Anchoring a molecular iron catalyst to solar-responsive WO3 improves the rate and selectivity of photoelectrochemical water oxidation

Article abstract of DOI:10.1021/ja4086808

Molecular catalysts help overcome the kinetic limitations of water oxidation and generally result in faster rates for water oxidation than do heterogeneous catalysts. However, molecular catalysts typically function in the dark and therefore require sacrificial oxidants such as Ce⁴⁺ or S₂O₈^2- to provide the driving force for the reaction. In this Communication, covalently anchoring a phosphonate-derivatized complex, Fe(tebppmcn)Cl₂ (1), to WO₃ removes the need for a sacrificial oxidant and increases the rate of photoelectrochemical water oxidation on WO₃ by 60%. The dual-action catalyst, 1-WO₃, also gives rise to increased selectivity for water oxidation in pH 3 Na ₂SO₄ (56% on bare WO₃, 79% on 1-WO ₃). This approach provides promising alternative routes for solar water oxidation.

Full text of DOI:10.1021/ja4086808

Communication
pubs.acs.org/JACS
Anchoring a Molecular Iron Catalyst to Solar-Responsive WO3
Improves the Rate and Selectivity of Photoelectrochemical Water
Oxidation
Benjamin M. Klepser and Bart M. Bartlett*
Department of Chemistry, University of Michigan, 930 North University Avenue, Ann Arbor, Michigan 48109, United States
S
* Supporting Information
To overcome these challenges, we report the synthesis of a
ABSTRACT: Molecular catalysts help overcome the
kinetic limitations of water oxidation and generally result
in faster rates for water oxidation than do heterogeneous
catalysts. However, molecular catalysts typically function in
the dark and therefore require sacrificial oxidants such as
modified molecular iron catalyst, Fe(tebppmcn)Cl2 (tebppmcn
= tetraethyl N,N′-bis(2-methylpyridyl-4-phosphonate)-N,N′-
dimethylcyclohexyldiamine), (1) with a phosphonate linkage
to attach the molecule to WO3 photoelectrodes for water
oxidation. The [Fe(bpmcn)]2+ (bpmcn = N,N′-bis(2-methyl-
pyridyl)-N,N′-dimethylcyclohexyldiamine) complex without the
tether (2) was first reported by Chen and Que19 as a
hydrocarbon oxidation catalyst,20 and has also been reported as
a dark water oxidation catalyst by Fillol et al.8 New to this
manuscript, we find that covalently anchoring the molecular iron
complex to the surface of WO3 electrodes increases both the
photocurrent density (jph) generated in semiconducting WO3
and the Faradaic efficiency (η(O2)) for water oxidation. This
result suggests that light-absorbing semiconductors can be used
to photo-oxidize the molecular complex, and this strategy could
be implemented to photocatalyze other oxidation reactions with
molecular catalysts in place of sacrificial chemical oxidants.
Recently, a molecular ruthenium catalyst tethered to Fe2O3
was communicated.21 However, in this report, no evidence for
oxygen evolution is presented. Furthermore, this report did not
demonstrate prolonged stability of the anchors under illumina-
tion or under an applied bias in aqueous solutions. Carboxylate
anchors were used to attach the Ru complex to the Fe2O3 surface,
and carboxylate anchors typically suffer from weak surface
attachment in aqueous solutions.22 Our work focuses on
phosphonate anchors that demonstrate improved stability,23
and oxygen evolution has been monitored and correlated to
greater Faradaic efficiency for the modified electrodes over 3 h of
photoelectrolysis using simulated solar illumination.
With regard to synthesis, there are very few examples of
modifying the para position on the pyridine rings for the N,N′-
bis(2-methylpyridyl)-N,N′-dimethylethylenediamine class of
ligands.24 However, adapting the traditional syntheses used to
generate this class of ligands was successful for preparing the
desired phosphonate-modified complexes. After quantitative
transformation of the ortho −CH2OH substituent on the
pyridine ring to −CH2Cl, 2 equiv of this o-chloromethylpyridine
reacted moderately (up to 62%) with the deprotonated
secondary diamine to afford the desired modified ligand in two
steps from the hydroxymethylpyridine as depicted in Scheme 1.
Metalation proceeded in high yields, and was fast as characterized
2−
Ce4+ or S2O8 to provide the driving force for the
reaction. In this Communication, covalently anchoring a
phosphonate-derivatized complex, Fe(tebppmcn)Cl2 (1),
to WO3 removes the need for a sacrificial oxidant and
increases the rate of photoelectrochemical water oxidation
on WO3 by 60%. The dual-action catalyst, 1-WO3, also
gives rise to increased selectivity for water oxidation in pH
3 Na2SO4 (56% on bare WO3, 79% on 1-WO3). This
approach provides promising alternative routes for solar
water oxidation.
verall, water photolysis has been proposed as a potentially
O
viable route to generating hydrogen fuel. However,
catalyzing the water oxidation half-reaction remains a major
obstacle. Both molecular and heterogeneous catalysts have been
developed for catalyzing water oxidation, but both suffer from
drawbacks.1,2 Molecular water oxidation catalysts (WOCs) were
first reported by Meyer and co-workers with the discovery of the
ruthenium “blue dimer”,3 and they have garnered renewed
interest with the development of metal complexes capable of
oxidizing water as fast as enzymes.4,5 The first report of a non-
noble metal molecular WOC was a manganese dimer developed
by Crabtree, Brudvig, and co-workers.6 Since this report, there
are several examples of iron,7,8 cobalt,9−11 and copper5,12 WOCs.
Molecular catalysts are attractive because they are easily
modified13 and reaction kinetics are measured straightfor-
wardly.14 However, many molecular water oxidation catalysts
are not photocatalysts without adding a photosensitizer.
Consequently, most rely on using sacrificial oxidants.3−14 The
only examples of molecular complexes oxidizing water without
sacrificial oxidants require electrochemical oxidation of those
catalysts.5,12 In contrast, heterogeneous photocatalysts such as
TiO2, WO3, Fe2O3, etc. are inherently photoactive and can
perform overall water splitting (with an additional bias if
necessary).15 The distinct advantage of metal oxide semi-
conductors is that they generally exhibit excellent chemical
stability in aqueous solution. However, semiconductor photo-
catalysts are not always kinetically selective for water
oxidation,16,17 and elucidation of the reaction mechanisms is
much more challenging.18
Received: August 21, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja4086808 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
Scheme 1. Synthesis of Fe(R-bpmcn)Cl2
Figure 2. Homogeneous water oxidation of 12.5 μM complex 2 (black)
and 1 (red) in 12.5 mM (NH4)2Ce(NO3)6 aqueous solutions.
films, the Raman spectrum (Figure S2) shows that complex 1 is
bound to the WO3 electrode. In addition, the cyclic voltammetry
of the freshly modified electrodes (Figure S3) exhibits a set of
redox waves at −0.18 V vs Fc+/0, which is consistent with solution
electrochemistry, but is covered by a redox event from WO3 at
−0.3 V vs Fc+/0. There is also evidence of a slight anodic peak at
0.75 V vs Fc+/0, which is consistent with irreversible Fe3+/4+
oxidation in acetonitrile electrolytes containing 1. These
assignments have been verified by low temperature EPR
spectroscopy in nonaqueous electrolytes (Figure S4). Coulom-
etry carried out at 0.43 V vs Fc+/0 reveals the surface coverage of
the complex is ∼15 nmol/cm2 (Figure S5). Because catalyst 1
should be an electrocatalyst, the dark linear sweep voltammetry
of 1-WO3 was compared to that of bare WO3, and no noticeable
difference was observed between the two electrodes (Figure S6).
This result is most likely due to low surface coverage of 1 such
that the catalyst is unable to out-perform the high surface area
semiconductor. Also noteworthy is that 1 tethered directly to the
underlying FTO conducting substrate behaves as an electro-
catalyst only at high bias (>1.6 V vs RHE).
Confident that the WO3 electrode is modified with 1, the
photoelectrochemical performance of modified electrodes (1-
WO3) is directly compared to the same films prior to
modification in pH 3 Na2SO4. The 1-WO3 electrodes exhibit a
significant, reproducible increase in photocurrent of ∼60%. The
linear sweep voltammetry for a single representative film before
and after anchoring the complex is illustrated in Figure 3.
Furthermore, we have carried out three control experiments: if a
WO3 electrode is soaked in acetonitrile in the absence of 1, with
only FeCl2, or with only the tebppmcn ligand without iron
present, and then dried, there is no increase in the photocurrent
density on WO3 (Figure S7). Therefore, anchoring complex 1 is
necessary to observe an increase in photocurrent density on the
electrode. The saturated photocurrent is consistent with excellent
charge-carrier separation in WO3, and the increase in photo-
current is indicative of a faster chemical reaction in the presence
of the molecular catalyst. These findings also demonstrate that
the photogenerated holes from WO3 are rapidly oxidizing the
molecular complex on the surface in place of a sacrificial oxidant.
To rule out the possibility that WO3 performance alone results
in faster reaction rate, we varied the thickness of the film to
control the photocurrent density.26 The optimized thickness of
WO3 was ∼1.8 μm by SEM imaging (Figure S8), and the thinner
film was ∼0.6 μm. Although thinner films gives rise to lower
Figure 1. X-ray crystal structure of 1 where C = white, H = gray, N =
blue, O = red, P = magenta, Cl = green, Fe = yellow.
by a rapid color change from orange to purple or yellow for
complexes 1 and 2, respectively.
X-ray crystallography shows the desired structure of complex
1, in Figure 1, and elemental analysis confirms the purity. The
solution electrochemical behaviors of 1 and 2 (Figure S1) are
similar to what has been reported for similar complexes.25
Furthermore, complexes 1 and 2 exhibit similar reactivity toward
water oxidation in the presence of the sacrificial oxidant,
(NH4)2Ce(NO3)6, as reported by Fillol et al.,8 illustrated in
Figure 2. For example, after 30 min of stirring in the dark,
complexes 1 and 2 show 85 and 113 turnovers, respectively.
Despite having slightly lower stability relative to complex 2, the
initial rate of water oxidation is comparable for complexes 1 and
2.
The phosphonate-modified complex, 1, was anchored to WO3
by soaking the electrodes in 0.50 mM 1/acetonitrile at 80 °C for
8 h. WO3 electrodes were prepared by a sol−gel synthesis as
previously reported.26 Although the UV−vis diffuse reflectance
of the films is largely unchanged due to the thin nature of the
B
dx.doi.org/10.1021/ja4086808 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
Figure 3. Chopped light linear sweep voltammetry of a WO3 film as
prepared (black) and after modification with 1 (red) in pH 3 0.1 M
Na2SO4, 100 mW cm−2 AM 1.5G illumination, 20 mV/s, Pt CE, SCE
RE.
photocurrent density, as expected, thinner films of WO3 still
demonstrated a 50% increase in photocurrent density when
modified with 1 (Figure S9). This suggests that the role of the
semiconductor is primarily to photooxidize the molecule to form
the catalytically active species. Moreover, this results shows that
the rate is limited by the rate of catalyst formation.
Furthermore, it has been demonstrated previously that
complex 2 loses reactivity under more basic (pH >7)
photocatalytic reaction conditions.27 Therefore, the stability of
1-WO3 was examined under various pH. Upon adjusting the pH
of the electrolyte from 1 to 7, we observed that the jump in
photocurrent decreases slightly as pH increases from 1 to 5
(Figure S10). However, increasing the pH to 7 results in a large
decrease in photocurrent enhancement (from ∼50% to ∼20%).
This decreased performance suggests that either the complex is
becoming less active or is desorbing from the WO3 surface.
Alternatively, this decrease at pH 7 could be attributed to the
relative instability of WO3 above pH 6. To distinguish among
these possibilities, we are now synthesizing different anchors that
are known to be stable on oxide semiconductor surfaces and
different semiconductors that are stable over a wider pH range.
One problem typically encountered with WO3 is that water
oxidation is slow under acidic conditions causing electrolyte
oxidation to become a competitive side reaction, as has been
previously studied.16 Accordingly, the Faradaic efficiency for O2
evolution is significantly reduced. However, a molecular catalyst
is generally more selective toward a specific reaction than
heterogeneous oxide-based catalysts, and 1 oxidizes water in the
presence of Ce4+ (1.72 V vs RHE). As such, we predicted that the
Faradaic efficiency for O2 evolution on 1-WO3 electrodes should
be higher than that of the bare WO3 electrode.
Figure 4. Faradaic efficiency of bare (a) and 1-modified (b) WO3
electrodes in pH 3 Na2SO4 under 100 mW cm−2 AM 1.5G electrolysis at
1.23 V vs NHE, Pt CE, SCE RE; the theoretical maximum O2 based on
charge passed (black) and the amount of O2 produced (red).
films are presented in Figure S11. Furthermore, 1-WO3
maintains high Faradaic efficiency for up to 12 h under
illumination (Figure S12). The Faradaic efficiency was not
improved after modifying WO3 with only the tebppmcn ligand
(Figure S13), implying that iron is required.
To verify the increased efficiency of the modified 1-WO3
electrodes further, the quantity of other non-O2 oxidized species
−
(e.g., S2O82−, H2O2, or HSO5 ) was determined spectroscopi-
cally by adding Fe(SCN)2 and excess NaSCN, then measuring
the concentration of the deep red oxidized species in solution.16
The efficiency for generating other species is η(non-O2) = 22%
and η(non-O2) = 37% for 1-WO3 and WO3 electrodes
respectively. These long-term experiments demonstrate the
higher reactivity of the catalyst toward water oxidation on the
electrode surface under illumination. Additionally, the catalyst 1-
WO3 is performing the photooxidation of water in the absence of
other sacrificial oxidants.
In solution, the previously reported, unmodified complex 2
decomposes by ligand hydrolysis or oxidation,8 and we note that
after photoelectrolysis for 3 h, the Raman and UV−vis spectra
(Figure S2), cyclic voltammetry (Figure S3), and XP
spectoscopy (Figure S14) show no features consistent with the
presence of 1 on the WO3 surface. These observations do not
preclude the presence of a trace amount of catalyst remaining
bound to the surface, but we also acknowledge that complex 1
tethered to WO3 may be a mere precatalyst for water oxidation.
Regardless, the rate enhancement and selectivity are noteworthy.
In order to provide evidence that a catalytically active species
Oxygen detection, measured with a fluorescence lifetime
probe, indeed shows a significant increase in Faradaic efficiency
of 1-WO3 electrodes relative to the control WO3 electrodes
(η(O2) = 79 9% for 1-WO3 and η(O2) = 56 7% for WO3,
summarized in Table S1), and representative examples are
depicted in Figure 4. With a catalyst present, there is a large
increase in the selectivity of the modified electrode toward water
oxidation. To demonstrate the repeatability of film production
and performance, O2 evolution data for two different 1-WO3
C
dx.doi.org/10.1021/ja4086808 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
(3) Gersten, S. W.; Samuels, G. J.; Meyer, T. J. J. Am. Chem. Soc. 1982,
104, 4029.
(4) Duan, L.; Bozoglian, F.; Mandal, S.; Stewart, B.; Privalov, T.;
Llobet, A.; Sun, L. Nat. Chem. 2012, 4, 418.
(5) Barnett, S. M.; Goldberg, K. I.; Mayer, J. M. Nat. Chem. 2012, 4,
498.
(6) Limburg, J.; Vrettos, J. S.; Liable-Sands, L. M.; Rheingold, A. L.;
Crabtree, R. H.; Brudvig, G. W. Science 1999, 283, 1524.
(7) Ellis, W. C.; McDaniel, N. D.; Bernhard, S.; Collins, T. J. J. Am.
Chem. Soc. 2010, 132, 10990.
remains on the WO3 surface after oxygen evolution, the Faradaic
efficiency of a 1-WO3 electrode was measured for 3 h, then the
electrolyte was discarded, fresh degassed electrolyte was
introduced into the cell, and oxygen evolution from the same
1-WO3 film was repeated for another 3 h. After this second
photoelectrolysis, the increase in Faradaic efficiency for oxygen
evolution remains (Figure S15). Moreover, when the Faradaic
efficiency of an unmodified WO3 electrode was measured in the
presence of freely diffusing 0.5 μM 1 (10 nmol in 20 mL) there is
no improvement of the Faradaic efficiency for water oxidation
(Figure S16). Although we cannot say unequivocally that 1 is the
true catalyst these results support the assertion that the active
species remains on the WO3 surface and is not freely diffusing in
the electrolyte.
(8) Fillol, J. L.; Codola,
Costas, M. Nat. Chem. 2011, 3, 807.
́ ́
Z.; Garcia-Bosch, I.; Gomez, L.; Pla, J. J.;
(9) McCool, N. S.; Robinson, D. M.; Sheats, J. E.; Dismukes, G. C. J.
Am. Chem. Soc. 2011, 133, 11446.
(10) Wasylenko, D. J.; Palmer, R. D.; Schott, E.; Berlinguette, C. P.
Chem. Commun. 2011, 47, 4249.
One final point of note is that at pH 3, any iron species
removed from the electrode surface is soluble in water. ESI mass
spectrometry of the electrolyte following photoelectrolysis only
shows peaks corresponding to 1, including the hydrolysis of the
phosphonate esters, as mostly FeV or FeVI fragments (Figure
S17). A CV of the supporting electrolyte further supports high
valent iron dissociating during photoelectrolysis (Figure S18).
Although some of the complex dissociates from the surface, SEM
imaging also shows no obvious evidence of new nanoparticles
growing on the WO3 surface or change in the morphology of
WO3 after oxygen evolution (Figure S19). These data support
the notion that the catalyst is not decomposing to form Fe2O3
under illumination in the presence of oxidative holes from WO3.
In conclusion, we have demonstrated successfully combining
an earth-abundant molecular catalyst, Fe(tebppmcn)Cl2 (1),
with the semiconducting photoelectrode, WO3, dramatically
increases the rate and selectivity of photoelectrochemical water
oxidation. Furthermore, we remove the need for a sacrificial
chemical oxidant such as Ce4+. These results reveal the promise
of enhancing water oxidation using catalytically active molecular
complexes anchored onto stable, solar-responsive semiconduc-
tors.
(11) Hong, D.; Jung, J.; Park, J.; Yamada, Y.; Suenobu, T.; Lee, Y.;
Nam, W.; Fukuzumi, S. Energy Environ. Sci. 2012, 5, 7607.
(12) Zhang, M.-T.; Chen, Z.; Kang, P.; Meyer, T. J. J. Am. Chem. Soc.
2013, 135, 2048.
(13) Zong, R.; Thummel, R. P. J. Am. Chem. Soc. 2005, 127, 12802−
12803.
(14) Representative examples: (a) Wasylenko, D. J.; Ganesamoorthy,
C.; Koivisto, B. D.; Henderson, M. A.; Berlinguette, C. P. Inorg. Chem.
2010, 49, 2202. (b) Concepcion, J. J.; Jurss, J. W.; Templeton, J. L.;
Meyer, T. J. J. Am. Chem. Soc. 2008, 130, 16462. (c) Duan, L.; Xu, Y.;
Tong, L.; Sun, L. Chem. Sus. Chem. 2011, 4, 238.
(15) (a) Fujishima, A.; Honda, K. Nature 1972, 238, 37. (b) Butler, M.
A.; Nasby, R. D.; Quinn, R. K. Solid State Commun. 1976, 19, 1011.
(c) Hardee, K. L.; Bard, A. J. J. Electrochem. Soc. 1976, 123, 1024.
(16) Mi, Q.; Zhanaidarova, A.; Brunschwig, B. S.; Gray, H. B.; Lewis, N.
S. Energy Environ. Sci. 2012, 5, 5694.
(17) (a) Hill, J. C.; Choi, K.-S. J. Phys. Chem. C 2012, 116, 7612. (b) Mi,
Q.; Coridan, R. H.; Bruschwig, B. S.; Gray, H. B.; Lewis, N. S. Energy
Environ. Sci. 2013, 6, 2646.
(18) Klahr, B.; Gimenez, S.; Fabregat-Santiago, F.; Bisquert, J.;
Hamann, T. W. Energy Environ. Sci. 2012, 5, 7626.
(19) Chen, K.; Que, L., Jr. Chem. Commun. 1999, 1375.
(20) (a) Costas, M.; Tipton, A. K.; Chen, K.; Jo, D.-H.; Que, L., Jr. J.
Am. Chem. Soc. 2001, 123, 6722. (b) Costas, M.; Que, L., Jr. Angew.
Chem., Int. Ed. 2002, 41, 2179.
(21) Chen, X.; Ren, X.; Liu, Z.; Zhuang, L.; Lu, J. Electrochem. Commun.
2013, 27, 148.
(22) Representative examples: (a) Gillaizeau-Gauthier, I.; Odobel, F.;
Alebbi, M.; Argazzi, R.; Costa, E.; Bignozzi, C. A.; Qu, P.; Meyer, G. J.
Inorg. Chem. 2001, 40, 6073. (b) Park, H.; Bae, E.; Lee, J.-J.; Park, J.;
Choi, W. J. Phys. Chem. B 2006, 110, 8740.
(23) Representative examples: (a) Chen, Z.; Concepcion, J. J.; Jurss, J.
W.; Meyer, T. J. J. Am. Chem. Soc. 2009, 131, 15580. (b) Lee, S-H. A.;
Zhao, Y.; Hernandez-Pagan, E. A.; Blasdel, L.; Youngblood, W. J.;
Mallouk, T. E. Faraday Discuss. 2012, 155, 165. (c) Giokas, P. G.; Miller,
S. A.; Hanson, K.; Norris, M. R.; Glasson, C. R. K.; Concepcion, J. J.;
Bettis, S. E.; Meyer, T. J.; Moran, A. M. J. Phys. Chem. C 2013, 117, 812.
ASSOCIATED CONTENT
* Supporting Information
■
S
Experimental procedures, UV−vis and Raman spectra, cyclic
voltammograms, O2 evolution data, SEM images, and a CIF file.
This material is available free of charge via the Internet at http://
pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
bartmb@umich.edu
■
Notes
(24) (a) Delroisse, M.; Rabion, A.; Chardac, F.; Tet
́
ard, D.; Verlhac, J.-
The authors declare no competing financial interest.
B.; Fraisse, L.; Seris, J.-L. J. Chem. Soc., Chem. Commun. 1995, 949.
́
(b) Mialane, P.; Tchertanov, L.; Banse, F.; Sainton, J.; Girerd, J.-J. Inorg.
Chem. 2000, 39, 2440. (c) England, J.; Gondhia, R.; Bigorra-Lopez, L.;
Petersen, A. R.; White, A. J. P.; Britovsek, G. J. P. Dalton Trans. 2009,
5319.
(25) Coates, C. M.; Hagan, K.; Mitchell, C. A.; Gorden, J. D.;
Goldsmith, C. R. Dalton Trans. 2011, 40, 4048.
ACKNOWLEDGMENTS
■
This research was supported by a grant from the United States
Department of Energy (DE-FG02-11ER16262). We thank Dr.
Jeff W. Kampf for assistance with X-ray crystallography.
(26) Santato, C.; Odziemkowski, M.; Ulmann, M.; Augustynski, J. J.
Am. Chem. Soc. 2001, 123, 10639.
(27) Chen, G.; Chen, J.; Ng, S.-M.; Man, W.-L.; Lau, T.-C. Angew.
Chem., Int. Ed. 2013, 52, 1.
REFERENCES
■
(1) Reviews of molecular WOCs: (a) Brimblecombe, R.; Dismukes, G.
C.; Swiegers, G. F.; Spiccia, L. Dalton Trans. 2009, 9374. (b) Duan, L.;
Tong, L.; Xu, Y.; Sun, L. Energy Environ. Sci. 2011, 4, 3296.
(2) Reviews of heterogeneous WOCs: (a) Walter, M. G.; Warren, E. L.;
McKone, J. R.; Boettcher, S. W.; Mi, Q.; Santori, E. A.; Lewis, N. S.
Chem. Rev. 2010, 110, 6446. (b) Kudo, A.; Miseki, Y. Chem. Soc. Rev.
2009, 38, 253.
D
dx.doi.org/10.1021/ja4086808 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX