Molecular Structures and Vibrational IR Spectra of GeX and XGeY

---

12189

J. Phys. Chem. 1993,97, 12189-12192

Molecular Structures and Vibrational IR Spectra of GeX and XGeY Species (X, Y = 0, S, Se) by ab Initio Hartree-Fock and Post-Hartree-Fock Study Jerzy Leszczyhski' Department of Chemistry, Jackson State University, 1400 Lynch Street, Jackson, Mississippi 3921 7 Jozef S. Kwiatkowskit Instytut Fizyki, Uniwersytet M . Kopernika, ul. Grudziadzka 5, 87- 100 Toruh, Poland Received: April 13, 1993; In Final Form: September 24, 1993"

Three diatomic GeX and six triatomic linear XGeY (X, Y = 0, S,Se) molecules were studied by theoretical methods. Equilibrium geometries of the studied species were located a t the HF and MP2 levels using valence triple-{ basis set with two sets of d-polarization functions (TZP) on all elements. The predicted molecular parameters, dipole moments, rotational constants agree well with the available experimental data. IR spectra calculated a t the M P 2 / T Z P level for oxo and thio species reproduce accurately experimental harmonic vibrational frequencies and intensities, and this accuracy strongly suggests that the same level of prediction furnishes reliable data for still elusive triatomic seleno compounds. An analysis of the predicted molecular bond distances justifies the observed experimental reversal in the trend of diatomic and triatomic bond stretching force constant from C to Si and Ge.

TABLE I: Predicted and Experimental Molecular

Introduction

Parameters of GeX (X = 0, S, Se)

There is still considerably less amount of data available for germanium compounds than for their lighter analogues. Though the known chemistry of germanium and silicon exhibits many similarities, there are some important differences between these two group IVAelements. As a matter of fact the analogy between these compounds seems to be overemphasized, and for example Lesbre et al. furnished more than 10 examples of significant chemical differences in organochemistryof silicon and germanium compounds. More information becomes currently available for model germanium compounds. Triatomic OGeO* and SGeS3 species were studied by matrix isolation IR techniques. Both molecules were found to possess D-h linear structure^.^^^ Recent experiments with atomic germanium reactions with oxygen and sulfur in argon matrices and identifications of the reaction products by infrared spectroscopic techniques suggested formation of the linear XGeY species (X, Y = 0, S).4 In the IR studies by Hassanzadeh and Andrews the analysis of the infrared spectra of all stable isotopes of SGeS, SGeO, and OGeO as well as GeO and GeS were performed and the fundamental frequencies were a ~ s i g n e dAb .~ initio Hartree-Fock (HF) level calculations by Koppe and Schnockel predicted bond lengths for all these specie^.^ However, the other experimental molecular parameters for the triatomic molecules are still unknown. The experimental data for GeX (X = 0, S) are richer than those for corresponding triatomic molecules and include information on their bond distances, dipole moments, rotational constants, and IR frequencies.s-l7 This gives an opportunity to test various levels of calculations necessary to obtain accurate predictions for these systems and to adopt the most reliable level to study properties of experimentally elusive compounds. An unique property of the germanium oxo and thio species was noticed by Hassanzadeh and A n d r e ~ s . The ~ analysis of Ge-O and G e S bond stretching force constants indicated that bond lengths in XGeX molecules with presumably lower bond f Visiting (1991/1992) the Department of Chemistry, Jackson State University, Jackson. MS 39217. Abstract published in Aduoncr ACS Absrrocrs, October 15, 1993.

0022-3654/93/2091- 121 89$04.00/0

quantity

HF

calculation level MP2

experimentb

7 4 ~ ~ 1 6 0

r(Ge0) 1

B EHF EMP2

r(GeS) P

B EHF EMPZ

159.12 3.881

164.67 2.575 (4.203) 15 179.265 14 173.988 21 48.348 90 2 148.345 68 2 148.777 13

164.0W 162.7 2.806c 3.282 38(30) (4.074)c 14 289.55lC 14 514.703(15)

74Ge32S 202.79 1.970 (2.952) 5671.958 5506.032 2 471.047 18 2 471.046 67 2 471.490 52 199.81 2.758

74Ge*oSe 214.60 1.546 (2.548) 2925.178 2857.742 4471.25604 4471.255 73 4471.618 62

201.4 2.00(6) 5593.1019(22y

r(GeSe) 212.11 U 2.401

213.5 1.648(50) . .

B EHF

2888.21 14(20)"

EMPZ

Bond lengths in r in picometers,dipole momentsp in debye, rotational constants B in megahertz, energies in -au. Dipole moments calculated from the HF densities at the MP2 geometries in parentheses. b From ref 26. c 6-31 1G (3df) basis set on the oxygen atom. Dunham coefficient.

orders are shorter than those in GeX species. Theoretical verification of this observation augmented by inclusion of the information for selenocompounds is presented in this paper. The main goal of the present study is to determine structures and properties of the series of two and three atomic germanium compounds and also to predict trends in their molecular geometries and properties on substitution with progressively heavier atoms (0,S, Se) down the periodic table, using uniform, HartreeFock, and MP2 levels of ab initio theory. Also we shall furnish accurate, theoretically calculated molecular parameters as well as harmonic vibrational wavenumbers and intensities for seleno compounds to guide experimental studies on this species. 0 1993 American Chemical Society

Leszczyfiski and Kwiatkowski

12190 The Journal of Physical Chemistry, Vol. 97, No. 47, 1993 TABLE 11: Predicted Molecular Parameters of XGeY (X, Y = 0, S, Se) calculation level quantityg

HF

r(GeX) P

B EHF EMP2 r(GeY) r(XGe) P

B EHF EMP2

calculation level MP2

HF

calculation level MP2

HF

MP2

16074Gel60 158.72 162.96 0.0 0.0 (0.0) 6271.39 5948.94 2 223.173 72 2 223.170 15 2 223.865 06

32Sl4Ge32S 198.52 200.36 0.0 0.0 (0.0) 2005.3 5 1968.78 2 868.594 84 2 868.594 46 2 869.322 51

80Se74Ge80Se 21 1.34 212.38 0.0 0.0 (0.0) 707.902 701.95 6 869.018 37 6 869.018 30 6 869.586 25

16074Ge32S 198.24 200.42 158.81 162.96 1.574 1.095 (1.688) 3278.38 3 178.26 2 545.884 66 2 545.882 71 2 546.594 37

16014Ge80Se 211.13 212.09 158.84 163.06 2.037 1.560 (2.231) 1833.44 1799.15 4 546.096 41 4 546.094 56 4 546.693 79

32S74Ge80Se 211.26 21 1.94 198.69 200.65 0.448 0.498 (0.533) 1180.466 1194.419 4 868.806 69 4 868.806 45 4 869.454 21

a Bond lengths r in picometers, dipole moments p in debye, rotational constants B in megahertz, energies in -au. Dipole moments calculated from the H F densities at the MP2 geometries in parentheses.

TABLE 111: Predicted' and Observed Wavenumbersb (u, cm-l) and Absolute Intensities ( A , km/mol) of GeX and XGeY (X, Y = 0, S, Se) calcd HF V

A

calcd MP2 74~el60 958, 959c 13, 14c

1141 110

exptl

HF

97 1

630 80

MP2 74Ge32S 578 25

520 0 148 18 685 155

32Sl4Ge32s 496 0 132 6 682 97

392 8 182 56 1107 123

16074Ge80Se 394 6 163 21 1035 78

16074~el60 VI

A v2

A v3

A VI

A y2

A y3

A

1000 0 252 148 1172 90 588 19 203 68 1107 115

924 0 215 58 1135 90 16014Ge32S 567 12 179 27 1047 78

calcd

1048

984

exptl

HF

567

440 38

653

296 0 107 6 506 95 368 6 126 12 637 125

MP2 74Ge80Se 406 12

exptl

80Se74Ge80Se 292

0 101 1 521 62 WYie80Se 365 2 115 3 63 1 80

Wavenumbers are ordered in a standard spectroscopic ordering. Experimental data taken from ref 26. The intensities of the degenerated modes ( 4 are multiplied by a factor of 2. e6-311G(3df) basis set on oxygen.

TABLE I V Observed' and Calculated (MP2) Shifts (cm-1) of Fundamental Vibrations in GeO and GeS on Isotopic Substitution GeS

GeO 70~e160 72~e160 73~e160 l6GeI60 lOGe180 l2GeI80 73Ge180 74Ge180 l6Gel80

exptl +4.9 +2.3 +1.3 -2.1 -40.1 -42.7 -43.8 -45.1 -47.5

calcd +5 +2 +1 -2 -40 -43 -44 -45 -47

10Ge32S 72Ge32S '3Ge"S 16GG22S l0Ge"S 12Ge34S 73Ge34S 14Ge"S 16GG4S

exptl

calcd

+4.9 +2.4 +1.3 -2.2 -6.8 -9.3 -10.4 -11.7

+5 +2 +1 -3 -7 -10 -11 -12

-14.0

-15

Experimental data taken from ref 3.

Method Calculations were performed on the title compounds using a b initio LCAO-Mo method.'* Equilibrium structures were initially computed at the Hartree-Fock level using standard gradient optimization techniques.19 The optimized molecular parameters were used as input data for geometry optimizations performed at the second-order Moller-Plesset perturbation theory MP2(full) level.20 The basis set used in the reported calculations was standard McLean-Chandler's valence triple-{augmented by two sets of six d-polarization functions on oxygen ({I = 0.646, si =

2.584) and sulfur (5; = 0.325, t 2 = 1.300),21 while partially uncontracted [43311 1/43 111/4*] Huzinaga's basis set supplemented with two sets of six d-polarization functions was used for germanium (tl = 0.108, si = 0.382) and selenium (ti = 0.144, 5; = 0.489).22 For the convenience; this modified basis set will be refereed to as TZP. The optimizations of molecular geometries were carried out within the appropriate (D,,, for XGeX and C,, for GeX and XGeY species, respectively) symmetry. At the HF/TZP and MP2/TZP optimized geometries, the harmonic vibrational wavenumbers and absolute intensities were calculated at the same levels. Additional geometry optimizations and vibrational frequency calculations were carried out for GeO a t the MP2 level using 6-3 1lG(3df) basis set on oxygen and previously described TZP basis set with two sets of polarization d functions on germanium. All calculations reported were performed with the GAUSSIAN 90 and GAUSSIAN 92 programs.23 Using the force constant matrices of the natural isotopic compounds, the harmonized wavenumbers of other isotopic species were calculated with the PACK program.24 Results and Discussion Previous studies on gallium and selenium compounds25 confirmed the reliability of calculations performed a t thelevel applied

The Journal of Physical Chemistry, Vol. 97, No. 47, 1993

GeX and XGeY Species

12191

TABLE V: Observed and Calculated (MP2) Shifts (cm-I) of Antisymmetric Stretching Fundamental Vibrations in 0 0 , SGeS, and OCeS on Isotopic Substitution OGeO

OGeS

exptl

calcd

exptl

calcd

exptl

calcd

+9.0 +4.4 +2.2 -4.1

+10 +5

+8.5 +4.2 +2.1 -4.0 +3.6 -0.8 -2.8 -4.9 -8.9 -1.7 -6.1 -8.3 -11.3 -14.4

+9 +5 +3

+6.1 +3.0 +1.5 -2.8 +5.8 +2.7 +1.2 -0.3 -3.0 -35.8 -39.2 -40.9 -42.3 -45.3

+7 +3 +2 -3 +6 +3 +1 0 -3 -38 -42 -44 -45 -49

-9.0 -13.5 -1 5.7 -17.9 -22.0 -3 1.9 -36.6 -38.9 -41.2 -45.5 a

SGeS

+2 -5 -10 -15 -18 -20 -24 -35 -40 -43 -45 -50

-5 +4 0 -3 -5 -9 -1 -6 -8 -10 -15

Experimental data taken from ref 4.

in the present study for the systems within the fourth-period elements. Additional confirmation of the accuracy of the predicted data for compounds containing germanium could be obtained by comparing experimental data available for model germanium species with the calculated parameters. Fortunately, the experimental geometries, dipole moments and rotational constants are known for germanium monoxide, monosulfide and monoselenide (GeX, X = 0, S, Se).26 In Table I we gathered the calculated molecular parameters for the studied molecules and compared them with the corresponding experimental values. Overall, good agreement between predicted at the HF and MP2 approximations and experimental bond lengths and rotational constants for all GeX systems should be noticed. The differences between the calculated and experimental distances are in the order of 0.02 A. The experimental data lie between H F and MP2 values, being slightly closer to the parameters calculated at the electron correlated level. The largest difference between predicted and experimental parameters is observed for germanium monoxide (1.2% and 2.3% for bond distance and rotational constant, respectively), those for the bond lengths of the S (Se) compounds amounts to 0.7% (0.6%) while rotational constants deviate by 1.6% (1.4%). One can expect similar accuracy of the calculated molecular geometries and rotational constants for the linear triatomic systems presented in Table 11. Also the deviations of the computed dipole moments from the experimental value for GeO (0.60 and 0.71 D a t the HFand MP2 levels, respectively) are much larger than those for GeS (0.76 and 0.03 D) and GeSe (0.75 and 0.19 D). In all cases the H F level dipole moments are larger, while the MP2 dipole moments are smaller than the corresponding experimental data. From very recent studies on f~rmamide,~’ and ketene,*5c it is obvious that only an electron-correlated level could reliably approximate the experimental dipole moments. However, for these model systems, accuracy within 0.1 D required a much larger basis sets of polarization functions than that applied in the present study, and the assumed accuracy was accomplished at the MP2/6-311G(3df,2p) level. Likewise, our present study concludes that there is a notable improvement of the predicted molecular parameters for GeO upon addition of the set of d- and f-polarization functions (631 lG(3df)) to the oxygen basis set. With this basis set at the MP2 level the differences between experimental and theoretical data are 0.8%, 1.6%, and 0.48 D for the bond distances, rotational constants, and dipole moments, respectively. An addition of the polarization functions to the basis set of germanium should further improve theoretical predictions. The only sources of experimental data for triatomic XGeY (X, Y = 0, S) species are the matrix IR spectra.24 Also IR data supplement experimental microwave information available for germanium monoxide and monosulfide. Calculated (unscaled)

harmonic vibration frequencies and IR intensities and the corresponding experimental parameters are given in Table 111. Fair agreement of the values predicted at the MP2 level and the experimental parameters has been noticed for GeO and GeS, the differences between unscaled theoretical wavenumbers and their experimental counterparts being 13 and 9 cm-l for germanium monoxide and sulfide, respectively. The analogous differences between the H F and experimental wavenumbers amount to 170 (GeO) and 63 cm-l (GeS). Interestingly, thedifferences between two sets of theoretical wavenumbers calculated at the MP2 and H F levels decrease notably with the increase of atomic mass of the 6th group element and amount only to 34 cm-l for GeSe. A similar effect is also observed for the triatomic species. However, there are noticeable differences between intensities calculated at the MP2 and H F approximations. Surprisingly, the MP2 level vibrational frequency for germanium monoxide is lower than that recorded experimentally. Since additional polarization functions on oxygen improved the predicted molecular parameters and dipole moment for GeO, we carried out further calculations of the vibrational frequencies for this species at the MP2 level using this 6-31 lG(3df) basis function on oxygen, while polarization functions on germanium were represented by two sets of d functions. However, the calculated wavenumber improves only by 1 cm-l compared with the results of the previous MP2 level calculations, still being 12 cm-1 lower than the experimental one. The observed difference of 87 cm-1 between thevalue predicted at the MP2 level and the experimental frequency for OGeO suggests a noticeable anharmonicity of this molecule. One can expect that the vibrations in SGeS molecule, where the oxygens are replaced by heavier sulfurs, are better approximated by a harmonic model. This assumption can be confirmed by the decrease of the differences between theoretical and experimental vibrational wavenumbers to 29 cm-I, while for the mixed OGeS compound the analogous difference amounts to 63 cm-I. We believe that one expect even better accuracy of the predictions for selenogermane species. In Tables IV and V we gathered the calculated and experimental shifts of the fundamental modes upon isotopic substitutions for oxo and thio species. The accuracy of the theoretical data is within one wavenumber for both diatomic molecules and for SGeS and within 5 cm-I for OGeO and OGeS species. As it was pointed out by Hassanzadeh and Andrews for oxo and thio species: the striking feature of these germanium compounds is that there is a reversal in the trend of diatomic and triatomic bond stretching force constant from C to Si and Ge. The diatomics GeO and GeS are characterized by smaller force constants than those in OGeO and SGeS molecules. This feature could be rationalized by the predicted molecular parameters of these molecules. Moreover, inclusion of selenocompounds in this

12192 The Journal of Physical Chemistry, Vol. 97, No. 47, 1993

prediction would give more details about germanium bonds with 6th group elements. Bond distances (Table I and 11) calculated at both H F and MP2 levels are longer for diatomic molecules than those for triatomic species by 0.01,0.02, and 0.02 A (MP2 level) for X = 0,S,and Se, respectively. Also Koppeand Schnockel concluded based on the H F level calculations that the G e S bond lengths in GeS is 0.014 A longer than that in SGeS; however, they predicted that the Ge-0 bond distance is 0.008 A shorter in GeO than in OGeO.3 From the combined IR spectroscopic and ab initio studies at the HF level the same authors concluded that there are virtually no differences between Si-0 and S i 4 multiple bonds in the di- and triatomic S i x and XSiX species.29 This characteristic of heavier (Si, Ge) group IV elements is in contrast with the properties of the analogous carbon compounds where C-0 and C-S bonds are shorter by ca. 0.03 A in diatomicsystems than in the triatomic counterpart^.^^ Interestingly, though monotonic change of bonding properties of the group IV elements has been observed, the trend in bond lengths of the two group VI elements (Sand Se) in diatomic GeX and triatomic XGeX species seems to be the same. ,

Leszczydski and Kwiatkowski Bm, A.; Ogden, J. S.; Orgee, H. J. Phys. Chem. 1974, 78, 1763. Koppe, R.; Schnockel, H. J . Mol. Scrucr. 1990, 238,429. Hassanzadeh, P.; Andrews, L. J. Phys. Chem. 1992, 96, 6181. Meyer, 8.;Smith, J. J.; Spitzer, K. J. Chem. Phys. 1970,53,3616. Meyer, B.; Jones, Y.; Smith, J. J. J. Mol. Spectrosc. 1971,37, 100. Torring, T. Z . Naturforsch. 1966, 210, 287. Hoeft, J.; Lovas, F. J.; Tiemann, E.; Tischer, R.; Torring, T. Z. Naturforsch. 1969, 240, 1217. (9) Sticda, W. U.; Tiemann, E.; Torring, T.; Hoeft, J. Z . Naturforsch. 1976, 31a, 374. (10) Raymonsa, J. W.; Muenter, J. S.; Klemperer, W. A. J . Chem. Phys. 1970, 52, 3458. (1 1) Stephenson, D. A.; Dickinson, J. T.; Zorn, J. C. J . Chem. Phys. 1970, 53, 1529. (12) Shapiro, C. V.; Gibbs, G. C.; Laubengayer, A. V. Phys. Rev. 1932, 40, 354. (13) Borrow, R. F. Proc. Phys. SOC.,London 1941, 53, 116. (14) Drummond, G.; Barrow, R. F. Proc. Phys. Soc., London 1937,49, 543. (15) Uehara, H.; Horiai, K.; Sueoka, K.; Nakagawa, K. Chem. Phys. Lett. 1989, 160, 149. (16) Marino, Ch. P.; Guerin, J. P.; Nixon, E. R. J . Mol. Spectrosc. 1974, 51, 160. (1 7) Hoeft, J.; Lovas, F. J.; Tiemann, E.; Torring, T. J . Chem. Phys. 1970, 53, 2736. (18) See,for example: Hehre, W. J.; Radom, L.; Schleyer P. v. R.; Pople,

Conclusions

J. A. Ab Initio Molecular Orbital Theory;John Wiley and Sons: New York,

The molecular parameters calculated for GeX and XGeY species are predicted well by the HF/TZP and MPZ/TZP level ofthmry, thoseat thecorrelated level beingslightlymoreaccurate. The electron-correlated level significantly improves prediction of the dipole moments and IR spectra though, for seleno compounds differences between H F and MP2 level harmonic frequencies are much smaller than for oxocompounds. Still elusive triatomic selenospecies are characterized by a reliable set of molecular and IR spectral parameters. The predicted bond distances of germanium diatomic and triatomic species allow an explanation of the experimentally observed trend of bond stretching force constant that is opposite to that displayed by their carbon counterparts.

Schlegel, H. B. J . Comput. Chem. 1982.3, 314. Maller, C.; Plesset, M. S. Phys. Reu. 1934, 46, 618. McLean, A. D.; Chandler, G. S. J . Chem. Phys. 1980, 72, 5639. Huzinaga, S.; Andzelm, J.; Klobukowski, M.; Radzio-Andzelm, E.; Sakai, Y.;Tatewaki, H. GaussianBasis Setsfor Molecular Orbital Calcularion; Elsevier: New York, 1984. (23) (a) Frisch, M. J.; Head-Gordon, M.; Trucks, G. W.; Foresman, J. B.; Schlegel, H. B.; Raghavachari, K.; Robb, M.; Binkley, J. S.; Gonzalez, C.; Defrees. D. J.; Fox, D. J.; Whiteside, R. A,; Seeger, R.; Melius, C. F.; Baker, J.; Martin, R. L.; Kahn, L. R.; Stewart, J. J. P.; Topiol, S.;Pople, J. A. GAUSSIAN 90, Reuision W,Gaussian, Inc.: Pittsburgh, PA 1990. (b) Frisch, M. J.; Trucks, G. W.; Head-Gordon, M.; Gill, P. M.W.; Wong, M. W.; Foresman, J. B.; Johnson, B. G.; Schlegel, H. B.; Robb, M.; Repiogle, E. S.; Gomperts, R.; Andres, J. L.; Raghavachari, K.; Binkley, J. S.;Gonzalez, C.; Martin, R. I.; Fox, D. J.; Defrees, D. J.; Baker, J.; Stewart, J. J. P.; Pople, J. A. GAUSSIAN 92, Reuision A; Gaussian, Inc.: Pittsburgh, PA, 1992. (24) KuBulat, K. Ph.D. Dissertation, University of Florida, 1989. Person, W. B.; KuBulat, K. Unpublished work (private information). (25) (a) Leszczynski,J.; Hale, B.; Leszczynska, D. Inr. J. Quantum Chem. QCS 1991,25,451. (b) Leszczynski, J.; Kwiatkowski, J. S.;Leszczynska, D. Chem. Phys. Lett. 1992, 194, 157. (c) Leszczynski, J.; Kwiatkowski, J. S. Chem. Phys. Lett. 1993,201,79. (d) Leszczynski, J.; Kwiatkowski, J. S. J . Phys. Chem. 1993,97,1364. (e) Leszczynski, J.; Kwiatkowski, J. S. J. Phys. Chem. 1992, 96,4148. (26) Lovas, F. J.; Tiemann, E. J . Phys. Chem. ReJ Data 1974, 3, 609. (27). (a) Kwiatkowski, J. S. Leszczynski, J. J. Mol. Struct. 1992,270,67. (b) Kwiatkowski, J. S.Leszczynski, J. J . Mol. Struct., in press. (28) (a) Kwiatkowski, J. S. Leszczynski, J. Inr. J . Quantum Chem., Quantum Chem. Symp. 1992,26,421. (b) Kwiatkowski, J. S. Leszczynski, J. J. Mol. Spectrosc. 1993, 157, 540. (29) Schnockel, H.; Koppe, R. J. Am. Chem. Soc. 1989, 111, 4583.

Acknowledgment. This work was supported in part by the National Science Foundation with the Grant RII-8902064 (U.S.A.)and in part by the Committee for Scientific Research (Poland) with the Grant KBN-2P303.031.05. The authors wish to thank the Mississippi Center for Supercomputing for an allotment of computer time to perform some of the calculations which are presented here. References and Notes (1) Lesbre, M.; Mazerolles, P.; Satge, J. The Organic Compounds of Germanium; John Wiley & Son: New York, 1971.

1986. (19) (20) (21) (22)