Langmuir 1994,10, 443-4497
4483
A Combined Investigation of Four Colloids for Artificial Photosynthesis by Scanning Electron Microscopy, Quasielastic Light Scattering, and Bis-heteroleptic Styrene-Attached Ruthenium Complexes, Adsorbed at the WaterKolloid Interface Peter Schwarz,? Stefan Bossmann,$ Andreas Guldner,? and Heinz Diirr*9t Fachbereich 11.2, Organische Chemie der Universitat des Saarlandes, 66123 Saarbrucken, Germany, and Lehrstuhl fiir Umweltmesstechnik am Engler-Bunte Institut der Universitat Karlsruhe, 76128 Karlsruhe, Germany Received March 8, 1994. In Final Form: August 19, 1994@ In this report we discussthe combined applicationof scanning electronmicroscopy and quasielasticlight scattering as well as UV/vis titrations employing [Ru(bpy)3I2+and especially designed bis-heteroleptic sensor ruthenium complexes [Ru(bpy)2(MS-R)I2+, [Ru(bpy)2(DS-R)I2+, [Ru(tap)2(MS-R)I2+, and [Ru(tap)a(DS-R)I2+(bpy = 2,Z’-bipyridine, tap = tetraazaphenanthrene) possessing mono- and bis-styrene-attached tetraazaphenanthrene units as third ligands for the characterization of Ru, RuO2, IrO2, and MnOz colloids, as they are commonly used in model systems for artificial photosynthesis. The elucidation of the binding mechanisms and orientationand the arising supramolecularphotoelectron transfer propertiesof ruthenium sensitizers adsorbed on metal colloids can be regarded as tools for the generation ofhighly efficient systems for artificial photosynthesis. We describe in this paper a simple and application-orientedmethod for the evaluation of the catalyst’seffective surface in water solution and characteristic differences in the binding and interactionmodes of [Ru(bpy)3I2+to the four colloidal catalysts and the bis-heteroleptic sensorcomplexes listed above with the ruthenium colloid. In consequence, supramolecular systems for directed photoelectron transfer toward metal colloids in artificial model systems for photosynthesis can be designed in a simple and efficient manner. 1. Introduction
In the last two decades there has been an enormous effect to achieve the goal of an artificial system for photosynthesis using visible light as the driving force. The definition for artificial photosynthesis used here is a system which converts the energy of visible light directly into chemically stored energy, for instance in high-energy compounds as dihydrogen or methane. Scheme 1shows the main features of such a hypothetical system. It consists of a sensitizer, electron relay, and catalyst triad for water reduction in physical analogy to the naturaloccurring photosynthesis. The electron hole located on the sensitizer has to be refilled by the oxidation of water. A close contact of sensitizer and catalysts would certainly increase the efficiency of an electron transfer between them. Without catalysts, the bimolecular rate constants for water oxidation and reduction would be in the range of lo5M-l s-l instead of lo9M-’ s-l, and the recombination of photogenerated primary redox products would be the main reaction pathway.2 Another very important function of the catalysts is their role as an interface between the photo-one-electron transfer processes induced by the sensitizer and the thermal multielectron transfer processes in the water oxidation as well as in the reduction of suitable substrates.
* Author to whom correspondence should be addressed. t Universitat des Saarlandes.
Universitat Karlsruhe. *Abstract published inAdvanceACSAbstracts, October 1,1994. (1)Wasielewski, M. R. Chem. Rev. 1992,92, 435-61. Kalyanasundaram, K.Coord. Chem.Rev.1982,46,159.Kalyanasundaram, K.; Griitzel, M. Coord. C h m . Rev. 1986,69,57-125.Gust, D.; Moore, T. A,; Moore, L. A.; Devadoss, C.; Lidell, P. A.; Hermanet, K; Nieman, R. A.; Demanche, L. J.;DeGraziano, J. M.; Gouni, I. J. Am. Chem. SOC. 1992,114,3590. (2)Hofhann, W. Solar Znduzierte Redozreaktionen in Mikrohetemgenen Systemen; Nukem: Hanau, 1980. Bossmann, S.; Diirr, H.; Mayer, E. Z. Naturforsch. 1993,48b,369-386.
In sacrificial systems for dioxygen photoevolution from water, a sacrificial electron acceptor is used in order to detain the back-electron transfer. [Ru(bpy)alC12and many [cO(N3)5ofits relatives can be employed as ~ensitizers.~ C1)IClp is one of the most common sacrificial electron acceptors.* Colloid d-electroncontaining metals and metal oxides such as Ru, Fe20dFe300H20, RuO2, Co203, IrO2, and Ni(OH), (x = 2-3) have been proven to be efficient catalysts for sacrificial water ~ x i d a t i o n .Mn0$ ~ and di-, tri-, and tetranuclear Mn complexes have been employed as well with mixed success, mimicking the naturaloccurring enzyme-metal complex for water oxidation.’ A working system of artificial photosynthesis can only be achieved if all the reaction steps involved possess a sufficient yield. Therefore, supramolecular systems as physical models for the photosynthetic reaction centers might permit enhanced reaction rates and directed electron transfel.8 as indicated by Scheme 1. In these supramolecular colloid-metal complex systems the type of interaction between sensitizer and catalyst, as (3)Juris,A.;Barigeletti, F.; Campagna, S.; Balzani, V.; Belser, P.; Zelewsky, A. v. Coord. Chem.Rev. 1988,84,85.Meyer, T. J.Pure Appl. Chem. 1990,62,1003. (4)Mills, A.; Dodsworth, E.; Willims, G. Zrt.org. Chim. Acta 1988, 150,101. (5)M. Gratzel, J. K Angew. Chem. 1979,91,8; Angew. Chem., Znt. Ed. Engl. 1979. M. Gratzel, J. K.; Gratzel, M.; Kiwi, J. Chimiu 1979, 33,289. Lehn, J. M.; Sauvage, J. P.; Ziessel, R. Angew. Chem. 1979; Angew. Chem., Znt. Ed. Engl. 1979,18,701. Harriman, A.;Thomas, J. M.; Millward, G. R. New J. Chem. 1987,11, 12. Glikman, T. S.; Shcheglova, I. S. Kinet. &tal. lW8,9,461.Harriman, A.;Pickering, I. J.; Thomas, J. M. J. Chem. Soc., Faraday Trans. 1 1988,84,2795. Hurst, J. K;Zhou, J.;Lei, Y . Zrt.org. Chem. 1992,31,1010-1017. (6)Luneva, N. P.:Shafkovich, V. Y.; Shilov, A. E. J. Mol. Catal. 1989,52,49-62. (7)Ramarai. R.: Kiva,, A,:. Kaneco. M. J. Chem. Soc.. Faradar Trans 1 1987,83,1639.’ (8)Lehn, J. M.Angew. Chem. l988,100,91;Angew. Chem.,Int. Ed. Engl. 1988, 27, 90; Supramolecular Photochemistry; Balzani, V., Scandola, F., Eds.; Horwood Chichester, 1990; Supramolecular Photochemistry; Balzani, V., Ed.; Reidel: Dordrecht, 1987.
0743-7463/94/2410-4483$04.50/00 1994 American Chemical Society
Schwarz et al.
4484 Langmuir, Vol. 10, No. 12, 1994 Scheme 1. Proposed Scheme for the Artificial Photodecomposition of WateP ‘c hv
+ H+
‘I2H2+ O H
1
H20
semiconductor sensitizer electron catalyst for or metal relay water reduction colloid The sensitizer S absorbs (visible)light and induces thermal reactions. At the reductive branch of the cycle an electron delay mediates electron transport t o the catalyst for water (or substrate) reduction. At the oxidative side of the cycle the supramolecular interaction of the sensitizer and a suitable metal or metal oxide colloid enables the transport of an electron toward the adsorbed, Dreviouslv sensitizer. The actual oxidation of water can be performed at the active site of the metal or metal oxide - photooxidized, . colloid. (I
for instance binding modes, directed adsorption of bisheteroleptic sensitizers at the surface of a catalyst, and the influence of the binding process in the photophysical properties of the sensitizers become very important for the optimization of the water oxidation or r e d u ~ t i o n . ~ Besides their excellent properties in light absorption and conversion, ruthenium complexes are versatile tools for the investigation of microheterogeneous structures as DNA’s, starburst dendrimers, micelles,1° SiOZ,” andhighenergy metal surfaces. Bis-heteroleptic ruthenium complexes proved to afford here the most detailed information about the intrinsic motion and electron transfer properties of the systems investigated as well as the binding modes with different substrates such as DNA, micelles,and SiOzTi02 surfaces. The strategy for these measurements is the design of a ligand possessing a high affinity to the surface or microstructure to be p r ~ b e d . In ~ order to have an unambiguous interpretation of the photophysical measurements, the use of at least one independent investigation method to obtain information about the unperturbed microheterogeneous structure-in our case the colloid in the aqueous system-is requested. The characterization of colloids has been performed by a variety of methods such as different microscopy or diffraction techniques,12spectroscopy of emitted photons, electrons, or neutrons,13 the application of laser-induced photophysical and photochemical effects,14 or infrared spectroscopy.15 An intrinsic problem of all these approaches is the appearance of change in the colloids, induced by the investigation method. Therefore the estimation of the “real” size distribution and surface area of colloids in solution (especially)water remains difficult. (9) Diirr, H.; Bossmann, S.; Beuerlein, A. Photochem. Photobiol.,A : Chem. 1993, 73, 233-245. (10) Turro, N. J.;Barton, J. K ; Tomalia, D. A.Acc. Chem.Res. 1991, 24, 332. (11)Willner, I.; Otvos,J. W.; Calvin, M. J.Am. Chem. SOC.1981, 103,3203. Willner, I.; Yang, J. M.; Laane, C.; Otvos,J. W.; Calvin, M. J. Phys. Chem. 1981,85, 3277. Willner, I.; Degani, Y. Zsr. J. Chem. 1992,22,163. Willner, I.; Degani, Y. J. Chem. Soc., Chem. Commun. 1982, 761. Willner, I.; Maidan, R.; Mandler, D.; Diirr, H.; Dorr, G.; Zengerle, K. J.Am. Chem. SOC.1987,109, 6080. (12) Somorjai, G. A. Principles of Surface Chemistry; Prentice-Hall: Englewood Cliffs, NJ, 1972. (13) Adamson, A. W. Physical Chemistry of Surfaces; John Wiley & Sons: New York, 1990; pp 329-363. (14) Chatzi,E.G.;Urban,M. W.;Ishida,H.;Loenig, J.L.;Laschewski, A.; Rothberg, L. Langmuir 1988,4,846. Damme, H. v.; Hall, W. K. J. Am. Chem. SOC.1979,101,4373. Thomas, J. K. J. Phys. Chem. 1987,
91. - - ,267. -(15) Ishikawa, T.; Nitta, S.;Kondos, S.J.Chem.Soc.,Faraday Trans. 1 1986,82, 2401.
The combination of at least two completely independent techniques for the acquisition of experimental data is therefore required to avoid possible misinterpretations of the data obtained. According to these investigation strategies outlined above, we report the following in this paper: (1) The synthesis of four catalysts for water oxidation [Ru, RuOz, IrOz, and MnOzl and their characterization by scanning electron microscopy; (2) the evaluation of the effective particle size of these catalysts in solution by applying a laser light scattering technique; (3) UV-vis adsorption studies of a novel class of mono- and bis-styrene-attached bis-heteroleptic ruthenium complexes and tris4bipyridinelruthenium a t the high-energy surface of the Ru colloid in order to give a contribution to the design of a supramolecular system for artificial photosynthesis. 2. Results
Synthesis and Characterization of Four Colloidal Catalysts for Water Oxidation. Four catalysts for water oxidation [Ru,16Ru02,17IrOz,l6MnOzl81as possible colloids for the adsorption of ruthenium sensitizers have been prepared according to standard procedures. These procedures for catalyst preparation have been adapted and optimized in terms of long-term stability against degradation reactions and catalytic efficiency over the last decade in our groups in order to be used as versatile tools in sacrificial photochemical systems for water and carbon dioxide reduction as well as water oxidation.18 The ruthenium and iridium dioxide colloids were prepared by sodium citrate reduction of RuC13-3 HzO or NazIrC16. The RuOz sol was prepared by ultrasonic treatment from commercial RuOz. The MnOz colloid was synthesized using KMn04 as starting material and N H 3 as reducing agent. These four colloidswere expected to differ in their actual concentration, shape, and accessible surface in the colloid solution from the preparation processes. Hereby the complication arises that standard methods for the characterization of porous surfaces, as for instance Nz (16) Haniman, A.; Naher, G.; Mosseri, S.; Neta, P. J. Chem. Soc., Faraday Trans. 1 1988,84, 2821-29. Mayer, E. Diploma Thesis, Universitlt des Saarlandes, Saarbriicken, 1984. (17) Cuy,E. J.J.Phys. Chem. 1920/21,25,415-17. The preparation of the RUOZcolloid was performed according to the procedure described for a F’t colloid in: Frank, A. J.; Willner, I.; Goren, Z.; Degani, Y. J . Am. Chem. SOC.1987,109,3568. (18) Heppe, G. Thesis, Universitat des Saarlandes, Saarbriicken, 1994.
Four Colloids for Artificial Photosynthesis absorption according to the BET method,lg cannot be adapted to determine the properties of colloids in water solution. The knowledge of these parameters can be regarded as essential for the fine tuning of the very complex redox processes in systems for artificial photosynthesis. Scanning Electron Microscopy The first step toward the characterization of our four catalysts in water solution was the evaluation of the shapes of the catalyst particles by employing scanning electron microscopy. The catalysts were dehydrated in high vacuum. The removal of the solvent-due to the experimental constraints of the method used here-caused, as previously expected, coagulation of the catalysts. Therefore, characteristic values, like the diameter of the catalyst's particles, cannot be derived in principle from these measurements. However, the important information from that first approach is the determination of the shape of the subunits in coagulated clusters. All four colloids belonged to the class of isometric spherocolloids.
3. Investigation of the Particle Size by Quasielastic Light scattering21 Measurement of the Average Particle Size. In the next step of colloid investigation in water solution, the technique of quasielastic light scattering has been applied. In our experimental setup, an He/Ne laser ( A = 632.8 nm) has been used. The reflection of the coherent laser light from a transparent colloidal solution (a = 45" between the light source and the solution surface) serves as the principle of measurement. Hereby the phase of the reflected and scattered laser light is dependent on the position of the scattering particle to the light beam as well as the distance between the particles and the detector (photomultiplier). Therefore, interference between different colloidal particles in a colloidal solution can be observed. The Brownian motion of the particles in the liquid modulates the interference and thus the intensity monitored by the detector in a given time interval. These light intensity fluctuations can be described by time-dependent correlation functions: The intensity correlation function g ( t ) is of the form
where I(t)= intensity in the detector at time t . Assuming a monodisperse suspension of hard, ball-shaped particles, we can use an exponential dropoff of the correlation function between At = 0 and dt = 0. The resulting equation is
where t b = particle size dependent relaxation time of the Brownian motion and A,B = apparatus constants. If the REM investigation indicates a polydisperse system, the integral in eq 2b can be solved numerically.
where s = particle size, a,b = interval of particle size from ) distribution of the particle size. REM, and ~ ( s = (19)Wedler, G.Lehrbuch der Physikalischen Chemie;VCH Verlagsgesellschafk Weinheim, 1985. (20)Naumer, H.; Heller, W. Untersuchungsmethoden in der Chemie; Thieme: Stuttgart, 1986;pp 311-26. (21)McNeil-Watson, F. K. A . New Instrument For Particle Size Analysis and Electrophoretic Mobility Measurement Using Photon Correlation Spectroscopy; Malvern Instr.: Miinchen, 1988.
Langmuir, Vol. 10, No. 12, 1994 4485 Table 1. Diameter and Surface Areas in Solution of the Four Catalyete for Water Oxidation colloid Ru RuO2 IrOz MnOz
diametee [nml 106 326 102 127
surface [m2/gl 87
50 32 200
surface [m2/mol] 8793 6653 7022 17378
diametel.b [nml 232 1017 203 242
The measurements were taken 1 day after the preparation of the catalysts. The second measurements of the diameter of the colloid catalysts were taken 155 days after preparation.
The dependence of t b on the diffusion coefficient and the geometry factors from the experiment is given by
(3) where D = diffusion coefficient and q is defined as
(4) where a = scattering angle, q = length of the scattering vector, 77 = viscosity of the colloidal solution, and 4 = wavelength of incoming light. Finally, the diameter d of the colloidal particles can be calculated using the simple EinsteidSmoluchowski equat i ~ n : ~ ~
where K = Boltzmann constant, T = temperature [Kl, and 17 = dynamic viscosity of the colloidal solution. In Table 1 the diameters for the four colloids calculated by this method are presented. As it becomes clear immediately from these data, the averaged diameter of the colloidal particles differs by approximately a factor of 3. The age ofthe colloidal solutions was 10-14 days during the first series of measurements. The same experiment has been repeated by employing the original stock solutions used during all the measurements reported in this publication after 5 months (155 days). There has been a significant aging of the colloids: The average diameter of all catalysts was found to be 2-3 times larger than in the first investigation. An important point here is that the aging occurred (in first order) proportional to the size of the particles directly after preparation. Another important observation might be that the aging velocity was only little dependent on the chemical nature of the colloids: The ruthenium metal colloid showed in principle the same behavior as that of the metal oxides IrOz, RuOz, and MnOz. Measurement of the Real Concentration of the Colloids in Solution. Dependent on the synthetic procedures used for the synthesis of colloidal catalysts for artificial systems for the photooxidatiodphotoreduction of water, there is always some loss of material during the preparation. Especially during the filtration and ion exchange operation necessary for the preparation of longterm stable colloidal solutions, some material is removed. In addition, the chemical reaction (usually controlled reduction of metal chlorides in high oxidation states) itself might be incomplete: ESCA studies of colloidal platinum catalysts and platinum catalysts on aTiOzsupport showed the presence of platinum(I1) specie^.^^^^ Consequently, some nonreacting starting material might remain in every reaction, leading to colloidal catalysts for solar energy
Schwarz et al.
4486 Langmuir, Vol. 10, No. 12, 1994
R=H R=OCH,: R=CN
m
I.
L.
m
R=H R=OCH,: R=CN:
MSH MSM MSC
..
..
DSH DSM DSC
I_\
IRU(TAP)~(TAP-RI” / 7
/
\
R=H MSH R=OCH3: MSM R=CN MSC
IRu(bpy),(PPB-RI’* R=H: R=OCH3: R=CN:
PPB PPB-CI PPB-CN
Figure 1. Chemical structures of the bis-heteroleptic sensitizers employed in this study.
Table 2. (Photo)physicalProperties of the Novel Mono-and Bis-styrene Attached Bis-heteroleptic Complexes [Ru(bpy)2(L)12+ and [Ru(tap)2(L)la+as well as [Ru(bpy)@
[R~(bpy)31~+
452
613
610
4.16
1.28
-0.82
-1.32
0.77
254 4.52 1.50 -0.60 -0.70 1.40 [R~(bpy)z(MSH)l~+ 395 730 448 4.54 1.49 -0.61 -0.64 1.46 [R~(bpy)z(MSM)l~+ 417 722 76 4.58 1.50 -0.60 -0.73 1.37 [Ru(bpy)2(MSC)I2+ 385 737 4.81 420 724 1995 1.46 -0.31 -0.80 0.97 [Ru(bpy)z(DSH)I2+ a [Ru(bpy)z(DSM)lZ+ a 463 727 3991 4.74 1.48 -0.44 -0.61 1.21 208 4.56 1.44 -0.39 -0.88 0.95 [R~(bpy)z(DSC)l~’ a 395 729 1536 4.35 1.40 -0.79 -0.80 1.39 [Ru(tap)z(MSH)lZ+ 399 603 602 4.31 1.56 -0.63 425 604 -0.87 1.32 [Ru(tap)z(MSM)I2+ 375 604 255 4.18 1.52 -0.66 -0.85 1.33 [Ru(tap)z(MSC)I2+ 4.64 1.80 -0.41 -0.85 1.36 443 708 3951 [Ru(tap)z(DSH)l2+ [Ru(tap)z(DSM)I2+ a 438 710 305 5.64 1.76 -0.50 -0.86 1.40 [Ru(tap)2(DSC)l2+ a 416 605 347 4.54 1.76 -0.45 -0.66 1.55 CH&N/H20 1:l (v/v). All redox potentials were measured by cyclic voltammetry employing the PF6- salts in 1.0 M TBABF4 solution in CH3CN. T h e excited state redox potentials were calculated by the equation E”(Ru*/R+) = E’(Ru/Ru+) + Eoo(Ru*). Eo0 was measured by low-temperature (77 K)emission spectroscopy in ethanol glass matrices.
conversion and other purposes. To measure the real concentration of our colloids in solution, 1 mL of each stock solution was taken, and a careful freeze-drying process was followed. Then the weight of the residue was measured by gravimetry.22 A Simple Model for the Estimation of the Real Surface in Solution. To gain more insight into the real behavior of our system for solar energy conversion,a simple (22) Jander, G.; Blasius, E. Einfihrung in das Anorganisch Chemische Praktikum, 10. Auflage; Hirzel: Stuttgard, 1977.
model for the calculation of “real surface areas” of our colloidal particles in solution has been developed. It uses data input for the weight of the residue and the averaged diameter of the colloidal particles in solution. The model developments begin with the results from the REM investigation: all four colloids employed here are built from spherical particles only. Furthermore, there exist in the inner region of the colloidal particles only units in simple cubic packing. As it is clear from simple crystal models, the space demand of the spherical colloids
Four Colloids for Artificial Photosynthesis
cannot exceed 52% in the simple cubic packing m0de.~3 On the other hand, this packing mode allows for a maximum of water, incorporated in the structure of the colloidal particles. The averaged particle diameter provides the total volume of the hydrated colloidal catalysts in solution. Thus the knowledge of the elementary unit of each metal or metal oxide employed as colloidal catalyst in our investigations permits the calculation of the volume of the spherical units in solution, assuming the simple packing model. With the known concentration of the colloids [g L-ll in solution, the molar concentration per milliliter of solution can be easily derived. From the calculated surface area of the particles and the concentration of the particles in solution [g L-l or mol L-ll, the surface areas of our colloids in solution can be calculated easily. The values obtained for our four colloids are presented in Table 1. Interestingly,the surface areas per gram and mole obtained from this model are relatively small, corresponding with the relatively large diameters of the colloids in solution. These results support earlier findings on T i 0 2 platinum catalysts, demonstrating that the catalytic activity cannot be regarded as a simple function of size and surface area of the catalyst^.^ Furthermore, the ability of the metal and metal oxide catalysts to mediate electron transfer from the contact regions with the photooxidized sensitizers or photoreduced electron relays can be regarded as a most important factor in the reactivity of our systems for artificial photosynthesis. 4. Interaction of the Colloids with Ruthenium Complexes. The possibility of an excited sensitizer undergoing an electron exchange reaction directly with a metal or metal oxide colloid is limited by the diffusion properties in the microheterogeneous system. The simple EinsteidSmoluchowsky equation (5)3apermits the estimation of the averaged diffusion length of an excited sensitizer in a given medium. For instance, the average m, diffusion length is only approximately 1 x assuming a diffusion coefficient of m2s-l and a lifetime of the excited (triplet) state of the sensitizer of 1 x s. Consequently, the efficiency of the photoelectron transfer reaction is severely hampered. Adsorption of ruthenium complexes on semiconductors and (modified) Si02 surfaces was studied intensively (among others) by Kalyanasundaram, G r a t ~ e 1 ,and ~~ Willner.ll In a recent study we have demonstrated that the attraction between Si02 or SiO2-TiO2 colloids and bisheteroleptic ruthenium complexes of the type [Ru(bpy)z(PPB)I2+follows an electrostatic adsorption mechanism.2b From W/vis, fluorescence, and quenching studies, we concluded that these ruthenium complexes show an orientation at the interface to Si02: The smaller 2,2’bipyridine ligands are oriented toward the surface and the larger styrene-attached PPB ligand toward the bulk solution. In this report we present the results from novel bisheteroleptic and tris-homoleptic ruthenium complexes possessing a larger space demand in interaction with the colloids described and characterized in parts 2 and 3. Electrophoresis results confirmed that all four colloids are negatively charged. Furthemore, the charge density of the colloids was found to be comparable to that of the Si02 and Si02-Ti02colloids investigated earlier (0.18 +/(23)Atkins, P. W. Physikalische Chemie;Verlag Chemie: Weinheim, 1990. (24)Kalyanasundaram, K.; Gratzel, M. Coord. Chem.Rev. 1986,69, 57-125. Pileni, M.; Gratzel, M. J.Phys. Chem. 1980,84, 2402.
Langmuir, Vol. 10, No. 12, 1994 4487 0.04 Cm-2).25The question, and its answers, of whether the same type of interaction can be found between our new ruthenium complexes and the colloidal systemstailormade for the sacrificial reductiodoxidation of water or carbon dioxide, possessing bigger particle sizes, is especially important for the design of a system for artificial photosynthesis in high photoconversion yields. If the same (or a similar) type of mechanism can be found, the generation of a tailor-made system complies with the necessity for long-term stable catalysts. Ruthenium Complexes Employed in This Study. A new series of bis-heteroleptic ruthenium complexes has been designed to serve as sensor complexes at the interface between the colloids and the bulk solution. In addition, the standard complex [Ru(bpy)3I2+was used for comparison purposes. Our new bis-heteroleptic ruthenium complexes consist of mono- and bis-styrene-attached 1,4,5,8tetraazaphenanthrenes and 2,2’-bipyridine or 1,4,4,8tetraazaphenanthrene as basic ligands. The synthesis of the ruthenium complexes will be described in a separate publication. The development of these sensor sensitizers has been aided by molecular modeling? The difference in size between the styrene-tetraazaphenanthrene units and the basic ligands is large. This should enhance a possible orientation ofthe complexes at the interface. In addition, even the styrene-attached complexes can be considered to be highly The difference in the electronic densities of 2,2-bipyridine and the mono- and bis-styreneattached 1,4,5,84etraazaphenanthreneson one hand and the latter ligand 1,4,5,&tetraazaphenanthrene on the other, as basic ligands is not very large. Therefore, electronic perturbations in the excited state can be regarded as small, and the assumption that the observed MLCT (metal to ligand charge transfer) transition of the complex arises from the addition of local rutheniumligand-MLCT transitions is permitted. The position of the noncoordinated aromatic nitrogen atoms in the heterocyclic ring systems of all 1,4,5,84etraazaphenanthrenes is favorable to form hydrogen acceptor bonds, for instance from polar M-0-H (m = metal) groups, at the interface in addition to the Coulomb interaction. The tuning of the electronic and photophysical properties of the ruthenium complexes is possible by choosing different substituents at the styrene-attached benzene ring(s). The structure of the ruthenium complexes employed in this study is shown in Figure 1. In Table 2 some photophysical properties of our new ruthenium complexes [Ru(bpy)2(R-tap)l2+,[Ru(bpy)z(Rtap-R)I2+,[Ru(tap)2(R-tap)12+,[Ru(tap)2(R-tap-R)l2+, and [Ru(bpy)3I2+are listed. As it becomes clear, the emission maxima and the lifetimes of the excited MLCT states can be especiallytuned toward very long or very short lifetimes by different substitution of the styrene-attached benzene rings. Long lifetimes of the excited states favor both the ability to exhibit environmentally dependent sensor properties and efficient photoelectron transfer reactions. In addition, the (photo)redox potentials of the styreneattached ruthenium complexes are very suitable for systems for artificial photosynthesis: The redox transitions Ru2+I3+and Ru2+*I3+are approximately 0.3 V more positive in the styrene-attached ruthenium complexes than the ones of [Ru(bpy)3I2+. At the same time, the ’
(26)Braun, R. Thesis in preparation, Universitat des Saarlandes, - Saarbrtkken, 1993. (26)“he force field and MIND0 calculations were performed using CHARMm. Version 2.1.89 0131 (C) (Polvizen Cow.) (27)FromlH-NMR measurementa, the ;:entation ofthe substituents at the double bonds of the styrene-attached tetraazaphenanthrenes was found to be in the trans position.
Schwarz et al.
4488 Langmuir, Vol. 10, No.12, 1994
thermal and photochemical reduction potentials for the redox transitions Ru2+11+and Ru2+'l+ are approximately 0.7 V more positive due to the more electron-deficient ligands in the ruthenium complexes. From these potential levels excellent photoreduction properties of the ruthenium complexes arise. According to the (photo)redox potentials, both transitions R U ~ + ~ + and Ru2+*11+ can be considered very favorable for water oxidation processes, whereas Ru2+*I3+ and R U ~ +permit ~+ the (photochemical) reduction of water and other substrates. Therefore, the interaction of the bis-heteroleptic ruthenium complexes [Ru(bpy)2(R-tap)12+,[Ru(bpy)2(Rtap-R)I2+,[Ru(tap)2(R-tap)12+, and [Ru(tap)2(R-tap-R)I2+ presented here with microheterogeneous surfaces is of specialinterest for applied photochemical systems for solar energy storage.
5. Interaction of the Sensor Complexes with Colloids [Ru(bpy)s12+as Reference Compound. First, the lMLCT absorption band of the well-known reference complex [Ru(bpy)3I2+has been monitored during its interaction with our four colloids. The mass ratio M, between the complex and the colloid employed has been calculated according to the simple equation
m (sensitizer)
M, = m (colloid)
I . 008
279
(6)
One reason for this approximation follows from molecular modeling: for most of the ruthenium complexes a (firstorder) linear relationship between the molar mass of the complexand the calculated space demand has been found. The second reason for choosing this simple ratio is the presence of different types of ruthenium-colloid interactions. According to the paradigm of Turro et al.,28the coexistence of surface-bound ruthenium complexes and complexes bound in the Stern layer as well as the GouyChapman layer makes it tremendouslydifficult to estimate the exact fraction of the surface-boundspecies. Therefore an adsorption isotherm such as the commonly known Langmuir isotherm cannot be applied successfully in this case. On the other hand, only minor mistakes arise from the differences in the molar masses of our colloids used here. The differences in the curves obtained remain still evident, even regarding a scaling factor of 2 in the molar masses. According to standard calculations performed earlier,2athe monomolecular coverage of our colloids has been estimated to be in the range between M,= 3 and M, = 12. As it becomes clear from Figure 3, the most remarkable changes were found for the Ru colloid in interaction with [Ru(bpy)3I2+.These changes in the absorption coefficient of a ruthenium complex are generally indicative of a binding of the complex a t the colloidlwater interface. A maximum in the absorption coefficients was monitored around Mx = 5. The occurrence of a maximum in the absorption coefficient found dependent upon the loading density M,can be interpreted as arising from the adsorption of ruthenium complexes at the water-colloid interface, which offers a new deactivation pathway for the excited 'MLCT states of these sensitizers, and consequently,an increase in the optical absorption can be monitored.29 Then, a subsequent decrease in the absorption coefficients monitored a t increasing density of the (28) "0, N.J.; Barton,J. IC;Tomalia,D. A. Ace. Chem.Res. 1991, 24,332. (29) Garoff, S.; Weitz, D. A.; Alvarez, M. S.; Gersten, J. I. J. Phys. Chen. 1984,81,5190.
e. e00
1
Figure 2. W-vis absorption spectrum of [Ru(bpy)2(MSM)I2+ in water.
I o
1 14500 14000
5
0
5
-
MnOt
t
1 0 1 5 2 0 2 5 3 0 3 5 4 0 45 5 0
M,
Figure 3. Extinction coefficient of the standard ruthenium sensitizer [Ru(bpy)3I2+in the presence of Ru, IrO2, Ru02, and MnO2 colloids (see Table 1.) As it becomes immediately clear, the most interesting changes in the absorption coefficient[e = 14 500 cm2/mmolin water] were seen in interaction with Ru and IrO2. In our W/vis titrations, the Ru sol was chosen because of the strong dependence of e on its concentration in the system.
ruthenium complex bound to the colloid was found. The slope obtained for [Ru(bpy)3I2+at the Ru colloid is typical for plurimolecular absorption ~ o n d i t i o n s . ~ ~ * ~ Only moderate changes in the absorption coefficientof [Ru(bpy)3I2+were seen in interaction with the three metal oxide colloids, except for a significant enhancement of the absorption coefficient at IrO2. In all cases investigated here the absorption maximum of its 'MLCT band [A, = 452.5 f 1nm] did not change significantly.
Four Colloids for Artificial Photosynthesis 428
Langmuir, Vol. 10, No. 12, 1994 4489
1
-
1
.
1
.
1
.
-
426 B
-
424
:
422 -m
418 -
[RU(~~~)~(MSM)I~+
420
.
-
B B O '
416.
0
El
D '
I
'
'
10
20
'
30
50
40
M X
406
1
.
1
.
I
.
I
.
.m 404 - 0
-
402
398
-
396
-
400
0
68
B 6
G I
394
.
1
10
0
.
1
20
.
30
I
I
B
-
.
40
50
392 390
386
D
L I
I .
LI 1
.
I
For all further measurements, the Ru colloid was employed because it exhibited the most significant effects in interaction with [Ru(bpy)s12+.Furthermore, it is ofgreat interest as a catalyst in systems for artificial photosynthesis, and its charge density has been found to be the highest of the colloids investigated. Bis-heteroleptic Ruthenium Complexes Containing 2,P'-Bipyridine as Basic Ligand at the Ruthenium ColloiWater Interface. The bis-heteroleptic ruthenium complexes using 2,2'-bipyridine and 1,4,5,8tetraazaphenanthrene as basic ligands were synthesized t o clarify the binding mechanism of the complexes at the colloidlwater interface. In the case where a purely electrostatic mechanism takes place, the resulting slopes of the measurement curves obtained by using bis-heteroleptic complexes featuring 2,2'-bipyridine and 1,4,5,8tetraazaphenanthrene as basic ligands should be very similar. Such a mainly electrostatic binding mechanism has been found in our earlier investigations employing
.
I
.
ruthenium-bis(bipyridine)-styrene-pyridyl-pyrimidine complexes [Ru(bpy)2(PPB-2)12+and Si02 or SiOzT i 0 2 colloid^.^^^^ The photophysical properties of our sensor complexes in the presence of the ruthenium colloid and dependent upon the loading density M,should differ significantly, if contributions other than purely electrostatic forces to the binding mechanism of the complexes to the surface of the colloids dominate, such as the occurrence of hydrogen bonding, dipolar-dipolar interactions, or electronic exchange effects. The structure of tetraazaphenanthrene as ligand is able to support these binding mechanisms because of its nonbinding heterocyclic ring nitrogens in favorable positions for hydrogen acceptor bonds, its low electronic density in its ll electron system, and the resulting dipole of the ligand, which have been obtained from force field calculations. If the binding of the complexes is dominated by forces other than purely electrostatic, the (modified) tetraaza-
Schwarz et al.
4490 Langmuir, Vol. 10, No. 12, 1994 1 .oo
0.95
-
-
0.90
-
-
0.85
-
-
0.80
0.75
-
0.70 0
I
.
10
I
.
20
I
.
30
I
I
.
50
40
MX
0.70 0
I
10
.
1
.
20
1
30
.
1
40
.
[Ru(bpy),(MSH)l2+
-
50
MX 1 .oo
0.95 0.90 0.85
e~ I R U ( ~ ~ ~ ) ~ ( M S C ) I ~ +
0.80
0.75 0.70 0
10
20
30
40
50
MX
Figure 5. Dependence of the normalized molar absorption coefficients on the sensitizer concentration for the mono-styreneattached bis-heteroleptic sensitizerspossessing 2,2'-bipyridine as basic ligand. The data shown in Figure 4 and Figure 5 arise from the same titration experiment. phenanthrene ligand is expected to be oriented toward the surface of the ruthenium colloid. Therefore, measurements of the UV-vis maxima of the lMLCT transitions of our complexes have been performed in the presence of the ruthenium colloid. If a n orientation of the bisheteroleptic complexes at the surface of the colloids exists, characteristic changes in both the 'MLCT absorption maxima and the optical absorption coefficients can be expected. The lMLCT transitions of both 1,4,5,8-tetraazaphenanthrene and its styrene-substituted derivatives are sufficiently different from each other in ruthenium complexes to permit in principle the same analysis as that employing 2,2'-bipyridine as basic ligand. Due to the design of our sensor complexes, the observed shifts permit straightforward conclusions on the binding mechanism and orientation of the complexes at the ruthenium colloid. Mono-styrene-Substituted Sensor Complexes. Figure 4 exhibits the dependence ofthe lMLCT absorption maxima of the bis-heterolepticcomplexes [Ru(bpy)z(MSM)l2+,[Ru-
(bpy)2(MSH)I2+,and [Ru(bpy)2(MSC)12+ in interaction with the ruthenium colloid a t different loading densities M,. For all three different complexes a significant hypsochromic shift at increasing M , values can be monitored. At the same time, the absorption coefficient increases until a plateau is reached, as shown in Figure 5. From these measurements one can clearly see that the design of the sensor complexes was successful: very significant changes in both the wavelength of the lMLCT transitions and the absorption coefficientscan be seen in interaction with the colloid. Note that the lMLCT absorption maxima occur at very different wavelengths due to the tuning by substitution of the benzene ring of the styrene attached to the tetraazaphenanthrene ligand of the bis-heteroleptic ruthenium complex. From the comparison ofthe 'MLCT absorption maxima without and in the presence of the ruthenium colloid, two conclusionsfollow immediately: (1)The complexis binding preferentially with the modified tetraazaphenanthrene ligand toward the surface. (2) There are other binding
Four Colloids for Artificial Photosynthesis
hrmaxl [-I
Langmuir, Vol.10,No. 12, 1994 4491
402 I
1
I
m
-
400 - 0
398
396
-
Gl
[RU(~~~)~(DSC)I~+ 0
G
394
l
IJ
Q
I
mechanisms than purely electrostatic interactions dominating. These would lead to the opposite orientation of the bis-heteroleptic complexes at the surface of the colloid. It is well known from literature that 2,2'-bipyrazine binds to the surface of colloidal ruthenium, but the mechanism still remains unclear. 1,4,5,8-Tetraazaphenanthrene possesses the same position for the aromatic nitrogens in the heterocyclic ring system. As it becomes clear from our measurements, the most striking changes in the optical properties of our complexes appear in the M, region from 0 to 10. This range is apparently consistent with the calculated value for monomolecular coverage of the surface from our REM and light scattering experiments and our force field calculations. Bis-styrene-Substituted Sensor Complexes. As Figures 6 and 7 indicate, qualitatively the same dependence of the 'MLCT maxima and the absorption coefficients were measured when the bis-heteroleptic ruthe-
I
Q I
nium complexes possessing one bis-styrene-substituted tetraazaphenanthrene ligand ([Ru(bpy)z(DSM)lzf,[Ru(bpy),(DSH)lZ+,and [Ru(bpy)z(DSC)P)were employed. Here also the position of the 'MLCT transitions can be tuned successfully by changing the substitution of both styrene-substituted units. The main conclusion from these measurements is in principle the same as from the monostyryl-substituted ruthenium complexes: The modified tetraazaphenanthrene ligand binds toward the surface. Bis-heteroleptic Ruthenium Complexes Featuring 1,4,S,8-Tetraazaphenanthreneas Basic Ligand at the Ruthenium ColloidNater Interface. In the next step bis-heteroleptic ruthenium complexes employing 1,4,5,8-tetraazaphenanthrene as basic ligand and styrenesubstituted tetraazaphenanthrenes as a third ligand have been used as sensor complexes in interaction with the ruthenium-colloid under exactly the same conditions.The comparison between the slopes of the monitored 'MLCT absorption maxima and the optical absorption coefficients
Schwarz et al.
4492 Langmuir, Vol. 10,No. 12, 1994
0
0.60 0
1.00
10
20
30
1
I
I
40
0.95 0.90 0.85
0.80
0
0.75 0.70
0.65 0.60 0
I
I
1
10
20
30
40
30
40
MX
1::;
1
0.65
0.60 0
10
20 M X
Figure 7. Dependence of the normalized molar absorption coefficients on the sensitizerconcentration for the bis-styrene-attached bis-heteroleptic sensitizerspossessing 2,2’-bipyridineas basic ligand. The data shown in Figure 6 and Figure 7 arise from the same titration experiment. obtained from the bis-heteroleptic complexes possessing 2,2’-bipyridineand 1,4,5&tetraazaphenanthreneas basic ligand can provide further insight into the dominating binding mechanism. The question, and its subsequent answer, is whether the styrene-substituted or the unsubstituted tetraazaphenanthrenes are the better binders as ligands in ruthenium complexes to the ruthenium colloid for a special reason: if the unsubstituted ligand binds preferentially, a mechanism dominated by hydrogen bonding or dipolar interactions can be assumed. On the other hand, if the styrene-substituted ligand interacts more strongly with the high-energy surface of the ruthenium colloid, there is evidence for an electronic contribution to the binding mechanism. Mono-styrene-Substituted Sensor Complexes. Figures 8 and 9 present the experimental results from [Ru(tap)z(MSM)I2+,[Ru(tap)z(MSH)lZ+, and [Ru(tap)z(MSC)lZ+at the ruthenium colloidwater interface. The hypsochromic shift of the lMLCT absorption maxima a t increasing
complex concentration in the system and the hyperchromic shifts in the absorption coefficientswere found to be much weaker compared to the case for [Ru(bpy)z(MSM)lZ+, [Ru(bpy)2(MSH)I2+,and [Ru(bpy)~(MSC)l~+ featuring 2,Ybipyridine as basic ligand. AB it follows from the observed shiRs in the absorption maxima and the optical absorption coefficients, the styrene-substituted ligands bind preferentially to the colloid surface. However, from our experiments we consider this orientation to be not as complete as with [Ru(bpy)z(MS-R)I2+and [Ru(bpy)z(DS-R)lZ+(R = M,H,C) as sensor complexes. Furthermore, a distinct dependence of the orientation behavior of the bis-heteroleptic ruthenium complexes [R~(tap)z(MS-R)1~+ on the substitution of the attached styrene groups follows the slopes: whereas [Ru(tap)z(MSH)I2+and [Ru(tap)z(MSC)Iz+show in principle the same trend in their ‘MLCT absorption maxima and absorption coefficients, depending on their concentration in the system, [Ru(tap)z(MSM)lZ+does not exhibit
Four Colloids for Artificial Photosynthesis
&,bI
434
Langmuir, Vol. 10, No. 12, 1994 4493
.
I
I
I
mm
432 430
-
I
B
426 428
424
405 4 m a x 1 [m] 403
-mm
-: -
a B Q I
1
1
I
1
I
Q
401 399
m 0
393 39 1
418
[Ru(tap),(MSM)IZ+
-. -
a
397 - m 395
m
-
-
I
I '
I
I
.
I
I
416 414 412 41 0 408
m
406
0
m
404 402
I
1
I
MX Figure 8. 'MLCT absorption maxima of the three bis-heteroleptic complexes possessing one bis-styrene-tetraazaphenanthrene ligand and two tetraazaphenanthrene units as basic ligands in the presence of the Ru colloid at different concentrations in the system. a similar behavior. The comparison of the optical absorption coefficients, however, yields a distinct similarity between [ R ~ ( t a p ) ~ ( M s Mand ) l ~ +[Ru(tap)2(MSH)I2+.Both complexes show a clear hypochromic shift in the presence of the colloid. [Ru(tap)z(MSC)lZ+does not show this effect and reaches a plateau region a t much smaller M,values (M,= 7 instead of Mx= 12). According to the model of Garoff, Weitz, and G e r ~ t e n , 2as ~ discussed above, the experiment can be interpreted in terms of a n overlap of two effects: The orientation of the bis-heteroleptic complexes at the colloidwater interface and the electronic interaction ofthe ruthenium complexes and the ruthenium colloid. Both effects influence the optical absorption coefficients in different directions and cannot be clearly separated from this experiment. Again, remarkable optical property changes of the adsorbed sensor complexes were found in the M, region from 0 to 10, consistent with our model described above.
Bis-styrene-Substituted Sensor Complexes. Finally, the bis-styrene-substituted bis-heteroleptic ruthenium complexes have been employed in the titration experiments: Both [Ru(tap)z(DSH)I2+ and [Ru(tap)2(DSC)I2+exhibit experimental behavior strikingly similar to that of [Ru(tap)2(MSH)I2+and [ R U ( ~ ~ ~ ) ~ ( M S CTherefore )I~+. the same conclusions as discussed in the previous section can be derived. 6. Discussion The investigation presented here consists of two parts: In the first step the real surface of four metal and metal oxide colloids in water as solvent has been estimated by the combined use of quasielastic light scattering, scanning electron microscopy, and gravimetry. The colloids, prepared according to an optimized synthesis protocol to ensure high conversion yields in systems for artificial photosynthesis, were found to possess surprisingly large particles and consequently relatively small surfaces.
Schwarz et al.
4494 Langmuir, Vol. 10,No. 12, 1994
0.80 0.75
;;iI::
0.70 0
1.00
r
I
1
1
10
M 2 0X
30
I
1
1
20
30
;
m
40
0.95 0.90
0.85 0.80
0.75 0.70 0
10
M
40
X
Figure 9. Dependence of the normalized molar absorption coefficientson the sensitizer concentration for the bis-styrene-attached bis-heteroleptic sensitizers possessing tetraazaphenanthrene as basic ligand. The data shown in Figure 8 and Figure 9 arise from the same titration experiment. As reported above, the reduction of metal chlorides to In the second step these negatively charged spherometal colloids remains incomplete under conditions colloids were investigated by using ruthenium complexes favorable for the synthesis of catalysts in artificial systems as sensors at the colloidlwaterinterface. [Ru(bpy)sI2+was for solar energy conversion. Therefore, Ru-OH species used as a standard ruthenium sensitizer. Then a new at the colloid surface could exist. type of ruthenium complex possessing one mono- and bisIn conclusion, the binding affinity of the ligands in bisstyrene-substituted 1,4,5,84etraazaphenanthreneligand heteroleptic ruthenium complexes to the surface of the and 2,2'-bipyridine or 1,4,5,84etraazaphenanthreneas basic ligands was employed to investigate the binding ruthenium colloid has been found to be behavior a t the ruthenium colloidlwater interface. Surprisingly, our experiments provide clear evidence R-tap-R R-tap > t a p > bpy for the hypothesis that the bis-heteroleptic complexes are It is further interesting that [Ru(tap)2(MSH)I2+and [Ruoriented with their mono- and bis-styrene-substituted 1,4,5,84etraazaphenanthreneligands toward the surface (tap)2(DSH>l2+ as the complexes possessing unsubstituted of the ruthenium colloid. With the assumption of a purely styrene groups attached to tetraazaphenanthrene do not electrostatic model the smaller basic ligands 2,Y-bipyriyield significant shifts in their IMLCT absorption in the are titration experiments, but the electron donor and acceptor dine and unsubstituted 1,4,5,8-tetraazaphenanthrene found to be bound next to the surface. Therefore, the substituted ruthenium complexes do yield significant occurrence of hydrogen-bonding or dipole-dipole interacshifts. This can provide evidence for the contribution of an electronic binding mechanism between the ruthenium tions can be assumed, together with electronic effects complex and the high-energy surface. These effects have which leads to the observed orientation.
-
Four Colloids for Artificial Photosynthesis
Langmuir, Vol. 10,No. 12, 1994 4495
0
412 410 .-
0
20
10
30
40
30
40
MX 440 446 444 442 440 438
0
10
20
Figure 10. For comparison purposes the UV-vis titration data from two bis-styrene-substitutedsensitizers in the presence of
the Ru colloid are shown.
0.75
1.o
t I
1
1
0
.
*
[R~(tap)~(DSc)l~+
0.7
1
1
1
Figure 11. In analogy to Figures 4,6, and 8, the normalized molar absorption coefficients from the titration experiment employing
two bis-styrene-substitutedsensitizers possessing tetraazaphenanthrene as basic ligand are presented. The occurrence of an additional electronic interaction in been found to be responsible for the binding of styrene and polystyrene a t the high-energy surfaces of metals. the case ofthe styrene-substituted tetraazaphenanthrenes
4496 Langmuir, Vol. 10, No. 12, 1994
Schwarz et al.
1.o
0.9
0.8
0.7
0.6 2
0
4
8
6
10
12
A h [nm]
A h [nml
[Ru(tap),(MSH)I'+
0.8 -
[RU(~~~)~(DSH)I~'
IB
IB
Io
t
0.7( . 1 . 1 1 . 1 . 1 . 1 . - 3 - 2 - 1
0
1
2
3
4
-
3
-
2
-
1
0
1
2
3
4
A h [nml A h [nml Figure 12. Plots ofthe normalized molar absorption coefficientsvs the shiftsin the 'MLCT absorption maxima revealing similarities and differences in the absorption mechanisms of mono- and bis-styryl-substitutedbis-heterolepticruthenium complexesemploying 2,2'-bipyridine and tetraazaphenanthrene as basic ligands. Whereas a linear dependence can be found for the series [Ru(bpy)zr2 = 0.946 for [Ru(bpy)z(MSH)I2+) and for the first seven to eight data pairs of the series (MS-R)I2+(y = 1.0188 - 0.02564~~ (y = 0.956 - O.O309x, 1.2 = 0.906 for [Ru(bpy)z(DSH)I2+), no simple dependence could be derived for the mono[Ru(bpy)~(DS-R)1~+ and bis-styryl-substitutedbis-heteroleptic complexes employing tetraazaphenanthrene as basic ligand. Further explanations are given in the discussion. could also explain why all complexes are preferentially Scheme 2. Dominating Binding Mode of the oriented with their largest ligands toward the colloid Bis-Heteroleptic Complexes [R~(bpy)z(R-tap-R)l~-~ surface. Another interesting observation from [Ru(tap)z) l ~=+H, C) (MSR)I2+(R = H, M, C) and [ R ~ ( t a p ) ~ ( D s R(R is that the (photo)physical properties in solution as well as in microheterogeneous systems can be effectivelytuned by choosing the right substitutes at the styrene unit(s). In Figure 12 the correlation of the relative hypochromic shifts dependent upon Mx[4Mx/40))1 and the shift of the lMLCT maximum wavelengths M for four bis-heteroleptic hv ruthenium complexes possessing unsubstituted styrene rlmmrrrw units is shown. Two general cases are discernible: ruthenium (1)In the case of 2,2'-bipyridine as basic ligand, a linear colloid correlation of these two main effects from the optical measurements can be found. For [Ru(bpy)2(MSH)I2+a linear relationship is evident for the whole region of measurement from the plot, whereas for [Ru(bpy)~(DSH)I2+a plateau region is reached. The linear correlation supports the assumption that the hypochromic effects arise mainly from the orientation of the complexes at the interface colloid/water and that the additional effects are of minor importance. (2) For all bis-heteroleptic ruthenium complexes possessing 1,4,5,84etraazaphenanthreneas basic ligand, a a As it became evident from our Uvlvis titrations,the complex binds with its modified tetraazaphenanthrene ligand toward linear correlation between c(MX/e(O))and M cannot be the colloid's surface. This orientation decreases with increasing observed. This indicates a more complicated binding loading density at the surface. behavior including competitive binding of styreneExperimental Section substituted and unsubstituted 1,4,5,8-tetraazaphenanthrene units as well as an electronic exchange reaction [Ru(bpy)slClz,RuOz, KMn04, and NazIrCls have been purbetween excited ruthenium complexes. chased from Aldrich; bidistilled water was used in all experiments.
-
Four Colloids for Artificial Photosynthesis
Langmuir, Vol. 10, No. 12, 1994 4497
efficienciesin our artificial systems for solar energy conversion.ls (2,2’-bipyridine)-2-s~l-l,4,5,atetr~aphenanthreneruthenium- All colloids were long-term stable with and without the presence of the ruthenium complexes, and consequently, no interference (11) chloride [Ru(bpy)z(MSH)I2+,bis(2,2’-bipyridine)-2+-methoxystyryl)-l,4,5,8-tetraazaphenanthreneruthenium(II) chloride of the colloid stability and the optical measurements has been [Ru(bpy)~(MSM)1~+, bis(2,2’-bipyridine)-2-@-cyanostyryl)-l,4,5,8- observed. The scanning electron microscope employed was of a tetraazaphenanthreneruthenium(I1)chloride [Ru(bpy)2(MSC)12+, CamScan Series 4-type. The setup for the QELS measurements bis(2,2’-bipyridine)-2,7-distyryl-l,4,5,8-tetraazaphenanthreneru- consisted of a H e m e laser (L = 632.8 nm) and photomultiplier thenium(I1)chloride [Ru(bpy)2(DSH)I2+, bis(2,2’-bipyridine)-2,7detection system. The experiments were performed under bis@-methoxystyryl)-1,4,5,8-tetraazaphenanthreneruthenium- computer control. The typical diffusion coefficients for the (11) chloride [Ru(bpy)z(DSM)lz+,bis(2,2’-bipyridine)-2,7-bis(psystems investigated were derived from the Einstein-Smolucyanostyry)-1,4,5,8-tetraazaphenan~eneruthenium(~) chloride chowsky equation. [Ru(bpy)z(DSC)lZ+,bis( 1,4,5,8-tetraazaphenanthrene)-2-styrylA Mettler B6 was used i n the gravimetric experiments. 1,4,5,8-tetraazaphenanthreneruthenium(II) chloride [Ru(bpy)zThe UV/vis spectra have been taken using a Kontron Uvikon (MSH)I2+,bis(l,4,5,8-tetraazaphenan~ene)-2-(p-methoxys~l)680 spectrometer, and all spectra are corrected for the self1,4,5,8-tetraazaphenanthreneruthenium(II) chloride [Ru(bpy),(MSM)I2+,bis(1,4,5,8-tetraazaphenanthrene-2-(p-cyanostyryl)- absorption ofthe colloid. The force field calculations were carried out using CHARMm, Version 2.1.89 0131 (c) (Polygen Corp.). 1,4,5,8-tetraazaphenanthreneruthenium(II) chloride [Ru(bpy)z(MSC)I2+,bis(1,4,5,8-tetraazaphenanthrene)-2,7-distyryl-l,4,5,8- The maximum deviation in the UV/vis absorption experiments has been determined to be fl nm in the measured IMLCT tetraazaphenanthreneruthenium(I1)chloride [Ru(bpy)z(DSH)lZ+, absorption maxima and 5 re1 % in the molar absorption bis(1,4,5,8-tetraazaphenanthrene~2,7-bis~-methoxystyryl)-l,4,5,8coefficients and the QELS experiments. The error in the tetraazaphenanthreneruthenium(I1) chloride [Ru(bpy)z(DSM)lZ+, and bis(1,4,5,8-tetraazaphenanthrene)-2,7-bis(p-cyanostyryl)- gravimetric measurements was 2 re1 %. 1,4,5,8-tetraazaphenanthreneruthenium(II)chloride [Ru(bpy)z(DSC)I2+have recently been performed in our l a b o r a t ~ r i e s . ~ ~ Acknowledgment. Generous financial support from The syntheses of the bis-heteroleptic ruthenium complexes bis-
The syntheses of the colloids Ru, RuOz, MnOz, and IrOz were performed according to literature prwedures.l6-l8 An optimization has been performed in order to obtain maximum catalytic
(30) Guldner, A. Thesis, Universitat des Saarlandes, Saarbriicken, 1990.
the German Bundesministerum fiir Forschung and Technologie BMFT for this research is gratefully acknowledged. The authors thank Ms. Janik for many helpful discussions and Dr. Thieme from the “Institut fur neue Materialien”, University of Saarland, for technical assistance.
