J. Phys. Chem. 1983, 87, 5395-5401
5395
Quenching of Ru(bpy),2+ Ionically Bound to Porous Vycor Glass by 02,N20, and SO, Steven Wolfgang and Harry D. Gafney" Department of Chemistiy, City University of New York, Queens College, Flushing, New York 11367 (Received: October 7, 1982; I n Final Form: May 12, 1983)
R~(bpy),~ binds + to anionic silanol sites of porous Vycor glass, PVG. The absorption and emission spectra of the adsorbed complex are equivalent to spectra of aqueous and alcoholic solutions of the complex and, when excited with a low-intensity pulse, the luminescence of the adsorbed complex decays with a rate constant of (1.35 i 0.04) X lo6 s-l. When samples of PVG impregnated with the complex are excited in the presence of Oz, NzO, and SOz, the MLCT state of the complex is quenched, and the quenching parallels adsorption of the gases onto the PVG. Quenching by NzO and SOz occurs by electron transfer, but formation of the redox pair proceeds by different pathways. NzO is weakly adsorbed and quenches by a dynamic process which suggests that adsorbed NzO migrates on the surface of PVG. O2and SOp are more strongly adsorbed and quench by a static process. Adsorption of these gases is random and analysis of the static quenching data yields a quenching distance r of 11 i 1 5 r 5 17 h 2 8, for Oz. Since SOz quenches by an electron transfer mechanism, similar analysis suggests the occurrence of long-range, 27 i 2 A, electron transfer in PVG. Introduction Since recognition of its photoinduced redox chemistry,l tris(2,2'-bipyridine)ruthenium(II), R ~ ( b p y ) , ~has + , been extensively used as a means to explore electron transfer reactions.2-6 The majority of these experiments have focused on reactions in homogeneous fluid solution where the products are transitory due to a thermally activated reverse reaction. Recent experiments have begun to explore the photophysical properties of the complex in heterogeneous media where the underlying strategy is to control reversibility by interposing phase b ~ u n d a r i e s . ~ Studies of the complex in media as diverse as semiconductor surface^,^,^ cation exchange resins,1° zeolites,'l m i c e l l e ~ , and ~ ~ Jmembraned4 ~ have established behavior quite different from that found in homogeneous fluid solution. Whether these media and the interfacial phenomena which they introduce will prove to be effective in controlling reaction reversibility remains to be established, but they do open new experimental possibilities. An immediate advantage of heterogeneous media, perhaps as important as mediating reversibility, is circumventing the problem of reagent solubility inherent in homogeneous media. The latter is particularly relevant to gaseous reagents where the solubility of a gas in a particular solvent may not be sufficient to efficiently quench the MLCT state of the complex and/or solvolysis of the gas molecule may radically change its properties. In this paper, we describe the use of porous Vycor glass, PVG, as a support for Ru(bpy)32+and the quenching of the adsorbed complex's MLCT state, designated *Ru(bpy),2+(ads),by 02,SO2,and
N20. (1) Gafney, H. D.; Adamson, A. W. J. Am. Chem. SOC.1972,94,8238. (2) Demas, J. N.; Adamson, A. W. J. Am. Chem. SOC. 1973,95,5159. (3) Young, R. C.; Meyer, T. J.; Whitten, D. G. J. Am. Chem. SOC.1975, 97,4781. (4) Bock, C. R.; Meyer, T. J.; Whitten, D. G. J. Am. Chem. SOC.1974, 96,4710. (5) Navon, G.; Sutin, N. Inorg. Chem. 1974,13,2159. (6) Laurence, G. S.; Balzani, V. Inorg. Chem. 1974,13,2976. (7) Gratzel, M. Acc. Chem. Res. 1981,14,376. (8) Clark, W. D. K.; Sutin, N. J . Am. Chem. SOC.1977, 99, 4676. (9) Duonghong, D.; Borgarello, E.; Gratzel, M. J. Am. Chem. SOC.1981, 103,4685. (10) Thorton, A. T.; Laurence, G. S. J. Chem. Soc., Chem. Commun. 1978,408. (11) DeWilde, W.; Peeters, G.; Lunsford, J. H. J. Phys. Chem. 1980, 84, 2306. (12) Calvin, M. Acc. Chem. Res. 1978,11, 369. (13) Rodgers, M. A. J.; Becker, J. C. J. Phys.Chem. 1980,84,2762. (14) Lee, P. C.; Meisel, D. J. Am. Chem. SOC.1980,102,5477.
Porous Vycor glass is a surface hydroxylated, transparent (A I290 nm), porous glass which has a myriad of 70 f 21 A diameter pores or cavities interconnected throughout the glass in a random three-dimensional array.I5J6 Like silica gel, PVG has surface silanol groups, although the number/unit area, the silanol number, depends on hydration and previous heating. The silanol groups are slightly acidic and data gathered in these experiments show that Ru(bpy)gl+ cation exchanges onto PVG and binds to the anionic silanol groups. Adsorption does not disrupt the primary coordination sphere of the adsorbed complex and it retains properties equivalent to those found in fluid solution. This paper describes the quenching of the LMCT state of the R ~ ( b p y ) ~ ~ + ( by a d 02, s ) SO2,and NzO. The data are interpreted within a model in which the quencher is adsorbed and the occurrence of static or dynamic quenching reflects the strength of adsorption onto PVG. NzO and SO2 quench by an electron transfer mechanism and the lower limit of the quenching radius of the latter, r = 27 f 2 A, suggests the occurrence of long-range electron transfer. Experimental Section [Ru(bpy),]ClZ.3Hz0 was prepared according to the procedure described by Palmer and Piper.17 Absorption and emission spectra of samples twice recrystallized from distilled water were in excellent agreement with published spectra.l8 The gases used in these experiments were obtained from the Linde Corporation and were used without further purification since each had a purity level of 299%. All solutions were prepared with water distilled in a Corning distillation apparatus. Pieces (25 mm X 25 mm X 5 mm) of Corning's code 7930 PVG were extracted with acetone followed by distilled water in a Soxhlet extractor and then heated in a muffle furnace at 550 "C for 72 h to remove carbonaceous impurities. After being heated, the pieces were stored in a desiccator under vacuum. For a number of these pieces, the number of cavities/g of PVG was determined by water ~
~~-
(15) Elmer, T. H. et al. J. Am. Ceram. S O ~1970,53, . 171, and references therein. (16) Janowski, V. F.; Heyer, W. 2. Chem. 1979,19, 1. An extensive listing of references on the properties of PVG. (17) Palmer, R. A.; Piper, T. S. Inorg. Chem. 1966,5,864. (18)Demas, J. N.; Crosby, G. A. J. Am. Chem. SOC.1971,93,2841.
0022-3654/83/2087-5395$01.50/00 1983 American Chemical Society
5396
'
The Journal of Physical Chemhtty, Vol. 87, No. 26, 1983
uptake to ensure internal consistancy. These measurements yielded (1.2 f 0.3) X l0l8 cavities/g which is in excellent agreement with previous measurementdg and the claimed variation, ca. 30%, in the cavity size. Cleaned pieces of PVG were impregnated by adsorption of aqueous solutions of [Ru(bpy),]Clz and then dried in a vacuum oven a t 40 "C under reduced pressure until a constant weight, within 0.2 mg of the original weight of the untreated PVG, was obtained. The number of moles of complex adsorbed was determined from the difference in absorbance of the initial solution and the solution after impregnation. Generally, three pieces were impregnated simultaneously, and, after drying, absorption spectra indicated that, within experimental error, each piece contained the same amount of complex. Although other samples were prepared in the same way, a variation in the amount of complex adsorbed did occur from group to group. Consequently, the pieces of PVG used in these experiments contain (1.0 f 0.2) X lo4 mol of Ru(bpy),2+/g of PVG. The impregnated samples were rigidly mounted with a Teflon holder in a 4 cm X 3 cm X 1cm Pyrex cell. Various pressures of the different gases were equilibrated with the samples and the adsorption isotherm of each gas was measured a t room temperature, 22 f 1 "C, by the manometric method of Ross and Olivier.20 The procedure was repeated with each sample to obtain the adsorption isotherms concurrent with the quenching data. To correct for adsorption onto surfaces other than the impregnated sample, which occurs principally with SOz, adsorption measurements were made in the presence and absence of the samples. Room temperature emission intensity measurements were made on a Perkin-Elmer Hitachi MPF-2A emission spectrophotometer. The cell holder of the spectrophotometer was adapated to rigidly hold the rectangular cell and the impregnated PVG in it at an angle of 50" relative to the 420-nm exciting light. The emission was monitored a t 620 nm from the surface of the sample at a right angle to the excitation. Luminescent lifetimes of Ru(bpy)?+(ads) were obtained by means of signal-averaging techniques. The sample was excited perpendicular to a 25 mm X 25 mm face with an Ortec Model 9352 nanosecond light pulser powered by an Ortec Model 9290 power supply. With a flow of prepurified Nz (Linde) through the spark chamber and an 8-kV charging voltage, 2-ns (fwhm)pulses of principally 337- and 358-nm light are produced a t a 20-kHz repetition rate. The 620-nm emission was monitored at 90" to the excitation pulse through a Bausch and Lomb Model 33-86-76 0.25-m grating monochromator with an RCA 31034A photomultiplier housed in a Pacific Photometric Model 3470 thermoelectrically cooled housing equipped with an AD6 amplifier/discriminator, and powered by a Tennelec Model TC941 high-voltage power supply. The discriminator output pulses are fed into a PAR Model 115 wideband preamp and then into a PAR Model 162 boxcar averager equipped with a Model 164 gated integrator. The output of the boxcar averager is displayed on a Hewlett-Packard Model 7044B X-Y recorder. These data were then plotted according to first-order kinetics, In I,, vs. time, and over the 1800-11s span of the decay measurement yielded linear plots. The rate constants for emission decay were calculated from the slopes of these plots and the lifetimes were taken as the time (19) Kennelly, T.; Gafney, H. D. J.Inorg. Nucl. Chem. 1981,43, 2988. (20) Ross, S.; Olivier, J. P. "On Physical Adsorption"; Wiley-Interscience: New York, 1964; Chapter 2, Section 1, p 31.
Wolfgang and Gafney
required for the emission intensity to decrease to l / e of the original intensity. To ensure the reliability of these decay measurements, the emission decay of degassed 5 X 10" M aqueous solutions of [Ru(bpy),]Cl, was measured periodically during the course of these experiments with PVG. For these solutions, lifetimes of 600 f 20 ns, which are in excellent agreement with the literature value: were obtained. Electronic absorption spectra were recorded on a Cary 14 or Techtron 635 spectrophotometer. The resonance Raman spectrophotometer has been previously described.21 The {potential of PVG in distilled water was determined at 25 f 0.1 "C with a Zeta Meter Inc. Model ZM-77 { meter. During adsorption of the complex, the [Cl-] in the aqueous phase was monitored with an Orion chloride ion specific electrode, calibrated with standard NaCl solutions, and its response was displayed on a Beckman SS-2 pH meter.
Results PVG develops a { potential of -26 f 2 mV, indicative of an anionic surface, when pulverized and dispersed in distilled water. To determine if Cl- is coadsorbed onto this anionic surface, we placed a 1.2548-g sample of clean, dry PVG in 25 mL of 1.03 X M Ru(bpy),Cl,. Monitoring the aqueous phase spectrally and with a chloride ion specific electrode shows that after 24 h 64% of the Ru(bpy)32+is adsorbed, whereas N20. When samples of PVG impregnated with Ru(bpy):+ are exposed to 02,SO2, and N20, lifetime and intensity quenching occur. The relative amount of each type of quenching depends on the specific gas, but all exhibit a similar dependence of intensity quenching on the equilibrium gas pressure, Pq The relative emission intensity, I o / I ,where Io is the emission intensity measured in vacuo and I is the intensity measured in the presence of the gas, is initially a function of P,, but at higher pressures becomes independent of Pq. With O2 and SO2, plots of I o / I vs. P vary by 118% for different samples of impregnated P V 6 and, in both shape and uncertainty, resemble the adsorption isotherms of these gases. With N20, however, there is a significant variation in the amount of quenching. Under the same pressure of N 2 0 , the relative Ru(bpy)82+(ads)emission intensities differ as much as 70% for different samples of impregnated PVG. In spite of the differences, the shape of the plots of Io/I vs. Pqare similar and the initial dependence on P , which is linear with O2 and SO2, is sigmoidal with The variation in quenching, which is independent of the amount of Ru( b ~ y ) ~ ~ + ( aisd similar s), to the variation in the amount of N,O adsorbed/gram of impregnated PVG, and, in both shape and uncertainty, each plot resembles the adsorption isotherm of N20. The maximum fraction quenched, designated f,, in Table I and taken as 1- (I0/I)-lwhere I o / I is independent of P,, also parallels the adsorption coefficients of these gases. A priori, *Ru(bpy);+(ads) could be quenched by either a gaseous molecule which collides with the adsorbed complex, but itself is not adsorbed, or an adsorbed molecule. Since quenching by each gas parallels their respective adsorption isotherms, the latter appears to be the most likely and the quenching data are plotted vs. the moles of gas adsorbedlgram of impregnated PVG, mads,in Figure 3. As indicated by Figure 3a, the plots of quenching by N 2 0 are reproducible functions of madswhich establishes that the variation in the amount of quenching by N 2 0 is a consequence of different amounts of N 2 0 adsorbed by the various impregnated samples of PVG. Reproducibility with respect to madsrather than the pressure of N20 is also consistent with the above suggestion that quenching of Ru(bpy)32+(ads)is limited to adsorbed NzO. The luminescent decay of *Ru(bpy),2+(ads)in vacuo, p < 5 x lo4 torr, at room temperature, 22 f 1 "C, is exponential with a rate constant of (1.35 f 0.05) X lo6 s-l and a lifetime of 740 f 20 ns. Under various pressures of these gases, the decays remain exponential, although the lifetimes, 7,are shorter. As shown in Figure 3a, lifetime , N 2 0 is within experimental error of quenching, T ~ / T by the intensity quenching. The equivalence establishes a dynamic quenching process and is interpreted within the Stern-Volmer model, Io/I = 1 + Ksvmads,where Ksvis the
Wolfgang and Gafney
D
130-40
m
90-3.0
20
40
60
, , 390
d20.
(26) Brunauer "The Adsorption of Gases and Vapors"; Princeton University Press: Princeton, NJ, 1945; Vol. 1.
Figure 3. Lifetime and intensity quenching of (1.5 f 0.5) X mol of Ru(bpy);+/g of PVG by (a) N,O, 5.86-9 sample T ~ / T(0)and I,/I (0),6.25-g sample T ~ / T(M) and I o / I (0);(b) SO2, 4.47-g sample T ~ / T (0)and I o / I (U);and (c) 02, 5.38-gsample T ~ / T(0)and I o / I (M). Ail measurements are at room temperature, 22 f 1 O C .
Stern-Volmer constant and mads is the moles of NzO adsorbed from the adsorption isotherm. Least-squares analysis of the initial linear dependence on madsyields Ksv = (3.9 f 0.6) X lo3 g of PVG/mol, and dividing the latter by T ~ 740 , f 20 ns, yields a bimolecular quenching rate constant for N20 of (5.3 f 0.1) X lo9 g of PVG/(mol 9). Figure 3, b and c, shows quite different results for SOz and Oz. At 1 atm of each gas, the difference in lifetime and intensity quenching indicates that 89 f 1% of the quenching by SOz and 70 f 1% of the quenching by Oz occur by a static process. At low pressures, both gases exhibit dynamic quenching which is dependent on madsand in turn the equilibrium pressure of each gas. Least-squares analyses of these initial dependencies yield Stern-Volmer constants of (8.5 f 1.3) X lo4 and (5.6 f 0.8) X lo5 g of PVG/mol for O2 and SO2, respectively. Dividing these values by T ~740 , f 20 ns, yield bimolecular quenching rate constant of (1.1f 0.2) x 10" g of PVG/(mol s) for Oz and (7.6 f 1.3) X 10" g of PVG/(mol s) for SOz. In part, the purpose of these experiments is to explore the possibility of initiating a reaction within these gases, particularly SO2, by a photoinduced electron transfer. However, absorption spectra recorded before and after these quenching experiments give no indication of a net formation of either Ru(bpy)?+ or [R~(bpy)~(bpy-)I+.'~
Discussion Regardless of whether a static or dynamic process is involved, the fundamental question is whether quenching of *R~(bpy),~+(ads) occurs by electron transfer or energy transfer. In the absence of detectable redox products, the
Quenching of Ru(bpy),'+ by
O,, N,O, and SO,
distinction is based on the energetic feasibility of each process. A mechanistic distinction is possible with N2O and SO2, but not with 02. Previous studies of the quenching of *Ru(bpy):+ by O2 and the resulting chemistry in fluid solution have led to different conclusions regarding mechanism. Demas and c o - ~ o r k e r spropose ~~ an energy transfer mechanism whereas Srinivasan and co-workersZ8propose an electron transfer mechanism. PVG is a hydrated surface with a surface pH of 4-5.29 If we assume that adsorption does not change the O2 excited singlet state energies30 and the reduction potential of adsorbed O2 is equivalent to that in pH 4-5 aqueous solution, -0.63 V,31,32a similar ambiguity exists with the adsorbed species which precludes a mechanistic distinction. Recent calculations indicate that the excited states of NzO lie a t energies 15.4 eV above the ground state,33although a very weak absorption in the spectrum of N20 has been reported a t 4.05 eV.34 If we assume, in view of the low extinction coefficient, that the 4.05-eV absorption positions the lowest triplet excited state, energy transfer from *R~(bpy),~+(ads) is endergonic by 22.0 eV. Similar arguments lead to the same conclusion regarding quenching by SO2, where the lowest lying triplet state is 3.20 eV above the ground Consequently, quenching of *R~(bpy),~+(ads) by N 2 0 and SOz must proceed by an electron transfer mechanism. In aqueous solution SO2is a reductant with an oxidation potential which increases with increasing basicity.36 Extrapolating this behavior to SO2adsorbed onto PVG suggests reductive quenching. *R~(bpy),~+(ads) + SOz(ads)
& [R~(bpy)~(bpy-)]+(ads) + S02+(ads) (1) kth
[R~(bpy)~(bpy-)]+ and SO2+represent the reduced complex with an e- in the bipyridine P* orbital and the oxidized quencher, respectively. Previous experiments in this laboratory show that [Ru(bpy)2(bpy-)]+ has an unusually long lifetime when adsorbed onto PVG,lg but spectra recorded before and after these quenching experiments do not indicate a net formation of the reduced complex. This suggests, as indicated in reaction 1,that a thermal backreaction complements the quenching reaction. In contrast, the use of N 2 0 as an electron scavenger and its mild oxidizing properties toward low-valent transition-metal complexes3' suggests oxidative quenching *R~(bpy),~+(ads) + N20(ads)
R~(bpy),~+(ads) +
N20-(ads) (2) Similar to the results with SOz, the absence of a spectral (27) Demas, J. N.; Diemente, D.; Harris, E. W. J. Am. Chem. SOC. 1973, 95,6864. (28) Srinivasan, V. S.; Podolski, D.; Westrick, N. J.; Neckers, D. C. J. Am. Chem. SOC.1978, 100, 6513. (29) Goonetiollake, H.V.; Gafney, H.D., unpublished observations. (30) Near-UV and visible spectra of various molecules adsorbed onto PVG resemble spectra of the molecules in aqueous or alcoholic solution; Kennelly, T. Ph.D. Thesis, City University of New York, June, 1980. (31) Eriksen, J.; Foote, C. S.; Parker, T. L. J . Am. Chem. SOC.1977, 99, 6455. (32) Divisek, J.; Kastening, B. J. EEectroanuL Chem. 1975, 65, 603. The value E = -0.634 for H+ + e- + O2 s HOz is calculated by assuming Eo = -0.33 V, pK, = 4.9, and a surface pH of 4.9. (33) Winter, N. Chem. Phys. Lett. 1975, 33, 300. (34) Sponer, H.; Bonner, L. G. J. Chem. Phys. 1940,8,33. (35) Brand, J. C. D.; Jones, V. T.; DiLauro, C. J. Mol. Spectrosc. 1973, 45, 404. (36) Cotton, F. A.; Wilkinson, G. "Advanced Inorganic Chemistry"; Wiley-Interscience: New York, 1972; 3rd ed, p 443. (37) Banks, R. G. S.; Henderson, R. J.; Pratt, J. M. J . Chem. SOC.A 1968, 2286.
The Journal of Physical Chemistry, Vol. 87,
No. 26, 1983 5399
change indicative of Ru(bpy),3+suggests a complementary thermal reverse reaction which is competitive with the decomposition of N20-.38 Although N 2 0 and SO2 quench by electron transfer processes, the formation of the donor-quencher pair proceeds by different pathways. With NzO, the equivalence of intensity and lifetime quenching suggests a collisional process whereas the dominance of intensity quenching by SO2 and O2 implies formation of a donor-quencher pair prior to excitation. The physical arrangement of donor and quencher in these experiment~-*Ru(bpy)~~+ is adsorbed onto the support while the quencher is in equilibrium between the vapor and adsorbed phases-suggests two possibilities for dynamic quenching by N20. Dynamic quenching is due to the flux of gaseous molecules onto the surface, or is due to adsorbed NzO molecules which, through long-range interactions or migration, quench the adsorbed complex. A number of observations, although individually not conclusive, collectively support dynamic quenching by N20 molecules adsorbed onto the surface rather than gaseous N 2 0 molecules impinging on the surface. The flux of N20 molecules onto the surface per second per cm2 is P/ (2amkT)1/239where P is the gas pressure and m, k , and T are, respectively, the molecular mass, Boltzmann constant, and temperature which in these experiments is 295 K. Letting P = 0.1 atm and substituting the other appropriate values yields a flux of 5.1 X mol of NzO/(s cm2). However, this represents the flux onto a principally vacant surface since the fractional surface coverage due to Ru(bpy);+(ads) is 50.01. Therefore, assuming that a Ru(bpy),2+(ads)-N20 collision is necessary for quenching, the rate of quenching by N 2 0 molecules impinging on the surface is 15.1 X mol/(s cm2). This rate of quenching by impinging NzO molecules is also equal to k b (mNzO)(.m*Ru(bpy)32+(ads)) where k b is the bimolecular quenching rate constant and mN0 and miRu(bpy)32+(ads) are the moles of gaseous N 2 0 and *hu(bpy):+(ads), respectively. Since ca. 1 cm2 of the impregnated sample is exposed to the exciting light, an upper limit of m:Ru(bpy)2+(a&) is the moles of Ru(bpy):+(ads) per cm2 of PVG or 5.71 X mol/cm2. If we take this value, mN = 6.6 X lo4 mol (P = 0.1 atm), and a rate of 5.1 X lo-%mol/(s cm2), the above equality yields a bimolecular rate constant of 1.4 X lo7 mol-l s-l. Comparing the calculated value with the experimental value, k b = 4.7 X lo9 mol-' s-l (Table I), suggests the rate at which N20 molecules impinge onto the surface is not sufficient to account for the observed quenching rate constant. Furthermore, since the number of collisions with the surface, P/(2rmkT)li2, increases linearly with increasing pressure, dynamic quenching due to the flux of gaseous NzO molecules onto the surface would exhibit a linear dependence on pressure. Yet, as shown in Figure 3a, quenching is initially a function of mads, which is proportional to P, but becomes independent of P. The dependence suggests that adsorption, rather than the flux onto the surface, is the limiting event. Also, quenching by N 2 0 is not a reproducible function of pressure, but rather a reproducible function of the moles of N 2 0 adsorbed by the different samples of impregnated PVG. These observations lead us to conclude that dynamic quenching of *Ru(bpy),2+(ads)principally involves NzO molecules which are themselves adsorbed onto PVG. Dynamic quenching by coadsorbed molecules, as noted in a recent report by Thornton and Laurence,lo implies (38) Anbard, M. Adu. Chem. Ser 1965, No. 50, 67. (39) Atkins, P. W. "Physical Chemistry";W. H. Freeman: San Francisco, 1978; p 804.
5400
The Journal of Physical Chemistry, Vol. 87, No. 26, 1083
Wolfgang and Gafney
either long-range interactions or migration of the adsorbed species on the surface. Since NzO(ads) quenches by an aE02 electron transfer mechanism, a long-range interaction implies long-range electron transfer where the driving force for the transfer arises from the photoinduced *Ru(b~y)~~+(ads)-N~O(ads) redox couple. Recent experiments have examined the relation between driving force and electron transfer distance. In the thermal back-reaction between Ru(bpy)gS+and homologues of methyl viologen, which have driving forces of 0.1 to 0.6 eV, Guarr and McLendon find that the mean separation between the reactants ranges from a contact distance to 10 8,in glycerol at 0 0C.40Experiments in this laboratory indicate that mods x lo6 in the thermal back-reaction between photogenerated , 2;o , 4 -0 6.0 R ~ ( b p y ) ~ ~ + ( aand d s ) [R~(bpy)~(bpy-)]+(ads), i.e., [Ru(bpy)z(bpy-)]+(ads),which has a driving force of 2.5 eV, the mean separation between the reacting molecules is 120 A.19 Although the one-electron reduction potential of N 2 0 is not known, N20 is a mild oxidant and the driving force for oxidative quenching by NzO, reaction 2, must be 20.84 / eV. Since R ~ ( b p y ) , ~is+a significantly stronger oxidant than NzO and [R~(bpy)~(bpy-)]+ is a stronger reductant than *Ru(bpy)?+, a conservative upper limit of the *Ru(bpy)32+(ads)-Nz0(ads)driving force is 1 2 . 5 eV. Consequently, the above correlations between driving force and electron transfer distance suggest, in our opinion, that the 0.5 driving force for oxidative quenching by NzO, 0.84 eV 5 E I 2.5 eV, is not sufficient for long-range electron transfer. Since the rates of adsorption and desorption indicate 12 24 36 that NzO is weakly adsorbed, a more likely explanation for dynamic quenching is that NzO(ads)migrates among many Figure 4. Static quenching of Ru(bpy),*+(ads) by (a) SO2 and (b) O2 sites some of which are close enough to * R ~ ( b p y ) ~ ~ + ( a d s ) (see text). for quenching to occur. The measured rate constant cavities or pore openings of PVG. Although both gases (Table I), which is the product of the surface diffusion rate exhibit some lifetime quenching a t low pressure, Figure and the quenching efficiency, indicates that the rate of 3, b and c, illustrates the dominance of static quenching. diffusion of NzO(ads) on PVG is 14.7 X lo9 mol-l s-l. Static quenching can involve formation of a precursor Surface migration of species physiadsorbed onto porous complex, but adsorption of SO2 and O2 is independent of supports has been noted by Beeck41and others.42 Howwhether the PVG contains Ru(bpy)?+ and the absorption ever, since the maximum fraction quenched by N20, Table spectrum of the adsorbed complex is unaffected by adI, is always less than unity, not all R ~ ( b p y ) ~ ~ + ( aisdacs) sorption of these gases. Rather, adsorption of these gases cessible to NzO(ads). In the quenching of silica-supported appears to be a random process in which some of the gas Mo species, Iwasawa and Ogasawarae report similar results molecules are adsorbed onto sites sufficiently close to and suggest that the region of the Stern-Volmer plot with Ru(bpy)?+(ads) to cause static quenching. Being random, zero slope corresponds to adsorbed species which for some quenching is analogous to static quenching observed in unspecified reason cannot interact with the adsorbed gas. frozen homogeneous solution and amenable to a similar Figure 1shows that Ru(bpy)?+ penetrates ca. 0.5 mm into analysis, although concentrations are now expressed as PVG. Whether this represents an actual penetration into moles adsorbedlgram of adsorbent. The general expresinterior cavities or deviations from planarity in the PVG sion for both static and dynamic quenching is44 surface such as pore openings is not clear, but either possibility suggests that not all Ru(bpy)32+(ads)is equally I o / I = (1 + Ksv(Q)) exp(k’(Q)) (3) accessible as would be the case on a uniform flat surface. I and Io are the steady-state emission intensities in the The implication is that N20(ads) migration is not presence and absence of the gas, Ksv is the Stern-Volmer throughout the entire pore volume, but is limited to the constant for dynamic quenching, k’ is the constant for outer surfaces of PVG and quenching of Ru(bpy)?+(ads) static quenching, and Q represents the moles of gas adon these outer surfaces. sorbedlgram of PVG. Substituting r o / r for the dynamic SOz and Oz are more strongly adsorbed and the extent component, 1 + Ksv(Q), yields of quenching, f,, in Table I, parallels the adsorption coefficients of the gases. The larger fraction quenched In ( ( I o / N T o / ~ = )K I Q ) (4) indicates that, when adsorbed onto PVG, the distribution and, as shown by Figure 4, the data are in accord with the of SO2 and O2 resembles that of the complex, and like equation. k’is evaluated from the initial slope and taking Ru(bpy)32+,both are capable of binding to sites within the k’ = AN where A is the surface area of the adsorbent and N is the molecules of gas adsorbedlunit surface area of (40) Guarr, T.; McLendon, C. Abstracts of the 182nd National MeetPVG, the donor-quencher interaction distance, r, is related ing of the American Chemical Society, New York, August, 1981; Inorto the slope of Figure 3 by ganic, Paper No. 20.
1
I
I
/
~~
~~~
~
(41) Beeck, 0. Adu. Catal. 1950, 2, 151. (42) Bikerman, J. J. “Physical Surfaces”;Academic Press: New York, 1970; pp 335-6. (43) Iwasawa, Y.; Ogasawara, S. J. Chem. Soc., Faraday Trans. I 1979, 75, 1465.
r = (k’/frriV)ll2 (44) Frank, J. M.; Vavilov, S. I. 2.Phys. 1931, 69, 100.
(5)
Quenching of Ru(bpy);+
by 0,, N,O, and SO2
The Journal of Physical Chemistry, Vol. 87, No. 26, 7983 5401
f is the fraction of the total surface area covered by the implies a distribution similar to that of the adsorbed adsorbed quenching gas. Since the extent of penetration complex. Since SO2 quenched Ru(bpy),2+(ads) by an of O2 and SO2into PVG is not known and therefore f is electron transfer mechanism, random adsorption within not known, the specific values of r corresponding to each the volume of PVG impregnated with the complex implies an electron transfer distance of 27 f 2 A. Although still gas are not accessible. However, limits of r can be calculated by assuming a distribution of the adsorbed gas. The quite large, the distance is surprisingly similar to that recently reported by Miller, Hartman, and Abrash.4’ The upper limit, f = 1, assumes that O2 and SO2 diffuse latter find electron transfer distances of ca. 25 A in the throughout the pore volume and is the entire surface area, 130 m2/g of PVG.15 Since O2 and SO2 quench 70 f 1% quenching of N,iV,iV’,iV’-tetramethyl-p-phenylenediamine and 89 f 1% of the adsorbed complex, respectively, the by phthalic anhydride in a 2-methyltetrahydrofuran glass lower limit of surface area to which the gas is adsorbed is a t 77 K. taken as the surface area within that volume of PVG imAdditional data in a variety of media are essential to pregnated with Ru(bpy)p. Ru(bpy)?+(ads) is uniformly correlate electron transfer distance with the driving force distributed on the outer surfaces of PVG and grinding the of the quenching reaction and to understand, in view of impregnated samples shows that the complex penetrates these long electron transfer distances, the role of the me0.5 mm into the glass. Since the latter corresponds to dium. impregnation of 28% of the total sample volume, the lower Conclusion limit o f f is taken as 0.28. Using these limits for f, the The data gathered in these experiments show that the slopes of Figure 3 yield 11 f 1A Ir I17 f 2 A and 27 MLCT state of Ru(bpy)32+,ionically bound to porous f 2 A Ir I42 f 4 A for quenching distances for O2and Vycor glass, is quenched by N20, 02, and SOz which are SO2, respectively. also adsorbed onto the glass. Although quenching is limAlthough the data gathered in these experiments do not ited to adsorbed species, quenching occurs by static and specify whether quenching by O2 involves energy or elecdynamic processes depending on the strength of adsorptron transfer, considering the difference in media, the tion. N20,which quenches by a dynamic electron transfer estimated range of r is in good agreement with previous process, is weakly adsorbed and appears to migrate on the measurements of the energy transfer quenching radius of surface of PVG whereas the more strongly adsorbed O2 and 02.In the quenching of naphthalene triplet state by O2 SO2quench by a static process. The quenching distance in a 3-methylpentane glass at 77 K, Siegel and Judekis find a quenching radius of 10.5 A.45 Bowen and M e t ~ a l f ~ ~ for O2 is similar to energy transfer distances previously found in frozen homogeneous solution. However, static report that in the quenching of anthracene fluorescence electron transfer quenching by SO2suggest the occurrence by O2 in paraffin solvents of different viscosity the O2 of long-range, 27 f 2 A, electron transfer in PVG. quenching radius is 20 A. The range of r for the quenching by O2is reasonable in view of these past results and lends Acknowledgment. Support of this research by the Resome credence to the estimate of the distribution of the search Foundation of the City University of New York, the adsorbed gas, but the range of r calculated for SO2 under Dow Chemical Company’s Technology Acquisition Prothe same approximations, particularly the upper limit, is gram, and the donors of the Petroleum Research Fund, exceptionally large. However, the rapid rate of adsorption administered by the American Chemical Society, is of SO2,1min, suggests that SO2does not diffuse into the gratefully acknowledge. H.D.G. thanks the Andrew W. interior cavities of PVG and we discount the upper limit. Mellon Foundation for a Fellowship during 1981-1982 and On the other hand, SO2 does quench ca. 90% of the adthe Corning Glass Company for the samples of porous sorbed complex. If we assume a random adsorption-no Vycor glass. evidence is available to suggest otherwise-%% quenching Registry No. Ru(bpy)gP+,15158-62-0; 0,, 7782-44-7; N20, 10024-97-2; SO,,7446-09-5. (45) Siegel, S.; Judeikis, H. S. J. Chern. Phys. 1968, 48, 1613. (46) Bowen,E. J.; Metcalf, W. S. Proc. R. SOC.London, Ser. A 1961, 206,937.
(47) Miller, J. R.; Hartman, K. W.; Abrash, S. J.Am. Chem. SOC.1982, 104, 4286.
