J . Phys. Chem. 1992,96, 7422-7424
7422
Quantum Chemical Molecular Models of Oxides. 3. The Mechanism of Water Interaction with the Terminal OH Group of Silica A. G. Pelmenschikov,+G. Morosi, and A. Gamba* Dipartimento di Chimica Fisica ed Elettrocliimica and Centro del CNR per lo Studio delle Relazioni fra Struttura e Reattivita’ Chimica, Universita’ di Milano. Via Golgi 19, I-20133 Milano, Italy (Received: March 3, 1992)
New forms for the adsorption of water on the terminal OH group of silica are suggested, with the water interacting with the OISiOH fragment both as a proton acceptor and as a proton donor simultaneously. The energy preference for these forms with respect to the 03SiOH-OHz one, proposed in previous quantum chemical studies, is supported by ab initio 6-31G*, 6-31G**//6-31G*, and MP2/6-31GS*//6-31G* calculations. The new forms are confirmed also by comparison of their calculated vibrational frequencies with the observed ones for the adsorption complex.
Introduction An important property of water and methanol in intermolecular interactions is their bifunctionality, which leads the molecules to form hydrogen-bonded polymers in solution, with the OH group acting both as a proton donor and as a proton acceptor simultaneously.I4 According to our recent nonempirical quantum chemical study,5 this property plays a significant role also in methanol adsorption of the silica surface: the comparison of experimental IR data on the adsorption with the predicted C-H stretching frequencies for hypothetical surface structures, using a scaling procedure to correct the ab initio 3-21G frequencies, showed that at low coverage the
of OH-OHZ fragment rotation around the S i 4 axis on the energy of IC. The model geometries, optimized under the C, symmetry constraint, with the additional geometry restriction on models IC and I d allowing the formation of only one hydrogen bond in these cases, are presented in Figure 1. The remaining parameters of the Si(OH)4 molecule are fud and made qual to those calculated for the molecule with the 6-31G* basis set?’ As in ref 27, OSiO angles are assumed tetrahedral. Owing to the very small 6-31G* energy difference between ICand IC’(0.3 kJ/mol; see Table I), the effect of OH.-OHZ rotation around the S i 4 axis in form IC is not taken into account in the following. The calculations with the 6-31G’ basis set show the instability of form lc: its reoptimization, allowing H.4-H and H-0-X -0, ’0-H.. angles to vary, turns ICinto lb without an energy barrier. SimSI 0-CH, ilarly, structure la is found to be the only stable one, when the optimization procedure starts from a structure with the values -0’ >O.. H’ of the geometry parameters of lc’, but with the HzO molecule lying in the plane of C, symmetry. Deeper examination of structure is the most likely among the possible physically adsorbed structures la, lb, and IC (Table I) shows that the basis set exmethanol species. The energy preference for this adsorption form, tension by gorbital addition to the hydrogen atoms (6-3 1G**/ in comparison with the other ones, was supported in ref 5 by /631G* level of approach) does not practically change the relative adsorption energy calculations with a larger basis set (6-3 1G*). stability of the structures, while the additional correction for In light of these results and of the similarityof acid-base properties electron correlation effects at the MP2 level causes a significant of methanol and water, one might expect the surface structure increase (7 kJ/mol) of stability of both l a and lb with respect SiOH--OHz, generally assumed for the adsorption of water on to IC. So, the extension of the HjSiOH molecular model of the the terminal OH group of silica,b1zto be less stable than structures terminal OH group by inclusion of the oxygen atoms bound to with extra hydrogen bonds linking one or two water hydrogen the SiOH fragment is evidence of the instability of structure IC. atoms with the neighboring oxygens of the SiOH fragment. In Obviously, the greater stability of the new hypothetical structures view of the great interest, during the last two decades, in the determination of the structure of this complex by experi~nental’~-~~ la and lb is the result of the extra hydrogen bonding. However, the calculated hydrogen bond lengths (Figure 1) show that the and theoretical6-l2methods, the above-mentioned hypothesis is bond in structure lc is stronger than any bond in la and lb. This verified in the present quantum chemical study. effect might be induced by the forced distortion of the 0-H-0 Models and Method fragments in la and l b in comparison with the optimal structure (cf. Figure l), when the 0-H-0 angle is near to 180°.8J19z3In Equilibrium geometries and harmonic vibrational frequencies structure lb the additional hydrogen bonds may be classified as are calculated by the ab initio SCF MO method with the 6-3 lG* very weak ones. basis set using the Gaussian 88 packagez3with the gradient opThe observed shiftd7J8of the water combination frequencies timization procedure. To check the effects of basis set extension v, i-v k n d and v, + vas and of the surface OH group stretching and of electron correlation, the adsorption energies are recalculated frequency v under the influence of the adsorption are presented at the 6-31G**//6-31G* and MP2/6-31G**//6-31G* levels. in Table 11, together with the corresponding values estimated for Considering the structural features of the adsorptional mechanism structures la, lb, and ICusing the calculated harmonic vibrational molecule is adopted as a model under investigation, the frequencies (Table 111). Unlike la and lb, structure ICdoes not of the terminal OH group, unlike the H3SiOH molecule presatisfy the considerable mixing” of the SiOH and H 2 0vibrations dominantly used for this purpose in recent nonempirical studies in the complex (see the assignments in Table 111). Moreover, (see, for example, refs 7, 11, and 24-26). considering that the harmonic approach with the 6-3 lG* basis Results and Discussion set usually overestimates the frequencia by 10-15%,23the observed value of the red shift of water v, + v,, frequency cannot be exThe suggested adsorption forms la and l b are compared with plained by the formation of structure IC. With regard to the other IC (see Figure 1). Model IC’is also calculated to check the effect two structures la and lb, the former is in better agreement with the experimental data. So,in addition to the energy calculations, Permanent address: Institute of Catalysis, The Siberian Branch of the the vibrational frequency calculations give further support to Russian Academy of Science, 630090 Novosibirsk, Prosp. Lavrentieva 5, Russia. structures la and lb with respect to IC.
.
0 1992 American Chemical Society
The Journal of Physical Chemistry, Vol. 96, No. 18, 1992 7423
Quantum Chemical Molecular Models of Oxides TABLE i: T a l bergks (Mrccr)rad Ad8orpth E.cbcr (U/mol)
6-31G* molecule la lb le IC'
Si(OH),' H20
total energ -666.908 86 -666.909 05 -666.904 72 -666.904 34 -590.883 92 -76.010 74
a& energ 37.3 37.8 26.4 26.1
6-31G**//6-31G* total energ ads energ -666.950 29 36.5 -666.95043 36.9 -666.946 27 25.9
MP2/6-3 1G**//6-3 1G* total energ ads energ -667.964 46 47.1 -667.964 99 48.5 -667.958 09 30.4
-590.91282 -76.023 57
-591.727 14 -76.21936
'Geometry parameters equal to those optimized with the 6-31G* basis set and S4 symmetry restriction:' but the energetically lipsed-staggercd conformation of the two interacting SiOH groups is assumed (cf. structure la in Figure 1). TABLE Ilk Calcuhted Harmonic Vibrational Freqaencies (cm-I)
TABLE 11: Cdeplrtal rpd Obsmed Sbifta of Vihtiolul Freqmmcks (em-')
mode H~OU,
calcd characteristie HzOA(U,+~) H,OA(U, + vu) SiOH Au
la
lb
le
-21 -30 -111
-14 -50 -111
-17 -14 -146
O
M
-19 -2Sb -100,b-8JC
'The combination frequencies are sums of the corresponding normal vibrational frequencies (Table 111). *Reference 17. cReferencc 18.
H20 vu
H20uw
SiOH u
SiOH uw
X X
H 10
ec-
IC'
1. Molecular models of adsorption (C, point symmetry). In the and le', ~ 0 - H - 0= 180° and LH-0-X optimization of structures IC = 123O, corresponding to the optimal values for this form of adsorp-
ti~n,'."*~~ are assumed. Owing to the assumptions made, the present study can be considered a first approach to the investigation of structures la and lb. Their further discrimination would require removal of the symmetry constraints, to enlarge the model by including the nearest atoms to the 03SiOH fragment and to consider the OSiO angle deviation from the tetrahedral value on the real surface. As the energy difference between models la and l b is quite small, the inhomogeneity of local geometries on the surface might allow the existence of both forms.
CdIRlion In contrast to previous Quantum chemical studies, nonempirical calculations at the 6-31G*, 6-31G**//6-31G*, and MP2/631G**//6-31G* levels on extended molecular models of the terminal OH group of silica show instability of the 03SiOH-OH2 surface structure. According to the new hypothesis, the stabilivltion of water on the 0 8 i O H group occurs through simultaneous interactions of the water, namely, as a base to the hydrogen of
molec la lb le
H20 la lb le
H2O la lb le
H20
la lb le SiOH la lb le SiOH
frea 4178 4154 4178 4188 4050 4054 4066 4070 1815 1846 1819 1826 4025 4025 3990 4136 1111 1101 1091 1043
assignment'
'Numbers in parentheses indicate the contributions to each normal mode from the normalized sums of the squares of the eigenvectors of each atom for each mode.' the SiOH group and as an acid to the nearest oxygens of the SiOH group, through one or two of its hydrogens. In this view the bifunctional property of the water is assesfed,as in solutions. Our hypothesis is strengthened as the new structures give a better agreement with IR data on the adsorption complex. In our recent quantum chemical studySa similar mechanism was found as the most preferable one in the first stages of the physical adsorption of methanol on the silica surface. It should be noted that the possibility for the extra hydrogen bonding of water by one hydrogen atom, on the terminal and geminal OH groups of silica, was discussed by Sauer and Schroder,lz on the basis of ab initio SCF calculations with the 4-31G basis set and of the nonempirical transferable potentials method.28 The 03SiOH-OH2complex was found neither unstable nor less stable than the one with the extra hydrogen bonding. Maybe, this is the reason why this possibility has not been considered in successive theoretical investigations, at higher levels of quantum chemical theory, of the adsorption of water as well as of methanol on the bridging O H As it was concluded by Pelmenschikov et al.30J1and recently confiied by Teunissen et al.32on the basis of quantum chemical studies, the coordination in zeolites of ammonium ion with the corresponding lattice counterion A104- occurs through the formation of two or three strong hydrogen bonds. In contrast to the interaction of methanol with the terminal O H group, the absorption of ammonia on the bridging OH group leads to formation of the protonated molecule. As it was shown in ref 30, the formation of more than one hydrogen bond between NH4+and AlOi ions is the necessary condition to compensate the energy consumption for the proton transfer from the surface to the molecule, as the solvation effect is the condition for ammonium ion formation in solution. Acknowledgment. A.G.P. gratefully acknowledges a fellowship for the year 1992 from Unione Industriali di Como.
J. Phys. Chem. 1992, 96, 7424-7430
7424 Registry No. H 2 0 , 7732-18-5; S O 2 , 7631-86-9.
References and Notes (1) Walrafen. G. E.; Luck, W. A. P. In Structure oJwater ond aqueous solutions; Luck, W. A. P., Ed.; Verlag Chemie: Weinham, 1974; p 222. (2) Hadzi, D.; Bratoz, S. In The Hydrogen Bond, Schuster, P., Zundel, G., Sandorfy, C., Eds.; North Holland Publication: Amsterdam, 1976; pp 565,
613. (3) Shivaglal, M. C.; Singh, S.Int. J . Quantum Chem. 1989, 36, 105. (4) Brakaspathy, R.; Singh, S.Chem. Phys. Lett, 1986, 131, 394. (5) Pelmenschikov, A. G.; Morosi, G.; Gamba, A. J . Phys. Chem. 1992, 96, 2241. (6) Zhidomirov, G. M.; Kazansky, V. B. Adu. Catal. 1986, 34, 131. (7) Ugliengo, P.; Saunders, V. R.; Garrone, E. J . Phys. Chem. 1990,94, 2260.
( 8 ) Hobza, P.; Sauer, J.; Morgeneyer, C.; Hurych, J.; Zahradnik, R. J. Phys. Chem. 1981,85,4061. (9) Sauer, J.; Zahradnik, R. Int. J. Quantum Chem. 1984, 26, 793. (10) Sauer, J.; Schrcder, K.-P. Chem. Phys. Lett. 1984, 107, 530. (11) Chakoumakos, B. C.; Gibbs, G. V. J . Phys. Chem. 1986, 90, 996. (12) Sauer, J.; Schroder, K.-P. 2.Phys. Chem. (Leipzig) 1985,266,379. (13) Peri, J. B.; Hensley, A. L. J . Phys. Chem. 1986, 72, 2926. (14) Hertl, W.; Hair, M. L. Nature 1969, 223, 1950. 115) Knozinger. H. In The Hvdroeen Bond: Schuster. P.. Zundel. G.. Sandoify, C., F b:;North Holland P u b i t i o n : ’Amsterdam, 1976; p 1263: (16) Klier, K. J . Chem. Phys. 1973, 58, 737. (17) Klier, K.; Shen, J. H.; Zettlemoyer, A. C. J. Phys. Chem. 1973, 77, 1458.
(18) Kazansky, V. B.;Gitscov, A. M.; Andreev, V. M.; Zhidomorov, G. M. J. Mol. Carol. 1978. 4. -135. -(19) Anderson, J. H.; Wickersheim, K. A. Surf Sci. 1964, 2, 252. (20) Zhdanov, S. P.; Kcsheleva, C. S.; Titova, I. I. Langmuir 1987,3,960. (21) Hoffman, P.; Knozinger, E. Surf.Sci. 1987, 188, 181. (22) Fubini, B.; Bolis, V.; Giamello, E. Inorg. Chim. Acta 1987, 138, 193. (23) Frisch, M. G.; Head-Gordon, M.; Schlegel, H. B.; Raghavachari, K.; Binkley, J. S.;Gonzales, C.; Defrees, D. J.; Fox, D. J.; Whiteside, R. A.; Seeger, R.; Melius, C. F.; Baker, J.; Martin, R.; Kahn, L. R.; Stewart, J. J. P.; Fluder, E. M.; Topiol, S.; Pople, J. A. Goussian; Gaussian, Inc.: Pittsburgh, PA, 1988. (24) Mix, H.; Sauer, J.; Schroder, K.-P.; Merkel, A. Collect. Czech. Chem. Commun. 1988,53, 2191. (25) Ugliengo, P.; Saunders, V. R.; Garrone, E.J . Phys. Chem. 1989,93,
-. .
5210. (26) (27) (28) 6375. (29)
Sauer, J. J . Phys. Chem. 1987, 91, 2315. Sauer, J. Chem. Phys. Lett. 1983, 97, 275. Sauer, J.; Morgeneyer, C.; Schroder, K.-P. J . Phys. Chem. 1984,88,
Ugliengo, P.; Garrone, E.; Ferrari, A. M.; Sauer, J.; Bleiber, A. Proceedings of the 7th International Congress on Quantum Chemistry, Menton, 1991; p 145. (30) Pelmenschikov,A. G.; Paukshtis, E. A.; Zhanpeisov, N. U.; Pavlov, V. I.; Zhidomirov, G. M. React. Kinet. Catal. Lett. 1987, 33, 423. (31) Paukshtis, E. A.; Pankratiev, Y.D.; Pelmenschikov, A. G.; Burgina, E. B.; Turkov, V. M.; Yurchenko, E. N.; Zhidomirov, G. M. Kinet. Carol. 1986, 27, 1440. (32) Teunissen, E. H.; van Duijneveldt, F. B.; van Santen, R. A. J . Phys. Chem. 1992, 96, 366.
Phase Transitions in Cm and the Related Microstructure: A Study by Electron Dlfftaction and Electron Microscopy G. Van Tendelm,**+C. Van Heurcht J. Van LanduyGt S. Amelinckx? M. A. Verheijen,t P. H. M. van Loosdrecht,t and G. Meijert Universiteit Antwerpen (RUCA) Groenenborgerlaan 171, B2020 Antwerp, Belgium, and Research Institute of Materials, University of Nijmegen, 6525 ED Nijmegen, The Netherlands (Received: February 19, 1992)
The phase transition in Cm and (related) lattice defects are studied by low-temperature electron microscopy and electron diffraction. The microstructure of the room-temperature face-centered-cubic (fcc-a,,) phase is very similar to that of a low stacking fault energy fcc alloy; micro twins and stacking faults on the { 111)planes are the main defects. The phase transition fcc-a. 4 simple cubic (sc) at 249 K, is confirmed by single-crystal diffraction and the space group of the sc phase was unambiguously determined from the systematic extinctions as Pa3. Moreover a second phase transition, sc 4 fcc-2ao, is discovered. It occurs at a slightly lower temperature. It is suggested that in the sc phase the molecules still have a rotational degree of freedom about their respective (111) rotation axis, of which the orientation pattern is already fixed by the Pa3 space group. In the fcc-2a,, phase the rotation angle is found to be frozen in and to alternate between +cp and -cp along the (100) directions. The domain structure of the sc phase consists of eight variants (four translation variants for each of the two orientation variants) present in rather ill-defined regions. As a consequence of the ease of rotation of the molecules, no sharp interfaces are formed between different orientation domains.
1. Introduction It seems by now well established that in the room-temperature phase of Cm the quasi-spherical molecules are packed in a face-centered cubic arrangement.’-” When the material is contaminated with residual solvent (e.g., toluene) or in the presence of a significant fraction of other fullerenes, such as CTO,a faulted hexagonal phase may be stabilized at room temperature. However, when such material is moderately heated in the electron microscope vacuum, it transforms ‘in situ” into a somewhat faulted facecentered cubic a r r a ~ ~ g e m e n t . ~ It has been demonstrated recently by the use of X-ray powder diffraction that below 249 K the lattice becomes primitive cubic. This transformation is attributed to the loss of orientational degrees of freedom of the Cs0molecules on cooling below the transition temperature of 249 K.s-8
’Research Universiteit Antwerpen. Institute of Materials. 0022-3654/92/2096-7424$03.00/0
It is the purpose of this paper to confirm and complement these powder data by observations in the electron microscope on single-crystal specimens of high purity Cmras well as to discuss defects in the basic structure and in the low-temperature orientationally ordered phase. Moreover we shall demonstrate the existence of a second low-temperature phase which is face centered with a lattice parameter 2ao. 2. Crystal Production The Cm crystals used in this study were prepared according to descriptions given in ref 9, whereby carbon soot is produced in a dc arc discharge between two high-purity graphite electrodes in a 0.2-atm He environment. Soxhlet extraction of this soot in boiling toluene is used to separate (mainly) Cm and C.l0from the rest. This extract contains Cm and C70in a 1O:l ratio, and this unseparated material is used in the experiments described in section 3.1.2; all other experiments were performed on pure crystals. Liquid column chromatographyI0 is used to obtain C60with a 0 1992 American Chemical Society
