Photochemistry and Radiation Chemistry - American Chemical Society

---

21

Downloaded by STANFORD UNIV on February 18, 2015 | http://pubs.acs.org Publication Date: April 17, 1998 | doi: 10.1021/ba-1998-0254.ch021

Photochemically Induced Charge Separation in Electrostatically Constructed Organic-Inorganic Multilayer Composites Steven W. Keller , Stacy A. Johnson, Edward H. Yonemoto, Elaine S. Brigham, Geoffrey B. Saupe, and Thomas E. Mallouk* 1

Department of Chemistry, The Pennsylvania State University, University Park, PA 16802

A multilayer film growth technique, in which single anionic sheets de­ -rived from inorganic solids are interleaved with cationic polyelectrolytes, has recently been developed. This method allows for the growth of con­ -centric monolayers of redox-active polymers on high-surface-area silica supports, andfor vectorial electron transfer reactions through the layers of the "onion". Transmission electron microscopy was used to probe the morphology of these lamellar heterostructures. Photoinduced charge separation has been observed in composites consisting of an inner poly­ -cationic layer of poly(styrene-co-N-vinylbenzyl-N'-methyl-4,4'-bipyri­ -dine ) (PS-MV ), and an outer polycationic layer of poly[Ru(bpy) (v­ -bpy)] , vbpy = 4-vinyl-4'-methyl-2,2'-bipyridine, bpy = 2,2'-bi­ -pyridine, which are separated by a thin inorganic sheet of a-Zr(PO ) . The thickness of the individual polymer layers was determined by ellip­ sometry for equivalent structures on planar supports. Electron transfer quenching of the Ru(II) polymer luminescence occurs upon addition of a solution-phase, reversible electron donor, disodium methoxyaniline­ -Ν,Ν'-diethykulfonate (MDESA ). In the absence of an inner viologen layer, this simple donor-acceptor charge-separated state decays in sev­ -eral microseconds to regenerate the ground state. In the triad system, which contains an inner viologen polymer layer, rapid electron transfer from Ru(I) to viologen creates a charge-separated state with a half-life 2+

2

2+

2-

4 2

2-

Current address: Department of Chemistry, University of Missouri-Columbia, Columbia, MO 65211 * Corresponding author. 1

©1998 American Chemical Society

In Photochemistry and Radiation Chemistry; Wishart, J., et al.; Advances in Chemistry; American Chemical Society: Washington, DC, 1998.

359

360

PHOTOCHEMISTRY AND RADIATION CHEMISTRY

of 21 μ£. The simultaneous second-order decay (Kemmb = I X I 0 M s' ) of signals from both MDESA" and reduced MV' is consistent with escape of the former from the Ru(II) polymer and subsequent diffusion to MV' sites. Quantum yields for charge separation are ca. 30%. 9

1

_ i

+

Downloaded by STANFORD UNIV on February 18, 2015 | http://pubs.acs.org Publication Date: April 17, 1998 | doi: 10.1021/ba-1998-0254.ch021

+

T h e reaction centers found in natural photosynthetic systems are remarkable not only for their unit quantum yields, but also for the long lifetimes of their charge-separated states (1-5). This feat is accomplished by an intricate vectorial arrangement of redox-active molecules that absorb light and separate charge across a lipid bilayer membrane. The membrane proteins that confine the reac­ tion center play a key role in maximizing the branching ratio between forwardand back-electron transfer rates for each successive step in the charge-separa­ tion process (6-9). The determination of the crystal structures of bacterial photosynthetic reaction centers has given chemists a blueprint from which to design biomimetic, artificial photosynthetic systems. In these designs, the energetics of elec­ tron transfer reactions, and the strength of electronic coupling between redoxactive subunits, are key parameters in controlling electron transfer rates. The dependence of electron transfer rates on energetics, which was first predicted theoretically by Marcus (10-12), has now been studied experimentally in a host of geometrically well defined donor-acceptor systems (13-27). Because electronic coupling strength is a strong function of intermolecular distance, there have also been many fundamental studies of the distance dependence of electron transfer rates in various media (28-57). In the last decade, many of the fundamental questions regarding electron transfer in both biological and model donor-acceptor systems have been sub­ stantially answered. Thanks to these fundamental studies, it is now possible to make elegant supermolecules that rival natural reaction centers in terms of their electron transfer rates and quantum yields (58-64). The success of these biomimetic systems is extremely impressive. Still, the synthesis of these multicomponent molecules is demanding, and it is difficult to couple them to catalytic particles in order to produce energy-rich chemicals from transiendy stored free energy. In order to achieve similar control over the distance (and therefore over electron transfer rates) between subunits, but with less synthetic effort, self-assembling and microheterogeneous photoredox systems have been investi­ gated. These organizing media include porous glasses (65-67), micelles (68-70), zeolites (71, 72), and polyelectrolytes (73), to name a few. Catalytic activity, typically in the form of water reduction to hydrogen, has manifested in some of these materials (65, 68, 74). However, most microheterogeneous media do not offer much flexibility in designing complex electron transfer chains, because typically only two moieties can be juxtaposed in a controlled fashion. A new technique for growing multilayer thin films composed of oppositely charged polyelectrolytes has been described by Decher and co-workers

In Photochemistry and Radiation Chemistry; Wishart, J., et al.; Advances in Chemistry; American Chemical Society: Washington, DC, 1998.

21.

KELLER E T AL.

361

(75-80). Recently, we (81) and others (82-84) have shown that similar heterostructures can be prepared by using two-dimensional inorganic sheets (made by exfoliation of various lamellar solids) in place of the organic polyanion. This technique offers a potentially powerful alternative to the construction of multicomponent electron transfer systems, because it can, in principle, be used to stack up an arbitrary number of redox-active polymers without interpénétration (85). This chapter describes the preparation and photochemistry of simple multilayer composites on high-surface-area silica. Specifically, the synthesis and electron transfer kinetics of systems containing a polycationic sensitizer, poly[Ru(bpy) (vbpy)(Cl) ] (1), (abbreviated [Ru(bpy) ] ; bpy = 2,2'-bipyridine and vbpy = 4-vinyl-4'-methyl-2,2'-bipyridine), and an electron-acceptor polycation poly[(styrene-co-N-vinylbenzyl-N -methyl-4,4 -bipyriine (2), (PS-M\^ ) are presented. Using a solution-phase electron donor, 3, as the third electroactive component, it was possible to prepare and study the photoinduced electron transfer reactions of several different diad and triad combinations. 2

Downloaded by STANFORD UNIV on February 18, 2015 | http://pubs.acs.org Publication Date: April 17, 1998 | doi: 10.1021/ba-1998-0254.ch021

Photochemically Induced Charge Separation

2

3

2+

/

n

,

+

™CH -CH2

SO3" The single layers of inorganic phosphate in this system physically separate the oxidized and reduced species in the polymer layers, but they are thin enough to allow electron transfer between polymers on opposite sides. The growth of these composites on solid supports also, in principle, makes possible the incorporation of catalytic sites at which energy-storing chemical reactions could occur.

Experimental Details General Materials. 4,4'-Bipyridine (Aldrich) was purified by dissolv­ ing it in dichloromethane and stirring with activated carbon for 30 min. The mixture was passed through a short silica-alumina column and eluted with In Photochemistry and Radiation Chemistry; Wishart, J., et al.; Advances in Chemistry; American Chemical Society: Washington, DC, 1998.

362

PHOTOCHEMISTRY AND RADIATION CHEMISTRY

additional dichloromethane, and the solvent removed under reduced pressure. ρ -Anisidine was purified by sublimation, and sodium 2-bromoethanesulfonate was recrystallized before use. A l l other chemicals and organic solvents used were Aldrich reagent or high-performance liquid chromatography ( H P L C ) grade and were used without further purification. Distilled water was passed through a Millipore deionizer to a resistance of 18.2 M i l .

[Ru(2,2' - bipyridine) (4 - vinyl - 4' - methyl - 2,2' - bipyridine)](PF ~) . Ru(2,2 -bipyridine) Cl -2H 0 (86) (0.5 g, 0.96 mmol) and 42

Downloaded by STANFORD UNIV on February 18, 2015 | http://pubs.acs.org Publication Date: April 17, 1998 | doi: 10.1021/ba-1998-0254.ch021

6

,

2

2

2

2

vinyl^'-metbyl^^'-bipyridine (87) (0.22 g, 1.12 mmol) were refluxed in 70 m L of a 4:6 mixture of H 0 / E t O H for l h and allowed to cool to room tempera­ ture. The solvent mixture was removed under reduced pressure. The remaining solid was redissolved in water only and added to an aqueous solution of N H P F , from which a red-orange precipitate immediately formed. The prod­ uct was collected by vacuum filtration, and washed with water and then ether to remove excess 4-vinyl-4 -methyl-2,2'-bipyridine that did not chelate. The P F " salt was dissolved in a minimum amount of acetonitrile and chromatographed on silica gel, using 5 : 4 : 1 CH CN/H 0/saturated aqueous K N 0 as the eluant, which yielded one major fluorescent band. The eluted band was collected and the solvent removed to the point where the K N 0 started to crystallize. At this point, acetone was added to precipitate the rest of the K N 0 . Some ether can be added to the cold solution, but excess will begin to precipitate the product and color the K N 0 crystals. The K N 0 was filtered off and washed with acetone. The bright red acetone solution (containing the nitrate salt) was evaporated and the sohd dissolved in water, reprecipitated by adding a concen­ trated aqueous solution of N H P F , and vacuum filtered. The precipitate was then washed with water to remove excess N H P F and K N 0 , then with ether, and then air dried to yield 0.61 g (84%). 2

4

6

/

6

3

2

3

3

3

3

3

4

6

4

6

3

Poly[Rii(2,2' - bipyridine) (4-vinyl-4' - methyl-2,2' - bipyridine) dichloride]. Compound 1 was prepared via free-radical polymerization of 2

the monomer in acetone using azo-bis-isobutyronitrile (AIBN) as an initiator after Abruna and co-workers (88). Specifically, 0.100 g (0.11 mmol) of [Ru(2,2'bipyridine) (4-vinyl-4 -methyl-2,2 -bipyri and 5.0 mg of A I B N were dissolved in 5 m L of acetone, transferred to a Pyrex tube (13 mm o.d.), and subjected to three freeze-pump-thaw cycles to remove dissolved oxygen. The tube was sealed under reduced pressure and heated at 60 °C for 72 h. After the tube was cooled to room temperature it was broken open and the orange-red solution collected; no sohd precipitate was evident. A n ion-ex­ change column was prepared with Dowex anion exchange resin by loading the column with 1 M HC1 and washing with water until the p H of the eluant was neutral. The acetone solution was introduced to the column, eluted with H 0 , and the small amount of residual acetone removed by rotary evaporation to 2

,

/

2

In Photochemistry and Radiation Chemistry; Wishart, J., et al.; Advances in Chemistry; American Chemical Society: Washington, DC, 1998.

21.

K E L L E R ET AL.

363

Photochemically Induced Charge Separation

form the water-soluble C l ~ salt (1). The H - N M R peaks observed in de-di­ methyl sulfoxide (rf -DMSO) were broadened significantly from those of the monomer. 1

6

Poly(chloromethylstyrene-co-styrene). Polystyrene was chloromethlyated using the procedure of Merrifield (89). Polystyrene (25.0 g Aldrich, 45,000 M W ) was dissolved in 150 m L of C H C 1 in a 250-mL round bottom flask equipped with an addition funnel and placed in an ice bath at 0 °C. The entire volume was flushed with A r for 30 min and continuously during the reaction. To the addition funnel was added 3.75 m L of SnCL* (0.032 mol) and 25 m L (0.329 mol) of chloromethyl methyl ether, also at 0 °C. The SnCLi-chloromethyl methyl ether solution was added dropwise to the polystyrene solution with constant stirring over 10 min, resulting in a reddish-brown solution. After an additional 30 min, the reaction was quenched by dropwise addition of H 0 , the reddish coloration disappeared, and the solution was subsequently washed three times with H 0 . The cloudiness of the organic layer (due to trace amounts of S n 0 ) was eliminated by vacuum filtration. The solvent was removed by rotary evaporation, which first left an oily substance and finally a glassy sohd. After further drying under reduced pressure, the sohd puffed up and was easily recoverable. H - N M R spectroscopy in C D C 1 showed two broad peaks in the aromatic region, a broad singlet in - 5 ppm ( C H C l ) , and several unresolved, overlapping peaks resembling profile views of snakes who had swallowed ele­ phants i n the aliphatic region. From ratios of aromatic to chloromethylene peaks, the amount of chloromethylation was estimated at 35%, which was con­ firmed by elemental analysis.

Downloaded by STANFORD UNIV on February 18, 2015 | http://pubs.acs.org Publication Date: April 17, 1998 | doi: 10.1021/ba-1998-0254.ch021

3

2

2

2

1

3

2

2V-Methyl-4,4'-bipyridine. W e dissolved 1.5 g (10 mmol) of 4,4'bipyridine in 150 m L of dichloromethane to which 5.0 g (35 mmol) of iodomethane (Fisher Scientific) was added. The solution was stirred for 2 h, after which a yellow precipitate was evident. Stirring for an additional 8 h at room tempera­ ture completed the precipitation, and the resulting yellow sohd was collected, washed with C H C 1 and diethyl ether, and dried in air. H - N M R in d - D M S O resulted in four sets of doublets in the aromatic region. N-Methyl-4,4'-bipyridine iodide was converted to the PF "salt by dissolving it in water and adding an aqueous solution of N H P F . The white sohd was filtered, washed with water and ether, and dried in vacuum. 2

1

2

6

6

4

6

Poly(styrene^o-N-vroylbenzyl^

di-

chloride (2). N-Methyl-4,4 -bipyridine was attached to chloromethylated polystyrene via a quaternization reaction. Chloromethylated polystyrene (1.00 g) and 1.26 g of IV-methyl-4,4'-bipyridine were dissolved in 100 m L of CHCI3 and 100 m L of C H C N . The pale yellow solution was refluxed for 72 h, after which much of the solvent was removed. An ion-exchange column was prepared /

3

In Photochemistry and Radiation Chemistry; Wishart, J., et al.; Advances in Chemistry; American Chemical Society: Washington, DC, 1998.

364

PHOTOCHEMISTRY AND RADIATION CHEMISTRY

with Dowex anion exchange resin by loading the column with 1 M H C l and washing with water until the p H of the eluant was neutral. The sohd residue was taken up in C H C N and C H O H , placed on the ion-exchange column, and eluted with water. The solvents were removed by rotary evaporation and the sohd dissolved in 80:20 H 0 : C H O H to make the stock solution. 3

3

2

3

(3). Sodium

Disodium Methoxyaniline-N-N'-diethylsutf^

2-bromoethanesulfonate (20.0 g, 0.189 mol) and p-anisidine (11.67 g, 0.095 mol) were placed in a 500-mL round-bottom flask and dissolved in 400 m L of dry ethanol. To this solution was added 13.13 g of K C 0 , and the mixture was heated at reflux for 12 h. The reaction was cooled and any solids remaining were filtered off, collected, and recrystallized first from boiling absolute ethanol, and finally from 20:80 water:ethanol mixtures, before being dried in air.

Downloaded by STANFORD UNIV on February 18, 2015 | http://pubs.acs.org Publication Date: April 17, 1998 | doi: 10.1021/ba-1998-0254.ch021

2

3

Layering Procedure. We have previously described briefly the pro­ cedure for growing the multilayer films on planar (81) as well as on smallparticle supports (90). The siliea-based composites were prepared by initially derivatizing high-surface-area silica (Cab-O-Sil, Kodak; 150 m /g) with 3aminopropyltriethoxysilane. We have also used 4-aminobutyldimethylmethoxysilane (United Technologies) as a nonpolymerizable anchoring agent. No signifi­ cant differences in the electron transfer kinetics have been observed between the two different composites. In a typical anchoring procedure, 5.0 g of CabO-Sil was suspended in 150 m L of anhydrous toluene in a poly(tetrafluoroethylene) bottle, to which was added 5.0 m L of silane. The solution was deaerated with flowing Ar, and the bottle sealed and heated with stirring at 60 °C for 72 h. The derivatized silica was isolated by centrifugation with intermittent washings with toluene, methanol, and finally water, and dried in air at 80 °C. Suspending the derivatized particles in water protonates the amine terminus and creates a cationic surface. Small-particle a - Z r ( H P 0 ) (a-ZrP) was synthe­ sized by using the method of Berman and Clearfield (91 ) and exfoliated with tetrabutylammonium hydroxide (TBAOH) to a constant 8 ^ p H ^ 8.5. The faindy milky suspensions (-1-5 mmol/L) were centrifuged before use to remove any large particle agglomerates from the solution. Silane-derivatized Cab-O-Sil was suspended in exfoliated ot-ZrP solution (~1 g of silica per 200 m L of solution) and stirred for 12 h. The sohd was isolated by centrifugation with three subsequent washings using 50 m L of water, before being dried in air at 75 °C. Deposition of the polycationic sensitizer was performed similarly, by dispersing typically 1 g of α-ZrP-anchored Cab-O-Sil in 250 m L of 10~ M aqueous solutions of 1. If smaller volumes of polyelectrolyte solution (50 m L per gram of composite) were used in this step, the resulting loading of polymer on the surface decreased by approximately 65%. Larger volumes aid the removal of the tert-butyl alcohol (TBA ) into the solution because they bias the ion-exchange equilibrium i n favor of surface-adsorbed 2

4

2

5

+

In Photochemistry and Radiation Chemistry; Wishart, J., et al.; Advances in Chemistry; American Chemical Society: Washington, DC, 1998.

Downloaded by STANFORD UNIV on February 18, 2015 | http://pubs.acs.org Publication Date: April 17, 1998 | doi: 10.1021/ba-1998-0254.ch021

21.

K E L L E R E T AL.

Photochemicaïly Induced Charge Separation

365

polyeleetrolyte. We note that this mass-action effect is important with highsurface-area substrates, where bulk amounts of counterions are released, but is probably unimportant with low-area planar substrates. In order to deposit a thicker layer of the MV^-containing polymer (2), the deposition was carried out in 1 M NaCl. Higher ionic strength results in coiling of the polymer, as explained below. Under these conditions there was always excess polymer in the supernatant solution, and the sohd was washed and centrifuged until the supernatant was clear. Loadings of the polymers were followed qualitatively by UV-visible diffuse reflectance spectroscopy of the dried powders, and quantified by measuring the change in polymer concentra­ tion in solution after each deposition step.

Instrumentation. Transmission electron micrographs were taken using a J E O L 1200EX II microscope operating at an accelerating voltage of 120 kV. Samples were prepared by dispersing the dried powders on carbon coated copper grids. Solution UV-visible spectra were collected on a HewlettPackard 8452A Diode Array Spectrometer, and diffuse reflectance spectra on a Varian DMS-300 equipped with an integrating sphere attachment. A l l H N M R spectra were recorded on a Bruker AM-300 spectrometer. Film thick­ nesses were measured with a Gaertner two-wavelength ellipsometer, using Si(001) wafers, cleaned as described elsewhere (81), as substrates. Typically, 5-8 spots per wafer were measured and averaged. Transient flash photolysis and time-resolved luminescence experiments were performed using a system similar to one described previously (92), which is diagramed in Figure 1. Aqueous suspensions of ca. 100 mg of sohd sample 1

Mechanical Shutter

Figure 1. Schematic of the nanosecond transient diffuse reflectance and transient fluorescence instrumentation.

In Photochemistry and Radiation Chemistry; Wishart, J., et al.; Advances in Chemistry; American Chemical Society: Washington, DC, 1998.

Downloaded by STANFORD UNIV on February 18, 2015 | http://pubs.acs.org Publication Date: April 17, 1998 | doi: 10.1021/ba-1998-0254.ch021

366

PHOTOCHEMISTRY AND RADIATION CHEMISTRY

were placed in 1-em quartz cuvettes, and continually stirred and purged with argon throughout the experiment. The second harmonic from a Spectra Physics Quanta Ray N d : Y A G laser [λ = 532 nm, -20 ns full width of half-maximum ( F W H M ) ] was used as the photoexcitation source in all experiments. A n Oriel 150-W H g - X e lamp provided the analyzing light. To improve throughput to the detector, a high-voltage pulse generator (Kinetic Instruments, Austin, TX), which increased the brightness of the white light a factor of 50, was used. The diffusely reflected light from the sample was collected and collimated by two lenses and focused into an Oriel Model 77250 grating monochromater (with appropriate optical filters), to which was attached a Hammumatsu Model R928 photomultiplier tube (PMT) biased at 600-800 V with a Bertan Model 205B-03R high-voltage power supply. Signals from the P M T were recorded with a Tektronics T D S 5 4 0 - A digital oscilloscope. Timing of the laser flash lamps and Q-switch, mechanical shutter, and lamp puiser was controlled by inhouse built electronics. Data collection was initiated by impinging ~5% of the incident laser beam on a photodiode trigger, and typically 20-30 shots were averaged by the oscilloscope before being sent to a personal computer (PC) for further analysis. Because of the diffuse reflectance geometry of the experiment, raw voltage data from the P M T was first converted to â(absorbance), as i n a transmission experiment, via: ΔΑ = l o g ( W o ) where A is absorbance, V is the averaged P M T voltage before the laser trigger, and V is the signal voltage. These data were then converted to relative absor­ bance, -Δ/// , which is the quantity that is proportional to the amount of excited-state species, by: 0

0

-Δ///ο = 1 - 1 0 "

ΔΑ

For spectra, an additional empirical correction was included (93) to account for the attenuation of the shorter-wavelength signals due to increased scattering of the exciting and analyzing light.

Results and Discussion In order to probe the microstructure of "onion" structures containing polyelec­ trolytes 1 and 2, model composites containing the same components were grown on planar Si supports, and the thickness of each successive layer was

In Photochemistry and Radiation Chemistry; Wishart, J., et al.; Advances in Chemistry; American Chemical Society: Washington, DC, 1998.

21.

KELLER ET AL.

367

Photochemically Induced Charge Separation

measured by ellipsometry. Shown in Figure 2a is the thickness evolution of a multilayer film of 1 separated by sheets of α-ZrP. The average thickness of both 1 and the α-ZrP is ca. 8 A, and it is fairly consistent throughout the entire film. Although this value is reasonable for an α-ZrP sheet, it is smaller than that expected for the Ru(bpy) complex itself. Estimates of the surface coverage (assuming a cross-sectional area of 150 A for the Ru(bpy) repeat unit, and 100 A /g for the area of exposed Cab-O-Sil) indicate that this layer is incomplete, perhaps because of inefficient charge-compensation by the embedded dication, which must compete with T B A for surface ion-exchange sites on the ot-ZrP. Figure 2b shows the dependence of the thickness of the P S - M V * layer on ionic strength; data for films prepared in both high ionic strength (1 M N a C l solution) and in zero ionic strength are presented. The average thicknesses for 2 are ca. 38 A and 7 A, respectively, in agreement with the findings of Decher and co-workers, who controlled the thicknesses of poly(allylamine) and poly(styrenesulfonate) layers by addition of NaCl to the polyelectrolyte solutions (78). The explanation for this effect is that polyelectrolytes adopt a more coiled conformation in high-ionic-strength solutions, because like charges can be screened effectively by the excess ions. The coiled conformation persists upon deposition onto the sohd substrate, producing thicker layers. Polymers that are completely extended, in zero-ionic-strength media, are deposited as thinner layers onto planar substrates. Figure 3a shows a transmission electron micrograph of the pristine silica, illustrating clearly that the nominally 200-A-diameter spheres are ag­ glomerated into irregularly shaped particles. Examination of many different particles indicates that some additional degree of agglomeration occurs during the anchoring process, likely due to partial polymerization of the trialkoxysilane anchoring agent. However, the surface area, measured by nitrogen adsorption (Brunauer-Emmett-Teller (BET) method) drops only slightly after this treatment, from 150 m /g to 146 m /g. Micrographs of the 1/2/ Cab-O-Sil composite (slash marks indicate a layer of α-ZrP) in Figure 3b show further agglomeration, as well as visible evidence of multilayer forma­ tion. Although we were unable to image the individual layers that make up the composite, there are significant textural differences. There appears to be a mottled coating around most of the sihca particles in Figure 3b, the thickness of which is 50-80 A. This thickness is reasonable for the 1/2/film that is present. In addition, there are some regions on the particles that remained uncoated, possibly because of incomplete formation of the anchor­ ing layer on the irregularly shaped substrate. To alleviate this complication, we have begun synthesizing S1O2 particles within reverse micelles (94), and preliminary T E M images confirm a high degree of monodispersity and good separation of individual spheres in samples prepared in this fashion. 3

2+

2

3

2

+

Downloaded by STANFORD UNIV on February 18, 2015 | http://pubs.acs.org Publication Date: April 17, 1998 | doi: 10.1021/ba-1998-0254.ch021

2

2

2

Deposition of the redox-active polymers onto high-surface-area sihca was monitored using diffuse reflectance UV-visible spectroscopy. Shown in Figure 4a are the spectra for the PS-MV^/Cab-O-Sil composites, both coiled and

In Photochemistry and Radiation Chemistry; Wishart, J., et al.; Advances in Chemistry; American Chemical Society: Washington, DC, 1998.

368

PHOTOCHEMISTRY AND RADIATION CHEMISTRY

(a)

Downloaded by STANFORD UNIV on February 18, 2015 | http://pubs.acs.org Publication Date: April 17, 1998 | doi: 10.1021/ba-1998-0254.ch021

100

Layer Number (b) A

H

1

:

0

1

2

1

3 4 Layer Number

1

!

5

6

r 7

Figure 2. (a) Ellipsometric data for multilayer composites of 1-1-1- . . . grown on a planar Si wafer, (b) Ellipsometric data for ZrP-2 multilayerfilmsdeposited with (circles ) and without (squares ) added electrolyte. The average thicknesses are 38 and 7 A, respectively.

In Photochemistry and Radiation Chemistry; Wishart, J., et al.; Advances in Chemistry; American Chemical Society: Washington, DC, 1998.

Downloaded by STANFORD UNIV on February 18, 2015 | http://pubs.acs.org Publication Date: April 17, 1998 | doi: 10.1021/ba-1998-0254.ch021

21.

K E L L E R E T AL.

369

Photochemicaïly Induced Charge Separation

Figure 3. Transmission electron micrographs of (a) reagent Cah-O-Sil, and (b) the 1/2/Cab-O-Sil composite.

uncoiled, before the addition of polymer 1 to the surface. It is clear that the coiling of 2 in solution results in increased loading on the high-surface-area particles, as was the case for the planar supports. The results of addition of a layer of α-ZrP and a layer of 1 to both composites are shown in Figure 4b. Roughly the same amount of [Ru(bpy) ] is deposited regardless of the thick­ ness of the first layer, although there is some loss of PS-MV " " absorption (-15%) after the two subsequent steps. The loadings of the polymers were determined quantitatively from solution UV-visible data of the supernatant solutions from the deposition steps, and typically were 1-2 Χ 10" mol/g composite. The redox-potential diagram for the donor-Ruibpy^^-MV " " triad system in Figure 5 is helpful in understanding the sequence of electron transfer events in these composites. Although excitation of the sensitizer is always the initial step in the overall process, there are several possible pathways for subsequent reactions. Transient absorbance measurements on the donor-sensitizer and 3

2+

n

2

1

5

2

1

In Photochemistry and Radiation Chemistry; Wishart, J., et al.; Advances in Chemistry; American Chemical Society: Washington, DC, 1998.

Downloaded by STANFORD UNIV on February 18, 2015 | http://pubs.acs.org Publication Date: April 17, 1998 | doi: 10.1021/ba-1998-0254.ch021

370

PHOTOCHEMISTRY AND RADIATION CHEMISTRY

Wavelength (nm) Figure 4. (a) UV-visibk diffuse reflectance spectra of 2(coiled)/Cab-OSil and 2(uncoiled)/Cab-0-Sil composites, and (b) the same samples after deposition

ofl.

sensitizer-acceptor diad composites (1/Cab-O-Sil with M D E S A and 1/2 (coiled)-Cab-O-Sil, respectively) help to elucidate the operative electron transfer pathway in the triad system. The transient absorption spectra of the l/2(coiled)/Cab-0-Sil diad shown in Figure 6 indicate that no oxidative quenching of the *Ru(II) occurs under these conditions. A l l of the spectral features can be attributed to metal-toligand charge transfer ( M L C T ) absorption and decay. The lifetimes of the 360nm absorption and the 450-nm bleaching are identical and similar to 1/CabO-Sil alone. More direcdy, no-peak associated with reduced viologen species is seen at ca. 390-400 nm (95). We postulate that although the reduction of MV " " by *Ru(II) in reaction 1 is energetically favorable by ~ 300 mV and occurs 2

1

In Photochemistry and Radiation Chemistry; Wishart, J., et al.; Advances in Chemistry; American Chemical Society: Washington, DC, 1998.

21.

KELLER ET AL.

-1.50

V

-1.08 V

Ru(II/I)

Ru(II*/III)

*

OS

,

/zv!

Φ

^recomb

0.0V S C E Downloaded by STANFORD UNIV on February 18, 2015 | http://pubs.acs.org Publication Date: April 17, 1998 | doi: 10.1021/ba-1998-0254.ch021

\

M V + V M V 2+

> 4-»

>

371

Photochemically Induced Charge Separation

1^1

S +0.50 1/S MDESA eu 2/

#

• +0.60 V

Ru ΙΓ7Ι)

Ru(II/III)

•+1.02V

Figure 5. Redox potential diagram of the MDESA-Ru(bpy ) -MV triad system, with the important electron transfer steps indicated. All potentials are referenced to SCE in aqueous solution. 3

rapidly in aqueous solution (k = 5 Χ 10 M " s" ) (96), the intervening 10 Â of insulating α-ZrP slows the electron transfer reaction to the point where it cannot compete with M L C T decay of *Ru(II). 8

M V * - + *Ru(II) — M V

1

1

+

+ Ru(III)

(1)

The other diad system, consisting of sensitizer 1 deposited on Cab-OSil and electron donor 3 in solution, shows more interesting photochemistry. Referring to Figure 5, the redox potential of 3 allows for the reductive quench­ ing of excited-state *Ru(II) via reaction 2. The formal potential of 3 is +0.50 V versus S C E (standard calomel electrode) in aqueous solution (0.1 M K C l , Pt working electrode); by way of comparison, the formal potential of the * R u (II)-Ru(I) couple is ca. +0.6 V (97, 98). M D E S A " + Ru(II)* — M D E S A " + Ru(I)

(2)

2

Shown in Figure 7 are the Stern-Volmer plots of emission intensities and hfetimes, monitored at 630 nm, as a function of M D E S A concentration. Both the static (intensity) and the dynamic (lifetime) components are nonlinear and indicate that the quenching mechanism is complicated. The extent of the static reaction (attributed to M D E S A " anions associated with Ru(bpy) cations) 2

3

2+

In Photochemistry and Radiation Chemistry; Wishart, J., et al.; Advances in Chemistry; American Chemical Society: Washington, DC, 1998.

Downloaded by STANFORD UNIV on February 18, 2015 | http://pubs.acs.org Publication Date: April 17, 1998 | doi: 10.1021/ba-1998-0254.ch021

372

PHOTOCHEMISTRY AND RADIATION CHEMISTRY

300.0

350.0

400.0

450.0

500.0

550.0

Wavelength (nm) Figure 6. Transient diffuse reflectance spectra for the [Ru(bpy) ] -(PS-MV *, coiled)-Cab-OSil composite taken 200 ns (circles), 500 ns (trian­ gles), and 1 μβ (squares) after the laserflash.The peak at 360 nm corresponds to the absorption of the * Ru(II) excited state and decays with a lifetime of ca. 550 ns. 2Jt

n

2

1.7 1.6 1.5 1.4 1.3°^ 1.2 1.1 1

0.9 0

1

2

3

4

5

6

[MDESA] (mM) Figure 7. Lifetime (t) and steady-state emission intensity (I) data for the [Ru(bpy) ] -Cab-OSil composite suspended in MDESA solution. Lifetime data were obtained by fitting the decay of the 630 nm *Ru(II) transient observed with 532-nm laser excitation. 2+

3

n

In Photochemistry and Radiation Chemistry; Wishart, J., et al.; Advances in Chemistry; American Chemical Society: Washington, DC, 1998.

21.

K E L L E R E T AL.

PhotochemwaUy Induced Charge Separation

levels off at 3.5 m M donor, and quenches roughly 40% of the *Ru(II) centers. Some of the * R u ( b p y ) ions may not be completely accessible to the solutionphase donors, accounting for the incomplete static quenching. Dynamic quenching yields were measured from emission lifetimes (also at 630 nm). The dynamic reaction quenches the luminescence of about 65% of the remaining *Ru(II), with a rate constant at low donor concentrations of k = 3.7 Χ 10 M " s" . There is an additional complication in that upon addition of donor the emission decay is biexponential, with one component