Revised MNDO parameters for silicon

Organometallics 1986, 5, 375-379

375

Revised MNDO Parameters for Silicon Michael J. S. Dewar,* James Friedheim, Gilbert Grady, Eamonn F. Healy, and James J. P. Stewart Department of Chemistry, The University of Texas at Austin, Austin, Texas 78712 Received May 24, 1985

MNDO has been reparametrized for silicon. The results for a wide variety of silicon-containingcompounds are in much better agreement with experiment. Enthalpies of reaction and activation are compared with results from recent high level ab initio calculations.

Introduction While MNDO calculations, using the original' parameters for silicon, have given satisfactory results in many cases,2recent extensive calculations, here and el~ewhere,~ have revealed certain inadequacies. In particular, calculations with the previous parameters showed an undue preference for divalent silicon. One manifestation of this was to be seen in reactions involving SiRz species. Reactions involving the formation of such silylenes were invariably predicted to be much too exothermic (see Table 11). The original MNDO version also performed badly for compounds containing multiply bonded silicon, greatly underestimating the strengths of the multiple bonds. For example, silaethylene and silacetylene were predicted to have bond orders of 1 and 2, respectively (Table VII). Large errors also occurred in calculations for compounds of silicon with other heteroatoms, most notably oxygen. We have now reoptimized the parameters for silicon, taking advantage of a new and much more efficient algorithm for parametrization. The increased speed of the new procedure allowed far more reference functions to be included in the basis set. In addition the 6, and 6, parameters as well as the orbital exponents ls and lpwere uncoupled and allowed to optimize independently.

Procedure The sum of squares of the differences between the calculated values for the properties and the reference values was used to define the error function SSQ. These properties included heats of formation, ionization energies, dipole moments, and geometries. The first derivatives of the heat of formation and ionization energies with respect to the various parameters were calculated analytically. The derivatives of the dipole moments were evaluated by finite difference, and the derivative of the energy with respect to geometry was used as a measure of the deviation of the calculated geometry from the experimental. A knowledge of these derivatives allowed the calculation of the derivative of the SSQ with respect to the various parameters. If the surface were purely parabolic and the exact Hessian, or second derivative, matrix was known, then the optimum set of parameters (P? could be evaluated in one step from the current set (P)using the relationship p' = p - GH-1 (1) Dewar, M. J. S.; Rzepa, H. S.; McKee, M. L. J . Am. Chem. Soc.

1978, 100, 3607.

( 2 ) Dewar, M. J. S.; Healy, E. F. Organometallics 1982, 1, 1705. (3) Verwoerd, W. S. J. Comput. Chem. 1982,3,445.

0276-7333/86/2305-0375$01.50/0

Table I. Revised MNDO Parameters f o r Silicon optimized parameters value derived " m e t e r s value UJeV -37.037 5330 Eh,,(298 K)/kcal mol-' 108.39 U,,/eV -27.7696780 Eel/eV -82.839 422 0 ra/au 1.315 9860 Dl/au 1.258034 9 r,/au 1.7099430 D2/au 0.978 582 4 &lev -9.0868040 pJau 0.360896 7 0.366 424 4 Pp/eV -1.075827 0 pz/au ar/A-' 2.205 316 0 p3/au 0.450 674 0 g,/eV 9.8200000 g,,/eV 7.310 000 0 ga,/eV 8.3600000 gp2/eV 6.540 OOO 0 h,,/eV 1.3200000

Table 11. Heats of Formation f o r a Variety of Silylenes a n d Silanes Using t h e Original Parameters MNDO obsd 86.40" 70.30 SiH 26.31 58.60b SiH2 -216.48 -147.90" SiF2 -39.60" -105.63 SiC12 -12.20c SiBr2 -51.64 SiH4 1.20 7.300 -15.53 -7.77d SiH3(CH3) -19.98d SiH2(CH3)2 -41.54 -37.43d -66.46 SiH(CH& -90.10 -57.100 Si(CH& -43.62' SiHK&H& -56.61 -85.49 -48.We SiH(C2H5)3 -64.40' Si(C2HA -1 11.03 SizHB 11.5 7.30" SisHs 32.5 29.90"

Substituted Silicon error -16.10 -32.29 -68.48 -66.04 -38.44 -6.10 -7.76 -21.56 -29.03 -31.50 -12.99 -37.69 -47.60 4.20 2.60

"Wagman, D. D.; Evans, W. H.; Parker, V. B.; Schumm, R. H.; Halow, I.; Bailey, S. M.; Churney, K. L.; Nuttall, R. L. J . Phys. Chem. Ref. Data, Suppl. 1982, 11, 2. *Vanderwielen, A. J.; Ring, M. A.; O'Neal, H. E. J. Am. Chem. SOC.1975, 97, 993. 'Schafer, H.; Braderreck, H.; Morcher, B. 2.Anorg. Allg. Chem. 1967, 352, 122. dPedley, J. B.; Iseard, B. S. 'CATCH Tables for Silicon Compounds"; University of Sussex, 1972. 'Pedley, J. B.; Rylance, J. Sussex-N.P.L. Computer Analysed Thermochemical Data, University of Sussex, 1977.

where G is the f i i t derivative matrix and H-' is the inverse Hessian. However, because of the more complex nature of the surface and the advisability of choosing small step sizes, a more conservative approach was adopted. From a knowledge of the derivatives of the SSQ for two slightly different sets of parameters an approximate Hessian was constructed. A line search along the search direction yielded the optimum step size, and this information was 0 1986 American Chemical Society

376 Organometallics, Vol. 5, No. 2, 1986

Dewar e t al.

Table 111. Calculated Heats of Formation, Ionization Potentials, and Dipole Moments for Molecules Containing Silicon AHf/kcal-mol-' IP 1eV diuole moment/D" compound SiH SiH, *SiH4 (CH3)SiH3 * (C2H5)SiH3 (CH3)2SiH2 (CzHdzS1Hz (CH3),SiH (C,H5)3SiH *(CH,),Si (C2H5)4Si *H,C=CHSiH, silylacetylene SiBr H3SiBr *(CH3),SiBr SiBr, *SiBr, Sic1 H3SiC1 (CH3),HSiC1 (CH3),SiC1 *SiCl, H2SiC1, (CH,)SiCl,H (CH3),SiC1, HSiC1, *SiC14 (CH3)SiC1, SiF *SiF2 H3SiF H,SiF, HSiF, *SiF4 (CH3)SiH2F (CH3)SiHF2 Si1 Si0 (SiHJSO (CH3),SiI * (CH,),SiOH 1,I-dimethvlsilacvclobutane *HN(Si(CH3),), *Si,H, *Si&CHde Si3Hs

calcd 90.2 64.3 1.2 -14.0 -21.6 -28.8 -44.0 -44.3 -64.0 -59.1 -82.0 6.3 34.7 57.8 -18.0 -62.3 11.2 -50.4 29.7 -43.9 -73.0 -87.4 -46.3 -83.5 -97.4 -111.2 -117.9 -147.6 -131.2 -28.8 -164.9 -96.4 -192.6 -285.1 -370.4 -110.3 -205.5 92.1 -22.6 -112.5 -34.5 -121.9 -33.7 -123.2 22.4 --73.7 31.8

obsd 86.4b 58.6' 7.3d

calcd 5.97 7.34 11.93 11.53 11.14 11.40 11.09 11.40 11.05 11.33 10.90 10.23 10.90 6.83 11.22 11.19 9.43 12.62 6.65 12.03 11.80 11.74 9.83 12.52 12.33 12.22 13.15 13.81 12.87 6.02 9.94 11.98 12.22 14.00 15.83 11.63 11.90 7.16 10.70

-7.8' -20.d -43.6h -37.d -48.0h -57.1b -64.4h -1.96 50.0b -70.0k 12.2h -99.3b 45.36 -75.96 -69.gh -84.6k -39.6b -96.0k -107.1k -122.6 -157.0* -136.7h 1.7' -147.9' -194.0d -283.0d -385.gb

10.35 10.92 10.70 9.69 9.62 9.71 9.48

-119.4k -33.0h -113.gh 19.Zb -86.ah 29.9'

obsd

calcd 0.69 0.13

obsd

0.14 0.10 0.18 0.23 0.17 0.21

0.748 0.81 0.75

12.36e

0.525

10.29' 10.4W

0.07 0.56 1.07 3.38 3.44 3.24 4.04

0.66 0.3168 1.31

10.90' 11.61"'

11.70" 11.94" 12.03'

12.85"

2.95 3.31 3.28 3.24 3.82 3.42 3.59 3.69 2.70

0.86

3.15 0.36 1.24 1.32 2.12 2.70

1.23 1.27 1.55 1.27

1.31

1.17

16.45'' 1.43 2.28 3.03 2.31 2.70 1.34 0.80 0.29

1.71 2.11 3.109 0.249

0.41

0.45

Except where noted all experimental dipole moments were obtained from: Nelson, R. D.; Lide, D. R.; Maryott, A. A. Natl. Stand. Ref. Data Ser. ( U S . , Natl. Bur. Stand.) 1967: NSRDS-NBS 10. bWagman, D. D.; Evans, W. H.; Parker, V. B.; Schumm, R. H.; Halow, I.; Bailey, S. M.; Churney, K. L.; Nuttall, R. L. J . Phys. Chem. Ref. Data, Suppl. 1982, 11, 2. cVanderwielen, A. J.; Ring, M. A.; O'Neal, H. E. J.Am. Chem. SOC. 1975,97,993. dStull, D. R.; Prophet, H. Natl. Stand. Ref. Data Ser. (U.S., Natl. Bur. Stand.) 1971, NSRDS-NBS 37. ePullen, B. P.; Carlson, T. A,; Moddeman, W. E.; Schweitzer, G. K.; Bull, W. E.; Grimm, F. A. J . Chem. Phys. 1970,53, 768. fPedley, J. B.; Iseard, B. S. "CATCH Tables for Silicon Compounds"; University of Sussex, 1972. gMcClellan, A. L. "Tables of Experimental Dipole Moments"; W. H. Freeman: San Francisco, 1963. Pedley, J. B.; Rylance, J. Sussex-N.P.L. Computer Analysed Thermochemical Data, University of Sussex, 1977. 'Jonas, A. E.; Schweitzer, G . K.; Grimm, F. A.; Carlson, T. A. J . Electron Spectrosc. 1972,1, 29. jWeidner, U.; Schweig, A. J . Organomet. Chem. 1972,39, 261. Cox, J. D.; Pilcher, G . "Thermochemistry of Organic and Organometallic Compounds"; Academic Press: London, Ser. A 1970, New York, 1970. 'Green, J. C.; Green, M. L. H.; Joachim, P. J.; Orchard, A. F.; Turner, D. W. Philos. Trans. R. SOC. 1971,67, 3425. "Frost, D. C.; Herring, F. G.; Katrib, A.; McLean, R. A. N.; A 2 6 8 , l l l . "Cradock, S.; Whiteford, R. A. Trans. Faraday. SOC. A. 1971, 641. Drake, J. E.; Westwood, N. P. C. Can. J . Chem. 1971, 49, 4033. "Bassett, P. J.; Lloyd, D. R. J . Chem. SOC. a

molecule SiH SiH, SiH4 (CH,)SiH, silylacetvlene

Table IV. Calculated Charge Distributions m o1ecu 1e

Si Si Si Si Si Si Si Si

atom (charge, e) (+0.417), H (-0.417) (+0.947), H (-0.473) (+1.603), H (-0.401) (+1.481), C (-0.367), H (-0.375) (+1.626), C (-0.530, -0.076), H (-0.389, +0.146) (+1.007), H (-0.336) (+0.721), C (-0.238) (+1.534, -0.522), H (-0.385, -0.104)

used to update the Hessian matrix. By repeating this procedure an optimum set of parameters was obtained.

SiH,Br SiH3C1 SiH,F SiH,I Br,SiF C1,SiF Si0 (SiH,),O

atom (charge, e) Si (+1.562), H (-0.353), Br (-0.505) Si (+1.610), H (-0.3581, C1 (-0.537) Si (+1.665), H (-0.381), F (-0.527) Si (+1.383), H (-0.341), I (-0.359) Si (+1.459), F (-0.446), Br (-0.338) Si (+1.633), F (-0.458), C1 (-0.392) Si (+0.746), 0 (-0.746) Si (+1.706), H (-0.404), 0 (-0.987)

The algorithm used was a modification of the DavidonFletcher-Powell (DFP) p r ~ c e d u r e . ~

Revised MNDO Parameters for Silicon

Organometallics, Vol. 5, No. 2, 1986 377

Table V. Calculated (Observed) Geometrical Parameters bond lengths (A) calcd (obsd)

molecule

bond angles (deg) calcd (obsd)

SiH SiH2 SiH, (CHJSiH,

SiH SiH SiH Sic

1.374 1.380 1.376 1.801

(1.520) (1.516) (1.481) (1.869)

H2C=CHSiH2

Sic SiH SiBr SiBr SiF SiBr SiH SiBr SiH SiH Sic1 SiH Sic1 SiH Sic1 Sic SiH Sic1 Sic1 SiH Sic1 SiF Sic Sic1 SiF SiH SiF SiH SiF SiF Sic Si1 SiH Si1 Si1 SiSi SiH Si0

1.772 1.380 2.228 2.219 1.585 2.197 1.366 2.210 1.368 1.369 2.103 1.369 2.112 1.365 2.097 1.791 1.370 2.101 2.087 1.367 2.094 1.572 1.793 2.090 1.595 1.379 1.578 1.375 1.593 1.594 1.803 2.395 1.370 2.386 2.406 2.173 1.385 1.615

(1.853) (1.475) (2.210) (2.153) (1.560) (2.170) (1.494) (2.171) (1.481) (1.561) (2.064) (1.485) (2.049) (1.480) (2.033) (1.850) (1.467) (2.040) (2.019) (1.465) (2.019) (1.520) (1.876) (2.021) (1.596) (1.480) (1.590) (1.447) (1.562) (1.574) (1.812) (2.451) (1.487) (2.437) (2.387) (2.327) (1.486) (1.634)

BrSiH3 BrSiF, Br3SiH Br,SiF ClSiH ClSiH, C12SiH2 C12HSi(CH3) C1,SiH C1,SiF C13Si(CH3) FSiH, F2Si F,SiH F3Si(CH3) ISiH ISiH, ISiF, Si2H, (SiH3)20

ref (I

HSiH

97.3

(92.1)

b C

HCH HSiH SiCC HSiH HSiBr FSiF

106.9 109.0 128.4 110.4 106.9 106.6

(107.7) (108.2) (122.9) (108.7) (107.9) (108.5)

g

BrSiBr

108.9

(111.6)

h

BrSiBr

108.7

(111.6)

h

HSiCl

99.0

(102.8)

I

HSiCl

106.9

(108.7)

f

ClSiCl

108.6

(109.7)

j

CSiH CSiCl ClSiCl ClSiCl HSiCl ClSiCl FSiCl

115.2 109.8 107.0 108.5 110.5 109.1 109.9

(110.9) (109.8) (108.8) (110.6) (108.3) (109.4) (109.5)

k

d

e

f

1 l

m

HSiF

108.4

(108.4)

n

FSiF FSiF

96.9 105.1

(100.8) (108.3)

P

HCSi

112.0

(110.0)

Q

HSiI HSiI

101.0 108.4

(102.7) (108.4)

f

HSiH SiOSi OSiH HSiH

107.7 177.4 108.9 109.7

(107.8) (144.1) (109.9) (109.1)

0

r

S

t U

"Rosen, B. 'Spectroscopic Data Relative to Diatomic Molecules"; Pergamon Press: New York, 1970. bDubois,I. Can. J . Phys. 1968, 46, 2485. 'Dang-Nhu, M.; Pierre, G.; Saint-Loup, R. Mol. Phys. 1974,28,447. dKilb, R. W.; Pierce, L. J. Chem. Phys. 1957,27, 108. 'OReilly, J. M.; Pierce, L. J . Chem. Phys. 1961,34, 1176. fKawley, R.; McKinney, P. M.; Robiette, A. G. J . Mol. Spectrosc. 1970, 34, 390. SSheridan, J.; Gordy, W. J . Chem. Phys. 1951, 19, 965. hHolm, R.; Mitzloff, M.; Hartmann, H. Z. Naturforsch., A: 1968, 23A, 1819. 'Herzberg, G.; Verma, R. D. Can. J. Phys. 1964,42,395. jDavis, R. W.; Gerry, M. C. L. J. Mol. Spectrosc. 1976,60,117. LEndo,K.; Takeo, H.; Matsumura, C. Bull. Chem. Soc. Jpn. 1977,50,626. 'Holm, R.; Mitzloff, M.; Hartmann, H. Z. Naturforsch., A 1967,22A, 1287. "Mockler, R. C.; Bailey, J. H.; Gordy, W. J. Chem. Phys. 1953, 21, 1710. "Georghiou, C.; Baker, J. G.; Jones, S. R. J. Mol. Spectrosc. 1976, 63, 89. "Shoji, H.; Tanaka, T.; Hirota, E. J . Mol. Spectrosc. 1973, 47, 268. PHoy,A. R.; Bertram, M.; Mills, I. M. J. Mol. Spectrosc. 1973, 46, 429. qDurig, J. R.; Li, Y. S.; Tong, C. C. J . Mol. Struct. 1972,14,255. rBillingsley, J. Can. J. Phys. 1972,50, 531. "ams, L. C.; Jache, A. W. J . Chem. Phys. 1967, 47, 1314. 'Shotton, K. C.; Lee, A. G.; Jones, W. J. J . Raman Spectrosc. 1973, I , 243. "Almenningen, A,; Bastiansen, 0.;Ewing, V.; Hedberg, K.; Traetteberg, M. Acta Chem. Scand. 1963, 17, 2455.

Table VI. Vibrational Frequencies compd SiHD, SiHF3 SiHCl, Si2HD, SiH,

assignt

A T E

T CH3SiH3

A1

E E A1 A1 E

descriptn Si-H stretch Si-H stretch Si-H stretch Si-H stretch symmetric stretch asymmetric stretch asymmetric bend asymmetric bend SiH, stretch SiH, stretch SiH, bend SiH, deformation Si-C stretch SiH, rock

MNDO. cm-I 2360 2404 2364 2349 2399 2347 929 991 2382 2345 1004, 958 958 803 634

exDtl. cm-' 2187.2 2316.8 2260.3 2162.5 2185 2189 972 913 2169 2166 946 .

946 701 545

MNDO. -, cm-' 2374 2338 1011

~

I

ref compd a CH,C=Ca SjH, a a b

c

assignt descriptn A Si-H stretch E Si-H stretch A Si-H deformation E Si-H deformation E SiH, rock A Si-C stretch E Si-C-C bend HSiOHa' 0-H stretch (trans) a' Si-H stretch H-Si-0 bend a' a' Si-0 stretch Si-0-H bend a' a" torsion cyclopropylSi-Ha stretch SiHDl Si-Hs stretch

~~

~

exatl. -~~ --, cm-I 2182 2182 943

960

937

749 610 189 4144 2287 975 1002 839 564 2354 2356

697 523 132 3650 1872 937 851 723 659 2162.7 2157.8

ref d

e

f

"McKean, D. C.; Torto, I.; Morrisson, A. R. J . Phys. Chem. 1982,86, 307. bKattenberg, H. W.; Oskam, A. J. Mol. Spectrosc. 1974,49, 52. Wilde, R. E. J . Mol. Spectrosc. 1962, 8, 427. dCradock, S.; Koprowski, J.; Rankin, D. W. H. J. Mol. Struct. 1981, 77, 113. 'Ismail, Z.K.; Hauge, R. H.; Fredin, L.; Kauffman, J. W.; Margrave, J. L. J. Chem. Phys. 1982, 77, 1617. (McKean, D. C.; Morrisson, A. R.; Dakkouri, M. Spectrochim. Acta, Part A 1984, 40A, 771. e

378 Organometallics, Vol. 5, No. 2, 1986

Dewar et al. Table VI1

H3SiSiH3 H2Si=SiH2 HSiFSiH

original MNDO mrb d(Si-Si)c

mfb

22.9 48.4 126.5

22.4 74.9 127.0

2.284 2.670 2.10

nGlidwell, C. J . Organomet. Chem. 1981, 217, 11.

revised MNDO d(Si-Si)'

ab initio d(Si-Si)'

ref

2.352 2.083

a

2.173 1.967 1.806

a

kcal/mol. < I n A.

Table VIII. Calculated Reaction Energies

Results and Discussion Table I shows the final set of parameters obtained for silicon in the usual notation. One unexpected result was the relative ordering of the Cs and lP parameters, the value for the latter being larger. However, this is compensated for by the fact that the value for 0,is much more negative than that for Pp, these parameters being coupled in their contribution to the one electron matrix. The one-center, two-electron parameters (g and h ) were kept constant at their previous values.' Table I1 gives the errors in the calculated heats of formation (AI&) for selected silicon compounds using the original parameters. The emphasis here is on those molecules for which the original parameters performed badly, but other compounds are also included to facilitate comparison with the present work. In addition to the previously stated large negative errors for Si" compounds, it can also be seen that progressive methyl and ethyl substitution in silane (SiHJ resulted in increasing negative errors in the heats of formation. This problem has been rectified with the new parameterization. The remaining tables, unless otherwise stated, aU contain results obtained with the revised set of parameters. Table I11 shows the heats of formation (AHf),first ionization potentials derived by using Koopmans' theorem (Il),and dipole moments (p)for a broad range of siliconcontaining compounds. The molecules included in the basis set are indicated by asterisks. Where possible, experimental values are included for comparison. The errors in the calculated heats of formation, while predictably larger than those for the organic elements (CHON),5represent a big improvement over those given by the previous set of parameters (see Table 11). Errors greater than 20 kcal/mol now occur only in one or two exceptional cases. The errors in the calculated ionization energies are large, varying from 0.17 to 1.78 eV. Similar errors have also been found for the other third-period elements6,' and, in the m e of large negative errors, have been attributed to use of the core approximation in MNDO. In any case, all attempts to rectify this problem in the case of silicon resulted in an impairment of the results for other properties. The errors in the calculated dipole moments are comparable with those reported for S6 and (21.' They show no trend. The calculated charge distributions (Table IV) give formal charges much greater than what would be expected from the standard electronegativity values and are also greater than those predicted by the original MNDO parameters. This latter fact is attributable to the value for the U, parameter, which is 3.5 eV more positive than the original value. Though unsatisfactory, this failing was considered less important than the dramatic improvements in the calculated heats of formation (Tables I1 and 111). (4) Fletcher, R.; Powell, M. J. D. Comput. J . 1963,6, 163. Davidon, W. C. Zbid. 1968, IO, 406. (5) Dewar, M. J. S.; Thiel, W. J. Am. Chem. Soc. 1977, 99, 4899, 4907. (6) Dewar, M. J. S.; McKee, M. L. J. Comput. Chem. 1983, 4, 84. (7) Dewar, M. J. S.; Rzepa, H. S. J . Comput. Chem. 1983, 4 , 158.

--+ --+ -

reactn H2C=SiHz H,CSiH HCESiH H2C=Si H2Si=SiH2 H,SiSiH H2C=Si + H2 HzC=SiHz H2C=SiH2 H2 H3CSiH3 HC=SiH Hz H2C=SiH2 H2 S i 0 H2Si=0 H2Si=0 HSiOH(trans) H 2 0 H2Si=0 H,Si(OH)2

+ +

MNDO, kcal/mol +9.2 -20.0 +15.3 -43.3 -52.1 -63.3 -2.9 -9.6 -73.1

ab initio -0.4 -49.1 -8.1 -31.0 -56.9 -71.7 2.5 -3.7 -72.6

ref a

b a C C C

d d d

aYoshioka, Y.; Goddard, J. D.; Schaefer 111, H. F. J. Am. Chem. SOC. 1981, 103, 2452. *Hoffmann, M. R.; Yoshioka, Y.; Schaefer 111, H. F. J.Am. Chem. SOC.1983, 105, 1084. 'Gordon, M. S.; Pople, J. A. J . Am. Chem. SOC.1981, 103, 2945. dKudo, T.; Nagase, S. J . Phys. Chem. 1984, 88, 2833.

Table IX. Calculated Enthalpies of Activation

--

reactn H2Si=0 HSiOH(trans) HSiOH(trans) HSiOH(cis) H2Si=CH2 HSiCH3

MNDO, kcal/mol 79.2 5.4 59.1

ab initio 60.8 9.3 41.0

ref a a

b

"Kudo, T.; Nagase, S. J . Phys. Chem. 1984, 88, 2833. bYoshioka, Y.; Goddard, J. D.; Schaefer 111, H. F. J . Am. Chem. SOC.1980, 102, 7644.

Table V compares the calculated geometries with experiment. The errors in the bond lengths are slightly greater than those found with the previous set of parameters, while those for the bond angles are similar (1.8'). Si-C and Si-H bonds are consistently predicted to be too short, the average errors being 0.06 and 0.125 A, respectively. While r,iuch greater than the corresponding errors for the organic dements (CHNO), these discrepancies in S i 4 and Si-H bonds are systematic. The errors in lengths of bonds between silicon and halogens are comparable with those given by MNDO for compounds of other third-period elements. Table VI compares the calculated and observed vibrational frequencies for a number of silicon-containingcompounds. As in other cases,8 the calculated values are systematically (only one vibrational mode out of the 29 reported is predicted too low in frequency) too high, by -9%. Table VI1 gives the heats of formation (AHf) and bond lengths for Si-Si multiply bonded species and compares these with the results obtained from the previous set of parameters and with ab initio results. A more thorough MNDO investigation of disilenes and related molecules will be the subject of a future paper; however, it should be noted here that the revised MNDO bond lengths show the expected decrease with the increase in bond order. This was not the case with the original parameters. Tables VI11 and IX involve comparisons of our calculated values, not with experiment, but with recent ab initio results. Table VI11 compares heats of reaction for a number of rearrangements of silicon compounds and for (8) Dewar, M. J. S.; Ford, G . P. J . Am. Chem. SOC.1977, 99, 1685.

Organometallics 1986,5, 379-380

some additional reactions. The agreement is highly satisfactory, especially in view of the fact that most of the ab initio calculations were carried out at a very high level. The calculated enthalpies of activation reaction also show good agreement, except for reactions involving hydrogen migration. MNDO is well-known to give values that are much too large in such cases. Acknowledgment. This work was supported by the Air Force Office of Scientific Research (Contract No. F49620-83-C-0024) and The Robert A. Welch Foundation (Grant F-126), The National Science Foundation (Grant CHE82-17948), and the Tektronix Foundation. The calculations were carried out by using a DEC VAX 11/780 computer purchased with funds provided by the National Science Foundation and The University of Texas at Austin. Registry No. SiH, 13774-94-2;SiH2,13825-90-6;Si, 7440-21-3;

N

a

379

SiF2, 13966-66-0; SiC12, 13569-32-9; SiBr2, 14877-32-8; SiH,, 7803-62-5; SiH3(CH3), 992-94-9; SM2(CHJ2,1111-74-6;SiH(CH3),, 993-07-7; Si(CH3)4,75-76-3; SiHz(C2H5)2, 542-91-6; SiH(C2H5),, 617-86-7;Si(C2H5),,631-36-7;Si2& 1590-87-0;Si3H8,7783-26-8; (C2H5)SiH3,2814-79-1; H2C=CHSiH3, 7291-09-0; HCZCSi, 1066-27-9; SiBr, 12350-21-9; H3SiBr, 13465-73-1; (CH3),SiBr, 2857-97-8;SiBr,, 7789-66-4;SiC1,13966-57-9;H,SiCl, 13465-78-6; (CH3),HSiC1,1066-35-9;(CH,)#iCl, 75-77-4; H2SiC12,4109-96-0; (CH3)SiC12H,75-54-7; (CH3)2SiC12, 75-78-5; HSiCl,, 10025-78-2; Sic4,10026-04-7; (CH3)SiC13,75-79-6; SiF, 11128-24-8; H,SiF, 13537-33-2;H&IiF2,13824-36-7;HSiF3,13465-71-9;Sip,, 7783-61-1; (CH3)SiH2F, 753-44-6;(CH3)SiHF2, 420-34-8; SiI, 13841-19-5;SiO, 10097-28-6; (SiH,)SO, 99583-30-9; (CH3),SiI, 16029-98-4; (CH3),SiOH, 1066-40-6;HN(Si(CH3)3)2, 999-97-3;Si2(CH3)6, 145014-2; (SiH3)20,13597-73-4;ISiH, 36098-67-6;ISiH,, 13598-42-0; ISiF,, 16865-60-4; SiHD,, 13537-02-5;Si2HD5,77815-94-2;HSiOH (trans),83892-34-6;cyclopropyl-SiH02,92917-37-8;H2Si=SiH2, 15435-77-5; HSi