J . Phys. Chem. 1994, 98, 1006-1009
1006
Proton Diffusion in the Pores of Silicate Sol-Gel Glasses John McKiernan, Eric Simoni,? Bruce Dunn,’ and Jeffrey I. Zink’ Department of Chemistry and Biochemistry and the Department of Materials Science and Engineering, University of California, Los Angeles, Los Angeles, California 90024 Received: October 25, 1993”
The rate of the recombination of protons with pyranine, or 8-hydroxy 1,3,6-trisulfonated pyrene, in the pores of sol-gel matrices prepared from tetraethoxysilane is measured by using a p H jump method. Irradiation of pyranine with a 10-ns laser pulse generates a proton concentration of about lP M within the time of the pulse. The rate of the ground-state back-protonation reaction is measured in sols, gels, and xerogels by monitoring the time dependence of the absorbance of the deprotonated form of pyranine after the excitation pulse. The rate constants in the gel and aged gel differ only slightly from those in aqueous solutions. In air-dried xerogels the rate constant is 4.6 X 1Olo M-l s-l. The activation energy of the back-protonation reaction is calculated from the temperature dependence of the rate constant. It is 8.4 kcal/mol, about 2.4 times larger than that in aqueous solution.
SCHEME 1
Introduction
The sol-gel process is a method by which inorganic glasses (such as silicates or aluminosilicates) can be prepared from solution at room temperature by hydrolysis and polycondensation of alkoxide precursors.’ Because it is a low-temperature process, organic and biological molecules with poor thermal stability can be encapsulated intact in the inorganic glass.2 Many recent studies have shown that these molecules retain their optical properties.24 Molecules encapsulated in the gel-glass can react with other molecules in solution or in the gas phase when the material is exposed to those environments because the matrix is very porous. Experimental evidence of this response was obtained in aluminosilicate and silicate ~erogels.~J’Investigation of sol-gel materials as the active element of optically based chemical sensors is currently an active area of research. Even enzymes and other proteins may retain their characteristic activities in sol-gel g l a s ~ . ~ -Thus, l ~ the prospects for using the encapsulants for a new generation of highly specific sensors are very promising. In considering the doped sol-gel glasses for sensor applications, a fundamental understanding of diffusion processes in these materials is essential. Although the prior studies show that the porosity allows transport of small molecules into and out of the glass, very little is known about the rates of diffusion. One specific study used electrochemical methods to determine the diffusion coefficient of ferrocene in silica and zirconia sols and gels.13 Another approach has been to use refractive index changes during leaching to determine diffusion coefficients in TiOz-SiOz and GeO2-SiOz systems.1615 In this paper we report the measurement of the rate of recombination of protons with pyranine, or 8-hydroxy 1,3,6trisulfonated pyrene, in the pores of sol-gel matrices prepared from tetraethoxysilane. The photochemistryI”l7 and pHindicator propertiesl8-I9 of this molecule has been studied. The pH jump technique, which can generate a proton concentration of up to 10-4 M within 100 ps by irradiation of the pyranine with a laser pulse, is used.zs22 The rate of the ground-state backprotonation reaction of the pyranine is measured by monitoring the time dependence of the absorbance of the deprotonated form of the pyranine after the excitation pulse. The results are interpreted in terms of a second-order diffusion-controlled reaction.22 The activation energy is calculated from the experimentally determined temperature dependence of the rate conCurrent address: Laboratoire de Radiochimie, Universitt Paris 11, Bat. 100 IPN, 91406 Orsay, France. Abstract published in Aduance ACS Abstracts, December 15, 1993. ?
0022-365419412098- 1006$04.50/0
pKa= 0.4
&OH*-
PyOH
PyO*+H+
-
PyO
pKa= 1.8
+ H+
stant. The rate constants in aqueous solution and in the xerogel are compared. The pH Jump Method
The pyranine molecule is employed as a proton generator. The acid dissociation constant, pK,, of the protonated molecule (PyOH) in the ground state is about 7.8. This value can be compared to the acid dissociation constant of the first excited singlet state, pK,*, which is about 0.4.20-22 This difference is used to create a sudden increase in the concentration of protons, [H+]. The proton pulse is achieved by irradiation with a laser pulse, which converts the proton emitter from a weak acid to a strong acid. Within 100 ps the molecule dissociates, releasing the protons into the medium. Steady-state and kinetic measurements of the protonation/ deprotonation have been reported for pyranine in aqueous solution.21-22The proton pulse can lower the pH by 3-5 pH units within 100 ps. This concentration change in transitory; the molecular relaxes to its more basic ground state within 6 ns, and the time constant for recombination of protons with the groundstate deprotonated pyranine is about 1 ps. The dynamics of this recombination can be studied to gain information about the surrounding medium. The overall process is summarized in Scheme 1. In our experiments, the back-protonation reaction is studied by following the variation of the concentration of the deprotonated form with time by measuring its transient absorption after the excitation pulse. The deprotonated form has a broad, strong absorption band centered about 460 nm. Experimental Section Sample Reparation. The sol-gel matrix is obtained by hydrolysis and polycondensation of tetraethoxysilane (TEOS). In the preparation, the initial sol was prepared by sonicating a 0 1994 American Chemical Society
Proton Diffusion in Silicate Sol-Gel Glasses mixture of 30 mL of TEOS, 10 mL of water, and 20 mL of 10 mM phosphate buffer (pH 6.5). The gels were doped with the pyranine (1.5 X 10-4 M concentration in the sol) during the synthesis of the sol. The sol was poured into polystyrene cuvettes immediately following the synthesis. The gelation time was less than 1 h. After gelation the materials were aged for 1 month. The xerogel was obtained by slow evaporation of solvent from the aged gel at ambient conditions. When the drying and shrinking of the gel were complete, the concentration of pyranine in the xerogel was approximately 1.2 X M. Note that in this paper the term xerogel is used to describe the samples which were airdried and exhibited no further weight loss over time. Optical Measurements. The transient absorption of the pyranine anion versus time was measured by using a pumpprobe technique. The pump pulse stimulates the pH jump, and the probe follows the changein theabsorbance of thedeprotonated form of the pyranine. The pump pulse was produced by an excimer laser (Lambda Physik) operating at 20 Hz (A = 308 nm) with a pulse duration of 10 ns. The pulse energy at the sample was about 7 mJ. Theabsorbancechange of thesamplewas monitored at 465 and 476 nm by using a continuous wave (cw) Ar+ or Kr+ laser focused through the sample with a 20-cm-focal length lens. The probe beam, having an approximate spot size of 100 pm, passed through on the side of the sample that was irradiated by the pump pulse. The pump (focused into the sample with a cylindrical lens) and probe beams crossed in the sample at a 90° angle. The probe beam was passed through a 1/4-m monochromator (set at the probe laser wavelength to eliminate fluorescence and light from the pump beam) and detected with a 1P28 photomultiplier tube terminated at 50 n. The photomultiplier tube voltage was typically 650 V. The detection system was checked to make sure that the tube response was linear with respect to the incident intensity. The response time of the instrument was about 10 ns. The transient absorption signal was recorded by using a Tektronix RTD 7 10transient digitizer with time resolution of 5 ns/channel. The data for each run were averaged over 256 pulses. For each experiment two additional measurements were made. One measurement was made in which the probe beam was blocked (corresponding to 0%transmittance). The second measurement, corresponding to 100% transmittance, was made with the pump beam blocked. These were used to convert the recorded data from voltages to percent transmittance. The transmittance data were then converted to absorbance by using the formula A = -log % T. When the data were fitted, only the portion of the transient absorption corresponding to the first four lifetimes of the signal were used. Beyond that point the signal-to-noise ratio was too poor for the data to be meaningful. Results A series of measurements at room temperature were performed
on samples at different stages of the sol-gel process. The initial sol, the aged gel, and the air-dried xerogel were examined. The effect of temperature on the recombination rates was studied on air-dried xerogels at temperatures between -6 and +20 OC. Control experiments were carried out by making the same measurements on sols, aged gels, xerogels, and heat-dried xerogels that did not contain pyranine. No transient absorption signals were detected on the blank samples. Xerogels. The most detailed studies were carried out on airdried xerogels. A typical room temperature transient absorbance, monitored at 476 nm, of a 1 month old xerogeldoped with pyranine is shown at the bottom of Figure 1. A sharp increase of the absorbance occurs on the time scale of the laser pulse. It is caused by the increase of the concentration of the deprotonated form of the pyranine. As the back-protonation occurs after the laser pulse, the concentration decreases to its equilibrium value on the
The Journal of Physical Chemistry, Vol. 98, No. 3, 1994 1007
Gel
4
nI
Xerogel
0.5
1
1.0
1.5
2.0
2.5
3.0
Time (psec) Figure 1. Time dependence of the absorbance of the deprotonated form of pyranine measured at 476 nm in the sol, gel, and xerogel. The traces
are offset for clarity.
24.51
22.5
-
22.0-
-45 . 12 340
350
360
370
380
UT (1 0.' x K-') Figure 2. Plot of the natural logarithm of the rate constant of the
protonation back-reaction versus the reciprocal of the temperature.
time scale of microseconds. The laser pulse would appear to be a 6 function on the time scale shown in the figure. The time dependence of the absorbance of xerogels at controlled temperatures between -6 and +20 O C shows similar behavior. There is no discernible change in the rise time of the absorbance, but the time required for decay decreases as the temperature increases. The experimental results, expressed in terms of rate constants, are used in Figure 2 and are interpreted in the discussion section. To quantify the water content in xerogels, thermogravimetric analyses were performed. The materia1 exhibits a gradual loss in weight beginning a t 69 OC that is completed a t 120 O C . The total weight change is 25 f 5%. It is important to note that over the temperature range in which the measurements were performed (-6 to +20 "C), there is no change in weight. Sol and Aged Gel. In order to compare the protonation rate constants for the reaction of the molecule encapsulated in gels during their sol-gel-xerogel transformation, transient absorption
1008 The Journal of Physical Chemistry, Vol. 98, No. 3, 1994
TABLE 1 rate constant‘ (M-I s-l)
rate constant‘
medium medium 8.5 X loLo gel (buffer) solution xerogel (buffer) 7.9 X 1Olo sol (buffer) The experimental uncertainty is about &lo%.
(M-1 s-l)
7.6 X loLo 4.6 X 1O1O
1.4
+ -
1.0
+ 0.8 b?
0.6
0.41 0.2
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
Time (psec) Figure 3. Relaxation kinetics analysis of the transient absorption of a gel containing pyranine. Error bars show theuncertaintyin the calculated value. Data shown are from the ‘gel” curve in Figure 1.
measurements were also made on the initial sol and on aged gels. Plots of the absorbance as a function of time for a typical sol and gel are shown in Figure 1. Again, the decrease in the absorbance occurs on a microsecond time scale. The quantitative results, expressed in terms of rate constants, are given in Table 1. Discussion
1. Analysis of the Rate constant of the Back-Protonation Reaction. The goal of this study is to determine the rate of recombination of protons with pyranine in sol-gel materials during processing. In the case where the concentration of protons generated by the proton jump is approximately equal to the initial concentration of protons in the solution, the change of the concentration with time can be fit by the following expression:
where xo is the initial change in the concentration of the deprotonated form of pyranine, xt is the change in the concentration of the deprotonated form of pyranine, x, is the change at timet, [H+]ois the initial concentration of protons in the medium, and k is the rate constant of the back-protonation reaction.22 This analysis is used to interpret the experimental decay profiles shown in Figure 1. A typical plot of the log function versus time is shown in Figure 3. The pH in the sol and gel was taken to be 4.9; this value gave the best fit to the experimental data. Other measurements in our laboratory using pH indicators showed that the pH in the material was approximately 5.0. The values of the rate constant obtained from this kinetic analysis are listed in Table 1. Note that the rate constants change by approximately a factor of 2 as the system evolves from the sol to the xerogel. The rate constants in the sol and in aqueous solution are equal within the experimental uncertainty of the measurements. 2. Calculation of the Activation Energy from the Temperature Dependence of the Rate Constant. The activation energy of the back-reaction is calculated from the temperature dependence of the rate constant. A plot of the natural log of the rate constant versus the reciprocal temperature is linear (Figure 2). The activation energy, the slope of the plot in Figure 2, is 8.4kcal/ mol, about 2.4 times larger than that measured in aqueous solution (3.5 kcal/ mol) .22
McKiernan et al. The measured activation energy, assuming a diffusion-controlled mechanism, includes contributions from three sources.22 Two of these, the contributions from the dielectric constant and the ionic strength, are small. The activation energy due to the mobility of the proton is the principal contribution to the activation energy, accounting for almost 80% of the total. The activation energy therefore reflects the process of proton transport through the medium. The activation energy for proton motion is much closer to that observed for water than for solid materials. Activation energies for ionic diffusion through solids are much higher, and this is inconsistent with the experimental value of the activation energy for the diffusion of protons obtained here. It is likely, therefore, that the mobile species are solvated on the interior surface of a pore and are not moving through the solid silica network. The magnitude of the water content in the xerogel as determined from the TGA measurements is consistent with this mechanism. 3. Comparison with an Aqueous Solution. The small difference between the values of k in the xerogel and those in water suggests that the proton is in a similar environment in the xerogel and in solution. The reprotonation, following the acidic dissociation which releases the proton, occurs at different rates in the different media. It is possible to understand this difference by considering the confinement of molecules in small pores. Recent experimental and theoretical work has shown that molecules confined in small pores exhibit changes in their thermodynamic, structural, and dynamic Sol-gel glasses have been used in the study of these effects because of the ability to control and characterize pore size, while the transparency enables optical techniques to be used to determine molecular motion. It has been shown that surface interactions, such as hydrogen bonding between the molecule and silanol groups, are particularly important in restricting molecular motion.25 The air-dried xerogels contain solvent molecules only on the walls of the pores, which are mainly OH groups with hydrogenbonded watere2* This situation would correspond to “surface molecules” treated in the molecular dynamics study by Brodka and Zerda.Z7 Their calculations for cyclohexane indicate that the diffusion coefficient for the surface molecules in small pores is less than that of pure solvent. The magnitude of this difference (30-50% depending upon temperature) is comparable to the difference in k values for the proton motion that were measured between the xerogel and the aqueous solution. The lower diffusion coefficient results from the different environment of the surface molecule as compared to a “center molecule” (in the center of the cavity). The simulations also showed that the diffusion coefficient for the center molecules is comparable to that of the bulk solvent. Our results are consistent with this as well; the rate constant for the aged gel, Le. prior to shrinkage and pore collapse, differs only slightly from that of the aqueous solution. Finally, it is interesting to note that the higher activation energy measured for the xerogel may be representative of the greater energy required to achieve proton migration along the surface of a small pore as compared to the center of the cavity. The similarity of the rate constants for diffusion-controlled proton recombination in the different media suggests that the value of the diffusion coefficient for the proton in the gel and air-dried xerogel is well within an order of magnitude of the diffusion coefficient of protons in water, which is 9.3 X 10-5 cm2 S-I.~O In contrast, Majors et al. measured the diffusion of D20 into a water-saturated porous cylinder of zirconia and found a diffusion coefficient of 6.5 X 10-6 cm2 s-l.Z9 This, however, was a macroscopic measurement of diffusivity in a porous ceramic where morphological considerations such as the size and structure of the pores are important. The measured diffusion coefficient thus includes the tortuosity of the material, or the relative difficulty of moving between pores. In the proton recombination experiment
Proton Diffusion in Silicate Sol-Gel Glasses this long-range factor would not affect the measured rate as the protons are not likely to be moving a new pores. Summary
The pH jump method was used to study the rates of the groundstate back-protonation reaction of pyranine trapped in a silica glass. The temperature dependence of the rate constant of the back-protonation reaction was measured, and the activation energy of this reaction was calculated. By comparison with the previously reported kinetic measurements done with the same molecule in aqueous solution, the recombination of protons in the xerogel seems to be affected by the walls of the pores. The rate of proton recombination with pyranine in an air-dried xerogel is 4.6X 1Olo M-1 s-1. Acknowledgment. This work was made possible by a grant from the National Science Foundation (Grant DMR-90-03080). E.S. thanks the UniversitC Paris 11 for partial financial support. References and Notes (1) Brinker, C. J.; Scherer, G. Sol-Gel Science: The Physics and Chemistry of Sol-Gel Processing, Academic Press: San Diego, 1989. (2) For reviews, see: (a) Dunn, B. D.; Zink, J. I. J . Mater. Chem. 1991, 1,903. (b) Avnir, D.;Braun,S.;Ottolenghi, M.'SupramolecularArchitecture in 1 and 2 Dimensions" ACS Symposium Series No. 499; American Chemical Society: Washington, D.C., 1992. (3) ,(a) Avnir, D.; Levy, D.; Reisfeld, R. J . Phys. Chem. 1984,88, 5956. (b) Avnir, D.; Kaufman, V. R.; Reisfeld, R. J. Non-Const. Solids 1985, 74, 395. (c) Levy, D.; Avnir, D. J . Phys. Chem. 1988, 92, 4737. (d) Levy, D.; Einhorn, S.;Avnir, D. J. Non-Cryst. Solids 1989, 113, 137. (4) (a) McKiernan, J.; Pouxviel, J.-C.; Dunn, B.; Zink, J. I. J . Phys. Chem. 1989, 93,2129. (b) McKiernan, J.; Yamanaka, S.;Dum, B.; Zink, J. I.J.Phys. Chem. 1990,94,5652. (c) Preston,D.;Pouxviel, J.-C.;Novinson, T.; Kaska, W. C.; Dunn, B.; Zink, J. I. J . Phys. Chem. 1990,94,4167. (d)
The Journal of Physical Chemistry, Vol. 98, No. 3, 1994 1009 McKiernan, J.; Yamanaka, S.;Knobbe, E.; Pouxviel, J.-C.; Parvaneh, S.; Dunn, B.; Zink, J. I. J . Inorg. Polym. 1991, 1 , 87. (5) (a) Tani, T.; Namikawa, H.; Arai, K.; Makishima, A. J. Appl. Phys. 1985,58,3559. (b) Matsui, K.; Nakazawa, T.; Morisaki, H. J. Phys. Chem. 1991, 95, 976. (6) Genet, M.; Brandel, V.; Lahalle, M.-P.; Simoni, E. C. R.Acad. Sci. Paris 1990, 311, 1321. (7) Pouxviel, J. C.; Dunn, B.; Zink, J. I. J . Phys. Chem. 1989, 93, 213. (8) Zuaman, R.; Rottman, C.; Ottolenghi, M.; Avnir, D. J. Non-Cryst. Solids 1990, 122, 107. (9) Braun, S.;Rappoport, S.;Zusman, R.; Avnir, D.; Ottolenghi, M. Mater. Lett. 1990, 10, 1. (10) Braun, S.; Rappoport, S.;Shtelzer, S.;Zusman, R.; Druckman, S.;
Avnir, D.; Ottolenghi, M. In Biotechnology, Bridging Research and Applications; Kamely, D., et. al., Eds.; Kluwer: Boston, 1991; p 205. (11) Ellerby, L.; Nishida, C.; Nishida, F.; Yamanaka, S.;Dunn, B.; Valentine, J. S.;Zink, J. I. Science 1992, 255, 1113. (12) Yamanaka, S.; Nishida, F.; Ellerby, L.; Nishida, C.; Dum, B.; Valentine, J. S.;Zink, J. I. Chem. Mater. 1992, 4, 495. (13) Audebert, P.; Griesmar, P.;Sanchez, C. J. Mater. Chem. 1991, I , 699. (14) Shingyouchi, K.; Makishima, A.; Tutumi, M.; Takerouchi S.J. NonCryst. Solids 1988, 100, 383. (15) Shingyouchi, K.; Makishima, A. J . Am. Ceram. Soc. 1988,71,682. (16) Kondo, K.; Fendler, J. H. Biochim. Biophys. Acta 1978,509,289. (17) Kondo, H.; Miwa, I.; Sunamoto, J. J. Phys. Chem. 1982,86,4826. (18) Wolfbeis, 0. S.;Furlinger, E.; Kroneis, H.; Marsoner, H. 2.Anal. Chem. 1983, 314, 119. (19) Clement, N. R.; Gould, M. Biochemistry 1981, 209, 1534. (20) Fbrster. Th.: Vblker. S. Chem. Phvs. Lett. 1975. 34. 1. (21) Gutman, M:; Huppert, D.; Pines, E. J . Am. Chem. soc. 1981, 103, 3710. (22) Pines, E.; Huppert, D. J. Phys. Chem. 1983,87,4471. (23) Huppert, D.; Kolodney, E. Chem. Phys. 1981, 63, 401. (24) Warnock, J.; Awschalom, D. D.; Shafer, M. W. Phys. Reu. B 1986, 34, 475. (25) Nikiel, L.; Hopkins, B.; Zerda, T. W. J . Phys. Chem. 1990,94,7458. (26) Brodka, A.; Zerda, T. W. J . Phys. Chem. 1992, 97, 5669. (27) Brodka, A.; Zerda, T. W. J. Phys. Chem. 1992, 97, 5676. (28) Hench, L. L.; West, J. Chem. Rev. 1990, 90, 33. (29) Majors, P.; Smith, D.; Davis, P. Chem. Eng. Sci. 1991, 46, 3037.
