5388
M:
J. Phys. Chem. 1991, 95,5388-5390
l.u)la
1.1nA
I.&
reaction enthalpy are determined solely from the five new modes of the complex. The effect of this assumption on the derived De value can be estimated by using the MP2/6-31ffi(d,p) vibrational frequencies in the harmonic approximation. Considering just the five new modes of the complex, the computed zero-point and thermal vibrational contributions to CLH3I4 are 2.4 and 1.7 kcal/mol, respectively, compared to 2.0 and 1.6 kcal/mol, respectively, if all the modes of the complex and monomers are taken into account. If the combined difference of 0.5 kcal/mol were to be applied as a correction to the determination of the experimental value of D, the mult would be 4.0 f 0.4 kcal/mol. (The 0.4 kcal/mol overestimation of the zero-point vibrational energy results mainly from the neglect of the decrease in the H-CI stretching frequency in the complex, which lowers the zero-point energy of this mode by 0.5 kcal/mol. Although the H-CI stretching frequency is overestimated at MP2/6-31 +G(d,p), the computed red shift of the H-CI band in the complex is 168 cm-', which is in satisfactory agreement with the experimental value of 155 c~n-'.'~)The computed value of De at MP4 obtained in the present work is 5.9 kcal/mol, which makes the complex much more stable than the estimated experimental value. Some of the limitations on the experimental determination were noted in ref 1. On the theoretical side, the effect of the basis set superposition error (BSSE)20 should be considered. This error should disappear as the basis set approaches completeness. To obtain an estimate of the upper bound on the BSSE, we have carried out a counterpoise The counterpoise correction at MP4 is 0.7 kcal/mol. If the total value of this correction were applied to the complex, the theoretical binding energy would be lowered to 5.2 kcal/mol. Part of the discrepancy between the experimental and the theoretical De values can be ascribed to the neglect of anharmonicity in the low-frequency modes used to obtain De from the experimental e l 4 . The magnitude of the thermal vibrational energy contribution would increase if these modes were treated anharmonically, leading to a greater binding energy De.
1. MPZ/C3I+G(d,p)optimized structures of CH,CN-HCI (C) and of the comsponding monomen HCI and CH3CN (M).The C-H bond lengths and CC-H angles are 1.087 A and 1 0 9 . 9 O , respectively, in CH,CN and 1.086 A and 109.8O, respectively, in CH,CN--HCI.
vibration of the methyl group is computed to occur at 352 cm-'. On the basis of their temperature dependence measurements, Ballard and Henderson' calculated a MI4 value of -3.3 f 0.3 kcal/mol for the formation of the complex. (It is not clear how these measurements were corrected for the temperature dependence of pure HCl background. There appears to be a residual absorbance present in the HCl rovibrational windows chosen in ref 1 for the integrated absorption measurements of the complex.) From this value, with the assumption that the vibrational modes in the "crsdo not change in the complex, and with a classical evaluation of the heat capacities of monomers and complex, they estimated a 0, value of 5.3 f 0.4 kcal/mol. In this approximation, the thermal vibrational energy computed classically from the heat However, the capacity as C,T contributes 2.9 kcal/mol to @I4. classical formula should not be used for vibrations at room temperature except for frequencies much below 200 cm-l, and its use for this complex leads to a significant error in the thermal vibrational energy contribution from the degenerate vibration at 350 cm-'. When evaluated in the harmonic approximation as1*
hE,T = RTCx,/(exp(x,) I
- 1)
where x, = hu,/kT, the thermal vibrational energy contribution from the fne new modes of the complex demases to 2.1 kcal/mol. This correction leads to an experimental De value of 4.5 f 0.4 kcal/mol. In addition, the assumption was also made in ref 1 that the zero-point and thermal vibrational energy contributions to the
Acknowledgment. These calculations were carried out on the Cray Y-MP8/864 at the Ohio Supercomputer Center. (19) Thomas, R. K.; Thompson,H. Proc. R. Soc. London, A 1970,316, 303. (20) Boys, S. F.; Bemardi, F. Mol. Phys. 1970, 19, 553.
(18) Pitzer, K. S. Quunrum Chemistry; Prmtia-Hall: Englewood Cliffs, NJ, 1961.
Infrared Evldence for Two Isolated Silanol Species on Activated Slllcas A. J. McFarlan and B. A. Morrow* Department of Chemistry, University of Ottawa, Ottawa, Ontario, Canada Kl N 6N5 (Received: March 27, 1991; In Final Form: May 29, 1991) Infrared spectros()8py has been used to study the OH stretching vibration of isolated silanol group on an aerosil and a precipitated silica which have been activated under vacuum at 450,600, or 800 OC. When the 450 OC activated samples are cooled to -191 O C , the normally asymmetric OH peak splits into two components having a peak maximum near 3750 or 3746 cm-' for the aerosil and precipitated silica, respectively, and a shoulder near 3738 em-'. The main peak is attributed to truly isolated SiOH groups and the low-wavenumber shoulder to pairs of isolated SiOH groups on adjacent silicon atoms which are sufficiently close to slightly perturb each other. The latter are preferentially eliminated as the temperature of activation is increased.
~~~
(1)
Zhuravlev, L. T. hngmuir 1987, 3, 316. 0022-3654/91/2095-5388$02SO/O
(2) Morrow, B. A.; McFarlan, A. J. J . Non-Crysr. Solids 1990, 120,61. (8
1991 American Chemical Society
The Journal of Physical Chemistry, Vol. 95, No. 14, 1991 5389
Letters
TABLE I: Frequcaey rad Bradwidth Data activation temo. O C r .
~~~~
sample teme. .. oc
450
600 800
22 -191 22 -191 22 -191
i,\
aerosil silica precipitated silica peak fwhh, peak fwhh, max. cm-1 cm-1 max. cm-1 cm-' 3747.2 3750.8 3747.3 3751.1 3748.0 3751.7
9.4 10.5 8.2 9.1 6.2 6.9
3743.5 3746.5 3746.2 3749.3 3747.7 3751.2
a a
11.1 13.3 6.8 8.3
asec text.
higher temperatures of activation, the low-wavenumber asymmetry disappears and the peak maximum shifts to higher wavenumber. Following activation at lo00 OC, the spectra of all silica types are very similar, having a symmetric sharp peak near 3749-3748 cm-I. At no point during thermal activation in the temperature range from 450 to 1000 OC has there been IR spectroscopic evidence for a distinct low-wavenumber component.6v7v11*'2 In this Letter, we report the high-resolution infrared spectra of two silicas that have been activated in the temperature range from 450 to 800 OC, followed by cooling to liquid nitrogen temperatures so as to narrow the bandwidth. We provide the first conclusive evidence for the existence of two distinct frequencies that can be attributed to different isolated silanol species. Experimental Seetion This work was carried out using an aerosil type silica, CabO-Sil HS5, and a nonporous precipitated silica from Rh6ne-Poulenc, France. The powder ( 14 mg, 5 mg/cm2) was compacted at about lo7 Pa into thin self-supporting disks, and it was then activated under vacuum for 1 h at 450,600, or 800 OC. The aerosil and precipitated silicas will be designated by the terms A-x and P-x, respectively, where x is the temperature of activation in degree Celsius. The BET (N2) surface areas of all of the activated samples were 325 f 5 mz/g for the aerosil and 285 f 5 m2/g for the precipitated silica. The IR cell consisted of a 22-mm-i.d., 25-mm-0.d. quartz tube, 15 cm in length, which had a 50-mm-diameter flat quartz flange at one end. A 25-mm ZnSe window was sealed with epoxy resin to one end, and a 50-mm NaCl window was attached to the flange by means of Apiezon H grease. Sample disks were placed in a slotted 22-mm-0.d. quartz holder which was placed in the middle of the quartz cell through the flanged end. The vacuum connection was via a 5-mm4.d. quartz tube situated 1 cm from the large flange. For sample activation, a 8-cm-long tubular quartz furnace was positioned over the narrow end of the cell so as to symmetrically cover the sample. After activation, the furnace was removed and the cylindrical cell was fitted snugly into a polystyrene cylindrical cup (1 2-cm i.d.) through which two 25-mm holes had been drilled. After activation of a given sample, the cup-cell assembly was placed in the beam of the spectrometer. By this means, infrared spectra of samples at room temperature (nominally 22 "C), or cooled by placing liquid nitrogen in the polystyrene cup, could be recorded without movement of the sample. Sample temperatures were measured in blank experiments by replacing the 50-mm NaCl window with a glass flange through which a thermocouple passed to the sample region. Sample activation temperatures were measured while evacuating the cell. For the liquid Nz cooled samples, 1 Torr of He was added to the (3) Kiselcv, A. V.;Lygin, V. 1. Infrorcd Specfro ofSurfice Compounds; Wiley: New York, 1975. (4) Tyler, A. J.; Hambleton, F. H.; Jockey, J. A. 1. Carol. 1969, 13, 35. (5) Armistead. C. 0.;Tyler, A. J.; Hambleton, F. H.; Mitchell, S. A,; Hockey, 1. A. J. Phys. Chcm. 1969, 73,3947. (6) Morrow, B. A.; Gay, 1. D.J. Phys. Chcm. 1988, 92, 5569. (7) Hoffmann. P.;Knazinger, E.S u r . Scf. 1987, 188. 181. (8) Burneeu, A.; Barr& 0.; Gallas, J. P.;Lavalley, J. C. Lungmuir 1990, 6, 1364. (9) Morrow, B. A.; McFarlan, A. J. Langmuir, in press. (IO) Gallas, J. P.;Lavalley, J. C.; Burneau, A,; Barr&, 0. Longmuir, in
press.
(11)
Morrow, B. A.; Cody, 1. A. 1. Phys. Chcm. 1973, 77, 1469. P.R.; Russel, B. G. J. Phys. Chcm. 1975, 79, 1276.
(1 2) Ryason,
r
3800
P-800
3i50
1
cm-'
Si00
3650
Figure 1. Infrared spectra (SiOH stretching region) of aerosil and pm cipitated silica after vacuum activation at the indicated temperature. Spectra were recorded with the samples at room temperature (22 "C).
cell to facilitate thermal conduction, and the measured temperature at the sample position in this pressure of He was 82 K or -191
"C. Infrared spectra were recorded with a Bomem DA3-02 FTIR spectrometer at a resolution of 0.5 cm-I.
Results and Discussion Infrared spectra of the isolated silanol peak on aerosil and precipitated silicas recorded at room temperature after vacuum activation at 450, 600, or 800 *C are shown in Figure 1. For all temperatures of activation, the peak in asymmetric to low wavenumber, but as the temperature of activation increases, the degree of asymmetry decreases as the peak narrows and the peak maximum shifts to higher wavenumber. Table I lists the wavenumber of the peak maximum and its full width at half-height (fwhh). (The fwhh for P-450 is not listed because of the very pronounced tail to low wavenumber; see further discussion below.) Similar spectra have been reported several times by ~ t h e r s , ~ ~ * ~ ' * ~ ~ and for either silica for any temperature of activation, there is no clear evidence that the asymmetry can be associated with a distinct second component to lower wavenumber of the peak maximum. It has previously been established that the peak maximum shifts to higher wavenumber as the temperature of the silica decreases.13J2When the same samples used to record the spectra shown in Figure 1 were cooled to near liquid nitrogen temperatures (82 K or -191 "C), the spectra shown in Figure 2 were observed. The peak maxima and the fwhh are given in Table I. The most striking observation in the spectra of P-450 and A-450 is the replacement of the continuous asymmetry to low wavenumber by a distinct shoulder near 3738 cm-', as indicated by the arrow in Figure 2. This shoulder is not evident for the 800 OC activated silicas, but it can still be seen as an inflection in the spectrum of P-600. Further, the fwhh data in Table I show that the apparent width of the peak increases as a result of cooling to -191 OC, even for the 800 OC activated samples. This is contrary to normal expectation and suggests that the isolated silanol peak is really a composite of at least two peaks whose separation increases as the temperature decreases, and this applies to both silicas. The fwhh data are particularly important because the shoulder is much less
5390 The Journal of Physical Chemistry, Vol. 95, No. 14, 1991
3800
37'50
cm-'
37bO
3650
Figwe 2. Infrared spectra of the same samples used in Figure 1 but after
cooling each disk to -191
OC.
P-450
I
Letters frequency between each component at the higher temperature. There is a precedent for the latter observation in the spectrum of silica in the near-infrared ngion.l3J4 A broad poorly resolved doublet near 4540 cm-I has been observed in the room-temperature spectrum of silica which has been attributed to a combination mode involving the isolated SiOH stretching frequency with low-frequency SiOH bending and/or a Si-OH stretching mode. When the sample is cooled to liquid nitrogen temperature, this combination band separates into two distinct peaks which have a greater separation than the room-temperature doublet. Theoretical calculations have indicated that the symmetric and antisymmetric OH stretching modes of a geminal pair of silanols would probably only be separated by a few wave number^.^^ Therefore, it is tempting to suggest that the low-wavenumber band can be assigned to one of the two expected o(0H) modes of geminal Si(OH)2 groups. However, solid-state %i NMR spectroscopy has shown that both single and geminal silanols persist even up to 800 or lo00 "C vacuum activation and that the relative number of geminal to single silanols on silicas of whatever origin does not vary significantly as a function of the temperature of activation.'620 Our results have shown that the 3738-cm-' shoulder is more intense on the precipitated silica than on the aerosil and that this is preferentially eliminated relative to the high-wavenumber band as the temperature of activation is increased. Therefore, it is unlikely that the low-wavenumberband can be exclusively attributed to geminal silanols. It is more reasonable to assign the low-wavenumber shoulder to a pair of vicinal isolated silanols (geminal or single) which are sufficiently far apart so as not to significantly interact via strong hydrogen bonding. It has been previously shown that when micromole doses of water are added to highly dehydroxylated silicas, a new isolated silanol band centered near 3742 cm-I is created. This has been attributed by Morrow et a1.2'*22 and by Hoffmann and Knozinger' to pairs of noninteracting vicinal silanols which are created as a result of the rehydration of strained siloxane bridges. This low-wavenumber peak can only be clearly seen at 20 OC on highly dehydroxylated silicas when the intensity of the 3748-cm-' band is low. These are the first isolated silanol types re-formed on rehydration, and because of their proximity, they are also probably the first types to be eliminated as the temperature of activation is increased above 450 OC. Finally, the activated precipitated silica has a greater silanol density than the aerosil? and it is not unreasonable that the number of vicinal weakly interacting pairs of silanols would be greater on the precipitated silica. Acknowledgment. We are grateful to N.S.E.R.C of Canada for financial support. (13) Tsygenenko, A. A. Russ.J . Phys. Chem. (Engl. Tronsl.) 1982,56, 1428. (14) Kustov, L. M.; Borovkov, V. Yu.; Kazanskii, V. B. Russ. J . Phys. Chem. ( E n d . Tronsl.) 1985. 59. 1314. (15)'Sa&r, J.; Sch;&ler, K.P.Z . Phys. Chem. (Lefpzig)1985,266,379. (16) Maciel, G. E.;Sindorf, D. W . J . Am. Chem. Soc. 1980, 102, 7606. (17) Sindorf, D. W.;Maciel, G. E.J . Am. Chem. Soc. 1983,105, 1487. (18) Fyfe, C. A.; Gobbi, G. C.; Kennedy, G. J. J. Phys. Chem. 1985,89, 211. (19) Sindorf, D. W.; Maciel, G. E. J . Phys. Chem. 1982,86, 5208. (20) Legrand, A. P.; Hommel, H.; Tuel, A.; Vidal, A,; Balard, H.;Papirer, E.; Levitz, P.; Czernichowski, M.; Erre, R.;Van Damme, H.; Gallas. J. P.; Hemidy, J. F.; Lavalley, J. C.; Barres, 0.;Burneau, A,; Grillet, Y . Ado. Colloid Inierfoce Sci. 1990, 33, 91. (21) Morrow, B. A.; Cody, I. A. J . Phys. Chem. 1975, 79, 761. (22) Morrow,8.A.; Cody, I. A.; Lee, L.S. M.J . Phys. Chem. 1976,80, 2761.
