IR Spectrum of CH3CN−BF3 in Solid Neon: Matrix Effects on the

1402

J. Phys. Chem. B 2007, 111, 1402-1407

IR Spectrum of CH3CN-BF3 in Solid Neon: Matrix Effects on the Structure of a Lewis Acid-Base Complex Audrey A. Eigner, John A. Rohde, Christopher C. Knutson, and James A. Phillips* Department of Chemistry, UniVersity of WisconsinsEau Claire, 105 Garfield AVenue, Eau Claire, Wisconsin 54701 ReceiVed: August 30, 2006; In Final Form: NoVember 5, 2006

We have observed several IR bands of CH3CN-BF3 in neon and nitrogen matrices. For the 11B isotopomer in neon matrices, we observed the BF3 symmetric deformation band (ν7) as a doublet at 600 and 603 cm-1, the BF3 symmetric stretching band (ν6) as a doublet at 833 and 838 cm-1, the BF3 asymmetric stretching mode (ν13) at 1281 cm-1 (partially obscured), and the C-N stretching mode (ν2) as a doublet at 2352 and 2356 cm-1. The nitrogen matrix data are largely consistent with those reported recently, though we do propose a refinement of one band assignment. Comparisons of the frequencies of a few key, structurally sensitive vibrational modes either observed in various condensed-phase environments or calculated for two minimumenergy gas-phase structures indicate that inert matrix media significantly alter the structural properties of CH3CN-BF3. Specifically, the B-N dative bond compresses relative to the gas phase and other concomitant changes occur as well. Furthermore, the frequency shifts depict structural changes that occur across the various matrix hosts in a manner that largely parallels the degree of stabilization offered by these inert media.

Introduction The structural properties of the CH3CN-BF3 complex have been the subject of numerous investigations dating back over 50 years.1-18 Earlier studies included a preliminary X-ray structure,1 some solution-phase IR investigations of the C-N stretching mode,2 and a normal-mode analysis.3 Somewhat later, these studies were extended with a more refined determination of the crystal structure,4 and a thorough vibrational analysis of the solid-state IR spectrum.5 Much recent interest in CH3CNBF3 arose after the determination of the gas-phase structure via rotational spectroscopy,6 in which the complex was found to have a B-N distance of 2.01 Å, about 0.8 Å shorter than that characteristic of a van der Waals complex, yet about 0.4 Å longer than that of a typical B-N dative bond.7 Moreover, the gas-phase B-N distance was 0.38 Å longer than that observed for the solid-state complex (1.63 Å),4 indicating that the donoracceptor bond contracted dramatically upon crystallization and the pyramidal distortion BF3 subunit increased accordingly.7 These observations spurred many computational studies of the complex as well.8-15 Interestingly, none of the calculated equilibrium structures9-13 agreed with the experimental gasphase structure.6 In our initial computational work,14 we found that the B-N distance potential of the complex was remarkably flat and anharmonic, and this led us to suspect that the discrepancy between the experimental and theoretical structures was genuine, the result of large amplitude motion on the B-N stretching coordinate. We have just recently extended this computational work and further validated this conclusion.15 Overall, our studies of CH3CN-BF3 and related systems have been motivated by a basic question that stems naturally from the observation of large gas-solid structural differences: What are the effects of solVents or other bulk, condensed-phase media? In this light, our experimental work has been concerned * To whom correspondence should be addressed. E-mail: phillija@ uwec.edu. Phone: 715-836-5399. Fax: 715-836-4979.

with the structural properties of this complex in inert, noble gas matrices, which are very much non-interacting “solvents” that offer very little in terms of dielectric stabilization for very polar molecules like CH3CN-BF3. The initial observation of the C-N stretching mode in solid argon indicated that the matrix effect on CH3CN-BF3 was extreme, driving the B-N bond to the point that the structure resembled that of the crystalline complex.16 However, we subsequently found this initial work to be in error when we re-examined the IR spectrum of CH3CN-BF3 in solid argon.17 However, on the basis of our frequency assignments, together with calculated frequencies for two minimum-energy gas-phase structures (at the B3LYP/augcc-pVQZ level) and an explicit examination of how the frequencies shift with changes in the B-N distance,14 we were able to infer that the argon matrix did significantly affect the structure of the complex. Just recently, Suzuki and co-workers published an extended set of assignments for the complex in solid argon, nitrogen, and xenon18 but did not address the issue of systematic structural changes across these media. In this manuscript, we present the IR spectrum of CH3CNBF3 in solid neon and solid nitrogen, and though the latter data are quite consistent with those in ref 18, we do offer a minor refinement for one of the band assignments. Moreover, a comparison of the shifts observed among the calculated gasphase frequencies,14 measured matrix frequencies,17,18 and measured solid-state frequencies5 indicates that inert matrix environments perturb the structure of CH3CN-BF3. That is, the inert environments cause the B-N bond to compress relative to the gas phase, and for the most part, the extent of this effect varies from matrix to matrix in accord with the degree of stabilization offered by these media. Experimental Section Neon Matrix Experiments. Neon gas (99.999%) was obtained from Praxair and used without further purification. Boron trifluoride (g99.5%) was obtained from Aldrich and also

10.1021/jp0656375 CCC: $37.00 © 2007 American Chemical Society Published on Web 01/25/2007

CH3CN-BF3 in Solid Neon used without further purification. Acetonitrile was also obtained from Aldrich and was dispensed through a solvent purification column and degassed via several freeze-pump-thaw cycles prior to use. Neon matrix spectra were obtained on a newly assembled optical cryostat system. The centerpiece of this system is a Janis SHI-4-5 optical cryostat, pumped by a Pfeiffer TSU 261 turbo molecular pumping station. Spectra were recorded using a Thermo/Nicollet Nexus 670 Fourier transform infrared (FTIR) spectrometer, from which an external IR beam enters the cryostat and reflects off a 45° gold-coated mirror before entering a purged housing that contains the detector and collection optics. For mid-IR experiments (400-4000 cm-1), an extended range KBr beamsplitter and DTGS/KBr detector were used, and the cryostat was fitted with KBr windows. Typically, 1-200 scans were averaged and the resolution was usually 1 cm-1 (though some 0.5 cm-1 data spectra were collected in order to deconvolute a peak that was obstructed by an unassigned impurity band; see below). For far-IR spectra (∼100-600 cm-1), a silicon beamsplitter and DTGS/HDPE detector were used, and the cryostat was fitted with silicon windows. Typically, 500 scans were averaged and the spectral resolution was 2.0 cm-1. We constructed a concentric, dual-deposition input line, based roughly on our previous design,17 in which a small section of 1/ in. o.d. stainless steel hypodermic tubing is approximately 16 centered in a semiflexible 1/8 in. o.d. stainless steel tube and aligned such that the inner tube is about 0.25 in. shorter than the outer tube. This allows the CH3CN and BF3 to undergo some mixing just prior to entering the cryostat chamber. Thus, separate gas mixtures were prepared (i.e., CH3CN in Ne and BF3 in Ne) with concentrations ranging from 1/200 to 1/800 and were flowed through separate variable leak valves (Granville-Phillips #203) and into the mixing portion of the deposition lines. Flow rates varied from 3 to 10 mmol/h, and the separate gas samples were usually flowed at nearly equal rates. However, in some instances, we changed the relative concentrations of CH3CN and BF3 by simply increasing or decreasing the flow at either one of the leak valves, and we did examine matrix samples with up to a 4-fold excess of each guest component. Optimum signals were obtained with a relative concentration of 1/1/800 (CH3CN/BF3/Ne), with a deposition rate of about 3 mmol/h at 5 K. Most samples were subsequently annealed at temperatures of 7-9 K to facilitate complex formation and aid in the identification and assignment of CH3CN-BF3 peaks. Nitrogen Matrix Experiments. Nitrogen gas (99.9999%) was obtained from Praxair and used without further purification. Nitrogen matrix spectra were obtained using the optical cryostat apparatus, FTIR, and the concentric dual-deposition flange described previously.17 The range of sample conditions was similar to that of the neon experiments, but optimum signals were obtained with a matrix composition of 1/1/400 (CH3CN/ BF3/N2). Also, deposition temperatures were warmer, ranging from 10 to 14 K. Spectra were recorded at 0.5 or 1.0 cm-1 resolution. Results Neon Matrix Spectra. Representative neon matrix spectra of three key regions are shown in Figures 1-4, and the frequencies of the bands we observed and assigned are summarized in Table 1. In total, we assigned six features (most were split into doublets) to various vibrational bands of the 10B or 11B isotopomers of CH3CN-BF3, and their frequencies are listed in Table 1. All of these bands were observed only when both CH3CN and BF3 were present in the matrix, and in addition,

J. Phys. Chem. B, Vol. 111, No. 6, 2007 1403

Figure 1. Representative spectra of the 500-900 cm-1 region for matrices containing CH3CN in Ne 1/400 (bottom), BF3 in Ne 1/400 (middle), and CH3CN and BF3 in Ne (top) 1/1/800. Bands assigned to the 1:1 complex are noted with asterisks.

assignments were based upon characteristic isotopic shifts and relative intensity ratios indicative of a 10B/11B isotopic pair. Others were made due to the absence of isotopic shifts (consistent with predictions12,18) and the fact that they exhibited consistent intensities relative to the other bands across a broad range of sample conditions. All bands assigned to the 1:1 complex also grew slightly when the sample was annealed to ∼8 K. In addition, a few other bands were observed only at slightly higher sample concentrations and/or after annealing. These peaks did not exhibit intensity ratios that were consistent with the other bands and are presumably due to larger aggregates of CH3CN and BF3 (e.g., (CH3CN)2-BF3 or CH3CN-(BF3)2), though we did pursue the assignment of these features to the point at which we could identify a specific carrier. The basis for the band assignments and other noteworthy features that were observed in these spectra are discussed in detail below. The BF3 Symmetric Deformation and Symmetric Stretching Region (500-900 cm-1). Representative neon matrix spectra for the 500-900 cm-1 region are displayed in Figure 1. This is where the BF3 symmetric deformation (or “umbrella”, ν7) and symmetric stretching (ν6) bands typically reside. The arrangement of these spectra (and the others) is as follows: The bottom trace is the spectrum obtained with only CH3CN in Ne, the middle shows BF3 in Ne, and the upper trace shows CH3CN and BF3 in Ne. Two prominent features are observed exclusively in the top trace near 600 cm-1, and they exhibit a consistent 4:1 intensity ratio across all sample conditions. They are observed as doublets (presumably due to different matrix trapping sites) at 584/586 and 600/603 cm-1. The isotopic splitting 16-17 cm-1 is consistent with that predicted12,18 and observed5,17,18 previously for the BF3 symmetric deformation (or umbrella) band and, together with the characteristic 4:1 intensity ratio, provides very convincing evidence that these peaks correspond to umbrella bands for the two naturally occurring isotopic forms of the complex. Thus, we assign the doublets at 584/586 cm-1 and 600/603 cm-1 to the 11B and 10B isotopomers of CH3CN-BF3, respectively. Another feature that appears exclusively in the top trace in Figure 1 is observed as a doublet at 833/838 cm-1, in the region of the BF3 symmetric stretching band. Excitation of the mode is forbidden by symmetry in free BF3, but this band gains intensity with the pyramidal distortion that results from complex formation. In fact, in one instance, the symmetric stretch of gasphase CH3CN-BF3 has been predicted to be nearly equal in strength with the umbrella band,12 and it is quite strong in solidstate IR spectra of various nitrile-BF3 complexes.5,19,20 This band exhibits consistent relative intensities with those assigned

1404 J. Phys. Chem. B, Vol. 111, No. 6, 2007

Figure 2. Representative spectra of the BF3 asymmetric stretching region for matrices containing CH3CN in Ne 1/400 (bottom), BF3 in Ne 1/400 (middle), and CH3CN and BF3 in Ne 1/1/800 (top). Bands assigned to the 1:1 complex are noted with asterisks.

Figure 3. Spectral deconvolution of the peak observed at 1281 cm-1. The resolution is 0.5 cm-1. The area of the 1318 cm-1 band was 0.30 in this spectrum.

to the umbrella band, and on this basis, we assign the 833/838 cm-1 doublet to the BF3 symmetric stretching bands of the complex. The isotopic shift is predicted to be only a few cm-1,12,18 but the pair of peaks does not exhibit a 4:1 intensity ratio. Thus, the doublet structure is presumably due to distinct matrix trapping sites, and this effect must obscure any small isotopic splitting. Thus, we quote a single assignment of 833/ 838 cm-1 for the ν6 band for both isotopic forms. In many experiments, we also observed another pair of peaks in the umbrella region at 618 and 633 cm-1. While these also have a splitting and 4:1 intensity ratio characteristic of the BF3 umbrella band, they were quite weak in the absence of annealing, and/or relatively high concentrations of CH3CN and/ or BF3 (though one of these bands is partially visible in Figure 1). These features did not have intensities that were consistent relative to the other bands assigned to CH3CN-BF3, indicating clearly that they arise from a carrier other than the 1:1 complex. The concentration dependence and overall signal strength provide some evidence that these may stem from larger clusters, such as (CH3CN)2-BF3 or CH3CN-(BF3)2, though we did not pursue the specific assignments of these bands. Nonetheless, we also note that these bands are shifted in a manner that is consistent with a contraction of the B-N bond that would presumably occur upon the addition of an additional CH3CN of BF3 moiety to the complex (this has been demonstrated explicitly for related systems21). Analogous bands that were shifted by nearly the same amount (∼30 cm-1) from the main umbrella bands were also observed in Ar and Xe matrices by Suzuki and co-workers17 but were assigned to 1:1 complex bands arising from minor trapping sites in the matrices. Indeed, a site splitting of 30 cm-1 is rather large in the absence of a strong, specific guest-host interaction,22 and in spite of the

Eigner et al. difference in host gas, the concentration dependence we observe is at odds with those assignments. We also note that we essentially overlooked the secondary features (at 632 and 646 cm-1) in our previous studies of CH3CN-BF3 in argon matrices17 because they were quite weak. The main difference in our experiments was a lower deposition temperature than that in ref 18 (8-15 K vs 15-20 K), which is again consistent with the idea that these satellite features are due to larger clusters. The BF3 Asymmetric Stretching Region (1150-1450 cm-1). Figure 2 displays representative spectra for the region of the BF3 asymmetric stretching band (ν13), in which we observed two major features at 1318 and 1281 cm-1 (the latter exhibits a distinct shoulder at approximately 1276 cm-1) as well as a few minor features at 1346, 1334, 1298, 1256, and 1217 cm-1. The two major features are split by 37 cm-1, reasonably consistent with the 10B-11B isotopic splitting observed (4245 cm-1 5,17,18), or predicted (48 cm-1 12,17), previously for the ν13 band of CH3CN-BF3, but the intensity ratio (relative to the 1318 cm-1 peak) nearly always exceeded 4:1. However, we also observed a very obvious shoulder near 1276 cm-1 on the 1281 cm-1 peak, 42 cm-1 to the red of the 1318 cm-1 band, which is somewhat more consistent with the isotopic shift expected for the ν13 band. Thus, we concluded that the band we sought was at least partially obscured by another peak. We do note that similar, unassigned features have been observed in argon matrices containing BF3 and H2O.23 Thus, we collected some additional, higher resolution (0.5 cm-1) spectra and performed a spectral deconvolution of the 1281 cm-1 feature. The results of this procedure are displayed in Figure 3. In the best case, the band was fit to four Gaussian functions (four distinct maxima were observed in the 0.5 cm-1 spectrum), which included two major features centered at 1281 and 1277 cm-1, as well as two minor features (smaller by a factor of 5 or greater in terms of overall area) centered at 1278 and 1280 cm-1. Of these peaks, only the Gaussian centered at 1281 cm-1 exhibited an area that was approximately 4 times that of the 1318 cm-1 peak. We also note that we saw no splittings in the 1318 cm-1 feature. Thus, we assign the 1318 cm-1 peak to the 10B component of ν13 and tentatively assign the 11B component to a frequency of 1281 cm-1, in spite of the fact that the isotopic splitting seems to favor the feature at 1277 cm-1. However, the complications arising from these overlapping bands prevent us from making a definitive assignment, and a discrepancy of 4 cm-1 does not affect the structural interpretation of these data (see below). As for the additional, minor bands observed in this region, none have consistent relative intensities with the other assigned bands and were typically stronger at higher concentrations and/ or after annealing. Thus, though we do not see any characteristic 10B/11B isotopic pairs among these peaks, it is possible that they arise from analogous bands in larger BF3/CH3CN aggregates, but we have no solid evidence implicating specific carriers for these features. The C-N Stretching Region (2200-2400 cm-1). Representative spectra of the C-N stretching region are displayed in Figure 4. Here, we observed only a single feature that was dependent on both BF3 and CH3CN, and it occurred as a doublet with peaks at 2352 and 2356 cm-1. This band exhibited relative intensities that were reasonably consistent with the other bands assigned to CH3CN-BF3, but peak area measurements in the region were complicated due to interference with background CO2. Nonetheless, this band is very near previous observations of this band in other matrix media,17,18 and no B-isotope shift

CH3CN-BF3 in Solid Neon

J. Phys. Chem. B, Vol. 111, No. 6, 2007 1405

Figure 4. Representative spectra of the C-N stretching region for matrices containing CH3CN in Ne 1/400 (bottom), BF3 in Ne 1/400 (middle), and CH3CN and BF3 in Ne 1/1/800 (top). The band assigned to the 1:1 complex is noted with an asterisk.

TABLE 1: Observed Frequenciesa of CH3CN-BF3 neon band ν7 ν6 ν13 ν2

description

nitrogen

10B

11B

10B

600/603b

584/586b

640/633b

BF3 sym. def. BF3 sym. stretch 833/838b BF3 asym. stretch 1318 1281c C-N stretch 2352/2356b

11B

625/618b 844/836b 1277d 1226d 2368

a In units of inverted centimeters. The experimental uncertainly is (1 cm-1. b These bands were observed as doublets, presumably due to distinct trapping sites. c Overlapping bands were observed. See text for discussion. d Center peaks of triplets: 1235/1226/1222 cm-1 for 11B, 1281/1277/1270 cm-1 for 10B. See text for discussion.

is expected.12,18 Thus, we assign the doublet at 2352/2356 cm-1 to the C-N stretching mode for both the 10B and 11B isotopic forms of the complex. Nitrogen Matrix Spectra. Representative spectra of matrices containing CH3CN and BF3 in solid nitrogen are displayed in Figure 5. These data are largely consistent with those published just recently, so we provide only a brief summary here, though we do offer a minor refinement for the ν6 assignment. The procedure for the nitrogen band assignments was identical to that for the neon spectra, with the only difference being that the nitrogen matrix data are complicated by the occurrence of rather large site splittings that make most peaks appear as multiplets. It has been noted that nitrogen matrices can cause larger perturbations on vibrational spectra than noble gases,22 and it does seem reasonable that multiple sites could be more likely with nonspherical host molecules. Nonetheless, we observed the umbrella bands (ν7) as doublets at 625/618 and 640/633 cm-1 for the 11B and 10B isotopomers, respectively, completely consistent with ref 18. We note that this site splitting is only 7 cm-1. We observed the BF3 symmetric stretching band as a doublet at 844/836 cm-1. Previously, the individual components were tentatively assigned to individual isotope peaks (835.8 cm-1 for 11B and 843.5 cm-1 for 10B),18 but these peaks clearly do not exhibit a 4:1 intensity ratio in our data, so as we noted above for the neon matrix data, we attribute the splitting to distinct trapping sties, which must obscure the small predicted isotopic splitting.12,18 We observed the BF3 asymmetric stretching bands (ν13) as triplets, with the 11B band at 1235/1226/1222 cm-1 and the 10B band at 1281/1277/1270 cm-1. The assignments given in ref 18 correspond to the largest peaks of the triplets (1277.8 cm-1 for 10B and 1235.0 cm-1 for 11B), but our results are otherwise consistent. We also observed the C-N stretch as a single, yet somewhat asymmetric shaped peak at 2368 cm-1, very consistent with the frequency of 2367 cm-1 given in ref 18.

Figure 5. Representative spectra for matrices containing CH3CN in N2 1/400 (bottom), BF3 in N2 1/400 (middle), and CH3CN and BF3 in N2 1/1/400 (top). Bands assigned to the 1:1 complex are noted with an asterisk.

Far-IR Experiments. We also collected far-IR spectra (100400 cm-1) of matrices containing CH3CN and BF3 in argon, nitrogen, and neon, in an attempt to see the B-N stretching frequencies. These bands have been predicted to be weaker than the others we have observed, with an intensity equal to roughly 40% of the ν6 and ν7 bands.12 However, we saw no reproducible features in this region that could be assigned to the complex, and though the far-IR experiments are less sensitive due to lower light levels, our inability to observe the B-N stretching band provides some indication it may be weaker than predicted. Discussion The measured frequencies for three key, structurally sensitive modes observed in this work differ markedly from those observed in (or calculated for) other media,5,16-18 and moreover, the shifts are consistent with structural differences that are induced by interactions between CH3CN-BF3 and its surrounding environment. That is, these data suggest that even inert matrix environments cause the B-N dative bond to compress to some extent relative to the gas phase, while other structural

1406 J. Phys. Chem. B, Vol. 111, No. 6, 2007

Figure 6. Trends in the frequency of the ν13 band of CH3CN-11BF3 across various media.

changes occur in accord with B-N bond contraction; the N-B-F angle increases, the B-F bonds increase slightly, and the C-N bond contracts. Indeed, this situation is quite similar to that for H3N-HCl,24 a partial proton transfer system for which the degree of proton transfer varies systematically across various media. The observations summarized below represent an extension of this phenomenon to a Lewis acid-base system. Perhaps the most unambiguous view of this situation is provided by the data for the BF3 asymmetric stretching mode (ν13), for which the frequency shifts that occur across various media (for the 11B isotopomer) are depicted in Figure 6. We note again that two minimum-energy gas-phase structures were identified at the B3LYP/aug-cc-pVQZ level of theory,14 and frequencies for both are plotted in Figure 6. The global minimum is a structure with a B-N distance of 2.32 Å, and the local minimum has a B-N distance of 1.92 Å (and is only 0.11 kcal/ mol higher in energy). Previously, we explicitly examined the structural dependence of the ν13 mode (as well as the others)14 and found that it shifted monotonically to lower frequencies as the B-N distance decreased. Moreover, the dynamic range of the shift in ν13 is quite large, as the difference between the B3LYP/aug-cc-pVQZ frequency for the 2.32 Å structure and the measured solid-state frequency (R(B-N) ) 1.63 Å) is nearly 200 cm-1. For the shorter, 1.92 Å gas-phase structure, this (calculated) gas-to-solid shift is still 100 cm-1. The measured matrix frequencies fall between those of the 1.92 Å gas-phase structure and the measured crystalline frequencies, and since this range of these shifts is much larger than typical “matrix shifts” (i.e., those arising from a purely physical effect),22 the data clearly indicate that the inert media induce a significant contraction of the B-N bond. Furthermore, the ordering of the ν13 frequencies is largely consistent with the degree to which the various matrix hosts interact with the CH3CN-BF3 complex and further the development of the B-N dative bond. Just recently, we have investigated the underlying reasons for such medium-induced structural changes,15 and in short, they stem from the fact that the complex becomes more polar at short distances, causing a stronger guesthost interaction, and there would be more positive feedback for this process with a more polarizable host. This, coupled with a very flat B-N distance potential such that energetic changes that stem from intermolecular interactions can affect intramolecular properties, enables subtle environmental changes to cause observable structural effects. The observed frequency ordering of the ν13 bands is: the following Ne > Ar > Xe g N2. Neon is clearly the smallest and least polarizable among these hosts, and the neon matrix frequency is higher than those for the other hosts, and nearest to the calculated gas-phase data. The other matrix bands lie reasonably close to one another, in the region

Eigner et al.

Figure 7. Trends in the frequency of the ν2 band of CH3CN-11BF3 across various media.

Figure 8. Trends in the frequency of the ν7 band of CH3CN-11BF3 across various media.

intermediate between the neon and solid-state data. However, in spite of their mutual proximity, the ordering of the frequencies is still fairly consistent with the magnitude of the guest-host interactions. Nitrogen is slightly more polarizable than argon (the polarizability Volume is 1.66 × 10-30 m3 for argon and 1.77 × 10-30 m3 for N2).25 The dielectric constant of liquid argon is slightly larger ((N2) ) 1.45 at -203 °C, (Ar) ) 1.53 at -191 °C),26 but nitrogen does have a nonzero quadrupole moment. Thus, it is reasonable that solid nitrogen would offer a bit more stabilization and cause the B-N bond to contract more than solid argon. As for xenon, the polarizability volume is much larger (4.0 × 10-30 m3)27 and the dielectric constant of the solid exceeds that of argon,28 suggesting that solid xenon should offer more in terms of stabilization than nitrogen, which is not consistent with the observed shifts. However, the ν13 frequencies differ by only 4 cm-1 between nitrogen and xenon, well within the range of a typical matrix shift,22 and such a small difference should not be interpreted in a structural context. Overall, though, the matrix frequencies suggest that these inert environments exert significant, observable structural changes in CH3CN-BF3 and for the most part that these effects vary in accord with the dielectric properties of the hosts. Analogous plots of the C-N stretching (ν2) and BF3 umbrella (ν7) frequencies across various media are displayed in Figures 7 and 8, and for the most part, these data also depict systematic structural differences in CH3CN-BF3 across the media. The C-N stretching mode monotonically blue-shifts as the B-N bond contracts,14 but the overall dynamic range of the shifts is much less than in the case of the ν13 band. Indeed, the matrix frequencies lie between the gas and solid-state data, but the range of these data is only (13 cm-1. In spite of this, the ordering of the Ne, Ar, and N2 bands is consistent with the ν13 data (only the Xe frequency differs), and thus, the data are consistent with significant matrix effects in the structure of CH3CN-BF3. The

CH3CN-BF3 in Solid Neon situation is quite similar for the ν7 data; however, this mode does not shift monotonically as the B-N bond decreases; at longer distances, the band shifts to lower frequencies,14,29 but at shorter distances (starting at about 2.0 Å), the band shifts back toward the blue.14 Thus, the 2.32 Å gas-phase band seems out of order among the matrix data, but it is consistent with the overall structural dependence of this band, and a clear trend is evident among the other frequencies. Again, the ordering of all the bands (except for the xenon frequency) follows the same trend that would be expected on the basis of matrix-induced structural changes across these media: 1.92 Å gas phase < Ne < Ar < N2. The xenon data are, however, quite inconsistent with this trend, and in this instance, the dynamic range of the shifts is about 40 cm-1, which makes this discrepancy difficult to dismiss in terms of a purely physical condensed-phase effect. The other data do vary systematically in accord with structural changes that would take place due to the dielectric properties of these media and, for certain, illustrate that, in general, inert matrix media do induce observable structural changes in CH3CN-BF3. Interestingly perhaps, there is no analogous trend among BF3 symmetric stretching frequencies (ν6) that is as clear as those for the other modes discussed above. Part of the reason may be issues with the spectral assignments (both the argon17,18 and xenon18 assignments are reported as “tentative”), but the main factor is simply that the mode is less sensitive to structure than those discussed above. A structural dependence has been predicted for ν6, and though it is significantly less sensitive to B-N distance,14 it shifts in a manner similar to ν7. That is, ν6 red-shifts from long B-N distances to about 1.75 Å and then blue-shifts, but the dynamic range of this effect is much less than that of ν7. We note that the observed nitrogen (844/836 cm-1) and neon (833/838 cm-1) frequencies are nearly coincident but are in between those observed for the solid (894 cm-1)5 and calculated for the 1.92 Å gas-phase structure (830 cm-1).14 Thus, in spite of the lack of sensitivity, the ν6 data are still coincident with the conclusions stated above. Conclusions We have measured several key, structurally-sensitive IR bands of CH3CN-BF3 in neon and nitrogen matrices. The nitrogen data are reasonably consistent with recently reported measurements.18 For neon, we observed six new bands that were assigned to the 10B and 11B isotopomers of CH3CN-BF3. A comparison of the frequencies of the BF3 asymmetric stretching mode (ν13), the C-N stretching mode (ν2), and the BF3 symmetric deformation mode (ν7) that have been observed in various environments and calculated for two minimum-energy gas-phase structures (at the B3LYP/aug-cc-pVQZ level) indicates that inert matrix media significantly alter the structural properties of CH3CN-BF3. Furthermore, the frequencies shift across the various matrix hosts (with the possible exception of xenon) in a manner that parallels the degree to which these media would be expected to stabilize the complex and further

J. Phys. Chem. B, Vol. 111, No. 6, 2007 1407 the development of the B-N dative bond. In turn, this suggests that there are subtle, systematic structural differences between the complex in the various media, such that the B-N bond contracts to a slightly greater extent relative to the gas phase as the dielectric stabilization of the host increases. Acknowledgment. J.A.P. acknowledges financial support from the National Science Foundation (#CHE-0216058 and #CHE-0407824) and a Henry Dreyfus Teacher-Scholar Award from the Camille and Henry Dreyfus Foundation. References and Notes (1) Hoard, J. L.; Owen, T. B.; Buzzell, A.; Salmon, O. N. Acta Crystallogr. 1951, 4, 379. (2) (a) Coerver, H. J.; Curran, C. J. Am. Chem. Soc. 1958, 80, 3522. (b) Purcell, K. F.; Drago, R. S. J. Am. Chem. Soc. 1966, 88, 919. (3) Beattie, I. R.; Gilson, T. J. Chem. Soc. 1964, 2292. (4) Swanson, B.; Shiver, D. F.; Ibers, J. A. Inorg. Chem. 1969, 8, 2183. (5) Swanson, B.; Shriver, D. F. Inorg. Chem. 1970, 9, 1406. (6) Dvorak, M. A.; Ford, R. S.; Suenram, R. D.; Lovas, F. J.; Leopold, K. R. J. Am. Chem. Soc. 1992, 114, 108. (7) (a) Leopold, K. R.; Canagaratna, M.; Phillips, J. A. Acc. Chem. Res. 1997, 30, 57. (b) Leopold, K. R. In AdVances in Molecular Structure Research; Hargittai, M., Hargittai, I., Eds.; JAI Press: Greenwich, CT, 1996; Vol. 2. (8) Barber, M.; Connor, J. A.; Guest; M. F. Hillier, I. H.; Schwarz, M.; Stacey, M. J. Chem. Soc., Faraday Trans. 2 1973, 69, 551. (9) Jurgens, R.; Almlo¨f, J. Chem. Phys. Lett. 1991, 176, 263. (10) Jiao, H. J.; Schleyer, P. v. R. J. Am. Chem. Soc. 1994, 116, 7429. (11) Jonas, V.; Frenking, G.; Reetz, M. T. J. Am. Chem. Soc. 1994, 116, 8741. (12) Cho, H.-G.; Cheong, B.-S. THEOCHEM 2000, 496, 185. (13) Mo, Y.; Gao, J. J. Phys. Chem. A 2001, 105, 6530. (14) Giesen, D. J.; Phillips, J. A. J. Phys Chem. A 2003, 107, 4009. (15) Phillips, J. A.; Cramer C. J. J. Phys Chem. B 2007, 111, 1408. (16) Beattie, I. R.; Jones, P. J. Angew. Chem., Int. Ed. Engl. 1996, 35, 1527. (17) Wells, N. P.; Phillips, J. A. J. Phys. Chem. A 2002, 106, 1518. (18) Hattori, R.; Suzuki, E.; Shimizu, K. J. Mol. Sruct. 2005, 750, 123. (19) Phillips, J. A.; Halfen, J. A.; Wrass, J. P.; Knutson, C. K.; Cramer, C. J. Inorg. Chem. 2006, 45, 722. (20) Phillips, J. A.; Giesen, D. J.; Wells, N. P.; Halfen, J. A.; Knutson, C. C.; Wrass, J. P. J. Phys. Chem. A 2005, 109, 8199. (21) See, for example: Fiaco, D. L.; Leopold, K. R. J. Phys. Chem. A 2003, 107, 2808. (22) (a) Dunkin, I. R. Matrix Isolation Techniques: A Practical Approach; Oxford University Press: Oxford, U.K., 1998. (b) Matrix Isolation Spectroscopy; Barnes A. J., Orville-Thomas, W. J., Mu¨ller, A., Gaufre´s, R., Eds.; D. Reidel Publishing Co.: Dordrecht, Holland, 1981. (c) Meyer, B. Low Temperature Spectroscopy: Optical Properties of Molecules in Matrices, Mixed Crystals, and Frozen Solutions; American Elsevier Pub. Co.: New York, 1971. (23) Evans, D. G.; Yeo, G. A.; Ford, T. A. Faraday Discuss. 1988, 86, 55. (24) Andrews, L.; Wang, X.; Mielke, Z. J. Phys. Chem. A 2001, 105, 6054. (25) Atkins, P.; de Paula, J. Physical Chemistry, 7th ed.; W.H. Freeman and Co.: New York, 2002; data from Bo¨ttcher, C. J. F.; Bordewijk, P. Theory of Electric Polarization; Elsevier: Amsterdam, 1978. (26) CRC Handbook of Chemistry and Physics, 65th ed.; CRC Press: Boca Raton, FL, 1984. (27) Allen, C. W. Astrophysical Quantities; The Athlone Press: London, 1963. (28) Amey, R. L.; Cole, R. H. J. Chem. Phys. 1964, 40, 146. (29) Nxumalo, L. M.; Andrzejak, M.; Ford, T. A. J. Mol. Struct. 1999, 509, 287.