, D. S.; Aldous, L.; Hardacre, C.; Compton, R. G. J.
Phys. Chem. C 2008, 112, 6551−6557.
(25) Gebert, E. J. Am. Chem. Soc. 1954, 76, 2049−2054.
(26) Guidelli, R.; Piccardi, G. Electrochim. Acta 1967, 12, 1085−1095.
(27) Popov, A. I.; Rygg, R. H.; Skelly, N. E. J. Am. Chem. Soc. 1956,
7
(
8, 5740−5744.
28) Iwamoto, R. T. Anal. Chem. 1959, 31, 955.
(29) Zittel, H. E. J. Electroanal. Chem. 1966, 11, 85−93.
(30) Dryhurst, G.; Elving, P. J. J. Electroanal. Chem. 1966, 12, 416.
(31) Elving, P. J. Anal. Chem. 1967, 39, 606−615.
1
1325
dx.doi.org/10.1021/ac402150y | Anal. Chem. 2013, 85, 11319−11325