Solar energy conversion using photochemical molecular devices: photocatalytic hydrogen production from water using mixed-metal supramolecular complexes

Abstract

Photocatalytic generation of hydrogen from water is an integral part of the next generation clean fuel technologies. The conversion of solar energy into useful chemical energy is of great interest in contemporary investigations. The splitting of water is a multi-electron process involving the breaking and making of chemical bonds which requires multi-component photocatalytic systems. Supramolecular complexes [{(TL)₂Ru(BL)}₂RhX₂](Y)₅ (where TL = terminal ligand, BL = bridging ligand, X = Cl⁻ or Br⁻, and Y = PF₆⁻ or Br⁻) have been synthesized and studied for their light absorbing, electrochemical and photocatalytic properties. The supramolecular complexes in this investigation are multi-component systems comprised of two ruthenium based light absorbers connected through bridging ligands to a central rhodium, which acts as an electron collecting center upon excitation. These complexes absorb light throughout the ultraviolet and visible regions of the solar spectrum. The supramolecular complexes possess ruthenium based highest occupied molecular orbitals (HOMO) and a rhodium based lowest unoccupied molecular orbital (LUMO). These molecular devices have been investigated and shown to function as photoinitiated electron collectors at the reactive rhodium metal center, and explored as photocatalysts to generate hydrogen from water in an aqueous solution in the presence of an electron donor.

---

Broader context

The work described herein provides basic chemical studies of the process of light to energy conversion via the collection of reducing equivalents in a molecular system. This system further delivers the reducing equivalents to a water substrate to produce hydrogen. The harvesting of solar energy efficiently by molecular systems is largely restricted by the lack of a fundamental understanding of multielectron chemistry. It is further severely restricted by the lack of systems that undergo multielectron photochemistry. Systems that harvest visible light efficiently typically undergo single electron reactions which is energetically prohibitive for solar energy water splitting. The first visible light induced photoinitiated electron collection and hydrogen generation from water by polyazine derived Ru, Rh, Ru heteronuclear water soluble system is described. This article places this unique chemistry in the context of the very few, typically organic soluble systems even known to under photoinitiated electron collection. The reported system is unusual in its ability to use low energy visible light to drive multielectron photochemistry, ability to function in water, and further ability to deliver the photocollected electrons to a substrate leading to hydrogen production from water.

Introduction

The need for renewable energy sources has driven an exploration for alternative energy. The present consumption rate of fossil fuels makes it necessary to explore alternative energy sources. One appealing alternative energy source is solar energy. The harvesting of solar energy can occur through the conversion of light energy into heat energy in thermal conversion schemes, or electrical potentials in solar cells, or fuels through light to chemical energy conversion schemes. The conversion of light energy into a fuel focuses on the conversion of a widely available chemical feedstock into a transportable fuel. The chemical feedstocks commonly suggested are carbon dioxide (via artificial photosynthesis) and water (via solar water splitting). Water is an attactive chemical feedstock for solar energy conversion as the fuel produced is hydrogen. Molecular hydrogen is an attractive nonpolluting clean burning fuel with a combustion energy of −282 kJ mol⁻¹ and with potential as a long-term replacement for natural gas.¹

The solar spectrum at the earth's surface consists of the infrared, visible and ultraviolet regions of the electromagnetic spectrum. Solar water splitting is a process by which light energy from the sun is harvested to convert water into hydrogen and oxygen.^2–4 The splitting of water is an energetically uphill process which requires 1.23 V of energy.⁴ Much of the visible and ultraviolet regions of the solar spectrum, that reaches the earth from the sun, has sufficient energy to drive the splitting of water into hydrogen and oxygen. Yet, this process does not happen by direct irradiation with sunlight because of the transparent nature of the water. One molecule of water yields, in a balanced chemical equation, one equivalent of hydrogen gas and a half equivalent of oxygen gas. The overall reaction for water splitting and the related energetics of the half reactions are represented below, where E is the redox potential vs. NHE at pH = 7.

2H⁺_(aq) + 2e⁻ → H_2(g)  E = −0.41 V

H₂O_(l) → ½O_2(g) + 2e⁻ + 2H⁺_(aq)  E = −0.82 V

H₂O_(l) → H_2(g) + ½O_2(g)  E = −1.23 V

The energy requirements for water splitting through a multielectron process is much less (1.23 V)^4–6 relative to water splitting through a single electron process (∼ 5 V). Fujishima and Honda showed the light activated water splitting reaction over UV light irradiated TiO₂ semiconductors.⁷

The discovery of [Ru(bpy)₃]²⁺ (bpy = 2,2′-bipyridine) has inspired photochemical and photophysical studies of similar polyazine complexes for light to energy conversion research.^8–11Ru(II)-polyazine chromophores absorb light with a high efficiency throughout the UV and much of the visible regions, and possess long-lived metal-to-ligand charge-transfer (MLCT) excited states. The photochemical and photophysical properties of ruthenium polyazine complexes are easily tuned by the selection of the ligand set bound to the metal center. The prototypical [Ru(bpy)₃]²⁺ possesses intense intra-ligand (IL) transitions in the UV with MLCT transitions in the visible, Fig. 1.

Fig. 1 Electronic absorption (solid line) and emission spectra (dotted line) of [Ru(bpy)₃]²⁺ (bpy = 2,2′-bipyridine) in acetonitrile at room temperature.

The intense Ru(dπ) → bpy(π*) MLCT transition occurs at 450 nm in acetonitrile solution for [Ru(bpy)₃]²⁺. Excitation of a Ru polyazine type of light absorber (LA) leads to intersystem crossing and internal conversion, which results in population of the lowest lying triplet metal-to-ligand charge-transfer (³MLCT) excited state with unit efficiency, Fig. 2. This ³MLCT excited state displays a strong emission in deoxygenated acetonitrile solution centered at 630 nm with an excited-state lifetime of 860 ns at room temperature.¹² The ³MLCT state of ruthenium polyazine complexes is typically long-lived and can undergo excited state energy and electron transfer reactions. As the excited state of [Ru(bpy)₃]²⁺ is both a more powerful oxidizing and reducing agent than its ground state, it is capable of undergoing either oxidative or reductive electron transfer quenching.^10,12–16

Fig. 2 Energy state diagram for [Ru(bpy)₃]²⁺, (bpy = 2,2′-bipyridine, GS = ground state, ¹MLCT = singlet metal-to-ligand charge-transfer state, ³MLCT = triplet metal-to-ligand charge-transfer state, k_isc = rate constant for intersystem crossing, k_r = rate constant for radiative decay, k_nr = rate constant for non-radiative decay, k_rxn = rate constant for reaction).

The complex [Ru(bpy)₃]²⁺ and other light absorbing analogs possess ³MLCT states of sufficient energy to split water into hydrogen and oxygen, yet direct photocatalysis has not been observed. Water splitting is a multistep and multielectron process involving breaking and making of chemical bonds. Water splitting requires multicomponent systems containing proper catalytically active sites in addition to adequate energy input. Complicated multicomponent systems incorporating light absorbing units, electron relays, and redox catalysts have been extensively utilized in solar energy conversion schemes.

Nocera et al. have made a significant contribution in the area of photochemical hydrogen production. They have utilized metal–metal bonded dirhodium photocatalysts, such as [Rh₂(dfpma)₃(PPh₃)CO] (dfpma = MeN(PF₂)₂), which undergo photochemical transformations and catalytically produce hydrogen from hydrohalic acids in the presence of a halogen trap.^1,2,17–20

Multicomponent systems using polyazine light absorbers have been studied and shown to photocatalytically produce H₂. Lehn et al.^5,6 and Sutin et al.^21,22 have studied a four component system for hydrogen production using the well studied [Ru(bpy)₃]²⁺ light absorber (LA), a [Rh(bpy)₃]³⁺ electron relay, triethanolamine (TEOA) as a sacrificial electron donor (ED), and a colloidal Pt catalyst system. Photolysis results in the transfer of electrons from the excited state of ^*[Ru(bpy)₃]²⁺ to [Rh(bpy)₃]³⁺. The newly generated [Rh^II(bpy)₃]²⁺ is proposed to transfer electrons to colloidal Pt, which produces hydrogen in the presence of water. Oishi²³ and Bauer and Werner²⁴ studied the Wilkinson's catalyst analog [Rh^ICl(dpm)₃]³⁻ (dpm = diphenylphosphino-benzene-m-sulfonate), which when photolyzed in the presence of [Ru(bpy)₃]²⁺ and ascorbic acid produces hydrogen. Eisenberg et al. reported the photocatalytic hydrogen production activity of a system containing a series of luminescent platinum(II) terpyridyl acetylide complexes in the presence of dialkylated bipyridinium cations, an electron donor (TEOA), and colloidal Pt catalyst.^25,26

Developing suitable molecules that can absorb the solar spectrum and utilize the absorbed light energy for a productive purpose is a challenge and involves the art of molecular design and development. Photocatalytic systems for fuel production typically combine (i) a chromophore or LA, (ii) an electron collector (EC), and (iii) a catalytic site (CAT).⁹ Molecular devices applied in this forum should be able to perform multiple tasks including: the absorption of light over a broad region of the electromagnetic spectrum, storage of the absorbed energy for an appropriate length of time, transferring electrons to an active center, and catalyzing a suitable reaction to carry out the chemical transformation. A schematic orbital energy diagram for a photochemical molecular device for photointiated electron collection is represented in Fig. 3. Supramolecular assemblies which couple an ED to a LA and EC represent photochemical molecular devices for electron collection.

Fig. 3 Schematic representation of an orbital energy diagram showing the functioning of a molecular device for photoinitiated electron collection (LA = light absorber, BL = bridging ligand, EC = electron collector, ED = electron donor).

Photochemical electron collection in molecular complexes employ chromophores connected, through proper bridging ligands (BL), to an electron collector. Utilization of solar energy typically proceeds through MLCT excitation of the chromophore. The excited electron is transferred and held by a suitable EC. This process is repeated, providing two electrons collected at the EC site. Bridging ligands provide a point of attachment between the light absorbing chromophores and the electron collector. A variety of TL and BL can be designed to tune the properties and functioning of the molecular devices. Some of the representative TL and BL are given in Fig. 4.

Fig. 4 Chemical structures of some of the representative terminal ligands (TL) and bridging ligands (BL), where bpy = 2,2′-bipyridine, phen = 1,10-phenanthroline, Ph₂phen = 4,7-diphenylphenanthroline, tpy = 2,2′:6′,2″-terpyridine, dpp = 2,3-bis(2-pyridyl)pyrazine, dpq = 2,3-bis(2-pyridyl)quinoxaline, bpm = 2,2′-bipyrimidine, tppz = 2,3,5,6-tetrakis(2-pyridyl)pyrazine, and tatpp = 9,11,20,22-tetraazatetrapyrido[3,2-a:2′,3′-c:3″,2″-1:2‴,3‴-n]pentacene.

Despite the increasing interest in photochemical molecular devices for photoinitiated electron collection, working systems that store multiple electrons remain a challenge. The first functioning photoinitiated electron collector in a molecular system was reported by Brewer et al.²⁷ The supramolecular assembly couples two ruthenium LA units to a central iridium that functions to electronically isolate the two LA subunits, Fig. 5. The [{(bpy)₂Ru(dpb)}₂IrCl₂](PF₆)₅ (dpb = 2,3-bis(2-pyridyl)benzoquinoxaline) system, when excited with visible light in the presence of the electron donor, N,N-dimethylaniline (DMA), photochemically reduces by two electrons to form [{(bpy)₂Ru(dpb⁻)}₂IrCl₂]³⁺. MacDonnell et al.^28,30,31 studied the Ru–Ru bimetallic systems containing interesting extended π-conjugated bridging ligands which can collect up to four electrons during photolysis in the presence of an electron donor triethylamine (TEA), Fig. 5. Bocarsly et al.^29,32 studied a series of Pt(IV) centered trimetallic complexes of the form [(NC)₅M^II(CN)Pt^IV(NH₃)₄(NC)M^II(CN)₅]⁴⁻ (M = Fe, Ru, or Os) which undergo photoinitiated electron collection at the platinum metal center Fig. 5. The complexes, however, are not stable and disassemble to monometallic components with formation of the two electron reduced Pt(II) complex.

Fig. 5 Representative examples of photoinitiated electron collector molecular devices, [{(bpy)₂Ru(dpb)}₂IrCl₂]⁵⁺,²⁷[(phen)₂Ru(tatpp)Ru(phen)₂]⁴⁺,²⁸ and [(NC)₅Fe(CN)Pt(NH₃)₄(NC)Fe(CN)₅]⁴⁻.²⁹

Mixed-metal polyazine complexes containing ruthenium LA coupled to reactive metal sites have been explored. Sakai et al.^33–35 recently investigated a Ru–Pt bimetallic system capable of photochemically producing hydrogen from water in the presence of the electron donor, ethylenediaminetetraacetic acid (EDTA). In this system the platinum component, of the form cis-Pt^IICl₂ was linked to the ruthenium LA through an amide linkage on the phenanthroline ligand coordinated to ruthenium. This system functioned with Φ ≈ 0.01, and 5 turnovers in 10 h. This low turnover has been postulated to reflect decomplexation of the reactive metal to form colloidal platinum, which functions as the hydrogen generation site.³⁶ Rau et al.³⁷ reported a Ru–Pd bimetallic system that photochemically produces hydrogen in the presence of an electron donor, triethylamine (TEA). This system was shown to give 56 turnovers in ∼30 h. Recently, colloidal Pd formation was suggested for Ru–Pd molecular systems.³⁸

Brewer et al.^39,40 recently reported the photoinitiated electron collection at a metal center in [{(bpy)₂Ru(dpp)}₂RhCl₂](PF₆)₅ (dpp = 2,3-bis(2-pyridyl)pyrazine). When excited with visible light, this complex is photochemically reduced converting the Rh(III) to Rh(I), which undergoes subsequent chloride loss to form [{(bpy)₂Ru(dpp)}₂Rh^I]⁵⁺.³⁹ This complex photochemically reduces water to hydrogen using visible light excitation with a Φ ≈ 0.01 in the presence of the sacrificial electron donor DMA.⁴⁰ The ability of this system to undergo photoinitiated electron collection on the rhodium center with the molecular architecture remaining intact is unprecedented and allows use in multielectron photochemistry. One supramolecular complex used in our study as a photochemical molecular device is given in Fig. 6.

Fig. 6 Supramolecular complex [(bpy)₂Ru(dpp)RhBr₂(dpp)Ru(bpy)₂]⁵⁺ for photoinitiated electron collection and solar water splitting (LA = light absorber, BL = bridging ligand, and EC = electron collector). Hydrogens are omitted for clarity.

Brewer et al. developed a number of mixed-metal supramolecular complexes containing ruthenium or osmium polyazine light absorbers coupled through a variety of bridging ligands to a rhodium metal center with interesting photophysical properties, which have also shown to be efficient as light activated DNA photocleaving agents.^41–45

In this paper, we report our study of Ru,Rh,Ru molecular devices for solar energy conversion and generation of hydrogen from water in the presence of a suitable electron donor in aqueous medium.

Experimental

The supramolecular complexes [{(bpy)₂Ru(dpp)}₂RhCl₂](PF₆)₅ and [{(bpy)₂Ru(dpp)}₂RhBr₂](PF₆)₅ were prepared following literature procedures.^42,44 The water soluble complex, [{(bpy)₂Ru(dpp)}₂RhBr₂]Br₅, was prepared from [{(bpy)₂Ru(dpp)}₂RhBr₂](PF₆)₅ by metathesis with Et₄NBr in acetone. All solvents used were spectroscopic grade obtained from VWR Scientific and used without further purification. High purity DMA and TEOA were obtained from Aldrich Chemical Company and used without further purification.

Electronic absorption spectra were collected using a Hewlett Packard 8452A or Agilent 8453 diode array spectrophotometer. Emission spectra were recorded using a QuantaMaster Model QM-200-45E Fluorimeter from Photon Technology Inc. Cyclic voltammetry and constant potential electrolysis experiments were conducted with a BioAnalytical Systems (BAS) electrochemical analyzer. Potentials were referenced to a Ag/AgCl reference electrode (0.197 V vs. NHE). The supporting electrolytes were 0.1 M Bu₄NPF₆ in acetonitrile, or 0.1 M KCl in water.

The photocatalytic hydrogen production experiments were performed under argon atmosphere as follows. For the photolysis reaction in acetonitrile, the septa capped deoxygenated photolysis reaction cells contained 4.5 mL final volume of the reaction solution containing the metal complex (65 µM, [{(bpy)₂Ru(dpp)}₂RhCl₂](PF₆)₅, or [{(bpy)₂Ru(dpp)}₂RhBr₂](PF₆)₅), electron donor (1.5 M) and water (0.62 M, acidified to pH = 2 with triflic acid, final [CF₃SO₃⁻] = 110 µM). The solutions were photolyzed using 470 nm LED light array constructed locally with flat optical bottom cells.⁴⁵ For the photolysis reaction in water, the septa capped deoxygenated photolysis reaction cells contained 4.5 mL final volume of the reaction solution containing the metal complex (65 µM, [{(bpy)₂Ru(dpp)}₂RhBr₂]Br₅) and electron donor (1.5 M, neat or adjusted to pH = 7.9 with triflic acid, HBr or H₃PO₄). The solutions were photolyzed using 470 nm LED light. DMA or TEOA were used as the electron donors. After 4 h photolysis time, the hydrogen was quantitated by headspace analysis using gas chromatography. Samples of the headspace (100 µL) were injected into a series 580 GOW-MAC gas chromatograph equipped with a 5 Å molecular sieves column using ultra-high purity argon as the carrier gas, purchased from Airgas Inc. (Radnor, PA), equipped with a rhenium tungsten thermal conductivity detector. The signal was amplified with a Vernier Software instrument amplifier and recorded with Logger Pro 3.4.5 software. The system was calibrated for hydrogen signal sensitivity by hydrogen standard measurements. The total amount of hydrogen produced in a photolysis experiment was obtained by the sum of the hydrogen found in the headspace as well as the hydrogen in the solution. The amount of hydrogen in the solution was calculated according to Henry's Law using the reported solubility of hydrogen in solvents.^46,47 Each photocatalysis experiment was conducted three times with the reported µmol hydrogen being the average of the trials.

Results and discussion

Assembling supramolecular light harvesting devices

The supramolecular solar energy harvesting photocatalysts under study are multicomponent systems consisting of ruthenium based LAs, an EC rhodium reactive center, and BLs connecting the LA and EC. The multimetallic complexes [{(bpy)₂Ru(dpp)}₂RhCl₂](PF₆)₅, [{(bpy)₂Ru(dpp)}₂RhBr₂](PF₆)₅ and [{(bpy)₂Ru(dpp)}₂RhBr₂]Br₅ were prepared in good yields using our building-block approach^40,44 and counter anion metathesis. The synthetic strategies are depicted in Fig. 7. Initially, the terminal ligand bpy was reacted with ruthenium(III) chloride to form the mononuclear complex [(bpy)₂RuCl₂].⁴⁸ The bridging ligand dpp was coordinated to the ruthenium light absorbing center by combining with [(bpy)₂RuCl₂] followed by addition of excess NH₄PF₆ salt to form [(bpy)₂Ru(dpp)](PF₆)₂.⁴⁹ The [(bpy)₂Ru(dpp)](PF₆)₂ was assembled into the mixed-metal trimetallic complexes [{(bpy)₂Ru(dpp)}₂RhX₂](PF₆)₅ (where X = Cl⁻ or Br⁻) by reaction with RhCl₃·3H₂O or RhBr₃·3H₂O with careful control of stoichiometry.^42,44 The water soluble mixed-metal complex [{(bpy)₂Ru(dpp)}₂RhBr₂]Br₅ was prepared from [{(bpy)₂Ru(dpp)}₂RhBr₂](PF₆)₅ by anion metathesis with Et₄NBr in acetone. The metal complexes were assayed by electronic absorption spectroscopy and cyclic voltammetry.

Fig. 7 Building block approach synthesis of dpp bridged mixed-metal supramolecular complexes.

Electrochemistry and electronic absorption spectroscopy. The dpp bridged heterotrimetallic complexes show interesting electrochemical features important to their functioning as light activated electron collectors. A representative cyclic voltammogram of the complexes, e.g. [{(bpy)₂Ru(dpp)}₂RhBr₂](PF₆)₅ is depicted in Fig. 8.^39,40,44

Fig. 8 Cyclic voltammogram of [{(bpy)₂Ru(dpp)}₂RhBr₂](PF₆)₅ (bpy = 2,2′-bipyridine, and dpp = 2,3-bis(2-pyridyl)pyrazine) in 0.1 M Bu₄NPF₆ in CH₃CN using a platinum working electrode and a Ag/AgCl reference electrode.

The cyclic voltammograms can be generalized to display ruthenium based oxidations and Rh, BL and TL based reductions. Overlapping Ru^II/III based oxidation processes for both Ru centers is seen for [{(bpy)₂Ru(dpp)}₂RhBr₂]⁵⁺ at +1.6 V vs. Ag/AgCl illustrative of the electronically isolated nature of the two ruthenium sites.⁴⁴ The reductive side consists of a first reduction at −0.33 V due to two sequential one electron irreversible reductions of the rhodium center from Rh^III to Rh^II to Rh^I.^39,42,50 At more negative potentials, two reversible reduction events at −0.72 V and −1.02 V occur due to the two dpp^0/−reduction processes. Reduction of the bpy ligands occur at more negative potentials.

DeArmond et al.^51,52 and Brewer et al.^39,50 described the electrochemical properties of cis-dihalide rhodium(III) complexes. [Rh(bpy)₂Br₂]⁺ and [Rh(bpy)₂Cl₂]⁺ undergo ECECEE reductive electrochemical processes, Fig 9. Similar behavior is observed for the dpp ligand bridged trimetallic complex, [{(bpy)₂Ru(dpp)}₂RhBr₂]⁵⁺.³⁹ This is consistent with the observed electrochemical properties of the [{(bpy)₂Ru(dpp)}₂RhBr₂]Br₅. The electrochemical features clearly show that the supramolecular complexes possess Ru(dπ) centered highest occupied molecular orbital (HOMO) and the Rh(dσ*) centered lowest unoccupied molecular orbital (LUMO), Table 1. The water soluble complex [{(bpy)₂Ru(dpp)}₂RhBr₂]Br₅ shows an irreversible Rh^III/II/Ireduction at −0.42 V in aqueous medium, which is slightly more negative than the corresponding couple of the [{(bpy)₂Ru(dpp)}₂RhBr₂](PF₆)₅ in acetonitrile. The electrochemical properties predict a low energy Ru → dpp CT transition in the trimetallic complexes with a lowest lying Ru → Rh metal to metal charge transfer (MMCT) state.

Table 1 Electrochemical properties of 2,3-bis(2-pyridyl)pyrazine bridged supramolecular complexes^a

Complex	Solvent	E 1/2/V	Assignment	Ref.
a Potential reported vs. Ag/AgCl. Ep^c and Ep^a respectively indicates the value of cathodic and anodic current peak maxima. b In acetonitrile with 0.1 M Bu4NPF6 supporting electrolyte and Pt working electrode. c In water with 0.1 M KCl supporting electrolyte and glassy carbon working electrode.
[{(bpy)2Ru(dpp)}2RhCl2](PF6)5	CH3CN^b	+1.63	2Ru^II^/III	41
		−0.37 (Ep^c)	Rh^III^/II/I
		−0.76	dpp,dpp/dpp^−,dpp
		−1.00	dpp^−,dpp/dpp^−,dpp^−
[{(bpy)2Ru(dpp)}2RhBr2](PF6)5	CH3CN^b	+1.60	2Ru^II^/III	44
		−0.33 (Ep^c)	Rh^III^/II/I
		−0.72	dpp,dpp/dpp^−,dpp
		−1.02	dpp^−,dpp/dpp^−,dpp^−
[{(bpy)2Ru(dpp)}2RhBr2]Br5	CH3CN^b	+1.60	2Ru^II^/III
		−0.39 (Ep^c)	Rh^III^/II/I
		−0.66	dpp,dpp/dpp^−,dpp
		−0.90	dpp^−,dpp/dpp^−,dpp^−
[{(bpy)2Ru(dpp)}2RhBr2]Br5	H2O^c	+1.66 (Ep^a)	2Ru^II^/III
		−0.42 (Ep^c)	Rh^III^/II/I
		−0.83	dpp,dpp/dpp^−,dpp
		−1.08	dpp^−,dpp/dpp^−,dpp^−

---

Fig. 9 ECECEE mechanism for the electrochemical reduction of cis-dihalide rhodium(III) complexes.

The supramolecular ruthenium(II) polyazine complexes display rich photochemical and electrochemical properties. The electronic absorption spectrum of [(bpy)₂Ru(dpp)](PF₆)₂ is characterized by intense bpy (290 nm) and dpp (320 nm) ligand based π → π* transitions in the UV region, and overlapping absorption bands at 430 nm and 470 nm in the visible region due to the MLCT transitions containing Ru(dπ) → bpy(π*) and Ru(dπ) → dpp(π*) CT transitions respectively. The lower energy band is the Ru(dπ) → dpp(π*) CT transition due to the lower energy π* orbital of dpp in comparison to that of bpy.

The dpp bridged trimetallic complexes are efficient light absorbers throughout the visible and ultraviolet region of the spectrum. The dpp bridged trimetallic complexes [{(bpy)₂Ru(dpp)}₂RhCl₂](PF₆)₅, [{(bpy)₂Ru(dpp)}₂RhBr₂](PF₆)₅ and [{(bpy)₂Ru(dpp)}₂RhBr₂]Br₅ show nearly identical electronic absorption spectral features.^40,44 A representative electronic absorbance spectrum of [{(bpy)₂Ru(dpp)}₂RhBr₂]Br₅ in water is shown in Fig. 10. The absorbance at 408 nm (ε = 1.3 × 10⁴ M⁻¹ cm⁻¹) is assigned as a Ru(dπ) → bpy(π*) MLCT transition, and a Ru(dπ) → dpp(π*) MLCT transition appearing at 525 nm (2.2 × 10⁴ M⁻¹ cm⁻¹). In [{(bpy)₂Ru(dpp)}₂RhBr₂]Br₅, the bridging ligand dpp π* orbitals are stabilized and the Ru(dπ) → dpp(π*) MLCT transition red shifts relative to that of the monometallic complex [(bpy)₂Ru(dpp)](PF₆)₂. The important lower energy electronic transitions of the dpp ligated ruthenium(II) complexes are summarized in Table 2. Incorporation of the polyazine bridging ligand into the [Ru(bpy)₃]²⁺ chromophoric unit, and further formation of the dpp ligand bridged trimetallic complexes, leads to a modification of the photophysical properties. For example, [(bpy)₂Ru(dpp)]²⁺, and [{(bpy)₂Ru(dpp)}₂RhX₂]⁵⁺ display red-shifted absorptions and emissions, and shortened ³MLCT excited-lifetimes (τ) relative to the [Ru(bpy)₃]²⁺, Table 2.^12,53 The trimetallic complexes [{(bpy)₂Ru(dpp)}₂RhX₂]⁵⁺ (X = Cl⁻ or Br⁻) display light-absorbing properties dominated by the two Ru light absorbing units and redox properties that illustrate the Rh(dσ*) nature of the lowest unoccupied molecular orbital (LUMO), Fig. 11.

Table 2 Electronic absorption and emission spectroscopic data for mixed-metal supramolecular complexes^a

Complex	Solvent	λ max ^abs/nm	λ max ^em/nm	τ/ns
		^1MLCT	^3MLCT
a Spectral measurements were carried out in deoxygenated solution at room temperature. bpy = 2,2′-bipyridine and dpp = 2,3-bis(2-pyridyl)pyrazine. b Ref. 12. c Ref. 50.
[Ru(bpy)3](PF6)2	CH3CN	450	630	860^b
[(bpy)2Ru(dpp)](PF6)2	CH3CN	470	691	240^c
[{(bpy)2Ru(dpp)}2RhCl2](PF6)5	CH3CN	520	760	23
[{(bpy)2Ru(dpp)}2RhBr2](PF6)5	CH3CN	520	760	26
[{(bpy)2Ru(dpp)}2RhBr2]Br5	H2O	525	770	—

---

Fig. 10 Electronic absorption spectrum of [{(bpy)₂Ru(dpp)}₂RhBr₂]Br₅ (bpy = 2,2′-bipyridine, and dpp = 2,3-bis(2-pyridyl)pyrazine) in water at room temperature.

Fig. 11 Energy state diagram for [{(bpy)₂Ru(dpp)}₂RhX₂]⁵⁺, (bpy = 2,2′-bipyridine, dpp = 2,3-bis(2-pyridyl)pyrazine, X = Cl⁻ or Br⁻) incorporating low-lying triplet metal-to-metal charge-transfer state (³MMCT). GS = ground state, ¹MLCT = singlet metal-to-ligand charge transfer state, ³MLCT = triplet metal-to-ligand charge transfer state, k_isc = rate constant for intersystem crossing, k_r = rate constant for radiative decay, k_nr = rate constant for non-radiative decay, k_et = rate constant for electron transfer, k_nr′ = rate constant for non-radiative decay from ³MMCT state, and k_rxn = rate constant for reaction.

Photoinitiated electron collection. Photoinitiated electron collection is a process where light energy is used to collect reducing equivalents. In our supramolecular photointiated electron collectors, multiple chromophores absorb light and intramolecular electron transfer results in multiple electrons being collected at the Rh center. The electrons are then available for use in a subsequent multielectron chemical process. This multielectron photochemistry has been of interest in harvesting light energy as a means to produce multielectron-reduced substrates or multielectron reduced reactive site for photocatalysis. The photochemical and photophysical properties of Ru,Rh dyads, such as [(bpy)₂Ru^II(dpp)Rh^III(bpy)₂]⁵⁺, have been studied showing single electron transfer.^54–58 We have reported trimetallics coupling ruthenium based light absorbers to rhodium centers that undergo electron collection.³⁹ Typically such systems undergo efficient intramolecular electron-transfer quenching of the Ru-based MLCT excited state by the rhodium center. The supramolecular complex [{(bpy)₂Ru(dpp)}₂RhCl₂]⁵⁺ containing ruthenium based chromophores coupled to a rhodium center possesses a low lying ³MMCT transition state. The rhodium center in this complex functions as an intramolecular electron collector, collecting reducing equivalents at the reactive metal center, Fig. 12.^39,40

Fig. 12 Photoinitiated electron collection at the Rh center of the supramolecular complex [{(bpy)₂Ru(dpp)}₂RhBr₂]⁵⁺ (bpy = 2,2′-bipyridine, and dpp = 2,3-bis(2-pyridyl)pyrazine) forming the two electron reduced [{(bpy)₂Ru(dpp)}₂Rh^I]⁵⁺ through halide loss in the presence of an electron donor (ED).

Brewer et al. have shown that the trimetallic complexes [{(bpy)₂Ru(dpp)}₂RhCl₂](PF₆)₅ and [{(bpy)₂Ru(dpp)}₂RhBr₂](PF₆)₅ undergo photoinitiated electron collection at the Rh center.⁵⁹ Herein we show that the water soluble supramolecular complex [{(bpy)₂Ru(dpp)}₂RhBr₂]Br₅ is able to photochemically collect electrons in water. Upon photolysis in the presence of an electron donor, the complex undergoes two electron reduction to produce coordinatively unsaturated Rh^I based assembly by the loss of two bromide ligands. In the presence of triethanolamine, photolysis of [{(bpy)₂Ru(dpp)}₂RhBr₂]Br₅ in water leads to the formation of the Rh^I form of the complex, [{(bpy)₂Ru(dpp)}₂Rh^I]⁵⁺. The two electron reduced Rh^I form can also be produced by electrochemical reduction at −0.4 V. This reduction process occurs through two steps in which the Rh^III is first reduced to the Rh^II state followed by conversion to the Rh^I state. The Rh^II species is not stable to isolate and study. The photoinitiated electron collection as well as the electrochemical reduction leading to the same Rh^I form of the molecular device can be followed spectroscopically, Fig. 13.

Fig. 13 Electronic absorption spectra for the conversion of [{(bpy)₂Ru(dpp)}₂Rh^IIIBr₂]⁵⁺ to the two-electron reduced form [{(bpy)₂Ru(dpp)}₂Rh^I]⁵⁺ electrochemically by electrolysis at −0.4 V in deoxygenated 0.1 M KCl in water using carbon cloth working electrode (A), and photochemically by photolysis at 520 ± 10 nm in deoxygenated water with 7 mM TEOA (B).

Photocatalytic hydrogen production from water. The [{(bpy)₂Ru(dpp)}₂RhCl₂](PF₆)₅ is our first photocatalytic molecular device proven to be an efficient hydrogen generator from water when photolyzed at wavelengths consistent with the Ru(dπ) → dpp(π*) CT excitation in the presence of an electron donor, DMA.⁴⁰ Upon photolysis, light energy is absorbed by the chromophore fragment (bpy)₂Ru^II(dpp) through MLCT excitation. In the presence of the ED, DMA, excited state electron transfer from DMA to the complex in either the ³MLCT or ³MMCT state is feasible to generate the Rh^II form. This process is repeated to produce the multielectron reduced Rh^I species with a loss of two chlorides to form a coordinatively unsaturated square planar system.

Photoinitiated electron collection is seen by comparing the spectrophotochemistry to spectroelectrochemistry of [{(bpy)₂Ru(dpp)}₂RhBr₂]Br₅, Fig. 13. This represents the first molecular system shown to undergo photoinitiated electron collection in water. Upon the reduction of [{(bpy)₂Ru(dpp)}₂RhBr₂]⁵⁺ to the two electron reduced Rh^I species [{(bpy)₂Ru(dpp)}₂Rh^I]⁵⁺ the Ru(dπ) → dpp(π*) MLCT transition blue shifts from 525 nm to 480 nm.

Rh centered complexes that undergo photoinitiated electron collection have been demonstrated to reduce water to H₂. These are the only photoinitiated electron collectors that reduce water to H₂. In acetonitrile, in the presence of water, photolysis at 470 nm of [{(bpy)₂Ru(dpp)}₂RhCl₂](PF₆)₅ (65 µM) in the presence of DMA produced 8.2 µmol of hydrogen in 4 h, Table 3.⁴⁴ The bromide analog [{(bpy)₂Ru(dpp)}₂RhBr₂](PF₆)₅ generates a slightly higher amount of hydrogen (10.9 µmol) in comparison to the analogous chloride complex [{(bpy)₂Ru(dpp)}₂RhCl₂](PF₆)₅ in acetonitrile.⁴⁴ The quantum yields of these photocatalytic processes in acetonitrile medium using DMA as electron donor are ∼0.01. The hydrogen production is catalytic, with 38 turnovers in 4 h for [{(bpy)₂Ru(dpp)}₂RhBr₂]⁵⁺, Table 3.

Table 3 Photocatalytic hydrogen production from water by supramolecular mixed-metal complexes^a

Run	Complex	Solvent	Electron donor	pH (acid used to adjust pH)	TN	H2/µmol
a Photolysis was carried out using 470 nm LED arrays at room temperature for 4 h. b Calculated pH in CH3CN, assuming that their pKa values are not purturbed relative to aqueous condition. c Ref. 44.
1	[{(bpy)2Ru(dpp)}2RhCl2](PF6)5	CH3CN	DMA	9.1 (CF3SO3H)^b	28	8.2^c
2	[{(bpy)2Ru(dpp)}2RhBr2](PF6)5	CH3CN	DMA	9.1 (CF3SO3H)^b	38	10.9^c
3	[{(bpy)2Ru(dpp)}2RhBr2](PF6)5	CH3CN	TEOA	11.8 (CF3SO3H)^b	1.7	0.48
4	[{(bpy)2Ru(dpp)}2RhBr2]Br5	H2O	TEOA	9.8	1.0	0.30
5	[{(bpy)2Ru(dpp)}2RhBr2]Br5	H2O	TEOA	7.9 (CF3SO3H)	1.3	0.38
6	[{(bpy)2Ru(dpp)}2RhBr2]Br5	H2O	TEOA	7.9 (HBr)	2.6	0.74
7	[{(bpy)2Ru(dpp)}2RhBr2]Br5	H2O	TEOA	7.9 (H3PO4)	2.9	0.83

---

One challenge which is of economical and environmental significance is replacing the organic solvent system with water as the medium. To that end, we have developed a suitable water soluble complex [{(bpy)₂Ru(dpp)}₂RhBr₂]Br₅ for photocatalysis, and further researched alternative water miscible electron donors in the place of DMA. Hence, here we employed TEOA as an electron donor and water as the reaction medium and source of hydrogen. Blue light (470 nm) photolysis of the reaction cells containing TEOA (1.5 M) and [{(bpy)₂Ru(dpp)}₂RhBr₂]Br₅ (65 µM) in water are able to produce H₂. It is very interesting and noteworthy that the supramolecular complexes can be used in very different reaction conditions, such as employing different electron donors, organic solvent mediums, as well as in water medium, while still retaining photocatalytic activity. The aqueous system produces less H₂, 0.3 µmol in 4 h, but the catalyst retains functioning. The photocatalysis in water was further studied at different reaction conditions. Hydrogen production was suppressed in the presence of air consistent with known O₂ quenching of ³MLCT excited states. The hydrogen production can be improved by adjusting the pH. At a pH value near the pK_a of TEOA, pH = 7.9, a notable increase in hydrogen production is observed. For a 4 h reaction period, TEOA buffered with triflic acid, HBr or H₃PO₄ to pH = 7.9 respectively produced 0.38, 0.74 or 0.83 µmol of hydrogen from water. The use of acids with coordinating anions such as Br⁻ or CF₃SO₃⁻ produce lesser amounts of H₂ with respect to PO₄³⁻ supporting halide loss as being important in the catalytic process. The solution at pH = 1, acidified with triflic acid, did not produce a detectable amount of hydrogen. In addition, spectrophotochemical studies of the reaction system containing pH = 1 TEOA and [{(bpy)₂Ru(dpp)}₂RhBr₂]Br₅ showed no spectral change. This is consistent with protonated TEOA being a poor electron donor unable to reduce the complex. Local in situ production of colloidal Pd or Pt metals are reported to be responsible active centers for the photocatalytic production of hydrogen by Ru–Pd and Ru–Pt based metal complexes in the presence of an electron donor.^36,38 The catalytic activities of the colloidal metals can be inhibited by the addition of Hg. However, in our Rh based systems, addition of Hg did not have any effect on the efficiency of [{(bpy)₂Ru(dpp)}₂RhBr₂]⁵⁺ for the hydrogen generation.

Conclusions

Supramolecular mixed-metal complexes have been developed for photoinitiated electron collection and photocatalytic water splitting, which have been shown to display both properties. The photoinitiated electron collecting properties of mixed-metal supramolecular complexes can be tuned by selecting proper bridging ligands and electron collecting metal centers. The dpb ligand bridged Ru, Ir, Ru complex, [{(bpy)₂Ru(dpb)}₂IrCl₂](PF₆)₅, is the first molecular photoinitiated electron collector, and has been shown to collect up to two reducing equivalents at the bridging ligand sites. The development of the mixed-metal Ru,Rh,Ru complexes bridged by dpp provide rhodium centered molecular devices, such as [{(bpy)₂Ru(dpp)}₂RhCl₂](PF₆)₅, which collect reducing equivalents at the rhodium active site and photoreduce water. The supramolecular complexes allow for variation of individual sub-units in this multicomponent system by selecting different light absorbers, terminal ligands, bridging ligands, central metal centers, and counter anions. A method for photocatalytic water splitting to hydrogen fuel was developed, which has shown to be an efficient model system functioning under various reaction conditions. The supramolecular complex [{(bpy)₂Ru(dpp)}₂RhBr₂]Br₅, which is soluble in water, adds a new direction for our photocatalytic hydrogen production studies using a water medium and readily available cheap electron donors. This complex is the first molecular system shown to undergo photoinitiated electron collection in an aqueous medium. This is one of only a handful of photoinitiated electron collectors that delivers the electrons to a substrate producing a fuel. Functioning can be enhanced with an adjustment in solution pH to the pK_a of the electron donor, while very acidic solutions do not function due to protonation of the electron donor. The large excess of amines in this solution may be impeding catalyst functioning. Work in our lab is underway to study a series of these types of complexes in more detail as well as to extend the molecular architecture, as made possible by their supramolecular design, for a new generation hydrogen photocatalysts.

Acknowledgements

Acknowledgement is made to the Chemical Sciences, Geosciences and Biosciences Division, Office of Basic Energy Sciences, Office of Sciences, U. S. Department of Energy for their generous support of our research. Acknowledgement is also made to the financial collaboration of Phoenix Canada Oil Company which holds long term license rights to commercialize the technology.

References

1. M. J. Esswein and D. G. Nocera, Chem. Rev., 2007, 107, 4022–4047.
2. J. L. Dempsey, A. J. Esswein, D. R. Manke, J. Rosenthal, J. D. Soper and D. G. Nocera, Inorg. Chem., 2005, 44, 6879–6892.
3. J. Rosenthal, J. Bachman, J. L. Dempsey, A. J. Esswein, T. G. Gray, J. M. Hodgkiss, D. R. Manke, T. D. Luckett, B. J. Pistorio, A. S. Veige and D. G. Nocera, Coord. Chem. Rev., 2005, 249, 1316–1326.
4. A. J. Bard and M. A. Mary, Acc. Chem. Res., 1995, 28, 141–145.
5. M. Kirch, J. M. Lehn and J. P. Sauvage, Helv. Chim. Acta, 1979, 62, 1345–1384.
6. J. M. Lehn and J. P. Sauvage, New J. Chem., 1977, 1, 449–451.
7. A. Fujishima and K. Honda, Nature, 1972, 238, 37–38.
8. A. Juris, V. Balzani, F. Barigelletti, S. Campagna, P. Belser and A. Vonzelewsky, Coord. Chem. Rev., 1988, 84, 85–277.
9. V. Balzani, L. Moggi and F. Scandola, Supramolecular Photochemistry, NATO ASI Series 214, Reidel, Dordrecht, The Netherlands, 1987.
10. K. Kalyanasundaram, Coord. Chem. Rev., 1982, 46, 159–244.
11. S. Campagna, S. Serroni, F. Puntoriero, C. D. Pietro and V. Ricevuto, in Electron Transfer in Chemistry, ed. V. Balzani, 2001.
12. B. Durham, J. V. Caspar, J. K. Nagle and T. J. Meyer, J. Am. Chem. Soc., 1982, 104, 4803–4810.
13. G. F. Strouse, J. R. Schoonover, R. Duesing, S. Boyde, W. E. Jones and T. J. Meyer, Inorg. Chem., 1995, 34, 473–487.
14. P. A. Anderson, G. F. Strouse, J. A. Treadway, F. R. Keene and T. J. Meyer, Inorg. Chem., 1994, 33, 3863–3864.
15. T. J. Meyer, Pure Appl. Chem., 1986, 58, 1193–1206.
16. M. H. V. Huynh, D. M. Dattelbaum and T. J. Meyer, Coord. Chem. Rev., 2005, 249, 457–483.
17. T. R. Cook, A. J. Esswein and D. G. Nocera, J. Am. Chem. Soc., 2007, 129, 10094–10095.
18. A. J. Esswein, J. L. Dempsey and D. G. Nocera, Inorg. Chem., 2007, 46, 2362–2364.
19. A. J. Esswein, A. S. Veige and D. G. Nocera, J. Am. Chem. Soc., 2005, 127, 16641–16651.
20. A. F. Heyduk and D. G. Nocera, Science, 2001, 293, 1639–1641.
21. S. F. Chan, M. Chou, C. Creutz, T. Matsubara and N. Sutin, J. Am. Chem. Soc., 1981, 103, 369–379.
22. G. M. Brown, S. F. Chan, C. Creutz, H. A. Schwarz and N. Sutin, J. Am. Chem. Soc., 1979, 101, 7638–7640.
23. S. Oishi, J. Mol. Catal., 1987, 39, 225–232.
24. R. Bauer and H. A. F. Werner, Int. J. Hydrogen Energy, 1994, 19, 497–499.
25. P. Du, J. Schneider, P. Jarosz, J. Zhang, W. W. Brennessel and R. Eisenberg, J. Phy. Chem. B, 2007, 111, 6887–6894.
26. P. Du, J. Schneider, P. Jarosz and R. Eisenberg, J. Am. Chem. Soc., 2006, 128, 7726–7727.
27. S. M. Molnar, G. Nallas, J. S. Bridgewater and K. J. Brewer, J. Am. Chem. Soc., 1994, 116, 5206–5210.
28. C. Chiorboli, S. Fracasso, F. Scandola, S. Campagna, S. Serroni, R. Konduri and F. M. MacDonnell, Chem. Commun., 2003, 1658–1659.
29. C. C. Chang, B. Pfennig and A. B. Bocarsly, Coord. Chem. Rev., 2000, 208, 33–45.
30. C. Chiorboli, S. Fracasso, M. Ravaglia, F. Scandola, S. Campagna, K. L. Wouters, R. Konduri and F. M. MacDonnell, Inorg. Chem., 2005, 44, 8368–8378.
31. R. Konduri, H. W. Ye, F. M. MacDonnell, S. Serroni, S. Campagna and K. Rajeshwar, Angew. Chem., Int. Ed., 2002, 41, 3185–3187.
32. D. F. Watson, H. S. Tan, E. Schreiber, C. J. Mordas and A. B. Bocarsly, J. Phys. Chem. A, 2004, 108, 3261–3267.
33. H. Ozawa and K. Sakai, Chem. Lett., 2007, 36, 920–921.
34. H. Ozawa, Y. Yokoyama, M. Haga and K. Sakai, Dalton Trans., 2007, 1197–1206.
35. H. Ozawa, M. A. Haga and K. Sakai, J. Am. Chem. Soc., 2006, 128, 4926–4927.
36. P. Du, J. Schneider, L. Fan, W. Zhao, U. Patel, F. N. Castellano and R. Eisenberg, J. Am. Chem. Soc., 2008, 130, 5056–5058.
37. S. Rau, B. Schafer, D. Gleich, E. Anders, M. Rudolph, M. Friedrich, H. Gorls, W. Henry and J. G. Vos, Angew. Chem., Int. Ed., 2006, 45, 6215–6218.
38. P. Lei, M. Hedlund, R. Lomoth, H. Rensmo, O. Johansson and L. Hammarstrom, J. Am. Chem. Soc., 2008, 130, 26–27.
39. M. Elvington and K. J. Brewer, Inorg. Chem., 2006, 45, 5242–5244.
40. M. Elvington, J. Brown, S. M. Arachchige and K. J. Brewer, J. Am. Chem. Soc., 2007, 129, 10644–10645.
41. A. A. Holder, D. F. Zigler, M. T. Tarrago-Trani, B. Storrie and K. J. Brewer, Inorg. Chem., 2007, 46, 4760–4762.
42. A. A. Holder, S. Swavey and K. J. Brewer, Inorg. Chem., 2004, 43, 303–308.
43. S. Swavey and K. J. Brewer, Inorg. Chem., 2002, 41, 6196–6198.
44. S. M. Arachchige, J. Brown and K. J. Brewer, J. Photochem. Photobiol. A: Chem., 2008, 197, 13–17.
45. J. R. Brown, M. Elvington, M. T. Mongelli, D. F. Zigler and K. J. Brewer, Proc. SPIE-Int. Soc. Opt. Eng.: Sol. Hydrogen Nanotechnol., 2006, 6340, 634001–634010.
46. Purwanto, R. M. Deshpande, R. V. Chaudhari and H. Delmas, J. Chem. Eng. Data, 1996, 41, 1414–1417.
47. T. E. Crozier and S. Yamamota, J. Chem. Eng. Data, 1974, 19, 242–244.
48. B. P. Sullivan, D. J. Salmon and T. J. Meyer, Inorg. Chem., 1978, 17, 3334–3341.
49. C. H. Braunstein, A. D. Baker, T. C. Strekas and H. D. Gafney, Inorg. Chem., 1984, 23, 857–864.
50. S. C. Rasmussen, M. M. Richter, E. Yi, H. Place and K. J. Brewer, Inorg. Chem., 1990, 29, 3926–3932.
51. G. Kew, K. DeArmond and K. Hanck, J. Phys. Chem., 1974, 78, 727–734.
52. G. Kew, K. Hanck and K. DeArmond, J. Phys. Chem., 1975, 79, 1828–1835.
53. G. Denti, S. Campagna, L. Sabatino, S. Serrni, M. Ciano and V. Balzani, Inorg. Chem., 1990, 29, 4750–4758.
54. K. Kalyanasundaram, M. Gratzel and M. K. Nazeeruddin, J. Phys. Chem. A, 1992, 96, 5865–5872.
55. M. T. Indelli, C. Chiorboli, L. Flamigni, L. De Cola and F. Scandola, Inorg. Chem., 2007, 46, 5630–5641.
56. M. T. Indelli, F. Scandola, L. Flamigni, J. P. Collin, J. P. Sauvage and A. Sour, Inorg. Chem., 1997, 36, 4247–4250.
57. M. T. Indelli, F. Scandola, J. P. Collin, J. P. Sauvage and A. Sour, Inorg. Chem., 1996, 35, 303–312.
58. F. Scandola, R. Argazzi, C. A. Bignozzi and M. T. Indelli, J. Photochem. Photobiol. A: Chem., 1994, 82, 191–202.
59. S. M. Arachchige, J. R. Brown, E. Chang, A. Jain, D. F. Zigler, K. Rangan and K. J. Brewer, Inorg. Chem., 2009, 48, 1989–2000.

---