H2O2 generation by decamethylferrocene at a liquid|liquid interface

Article abstract of DOI:10.1002/anie.200801004

(Figure Presented) Hydrogen peroxide generation at a liquid|liquid interface occurs with a yield of 20% with respect to the concentration of reducing agent (decamethylferrocene). The liquid|liquid interface supplies electrons from the reducing agent and protons from the aqueous phase to drive the reduction of O₂ into H₂O₂, which is extracted into the aqueous phase during the course of reaction (see picture; DCE = 1,2-dichloroethane).

Full text of DOI:10.1002/anie.200801004

Angewandte
Chemie
DOI: 10.1002/anie.200801004
Hydrogen Peroxide Production
H2O2 Generation by Decamethylferrocene at a Liquid j Liquid
Interface**
Bin Su, Raheleh Partovi Nia, Fei Li, Mohamad Hojeij, Michel Prudent, ClØmence Corminboeuf,
Zdenek Samec, and Hubert H.Girault*
The reduction of molecular oxygen (O2) is a technologically
important research topic, particularly in the context of a
hydrogen economy. The O2 reduction reaction (ORR) can
proceed by a direct four-electron reduction to produce water
or a two-electron reduction to give hydrogen peroxide, with
the former pathway being highly desirable for fuel cell
applications. The development of fuel cells for the combined
production of electricity and hydrogen peroxide (H2O2) has
been proposed recently.[1] H2O2 is an industrially important
product that is used on a scale of about three million metric
tons per year worldwide. Its production is currently based
almost exclusively on the anthraquinone hydrogenation and
oxidation process.[2] Many alternative routes have also been
of the interface can be controlled by the distribution of
different salts. The main advantage of the present biphasic
system is that the ORR can be stopped at the formation of
H2O2, which can be directly extracted into the aqueous phase
during the reaction. H2O2 is a strong oxidant that readily
oxidizes ferrocene derivatives, which usually leads to the
absence of H2O2 in the final products of the homogeneous
oxidation of ferrocene derivatives by O2.[8,10]
When an interface is formed between an aqueous electro-
lyte containing hydrophilic ions and an organic solution
containing lipophilic ions, this interface becomes polarizable.
This leads to a polarization potential window, the width of
which is defined by the transfer reactions of the electrolyte
ions across the interface. The potential dependence for ion
transfer across the interface follows a Nernst equation:
developed, one of which is the electrochemical cathodic
[2–4]
reduction of O2
in the presence of molecular electro-
catalysts such as metalloporphyrins[5] and anthraquinones.[6]
Herein we present an electrochemical method for pro-
ducing H2O2 at a soft molecular interface rather than at a solid
electrode. This approach relies on controlling the interfacial
Galvani potential difference between two immiscible phases,
in other words the polarization at the liquid j liquid interface,
to allow the reduction of O2 to H2O2. We chose the ORR by
ferrocene derivatives, a reaction that has been known for
many years, to illustrate this principle.[7,8] Specifically, we
studied the reduction of O2 by decamethylferrocene (DMFc)
in 1,2-dichloroethane (DCE)[9] in contact with an aqueous
solution of sulfuric acid. As described below, the polarization
ꢀ
ꢁ
R T
zi F
aiO
aiW
DWO ꢀ ¼ DWO ꢀOi
þ
ln
ð1Þ
where DWO ꢀOi represents the standard transfer potential, in
other words the Gibbs energy of transfer expressed in a
voltage scale. For example, using Li2SO4 and bis(triphenyl-
phosphoranylidene)ammonium tetrakis(pentafluorophenyl)-
borate (BTPPATPFB) as the hydrophilic and lipophilic
electrolytes, respectively, in water and DCE results in a
potential window from À0.5 to 0.4 V, as shown by the dotted
line in Figure 1. This window is determined by the transfer of
Li+ and SO4 ions from water to DCE at positive and
2À
negative potentials respectively, since BTPPA+ and TPFBÀ
ions are too lipophilic to transfer first. If an ion having a
medium lipophilicity, for example decamethylferrocenium
(DMFc+) in DCE, is present it will transfer within the above
potential window, and this transfer can be monitored by cyclic
voltammetry, as shown by the full line in Figure 1.
[*] Dr. B. Su, R. P. Nia, Dr. F. Li, M. Hojeij, M. Prudent,
Prof. H. H. Girault
Laboratoire d’Electrochimie Physique et Analytique
Ecole Polytechnique FØdØrale de Lausanne (EPFL)
Station 6, 1015 Lausanne (Switzerland)
Fax: (+41)21-693-3667
E-mail: hubert.girault@epfl.ch
Homepage: http://lepa.epfl.ch
Prof. C. Corminboeuf
Laboratory for Computational Molecular Design
Ecole Polytechnique FØdØrale de Lausanne (EPFL)
BCH, 1015 Lausanne (Switzerland)
Prof. Z. Samec
Department of Electrochemistry
J. Heyrovsky Institute of Physical Chemistry
Academy of Sciences of the Czech Republic
Dolejskova 3, 18223 Prague 8 (Czech Republic)
[**] This work was supported by the EPFL and European COST Action
(D36/007/06). We thank Professors X. Hu and J. Waser for fruitful
discussions. The technical assistance of V. Devaud is also
acknowledged.
Figure 1. Polarization of a waterjDCE interface by various common
ions. The dotted and full lines show the potential window and transfer
of DMFc+ due to polarization by external voltages, respectively.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or fromthe author.
Angew. Chem. Int. Ed. 2008, 47, 4675 –4678
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4675

Communications
Alternatively, the interface polarization can be controlled
disappeared and has been replaced by a very large absorbance
in the UV range. Formation of the DMFc+ ion was also
confirmed by the cyclic voltammetric response of a platinum
microdisc electrode (diameter: 25 mm) in the organic phase, as
illustrated in Figure 2c. After 4 h of reaction, a steady-state
current wave, which consists of a larger cathodic steady-state
current (ISC) and a smaller anodic steady-state current (ISA), is
observed at the same potential as for DMFc in DCE. As
DMFc and DMFc+ ion have about the same diffusion
coefficient, the percentage of DMFc oxidized can be calcu-
lated from the ratio ISC/ISS to be 74%, that is, a resulting
concentration of 3.7 mm. Furthermore, the sum of the
magnitudes of ISC and ISA is very close to that of freshly
prepared 5 mm DMFc in DCE (ISS), as can be seen from
Figure 2c. This voltammetric result provides two indications.
First, DMFc is oxidized to the DMFc+ cation, which stays in
the DCE phase. This coincides with the full line shown in
Figure 1 in that the transfer of DMFc+ ion from DCE to water
only occurs at negative Galvani potential differences. Second,
it indicates that both DMFc and the DMFc+ cation are stable
over the course of the two-phase reaction and that no
decomposition takes place. This was also confirmed by mass
spectrometric measurements as the mass spectra do not
display any peaks for iron ions or a free cyclopentadienyl ring
(see Figures S1.1–1.5 in the Supporting Information).
by the distribution of ions, for example by dissolving a
hydrophilic and a lipophilic salt featuring a common ion
(either cation or anion) in water and in the organic solution,
respectively. In this way, the Galvani potential difference
across the interface is given by the Nernst equation for the
distribution of this common ion. As illustrated in Figure 1, the
Galvani potential difference across the interface can be varied
by employing different common ions. This method allows a
chemical control of the Galvani potential difference without
supplying an external voltage.
Figure 2a illustrates an equal-volume (2:2 mL), two-
phase reaction under static conditions using TPFBÀ as the
common ion. The Galvani potential difference across the
waterj DCE interface is fixed at potentials greater than
0.59 V. Although the standard ion-transfer potential of the
TPFBÀ ion is unknown, it is known to be more positive than
the Li+ cation, whose standard transfer potential is 0.59 V.[11]
A fresh solution of 5 mm DMFc in DCE is yellow. After 4 h in
contact with 5 mm Li2SO4, the DCE phase turns green, thus
indicating the oxidation of DMFc to DMFc+, whereas the
aqueous phase remains colorless. The two phases were then
separated and the UV/Vis spectrum of the DCE solution
measured. As can be seen in Figure 2b, this solution shows an
absorption band due to the DMFc+ cation (lmax = 779 nm)
whereas the absorption peak for DMFc (lmax = 425 nm) has
The isolated aqueous solution was titrated with NaI to
detect the formation of H2O2. Thus, 29.98 mg (corresponding
to 0.1m, a large excess) of NaI was added to 2 mL of the
solution and, as shown in Figure 2d, the solution changed
from colorless to pale yellow (flask 2). Adding NaI to an
aqueous solution containing 5 mm LiTPFB and 5 mm H2SO4
in a controlled titration did not lead to any color change
within the present experimental time scale, thus confirming
the presence of H2O2 in the aqueous solution. H2O2 is a strong
oxidant that can oxidize IÀ to I3À, which can be visualized by
À
adding starch to give a red-brown color (flask 3). I3 can be
also detected by UV/Vis spectroscopy, as shown in Figure 2b
(sharp absorption band at lmax = 352 nm). Taking a emax value
À
of 2.76 ” 104 mÀ1 cmÀ1,[12] the concentration of I3 can be
calculated to be 0.070 Æ 0.003 mm, which corresponds to that
of H2O2 formed.
Thus far, it can be concluded that a two-electron reduction
of O2 to H2O2 by DMFc occurs in the two-phase reaction:
2 DMFc þ O2 þ 2 Hþ ! 2 DMFcþ þ H2O2
ð2Þ
Two reaction mechanisms can be considered: either DMFc
reduces O2 heterogeneously to produce H2O2 in water, or
DMFc reduces O2 homogeneously in DCE to H2O2, which is
then extracted into the aqueous phase. Fomin has proposed a
mechanism for the reduction of O2 by ferrocene derivatives in
the presence of an acid based on the form of the experimental
rate equation and on some computed thermodynamic data.
This mechanism involves protonation of the ferrocene
derivative followed by the reduction of O2 by two protonated
ferrocene derivatives to give H2O2.[10] A similar mechanism
can be proposed in this biphasic system: the first step consists
of the protonation of DMFc to form the DMFc-H+ cation.
Once DMFc-H+ is formed in DCE, it can react homoge-
Figure 2. a) Two-phase reaction controlled by TPFBÀ ions at the begin-
ning (left) and after 4 h (right). b) UV/Vis spectra of the DCE phase
(full black) and the water phase before (dotted red) and after (full red)
treatment with 0.1m NaI after 4 h of the two-phase reaction; the
spectrumof freshly prepared 5 m m DMFc (dotted black) is also
included for comparison. c) CVs obtained with a 25-mmPt micro-
electrode of freshly prepared 5 mm DMFc (full line) and the DCE
phase after 4 h of the two-phase reaction (dotted line). d) Flask 1:
water phase after 4 h of the two-phase reaction in (a); flask 2:
flask 1+0.1m NaI; flask 3: flask 2+starch; flask 4: 5 mm
LiTPFB+5 mm H2SO4 +0.1m NaI+starch.
4676 www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 4675 –4678

Angewandte
Chemie
neously with O2 in DCE to produce H2O2 and also hetero-
geneously with O2 in water to produce H2O2 directly.
However, these two reaction pathways are difficult to
discriminate in this biphasic system.
A control experiment using 5 mm H2SO4 and 5 mm
LiTPFB in water with 0.5 mm DMFc and 5 mm BTPPATPFB
in DCE was carried out to elucidate the stoichiometry of the
reaction. Figure 4, which displays the UV/Vis spectra of the
Figure 3 illustrates the influence of the Galvani potential
difference on this two-phase reaction when employing differ-
ent common ions. The Galvani potential difference at the
Figure 4. UV/Vis spectra of the DCE (a) and water phases (b) after an
equal-volume two-phase reaction. Conditions: 5 mm H2SO4 and 5 mm
LiTPFB in water and 0.5 mm DMFc and 5 mm BTPPATPFB in DCE; the
two-phase system was first stirred for 10 min then left for 30 min for
phase separation. The separated aqueous solution was treated with
excess NaI.
two separated phases, shows that about 95% of the DMFc is
oxidized and 0.090 Æ 0.004 mm H2O2 is detected in water.
Considering that the stoichiometric ratio given by Equa-
tion (2) is 2:1, the above ratio of 5:1 indicates that either the
partition coefficient for the extraction of H2O2 into water is
less than unity or that part of the H2O2 formed is reduced by
DMFc and/or decomposes. The former possibility was
examined by performing a titration measurement of the
H2O2 partition between water and DCE using NaI. This
experiment reveals that partition of the H2O2 initially present
in water into DCE is negligible (see Figure S3.1 in the
Supporting Information). Reduction of H2O2 by DMFc and/
or decomposition must therefore account for the low
quantities observed. H2O2 is a strong oxidant in acidic
solution and it can readily oxidize DMFc to DMFc+.[10]
H2O2 can also decompose in a reaction that is catalyzed by
most transition metals and their compounds.[3]
In summary, we have shown that O2 reduction by DMFc
occurs in the absence of any noble metal catalysts at a
polarized waterj DCE interface. The rate of reaction is
controlled by the Galvani potential difference across the
interface, which has been determined chemically using
various salts with a common ion. The resulting concentration
of H2O2 has been measured and shows a yield of 20% with
respect to the concentration of the reducing agent (DMFc).
This work illustrates how a biphasic system can be used to
simultaneously supply electrons from nonaqueous donors and
protons from an aqueous phase to drive interfacial reactions.
Figure 3. Two-phase reaction controlled by different common ions
(TPFBÀ, TMA+, TEA+, TBA+, and BTPPA+ fromleft to right; 5 m m in
both phases) after a) 0 min, b) 62 min, c) 17.5 h, and d) 102 h.
waterj DCE interface is set at > 0.59, 0.160, 0.019, À0.225, and
< À0.53 V by the ions TPFBÀ, TMA+, TEA+, TBA+, and
BTPPA+,[11] respectively, as shown in Figure 1. It is clear from
the color change in Figure 3, as well as the UV/Vis spectra
(Figure S2.1 in the Supporting Information), that the reaction
rate follows the order TPFBÀ > TMA+ > TEA+ > TBA+
> BTPPA+. The reaction is very fast when TPFBÀ is used as
the common ion. The color change of the DCE solution from
yellow to green starts immediately at the interfacial region
upon contact of the aqueous solution with the DCE solution,
thereby indicating that the O2 reduction by DMFc occurs at
the interface. When TMA+ and TEA+ are used as the
common ion, however, the green color of the DCE phase can
only be seen after more than 10 h, and in the case of TBA+ the
DCE phase remains yellow after 102 h and only a weak
absorption band at lmax = 779 nm can be seen in the UV/Vis
spectrum. In contrast, the DCE phase remains yellow after
102 h with BTPPA+ as the common ion but the aqueous phase
is slightly green. Absorption measurements show that the
appearance of an absorption band at lmax = 779 nm in the UV/
Vis spectrum of the aqueous phase is concomitant with a
decrease of the absorbance of DMFc in the DCE phase
(lmax = 425 nm); only a negligible absorbance is observed for
the DMFc+ ion in the DCE phase. This suggests that the
DMFc+ ion formed in the BTPPA+-controlled experiment is
transferred from DCE to water, which coincides with the
illustration shown in Figure 1. Figure 3 clearly shows the
influence of the Galvani potential difference on this biphasic
reaction, which reflects the potential dependence of either the
proton partition or the heterogeneous O2 reduction, or both.
Experimental Section
All chemicals were used as received. DMFc (97%) and H2SO4 (5m)
were purchased from Sigma–Aldrich. Li2SO4 (> 98.0%), NaI
(> 99.5%), tetramethylammonium chloride (TMACl, > 98.0%),
tetraethylammonium chloride (TEACl, ꢀ 98%), tetraethylammo-
nium perchlorate (TEAClO4, ꢀ 99%), tetrabutylammonium chloride
Angew. Chem. Int. Ed. 2008, 47, 4675 –4678
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 4677

Communications
(TBACl, ꢀ 97%), tetrabutylammonium perchlorate (TBAClO4,
ꢀ 99%), bis(triphenylphosphoranylidene)ammonium chloride
2001, 40, 1776 – 1779; c) I. Yamanaka, T. Murayama, Angew.
Chem. 2008, 120, 1926 – 1928; Angew.Chem.Int.Ed. 2008, 47,
1900 – 1902; d) K. Otsuka, I. Yamanaka, Electrochim.Acta 1990,
35, 319 – 322.
(BTPPACl, ꢀ 98%), potassium tetrakis(4-chlorophenyl)borate
(KTPBCl, ꢀ 98%), DCE (ꢀ 99.8%), and starch (from potatoes)
were obtained from Fluka. Lithium tetrakis(pentafluorophenyl)bo-
rate (LiTPFB) diethyl etherate was provided by Boulder Scientific
Company. TMATPBCl and BTPPATPFB were prepared by meta-
thesis of 1:1 mixtures of TMACl and KTPBCl and BTPPACl and
LiTPFB, respectively, in a methanol/water mixture (2/1, v/v), followed
by recrystallization from acetone. All the aqueous solutions were
prepared with ultrapure water (18.2 MWcm).
[2] W. Eul, A. Moeller, N. Steiner in Kirk–Othmer Encyclopedia of
Chemical Technology, Vol.14 , 5th ed. (Ed.: A. Seidel), Wiley,
Hoboken, 2005, pp. 35 – 78.
[3] J. M. Campos-Martin, G. Blanco-Brieva, J. L. G. Fierro, Angew.
Chem. 2006, 118, 7116 – 7139; Angew.Chem.Int.Ed. 2006, 45,
6962 – 6984.
[4] W. R. Thiel, Angew.Chem. 1999, 111, 3349 – 3351; Angew.Chem.
Int.Ed. 1999, 38, 3157 – 3158.
[5] J. H. Zagal, M. A. Paez, J. F. Silva in N4-Macrocyclic Metal
Complexes (Eds.: J. H. Zagal, F. Bedioui, J.-P. Dodelet),
Springer, New York, 2006, pp. 41 – 75.
The two-phase reactions were performed in a small flask under
static conditions unless specified otherwise. A DCE solution (2 mL)
containing 5 mm DMFc was added first, followed by the addition of
2 mL of an aqueous solution containing 5 mm H2SO4. The aqueous
and organic common ion salt solutions were added at the same
concentration (5 mm).
[6] a) K. Tammeveski, K. Kontturi, R. J. Nichols, R. J. Potter, D. J.
Schiffrin, J.Electroanal.Chem.
2001, 515, 101 – 112; b) A.
The microelectrode cyclic voltammetric measurements were
performed with a CHI-900 potentiostat (CH Instrument, USA) with
a platinum electrode (diameter: 25 mm), a platinum wire, and a silver
wire as the working, counter, and reference electrodes, respectively. A
scan rate of 0.02 VsÀ1 was employed. The potential scale was referred
Sarapuu, K. Vaik, D. J. Schiffrin, K. Tammeveski, J.Electroanal.
Chem. 2003, 541, 23 – 29; c) F. Mirkhalaf, K. Tammeveski, D. J.
Schiffrin, Phys.Chem.Chem.Phys.
2004, 6, 1321 – 1327; d) K.
Vaik, A. Sarapuu, K. Tammeveski, F. Mirkhalaf, D. J. Schiffrin, J.
Electroanal.Chem. 2004, 564, 159 – 166; e) K. Vaik, D. J.
Schiffrin, K. Tammeveski, Electrochem.Commun. 2004, 6, 1 –
5; f) J. R. T. Johnsson Wass, E. Ahlberg, I. Panas, D. J. Schiffrin,
Phys.Chem.Chem.Phys. 2006, 8, 4189 – 4199.
to the Fc+/Fc (EFc+ vs. SHE = 0.64 V[13]) couple.
/Fc
The UV/Vis spectra were measured with an Ocean Optical
CHEM2000 spectrophotometer with a quartz cuvette (path length:
10 mm).
[7] a) R. Prins, A. G. T. G. Kortbeek, J.Organomet.Chem. 1971, 33,
C33 – C34; b) T. E. Bitterwolf, A. C. Ling, J.Organomet.Chem.
1972, 40, C29 – C32.
Received: March 1, 2008
[8] J. Lubach, W. Drenth, Recl.Trav.Chim.Pays-Bas 1973, 92, 586 –
592.
Published online: May 16, 2008
[9] V. J. Cunnane, G. Geblewicz, D. J. Schiffrin, Electrochim.Acta
1995, 40, 3005 – 3014.
[10] V. M. Fomin, Russ.J.Gen.Chem. 2007, 77, 954 – 960.
[11] A. Sabela, V. Marecek, Z. Samec, R. Fuoco, Electrochim.Acta
1992, 37, 231 – 235.
[12] R. O. Rahn, M. I. Stefan, J. R. Bolton, E. Goren, P.-S. Shaw,
K. R. Lykke, Photochem.Photobiol. 2003, 78, 146 – 152.
[13] D. J. Fermin, R. Lahtinen, Surfactant Sci.Ser. 2001, 95, 179 – 227.
Keywords: decamethylferrocene · hydrogen peroxide ·
interfaces · oxygen · reduction
.
[1] a) I. Yamanaka, T. Onizawa, S. Takenaka, K. Otsuka, Angew.
Chem. 2003, 115, 3781 – 3783; Angew.Chem.Int.Ed. 2003, 42,
3653 – 3655; b) V. R. Choudhary, A. G. Gaikwad, S. D. Sansare,
Angew.Chem. 2001, 113, 1826 – 1829; Angew.Chem.Int.Ed.
4678 www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 4675 –4678