THE JOURNAL OF CHEMICAL PHYSICS 124,

Transcription

1 THE JOURNAL OF CHEMICAL PHYSICS 124, Monobridged Si 2 H 4 M. C. McCarthy a Harvard-Smithsonian Center for Astrophysics, Cambridge, Massachusetts and Division of Engineering and Applied Sciences, Harvard University, Cambridge, Massachusetts Z. Yu Department of Chemistry, Harvard University, Cambridge, Massachusetts L. Sari b and H. F. Schaefer III Center for Computational Quantum Chemistry, University of Georgia, Athens, Georgia P. Thaddeus Harvard-Smithsonian Center for Astrophysics, Cambridge, Massachusetts and Division of Engineering and Applied Sciences, Harvard University, Cambridge, Massachusetts Received 19 August 2005; accepted 3 January 2006; published online 15 February 2006 The rotational spectrum of a new monobridged isomer of Si 2 H 4, denoted here as H 2 Si H SiH, has been detected by Fourier transform microwave spectroscopy of a supersonic molecular beam through the discharge products of silane. On the basis of high-level coupled cluster theory, this isomer is calculated to lie only 7 kcal/mol above disilene H 2 SiSiH 2, the most stable isomeric arrangement of Si 2 H 4, and to be fairly polar, with a calculated dipole moment of =1.14 D. The rotational spectrum of H 2 Si H SiH exhibits closely spaced line doubling, characteristic of a molecule undergoing high-frequency inversion. Transition state calculations indicate that inversion probably occurs in two steps: migration of the bridged hydrogen atom to form silylsilylene, H 3 SiSiH, and then internal rotation of the SiH 3 group, followed by the reverse process. The potential energy surface for this type of inversion is quite shallow, with a barrier height of only 2 3 kcal/mol. Searches for the rotational lines of silylsilylene, calculated to be of comparable stability to H 2 Si H SiH but about five times less polar =0.23 D, have also been undertaken, so far without success, even though strong lines of H 2 Si H SiH have been detected. The favorable energetics and high polarity of monobridged Si 2 H 4 with respect to either disilene or silylsilylene make it a plausible candidate for radioastronomical detection in sources such as IRC , where comparably large silicon molecules such as SiS, SiC 3, and SiC 4 have already been discovered American Institute of Physics. DOI: / I. INTRODUCTION Silicon hydrides are of considerable interest because most possess several isomers of unusual shape and comparable stability in which both silicon and hydrogen are bonded to several atoms. Determination of relative stabilities, structures, and other properties of these unfamiliar molecules has proven to be a computational challenge, even with high-level theoretical methods. Silicon hydrides are also of practical interest because they are thought to be important in silicon vapor deposition 1 and the production of semiconductors. The silicon-containing molecules SiH 4, SiS, and SiC and larger silicon carbides up to SiC 4 have been detected in rich astronomical sources, 2 suggesting that other silicon molecules may be found in space once precise laboratory frequencies exist. Silicon hydrides of the form Si 2 H m, m=2 4, have been the subject of many theoretical calculations, but few highresolution spectroscopic studies. To date only disilyene, Si 2 H 2, has been studied in much detail; the rotational spectra of the dibridged singlet Si H 2 Si, the most stable isomer, 3 and a low-lying monobridged isomer Si H SiH, calculated to lie about 8.7 kcal/mol higher in energy, were first measured by millimeter-wave absorption spectroscopy by Bogey and co-workers. 4,5 By means of isotopic substitution, precise experimental structures were subsequently reported for both. 6,7 A third singlet isomer with a vinylidenelike structure, H 2 SiSi, is predicted to lie only 2.9 kcal/mol above the monobridged isomer, but its rotational spectrum has never been measured, possibly owing to a small dipole moment calculated to be only 0.1 D and a low barrier 2.4 kcal/mol to rearrangement. 3 Little spectroscopic data exist for the isomers of Si 2 H 3 and Si 2 H 4 ; preliminary laboratory detections of a nearly planar isomer of H 2 SiSiH and a new isomer of Si 2 H 4 were recently reported by several of us. 8 Ab initio calculations of Si 2 H 4 have focused almost exclusively on the stability and structures of two isomers: disilene denoted here H 2 SiSiH 2 and silylsilylene H 3 SiSiH shown in Fig. 1. The most recent studies conclude that disilene is the ground state and that silylsilylene lies 5 10 kcal/ mol higher in energy, but there is little agreement as to the precise geometry of disilene, with roughly half prea Electronic mail: mccarthy@cfa.harvard.edu b Present address: Department of Physics, Fatih University, Buyukcekmece, Istanbul 34500, Turkey /2006/124 7 /074303/7/$ , American Institute of Physics

2 McCarthy et al. J. Chem. Phys. 124, FIG. 1. Structures of the three low-lying singlet isomers of Si 2 H 4 : 1 disilene, 2 the monobridged isomer here, and 3 silylsilylene. dicting a planar structure and the rest a trans-bent structure. 9 Although the two most stable isomeric arrangements of Si 2 H 2 possess bridged structures, the few bridged structures of Si 2 H 4 which have been calculated to date were found to be much less stable than disilene. 10,11 Motivated by the recent laboratory discovery of the two new silicon hydrides, highly correlated coupled cluster calculations were undertaken to better understand the structure and properties of Si 2 H 3 and Si 2 H For Si 2 H 4, five isomers were studied. These results confirm that disilene is the lowest-energy isomer with a trans-bent C 2h structure with 1 A g symmetry. In addition to the much studied silylsilylene, H 3 SiSiH, a new monobridged isomer H 2 Si H SiH 1 A 1 with a C 1 structure Fig. 1 was found to be a minimum on the potential energy surface with comparable stability; both isomers are predicted to lie about 7 kcal/mol above disilene. On the basis of the close agreement of the measured rotational constants to those calculated, one of the new silicon hydrides was identified as monobridged Si 2 H 4. The purpose of this paper is to give a full account of the laboratory observations, including i tabulations of individual line frequencies for the normal and the more abundant rare isotopic species, ii spectroscopic constants, and iii the evidence for the inversion doubling. Additional transition state calculations at the MP2/cc-pVTZ level of theory have also been done to better understand the observed inversion doubling. Details of an unsuccessful search for the rotational spectrum of H 3 SiSiH are also given. II. EXPERIMENT Monobridged Si 2 H 4 was first detected in a supersonic molecular beam with a Fourier transform microwave FTM spectrometer 13 in a large frequency survey GHz through the discharge products of silane heavily diluted 99.8% in a Ne buffer. Experimental parameters were chosen to optimize the production of the previously reported monobridged Si H SiH: a gas pulse 200 s long yielding a gas flow of cm 3 /min at STP, a discharge potential of 1000 V, and a stagnation pressure behind the pulsed valve of 3.2 atm. The density of unidentified lines is fairly low in this band, with a line roughly every 50 MHz, but near 12.2 GHz a fairly tight and symmetrical doublet stands out Fig. 2. Searches for additional lines at twice and three times the frequency quickly yielded similar doublets of comparable intensity, with a splitting that increases in proportion to the rotational quantum number J. The absence of lines at subharmonic intervals implies that these doublets are not produced by a molecule with more than two silicon atoms. Elemental tests using fully deuterated SiD 4 confirm that the new molecule contains hydrogen; a molecular complex with the Ne buffer gas is ruled out because the lines are produced with comparable intensity when Ar is the buffer gas. A dilute mixture of disilane Si 2 H 6 ; 99.8% in Ne also yielded lines of the new molecule, but the line intensities were only about one-half as strong as those produced with silane. The lack of a detectable Zeeman effect when a permanent magnet is brought near the molecular beam suggests that the carrier has an even number of H atoms: Si 2 H 3 and larger Si 2 H 2n+1 hydrides are open-shell radicals, while Si 2 H 2 and larger Si 2 H 2n hydrides are predicted to have singlet ground states triplet states are calculated to lie at least 10 kcal/mol higher in energy; Ref. 3. On the basis of the experimental rotational constant B eff, the most likely carrier is an isomer of Si 2 H 4 with a nonzero electric dipole moment along its least principal a inertial axis. Silylsilylene, H 3 Si SiH, is an obvious candidate, because it possesses the required dipole moment and if the lone hydrogen atom is freely rotating would possess C 3v symmetry, which in a rotationally cold molecular beam would give rise to a spectrum quite similar to that observed in Fig. 2, owing to closely spaced K = 0, 1 structure. However, this isomer can be rejected for three reasons: i its calculated rotational constant is too low by more than 10% B+C calc /2=5280 MHz, owing to a long Si Si bond of length of 2.40 Å; Ref. 14, ii only a single line would be produced by the fundamental rotational transition, owing to the absence of the K=1 levels for J=0, contrary to what is observed, and iii there is no theoretical or experimental evidence suggesting that the lone hydrogen atom undergoes free rotation i.e., the symmetry of silylsilylene is C s not C 3v. The intensity of the K a =0 lines is so high that it was possible to detect rotational transitions from the K a = ±1 levels of the new silicon hydride. The large derived A rotational constant 73 GHz indicates that the molecule is an asymmetric top very near the prolate limit = ; the symmetry of the molecule is probably low because the K a =±1 lines are quite weak in intensity with respect to the K a =0 lines, i.e., the K a = ±1 levels are not metastable owing to ortho-para nuclear-spin statistics, but rather are thermally populated in our rotationally cold molecular beam. Measurements of intensities of lines with K a = ±1 relative to those with K a =0 are consistent with thermal excitation, yielding a rotational temperature of T rot =1.5±0.5 K, in fairly good agreement with that of other molecules. 7 The large nonzero

3 Monobridged Si 2 H 4 J. Chem. Phys. 124, FIG. 2. Sample FTM spectra of the normal and one of the two 30 Si isotopic species of monobridged Si 2 H 4, showing line doubling from inversion and the characteristic double-peaked instrumental line shape. The instrumental line shape results from the Doppler splitting of the Mach 2 supersonic molecular beam interacting with the two traveling waves that compose the standing wave of the confocal mode of the Fabry-Pérot cavity. Integration times were approximately 1 min for the normal species and 20 min for the 30 Si isotopic species. inertial defect = 3.43 amu Å 2 indicates that the molecule is not planar. On the basis of the theoretical calculations described in Sec. III, searches for the 29 Si, 30 Si, and the fully deuterated isotopic species were undertaken. For each species, at least two rotational transitions in the K a =0 ladder were detected within ±1% of the predicted frequency shifts from the normal species, providing conclusive evidence that the carrier of the lines is H 2 Si H SiH and no other molecule. Even though only a few rotational lines have been observed for several isotopic species, the assignments are extremely secure: i the carrier of each line has been shown by elemental tests to be composed of only Si and H, ii each line also possesses closely spaced doublet structure with nearly the same splitting as that observed for the normal species see Fig. 2, iii line intensities are close to those expected from the natural abundance of Si 4.7% for 29 Si and 3.1% for 30 Si, and iv for the two 29 Si species, the lines also exhibit small additional splitting owing to the interaction of the nuclear spin I=1/2 with the small magnetic field induced by rotation confirmation that the carrier possesses a nonzero nuclear spin. We were unable to detect additional lines for the 29 Si and 30 Si species, because higher-j lines in the K a =0 ladder and those from the K a = ±1 levels are below the present detection sensitivity. The silicon isotopic species were observed in natural abundance, but for D 2 Si D SiD, SiD 4 was used in our discharge instead of SiH 4. Attempts to detect the singly deuterated isotopic species using equal mixtures of SiH 4 and SiD 4 were unsuccessful, presumably because the formation of Si 2 H 4, like that of Si H SiH, 7 is dominated by the reaction of a Si atom or ion with SiH 4. Following the identification of H 2 Si H SiH, an unsuccessful laboratory search was undertaken for silylsilylene H 3 SiSiH, under conditions that optimized the production of the new Si 2 H 4 isomer here. Searches for the 1 0,1 0 0,0 transition were based on the high-level coupled cluster calculations in Sec. III and covered a frequency range that corresponds to ±2% of the calculated rotational constants. III. COMPUTATIONAL DETAILS Self-consistent field SCF restricted closed-shell wave functions have been used for the zeroth-order descriptions of the ground state of the Si 2 H 4 isomers. Correlation effects were included using high-level coupled cluster theory. Coupled cluster with single and double excitations plus per-

4 McCarthy et al. J. Chem. Phys. 124, TABLE I. Calculated relative energies, rotational constants, and dipole moments of the low-lying Si 2 H 4 isomers. Rotational constants MHz e Structure a Rel.energy kcal/mol A B C D H 2 SiSiH 2 C 2h, 1 A g H 2 Si H SiH C 1, 1 A H 3 SiSiH C s, 1 A H-Si H 2 SiH C 2h, 1 A g H-Si H 2 SiH C 2v, 1 A a Properties calculated at the CCSD T /cc-pvtz level of theory. Atoms enclosed in parentheses are bridged atoms. turbative treatment of the triples CCSD T was used in all geometry optimizations, along with correlation-consistent polarized valence double- cc-pvdz and triple- ccpvtz basis sets. 15 The more recently developed coupled cluster method CCSD 2 ; Ref. 16 and the coupled cluster method with single, double, and full triple excitations CCSDT were also employed for single point energy calculations. Scalar relativistic corrections mass-velocity and one-electron Darwin terms have also been calculated by first-order perturbation theory. 12 At the CCSD T level of theory, five different minima have been located on the ground electronic potential-energy hypersurface of Si 2 H 4 in order of relative energy : disilene H 2 Si SiH 2,C 2h, 1 A g, silylsilylene H 3 Si SiH,C s, 1 A, a monobridged isomer H 2 Si H SiH,C 1, 1 A, a trans-like dibridged structure HSi H 2 SiH,C 2h, 1 A g, and a cis-like dibridged isomer HSi H 2 SiH,C 2v, 1 A 1. Geometrical parameters and other spectroscopic constants of the five isomers are reported elsewhere. 12 Relative energies, rotational constants, and dipole moments at the CCSD T /cc-pvtz level of theory are given in Table I. Monobridged Si 2 H 4 is found to be quite stable, lying only 6.64 kcal/mol higher than the disilene at the highest level of theory. At the same level, the much studied silylsilylene isomer is 0.22 kcal/mol less stable than the monobridged form. Inclusion of zero-point vibrational energies, scalar relativistic corrections, and effects of full coupled cluster excitations, however, reverses this energy ordering, with silylsilylene now lying slightly lower in energy by about 0.65 kcal/mol than the monobridged form see Table 8, Ref. 12. It is difficult to determine with certainty which isomer of the two is more stable because the energy difference between the two structures is quite small less than 1 kcal/mol and the energy ordering is sensitive to the size of the basis set and the level of theory. The two dibridged isomers are predicted to be about 14 kcal/mol above the monobridged form, the most polar of the low-lying isomers, with a calculated dipole moment of 1.14 D. At the CCSD T / aug-cc-pvdz level of theory, the components along the three inertial axes are a =0.84 D, b =0.75 D, and c =0.15 D. IV. DATA AND SPECTRAL ANALYSIS The measured rotational lines of monobridged H 2 Si H SiH and its isotopic species are given in Table II, and the derived spectroscopic constants in Table III. For the normal isotopic species, three rotational constants and two TABLE II. Measured rotational transitions of normal and isotopic H 2 Si H SiH in MHz. a J Ka,K c J Ka,K c H 2 Si H SiH H 2 Si H 29 SiH H 2 Si H 30 SiH H 2 29 Si H SiH H 2 30 Si H SiH D 2 Si D SiD 1 0,1 0 0, b b b b ¼ 2 1,2 1 1, ¼ 2 0,2 1 0, ¼ 2 1,1 1 1, ,3 2 1, ¼ 3 0,3 2 0, ¼ 3 1,2 2 1, a The estimated measurement uncertainty is 2 khz for the normal isotopic species and 5 khz for the silicon isotopic species and Si 2 D 4. Spectroscopic constants derived from the centriod of each rotational line. b Centriod of hyperfine-split line.

5 Monobridged Si 2 H 4 J. Chem. Phys. 124, TABLE III. Spectroscopic constants of normal and isotopic H 2 Si H SiH. Isotopic species Rotational Inertial defect constant Experiment a Theory b amu Å 2 H 2 Si H SiH A B C H 2 Si H 29 SiH B C H 2 Si H 30 SiH B C H 29 2 Si H SiH B C H 30 2 Si H SiH B C D 2 Si D SiD B C a Spectroscopic constants in MHz derived using Watson s A-reduced Hamiltonian; 1 uncertainties in parentheses are in the units of the last significant digits. For the normal isotopic species, the derived centrifugal distortion constants in MHz are D J = and D JK = For the five rare isotopic species, D J and D JK were constrained to the values of the normal isotopic species. The rotational constant A was constrained to the value for the normal species for the four Si species, while for D 2 Si D SiD, A was constrained to the theoretical value of MHz. b Calculated at the CCSD T /cc-pvtz level of theory. D J and D JK of the five quartic centrifugal distortion constants reproduce the data to the measurement uncertainty 2 khz. For the four rare Si isotopic species and D 2 Si D SiD, the A rotational constant and two distortion terms were constrained to the values of the normal isotopic species. Nearly all of the stronger b-type and c-type transitions are not accessible with our current spectrometer; these lie well above its present frequency ceiling 42 GHz. The experimental rotational constants of the new silicon hydride here are in extremely good agreement with those predicted for monobridged H 2 Si H SiH. As shown in Table III, the B and C rotational constants of the normal isotopic species are within 0.3% of those predicted at the highest level of theory, and the A rotational constant, which for a highly prolate molecule such as H 2 Si H SiH is sensitive to small deviations in the location of the hydrogen atoms, agrees to within 1.2%. For the four silicon isotopic species and D 2 Si D SiD, the experimental B and C constants also agree with those predicted to better than 0.4%. By comparing line intensities with those of the rare isotopic species of carbonyl sulfide OCS in a supersonic molecular beam of 1% OCS in Ar in the absence of a discharge, we estimate that of order H 2 Si H SiH molecules are produced per gas pulse. This abundance is 30 times less than that of monobridged Si H SiH Ref. 7 and about 100 times less than that of Si 3, 17 both of which are produced under similar experimental conditions, i.e., with silane heavily diluted in Ne. The lower abundance of more saturated silicon hydrides suggests that decomposition of silane in our discharge efficiently produces Si atoms, via sequential loss of H 2. V. BARRIER TO INVERSION A rigid monobridged Si 2 H 4 would produce structureless rotational lines, not closely spaced doublets as observed Fig. TABLE IV. Relative energies and molecular structures of four transition states of Si 2 H 4. Parameter a TS1 TS2 TS3 TS4 r Si Si /Å i b E c a Properties calculated at the MP2/cc-pVTZ level of theory. See Fig. 3 for the geometry of each transition state. b The imaginary vibrational frequency in cm 1. c Relative energy in cm 1 with respect to disilene. 2. Such doubling is characteristic of a molecule interconverting between two equivalent structures. The frequency separation between the two lines is so small of order several tens of kilohertz that it is only resolved in the present experiment because of the high spectral resolution / 10 6 of the FTM technique when applied to a supersonic molecular beam aligned along the axis of the Fabry-Pérot cavity. Because the inversion motion involves hydrogen, deuterium substitution should substantially decrease the inversion frequency and thereby reduce the observed splittings, owing to its two times heavier mass. That is precisely what is observed: at a spectral resolution of 5 khz, no line doubling was observed for any of the rotational lines of the fully deuterated isotopic species, Si 2 D 4. To better understand the observed inversion splittings, transition state calculations have been undertaken at the MP2/cc-pVTZ level of theory with correlation calculated with all electrons. Using efficient convergence methods i.e., the synchronous transit-guided quasi-newton method; Ref. 18, four transition states have been located between the geometrically optimized potential minima. Structures and relative energies for each transition state are given in Table IV; Fig. 3 shows Newman projections of the stable isomers and transition states, along with relative energies. To accurately determine the relative energies of the isomers and verify each transition state, all of the calculated energies have been corrected for zero-point vibrational energy. The simplest inversion motion tunneling of the lone hydrogen atom proceeds through a transition state TS1 with a substantial barrier 16.8 kcal/mol or 5890 cm 1 above the monobridged form, the highest of the four states studied here. Inversion via hindered rotation of the bridged and lone hydrogen atoms has also been considered. This transition state TS2 was previously reported by Sari et al., 12 and is a saddle point linking disilene and the monobridged isomer, but it too has an appreciable barrier height 7.4 kcal/mol or 2585 cm 1. A third pathway for inversion, with both hydrogen migration and internal rotation, may exist. In the first step, the bridged hydrogen migrates to the silicon which is singly bonded to two hydrogen atoms, forming silylsilylene via a transition state TS3 with a low barrier 2.5 kcal/mol or 895 cm 1. In the second step, internal rotation of the SiH 3 group exchanges hydrogen atoms; the silyl group is calculated to undergo nearly free rotation, with a barrier of only 275 cm 1. Via the reverse process, the symmetrically equivalent structure is formed. The potential energy surface is re-

6 McCarthy et al. J. Chem. Phys. 124, FIG. 3. Low-lying potential-energy surface of singlet Si 2 H 4, showing the relative energies in cm 1 of the three most favorable isomers and four transition states calculated at the MP2/ccpVTZ level of theory. Newman projections along the Si Si bond are also shown for each isomer or transition state; a dashed line indicates a bridged hydrogen atom and a short dash-long solid line for TS3 indicates a preference of the bridged hydrogen towards the back Si atom. markably shallow along this coordinate, even though the Si Si bonds of silylsilylene and the monobridged isomer differ by nearly 0.15 Å. The most significant dynamical variation between the two-step mechanism and either simple hydrogen tunneling or hindered rotation is the role of heavy atom motion. The large change in the Si Si bond length for the two-step mechanism implies that the inversion frequency should exhibit some dependence on the masses of the two silicon atoms. The observed splittings for different isotopic species see Table II appear to support this mechanism: substitution at either silicon atom systematically reduces the observed splitting by about 15%, e.g., from 21 khz for the fundamental rotational transition of the normal isotopic species to about 18 khz for the same transition of either the 29 Si or 30 Si isotopic species. VI. DISCUSSION For the proposed inversion mechanism, the observed splittings in the a-type rotational spectrum are expected to be small because these transitions involve no change in the inversion state, yielding only the difference between the inversion splitting of the upper and lower states; the same is true for the b-type spectrum. 19 The weaker c-type transitions, however, change the inversion symmetry, and therefore should exhibit large inversion splittings if the barrier is indeed low. Detection of these transitions, which are predicted to lie in the millimeter-wave band, may be possible either by double resonance, 20 in which the intensity of a strong centimeter-wave line is monitored as a function of the frequency of the millimeter-wave radiation, or by conventional absorption spectroscopy using a long path discharge source. In either case, the search range is probably not prohibitive, i.e., a few gigahertz at 79.5 GHz, the predicted frequency of the fundamental c-type transition 1 1,0 0 0,0. Our failure to detect H 3 SiSiH may result from our limited search range. If this is not the case, the implication is that the product of the dipole moment and the abundance of this molecule, N a, is at least 80 times smaller than that of H 2 Si H SiH. Owing to the fivefold lower dipole moment, this implies that H 3 SiSiH is at least 16 times less abundant than H 2 Si H SiH in our molecular beam. Because the two are calculated to be comparably stable, and because the abundances of isomers approximately correlate with energies, we might expect a priori similar abundances for the two isomers. Our failure to detect silylsilylene may indicate that it rearranges on the time scale of the supersonic expansion, either to the monobridged isomer or to disilene. The low calculated barrier for H 3 SiSiHÛH 2 Si H SiH isomerization combined with the detection here of the monobridged isomer and evidence for inversion doubling in its rotational spectrum suggests that H 2 Si H SiH may be the more stable of the two isomers, and that hydrogen migration from H 3 SiSiH may be facile. Several other theoretical calculations 21,22 also conclude that silylsilylene can rearrange via hydrogen migration with only a small energy barrier 5 kcal/mol, and that similar isomerization occurs for other silicon hydrides which contain a silyl group. For example, -silylsilylene, HSiSiMe 3, rearranges to HMe 2 SiSiMe via a disilenelike intermediate, HMeSi=SiMe If H 3 SiSiH is, in fact, stable on the time scale of the expansion, but its rotational lines are outside the expected search range or are significantly fainter than expected, detection may still be possible. With improvements in instrumentation and production efficiency, a factor of at least 3 in sensitivity may be within reach. It may be possible to determine a precise molecular structure for monobridged Si 2 H 4. Although partially deuterated isotopic species of disilicon hydrides are apparently not efficiently formed in our discharge from mixtures of normal

7 Monobridged Si 2 H 4 J. Chem. Phys. 124, and fully deuterated silane, these species should be detectable using partially deuterated silane. High yields of SiH 3 D, for example, can be synthesized from SiH 3 I with LiAlD 4 and purified via vacuum distillation. 24 The reaction of a Si atom with SiH 3 D should then yield the four singly deuterated species of Si 2 H 4 in comparable abundance. The line intensities of each species should be about one-fourth that of the normal species more than sufficient for detection. By combining these rotational constants with those for the rare Si isotopic species, all of the structural parameters of monobridged Si 2 H 4 five bond lengths and six angles should be determinable. In addition to monobridged Si 2 H 4, strong lines of nearly planar H 2 SiSiH, 12 Si 3, 17 and the more abundant rare isotopic species of Si H SiH Ref. 7 have now been detected in a large spectral search between 10 and 16 GHz through the discharge products of silane heavily diluted in Ne. A number of unidentified lines remain i.e., about one line per 200 MHz for a total of 30 lines, or about 1/3 of the total line density, and chemical and other assays establish that the carriers of many of these lines are silicon hydrides. These remaining lines indicate that still other silicon molecules of chemical and astronomical interest await laboratory detection with the present technique. Candidates include the triplet isomers of Si H SiH, dibridged isomers of Si 2 H 4, and other low-lying isomers of Si 2 H 3, including a monobridged form similar in structure to the isomer detected here, but without the lone hydrogen atom. This isomer is calculated to be of comparable stability to the nearly planar form and to be fairly polar, with a calculated dipole moment of 1.07 D see Table 1 of Ref. 12. H 2 Si H SiH may also be a good candidate for astronomical detection. It is calculated to lie only 7 kcal/ mol above disilene and to possess the largest dipole moment of the three low-lying isomers of Si 2 H 4. Because of the high cosmic abundance of silicon 25 and its only moderate depletion in the interstellar gas, more than ten silicon-bearing molecules have now been detected in galactic sources, 26 including silane and SiS a molecule with two second row elements, so abundant in the molecule-rich stellar envelope that surrounds the evolved carbon star IRC that a number of its rare isotopic species i.e., 29 Si 34 S and Si 36 S; Refs. 27 and 28 are readily detected. There is no evidence for monobridged Si 2 H 4 in recent spectral line surveys in this source, 27,29 but that is not surprising because the strongest lines of similarly sized silicon molecules such as rhomboidal SiC 3 Ref. 30 are in the 5 10 mk range, comparable to the sensitivity achieved in the spectral line surveys. On the basis of the spectroscopic constants in Table III, the radio lines of the most interest to astronomers can be calculated to high precision better than a few km s 1 in equivalent radial velocity up to 150 GHz, adequate for a deep search in the best astronomical sources. ACKNOWLEDGMENTS We thank C. A. Gottlieb, W. Caminati, and W. Klemperer for helpful discussions. This work is supported by NASA Grant No. NAG and NSF Grant Nos. AST , CHE , and CHE S. Park, F. Liao, J. M. Larson S. L. Girshick, and M. R. Zachariah, Plasma Chem. Plasma Process. 24, M. C. McCarthy, C. A. Gottlieb, and P. Thaddeus, Mol. Phys. 101, R. S. Grev and H. F. Schaefer III, J. Chem. Phys. 97, M. Bogey, H. Bolvin, C. Demuynck, and J.-L. Destombes, Phys. Rev. Lett. 66, M. Cordonnier, M. Bogey, C. Demuynck, and J.-L. Destombes, J. Chem. Phys. 97, M. Bogey, H. Bolvin, M. Cordonnier, C. Demuynck, J. L. Destombes, and A. G. Császár, J. Chem. Phys. 100, M. C. McCarthy and P. Thaddeus, J. Mol. Spectrosc. 222, M. C. McCarthy, P. Thaddeus, L. Sari, H. F. Schaefer III, and Z. Hu, Talk RE04, 58th OSU International Symposium on Molecular Spectroscopy, Columbus, OH, G. Olbrich, Chem. Phys. Lett. 130, K. Krogh-Jespersen, Chem. Phys. Lett. 93, G. Trinquier, J. Am. Chem. Soc. 112, L. Sari, M. C. McCarthy, H. F. Schaefer III, and P. Thaddeus, J. Am. Chem. Soc. 125, M. C. McCarthy, W. Chen, M. J. Travers, and P. Thaddeus, Astrophys. J., Suppl. Ser. 129, C. Pak, J. C. Rienstra-Kiracofe, and H. F. Schaefer III, J. Phys. Chem. A 104, Extensible Computational Chemistry Environment Basis Set Database, Version 1/29/01, as developed and distributed by the Molecular Science Computing Facility, Environmental and Molecular Sciences Laboratory which is part of the Pacific Northwest Laboratory, P.O. Box 999, Richland, WA 99352, USA, /forms/ basisform.html 16 S. R. Gwaltney and M. Head-Gordon, Chem. Phys. Lett. 323, M. C. McCarthy and P. Thaddeus, Phys. Rev. Lett. 90, C. Peng and H. B. Schlegel, Isr. J. Chem. 33, ; C.Peng,P.Y. Ayala, H. B. Schlegel, and M. J. Frisch, J. Comput. Chem. 17, R. P. Lattimer and M. D. Harmony, J. Phys. Chem. 53, V. N. Markov, Y. Xu, and W. Jager, Rev. Sci. Instrum. 69, P. Ho, M. E. Coltrin, J. S. Binkley, and C. F. Melius, J. Phys. Chem. 90, B. H. Boo and P. B. Armentrout, J. Am. Chem. Soc. 109, B. H. Boo and P. P. Gaspar, Organometallics 5, K. Ohno, H. Matsuura, Y. Endo, and E. Hirota, J. Mol. Spectrosc. 111, E. B. Jenkins, in Interstellar Processes, edited by D. J. Hollenbach and H. A. Thronson, Jr. Dordrecht, Reidel, 1987, p M. C. McCarthy, C. A. Gottlieb, and P. Thaddeus, Mol. Phys. 101, J. Cernicharo, M. Guélin, and C. Kahane, Astron. Astrophys., Suppl. Ser. 142, R. Mauersberger, U. Ott, C. Henkel, J. Cernicharo, and R. Gallino, Astron. Astrophys. 426, K. Kawaguchi, Y. Kasai, S.-I. Ishikawa, and N. Kaifu, Publ. Astron. Soc. Jpn. 43, A. J. Apponi, M. C. McCarthy, C. A. Gottlieb, and P. Thaddeus, Astrophys. J. 516, L103, 1999.