COMMUNICATION pubs.acs.org/JACS

Biomimetic Multifunctional Porous Chalcogels as Solar Fuel Catalysts Benjamin D. Yuhas,† Amanda L. Smeigh,† Amanda P. S. Samuel,† Yurina Shim,† Santanu Bag,† Alexios P. Douvalis,‡ Michael R. Wasielewski,† and Mercouri G. Kanatzidis*,†,§ †

Department of Chemistry and Argonne-Northwestern Solar Energy Research (ANSER) Center, Northwestern University, Evanston, Illinois 60208-3113, United States ‡ Department of Physics, University of Ioannina, 45110 Ioannina, Greece § Materials Science Division, Argonne National Laboratory, Argonne, Illinois 60439, United States

bS Supporting Information ABSTRACT: Biological systems that can capture and store solar energy are rich in a variety of chemical functionalities, incorporating light-harvesting components, electron-transfer cofactors, and redox-active catalysts into one supramolecule. Any artificial mimic of such systems designed for solar fuels production will require the integration of complex subunits into a larger architecture. We present porous chalcogenide frameworks that can contain both immobilized redox-active Fe4S4 clusters and light-harvesting photoredox dye molecules in close proximity. These multifunctional gels are shown to electrocatalytically reduce protons and carbon disulfide. In addition, incorporation of a photoredox agent into the chalcogels is shown to photochemically produce hydrogen. The gels have a high degree of synthetic flexibility, which should allow for a wide range of lightdriven processes relevant to the production of solar fuels.

T

he quest for renewable energy production from sunlight has led to an intense research effort in the field of solar fuels. Many approaches have focused on mimicking naturally occurring processes such as photosynthesis,1 where light-driven water splitting by the Photosystem II reaction center protein yields both oxygen and protons. In contrast, biological proton reduction to hydrogen is catalyzed by hydrogenases that generally occur in non-photosynthetic organisms, and recent research on these enzymes largely focuses on their hydrogen-producing capabilities,2 particularly as a function of different environments, such as one might encounter in nature3 or in a purely engineered design,4 with an applied electrochemical or photochemical potential driving the generation of hydrogen from protons. Using proteins and other biological structures as starting points, researchers have developed analogous inorganic and organometallic molecules that can catalyze hydrogen production from water.5 Many of these compounds have been shown to be reasonably successful at short-term hydrogen production,6 yet the major limitation of these molecules arises from catalyst deactivation, primarily by oxygen, which is both the other product of the water-splitting reaction and a major component of the atmosphere. It is clear, then, that any design motif that incorporates active catalysts into a larger framework that can serve to protect the catalysts from the adverse effects of oxygen may significantly improve long-term catalyst stability. r 2011 American Chemical Society

A newly emerging class of porous chalcogenide aerogels, or “chalcogels”, may be an ideal supramolecular structure for the integration of redox-active cofactors relevant to solar fuels production.7 Chalcogels are highly porous materials that, unlike the ubiquitous oxide-based aerogels,8 are based on chalcogenide materials such as sulfides, selenides, and tellurides. This allows for the possibility of strong visible light absorption, as well as a number of other interesting properties, such as desulfurization catalysis9 and heavy metal ion sequestration.7 Here, we present a new type of chalcogel that contains both redox-active transition metal clusters and light-harvesting photoredox molecules. The transition metal clusters are Fe4S4 cubane clusters that are known redox-active cofactors in enzymes. Synthetic molecular analogues of these bioinorganic Fe4S4 cofactors have existed for some time.10 We show that the Fe4S4 cubane clusters retain their redox activity when incorporated into a larger chalcogenide gel framework and that their redox states can be switched electrochemically. We also demonstrate the catalytic potential of these Fe4S4-functionalized chalcogels through electrocatalytic reduction of protons, provided by the weak organic acid lutidinium, and carbon disulfide (CS2), a more-readily reduced surrogate for CO2. Finally, we show that cationic light-harvesting photoredox dyes can be incorporated into the chalcogel framework by a simple ion-exchange process, that the excited states of these photoredox dye molecules are strongly quenched by the presence of the Fe4S4 cubane clusters in the gels, and that the dye-functionalized gels are capable of photochemically producing hydrogen. The chalcogels are multifunctional materials with high porosity that can be engineered and functionalized to have all of the components required for light-driven chemical catalysis, providing new integrated materials for solar fuels production. The slow, controlled metathesis reaction between the precursors Na4Sn2S6 3 14H2O and (Ph4P)2[Fe4S4Cl4] results in replacement of the terminal chloride ligands on the iron-sulfur cubane cluster with sulfur atoms from the tin sulfide cluster (Scheme 1), giving a polymerized network that condenses in a solid, spongy gel (Figure 1A,B). Electron microscopy reveals the spongy, porous nature of these chalcogels, which always appear amorphous under TEM as well as in capillary X-ray diffraction experiments. With multiple binding points on each Fe4S4 cluster, the reaction can occur at up to four points per cluster, as suggested by the equation

Received: December 23, 2010 Published: March 16, 2011 7252

dx.doi.org/10.1021/ja111275t | J. Am. Chem. Soc. 2011, 133, 7252–7255

Journal of the American Chemical Society

COMMUNICATION

Scheme 1. Reaction between [Sn2S6]4- and [Fe4S4Cl4]2Cluster Anions and Functionalization of the Framework with Light-Harvesting Ru(bpy)32þ

½Fe4 S4 Cl4 2- þ ½Sn2 S6 4- f f½Fe4 S4 x ½Sn2 S6 y gm- þ yCl- ð1Þ

The porous structures are formed without the use of templating agents. The pore sizes and porosity can be described as a mixture of macroporous and mesoporous structures, as evidenced by the nitrogen adsorption isotherms (Figure 1C), which exhibit Type IIb behavior.11 The resulting estimated surface areas range from 90 to 290 m2/g. Elemental analysis (Figure 1D) of the chalcogels consistently shows, in addition to all of the expected elements, signatures arising from the counter-cations of the precursors, Ph4Pþ and Naþ. This indicates that the chalcogel network is overall negatively charged ({[Fe4S4]x[Sn2S6]y}m-, eq 1 and Scheme 1). The remaining charge is balanced by either Ph4Pþ or Naþ. Characterization of the Fe/S clusters inside the chalcogels was accomplished by UV-vis and M€ossbauer spectroscopies. Although the chalcogels are stable in nearly all solvents with no observable leaching, when a large excess of benzenethiol is added to the solution, the thiol extrudes the Fe/S clusters from the gel framework, and eventually a uniform solution results. The characteristic UV-vis spectra in N,N0 -dimethylformamide (DMF) are shown in Figure 1F, with the well-known12 absorption maximum of [Fe4S4(SPh)4]2appearing at ∼460 nm. This is also a well-established method of detecting Fe4S4 clusters in their native proteins.13 Figure 1G,H shows zero-field 57Fe M€ossbauer spectra acquired at 40 and 10 K, respectively. At 40 K, the spectrum is characterized by two very closely spaced quadrupole doublets, suggesting a high degree of equivalence among the Fe centers of the clusters in the gels. The isomer shifts (δ) of the two doublets are found to be 0.49 and 0.47 mm/s (relative to R-Fe), with respective quadrupole splitting parameters (ΔEq) of 1.07 and 0.66 mm/s. This finding is consistent with previous M€ossbauer investigations on isolated, molecular analogues of the Fe4S4 cluster,14 as well as on native Fe4S4-bearing proteins, such as ferredoxins.15 At 10 K, however, we begin to observe a magnetic hyperfine splitting of the M€ossbauer signal, despite the absence of an applied magnetic field. This could be the effect of a residual paramagnetic state in the gels, or perhaps the effect of intercluster coupling within the chalcogels. In native proteins, the Fe4S4 cubane moiety is sometimes found in chains to facilitate electron transfer, such as in bacterial ferredoxins.15 However, to our knowledge, no zero-field M€ossbauer magnetic splitting has been reported in systems

Figure 1. (A,B) SEM and TEM images of a typical Fe4S4-Sn2S6 chalcogel. Insets: (A) real-size image of the chalcogel; (B) representative SAED pattern of the chalcogels. (C) Nitrogen adsorption/desorption isotherm at 77 K of the chalcogel. (D) Typical EDS spectrum of the chalcogels. (E) XRD patterns of the chalcogels, illustrating their amorphous character. (F) UVvis spectra of DMF solutions of chalcogels that have been exposed to excess benzenethiol, reflecting the absorption of the extruded [Fe4S4(SPh)4]2anion. The spectrum of the precursor (Ph4P)2[Fe4S4Cl4] is included for comparison. (G,H) M€ossbauer spectra of Chalcogel-1 at 40 and 10 K.

containing multiple Fe4S4 clusters. Nevertheless, it should be noted that the possibility exists for the two or more Fe4S4 clusters to be in very close proximity (i.e., separated by a single [Sn2S6]4unit), which could enable intercluster coupling. The chalcogels’ Fe4S4 cluster redox activity was examined with cyclic voltammetry (Figure 2). All CV experiments were performed on the chalcogels in the solid state. Sweeping the potential in a reductive direction revealed an initial reduction near -800 to -900 mV (vs Ag/AgCl) and a second event further negative, near -1500 to -1600 mV. The redox potentials of the Fe4S4 cubane clusters in solution have been well-studied,12 but there are comparatively fewer literature reports on the redox activity of these types of clusters in bound environments,16 aside from the native proteins themselves. The first reduction at -900 mV is assigned to [Fe4S4]2þ/þ core cluster reduction. The potential at which this occurs is in good agreement with previous studies in solution. While the first reduction wave exhibits reversible behavior in many of these solution-based studies, in our case the reduction is decidedly quasi-reversible, if not irreversible. Upon reversing the direction of the scan, an anodic wave is observed at least 600 mV more positive than the initial reduction wave. This large separation between the initial reduction and subsequent re-oxidation waves is nearly unprecedented and suggests either a capacitive effect in the gel or perhaps a charge transfer into the gel backbone. Furthermore, we observe that there is little to no dependence of the CV curves on solvent, electrolyte, or scan rate. This suggests that what we are measuring is due entirely to the solid gel on the surface of the electrode and is not a result of Fe4S4 clusters leaching into the electrolyte solution. The second observed reduction wave could be attributed to the [Fe4S4]þ/0 reduction. In solution, this reduction is always irreversible.17 However, the all-ferrous state of the cluster could be more easily stabilized when bound in the chalcogel. When the gels are prepared without Fe4S4 clusters, we see no comparable redox events (Figure 2B). 7253

dx.doi.org/10.1021/ja111275t |J. Am. Chem. Soc. 2011, 133, 7252–7255

Journal of the American Chemical Society

Figure 2. (A) CVs of Fe4S4-Sn2S6 chalcogel at varying scan rates. (B) Various chalcogel CVs scanned at 100 mV/s. Only the chalcogels synthesized with the Fe4S4 clusters exhibit redox activity. (C,D) CVs of chalcogels with lutidinium chloride (C) and CS2 (D) as substrates dissolved in solution. The scan rate is 20 mV/s in each case.

The reproducible redox activity of these Fe4S4-containing chalcogels coupled with their high degree of porosity make them excellent electrocatalyst candidates. We explored this possibility by probing the electrochemical reduction of two substrates, lutidinium ion and CS2, in the presence of the chalcogel-modified electrodes. Lutidinium was chosen because isolated Fe4S4 clusters in solution have been shown to reduce protons from weak organic acids such as lutidinium, pyridinium, and other similar sources,18 whereas CS2 was examined as a more easily reduced analogue of CO2, which has also been shown to undergo enhanced electrochemical reduction in the presence of artificial Fe4S4bearing molecules.19 The latter experiments involved free, isolated clusters in solution, not bound into a larger supramolecular framework as is the case in our chalcogels. Figure 2C shows the CVs for electrochemical reduction of lutidinium cations in acetonitrile in the presence and absence of Fe4S4-Sn2S6 chalcogels. A modest, ∼6-fold increase in current of the reduction wave is seen when the chalcogel is present on the working electrode, as compared to when no lutidinium is present in the electrolyte. Additionally, there is a clear shift in the potential (∼200 mV) at which reduction occurs when the chalcogel is present as opposed to a bare working electrode. A similar set of experiments is presented in Figure 2D, this time using CS2 as the substrate in DMF. In this case, the ∼10-fold current increase with the chalcogel present is greater than that observed in the lutidinium experiments. There is also an anodic shift in the potential at which the current begins to rise rapidly (∼200 mV). Further experiments with these and other substrates are ongoing, but our results suggest that the chalcogels could be effective electrocatalysts for the reduction of substrates relevant to solar fuels. As mentioned above, elemental analysis of the fully washed chalcogels always reveals a persistent quantity of the countercation associated with the chalcogel precursors. This indicates that the metal sulfide framework is anionic, which allows for the possibility for ion-exchange, specifically the exchange of the counter-cations such as light-harvesting photoredox dyes. Thus, to examine photodriven electron-transfer chemistry and potential photocatalysis in chalcogels, we functionalized chalcogels with Ru(bpy)32þ. Solution-based ion exchange results in the dye molecules displacing the existing cations (Naþ and Ph4Pþ) and remaining electrostatically bound to the chalcogel surface. Dye functionalization is confirmed by FTIR and EDS spectroscopy

COMMUNICATION

(Supporting Information (SI), Figure S1). No leaching of the dyes was observed. The dye-functionalized gels are placed in o-dichlorobenzene and sonicated, resulting in a thick slurry that is spuncast onto a glass substrate to give thin films of the gels. The Ru(bpy)32þ-functionalized films were then examined by transient absorption spectroscopy (for full details, see the SI). All of the samples show a clear bleaching of the ground state at 460 nm, and both gel samples exhibit multiphasic recovery of the Ru(bpy)32þ ground-state bleach. However, the amount of residual bleach remaining on a sub-nanosecond time scale is significantly larger in the Zn:Sn2S6 control gel than in the Fe4S4:Sn2S6 gel (48% and 11%, respectively). In the Ru(bpy)32þ-functionalized Fe4S4: Sn2S6 gels, nearly 90% of the excited state is quenched within a few picoseconds. This stands in stark contrast to Ru(bpy)32þ in solution, which exhibits essentially zero decay of the excited state on the time scale of interest because of its intrinsic 600 ns excitedstate lifetime. Figure 3A shows the decay of the ground-state bleach at 460 nm of Ru(bpy)32þ in both the Fe4S4:Sn2S6 and Zn:Sn2S6 gels, as well as Ru(bpy)32þ in solution. The rapid decay at 460 nm, on the <10 ps time scale, indicates an additional deactivation pathway is present in the Fe4S4:Sn2S6 gel, which is absent in the Zn: Sn2S6 gel. The residual, unquenched signal in the Zn:Sn2S6 gel is likely associated with the Ru(bpy)32þ excited state, as evidenced by nanosecond transient absorption measurements (SI, Figure S2). On the nanosecond time scale, an emissive feature at 620 nm is observed in both solution and the control Zn:Sn2S6 gel but is absent in the Fe4S4:Sn2S6 gel. The exact mechanism of excited-state quenching is still under investigation. Since we cannot observe any distinguishing spectroscopic features of Ru(bpy)33þ or Ru(bpy)3þ, it is difficult to say if quenching occurs primarily by oxidative or reductive pathways. However, the possibility of oxidative quenching of Ru(bpy)32þ by the Fe4S4 cluster is supported by an approximate calculation of the free energy for photodriven electron transfer from the photogenerated metal-to-ligand charge-transfer state of Ru(bpy)32þ to the Fe4S4 cluster using the Weller equation,20 which gives a favorable free energy change for this process of about ΔG = -90 meV. The proposed energy level diagram, constructed from the electrochemical and photophysical measurements, is shown in Figure 3B. Although ΔG is small, the Coulombic attraction between the anionic, anchored Fe4S4 clusters and the cationic Ru(bpy)32þ photoredox dye means that we expect these two functional groups to be positioned very close to each other, which should facilitate electron transfer by increasing the electronic coupling between them. As a practical demonstration of the potential photocatalytic capability of our photoredox dye-functionalized chalcogels, photochemical hydrogen evolution experiments were performed (Figure 3C). The solid Ru(bpy)32þ-functionalized gels were placed in an acetonitrile/water (70:30 v/v) solution containing 50 mM lutidinium chloride and 50 mM sodium ascorbate as a sacrificial electron donor and then continuously irradiated with a xenon lamp (λ > 300 nm). After 24 h, H2 gas was detected by gas chromatography, and the amount of H2 produced increased steadily over 4 days of illumination. There was no detectable H2 evolution when the dye-functionalized gels were kept in darkness or when a gel-free solution containing lutidine, ascorbate, and Ru(bpy)32þ was illuminated. Irradiation of control gels lacking Ru(bpy)32þ produced only a very small amount of H2 that only was observable above the baseline after about 2 days. The amount of H2 produced by the dye-functionalized gels was about 8-fold greater than that generated by gels with no photoredox dye present. Further 7254

dx.doi.org/10.1021/ja111275t |J. Am. Chem. Soc. 2011, 133, 7252–7255

Journal of the American Chemical Society

COMMUNICATION

’ ACKNOWLEDGMENT We thank Prof. Joseph Hupp for use of the potentiostat. Electron microscopy and elemental analysis were performed at the Electron Probe Instrumentation Center at Northwestern University. This work was supported as part of the ANSER Center, an Energy Frontier Research Center funded by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences, under Award No. DE-SC0001059. ’ REFERENCES

Figure 3. (A) Femtosecond transient absorption at 460 nm of Ru(bpy)32þ dye in three different environments: free in solution, in a control Zn:Sn2S6 gel, and in the redox-active Fe4S4-Sn2S6 gel. Within a few picoseconds, almost 90% of the excited state is quenched by the Fe4S4 clusters. (B) Proposed energy diagram for the Ru(bpy)32þ dye in the Fe4S4-Sn2S6 chalcogels, as calculated from cyclic voltammetry and spectroscopic experiments. (C) Hydrogen evolution for a variety of gels and conditions. The volume of H2 evolved is normalized to the dry mass of the chalcogels, where applicable. (D) Hydrogen evolution of Ru(bpy)32þ-functionalized gels, showing the beneficial effect of adding extra dye (100 mM) into the solution.

enhancement of the H2 output is observed when extra Ru(bpy)32þ is dissolved in the gel-containing solutions with the lutidinium and ascorbate (Figure 3D). Although the overall H2 output is rather small (∼0.2 mol % based on [Fe4S4]), it is reasonable to envision that through further optimization, such as manipulation of the reduction potentials of the clusters and/or the photoredox dyes, photochemical H2 production could be greatly enhanced. The chalcogel system presented here allows us to fabricate multifunctional and chemically integrated materials using a relatively facile synthetic method. We have shown that porous frameworks with controllable and tunable light absorption can be made to include both redox-active components important to catalysis and light-harvesting photoredox components necessary for solar energy conversion. The presence of the biomimetic redox-active clusters in the gels not only facilitates electrochemical reductions of various substrates but also provides a means of positioning photoredox molecules sufficiently close to the redox-active Fe4S4 clusters to promote rapid excited-state quenching of the photoredox dye excited states. The photoredox dye-functionalized gels are capable of producing hydrogen under photochemical conditions. The chalcogels have the distinct advantage of being easy to scale up as well as providing spatial separation between active redox groups. This could ultimately lead to a self-assembling, fully synthetic material that encompasses and emulates all of the active materials of biological systems involved in solar energy conversion.

’ ASSOCIATED CONTENT

bS

Supporting Information. Experimental and characterization details and spectra. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

[email protected]

(1) (a) Ferreira, K. N.; Iverson, T. M.; Maghlaoui, K.; Barber, J.; Iwata, S. Science 2004, 303, 1831. (b) Youngblood, W. J.; Lee, S.-H. A.; Maeda, K.; Mallouk, T. E. Acc. Chem. Res. 2009, 42, 1966. (c) Gust, D.; Moore, T. A.; Moore, A. L. Acc. Chem. Res. 2009, 42, 1890. (2) Vincent, K. A.; Parkin, A.; Armstrong, F. A. Chem. Rev. 2007, 107, 4366. (3) (a) Ihara, M.; Nishihara, H.; Yoon, K.-S.; Lenz, O.; Friedrich, B.; Nakamoto, H.; Kojima, K.; Honma, D.; Kamachi, T.; Okura, I. Photochem. Photobiol. 2006, 82, 676. (b) Fouchard, S.; Hemschemeier, A.; Caruana, A.; Pruvost, J.; Legrand, J.; Happe, T.; Peltier, G.; Cournac, L. Appl. Environ. Microbiol. 2005, 71, 6199. (4) (a) Dementin, S.; Belle, V.; Bertrand, P.; Guigliarelli, B.; Adryanczyk-Perrier, G.; De Lacey, A. L.; Fernandez, V. M.; Rousset, M.; Leger, C. J. Am. Chem. Soc. 2006, 128, 5209. (b) Jones, A. K.; Sillery, E.; Albracht, S. P. J.; Armstrong, F. A. J. Chem. Soc., Chem. Commun. 2002, 866. (5) (a) Tard, C.; Liu, X.; Ibrahim, S. K.; Bruschi, M.; De Gioia, L.; Davies, S. C.; Yang, X.; Wang, L.-S.; Sawers, G.; Pickett, C. J. Nature 2005, 433, 610. (b) Schmidt, M.; Contakes, S. M.; Rauchfuss, T. B. J. Am. Chem. Soc. 1999, 121, 9736. (c) Evans, D. J.; Pickett, C. J. Chem. Rev. 2003, 32, 268. (6) (a) Gloaguen, F.; Lawrence, J. D.; Rauchfuss, T. B. J. Am. Chem. Soc. 2001, 123, 9476. (b) Christou, G.; Hageman, R. V.; Holm, R. H. J. Am. Chem. Soc. 1980, 102, 7600. (7) Bag, S.; Trikalitis, P. N.; Chupas, P. J.; Armatas, G. S.; Kanatzidis, M. G. Science 2007, 317, 490. (8) Husing, N.; Schubert, U. Angew. Chem., Int. Ed. 1998, 37, 22. (9) Bag, S.; Gaudette, A. F.; Bussell, M. E.; Kanatzidis, M. G. Nat. Chem. 2009, 1, 217. (10) Rao, P. V.; Holm, R. H. Chem. Rev. 2004, 104, 527. (11) Sing, K. S. W.; Everett, D. H.; Haul, R. A. W.; Moscou, L.; Pierotti, R. A.; Rouquerol, J.; Siemieniwska, T. Pure Appl. Chem. 1985, 57, 603. (12) (a) Coucouvanis, D.; Kanatzidis, M. G.; Simhon, E.; Baenziger, N. C. J. Am. Chem. Soc. 1982, 104, 1874. (b) Wong, G. B.; Bobrik, M. A.; Holm, R. H. Inorg. Chem. 1978, 17, 578. (13) Moulis, J. M.; Meyer, J. Biochemistry 1982, 21, 4762. (14) (a) Holm, R. H.; Averill, B. A.; Herskovitz, T.; Frankel, R. B.; Gray, H. B.; Siiman, O.; Grunthaner, F. J. J. Am. Chem. Soc. 1974, 96, 2644. (b) Papaefthymiou, V.; Millar, M. M.; Munck, E. Inorg. Chem. 1986, 25, 3010. (15) (a) Surerus, K. K.; Chen, M.; van der Zwaan, J. W.; Rusnak, F. M.; Kolk, M.; Duin, E. C.; Albracht, S. P. J.; Muenck, E. Biochemistry 1994, 33, 4980. (b) Christner, J. A.; Munck, E.; Janick, P. A.; Siegel, L. M. J. Biol. Chem. 1981, 256, 2098. (16) Gorman, C. B.; Smith, J. C.; Hager, M. W.; Parkhurst, B. L.; Sierzputowska-Gracz, H.; Haney, C. A. J. Am. Chem. Soc. 1999, 121, 9958. (17) Cambray, J.; Lane, R. W.; Wedd, A. G.; Johnson, R. W.; Holm, R. H. Inorg. Chem. 1977, 16, 2565. (18) Henderson, R. A. Chem. Rev. 2005, 105, 2365. (19) (a) Tezuka, M.; Yajima, T.; Tsuchiya, A.; Matsumoto, Y.; Uchida, Y.; Hidai, M. J. Am. Chem. Soc. 1982, 104, 6834. (b) Tomohiro, T.; Uoto, K.; Okuno, H. J. Chem. Soc., Chem. Commun. 1990, 194. (20) Morandeira, A.; Fortage, J.; Edvinsson, T.; Le Pleux, L.; Blart, E.; Boschloo, G.; Hagfeldt, A.; Hammarstr€om, L.; Odobel, F. J. Phys. Chem. C 2008, 112, 1721. 7255

dx.doi.org/10.1021/ja111275t |J. Am. Chem. Soc. 2011, 133, 7252–7255

Supporting Information

Biomimetic Multifunctional Porous Chalcogels as Solar Fuel Catalysts

Benjamin D. Yuhas1, Amanda L. Smeigh1, Amanda P. S. Samuel1, Yurina Shim1, Alexios P. Douvalis2, Michael R. Wasielewski1, Mercouri G. Kanatzidis*,1,3 1) Department of Chemistry and Argonne-Northwestern Solar Energy Research (ANSER) Center, Northwestern University, Evanston, IL 60208-3113 2) Department of Physics, University of Ioannina, 45110 Ioannina, Greece 3) Materials Science Division, Argonne National Laboratory, Argonne, IL 60439

S1

Experimental Methods Synthesis of Na4Sn2S6·14H2O. The Na4Sn2S6·14H2O was synthesized using a modification of the literature method1. 60 mmol of Na2S·9H2O (14.4 g) was dissolved in 100 mL of water with vigorous stirring. 20 mmol (7.01 g) of SnCl4·5H2O is dissolved in a minimum amount of water (~5 mL) and slowly added dropwise to the Na2S solution with vigorous stirring. Yellow flakes form upon addition of the SnCl4, which eventually dissolve to form a transparent yellow solution. After addition of all the SnCl4, the mixture is covered and stirred overnight at room temperature. The clear yellow solution is then slowly added to 300 mL acetone with vigorous stirring. The acetone solution goes from clear to blue to green, and ultimately a viscous yellow oil is seen on the bottom of the flask with a white translucent solution on top. The flask is placed in a refrigerator for 2 days, and then removed. At this point, the top layer is now transparent. The top layer is decanted away, and 100 mL of acetone is added to the yellow oil with vigorous stirring. The mixture is then manually shaken until a white precipitate forms. The flask is placed back in the refrigerator for 24 h, and then the precipitate is filtered, washed with acetone, and dried under vacuum. The synthesis of (Ph4P)2[Fe4S4Cl4] was Synthesis of (Ph4P)2[Fe4S4Cl4]. performed in a nitrogen-filled glovebox following the established procedures2,3. 1.0 g FeCl2 (7.89 mmol), 1.56 NaSPh (11.80 mmol), 1.48 g Ph4PCl (3.94 mmol), and 0.32 g S (9.86 mmol) were dissolved in 40 mL acetonitrile and stirred for 45 minutes. The brown solution was filtered, and 200 mL of ether was added to the filtrate and allowed to stand overnight. Then, the solution was filtered and washed with ether, yielding a black crystalline product. These black crystals were re-dissolved in acetonitrile and filtered three times to remove excess NaCl. An excess of ether was added to the resulting amberbrown solution and again allowed to stand overnight. The resulting black crystals were filtered, washed with ether, and allowed to dry under nitrogen. Synthesis of Chalcogels. Chalcogel synthesis was performed in a nitrogen-filled glovebox. In a typical synthesis, 80 mg (0.1 mmol) of Na4Sn2S6·14H2O was dissolved in 9 mL of formamide in a scintillation vial. In a separate vial, 120 mg (0.1 mmol) of The (Ph4P)2[Fe4S4Cl4] was dissolved in 2 mL of N,N-dimethylformamide. (Ph4P)2[Fe4S4Cl4] solution was then added dropwise to the Na4Sn2S6·14H2O solution. The Na4Sn2S6 solution is shaken manually after each drop of (Ph4P)2[Fe4S4Cl4] is added, and the entire addition occurs over 30-45 minutes. When the addition of (Ph4P)2[Fe4S4Cl4] is complete, a viscous black liquid results. This liquid is poured into a clean vial, covered, and left undisturbed at 25°C for 7-14 days. After this time, the viscous liquid has solidified into a black gel. The gel is subjected to cleaning by solvent exchange with a 4:1 (v:v) mixture of ethanol and water (3x), followed by pure ethanol (3x). The solvent is exchanged every 24 hours, so that the entire cleaning procedure took about a week. Electron Microscopy and Elemental Analysis. Scanning Electron Microscopy was performed on a Hitachi S4800N-II instrument equipped with an energy dispersive spectroscopy (EDS) detector. Transmission Electron Microscopy was performed on a JEOL 2100F instrument equipped with an EDS detector. For TEM experiments, the

S2

powdered gel samples are soniccated in ether, and the resulting suspension is dropcast onto a holey carbon TEM grid. For EDS analysis, the chalcogels were examined in two forms. In one case, a slice of the fully cleaned wet gel (“alcogel”) is allowed to slowly dry out under nitrogen, resulting in a xerogel which is ground into a fine powder that is used for the imaging and EDS analysis. In the second case, the samples were analyzed as aerogels that had undergone critical point drying (see below). No significant change in elemental analysis was observed when using either SEM or TEM, or when examining a xerogel or aerogel sample. Critical Point Drying and Nitrogen Adsorption. A Critical Point Drying apparatus (Tousimis Autosamdri 815B Series A) was used to remove the solvent from the fully cleaned alcogels. Slices of the alcogel were placed into the critical point dryer, and the chamber was cooled with liquid carbon dioxide to 5°C. Liquid CO2 was then introduced into the chamber to exchange out the remaining ethanol solvent. Fresh liquid CO2 was introduced into the chamber every 20-30 minutes over the course of 7-8 hours in order to completely remove all traces of ethanol. Then, the supercritical drying was done at elevated temperature and pressure (41°C, 1400 psi) for 4 minutes, followed by the gaseous CO2 being slowly bled out of the chamber. The black aerogels are then immediately placed under nitrogen to minimize oxidation from the air. The aerogels can then be used for surface area measurements by nitrogen adsorption. The aerogels are placed in a sample tube and degassed at 75°C under vacuum overnight. Nitrogen adsorption/desorption isotherms were obtained at 77 K on a Micromeritics ASAP 2020 instrument. The surface area was measured using the Brunauer-Emmett-Teller (BET) model using a series of relative pressures P/P0 from 0.05-0.30. Mössbauer spectroscopy. 57Fe Mössbauer spectra were collected in transmission geometry, using a constant acceleration spectrometer equipped with a 57Co(Rh) source kept at room temperature. The Mössbauer sample holders were prepared and sealed in strict glove-box conditions under N2 atmosphere, and then transferred to a closed loop Mössbauer cryostat (ARS), where the measurements were done in high vacuum (10-6 Torr) conditions at variable temperatures. Cyclic Voltammetry. A conventional three-electrode setup was used for cyclic voltammetry. Slices of fully cleaned alcogels were placed on the surface of a glassy carbon working electrode by a doctor-blade technique. The working electrode was then placed in a 0.1 M tetrabutylammonium hexafluorophosphate electrolyte solution. The counter and reference electrodes were Pt wire and Ag/AgCl, respectively. Voltammograms were obtained by sweeping reductively at variable scan rates (10-100 mV/s). Dye Functionalization Experiments. A 10 mM ethanolic solution of (tris(2,2′bipyridyl)ruthenium(II) chloride hexahydrate (Ru(bpy)3+2) was added to a piece of a fully cleaned alcogel and allowed to stand for 24 hours. The solvent was removed and fresh dye solution introduced in the manner of a solvent exchange, for a total of three “washes” S3

with the dye solution. During dye functionalization, the chalcogel and its solution were protected from light with aluminum foil. Infrared Spectroscopy. Fourier-transform infrared spectroscopy (Nicolet 6700) was performed on ground powders of fully dried chalcogels (xerogels), mixed with cesium iodide. Transient Absorption. Pieces of the dye-functionalized gels were soniccated into 1,2-dichlorobenzene, forming a thick, viscous suspension. These suspensions were deposited onto glass substrates by spincoating, resulting in thin films of the gels. The photophysics of the dyes in these gels were then examined by various timescale transient absorption spectroscopies. Apparatus for the femtosecond transient absorption experiment has been described previously4. Frequency doubling of the fundamental resulted in pulses at 416 nm and 1.03 mW average power with ~150 fs time resolution. Films were illuminated through the deposited gel in a “front face” configuration. Each time point is a signal average over 5 s. Nanosecond transient absorption experiments are achieved by exciting the samples with 7 ns, 1.1 mJ, 416 nm using the frequency-tripled output of a Continuum Precision II 8000 Nd-YAG laser pumping a Continuum Panther OPO. The probe pulse, generated using a xenon flashlamp (EG&G Electro-Optics FX200), and pump pulse are overlapped on the film with the pump being focused to slightly larger than the probe. Again, a “front face” configuration is used. Kinetic traces are observed from 440-800 nm every 5 nm using a 416 nm long-pass filter, a monochromator, and photomultiplier tube (Hamamatsu R928) with high voltage applied to only 4 dynodes. Kinetic traces are recorded with a LeCroy Wavesurfer 42Xs oscilloscope interfaced to a customized Labview program (Labview v. 8.6.1). Spectra are built from the single wavelength kinetic traces taken every 5 nm. Each kinetic trace is representative of an average of 50 shots over a two microsecond time window. To increase the signal to noise ratio of the spectral profiles, 5-10 ns segments of data are averaged together and the median time reported as the time of the spectral slice. Photochemical Hydrogen Evolution. In a nitrogen-filled glovebox, pieces of the dye-functionalized gels were placed in a scintillation vial. Then, a 70:30 (v/v) acetonitrile/water solution containing 50 mM lutidinium chloride (proton source) and 50 mM sodium ascorbate (sacrificial electron donor) was added to the vial. The vial was sealed and flushed with pure N2 for 10 minutes. The vial was then continuously illuminated with a 100 Watt Xe lamp (Oriel). At regular intervals, the headspace in the vial was sampled and quantified with a gas chromatograph (Shimadzu GC-2014) equipped with a 5 Å molsieve column and a thermal conductivity detector. Argon was used as the carrier gas. Calibration was done with a 7% H2/93% N2 standard. References 1) Krebs, B., Pohl, S., Schiwy, W. Z. Anorg. Allg. Chem. 1973, 393, 241. 2) Coucouvanis, D., Kanatzidis, M.G., Simhon, E., Baenziger, N.C. J. Am. Chem. Soc. 1982, 104, 1874. 3) Wong, G.B., Bobrik, M.A., Holm, R.H. Inorg. Chem. 1978, 17, 578.

S4

4) Bullock, J.E., Vagnini, M.T., Ramanan, C., Co, D.T., Wilson, T.M., Dicke, J.W., Marks, T.J., Wasielewski, M.R. J. Phys. Chem. B 2010, 114, 1794.

Figure S1. FTIR spectra of pristine and Ru(bpy)32+-functionalized chalcogels. The pure Ru(bpy)32+ powder is included as a reference. At the right is a table of elemental amounts, in atomic percent, as determined by EDS spectroscopy before and after dye functionalization, as well as after washing the functionalized gels three times with ethanol.

S5

Figure S2. Nanosecond transient absorption spectra of Ru-bpy dye in three different environments: free in solution (A), in a control Zn:Sn2S6 gel (B), and in the redox-active Fe4S4-Sn2S6 gel (C). The indicated lifetimes at 620 nm were extracted from each series of spectra, and reflect the lifetime of the excited state emission of Ru-bpy. In solution and in the control gel, the lifetime is still observable on the nanosecond timescale, but it is not observable when the Ru-bpy is in the redox-active Fe4S4-Sn2S6 gel.

S6

2 Fe4S4:Sn2S6

0.06

2+

Ru(bpy)3 @ Fe4S4:Sn2S6

Kubelka-Munk units

Absorbance

A

1

0 400

500

600

700

Wavelength (nm)

800

B

0.04

0.02

0.00

500

1000

1500

2000

Wavelength (nm)

Figure S3. A) UV-VIS spectra of chalcogel thin films spun onto glass substrates, with and without Ru(bpy)32+ functionalization. B) Diffuse reflectance spectrum of a typical chalcogel.

S7

1.4 Chalcogel Before H2 Experiment

Absorbance (a.u.)

1.2

Chalcogel After H2 Experiment Ru-bpy Alone

1.0 0.8 0.6 0.4 0.2 0.0

400

500

600

700

800

Wavelength (nm)

Figure S4. UV-Visible absorption spectra of chalcogels that have been exposed to excess quantities of benzenethiol in N,N’-dimethylformamide, before and after hydrogen evolution experiments. The spectrum of the Ru(bpy)3Cl2 dye is also shown for comparison.

S8