A Series of Supramolecular Complexes for Solar Energy Conversion via Water Reduction to Produce Hydrogen: An Excited State Kinetic Analysis of Ru(II),Rh(III),Ru(II) Photoinitiated Electron Collectors

Mixed-metal supramolecular complexes have been designed that photochemically absorb solar light, undergo photoinitiated electron collection and reduce water to produce hydrogen fuel using low energy visible light. This manuscript describes these systems with an analysis of the photophysics of a series of six supramolecular complexes, [{(TL)2Ru(dpp)}2RhX2](PF6)5 with TL = bpy, phen or Ph2phen with X = Cl or Br. The process of light conversion to a fuel requires a system to perform a number of complicated steps including the absorption of light, the generation of charge separation on a molecular level, the reduction by one and then two electrons and the interaction with the water substrate to produce hydrogen. The manuscript explores the rate of intramolecular electron transfer, rate of quenching of the supramolecules by the DMA electron donor, rate of reduction of the complex by DMA from the 3MLCT excited state, as well as overall rate of reduction of the complex via visible light excitation. Probing a series of complexes in detail exploring the variation of rates of important reactions as a function of sub-unit modification provides insight into the role of each process in the overall efficiency of water reduction to produce hydrogen. The kinetic analysis shows that the complexes display different rates of excited state reactions that vary with TL and halide. The role of the MLCT excited state is elucidated by this kinetic study which shows that the 3MLCT state and not the 3MMCT is likely that key contributor to the photoreduction of these complexes. The kinetic analysis of the excited state dynamics and reactions of the complexes are important as this class of supramolecules behaves as photoinitiated electron collectors and photocatalysts for the reduction of water to hydrogen.


Introduction
The demand for alternative fuel sources is continually increasing. An attractive approach to this issue is the conversion of solar energy to chemical energy in the form of H 2 O splitting to produce H 2 fuel [1,2]. At neutral pH and 25 °C, H 2 O can be split into H 2 and O 2 via a multi-electron pathway that requires 1.23 V [3]. Sunlight provides an abundant amount of energy to the Earth's surface that contains the required energy to drive this thermodynamically uphill, multi-electron reaction. However, H 2 O does not absorb an appreciable amount of sunlight reaching the surface, therefore systems must be designed to efficiently absorb light and deliver appropriate charges to H 2 O. One means of achieving this goal is through the use of supramolecular complexes [4]. In this arena, supramolecular complexes are described as molecular machines comprised of multiple molecular components whose individual properties contribute to the overall functioning of the system [5]. Supramolecular complexes of interest in solar energy conversion schemes are photochemical molecular devices (PMDs) as they perform a specific light-driven task utilizing solar energy as the thermodynamic driving force for a desired chemical reaction. Engineering PMDs to perform specific, complex functions at the molecular level allows for the exploitation of these systems as potential photocatalysts. Systems can be designed to perform photoinduced vectoral electron transfer and charge migration between appropriate electron donor (ED), such as an electron rich, metal-based light absorber (LA), and electron acceptor (EA) sites. Generating this photoinduced charge separation and migration within PMDs is of considerable interest in the realm of solar energy conversion schemes [5].
Modifying the [{(bpy) 2 Ru(dpb)} 2 IrCl 2 ](PF 6 ) 5 trimetallic by changing the BL from dpb to dpp and the central metal from Ir(III) to Rh(III) generates [{(bpy) 2 Ru(dpp)} 2 RhCl 2 ](PF 6 ) 5 (dpp = 2,3-bis (2-pyridyl)pyrazine) [11]. This Ru(II),Rh(III),Ru(II) trimetallic complex displays orbital inversion with the LUMO now localized on the Rh(III) metal center and is the first reported PEC to collect multiple reducing equivalents at a central metal site while staying intact. Intramolecular electron transfer from the Ru(II)-based LAs to the Rh(III)-based EC subunit produces a doubly-reduced Rh metal center with the potential to deliver electrons to a substrate. Further modification of the [{(TL) 2 Ru(dpp)} 2 RhX 2 ] 5+ molecular components through halide variation, as well as TL variation, has generated a series of complexes functioning as PECs. In the presence of a sacrificial ED and H 2 O, many of the Rh centered PECs function as photocatalysts reducing H 2 O to H 2 [12][13][14][15][16]. Figure 1 displays an example of an ED-LA-BL-EC-BL-LA-ED structural motif for PEC and the required orbital energetics. Photoexcitation at 470 nm produces 7.2 ± 0.7 µmol of H 2 in an CH 3   Intermolecular electron transfer reactions have been widely studied focusing on the development of molecular photovoltaics [17][18][19][20]. Ru(II)-based polyazine LAs are efficient light absorbers throughout the UV and visible regions as photoexcitation populates 3 MLCT (metal-to-ligand charge transfer) The rate of excited state electron transfer depends on the thermodynamic driving force for these reactions [4,21,22]. The excited state oxidation (Equation 4) and reduction (Equation 5) potentials of the excited LA are calculated using the energy of the E 0-0 transition of the 3 MLCT emission and the ground state redox potentials.
In the equations above, LA is the Ru(II)-polyazine light absorber, E(LA/LA + ) is the ground state oxidation potential, E(LA/LA − ) is the ground state reduction potential, E(*LA + /LA) is the excited state oxidation potential, and E(*LA/LA − ) is the excited state reduction potential. Emission spectroscopy is often used to probe the rate of quenching of the emissive 3 MLCT excited states by a quenching species, such as an ED [23,24]. Supramolecular complexes take advantage of covalently coupled molecular components to promote photoinduced intramolecular electron transfer. Bridging a Ru(II)-based LA to an EA subunit (LA-EA) can afford excited state intramolecular electron transfer upon photoexcitation of the LA subunit, as shown in Equations 6 and 7.
Reported herein is a study of the excited state dynamics and a kinetic analysis of the quenching of the 3

Photophysical Properties
The [{(TL) 2 Ru(dpp)} 2 RhX 2 ] 5+ trimetallic complexes are efficient light absorbers throughout the UV and visible regions at room temperature in acetonitrile, Figure 3. The UV region is dominated by intense TL π→π* intraligand (IL) transitions, with the dpp BL π→π* IL transitions occurring at slightly lower energy. The visible region displays higher energy Ru(dπ)→TL(π*) CT transitions and lowest energy Ru(dπ)→dpp(π*) CT transitions. The lowest-lying MLCT transition is nearly isoenergetic in the series of complexes indicative of the similar Ru(dπ)→dpp(π*) CT nature of the optically populated state. These systems absorb more of the solar spectrum than typical [Ru(bpy) 3 ] 2+ based systems via enhanced molar absorptivity in the UV and visible with Ru(dπ)→dpp(π*) CT transitions that provide absorption in the low energy visible. The excited state properties of the Ru(II),Rh(III),Ru(II) trimetallic and model Ru(II),Ru(II) bimetallic complexes are summarized in Table 1. The trimetallic complexes of the design [{(TL) 2 Ru(dpp)} 2 RhX 2 ] 5+ (TL = bpy, phen, Ph 2 phen; X = Cl or Br) display weak emission and a short excited state lifetime of the Ru(dπ)→dpp(π*) 3 MLCT emissive excited state when compared with the model [(TL) 2 Ru(dpp)Ru(TL) 2 ] 4+ bimetallic complexes, which display the same Ru→μ-dpp 3 MLCT emissive state but lack a Rh-based EC metal center. The Ru(II),Ru(II) bimetallic complexes are used as model systems for photophysical studies due to the similar nature and energy of the emissive Ru→μ-dpp CT excited state. Terminal ligand variation has been shown to modulate this 3 MLCT emissive excited state (presumably from a contribution to the formally Ru(dπ) HOMO in this motif), therefore different Ru(II),Ru(II) bimetallics are needed for TL = bpy, phen, or Ph 2 phen [32]. Figure 4 displays the state diagram for the trimetallic complex [{(Ph 2 phen) 2 Ru(dpp)} 2 RhBr 2 ](PF 6 ) 5 . At RT, deactivation from the 3 MLCT excited state is dominated by intramolecular electron transfer to populate a low-lying, energetically close Ru(dπ)→Rh(dσ*) 3 MMCT excited state. This is supported by the observation of a Rh-based lowest unoccupied molecular orbital (LUMO) in electrochemical analyses of these systems [11][12][13]15,16]. The shortened excited state lifetime of the emissive 3 MLCT state in the Ru(II), Rh(III),Ru(II) motif at RT is ascribed to intramolecular electron transfer to populate a low-lying 3 MMCT state which quenches the 3 MLCT state at RT but not at 77 K [32]. Due to the similar energy and nature of the emissive 3 MLCT excited state for the Ru(II),Ru(II) bimetallic and Ru(II),Rh(III), Ru(II) trimetallic complexes, it is assumed that calculated rate constants for radiative (k r ) and non-radiative (k nr ) decay from the 3 MLCT excited state of the bimetallics are the same for the analogous trimetallics. Both the title trimetallics and the model bimetallic used as the model for each trimetallic possess not only the same Ru(dπ)→dpp(π*) 3 MLCT emissive state but also the same TL and the same (TL) 2 Ru II (∝-dpp) sub-unit. Under this assumption, the rate constant for intramolecular electron transfer (k et ) to populate the non-emissive 3  trimetallics, varying the halide from Cl to Br displays a decrease in Φ em and τ with a subsequent increase in k et . The inclusion of the weaker σ-donating Br stabilizes the 3 MMCT excited state and affords enhanced driving force and rate of intramolecular electron transfer to populate the 3 MMCT state. Choice of TL within the Ru(II),Rh(III),Ru(II) architecture also impacts the excited state properties with the energy, Φ em , and τ of the formally Ru(dπ)→dpp(π*) 3 MLCT excited state varying. The phen systems display enhanced rates of intramolecular electron transfer to populate the 3 MMCT state vs. bpy or Ph 2 phen. The phen systems have slightly higher energy 3 MLCT excited states which may provide a larger driving force for electron transfer to populate the 3 MMCT state.  ](PF 6 ) 5 (TL = bpy, phen or Ph 2 phen; X = Cl or Br). hν is energy of the photon, k isc is the rate constant for intersystem crossing, k r is the rate constant for radiative decay, k nr is the rate constant for non-radiative decay, k et is the rate constant for intramolecular electron transfer, and k rxn is the rate constant for a photochemical reaction.
At 77 K in a rigid glass matrix, the Ru(II),Rh(III),Ru(II) trimetallic and Ru(II),Ru(II) bimetallic complexes display similar emissive excited states with nearly equivalent lifetimes. The shape of the 3 MLCT emission profile sharpens in rigid media at 77 K and the emission maxima blue shift. This is consistent with electron transfer at RT to populate the 3 MMCT state from the emissive 3 MLCT state being impeded at 77 K in a rigid media.

Photochemical Properties
Photochemical reduction of these [{(TL) 2 Ru(dpp)} 2 Rh III X 2 ] 5+ trimetallic complexes illustrates their ability to undergo photoinitiated electron collection at the Rh(III) metal center to generate Rh(I) centered trimetallics. When illuminated at 470 nm in the presence of the sacrificial electron donor DMA, the electronic absorption spectrum displays a shift to higher energy of the Ru(dπ)→dpp(π*) CT transition due to the formation of a more electron-rich Rh(I) metal center. The increase in electron density at Rh destabilizes the dpp(π*) acceptor orbitals relative to Rh(III) resulting in an increase in the energy of the observed Ru(dπ)→dpp(π*) CT transition. The electronic absorption spectra generated through electrochemical reduction of the Rh(III) to Rh(I) metal center correlates well with the photochemical reduction of trimetallics [11]. The reduction of these complexes occurs via an ECEC mechanism analogous to previously studied [Rh(NN) 2 X 2 ] + , Figure 5 [34,35]. The photochemical reduction of the complexes [{(TL) 2 Ru(dpp)} 2 Rh III X 2 ] 5+ is critical to photocatalysis and involves many steps including possible reactions from the 3 [34,35]. Photochemically, this reduced species can be formed through intermolecular electron transfer from a sacrificial electron donor to the 3 MLCT or 3 MMCT excited states. Additionally, both excited states can undergo unimolecular or bimolecular deactivation. The present kinetic study will analyze the rate of intramolecular electron transfer described above, the rate of quenching of the 3 MLCT state via Stern-Volmer analysis and the rate of photochemical reduction via spectroscopic analysis. These Using our mechanism, unimolecular deactivations k 1 and k 4 include radiative, k r , and non-radiative, k nr , decay including relaxation mediated by solvent. Bimolecular deactivations, k 2 and k 3 , include electron transfer from DMA followed by rapid back electron transfer as well as other bimolecular deactivations by DMA.

Emission Quenching
The emissive nature of the 3 MLCT excited state provides a handle to study the excited state dynamics. This probe was used to study the rate of intramolecular electron transfer (k et ) as described above. Addition of the ED DMA provides a means to assay the kinetics of quenching of the 3 MLCT state by this ED. The sacrificial electron donor DMA has been shown to quench the 3 MLCT emissive excited state of Ru-polyazine complexes and [{(bpy) 2 Ru(dpp)} 2 RhCl 2 ] 5+ through bimolecular interactions [11]. The [{(TL) 2 Ru(dpp)} 2 RhX 2 ] 5+ trimetallic complexes reported herein undergo efficient excited state reductive quenching of the 3 MLCT emission. DMA is reported to quench the 3 MLCT emission of [Ru(bpy) 3 ] 2+ [23,36] and [Ru(bpz) 3 ] 2+ (bpz = 2,2′-bipyrazine) [37] with a rate constant of 7.1 × 10 7 M −1 s −1 and 8.4 × 10 9 M −1 s −1 , respectively. The 3 MLCT excited state of *[Ru(bpy) 3 ] 2+ and *[Ru(bpz) 3 ] 2+ have excited state reduction potentials of 0.82 V and 1.50 V vs. Ag/AgCl, respectively, while DMA has a ground state oxidation potential of 0.86 V vs. Ag/AgCl [36]. The thermodynamic driving force (E redox ) for reductive quenching of the 3 MLCT excited state is determined by the ground state oxidation potential of the electron donor (E(ED 0/+ )) and the excited state reduction potential of the Ru(II)-polyazine complex (E(*CAT n+ /CAT (n − 1)+ )), Equations 16 and 17: E(*CAT n+ /CAT (n − 1)+ ) = E 0-0 + E(CAT n+ /CAT (n − 1)+ ) (17) where E(CAT n+ /CAT (n − 1)+ ) is the ground state reduction potential of the complex and E 0-0 is the energy of the 0-0 transition between the excited state and the ground state. The E 0-0 energy is estimated using the observed 77 K emission maxima.     23 2.9 × 10 9 a Potential in V vs. Ag/AgCl, E(*CAT n+ /CAT (n − 1)+ ) is the excited state reduction potential; b Thermodynamic driving force calculated by measuring the difference between the excited state reduction potential of the complex and the ground state oxidation potential of the electron donor DMA (DMA 0/+ = 0.86 V vs. Ag/AgCl); c Rate constant for quenching of 3 MLCT excited state through bimolecular interactions with the electron donor DMA; d Values are reported k q rate constants; e From reference [33]; f From reference [34].
A Stern-Volmer analysis was performed to observe the 3 MLCT emission quenching of the trimetallic complexes using the electron donor DMA, Figure 6. All complexes show a linear Stern-Volmer relationship with reduction of the 3 MLCT excited state emission intensity varying linearly with increasing [DMA]. Equation 18 relates the ratio of the intensity of 3 MLCT emission in the absence (I 0 ) and presence (I) of DMA to the concentration of DMA added: where k 1 = k r + k nr [38]. The slope of the Stern-Volmer quenching plot contains the rate constant for quenching by DMA of the 3 MLCT state via bimolecular deactivation (k 2 ) or photoreduction (k q ) to form the reduced Rh(II) photoproduct. From these experiments, the values corresponding to the deactivation of the 3 MLCT excited state through bimolecular interactions with DMA (k q + k 2 ) were obtained and vary from 1.5 × 10 9 M −1 s −1 to 5.9 × 10 9 M −1 s −1 , indicative of efficient quenching of the 3 MLCT excited state, Table 2. Competing pathways for deactivation of the 3 MLCT excited state are radiative (k r ) and non-radiative (k nr ) decay to the 1 GS and intramolecular electron transfer (k et ) to populate the 3 MMCT excited state. The calculated rate constants for these unimolecular deactivation pathways (Table 1) are substantially smaller than the DMA bimolecular quenching rate constants. This observation suggests that in the presence of DMA, the dominating pathways of deactivation from the 3 MLCT excited state involve bimolecular quenching with the electron donor. While this Stern-Volmer analysis of the 3 MLCT excited state quenching does not permit the independent calculation of k q and k 2 , photochemical product formation studies enable extraction of the k q value and the subsequent value of k 2 is obtained below.

Product Formation
The upon reduction from Rh(III) to Rh(II) to Rh(I) proceeds with a smooth shift to higher energy of the Ru(dπ)→dpp(π*) CT transition. This photoproduct can be generated through excited state reductive quenching of the 3   Product formation to generate the reduced supramolecules can occur from the 3 MLCT or 3 MMCT excited states. Kinetic analysis first considers product formation from the 3 MLCT state. The Ru(dπ)→dpp(π*) 3 MLCT excited state can deactivate through unimolecular deactivation to the ground state (k 1 ), bimolecular deactivation through interaction with DMA (k 2 ), intramolecular electron transfer to populate the 3 MMCT state (k et ), or reductive quenching by DMA to produce the reduced species (k q ). Equation 19 relates the quantum yield of formation of the reduced species (Φ product ) to [DMA].
Plotting 1/Φ product vs. 1/[DMA] gives a linear relationship, with a slope of (k 1 +k et )/k q and an intercept of (k q +k 2 )/k q . The rate constant for unimolecular deactivation, k 1 , is the sum of k r and k nr and has been determined above. The rate constant for intramolecular electron transfer, k et , was obtained from our above emission analysis. This allows the determination of k q and k 2 , Table 3. Reduction of the [{(TL) 2 Ru(dpp)} 2 RhX 2 ] 5+ can also occur from the 3 MMCT excited state. The Ru(dπ)→Rh(σ*) 3 MMCT excited state can undergo multiple deactivation pathways including unimolecular deactivation (k 4 ), bimolecular deactivation with DMA (k 3 ) and reductive quenching of the excited state by DMA to produce the singly reduced species (k q2 ). The efficiency of Rh(II) product formation from the 3 MMCT state depends on the efficiency of populating the 3 MMCT state (Φ 3MMCT ). Equation 20 relates the quantum yield of formation of the reduced species (Φ product ) from the 3 MMCT excited state to [DMA].
A plot of 1/Φ product and 1/[DMA] is linear with a slope of (1/Φ 3MMCT )(k 4 /k q2 ) and an intercept of (k q2 + k 3 )/k q2 . Values obtained for Φ 3MMCT and k 4 /k q2 from these analyses are presented in Table 3. The Φ 3MMCT is given by the ratio of k et to k r +k nr determined from the emission of the [{(TL) 2 Ru(dpp)} 2 RhX 2 ] 5+ complexes above. A direct measure of k 4 is not provided so this analysis gives a ratio of k 4 /k q2 . The Ru(II),Rh(III) bimetallic complex [(Me 2 phen) 2 Ru(Mebpy-CH 2 CH 2 -Mebpy) Rh(Me 2 bpy) 2 ] 5+ (Me 2 phen = 4,7-dimethyl-1,10-phenanthroline; Me 2 bpy = 4,4'-dimethyl-2,2′-bipyridine) was studied via transient spectroscopy to provide k 4 = 7.1 × 10 9 s −1 [25]. This system shows a k et to populate the 3 MMCT state of 1.4 × 10 7 s −1 , similar in magnitude to our systems. The rate of back electron transfer from Rh(II) to Ru(III) to generate the ground state from the 3 MMCT state, k 4 , is expected to be fast for our complexes given the direct dpp coupling of the Ru and Rh centers in our systems vs. the Mebpy-CH 2 CH 2 -Mebpy linker in the previously reported system. Assuming k 4 for our systems is >7.1 × 10 9 s −1 , this calculates k q2 values of ca. 10 11 M −1 s −1 , an unreasonably large number.
This suggests supramolecule reduction occurs primarily out of the 3 MLCT state in our systems. The direct analysis of the contribution of the 3 MMCT state to product formation is not accessible via these methods. The analysis herein does highlight that any photoreduction via the 3 MMCT state would occur on the picosecond time scale.
Several pathways of deactivation of the 3 MLCT state impact the trimetallic complexes' ability to function as PECs and ultimately as solar energy conversion catalysts for water reduction. Deactivation of the 3 MLCT state to the GS is a dominant pathway both via non-radiative (k nr ), radiative (k r ), and bimolecular deactivation (k 2 ). The quenching of the 3 MLCT excited state of the trimetallic complexes [{(TL) 2 Ru(dpp)} 2 RhX 2 ] 5+ in CH 3 CN at RT is very efficient with rate constants 1-6 × 10 9 M −1 s −1 at the diffusion control limit. The rate of the associated photoreduction of the trimetallics by DMA is less, 2-4 × 10 8 M −1 s −1 , indicative of the often efficient back electron transfer prior to cage escape in Ru-polyazine systems. Nonetheless photoreduction occurs at a significant rate, 10 8 M −1 s −1 , providing for the rapid conversion of the Rh(III) supramolecules to reduced species. The variation of the halide bound to the Rh from Cl to Br provides for enhanced rates of photoreduction independent of TL (bpy, phen or Ph 2 phen). TL variation impacts observed rates as well. Emission quenching by DMA (k q + k 2 ) is most efficient for phen complexes with Ph 2 phen providing for the lowest rate of DMA quenching of the 3 MLCT excited state. The enhanced rate of quenching of the 3 MLCT state by DMA for TL = phen may be a result of efficient π-π interaction of the phen TL with the DMA electron donor placing the DMA near the Ru center.
The above Ru(II),Rh(III),Ru(II) trimetallic complexes are photocatalysts in the reduction of H 2 O to H 2 , ][DMAH + ] in CH 3 CN were photolyzed for 5 h using a 470 nm LED light source. Turnover numbers (TON) were measured as the mol of H 2 produced per mol of Rh catalytic center. The quantum efficiency of H 2 (Φ H2 ) was measured as mol of H 2 produced per mol of photons, multiplied by two given the formation of H 2 is a two photon and two electron process within our molecular architecture. Halide variation from Cl to Br displays more efficient H 2 production as suggested by the enhanced rates of reduced Rh product formation. Photocatalysts where TL = phen display the lowest amount of H 2 , consistent with the larger rate constant for bimolecular deactivation of the 3 MLCT excited state (k 2 ) inhibiting efficient formation of the reduced Rh species. While photocatalysts with TL = Ph 2 phen outperform TL = bpy or phen systems, the observed excited state rate constants do not vary greatly, suggesting additional factors impact photocatalytic functioning. The steric demands of the Ph 2 phen ligand may provide protection of the photoreduced Rh(I) center, decreasing unfavorable side reactions and therefore enhancing H 2 production.

Electronic Absorption Spectroscopy
Electronic absorption spectra were measured using a Hewlett-Packard 8452A diode array spectrophotometer with 2 nm resolution. Spectra were recorded at room temperature in spectral grade acetonitrile using a 1 cm path length cylindrical quartz cuvette (Starna Cells, Inc., Atascadero, CA, USA).

Steady State Luminescence Spectroscopy
The room temperature steady state emission spectra were measured in spectral grade acetonitrile using a 1 cm path length quartz cuvette equipped with a screw top (Starna Cells, Inc.; Atascadero, CA, USA). The instrument used to record the spectra was a QuantaMaster Model QM-200-45E fluorimeter from Photon Technologies International, Inc. The excitation source was a water-cooled 150 W Xenon arc lamp with the corresponding emission collected at a 90° angle using a thermoelectrically cooled Hamamatsu 1527 photomultiplier tube operating in photon counting mode with 0.25 nm resolution. The emission monochromator contained a Czerny-Turner style grating monochromator set to 1,200 line/mm 750 nm blaze.

Excited State Emission Quenching
Stock solutions of each trimetallic complex were prepared using spectral grade acetonitrile. Sample solutions were composed of a fixed final concentration of trimetallic complex (~30 μM) in a 1 cm quartz cuvette with increasing final concentrations of DMA ((2.4-0.2) × 10 −2 M) added to a new sample. DMA was injected into the sample in the dark using a syringe just prior to excitation from the 150 W Xe arc lamp light source. The steady state emission spectrum for each sample was obtained and a Stern-Volmer plot of I 0 /I vs.

Photochemical Product Formation
Sample solutions were composed of a fixed concentration of trimetallic complex (~25 μM) with increasing final concentrations of DMA ((4.0-0.33) × 10 −3 M) added to each sample. The electronic absorption spectra were measured after photolysis on a 470 nm LED array designed and constructed locally (flux = 2.83 × 10 19 photons/min) [40]. Data were plotted and extrapolated to zero time.

Photocatalytic Hydrogen Production
The photocatalytic hydrogen production experiments were performed using modifications of previously reported conditions [14]. The trimetallic stock solutions (92 μM) in CH 3 CN were combined with water (acidified to pH 2 using CF 3 SO 3 H) in air tight photolysis reaction cells that were deoxygenated using argon gas. The electron donor DMA was deoxygenated separately and injected into the reaction cells just prior to photolysis (final conditions:  [40]. The amount of hydrogen produced was monitored in real-time using a HY-OPTIMA™ 700 in-line process solid state hydrogen sensor from H2scan connected to the photolysis reaction cell. The sensor was calibrated by injecting known quantities of hydrogen into the photolysis cells and generating a calibration curve. The functioning of the sensor was verified by injecting a 100 μL sample from the reaction cell's headspace into a series 580 GOW-MAC gas chromatograph equipped with a rhenium-tungsten thermal conductivity detector and a 5 Å molecular sieves column using ultra-high purity argon gas. The gas chromatograph signal was amplified with a Vernier Software instrument amplifier and recorded using Logger Pro 3.4.5 software. The gas chromatograph was calibrated for hydrogen signal sensitivity by injecting known amounts of hydrogen gas and generating a calibration curve. The reported value for hydrogen production is the average of three experiments.

Conclusions
The kinetic analysis shows that both TL and halide bound to Rh impacts observed excited state dynamics. Variation of TL and halide bound to Rh impacts rates of reactions from the formally Ru→dpp CT excited states. The 3 MLCT states are longest lived for TL = Ph 2 phen and X = Cl and shortest for TL = phen and X = Br. The rate of intramolecular electron transfer, k et , to generate the Φ 3MMCT is large in all cases varying from 0.73-0.82. Quenching of the 3 MLCT states is very efficient and all complexes studied undergo photoinitiated electron collection to produce the Rh(I) complex. Many of these systems are known photocatalysts for H 2 O reduction to produce H 2 with high quantum yields and turnovers with respect to known supramolecular photocatalysts. The study of the rate of quenching of the 3 MLCT state by DMA shows rapid quenching near the diffusion control limit. Photoreduction occurs at a rate, k q , of (2-4) × 10 8 M −1 s −1 , leading to rapid reduction of the supramolecules. This is consistent with the high thermodynamic driving force for reduction of the trimetallics by DMA which is thermodynamically favorable by 0.49-0.60 V. Analysis of the kinetic requirements for photoreduction from the 3 MLCT and 3 MMCT states suggests photoreduction occurs primarily from the 3 MLCT state. These kinetic analyses provide considerable insight into the important excited state reactions of these Ru(II),Rh(III),Ru(II) supramolecular photoinitiated electron collectors, a class of molecules of interest as visible light induced photocatalysts for H 2 O reduction to H 2 .