18848
J. Phys. Chem. 1996, 100, 18848-18851
Effect of Deposition Temperature on the Mobility of Matrix-Isolated Species: HBr in Xenon D. Howard Fairbrother,† Dwayne LaBrake,‡ and Eric Weitz* Department of Chemistry, Northwestern UniVersity, 2145 Sheridan Road, EVanston, Illinois 60208 ReceiVed: July 16, 1996X
FTIR spectroscopy has been used to probe the mobility of matrix-isolated species by monitoring the formation of dimers and higher-order multimers in a 1:1600 HBr doped xenon matrix as a function of both the initial deposition temperature and subsequent annealing treatment. Upon subsequent annealing, matrices initially deposited at lower temperatures produce larger concentrations of multimeric HBr species compared to matrices formed by initial deposition at higher temperatures. These results are interpreted in the context of the effect of the microscopic morphology of a matrix on the mobility of matrix-isolated species. Time-resolved studies of the formation of multimers allow an approximate diffusion coefficient to be determined for HBr at 50 K. These results are consistent with HBr mobility in these rare gas matrices being defect mediated, which is enhanced in more amorphous matrices formed at lower deposition temperatures.
Introduction The mobility of atoms, radicals, and molecules in matrixisolated systems has been of interest from the earliest matrix isolation studies.1 Formation of dimers and even higher-order multimers often accompanies deposition of a molecular species and typically increases upon annealing of the matrix. Both of these observations implied a degree of mobility of the species of interest. The formation of molecular adducts, such as HCO and HNO, by reaction of H atoms with CO and NO in a matrix environment has been interpreted on the basis that at least some species were mobile under typical matrix conditions.2 The issue of mobility of small atoms in rare gas matrices has recently been revisited for O, H, and F atoms in a variety of cryogenic rare gas solids.3-6 One of the surprises in the study of H atom thermal mobilities was the effect of deposition conditions. Lower deposition temperatures produced matrices in which H atoms were more mobile than those deposited at somewhat higher temperatures.4 This behavior was attributed to the production of a more “crystalline” environment in matrices deposited at higher temperatures and a more “amorphous” environment resulting from lower temperature deposition. Deposition conditions and their effect on the morphology of the resulting matrix have also been shown to have an effect on the quantum yields for photolysis.7 These latter results have been interpreted as arising from the effect of crystallinity on the probability of cage exit versus in-cage recombination. In this paper we report an investigation of the effect of deposition temperature on the formation of dimers and higherorder multimers in the HBr:Xe system subsequent to annealing of these matrices. The propensity of HBr molecules to form dimers and multimers can be used as a probe of HBr mobility at the annealing temperature. We find that deposition temperature can have a significant effect on HBr mobility, and the observed change in mobility as a function of deposition temperature is attributed to differences in the morphology of the matrices. More amorphous matrices produced at lower * To whom correspondence should be addressed. † Present address: Department of Chemistry, University of California, Berkeley, CA 94720. ‡ Present address: 3M Co., M/S A003-1N-05, 11705 Research Blvd., Austin, TX 78759. X Abstract published in AdVance ACS Abstracts, November 1, 1996.
S0022-3654(96)02132-6 CCC: $12.00
deposition temperatures lead to a greater propensity for defectmediated mobility. Experimental Section Samples of HBr in Xe were prepared in preconditioned glass bulbs and allowed to mix overnight. For deposition, a 10 mL sample deposition chamber was connected in tandem to the sample mixing bulb and a 1/16 in. diameter sample deposition tube. The sample deposition chamber was isolated from the vacuum shroud of a closed cycle helium refrigerator (CTI Cryotronics, Model 22) by a solenoid valve and filled with approximately 480 Torr of sample. The sample chamber contents were then released through the 1/16 in. diameter deposition tube onto a BaF2 window held at the desired temperature by the closed cycle helium refrigerator and temperature controller (Lake Shore Cryogenics, Model 320). This process was repeated eight times, yielding a matrix ∼6 µm thick. Where the sample temperature is different from the deposition temperature, it was achieved by raising or lowering the temperature in 2 K increments, allowing 5 min for thermal equilibrium to be attained at each temperature point. The same procedure was followed for annealing. IR spectra were recorded at 0.5 cm-1 resolution using a Mattson "Polaris” FTIR with a HgCdTe semiconductor detector. Subsequent to deposition and/ or annealing, samples were cooled to 10 K at which point spectra were recorded. HBr, specified as 99.8% pure, was obtained from Matheson and subjected to several freeze-pump-thaw cycles. Xenon, specified as 99.99% pure, was obtained from Cryogenic Rare Gases and was used without further purification. Results The infrared (IR) spectra of HBr in Ar and Kr have been well-studied.8 These spectra are typically characterized by P(1) and R(0) lines with a weaker “Q branch”. For the Xe matrices in this study, absorptions assignable as the P(1) and R(0) lines were observed at 2508.7 and 2530.9 cm-1 (see Figure 1). As expected, these absorptions are shifted to somewhat lower frequencies as compared to those observed in lighter rare gas matrices. The separation of 22 cm-1 between the R(0) and P(1) absorption is in good agreement with a separation of 19 cm-1 in argon.8 The weak Q branch absorption at 2520.1 cm-1 is more prominent in the spectrum of the matrix deposited at 28 © 1996 American Chemical Society
Mobility of Matrix-Isolated Species
J. Phys. Chem., Vol. 100, No. 48, 1996 18849
Figure 1. FTIR spectra obtained from a 1:1600 HBr:Xe matrix initially deposited at 28 K and annealed in 5 K increments, for 30 min at each step, from 30 to 60 K. All spectra were recorded after cooling the matrix to 10 K.
Figure 2. FTIR spectra obtained from a 1:1600 HBr:Xe matrix initially deposited at 45 K and annealed in 5 K increments, for 30 min at each step, from 45 to 60 K. All spectra were recorded after cooling the matrix to 10 K.
TABLE 1: HBr Frequencies (cm-1) R(0) Q P(1)
2530.9 2520.1 2508.7
dimer/higher-order multimers dimer/higher-order multimers dimer/higher-order multimers
2492.3 2474.5 2455.7
K than at higher temperatures (see Figure 3). A number of absorptions due to HBr dimers and various higher-order multimers and their isomers have been identified in lighter rare gases.8 Virtually all of these features appear at lower frequencies than the parent absorptions. The major features observed in this study are compiled in Table 1. Though we have not performed a detailed study in an attempt to provide an exact correspondence between the dimer and multimer absorptions identified in lighter rare gas matrices and those observed in xenon, it is clear that there is a qualitative correspondence between features that are observed both before and after annealing in xenon and lighter rare gas matrices. The principal nonparent absorptions, due to dimer and higher-order multimer (multimeric species) absorptions, are at 2492.3, 2474.5, and 2455.7 cm-1. Inspection of Figures 1 and 2 demonstrates that the 2492.3 cm-1 absorption does not have the same intensity dependence as a function of annealing temperature as the other peaks. Additionally, the intensities of the 2474.5 and 2455.7 cm-1 peaks do not appear correlated. This indicates that it is likely that each of the three principal nonparent absorptions can be associated with a different absorbing species. This observation coupled with the behavior of the peaks on annealing and
Figure 3. FTIR spectra obtained for 1:1600 HBr:Xe matrices initially deposited at (a) 28, (b) 45, and (c) 60 K. All spectra were recorded after cooling the matrix to 10 K.
assignments for analogous peaks in the lighter rare gases suggests that the 2492 cm-1 peak is due to a dimer, the 2472 cm-1 peak is an absorption of a trimer, and the 2455 cm-1 peak is a different trimer isomer or a tetramer. The largest multimer with a distinct absorption feature that was assigned in the studies of the lighter rare gases was a tetramer.8 It should be noted that all IR spectra were recorded for 1:1600 HBr:Xe matrices at a temperature of 10 K after cooling, in 2 K increments, following the relevant annealing treatment. Figure 1 displays a series of spectra for a matrix deposited at 28 K before subsequent annealing, in 5 K increments, from 30 to 60 K. A comparison of the relevant IR intensities in Figure 3 indicates that monomeric HBr, identified by the P(1) and R(0) lines at 2508.7 and 2530.9 cm-1, is the primary chemical species present upon initial matrix deposition at 28 K. However, Figure 1 also indicates that subsequent annealing treatments produce a monotonic increase in the ratio of multimeric:monomeric HBr molecules. Experiments were also carried out to probe the time dependence of the multimer to monomer ratio for a matrix deposited and held at 28 K. These experiments encompassed a period of several hours following deposition, corresponding to the same total time required to increase the matrix temperature in Figure 1 from 28 to 60 K. These experiments reveal that although there are some small changes in the IR spectra over this time period, these timedependent effects are much smaller than the effects of increasing annealing temperature shown in Figure 1. Figure 2 shows the effect of increasing annealing temperature upon a matrix initially deposited at 45 K and annealed for 30 min at the indicated temperatures. Upon initial deposition the IR spectra indicate that, analogous to the situation at 28 K, the matrix is composed primarily of monomeric HBr. However, subsequent annealing of this matrix produces a much smaller increase in the multimer:monomer ratio compared to the analogous matrix deposited at 28 K (Figure 1). Thus, a comparison of Figures 1 and 2 reveals that, upon annealing both matrices to between 50 and 60 K, the matrix initially deposited at 28 K contains a far greater proportion of multimeric HBr species. Figure 3 shows the effect of initial deposition temperature, between 28 and 60 K, on matrix composition. Higher initial deposition temperature favor the formation of more multimeric HBr species. However, this variation is smaller than that associated with the effect of increasing annealing temperature shown in Figure 1. Figure 4 displays spectra for a matrix deposited at 60 K which is then held (annealed) at the same temperature for 3 h. This
18850 J. Phys. Chem., Vol. 100, No. 48, 1996
Figure 4. FTIR spectra obtained from a 1:1600 HBr:Xe matrix (a) initially deposited at 60 K and (b) then subsequently held (annealed) at 60 K for 3 h. All spectra were recorded after cooling the matrix to 10 K.
Figure 5. Time-resolved FTIR spectra for a 1:1600 HBr:Xe matrix initially deposited at 28 K and then annealed at 50 K for (a) 10 min, (b) 1 h, and (c) 17 h. All spectra were recorded after cooling the matrix to 10 K.
figure demonstrates that, in contrast to matrix deposition at either 28 K (Figure 1) or 45 K (Figure 2), followed by subsequent annealing at 60 K, matrices initially deposited and held at 60 K produce virtually no change in the multimer:monomer ratio. Figure 5 shows the time-resolved FTIR spectra for a matrix deposited at 28 K and annealed at 50 K for three time intervals. Upon initial deposition the spectra is dominated by absorption bands characteristic of monomeric HBr species. However, the relative concentration of multimeric HBr species increases steadily over the time scale of the experiment. After 17 h has elapsed since initial matrix deposition the strongest infrared bands are associated with multimeric HBr species. Discussion A comparison of Figures 1 and 5 indicates that, for matrices initially deposited at 28 K, long-term annealing at 50 K (Figure 5) results in the evolution of the spectrum into one that is virutally identical to that obtained as a result of annealing at 60 K for a much shorter time period. Figures 1, 2, and 4 also show that although subsequent annealing to 60 K leads to an increase in the multimer:monomer ratio, these changes become markedly less pronounced as the initial deposition temperature is increased from 28 to 45 K and then to 60 K. For all of these matrices growth of the dimer and multimer peaks is accompanied by a concomitant loss of intensity associated with the parent peak.
Fairbrother et al. These results show that the growth of the multimeric HBr peaks is clearly dependent on the thermal history of the system. In a perfect crystal, bulk diffusion should be independent of the thermal history of the system. The data imply that the mobility of HBr in a xenon matrix can be strongly influenced by deposition conditions. A similar observation was reported for the mobility of H atoms in xenon.4 For the H atom system this observation was attributed to a change in the morphology of the matrix as a function of deposition temperature. This morphology change, in turn, effected the mobility of a dopant. This could occur since a Xe matrix is composed of an ensemble of face-centered cubic microcrystalline grains surrounded by disordered regions called grain boundaries. In contrast, the large increase in the percentage of multimeric species that occurs as a result of 50 K annealing of the matrix deposited at 28 K versus the matrix deposited at 45 K provides evidence that the percentage of grain boundaries relative to microcrystals is larger in matrices deposited at 28 K than at 45 K. A trend of increasing crystallinity with increasing deposition temperature is also consistent with data for the matrix in Figure 3, where the lack of growth of significant dimer and multimer peaks on prolonged annealing at 60 K demonstrates that there is little HBr mobility associated with the monomer remaining in this matrix at 60 K. The conclusion that grain boundaries facilitate mobility is also consistent with the experimental observation of a greater degree of visible light scattering from matrices deposited at 28 K as compared to 45 K. The increased light scattering is likely due to an increase in sample density fluctuations on the length scale of visible light. This trend of increasing matrix crystallinity may also be reflected in a monotonic decrease in the full width at half-maximum of the HBr R(0) line at 2530.9 cm-1 as a function of increasing initial deposition temperature. This result is consistent with the idea that more amorphous matrices, produced using lower initial deposition temperatures, contain a wider variety of local environments for the monomeric HBr species. Thus, we conclude that lower deposition temperatures lead to more amorphous matrices which facilitate HBr mobility. These data imply that the potential energy surface of the grain boundary regions is such that the barrier height to thermally induced motion is effectively reduced compared to more crystalline regions. Figure 5 demonstrates that it is possible to directly monitor, in real time, diffusion at 50 K for a matrix deposited at 28 K. Though a precise determination of the diffusion coefficient for a diffusion-limited reaction rate constant cannot be obtained until the details of the associative loss mechanism are known, dimer and multimer formation in matrices have been treated as diffusion-limited processes.9 The best data for the time evolution of a relevant absorption was obtained for the peak at 2472 cm-1. In this model the diffusion-limited rate of formation of the 2472 cm-1 cm-1 peak can be represented by an equation of the following form:
kobs ) 4πrHBrD
(1)
where rHBr ) 0.7 Å is the HBr bond radius, and D (cm2 s-1) is the diffusion coefficient. kobs can be obtained by solving the relevant kinetic equation
1 1 ) + kobst CHBr(t) CHBr(0)
(2)
where CHBr(t) is the concentration of mobile HBr molecules at time t and CHBr(0) is the initial concentration of mobile HBr molecules. Since mass balance is most closely realized if it is assumed that the absorption intensity of an HBr molecule is
Mobility of Matrix-Isolated Species independent of whether it is a monomer or incorporated into a multimer, it is possible to obtain a concentration for the initial concentration of mobile HBr molecules from the difference between the initial and long-term values of the HBr monomer peak intensity. From this analysis, the best fits to the data at 40 K in Figure 5 yield a diffusion coefficient of D45 K ) 5 × 10-15 cm2 s-1 for a matrix deposited at 28 K. This result is qualitatively consistent with determinations of diffusion coefficients for small atoms in rare gas matrices. The diffusion coefficient for H atoms in xenon matrices at 40 K, where diffusion is believed to be defect mediated, varies with deposition conditions, and presumably matrix morphology, and is on the order of 10-13 cm2 s-1. In the case of oxygen atom mobility in xenon matrices the time dependence of the oxygen atom concentration was described by two bimolecular processes, which at 40 K were measured to be 7.3 × 10-15 and 2.0 × 10-17 cm2 s-1. The process with the larger diffusion coefficient was implicated as possibly being due to diffusion in more amorphous regions of the matrix while the smaller diffusion coefficient would then correspond to diffusion in more crystalline regions.5 Theoretical simulations by Raff et al.10 indicate that these diffusion processes are likely to be associated with defect-mediated diffusion. In contrast, the self-diffusion rate for Xe at 50 K can be approximated by extrapolating the results of a recent study at higher temperature to 50 K, yielding a diffusion coefficient of roughly 10-38 cm2 s-1.11 While the extrapolation is extreme, the result indicates that the Xe selfdiffusion coefficient should be many orders of magnitude smaller than that observed for HBr molecules. Thus, if Xe diffusion were necessary to create vacancies to facilitate HBr mobility, this would result in no observable diffusion of HBr on the experimental time scale of annealing. This suggests that HBr thermal diffusion at 50 K follows either a defect-mediated interstitial or substitutional mechanism. This behavior is a clear indication that bulk diffusion through crystalline Xe cannot be the dominant mechanism responsible for growth of multimeric species. However, the microscopic details of this type of mobility and the details of the interplay between matrix structure and mobility are just starting to be probed.10,12 Superficially, there is a potential alternative interpretation for the observed behavior. In this picture there is a similar distribution of trap sites produced for higher or lower temperature deposition, but higher temperature deposition allows the shallower trap sites to be depleted. There would then be a smaller population of HBr molecules in these shallower sites and therefore fewer mobile HBr molecules for a system deposited at a higher temperature when it was raised to the annealing temperature. Though depletion of low-energy trap sites could contribute to the effect observed in Figure 1, and is probably at least partially responsible for the enhancement in dimer and multimer peaks relative to lower temperature deposition (Figure 3), it is important to note that there is much more relative depletion of parent when the system is deposited at lower temperature and annealed than is observed subsequent to higher temperature deposition (cf. Figures 1 and 4). Similarly, there is a much greater relative proportion of nonparent absorption in the annealed matrices deposited at lower temperatures than in matrices deposited at higher temperature. Though there is likely to be a similar distribution of trap sites for the different deposition temperatures, these results argue against higher temperature deposition producing a depletion of the lower energy trap sites. In principle, molecules formerly in lower energy trap sites could be so mobile that they form “high polymers” or even clusters of “solid” HBr. Absorptions for these species have been identified in prior studies,8 and there is
J. Phys. Chem., Vol. 100, No. 48, 1996 18851 no evidence for their formation at the different deposition temperatures under experimental conditions. In addition, there is no significant depletion of parent for higher temperature deposition relative to lower temperature deposition as would be required by this explanation. Thus, all evidence points to the difference in deposition temperature leading to a fundamental alteration in the structure of the matrix and correspondingly in the potential energy surface that can be used to describe dynamical processes in these matrices. Conclusions The effect of deposition conditions on the mobility of HBr molecules is consistent with deposition conditions having a significant effect on matrix morphology. More amorphous matrices, produced by lower temperature deposition, facilitate mobility. The deposition temperature appears to have a greater influence on matrix morphology, and consequently on HBr mobility, than annealing. It is concluded that the magnitude estimated for the thermal diffusion coefficient for HBr (5 × 10-15 cm2 s-1) in Xe, at 50 K, for a matrix deposited at 28 K, is consistent with mobility being defect driven. These results also imply that better isolated, more thermally stable monomer species are likely to be obtained by higher temperature as opposed to lower temperature deposition since higher temperature deposition favors the production of more “crystalline” lattices. This study, taken in conjunction with the prior reports of the effect of deposition conditions on the mobility of H atoms in xenon,4 indicates control of deposition conditions may provide a means to affect mobilities in cryogenic solids. Since many reactions taking place in cryogenic solids are expected to be mobility limited, this in turn potentially provides a novel means to control rates for a variety of processes, including reactions, in this medium. Acknowledgment. We acknowledge support of this work by the United States Air Force, Air Force Systems Command, Phillips Laboratory, Edwards Air Force Base CA 93523-5000, under Contract F29601-91-C-0016. We also thank Dr. Todd Ryan and Dr. Paul House for useful advice and experimental assistance. References and Notes (1) Perutz, R. N. Chem. ReV. 1985, 85, 77, 97. Jacox, M. E. In Chemistry and Physics of Matrix Isolated Species; Andrews, L. E., Moskovits, M., Eds.; Elsevier: Amsterdam, 1989. (2) Ewing, G. E.; Thompson, W. E.; Pimentel, G. C. J. Chem. Phys. 1960, 32, 927. Robinson, G. W.; McCarthy, M., Jr. J. Chem. Phys. 1958, 28, 350. (3) Ryan, E. T.; Weitz, E. J. Chem. Phys. 1993, 99, 1004. Ryan, E. T.; Weitz, E. J. Chem. Phys. 1993, 99, 8628. (4) LaBrake, D.; Weitz, E. Chem. Phys. Lett. 1993, 211, 430. (5) Krueger, H.; Weitz, E. J. Chem. Phys. 1992, 96, 2846. (6) Kunttu, H.; Feld, J.; Alimi, R.; Becker, A.; Apkarian, V. A. J. Chem. Phys. 1990, 92, 4856. Feld, J.; Kunttu, H.; Apkarian, V. A. J. Chem. Phys. 1990, 93, 1009. Alimi, R.; Gerber, R. B.; Apkarian, V. A. J. Chem. Phys. 1990, 92, 3551. Danilychev, A. V.; Apkarian, V. A. J. Chem. Phys. 1993, 99, 8617. Danilychev, A. V.; Apkarian, V. A. J. Chem. Phys. 1994, 100, 5556. (7) Schriever, R.; Chergui, M.; Schwentner, N. J. Phys. Chem. 1995, 1991, 95, 6124. (8) Barnes, A. J.; Hallam, H. E.; Scrimshaw, G. F. Trans. Faraday Soc. 1969, 65, 3150, 3159, 3172. (9) Plonka, A. Time-Dependant ReactiVity of Species in Condensed Media; Springer-Verlag:Berlin, 1986. (10) Ford, M. B.; Foxworthy, A. D.; Mains, G. J.; Raff, L. M. J. Phys. Chem. 1993, 97, 12134. (11) Cowgill, D. F.; Norberg, R. E. Phys. ReV. 1976, B13, 2773. (12) Raff, L. M. J. Chem. Phys. 1992, 97, 7459.
JP962132D