Photovoltaics as an Experimental Tool for Determining Frontier Orbital


Photovoltaics as an Experimental Tool for Determining Frontier Orbital...

0 downloads 87 Views 2MB Size

Subscriber access provided by UNIV OF SCIENCES PHILADELPHIA

Photovoltaics as an Experimental Tool for Determining Frontier Orbital Energies and Photocatalytic Activity of Thiol Protected Gold Clusters Kevin G. Stamplecoskie, Goonay Yousefalizadeh, Lea Gozdzialski, and Hannah Ramsay J. Phys. Chem. C, Just Accepted Manuscript • DOI: 10.1021/acs.jpcc.8b00571 • Publication Date (Web): 08 Mar 2018 Downloaded from http://pubs.acs.org on March 8, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Photovoltaics as an Experimental Tool for Determining Frontier Orbital Energies and Photocatalytic Activity of Thiol Protected Gold Clusters Stamplecoskie, K.G.*, Yousefalizadeh, G., Gozdzialski, L., Ramsay, H. Department of Chemistry, Queen’s University, Kingston, Ontario, K7L 3N6, Canada

Abstract: Metal clusters are an emerging photonic material with a rapidly growing library of stable, atomically precise, clusters recently reported. The growing interest in metal clusters, such as thiol-protected Au clusters, is due in part to interest in this new material for light harvesting, catalysis and electron transfer studies, high efficiency photovoltaics (i.e. metal- cluster-sensitized solar cells (MCSSC) with longterm stability). Herein, we use cobalt redox couples as a tool to investigate the electronic properties (i.e. HOMO energy level) and the wavelength-dependent photocatalytic activities of glutathione protected gold clusters. To this purpose, we have synthesized a series of cobalt complexes with bipyridine and phenanthroline derivatives in order to study Au18(SR)14 and Au25(SR)18 sensitized solar cells. Using cobalt complex mediators as an alternative to the conventional high performance I/I3- electrolyte, long-term performance of MCSSC is achieved. Furthermore, these one-electron redox mediators conclusively demonstrate the unique optical activity of clusters, with excitation wavelength-dependent photocatalytic activity. Introduction TiO2 as a photocatalyst material was investigated to a great extent in the 1960s, but increasing global energy demand and the pursuit of clean, renewable energy has led to resurgence in the discovery of new materials for photocatalysis.1 Due to the wide bandgap of TiO2 (3.2 eV), a co-catalyst, capable of absorbing more of the solar spectrum is needed to increase solar energy conversion.1-3 Metal clusters, especially thiol stabilized gold clusters, have emerged as a new photocatalyst material. For example, clusters such as Au25SR18, have been reported as effective sensitizer for TiO2.4-5 Cluster chemistry offers a great deal of versatility in the choice of metals and stabilizing ligands.6-8 The variability in structure can subsequently afford a great deal of control over the unique optical and electronic properties of the isolated clusters.5, 9-14 Furthermore, the versatility in chemical composition of metal clusters allows for the design of catalysts with tuneable reactivity/catalytic activity.15-18 Atomically precise clusters are rapidly evolving because of their potential as highly active and selective catalysts/photocatalysts.18-20 Apart from their application, achieving a deep insight

ACS Paragon Plus Environment

The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

of the fundamental properties of metal clusters is of great importance to understanding the size dependent optical and electronic properties of metal materials.21-23 As photocatalysts, metal clusters can be used to convert light (usually solar radiation) into electricity or other forms of energy, like the potential energy stored in chemical bonds.24 A general scheme for a sensitized TiO2 catalyst is illustrated in Figure 1. At minimum, a highly efficient molecular sensitizer should; (1) be capable of absorbing a broad spectrum of solar radiation, (2) have a LUMO energy close to, but above, the conduction band edge of TiO2, to perform efficient electron transfer (k1) while not losing a great deal of energy through thermal relaxation, (3) have a sufficiently long excited state lifetime to facilitate electron transfer and sensitizer regeneration (k2), and (4) adhere to the polar surface of TiO2, often through carboxylic acid or other polar groups.

Fig. 1. General mechanism of electron transfer between a sensitizer and TiO2, including light excitation of the sensitizer (hn), electron transfer from the sensitizer to the conduction band (CB) of TiO2 (k1) and ground state dye regeneration (k2).

Metal clusters have been incorporated into devices for light-harvesting and other photonics applications.2, 25 The photovoltaic performance of several thiol protected gold clusters has been reported, with maximum solar energy conversion efficiencies currently reaching 2-4 %.25-27 The use of gold clusters in devices has come into question due to the unique challenges of maintaining long-term stability, especially when combined with other materials or in different device architectures. For example, on metal oxide supports, photo-excited thiol protected gold clusters can rapidly convert to larger nanoparticles, which have entirely different optical properties and drastically reduced light-harvesting capabilities.19 The most common redox mediator in liquid junction solar cells, which is used to regenerate sensitizers, is the I-/I3- redox couple. However, halogens, such as I-, are corrosive towards gold, making it necessary to use other electron donors. Using the Co2+/3+(bpy)3 redox mediator, the Kamat lab was able to demonstrate stable photovoltaic performance

ACS Paragon Plus Environment

Page 2 of 16

Page 3 of 16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

from metal-cluster-sensitized solar cells (MCSSC) with >2 % energy conversion efficiencies, most importantly over several hours of illumination.26 This seminal work was important in proving that stable and efficient photocatalysis can be achieved with thiol protected gold clusters. The search is now on for clusters that can increase this solar energy conversion efficiency, remain stable over long-term use, as well as to discover catalyst systems that can efficiently harvest solar radiation for high turnover, and/or selective photocatalysis.17, 28 HOMO/LUMO energy levels and excited state dynamics are critical factors in evaluating the potential for particular metal clusters in photocatalysis.22 The HOMO and LUMO of clusters like Au25SR18 have been estimated using electrochemical techniques.29-30 However, it can be challenging to determine the HOMO/LUMO energies of other clusters by cyclic voltammetry, especially as many clusters are not stable under oxidizing or reducing conditions for extended periods of time. Other techniques, such as absorbance and fluorescence spectroscopy, only provide an upper or lower limit to the HOMO-LUMO energy gap. Herein we use photovoltaic devices as a tool to experimentally determine some of the fundamental optical and electronic properties of clusters, as an alternative to conventional solution electrochemistry.25, 27, 31 In this article we investigate the photovoltaic performance of metal-cluster-sensitized solar cells (MCSSC) with the most common 25 Au atom gold cluster, Au25SR18, as the light absorbing sensitizer adsorbed on TiO2 photoanodes. The photovoltaic performance of Au18SR14 is also investigated. This glutathione protected Au18 cluster is, to the best of our knowledge, the most efficient atomically precise gold cluster in MCSSC reported to date. We report the photovoltaic performance of MCSSC using a series of CoII/III redox couples, where the redox potential of the complexes spans a range of 0.65-0.95 eV vs NHE. This series of CoII/III complexes are a valuable tool for experimentally approximating the HOMO energy level of clusters. Furthermore, the photovoltaic performance of clusters is investigated using the conventional I-/I3- redox couple as well. The long-term stability of devices made with both the CoII/III and I-/I3- redox mediators was evaluated, with a focus on corrosion of the metals by halides and other possible parasitic reactions. Synthesis of Glutathione Protected Au18 and Au25 Clusters. Au18SR14 and Au25SR18 were synthesized according to previously reported methods.32 Au18SR14 clusters were synthesized using 0.15 g HAuCl4·3H2O (99.9 %, Sigma-Aldrich) dissolved in 1.2 mL Methanol. Then 0.3 g glutathione (reduced, 98 %, SigmaAldrich) dissolved in 1.8 mL ultra pure water was added. While the solution was sonicated, its color became transparent, indicating the conversion of Au3+ to Au+. The solution was diluted to 96 mL by methanol and then stirred for 10min. Sodium cyanoborohydride was used as a mild reducing agent, which provides a slower

ACS Paragon Plus Environment

The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

reduction of Au(I) than NaBH4, affording sized-selective growth of Au18SR14. 4.5 ml of a 220 mM aqueous solution of NaBH3CN (95 %, SigmaAldrich) was prepared and added dropwise. The solution was left to stir vigorously for 30min. The product was collected through centrifugal precipitation and washed repeatedly with acetonitrile and re-dissolved in milliQ water. An absorbance spectrum of the product is shown in Figure 2a. There are several reported methods for synthesizing glutathione stabilized Au25SR18 clusters. Here, Au25SR18 clusters were synthesized using carbon monoxide as a reducing agent, in a COdirected synthesis.33 0.1 g HAuCl4·3H2O (99.9 %, Sigma-Aldrich) and 0.15g glutathione (reduced, 98 %, Sigma- Fig. 2.: Absorbance spectra of (a) aqueous Aldrich) were dissolved in 100 mL ultra solutions of synthesized Au18SR14 and Au25SR18 TiO2 active layers sensitized with gold pure water and stirred until the solution (b) clusters. was nearly colourless (5 minutes). The solution was then adjusted to pH 11 using 1 mM aqueous sodium hydroxide. The gaseous reducing agent, carbon monoxide, was bubbled through the solution for about 2 minutes, sealed, and left to stir overnight. The gold clusters were collected through centrifugal precipitation with acetonitrile and re-dissolved in aqueous media. The resulting glutathione protected Au cluster was characterized by UV-vis absorption and purified by RP-HPLC, following a report on separation of hydrophilic Aun(SG)m clusters.34 Separation was done through a linear gradient program (starting from 25 mM sodium phosphate buffer solution (pH = 6.9) transformed to 25 mM TBAClO4 methanol solution) using a C18 (Hypersil gold column 250 × 21.2 mm). Figure 2a shows the absorbance spectrum of separated Au25SR18 clusters. In a previous work, the atomically precise synthesis of these clusters has been confirmed by mass spectrometric analysis.35 Synthesis of CoII/III Redox Couples and Electrochemical Characterization. The Co (II/III) redox mediators were synthesized according to a previously reported procedure.36 Briefly, CoCl2·6H2O (1.0 g, 4.12 mmol) and bipyridine (99+ %, SigmaAldrich, 2.2 g, 13.94 mmol)) were dissolved in a minimum amount of methanol (100

ACS Paragon Plus Environment

Page 4 of 16

Page 5 of 16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

mL) and refluxed for 2 hours. Once cooled to room temperature, an excess of NH4PF6 (3.4 g, 21 mmol) was added and the precipitate was collected via vacuum filtration. Co(bpy)3(PF6)2 was then dissolved in a minimum amount of acetonitrile and precipitated with ethyl ether 2-3 times to remove excess chlorides. To synthesize the oxidized Co(III) form, Co(bpy)3(PF6)2 (502 mg, 0.6 mmol) was added to a solution of NOBF4 (502 mg, 3.1 mmol) in acetonitrile and stirred for 30 minutes. The solution was rotovapped and the residue was dissolved in a minimum amount of acetonitrile before adding an excess of NH4PF6. The product, Co(bpy)3(PF6)3, was precipitated using ethyl ether and collected and dried via vacuum filtration. No further purification was performed to separate the racemic mixture of cobalt complexes produced. The same method was used for the cobalt complexes using each of the phenanthroline ligands that are illustrated in their CoII form in Figure 1. Cyclic voltammetry was performed on the Co(II) complexes using a USB electrode from Pine Instruments with a pseudo Ag wire reference electrode. Ferrocene was also measured as a reference to calibrate the reference redox potential. CV measurements were performed on a solution of 1-2 mM of the cobalt complex in dry acetonitrile, with 0.1 M tetrabutylammonium perchlorate (Sigma, 99 %) as the supporting electrolyte. A scan rate of 15 mV/s was used. The cyclic voltammogram of the series of cobalt redox couples are shown in Figure 3. Solar cell fabrication. Solar cells were assembled according to a Fig. 3.: CV of a) [Co(5,6-(CH3)2-phen)3]n+ b) 12,15 previously reported method. [Co(bpy) ]n+ c) [Co(phen) ]n+ d) [Co(5-Cl3 3 Fluorine-doped tin oxide (FTO) glass phen)3]n+, e) [Co(5-NO2-phen)3]n+. Ferrocene substrate, used for both the working was used to calibrated Eo (Fc+/Fc)=0.63 V vs NHE.1 A conventional three-electrode system and counter electrodes, was sonicated comprising a glassy carbon electrode (GCE, 3 in a VersaClean solution for 30 minutes, mm diameter) as the working electrode, a Pt rinsed with MilliQ water and ethanol, wire counter electrode and a pseudo Ag reference electrode were used. Cyclic and baked at 500°C for ten minutes to Voltammetric scan was conducted at 15 mV s-1 remove any organic residue. The with 2.44 mV pulse. counter electrode was prepared by dropping an H2PtCl6 solution (2 mg in 1 mL of ethanol) on FTO glass and baking at 400°C for 15 minutes. For the working electrode, the clean FTO glass was immersed in a 40 mM TiCl4 solution at 70°C for 30 minutes, and then rinsed with water and ethanol. A TiO2 film (active layer, 0.1 cm2) was prepared on the glass by a doctor

ACS Paragon Plus Environment

The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

blade technique using TiO2 paste (Solaronix, Ti-Nanoxide T/SP) and baked at 80°C for 1 h and then 500°C for 1 h. A scattering TiO2 layer (Solaronix, Ti-Nanoxide R/SP) was prepared over the active layer by the same procedure. The photoanode with TiO2 was again immersed in 40 mM TiCl4 for 30 min at 70 °C, followed by baking at 500°C for 30 min. The TiO2 films were then sensitized with the different Au25 and Au18 clusters. The sensitizing solutions were prepared by dispersing the Au NC in water and adjusting the pH to 4 using a dilute acetic acid solution. This allows the negatively charged carboxylic group of the glutathione ligand to adsorb onto the TiO2 surface. The working electrodes were submerged in the sensitizing solution for 15 minutes, rinsed with water and ethanol and allowed to dry. The UV-vis of the adsorbed clusters on TiO2 (active only) are shown in Figure 2b. The two electrodes were assembled into a cell by heating with an iron and a Surlyn film (Solaronix, 60 µm thickness) as a spacer until sealed. A drop of the electrolyte solution was placed over a small hole drilled into the Pt-counter electrode and driven into the cell by a vacuum backfilling technique. The hole was sealed using additional Surlyn and a piece of cover glass. Indium metal (99.999 %, Sigma) was soldered to the contact points of both electrodes. The Co(bpy)2+/3+ electrolyte solution for solar cell measurements contained 0.2 M Co2+, 0.033 M Co3+, 0.1 M LiClO4, and 0.5 M 4-tert butylpyridine (TBP). Solubility of cobalt phenanthroline complexes are lower in acetonitrile. So electrolyte solutions are comprised of 0.1 M Co2+, 0.01M Co3+, 0.1 M LiCLO4, 0.2 M TBP in acetonitrile.36 The high performance electrolyte (HPE - I3-/I-) was purchased from Sigma-Aldrich. The photovoltaic performance, IV curves, and photocurrent as a function of time, were evaluated using a SolSimPVC system from Luzchem Inc. The systems meats Class AAA standards for AM 1.5G solar simulation and is equipped with a metrohm PGSTAT204 potentiostat/galvanostat. Tabulated data for the light harvesting efficiency (ƞeff), short circuit current (ISC), and open circuit voltage (VOC), for each of the Au25SR18 and Au18SR14 sensitized solar cells have been reported in Table S1 and S2 in supporting information.

ACS Paragon Plus Environment

Page 6 of 16

Page 7 of 16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry





Fig. 4.: Photovoltaic Performance of (a) Au25SR18, (b) Au25SR18 with 500 nm Long Pass Filter(LPF), (c) Au18SR14 sensitized solar cells and (d) Schematic energy diagram displaying the kinetic processes in the operation of a solar cell. kreg is the regeneration rate of the sensitizer by the CoII/III redox couple.

Results and Discussion Representative IV curves for Au25SR18 sensitized solar cells, under AM 1.5 G illumination, with each of the CoII/III redox mediators, as well as the I-/I3- redox mediator, are shown in Figure 4a. The highest photovoltaic performance for Au25SR18 sensitized cells (highest photocurrent, 1.5 mA/cm2, and highest ηeff =0.54 %) is clearly obtained when using the I-/I3- redox couple (Table S2). The relatively high photocurrent for the Au25SR18 with I-/I3- system is a result of the more negative redox potential for at least the first electron transfer of the I-/I3- couple, and therefore, a very favourable ΔG for electron transfer and regeneration of the sensitizer. However, in the Marcus theory analysis that follows (see below), it is only possible to accurately estimate redox potentials for single electron transfers, and therefore, not for the I-/I3- redox couple. Multiple electron transfer and complex kinetics for the I-/I3- electrolyte render it too complex for analysis of electron transfer rates, and as such were not performed. There is also a negative correlation between photocurrent and open circuit voltage for the series of CoII/III redox couples. This is not surprizing, as the free energy of electron transfer becomes

ACS Paragon Plus Environment

The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 16

smaller as the redox potential of the CoII/III approaches the HOMO energy level of Au25SR18, illustrated in Figure 4d. With the data presented in Figure 4a alone, one might conclude that the HOMO energy level of Au25SR18 is lower than the redox value of all of the CoII/III redox couples. However, when we perform the exact same experiment, but with a 550 nm long pass filter in the illumination path, a very different trend in photovoltaic performance is observed, Figure 4b. It is well known for Au25SR18 clusters that the visible absorbance features above 500 nm are due to HOMO-LUMO transitions, and that the shorter wavelength (higher energy) transitions are due to HOMO-1 and HOMO-2 to LUMO transitions.37 Therefore, by choosing only > 550 nm excitation we selectively excite only the HOMO-LUMO transition. These two experiments illustrate the wavelength-dependent photocatalytic activity of Au25SR18, whereby using full visible spectrum excitation, photocatalytic activity with all five different cobalt electrolytes is observed. However, by limiting the wavelength range to >500 nm, only the cobalt mediators with the highest redox potentials function to provide photocurrent. Photocatalytic activity originating from excitations shorter than 500 nm arise from photo-excited HOMO-x/LUMO transitions, whereas excitations >500 nm are attributed to HOMO/LUMO, or bandedge transitions. We rule out the possibility that it is rapid relaxation of excited electrons by the fact that the I-/I3- electrolyte maintains photocatalytic activity over all excitation wavelengths. Surprisingly, relaxation from the HOMO-x to HOMO energy levels must be slower than dye regeneration for the wavelength dependent photocatalytic activity to be observed. Among the CoII/III redox couples used, the only one capable of providing photocurrent with >500 nm excitation is complex ‘a’, dimethyl-phenanthroline substituted CoII/III. This is the CoII/III with the most negative redox potential of 0.65 eV. There are several very important findings for Au25SR18 demonstrated; (1) the relaxation of the higher energy transitions to the HOMO/LUMO excited state in Au25SR18 is slower than might be expected, and certainly slower than electron transfer to TiO2 (ket) and dye regeneration (kreg), (2) the largest contribution to light-harvesting for the Au25SR18 comes not from HOMO/LUMO excited state, but from higher energy excitations, and (3) the HOMO energy is estimated to be between 0.65 – 0.76 eV, the redox potential of complex ‘a’ and ‘b’. The validity of the estimation of the HOMO energy level can be evaluated using Marcus theory.38 The rate constant for regeneration of the HOMO of oxidized Au25SR18 by cobalt mediators is estimated using Eq 1 - Marcus theory of electron transfer.36, 39-40 41 𝑘"# =

%& ħ



)*+ , -&./0 1

exp (−

(78∆:), -./0 1

ACS Paragon Plus Environment

) (Eq. 1)

Page 9 of 16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

|Had | is the electronic coupling between the donor and acceptor states, ΔG is the reaction (in this case, electron transfer from cobalt redox shuttles to oxidized cluster) free energy, λ is the reorganization energy, kB is Boltzmann's constant and T is the temperature. Figure 5a shows the regeneration rate of the oxidized cluster (kreg) as a function of ΔG. In this study, we model the rapid electron transfer for regeneration of clusters (and corresponding CoII à CoIII). Activation energies are a minimum when λ = ΔG, and electron transfer is maximized for values of λ = 0.7-0.8 eV.36, 40 This estimates electron transfer rates to be greater than the diffusion controlled limit as illustrated by the dotted line in Figure 5a, for minimal free energies of electron transfer (as low as ≃ -0.1eV). Therefore, kreg is limited by diffusion, with a diffusion coefficient of ≃8×10-6 cm2s-1 for cobalt phenanthroline based electrolytes in acetonitrile.36 In short, Marcus theory analysis shows that electron transfer rates are diffusion controlled, and that observation of photocurrents can be used to accurately estimate cluster HOMO levels, and more importantly, that the lack of photocurrent is indicative of HOMO levels of excited clusters that are above the redox potential. From this we are able to reasonable estimate the HOMO level of Au25SR18 to be between 0.65 – 0.71 V vs NHE.



Fig. 5.: (a) Logarithmic plot of electron transfer rate versus free energy for regeneration of oxidized clusters and cobalt polypyridine redox mediators. The drawn curve is a fit according to the Marcus theory (Eq. 1) for regeneration kinetics. The light blue highlighted area on the curve corresponds to the electron transfer rates estimated for λ = 0.7-0.8 eV. The black and green dashed lines show ΔG for electron transfer from [Co(5,6-(CH3)2-phen)3]n+ and [Co(bpy)3]n+ mediators to the HOMO of oxidized cluster, respectively. (b) Energy diagram is illustrating important electron transfer processes governing photocurrent in a photovoltaic. kreg is the regeneration rate of the sensitizer by the redox mediator, krec1 and krec2 are back electron transfer processes for recombination of photoinjected electrons with the oxidized cluster and oxidized redox species, respectively

The Marcus theory analysis, and determination of kreg to be diffusion controlled (~107s-1), allows us to also evaluate electron recombination kinetics. In

ACS Paragon Plus Environment

The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

order to observe significant photocurrent, kreg must also be faster than krec1 and krec2 in Figure 5b. Voltage recovery for MCSSC shows that recombination rates are less than ≃ 10-1s-1 (see supporting information Figure S3). The competitive kinetics of recombination vs regeneration can be evaluated using the regeneration efficiency equation (Eq. 2) for the different cobalt bipyridine redox couples: ∅="> =

/?@A /?@A 8/?@BC

(Eq. 2)

where kreg is the first order rate constant for the regeneration of the cluster by the redox mediator and krec1 is the rate constant for the back electron transfer from the electrons in the TiO2 conduction band to the oxidized cluster molecules (Figure 5b). For example, if we assume that significant photocurrent is observed for a >98 % quantum yield of regeneration, this corresponds to a very reasonable estimate of krec1 of 2x105 s-1.36 To the best of our knowledge, Au18SR14 has provided the best photovoltaic activity of thiol-protected gold clusters to date. Representative IV curves for Au18SR14 sensitized solar cells, under AM1.5 G illumination, and with each of the CoII/III redox mediators, as well as the I-/I3- redox mediator, are shown in Figure 4c. In this case, the highest light harvesting efficiency is observed when employing the CoII/III(bpy)3 redox couple. In fact, the photocurrents observed for Au18SR14 with each of the redox mediators is higher than the corresponding performance for Au25SR18. While the visible light absorption of Au25SR18 is greater than Au18SR14, the higher performance for Au18SR14 has been previously determined to be due to the longer-lived excited state and better electron donating ability of Au18SR14.27 In other words, the visible transition for Au25SR18, while they provide more light absorption, contributes to a very small extent to electron transfer and solar energy harvesting. Importantly, from the relatively high photovoltaic performance for Au18SR14 with all of the redox couples, and as all redox couples induced photocurrent, it can be concluded that the HOMO energy of Au18SR14 is > 0.95 eV, as illustrated in Figure 4d. This is an important finding, as the experimental determination of the HOMO energy of Au18SR14 clusters can be difficult by typical electrochemical techniques. In a typical cyclic voltammetry experiment, clusters are subjected to long-term oxidation to determine the HOMO energy, and reduction to determine the LUMO. Herein, the rapid regeneration of the dye by electron transfer from the redox couple mitigates the redox stress on the stability of clusters. It is interesting to note that the highest ISC values for Au18SR14 MCSSC are not obtained for the I-/I3- or even CoII/III complex ‘a’ with the most favourable ΔG for regeneration. The best overall performance is observed when using the Co(bpy)3II/III complexes ‘b’ as the redox mediator. The discovery of Co(bpy)3II/III complexes marked a major advance for dye-sensitized solar cells. The Co(bpy)3II/III redox

ACS Paragon Plus Environment

Page 10 of 16

Page 11 of 16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

couple has a more positive redox potential than the I-/I3-, and so it can be used to increase VOC and overall efficiency, but it also leads to higher photocurrents here. This may be a result of corrosion of the clusters by the reactive I-, even in the short time before measuring photovoltaic performance. It is clear from the relatively low photocurrents obtained for each of Co complexes (when using Au25(SR)18 sensitized solar cells), that using phenanthroline derivatives as ligands reduce the rate of electron transfer (regeneration of ground state dye) and are a limiting factor in overall light harvesting capacity. For the metal clusters investigated, the Co(bpy)3II/III complex remains the best redox couple for maximizing light harvesting. The long-term stability of clusters as photocatalysts has come into question, especially as a sensitizer for TiO2. It has been shown that the thiol protected gold clusters adhered to TiO2, can readily convert to nanoparticles under visible illumination.19, 35 The conversion to nanoparticles is likely due to the oxidative stress caused by excited state electron transfer from clusters to TiO2, which is the primary electron Fig. 6. Long-term stability of photovoltaic transfer required for photocatalysis performance of Au25SR18 sensitized solar cells, using both high performance and [Co(bpy)3] n+ with metal oxide supports. Figure 6 electrolytes. presents the ISC vs time data for Au25SR18 sensitized cells with both the I- /I3- and Co(bpy)3II/III electrolytes. Previous reports have claimed a superior photovoltaic performance for MCSSC when using the more conventional I-/I3- redox couple.27 However, our findings indicate that the Co(bpy)3II/III complex provides the highest efficiency of solar energy conversion. Furthermore, even if it is possible to obtain higher photocurrents with the I-/I3- electrolyte, corrosion of metal particles with these halides severely reduces the stability of these systems within minutes. The Co(bpy)3II/III complex remains the best redox mediator for light-harvesting with Au clusters. In fact, photocurrents have the potential to be indefinitely stable when using the Co(bpy)3II/III electrolyte system, an extremely encouraging observation in terms of the potential for thiol protected gold clusters as photocatalysts. One of the advantages of clusters as a sensitizer is the long-excited state lifetime, much longer typically than electron transfer, which occurs on the ~1 ns timescale.31 Owing to the unique optical properties, such as long-lived excited states and visible light excitation, thiol protected gold clusters are attractive photoredox catalysis. For example, the HOMO energy level of > 0.95 eV for Au18SR14, determined

ACS Paragon Plus Environment

The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

from photovoltaic measurements, makes this cluster a desirable candidate for photo-oxidation of substrates. The long-term stability in photovoltaic performance indicates that with a suitable electron donor, long photocatalytic activity can be realized. Conclusion The photovoltaic performance of Au25SR18 and Au18SR14 MCSSC has been investigated with the conventional I-/I3- electrolyte and with a series CoII/III complexes as redox mediator with redox potentials spanning a large range. The use of photovoltaic cells as a tool for elucidating fundamental properties of metal clusters has been highlighted. HOMO and HOMO-x energy levels for Au25SR18 were estimated through this method, and most importantly, photocatalytic activity from photo-excited HOMO-x/LUMO transitions is illustrated, which is distinct from the HOMO/LUMO excitation of Au25SR18. An approximate HOMO energy of > 0.95 eV was also determined for Au18SR14, which has been difficult to estimate by conventional electrochemistry. The strong oxidizing potential of Au18SR14, long-term stability, and relatively high solar energy harvesting efficiency of the Au18SR14/Co(bpy)3II/III system are encouraging results when considering this particular cluster for future studies as a photoredox catalyst. Supporting Information. The supporting information contains further information on the absorbance/emission of Au25SR18 and Au18SR14 clusters, photovoltaic performance metrics for each cluster sensitized cell with each of the electrolytes, and further Marcus theory and electron recombination kinetics. This material is available free of charge via the Internet at http://pubs.acs.org. Corresponding Author Kevin G. Stamplecoskie [email protected] Acknowledgements The research presented herein was supported by the Natural Science and Engineering Research Council (NSERC) through the Discovery Grant program award number RGPIN-2016-05070. References 1. Kogo, A.; Sakai, N.; Tatsuma, T., Photoelectrochemical analysis of sizedependent electronic structures of gold clusters supported on TiO2. Nanoscale 2012, 14, 4217-4221.

ACS Paragon Plus Environment

Page 12 of 16

Page 13 of 16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

2. Sakai, N.; Tatsuma, T., Photovoltaic Properties of Glutathione-Protected Gold Clusters Adsorbed on TiO2 Electrodes. Adv. Matter. 2010, 29, 3185-3188. 3. Pernik, D. R.; Tvrdy, K.; Radich, J. G.; Kamat, P. V., Tracking the Adsorption and Electron Injection Rates of CdSe Quantum Dots on TiO2: Linked versus Direct Attachment. J. Phys. Chem. C 2011, 27, 13511-13519. 4. Kamat, P. V., Emergence of New Materials for Light–Energy Conversion: Perovskites, Metal Clusters, and 2-D Hybrids. J. Phys. Chem. Lett. 2014, 23, 41674168. 5. Jin, R., Atomically precise metal nanoclusters: stable sizes and optical properties. Nanoscale 2015, 5, 1549-1565. 6. Zhu, M.; Qian, H.; Jin, R., Thiolate-Protected Au20 Clusters with a Large Energy Gap of 2.1 eV. J. Am. Chem. Soc. 2009, 21, 7220-7221. 7. Zhu, M.; Qian, H.; Jin, R., Thiolate-Protected Au24(SC2H4Ph)20 Nanoclusters: Superatoms or Not? J. Phys. Chem. Lett. 2010, 6, 1003-1007. 8. Yu, Y.; Chen, X.; Yao, Q.; Yu, Y.; Yan, N.; Xie, J., Scalable and Precise Synthesis of Thiolated Au10−12, Au15, Au18, and Au25 Nanoclusters via pH Controlled CO Reduction. Chem. Mater. 2013, 25, 946-952. 9. Huang, T.; Murray, R. W., Luminescence of Tiopronin Monolayer-Protected Silver Clusters Changes To That of Gold Clusters upon Galvanic Core Metal Exchange†. J. Phys. Chem. B 2003, 30, 7434-7440. 10. Goh, J.-Q.; Malola, S.; Häkkinen, H.; Akola, J., Role of the Central Gold Atom in Ligand-Protected Biicosahedral Au24 and Au25 Clusters. J. Phys. Chem. C 2013, 42, 22079-22086. 11. Joshi, C. P.; Bootharaju, M. S.; Alhilaly, M. J.; Bakr, O. M., [Ag25(SR)18]−: The “Golden” Silver Nanoparticle. J. Am. Chem. Soc. 2015, 36, 11578-11581. 12. Bhat, S.; Chakraborty, I.; Baksi, A.; Narayanan, R. P.; Pradeep, T., Evolution of atomically precise clusters through the eye of mass spectrometry. RCS Nanoscience, 2016, 3, 343-385. 13. Wu, Z.; Jin, R., On the Ligand’s Role in the Fluorescence of Gold Nanoclusters. Nano Lett. 2010, 7, 2568-2573. 14. Yousefalizadeh, G.; Stamplecoskie, K. G., Norrish type I photochemistry as a powerful tool in the isolation of thiol protected Au25SR18 clusters. J. Photochem. Photobiol., A 2018, 353, 251-254. 15. Zeng, C.; Liu, C.; Pei, Y.; Jin, R., Thiol Ligand-Induced Transformation of Au38(SC2H4Ph)24 to Au36(SPh-t-Bu)24. ACS Nano 2013, 7, 6138-6145. 16. Kumawat, M. K.; Thakur, M.; Lakkakula, J. R.; Divakaran, D.; Srivastava, R., Evolution of thiol-capped gold nanoclusters into larger gold nanoparticles under electron beam irradiation. J. Photochem. Photobiol., A 2017, 95, 1-6. 17. Chen, Y.-S.; Choi, H.; Kamat, P. V., Metal-cluster-sensitized solar cells. A new class of thiolated gold sensitizers delivering efficiency greater than 2%. J. Am. Chem. Soc. 2013, 24, 8822-8825. 18. Choi, H.; Chen, Y.-S.; Stamplecoskie, K. G.; Kamat, P. V., Boosting the Photovoltage of Dye-Sensitized Solar Cells with Thiolated Gold Nanoclusters. J. Phys. Chem. Lett. 2015, 1, 217-223.

ACS Paragon Plus Environment

The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

19. Liu, S.; Xu, Y.-J., Photo-induced transformation process at gold clusterssemiconductor interface: Implications for the complexity of gold clusters-based photocatalysis. Sci. Rep. 2016, 6, 22742. 20. Kale, M. J.; Avanesian, T.; Christopher, P., Direct Photocatalysis by Plasmonic Nanostructures. ACS Catal. 2014, 1, 116-128. 21. Link, S.; El-Sayed, M.; Schaaff, T.; Whetten, R., Transition from nanoparticle to molecular behaviour: a femtosecond transient absorption study of a size selected 28 atom gold cluster. Chem. Phys. Lett. 2002, 3-4, 240-246. 22. Listorti, A.; O’Regan, B.; Durrant, J. R., Electron Transfer Dynamics in DyeSensitized Solar Cells. Chem. Mater. 2011, 15, 3381-3399. 23. Link, S.; Burda, C.; Wang, Z. L.; El-Sayeda, M. A., Electron dynamics in gold and gold–silver alloy nanoparticles: The influence of a nonequilibrium electron distribution and the size dependence of the electron–phonon relaxation. J. Chem. Phys. 1999, 3, 1255-1264. 24. Chen, Y. S.; Kamat, P., Glutathione Capped Gold Nanoclusters as Photosensitizers. Visible Light Induced Hydrogen Generation in Neutral Water. J. Am. Chem. Soc. 2014, 6, 6075-6082. 25. Stamplecoskie, K. G.; Swint, A., Optimizing molecule-like gold clusters for light energy conversion. J. Mater. Chem. A 2016, 4, 2075-2081. 26. Chen, Y. S.; Choi, H.; Kamat, P. V., Metal-Cluster-Sensitized Solar Cells. A New Class of Thiolated Gold Sensitizers Delivering Efficiency Greater than 2%. J. Am. Chem. Soc. 2013, 135 (24), 8822-5. 27. Abbas, M. A.; Kim, T.-Y.; Lee, S. U.; Kang, Y. S.; Bang, J. H., Exploring Interfacial Events in Gold-Nanocluster-Sensitized Solar Cells: Insights into the Effects of the Cluster Size and Electrolyte on Solar Cell Performance. J. Am. Chem. Soc. 2016, 138 1, 390-401. 28. Boschloo, G.; Hagfeldt, A., Characteristics of the Iodide/Triiodide Redox Mediator in Dye-Sensitized Solar Cells. Acc. Chem. Res. 2009, 11, 1819-1826. 29. Hesari, M.; Workentin, M. S.; Ding, Z., Near-infrared electrochemiluminescence from Au25(SC2H4Ph)18+ clusters co-reacted with tri-npropylamine. RSC Advances 2014, 56, 29559-29562. 30. García-Raya, D.; Madueño, R.; Blázquez, M.; Pineda, T., Electrochemistry of Molecule-like Au25 Nanoclusters Protected by Hexanethiolate. J. Phys. Chem. C 2009, 20, 8756-8761. 31. Stamplecoskie, K. G.; Chen, Y. S.; Kamat, P. V., Excited-State Behavior of Luminescent Glutathione-Protected Gold Clusters. J. Phys. Chem. C 2014, 118, 13701376. 32. Ghosh, A., Thumu Udayabhaskararao, and Thalappil Pradeep., One-step route to luminescent Au18SG14 in the condensed phase and its closed shell molecular ions in the gas phase. J. Phys. Chem. Lett. 2012, 3, 1997–2002. 33. Liu, Q.; Zhou, Z.; Qiu, G.; Li, J.; Xie, J.; Lee, J. Y., Surface Reaction Route To Increase the Loading of Antimicrobial Ag Nanoparticles in Forward Osmosis Membranes. ACS Sustainable Chem. Eng. 2015, 3, 2959–2966. 34. Yoshiki Niihori, a. C. U., a Wataru Kurashigea and Yuichi Negishi, Highresolution separation of thiolate-protected gold clusters by reversed-phase highperformance liquid chromatography. J. Phys. Chem. 2016, 6, 4251-4265

ACS Paragon Plus Environment

Page 14 of 16

Page 15 of 16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

35. Stamplecoskie, K. G.; Kamat, P. V., Size-Dependent Excited State Behavior of Glutathione-Capped Gold Clusters and Their Light-Harvesting Capacity. J. Am. Chem. Soc. 2014, 31, 11093-11099. 36. Feldt, S. M.; Wang, G.; Boschloo, G.; Hagfeldt, A., Effects of driving forces for recombination and regeneration on the photovoltaic performance of dye-sensitized solar cells using cobalt polypyridine redox couples. J. Phys. Chem. C 2011, 43, 2150021507. 37. Zhu, M.; Aikens, C. M.; Hollander, F. J.; Schatz, G. C.; Jin, R., Correlating the crystal structure of a thiol-protected Au25 cluster and optical properties. J. Am. Chem. Soc. 2008, 18, 5883-5. 38. Marcus, R. A., Electron transfer reactions in chemistry. Theory and experiment. Rev. Mod. Phys. 1993, 3, 599. 39. Kaneza, N.; Zhang, J.; Liu, H.; Archana, P. S.; Shan, Z.; Vasiliu, M.; Polansky, S. H.; Dixon, D. A.; Adams, R. E.; Schmehl, R. H., Electrochemical and spectroscopic properties of boron dipyrromethene–thiophene–triphenylamine-based dyes for dye-sensitized solar cells. J. Phys. Chem. C 2016, 17, 9068-9080. 40. Ørnsø, K. B.; Jó nsson, E. O.; Jacobsen, K. W.; Thygesen, K. S., Importance of the reorganization energy barrier in computational design of porphyrin-based solar cells with cobalt-based redox mediators. J. Phys. Chem. C 2015, 23, 12792-12800. 41. Kaduk, B.; Kowalczyk, T.; Van Voorhis, T., Constrained density functional theory. Chem. Rev. 2011, 1, 321-370. TOC Graphic



ACS Paragon Plus Environment

The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60



ACS Paragon Plus Environment

Page 16 of 16