Photochemical oxidation of water and reductionof polyoxometalate anions at interfaces ofwater with ionic liquids or diethyletherGianluca Bernardinia, Anthony G. Weddb, Chuan Zhaoc, and Alan M. Bonda,d,1

aSchool of Chemistry, Monash University, Clayton, Victoria 3800, Australia; bSchool of Chemistry and the Bio21 Institute of Molecular Science andBiotechnology, University of Melbourne, Victoria 3010, Australia; cSchool of Chemistry, University of New South Wales, Sydney, New South Wales 2052,Australia; and dARC Centre for Green Chemistry, Monash University, Clayton, Victoria 3800, Australia

Edited by* Allen J. Bard, The University of Texas at Austin, Austin, TX, and approved June 4, 2012 (received for review March 5, 2012)

Photoreduction of ½P2W18O62�6−, ½S2Mo18O62�4−, and ½S2W18O62�4−polyoxometalate anions (POMs) and oxidation of water occurswhen water–ionic liquid and water–diethylether interfaces areirradiated with white light (275–750 nm) or sunlight. The ionicliquids (ILs) employed were aprotic ([Bmim]X; Bmim ¼ ð1-butyl-3-methylimidazolium,X ¼ BF4,PF6Þ and protic (DEAS ¼diethanolaminehydrogen sulphate; DEAP ¼ diethanolaminehydrogen phosphate). Photochemical formation of reduced POMsat both thermodynamically stable and unstable water–IL interfacesled to their initial diffusion into the aqueous phase and subsequentextraction into the IL phase. The mass transport was monitoredvisually by color change and by steady-state voltammetry at micro-electrodes placed near the interface and in the bulk solutionphases. However, no diffusion into the organic phase was observedwhen ½P2W18O62�6− was photo-reduced at the water–diethyletherinterface. In all cases, water acted as the electron donor to give theoverall process: 4POMþ 2H2Oþ hν → 4POM− þ 4Hþ þO2. How-ever, more highly reduced POM species are likely to be generated asintermediates. The rate of diffusion of photo-generated POM− wasdependent on the initial concentration of oxidized POM and theviscosity of the IL (or mixed phase systemproduced in cases in whichthe interface is thermodynamically unstable). In the water-DEASsystem, the evolution of dioxygen was monitored in situ in theaqueous phase by using a Clark-type oxygen sensor. Differences inthe structures of bulk and interfacial water are implicated in the ac-tivation of water. An analogous series of reactions occurred uponirradiation of solid POM salts in the presence of water vapor.

electrochemistry ∣ water oxidation

It has been known for some time that photochemical reductionof polyoxometalate anions (POMs) in molecular solvents may

occur in the presence of an efficient electron donor such as 2-pro-panol or benzyl alcohol (1–6). In recent studies, we have foundthat photoreduction of tetracyanoquinodimethane (TCNQ) toTCNQ− and ½P2W18O62�6− to ½P2W18O62�7− or more extensivelyreduced POMs occurs in “wet” ionic liquids (ILs) where wateracts as an electron donor and is photo-oxidized to dioxygen (7, 8).Intriguingly, these processes do not occur in wet molecular organ-ic solvents or in neat water itself. The net reactions that describethe photo-irradiation of ½P2W18O62�6− in wet 1-butyl-3-methyli-midazolium tetrafluoroborate [Bmim][BF4] (Fig. S1) are sum-marized in Eqs. 1 and 2:

4½P2W18O62�6− þ hν → 4½P2W18O62�6−�; [1]

4½P2W18O62�6−� þ 2H2O → 4½P2W18O62�7− þO2 þ 4Hþ: [2]

However, while ½P2W18O62�7− is the detected product it is prob-able that more extensively reduced POMs are generated (8) asintermediates in the overall water oxidation reaction [Eq. 3]:

2H2O → O2 þ 4Hþ þ 4e−: [3]

The modified structure of water present as a solute in ILsrelative to that found in bulk water is postulated to facilitateH2O acting as an electron donor (7–10). However, while multi-step two- and four-electron-proton-coupled reaction schemeshave been postulated, full mechanistic details have yet to beestablished (8).

To date, photochemical reactions have rarely been carried outin ILs even though these low temperature molten salts that con-sist entirely of ionic species have emerged as an important classof solvent (11–18). Nevertheless, it is known that IL media canprovide a favorable solvent polarity effect for photo-induced elec-tron transfer (19). Both protic and aprotic ILs are often sparinglysoluble or immiscible in water, so certain mixtures of ILs andwater give rise to two phases. Furthermore, on examining ILsas a medium for photooxidation of water (7, 8), it was noted thateven if an IL is fully miscible with water, the kinetics of itsdissolution are often sufficiently slow that two phases may persistfor long periods of time. Consequently, studies on the photo-chemistry at both thermodynamically stable and unstablewater–IL interfaces should be possible in a regime in which sig-nificant structural kinetic and thermodynamic differences occurrelative to bulk solvents. Consequently, photooxidation of waterat interfaces may be even more favorable than in bulk wet ILs.

Solubility issues play an important role in determining whichphotoreactions may be studied at water–IL interfaces. Sodiumand potassium salts of polyoxometalate anions are often verysoluble in aqueous media. In contrast, tetraalkylammonium saltssuch as ½Bu4N�6½P2W18O62� are sparingly soluble in water (oreven insoluble) but are reasonably soluble in many ILs. However,dissolution of these salts in highly viscous aprotic ILs is oftenkinetically slow even though mM concentrations or above arethermodynamically accessible (20–22). Typically, reduced poly-oxometalate anions are substantially more soluble in IL media.For example, one-electron reduced ½P2W18O62�7− is highly solu-ble in many aprotic and protic ILs even in the presence of Kþcounter ions (23, 24). It is apparent that a water–IL interface witha POM present in either phase can provide an environmentwhere a variety of mass transport outcomes are possible.

ILs may act as both the solvent and electrolyte, so voltam-metric detection of electroactive POM species is possible by care-ful placement of electrodes in the bulk solution and near theinterface. A similar situation applies in an aqueous medium or

Author contributions: C.Z. and A.M.B. designed research; G.B. performed research; G.B.,A.G.W., C.Z., and A.M.B. analyzed data; and G.B., A.G.W., C.Z., and A.M.B. wrote the paper.

The authors declare no conflict of interest.

*This Direct Submission article had a prearranged editor.1To whom correspondence should be addressed. E-mail: alan.bond@monash.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1203818109/-/DCSupplemental.

11552–11557 ∣ PNAS ∣ July 17, 2012 ∣ vol. 109 ∣ no. 29 www.pnas.org/cgi/doi/10.1073/pnas.1203818109

Dow

nloa

ded

by g

uest

on

Apr

il 24

, 202

0

organic solvent phases, provided an electrolyte is added. In prin-ciple, voltammetric monitoring of a two-phase system will allowthe level of reactivity to be probed in both phases. An organic–aqueous interface provides another scenario to study photochem-istry in which the structure of interfacial water is again expectedto differ from that in bulk water. In this study we show that photo-oxidation of water in the presence of photoactive ½P2W18O62�6−,½S2Mo18O62�4−, or ½S2W18O62�4− anions occurs preferentially atwater–IL and water–diethylether interfaces. We emphasize theuse of electroanalytical techniques of transient and steady-statevoltammetry to establish the POM redox level [Eq. 2] and todetect dioxygen and proton concentrations (Eq. 3; the Clark andpH electrodes).

Results and DiscussionAs noted above, K6½P2W18O62� is highly soluble in water but hasvery limited solubility in the aprotic and protic ionic liquids usedin this study. However, the solubility of the reduced speciesformed in photo-irradiation experiments (8) is significantly high-er in the ILs. This is illustrated by cyclic voltammograms obtainedwhen K6½P2W18O62� was adhered to a GC electrode and placedin contact with the IL DEAS (Fig. S1). Increased currents weredetected upon repetitive cycling of the potential (Fig. S2), con-sistent with dissolution of the more soluble reduced salt of½P2W18O62�7− formed at the electrode surface. The mechanismappears to be analogous to that proposed for a POM-modifiedelectrode in contact with [Bmim][PF6] (20) and is described byEqs. 4 and 5:

K6½P2W18O62�S þDEAþIL þ e− ⇄ ½DEA�½K�6½P2W18O62�S

↓ dissolution

½P2W18O62�6− þ e− ⇄ ½P2W18O62�7−IL

þDEAþIL þ 6 Kþ

IL; [4]

½P2W18O62�7−IL þ ne− ⇄ ½P2W18O62�−ðnþ7ÞIL: [5]

Scanning to more negative potentials generates ½P2W18O62�8−ILand more highly reduced forms [Eq. 5]. Formation of soluble re-duced anions was confirmed visually by the appearance of a bluecolor in the IL near to the electrode surface.

In order to ascertain if photoreduction of ½P2W18O62�6− couldoccur at the thermodynamically unstable water–DEAS interface,photochemical experiments were initially undertaken on solu-tions contained in a soda glass vial with visual monitoring ofsolution color as a function of time (Fig. 1). In a previous study,it was shown that photoreduction of ½P2W18O62�6− does not occurin either bulk water or bulk DEAS (8). In the present two-phaseconfiguration, the upper phase was water, in which K6½P2W18O62�is highly soluble, and the lower phase was DEAS, in which the saltis essentially insoluble. Even though DEAS is highly miscible inwater, the dissolution process (in the absence of convection) isslow, allowing a two-phase system to exist for more than 10 h.

However, after sufficient time has elapsed, the upper phase canbe described as water-rich and the lower one as DEAS-rich.

To achieve the thermodynamically unstable two-phase condi-tion, an aqueous solution containing ½P2W18O62�6− was addedslowly and dropwise onto the top of the DEAS phase in the sodaglass vial using a Pasteur pipette whose tip rested against the edgeof the vial, as shown in Fig. 2A. This approach minimized inter-mixing of the two solvents by convection: Miscibility was achievedonly after standing for more than 10 h (predominantly by diffu-sion, a very slow process). Water has the lower viscosity (approxi-mately 1 vs: > 4;356 mPa s) and density (1.00 vs. 1.21 g cm−3)(25–27). Thus, careful dropwise addition of the aqueous solutiononto the IL surface was preferred to the alternative of additionof the IL onto water. The addition of water (5 mL) to DEAS(2 mL), in the manner described, leads to the formation oftwo phases, with the bottom one being the denser IL phase(Fig. 1A). Prior to irradiation with white light, cylic voltammetricmonitoring of the ½P2W18O62�6−∕7− process (Fig. 3) confirmedthat oxidized ½P2W18O62�6− remained present in the aqueousphase and that its concentration in the DEAS phase was negli-gible (below the detection limit of 10 μM). Note that electrolyteKCl (0.1 M) was also present in the aqueous phase in experimentsinvolving voltammetric monitoring of the aqueous phase. A singlephase was formed rapidly when the two-phase system was shakenvigorously. It follows that partial mixing of the two neat liquidsmust occur in the interfacial region (Fig. 1) in a time-dependentmanner.

In photochemical experiments, the initial water–DEAS two-phase system was irradiated from the side with white light(λ ¼ 275–750 nm). The light source was located at a distanceof 3 cm from the vial in a manner that ensured that both phaseswere exposed to the same light intensity (Fig. 2A). No change intemperature (�0.1 °C) was detected in either phase during thecourse of this experiment. Very similar results were obtainedusing UV light (275–320 nm) because the ½P2W18O62�6− aniononly absorbs weakly in the visible wavelength region.

Initially, both phases were colorless. Five to 10 min of irradia-tion with white light generated blue coloration solely in thewater–DEAS interfacial region (Fig. 1 B and C). The blue coloris characteristic of reduced ½P2W18O62�n− (n > 6). After 10 min,the color diffused slowly into the bulk water phase but not intothe DEAS phase (Fig. 1C). Previously, photoreduction of½P2W18O62�6− was shown to occur in aqueous media containingDEAS concentrations ≥ 0.08 M (8). It is likely that these condi-tions are present in the interfacial region in the present experi-ment. The time required for blue material to be detected inthe aqueous phase decreased with increasing ½P2W18O62�6− con-centration (0.1 to 3 mM), consistent with the extent of photore-duction being dependent on the polyoxometalate anion concen-tration. After 20 min of irradiation, some blue material was seento be transported into the DEAS phase (Fig. 1D). Continuedphotoreduction led to further diffusion of color into the IL phaseuntil, after about 1 h, this phase was uniformly blue (Fig. 1E).

Fig. 1. Photo-irradiation of a thermodynamically unstable two-phase water–DEAS system with a white light (275–750 nm) source located at a distance of3 cm from the side of the soda glass vial (see Fig. 2). The initial aqueous phase contained K6½P2W18O62� (2.0 mM) and KCl (0.1M). Irradiation time: (A) none; (B) 5;(C) 10; (D) 20; (E) 60 min.

Bernardini et al. PNAS ∣ July 17, 2012 ∣ vol. 109 ∣ no. 29 ∣ 11553

CHEM

ISTR

YSP

ECIALFEAT

URE

Dow

nloa

ded

by g

uest

on

Apr

il 24

, 202

0

Steady-state microelectrode voltammetric measurements onthe ½P2W18O62�6−∕7− couple were undertaken in the aqueousphase after 20 min of irradiation by using the photoelectrochem-ical cell shown in Fig. 2B. Results are displayed in Fig. 4. Close tothe interface, the anion was found to be at the one electron-re-duced redox level ½P2W18O62�7−. In the bulk phase, ½P2W18O62�7−was still the major component but with a significant concentrationof ½P2W18O62�6− still present. The magnitude of the limiting cur-rent was higher in the bulk aqueous solution than in the region

just above the interface (Fig. 4). This reflects the higher concen-tration of viscous DEAS in the interfacial region and correspond-ingly lower diffusion coefficient values. The diffusion coefficientfor ½P2W18O62�6− in water–DEAS mixed solvent media is a func-tion of DEAS concentration (Table S1).

The mass transport process is driven by the differential solu-bilities and distribution coefficients of the reactant and product aswell as their diffusion coefficients in the water and IL phases.K6½P2W18O62� is highly soluble in water but not in the IL; thereduced material is more soluble in the IL than in water. Theoverall mass transport observations are rationalized by the factthat the diffusion coefficient of ½P2W18O62�7− in the IL is twoorders of magnitude smaller than in molecular solvents (i.e.,the product anion diffused into the aqueous phase prior to itsextraction into the IL phase). Eventual transfer of ½P2W18O62�7−to the IL layer was confirmed by cyclic voltammetric experi-ments (Fig. 5).

The production of dioxygen via Eq. 2 in the upper aqueousphase in the water–DEAS system was monitored by a Clark-typeelectrode located 1 cm above the interface. The aqueous phasecontaining ½P2W18O62�6− was initially degassed for 10 min withdinitrogen to minimize the dioxygen concentration. This solutionwas then added to dinitrogen-degassed DEAS using the proce-dure described in Fig. 2A. The vial was then sealed from the at-mosphere (8). No significant change in [O2] from the backgroundlevel of about 120 μM was apparent initially but a significant in-crease occurred after about 6 min of irradiation with white light(Fig. 6). The delay is attributed to the time required for diffusionof O2 to the Clark electrode. After prolonged photolysis, theoxygen concentration reached a maximum value and then de-creased slowly (Fig. 6). ½P2W18O62�7− reacts very slowly withO2 (28) to regenerate ½P2W18O62�6−. Clearly, if this reaction werefast, ½P2W18O62�7− would not be detected. More highly reducedanions react rapidly with O2 so ½P2W18O62�− is the “sink” (28).

Two-phase irradiation experiments were also carried out underthe conditions described above but with [Bmim][PF6] as the IL.Water is only sparingly soluble in this aprotic IL and the systemnow represents a thermodynamically stable two-phase scenario.However, photoreduction again occurred preferentially at theinterface with initial diffusion of ½P2W18O62�7− into the aqueousphase followed by subsequent extraction into the IL phase.

Fig. 2. Schematic representation of the cell arrangement for photochemicalexperiments at a water–IL interface. (A) Dropwise addition of the aqueousphase onto the IL phase; (B) in situ voltammetric monitoring.

Fig. 3. Cyclic voltammograms (ν ¼ 0.15 V s−1) obtained at a GC electrodebefore irradiation with white light in (A) the aqueous phase (K6½P2W18O62�,0.3 mM; KCl, 0.1 M) and (B) the DEAS phase of the thermodynamicallyunstable two-phase system.

Fig. 4. Steady-state voltammograms (GC microelectrode; d ¼ 15 μm);ν ¼ 0.002 V s−1) obtained in the aqueous phase (K6½P2W18O62�, 0.3 mM; KCl,0.1 M) after 20 min of irradiation of the water–DEAS two-phase system.Voltammograms obtained just above the interface (‐‐‐) and in the bulk region(—) of the aqueous solution phase.

11554 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1203818109 Bernardini et al.

Dow

nloa

ded

by g

uest

on

Apr

il 24

, 202

0

In related experiments, two-phase systems were formed withacetonitrile (9% water) as the phase containing dissolvedK6½P2W18O62�, ½Bu4N�4½S2W18O62� or ½Bu4N�4½S2Mo18O62� and0.1 M Bu4NPF6 as the supporting electrolyte. The IL phasestested were DEAS, DEAP, [Bmim][BF4], or [Bmim][PF6]. Inall cases, photoreduction of the polyoxometalate anion occurredpreferentially at the interface to produce ½P2W18O62�7−,½S2Mo18O62�5−, or ½S2W18O62�5− followed by diffusion into thewater-acetonitrile phase (0.1 M Bu4NPF6, voltammetric detec-tion). Again, at longer times, the reduced anions were trans-ported into the IL phase. ½S2Mo18O62�5− is further reduced to½S2Mo18O62�6− on longer light exposure times. Unlike the otherPOM anions, ½S2Mo18O62�4− is photochemically reduced veryslowly to ½S2Mo18O62�5− in wet acetonitrile, confirming that the

role of reversible potential also is significant (8, 28). In generalterms the overall reaction is represented by Eq. 6:

4POMþ 2H2Oþ hν → 4POM− þ 4Hþ þO2: [6]

The significance of the interfacial region was supported by dataobtained in analogous experiments involving a thermodynami-cally stable (immiscible) aqueous-organic solvent two-phasesystem. A biphasic system consisting of an aqueous solution ofK6½P2W18O62� (0.3 mM; 0.1 M KCl) in contact with diethyl etherphase was examined; on this occasion the diethyl ether phase ison top (ρ ¼ 0.7134 vs. 1.0000 g cm−3 for water) (26). Irradiationwith white light for 10 min gave the outcome shown in Fig. 7.Steady-state voltammograms were obtained with a GCmicroelec-trode placed at three different locations (A–C) in the aqueousphase (Fig. 8). All voltammograms exhibit similar total limitingcurrents but reveal differences in the redox level in the threezones; the oxidation current is higher for the locations closerto the interface, indicating increased proportions of reducedanion ½P2W18O62�7−. No extraction into the ether phase occurredin this system as the POM salts are insoluble in ether, irrespectiveof redox level.

In addition, a decrease of pH from 6.35 to 3.90 occurred inregion A during the experiment shown in Fig. 7. This is consistentwith production of protons via Eq. 3. Dioxygen evolution was de-tected by a Clark-type electrode experiment. A control reactioninvolving irradiation of a suspension of K6½P2W18O62� in diethy-lether did not change color over 2 h. Photooxidation does notoccur for the salt dissolved in neat water.

It is apparent that the interface between the liquid phases playsa key role in photoreduction of the POM anion. This was againapparent in the experiment of Fig. S3 where reduced anion canbe detected at the water–diethylether interface after irradiationwith sunlight for 3 min. Longer exposure times led to diffusionof reduced product into the aqueous phase. Equivalent changesoccurred in the water–DEAS system upon exposure to sunlight;as with white light, diffusion into the aqueous phase occurred in-itially, then, after 4 h, the products extracted into the IL phase.

Photoreduction was also observed when solid polyoxometalatesalts were in contact with water vapor. Soda glass vials containingpowdered salts were used for these experiments. A micro dropletof water was adhered to the vial edge and immediately sealedfrom the atmosphere. It was placed on a hot plate so that eva-poration of water occurred followed, upon cooling, by uniformcondensation over the solid. Controls lacking the droplet andsubjected to high vacuum were treated equivalently. The vialswere exposed to sunlight (23–31 °C for 12 h per day) and thecolor changes monitored over a two-week period. Surface color

Fig. 5. Cyclic voltammograms obtained at a GC working electrode(ν ¼ 0.15 V s−1) in photo-irradiation experiments of a water–DEAS two-phasesystems initially containing K6½P2W18O62� (0.2 mM) in the aqueous phase. Theelectrode was located in the DEAS ionic liquid phase (see Fig. 2B). (A) Beforeirradiation with white light; (B) after 2 h irradiation and (C) after 4 h irradia-tion. The increase in faradic current is primarily due to the transfer of photo-generated ½P2W18O62�7− into the IL phase.

Fig. 6. Concentration of dioxygen in the aqueous phase in a sealed glassvial. Dioxygen was detected by a Clark-type electrode as a function of irra-diation time during photoilumination with white light of the two-phasewater–DEAS system. K6½P2W18O62� (0.3 mM) and KCl (0.1 M) was presentin the initial aqueous phase.

Fig. 7. Two phases water-diethylether system after irradiation with whitelight for 10 min. The ether phase is at the top with the aqueous phase atthe bottom. K6½P2W18O62� (0.3 mM) and KCl (0.1 M) was present in the initialaqueous phase. Steady-state voltammograms were obtained at the indicatedpositions (A, B, C) in order to evaluate the redox level of the polyoxometalateafter irradiation (see Fig. 8).

Bernardini et al. PNAS ∣ July 17, 2012 ∣ vol. 109 ∣ no. 29 ∣ 11555

CHEM

ISTR

YSP

ECIALFEAT

URE

Dow

nloa

ded

by g

uest

on

Apr

il 24

, 202

0

changes to green/blue were seen after a few hours forðBu4NÞ4½S2Mo18O62� and after a few days for K6½P2W18O62�. Thecontrols exhibited no change in color. The observations are con-sistent with photoreduction of the salts and oxidation of water.Similar experiments under 2-propanol vapour produced colorchanges within 1 h.

This work demonstrated that photoreduction of ½P2W18O62�6−,½S2Mo18O62�4−, and ½S2W18O62�4− anions occurred at the inter-faces of two-phase systems under conditions in which water actsas the electron donor. It is proposed that photo-oxidation ofwater is facilitated at an interface where the structure of wateris different from that in neat H2O or when water is dissolvedin a molecular organic solvent. In particular, hydrogen bondingis expected to be weaker when water is present in the gas phase orat an interface. While the electrochemical methodology used in

this study provides qualitative information on the overall reac-tion, quantitative photochemical data and the photophysics ofthe reactions are yet to be established. Thus, while Eq. 6 is writtenas an overall one-electron reduction step with formation of finalproduct POM−, it is probable that multi-electron reduced anionsare produced as intermediate photo products arising directlyfrom oxidation of water. They would then be expected to reactwith oxidized anion, Hþ, or O2 to generate POM− as the de-tected final product (8).

Quantum yields, ground and excited state potentials, plus theactivity of Hþ at the interface are needed to quantify the photo-chemistry. The present study raises a series of important photo-chemical, photophysical, and interfacial questions that are ingeneral beyond the scope of electrochemical methodology. Itis relevant that studies in acid-base behavior and catalysis inemulsions have also revealed the role that differences in structureplay in the chemistry of water at interfaces (29–32).

Materials and MethodsReagents. K6½α-P2W18O62�14H2O, ½Bu4N�4½α-S2W18O62�, and½Bu4N�4½α-S2Mo18O62� were synthesized according to literature procedures(28, 33–35). The aprotic ILs [Bmim][BF4] and [Bmim][PF6] (Merck, high puritygrade, ≥99.0%) were used as received. The protic ionic liquids DEAS andDEAP were synthesized and purified according to published procedures(36). Structures of all ILs are provided in Fig. S1.

Electrochemistry. In most cases, the POM salt was dissolved in the solventof interest. Voltammetric experiments in such single-phase solutions wereconducted at ð20� 1Þ °C according to the procedure described before (8).For electrochemistry in two-phase environments, a specially designed electro-chemical cell arrangement was developed for voltammetric monitoring ofreactions taking place upon irradiation with light in a two-phase water/ILsystem (Fig. 2B). Detailed electrochemical precedures are present in SI Text.

Photochemistry. Both phases and the interfacial regions of aqueous–IL oraqueous–diethylether systems were irradiated with white light (275–750 nm)using a Rofin Polilight PL6 Xe source placed 3 cm from the size wall of thesoda glass vial (see Fig. 2A). The oxygen concentration and pH values in aqu-eous media were estimated with a Clark-type gas sensor and a pH Meter.Further details of the light source, Clark sensor, and pH Meter are providedin ref. 8.

ACKNOWLEDGMENTS. The authors gratefully acknowledge financial supportfrom the Australian Research Council.

1. Argitis P, Papaconstantinou E (1985) Photocatalytic multielectron photoreductionof 18-tungstodiphosphate in the presence of organic compounds—production ofhydrogen. J Photochem 30:445–451.

2. WardMD, Brazdil JF, Grasselli RK (1984) Photocatalytic alcohol dehydrogenation usingammonium heptamolybdate. J Phys Chem 88:4210–4213.

3. Hill CL, Bouchard DA (1985) Catalytic photochemical dehydrogenation of organicsubstrates by polyoxometalates. J Am Chem Soc 107:5148–5157.

4. Papaconstantinou E, Dimotikali D, Politou A (1980) Photochemistry of heteropolyelectrolytes. The 18-molybdodiphosphate. Inorg Chim Acta 43:155–158.

5. Yamase T, Takabayashi N, Kaji M (1984) Solution photochemistry of tetrakis(tetrabu-tylammonium) decatungstate(VI) and catalytic hydrogen evolution from alcohols.J Chem Soc, Dalton Trans 793–799.

6. Papaconstantinou E (1989) Photochemistry of polyoxometallates of molybdenum andtungsten and/or vanadium. Chem Soc Rev 18:1–31.

7. Zhao C, Bond AM (2009) Photoinduced oxidation of water to oxygen in the ionic liquidBMIMBF4 as the counter reaction in the fabrication of exceptionally long semicon-ducting silver-Tetracyanoquinodimethane nanowires. J Am Chem Soc 131:4279–4287.

8. Bernardini G, Zhao C, Wedd AG, Bond AM (2011) Ionic liquid-enhanced photooxida-tion of water using the polyoxometalate anion ½P2W18O62�6− as the sensitizer. InorgChem 50:5899–5909.

9. Menjoge A, Dixon J, Brennecke JF, Maginn EJ, Vasenkov S (2009) Influence of water ondiffusion in imidazolium-based ionic liquids: A pulsed field gradient NMR study. J PhysChem B 113:6353–6359.

10. Raju SG, Balasubramanian S (2009) Aqueous solution of [bmim][PF6]: Ion and solventeffects on structure and dynamics. J Phys Chem B 113:4799–4806.

11. Stark A, MacLean BL, Singer RD (1999) 1-Ethyl-3-methylimidazolium halogenoalumi-nate ionic liquids as solvents for Friedel-Crafts acylation reactions of ferrocene. J ChemSoc, Dalton Trans 63–66.

12. Earle MJ, Seddon KR, Adams CJ, Roberts G (1998) Friedel-Crafts reactions in room tem-perature ionic liquids. Chem Commun 2097–2098.

13. Fischer T, Sethi A, Welton T, Woolf J (1999) Diels-Alder reactions in room-temperatureionic liquids. Tetrahedron Lett 40:793–796.

14. Earle MJ, McCormac PB, Seddon KR (1998) Regioselective alkylation in ionic liquids.Chem Commun 2245–2246.

15. Monteiro AL, Zinn FK, de Souza RF, Dupont J (1997) Asymmetric hydrogenation of2-arylacrylic acids catalyzed by immobilized Ru-BINAP complex in 1-n-butyl-3-methy-limidazolium tetrafluoroborate molten salt. Tetrahedron: Asymmetry 8:177–179.

16. Simon LC, Dupont J, de Souza RF (1998) Two-phase n-butenes dimerization by nickelcomplexes in molten salt media. Appl Catal A: Gen 175:215–220.

17. Dullius JEL, et al. (1998) Selective catalytic hydrodimerization of 1,3-butadiene bypalladium compounds dissolved in ionic liquids. Organometallics 17:815–819.

18. Chauvin Y, Olivier H, Wyrvalski CN, Simon LC, de Souza RF (1997) Oligomerization ofn-Butenes catalyzed by nickel complexes dissolved in organochloroaluminate ionicliquids. J Catal 165:275–278.

19. Kavarnos GJ, Turro NJ (1986) Photosensitization by reversible electron transfer:Theories, experimental evidence, and examples. Chem Rev 86:401–449.

20. Zhang J, et al. (2005) Voltammetric Studies on the reduction of polyoxometalateanions in ionic liquids. Inorg Chem 44:5123–5132.

21. Bond AM (2002) Broadening Electrochemical Horizons (Oxford Univ Press, Oxford, UK).22. Scholz F, Meyer B (1998) Electroanalytical Chemistry (Marcel Dekker, New York).23. Zhang J, Bond AM (2005) Practical considerations associated with voltammetric studies

in room temperature ionic liquids. Analyst 130:1132–1147.24. Zhang J, Bond AM, Richardt PJS, Wedd AG (2004) Voltammetric reduction of α- and

γ�-½S2W18O62�4− and α-, β-, and γ-½SiW12O40�4− : Isomeric dependence of reversiblepotentials of polyoxometalate anions using data obtained by novel dissolution andconventional solution-phase processes. Inorg Chem 43:8263–8271.

25. Sawyer DT, Sobkowiak A, Roberts JL (1995) Electrochemistry for Chemists (Wiley, NewYork), 2nd Ed.

26. Berstad DA, et al. (1988) Accurate measurements of the viscosity of water in thetemperature range 19.5–25.5 °C. Physica A 151:246–280.

Fig. 8. Steady-state voltammograms (GC microelectrode: d ¼ 15 μm;ν ¼ 0.002 V s−1) probing the redox level of ½P2W18O62�6−∕7− in the water-diethylether system after irradiation with white light for 10 min at threelocations in the aqueous phase (see positions A–C in Fig. 7). K6½P2W18O62�(0.3 mM) and KCl (0.1 M) was present in the initial aqueous phase.

11556 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1203818109 Bernardini et al.

Dow

nloa

ded

by g

uest

on

Apr

il 24

, 202

0

27. Lide DR (1990) CRC Handbook of Chemistry and Physics (CRC Press, Boca Raton, FL),70th Ed.

28. Bernardini G, Wedd AG, Bond AM (2011) Reactivity of one-, two-, three- and four-elec-tron reduced forms of α-½P2W18O62�6− generated by controlled potential electrolysis inwater. Inorg Chim Acta 374:327–333.

29. Cooper JB, Way DM, Bond AM, Wedd AG (1993) A green heteropoly blue: Isolation ofa stable, odd oxidation level in a Dawson molybdate anion, ½S2Mo18O62�5− . InorgChem 32:2416–2420.

30. Narayan S, et al. (2005) “On Water”: Unique reactivity of organic compounds in aqu-eous suspension. Angew Chem Int Ed Engl 117:3219–3219.

31. Gray-Weale A, Beattie JK (2009) An explanation for the charge onwater’s surface. PhysChem Chem Phys 11:10994–11005.

32. Beattie JK, McErlean CSP, Phippen CBW (2010) The mechanism of on-water catalysis.Chem Eur J 16:8972–8974.

33. Mbomekalle I-M, Lu YW, Keita B, Nadjo L (2004) Simple, high yield and reagent-savingsynthesis of pure α-K6P2W18O62•14H2O. Inorg Chem Commun 7:86–90.

34. Graham CR, Finke RG (2008) The classic Wells–Dawson polyoxometalate,K6½α-P2W18O62�•14H2O Answering an 88 year-old question: What is its preferred,optimum synthesis? Inorg Chem 47:3679–3686.

35. Richardt PJS, Gable RW, Bond AM, Wedd AG (2001) Synthesis and redox characteriza-tion of the polyoxo anion, γ�-½S2W18O62�4− : A unique fast oxidation pathway deter-mines the characteristic reversible electrochemical behavior of polyoxometalateanions in acidic media. Inorg Chem 40:703–709.

36. Zhao C, et al. (2008) Electrochemistry of room temperature protic ionic liquids. J PhysChem B 112:6923–6936.

Bernardini et al. PNAS ∣ July 17, 2012 ∣ vol. 109 ∣ no. 29 ∣ 11557

CHEM

ISTR

YSP

ECIALFEAT

URE

Dow

nloa

ded

by g

uest

on

Apr

il 24

, 202

0