Communication

pubs.acs.org/JACS

A Bio-Inspired, Small Molecule Electron-Coupled-Proton Buffer for

Decoupling the Half-Reactions of Electrolytic Water Splitting
Benjamin Rausch, Mark D. Symes, and Leroy Cronin*
WestCHEM, School of Chemistry, University of Glasgow, University Avenue, Glasgow G12 8QQ, U.K.
*
S Supporting Information

Scheme 1. Analogies between Natural Water Splitting during

ABSTRACT: Electron-coupled-proton buffers (ECPBs) Photosynthesis (Green Cycle) and an Artificial ECPB-
allow H2 and O2 evolution to be separated from each other Mediated Water Splitting Cycle (Blue Arrows)
in time during the electrolysis of water. Natural photo-
synthetic systems achieve an analogous feat during water
splitting and employ a range of intermediate redox
mediators such as quinone derivatives to aid this process.
Drawing on this natural example, we show that a low
molecular weight quinone derivative is capable of
decoupling H2 evolution from O2 evolution at scale during
electrochemical water splitting. This work could signifi-
cantly lower the cost of ECPBs, paving the way for their
more widespread adoption in water splitting.

H ydrogen has great potential as a clean fuel and energy

storage medium in the putative “hydrogen economy”.1
Currently, however, the majority of the world’s H2 is produced
abundant elements would be beneficial. However, given the
demands required of an effective ECPB (an appropriately
by the reformation of fossil fuels, which is neither sustainable positioned reversible redox wave, high solubility in water,
nor environmentally friendly.2 An alternative and sustainable stability in both oxidized and reduced forms, ability to buffer
source of H2 is water, via electrolysis driven by renewable the pH during water splitting, and low cost of the components),
power inputs such as wind and solar.3 If this approach is to it was far from obvious that the concept could be extended
become more widespread, it must be made more economical, beyond the polyoxometalates. Herein, we show that quinones
and hence there is an imperative to investigate new paradigms can be used as ECPBs, combining the requisites listed above
in electrolytic H2O splitting and electrolyzer design.4 with low molecular weight (MW), abundance of the
Recently, we introduced the concept of the electron-coupled- components, and production on an industrial scale, features
proton buffer (ECPB), which allows H2 and O2 to be produced which are essential if the promise of ECPB-mediated water
at separate times during electrolytic water splitting.5 By splitting is to be realized.
decoupling the oxygen-evolving reaction (OER) from the Quinone derivatives, such as plastoquinone and coenzyme
hydrogen-evolving reaction (HER) in this way, it may prove Q10, are known to act as mitochondrial redox carriers in the
possible to replace (or significantly improve the lifespan of) photosynthetic electron transport chain in bacteria and green
Nafion in proton electrolyte membrane electrolyzers, with plants.7 They,8 and other biomimetic redox mediators,9 have
implications for the durability and price of such devices.6 also recently been used as redox-active components in the
Likewise, decoupling the OER and HER gives greater flexibility electrolyte of dye-sensitized solar cells. Plastoquinone accepts
over when, and how fast, the products of water splitting are electrons from Photosystem II during water oxidation and
made. In an ECPB electrolysis cell, water is oxidized at the protons from the chloroplast stromal matrix, such that an
anode to give O2, protons, and electrons. At the cathode, these initially oxidized benzoquinone derivative undergoes a
protons and electrons are used to reversibly reduce and reversible two-electron, two-proton-transfer process to form a
protonate the ECPB (instead of making H2 directly), meaning reduced hydroquinone. These electrons and protons stored on
that no H2 is produced during the water oxidation phase. the hydroquinone are then ultimately used to generate H2
Subsequently, the reduced and protonated ECPB (hereafter equivalents in the form of NADH (Scheme 1).10−12 Inspired by
termed ECPB*) is reoxidized, regenerating oxidized ECPB and these natural examples, we hypothesized that water-soluble
releasing protons and electrons which combine at the cathode quinone derivatives could function as low MW, organic ECPBs
to give H2 that is completely free of O2. This train of events is in electrolytic cells, allowing effective decoupling of the OER
summarized in Scheme 1 and Figure 1. In our initial study,5 we from the HER.
used the all-inorganic polyoxometalate phosphomolybdic acid
(H3Mo12PO40, MW > 1800 g mol−1) as an ECPB but noted Received: July 14, 2013
that ECPBs with lower molecular weights and based on more

© XXXX American Chemical Society A dx.doi.org/10.1021/ja4071893 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX
Journal of the American Chemical Society Communication

Figure 1. ECPB-mediated H2O splitting with decoupled OER and HER. The colorless hydroquinone (shown in blue) is oxidized at a carbon
electrode (black rectangle) to form the red benzoquinone. During this process, two protons and two electrons are released per hydroquinone
molecule (right). The protons migrate across the membrane, while the electrons pass through the external circuit. The protons and electrons
recombine on the Pt electrode (gray rectangle) to form H2. When the process is reversed, H2O is oxidized to give O2, protons, and electrons on the
Pt electrode. The H+ and electrons are used to reduce and protonate the benzoquinone, re-forming the hydroquinone. Cyclic voltammetry (center)
shows a 10 mM solution of the hydroquinone sulfonate in 1.8 M H3PO4 (pH 0.7), at a modified glassy carbon working electrode (area 0.071 cm2),
using Pt counter and Ag/AgCl reference electrodes at a scan rate of 0.1 V s−1.

Herein we show that the commercially available 1,4-

hydroquinone derivative potassium hydroquinone sulfonate is
suitable as an ECPB. The 1,4-hydroquinone precursor is
manufactured by oxidative cleavage of diisopropylbenzene,
oxidation of aniline, or hydroxylation of phenol on a scale of 40
000−50 000 tons annually.13 By aromatic sulfonation, a sulfonic
group can be added to the aromatic ring system such that both
the reduced hydroquinone sulfonate and the oxidized
benzoquinone sulfonate are very water-soluble. This allows
access to suitably high concentrations of the ECPB in aqueous
solutions (0.5 M).
An effective ECPB should have a reversible redox wave
between the OER and HER onsets and should attenuate the Figure 2. Three-electrode i−V curves with an Ag/AgCl reference
large fluctuations in pH which would otherwise arise as a result electrode and a large-area Pt mesh counter electrode: (a) reduction of
of water oxidation and proton reduction being decoupled from ECPB or protons; (b) oxidation of ECPB* or water. Black line and
one another.5 The parent 1,4-hydroquinone/p-benzoquinone squares: 1.8 M H3PO4 (pH 0.7) and a Pt disc working electrode (area
redox couple lies at around +0.7 V (vs NHE) under standard = 0.0314 cm2). Red line and circles: 0.5 M ECPB/ECPB* in 1.8 M
conditions,14 and cyclic voltammetry (see the center of Figure H3PO4 (pH 0.7) and a glassy carbon working electrode (area = 0.071
1) shows that the redox couple of the sulfonate derivative used cm2).
in this work is at +0.65 V (all three-electrode potentials vs
NHE) in 1.8 M H3PO4 at pH 0.7 on a modified glassy carbon oxidation of the ECPB* on a glassy carbon proceeds at +0.90
electrode.15 The peak current for this wave increases linearly V, whereas the oxidation of water on a Pt electrode at 50 mA
with the square root of the scan rate (see Supporting cm−2 requires +2.23 V.
Information, Figure S1), indicating a solution-phase process Given that the use of a three-electrode configuration
that is limited by diffusion effects. minimizes the effects of resistance,17 it is possible to compare
The efficacy of hydroquinone sulfonate as an ECPB was then the expected energetics of a two-step ECPB system (where a
probed by examining the voltages required to produce a given glassy carbon electrode oxidizes and reduces the ECPB and a Pt
current density in the ECPB electrolysis cell. To be able to electrode alternately oxidizes water and reduces protons) with a
study both oxidations and reductions of the ECPB under single-step system that uses two Pt electrodes to produce O2
conditions where both forms were present at high concen- and H2 simultaneously, by simply summing the voltages
trations, a 0.5 M solution of hydroquinone sulfonate in 1.8 M required to produce a given current density. Hence, to oxidize
H3PO4 (pH 0.7) was oxidized until a statistical 50:50 mix of the water and reduce the ECPB at 50 mA cm−2 requires 2.23 −
oxidized and two-electron-reduced states (hereafter termed 0.23 = 2.00 V, with the reverse step (reoxidation of ECPB* and
ECPB/ECPB*) had been produced.16 The low pH was found simultaneous proton reduction to H2) needing +0.9 − (−0.39)
necessary in order to ensure adequate reversibility in the redox = 1.29 V. This gives an overall voltage requirement of 3.29 V
wave (Figures S2−S10). Figure 2 shows how the current for the two-step process, compared to 2.62 V (+2.23 −
densities obtained in a three-electrode configuration at a Pt (−0.39)) for the single-step reaction.
working electrode in 1.8 M H3PO4 compare with the current This gives an expected efficiency of 80% for the two-step
densities that can be achieved at a glassy carbon electrode in an process relative to the single-step process in the absence of
ECPB/ECPB* solution (see Supporting Information for resistive factors, although we note that the single-step reaction
details). At a benchmark current density of 50 mA cm−2, requires two precious metal electrodes, compared to just one
Figure 2a shows that reduction of ECPB to ECPB* on glassy precious metal electrode when an ECPB is used. Similarly,
carbon will proceed at 50 mA cm−2 when a bias of +0.23 V is Figure S11 shows that to reduce protons at 50 mA cm−2 on a
applied, while −0.39 V is required to reduce protons to H2 on glassy carbon electrode requires a potential of −0.77 V, which
Pt at this current density. Conversely, Figure 2b shows that the means that a single-step process for generating H2 on a carbon
B dx.doi.org/10.1021/ja4071893 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX
Journal of the American Chemical Society Communication

cathode and O2 on a Pt anode requires a total voltage of 3.0 V

to run at 50 mA cm−2 (giving the ECPB-mediated process 91%
efficiency relative to a one-step process that uses a single Pt
electrode). Moreover, as the two-step ECPB process uses two
smaller power inputs to split water (compared to a single larger
power input in a non-ECPB cell), using an ECPB reduces the
instantaneous power required to do a productive step in the
water-splitting reaction, potentially allowing lower power and
more diffuse energy inputs to be used to generate hydrogen
from water.
Next, we moved from the ideal situation found with three-
electrode configurations to a two-electrode configuration such
as that shown in Figure 1. Hence, a 0.5 M solution of ECPB/ Figure 4. (a) Two-electrode i−V curves for a two-compartment cell
ECPB* (in 1.8 M H3PO4) was placed into one compartment of containing a 0.5 M solution of ECPB/ECPB* in H3PO4 (1.8 M, pH
a two-compartment H-cell, and the second compartment was 0.7) in the counter electrode compartment and 1.8 M H3PO4 in the
filled with pure 1.8 M H3PO4 (pH 0.7). The two compartments working electrode compartment (red line and circles), or 1.8 M
were separated by a Nafion 118 membrane. A Pt “working” H3PO4 in both sides (black lines). In each case the gas-evolving
electrode (area = 0.031 cm2) was placed into the chamber filled electrode was a Pt-disc working electrode (area = 0.0314 cm2) in 1.8
with pure 1.8 M H3PO4, while a large area carbon felt was used M H3PO4. Using the ECPB, the counter electrode was a large-surface
for reduction/oxidation of the ECPB. Various potentials were carbon felt in the ECPB/ECPB* solution. When 1.8 M H3PO4 was
used in both compartments, data were collected using both a large-area
then applied across this cell, and the current densities Pt mesh (black line and triangles) and a large-area carbon cloth (black
(normalized to the area of the Pt electrode performing the line and circles) in the counter electrode compartment. (b) A
gas evolution reactions) were recorded as shown in Figures 3a representative trace showing %O2 expected in the cell headspace
and 4a. (calculated on the basis of the amount of charge passed, black line and
squares) compared to the %O2 in the headspace measured by GCHA
(red line and circles).

Figure S12), gas chromatography headspace analysis (GCHA)

indicated that the amount of hydrogen produced corresponded
to a Faradaic efficiency of 98% ± 7% for the HER (Figure 3b).
The amount of O2 originating from water splitting under these
conditions was determined to be <2% of that which would be
expected if the HER and OER were occurring concurrently (see
Supporting Information for details of the determination of
percentages of gases in the cell headspace). The pH changes
that might be expected when the HER is decoupled from the
OER were significantly diminished by using hydroquinone
Figure 3. (a) Two-electrode i−V curves for a two-compartment cell sulfonate as an electron mediator (see Supporting Information,
containing a 0.5 M solution of ECPB/ECPB* in H3PO4 (1.8 M, pH
0.7) in one compartment and 1.8 M H3PO4 in the other (red line and
Table S1), supporting the hypothesis that the 1,4-hydro-
circles), or 1.8 M H3PO4 in both sides (black line and triangles). In quinone sulfonate/p-benzoquinone sulfonate system accepts
each case the gas-evolving electrode was a Pt-disc working electrode and donates both protons and electrons during the decoupled
(area = 0.0314 cm2) in 1.8 M H3PO4. Using the ECPB, the other half-reactions of water splitting and thus acts as an ECPB.
electrode was a large-area carbon felt in the ECPB/ECPB* solution, Figure 4a shows current densities for O2 evolution and
while a large-area Pt-mesh was used in the cell containing 1.8 M simultaneous ECPB reduction. Current densities of 0.5 A cm−2
H3PO4 in both sides to avoid oxidative degradation of the anode. (b) could be achieved when an iR-corrected voltage of 1.90 V was
A representative trace showing %H2 expected in the cell headspace applied across a cell using a Pt anode and a carbon cathode/
(calculated on the basis of the amount of charge passed, black line and ECPB combination (red line and circles), but 2.78 V had to be
circles) compared to the %H2 in the headspace measured by GCHA applied across this cell to obtain a current density of 0.5 A cm−2
(red line and squares).
when no ECPB was present (black line and circles). The O2
headspace concentration found by GCHA indicated a Faradaic
As shown in Figure 3a, current densities of 0.5 A cm−2 could yield of 91 ± 5% for the production of O2. No H2 was detected,
be achieved when an iR-corrected voltage of 0.98 V was applied again suggesting that the ECPB/ECPB* mixture allowed the
across the ECPB cell (red line and circles). However, when no OER and HER to be fully decoupled.
ECPB* was used (black line and triangles), the current density In order to test the stability of this ECPB to repeated cycling,
only reached 0.5 A cm−2 when the cell bias was >2.5 V and a a 10 mM solution of potassium hydroquinone sulfonate was
large-area Pt mesh was used in place of the carbon felt/ECPB* alternately fully oxidized and fully reduced in a two-electrode,
combination. The use of a Pt mesh was essential when no two-compartment H-cell. Figure S13b shows the resulting plot
ECPB* was present to oxidize, as when carbon felt is polarized of charge passed vs cycle number, which shows that the ability
anodically in the absence of easily oxidized species the reaction to re-reduce the oxidized form of the ECPB decreased by
that occurs is oxidation of the carbon anode and not oxidation around 1% per cycle (over 20 cycles). This decrease in charge
of water to O2.18 When the ECPB/ECPB* mixture was storage capacity was matched by a color change to red, which
oxidized with a carbon felt electrode and 1.8 M H3PO4 was was recorded at regular intervals during electrochemical cycling
reduced concomitantly at a Pt mesh in an airtight cell (see of the solution by UV/vis spectroscopy (Figures S13a and
C dx.doi.org/10.1021/ja4071893 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX
Journal of the American Chemical Society Communication

S14). Exclusion of O2 made little difference to the rate of (2) Holladay, J. D.; Hu, J.; King, D. L.; Wang, Y. Catal. Today 2009,
discoloration, but less concentrated solutions were observed to 139, 244.
suffer slower degradation than more concentrated solutions. (3) (a) Lewis, N. S.; Nocera, D. G. Proc. Natl. Acad. Sci. U.S.A. 2006,
Taken together, these results suggest that the oxidized form of 103, 15729. (b) Gust, D.; Moore, T. A.; Moore, A. L. Acc. Chem. Res.
2009, 42, 1890. (c) Cook, T. R.; Dogutan, D. K.; Reece, S. Y.;
the ECPB is somewhat unstable with respect to oligomeriza- Surendranath, Y.; Teets, T. S.; Nocera, D. G. Chem. Rev. 2010, 110,
tion, as previously observed with benzoquinone derivatives 6474. (d) Chen, X.; Shen, S.; Guo, L.; Mao, S. S. Chem. Rev. 2010,
under highly acidic conditions.19 In contrast, the reduced 110, 6503. (e) Artero, V.; Chavarot-Kerlidou, M.; Fontecave, M.
hydroquinone form (ECPB*) could be stored for several days Angew. Chem., Int. Ed. 2011, 50, 7238. (f) Styring, S. Faraday Discuss.
in the dark without significant aerial oxidation or degradation, 2012, 155, 357.
as gauged by UV/vis (see Figure S15) and coulometry (after 8 (4) (a) Smith, D. W.; Oladoyinbo, F. O.; Mortimore, W. A.;
days under air in the dark, >98% of the charge stored in a 0.5 M Colquhoun, H. M.; Thomassen, M. S.; Odegard, A.; Guillet, N.;
sample of ECPB* could be recovered for H2 production). Mayousse, E.; Klicpera, T.; Hayes, W. Macromolecules 2013, 46, 1504.
The use of larger or more sterically hindered quinone (b) Ursua, A.; Gandia, L. M.; Sanchis, P. Proc. IEEE 2012, 100, 410.
derivatives would be expected to significantly retard oligome- (c) Trainham, J. A.; Newman, J.; Bonino, C. A.; Hoertz, P. G.;
Akunuri, N. Curr. Opin. Chem. Eng. 2012, 1, 204. (d) Carmo, M.; Fritz,
rization of the oxidized form of the ECPB, and should thus
D. L.; Mergel, J.; Stolten, D. Int. J. Hydrogen Energy 2013, 38, 4901.
extend the cyclability of the system. Larger ECPB molecules (e) Blankenship, R. E.; Tiede, D. M.; Barber, J.; Brudvig, G. W.;
could also be beneficial with regard to ECPB permeation Fleming, G.; Ghirardi, M.; Gunner, M. R.; Junge, W.; Kramer, D. M.;
through the membrane. Nafion was chosen as the membrane in Melis, A.; Moore, T. A.; Moser, C. C.; Nocera, D. G.; Nozik, A. J.; Ort,
these studies because it hinders the movement of anions while D. R.; Parson, W. W.; Prince, R. C.; Sayre, R. T. Science 2011, 332,
remaining freely permeable to protons. However, we found that 805.
the Nafion membrane was not entirely impermeable to this (5) Symes, M. D.; Cronin, L. Nature Chem. 2013, 5, 403.
ECPB/ECPB* couple, which was able to cross the Nafion (6) Barbir, F. Solar Energy 2005, 78, 661.
membrane at a rate of ∼2.7 × 10−6 mol h−1 (determined by (7) (a) Crane, F. L. Mitochondrion 2007, 7, S2. (b) Green, D. E.; Mii,
UV/vis spectroscopy, see Supporting Information and Figure S.; Kohout, P. M. J. Biol. Chem. 1955, 217, 551. (c) Green, D. E.;
Järnefelt, J.; Tisdale, H. D. Biochim. Biophys. Acta 1959, 31, 34.
S16). This could potentially lead to nonproductive cycling of (d) Saito, K.; Rutherford, A. W.; Ishikita, H. Proc. Natl. Acad. Sci.
the ECPB, which could be reduced and protonated on one U.S.A. 2013, 110, 954.
electrode, cross the membrane, and become reoxidized at the (8) Cheng, M.; Yang, X. C.; Zhang, F. G.; Zhao, J. H.; Sun, L. C.
other electrode. Larger quinone derivatives would be expected Angew. Chem., Int. Ed. 2012, 51, 9896.
to cross the membrane less readily, boosting overall efficiency. (9) Zhao, Y.; Swierk, J. R.; Megiatto, J. D., Jr.; Sherman, B.;
In conclusion, we have shown that the hydroquinone/ Youngblood, W. J.; Qin, D.; Lentz, D. M.; Moore, A. L.; Moore, T. A.;
benzoquinone redox couple has potential as the basis for a Gust, D.; Mallouk, T. E. Proc. Natl. Acad. Sci. U.S.A. 2012, 109, 15612.
system of low MW, organic ECPBs that can effectively (10) Brandt, U. Annu. Rev. Biochem. 2006, 75, 69.
decouple the OER from the HER during electrolytic water (11) (a) Renger, G.; Renger, T. Photosynth. Res. 2008, 98, 53.
splitting. This work not only extends the ECPB concept to (b) Siggel, U. Bioelectrochem. Bioenerg. 1976, 3, 302.
(12) Lubitz, W.; Reijerse, E. J.; Messinger, J. Energy Environ. Sci.
include organic molecules as redox mediators, but also 2008, 1, 15.
highlights parallels with the role of quinones in photosynthetic (13) (a) Hudnall, P. M. Ullmann’s Encyclopedia of Industrial
water splitting. Work to improve the stability and cyclability of Chemistry; Wiley-VCH Verlag GmbH & Co. KGaA: Berlin, 2000.
such organic ECPBs is ongoing in our laboratories.

■
(b) Weissermel, K.; Arpe, H.-J. Industrial Organic Chemistry; Wiley-
VCH Verlag GmbH: Berlin, 2007; pp 335−383.
ASSOCIATED CONTENT (14) Vanýsek, P. In CRC Handbook of Chemistry and Physics, 93rd
*
S Supporting Information ed.; Haynes, W. M., Ed.; CRC Press: New York, 2012−2013.
Experimental protocols for electrochemistry, gas analysis, (15) Chen, L.; Tang, Y.; Wang, K.; Liu, C.; Luo, S. Electrochem.
operation of the ECPB in alternative electrolytes, and stability Commun. 2011, 13, 133.
(16) Laviron, E. J. Electroanal. Chem. 1984, 164, 213.
tests. This material is available free of charge via the Internet at (17) (a) Hamann, C. H.; Hamnett, A.; Vielstich, W. Electrochemistry,
http://pubs.acs.org.

■
2nd ed.; Wiley-VCH: Weinheim, 2007. (b) Hernández-Pagán, E. A.;
Vargas-Barbosa, N. M.; Wang, T. H.; Zhao, Y. X.; Smotkin, E. S.;
AUTHOR INFORMATION Mallouk, T. E. Energy Environ. Sci. 2012, 5, 7582.
Corresponding Author (18) Engstrom, R. C.; Strasser, V. A. Anal. Chem. 1984, 56, 136.
lee.cronin@glasgow.ac.uk (19) Erdtman, H.; Högberg, H.-E. Tetrahedron 1979, 35, 535.
Notes
The authors declare no competing financial interest.

■ ACKNOWLEDGMENTS
We thank the EPSRC for funding. M.D.S. acknowledges
Glasgow University for a Kelvin Smith Fellowship. L.C. thanks
the Royal Society/Wolfson Foundation for a Merit Award.

■ REFERENCES
(1) (a) Armaroli, N.; Balzani, V. ChemSusChem 2011, 4, 21. (b) Olah,
G. A.; Prakash, G. K. S.; Goeppert, A. J. Am. Chem. Soc. 2011, 133,
12881. (c) Walter, M. G.; Warren, E. L.; McKone, J. R.; Boettcher, S.
W.; Mi, Q. X.; Santori, E. A.; Lewis, N. S. Chem. Rev. 2010, 110, 6446.
(d) Olah, G. A. Angew. Chem., Int. Ed. 2013, 52, 104.

D dx.doi.org/10.1021/ja4071893 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX