Dissociative Multiple Ionization Following Valence and Si:2p Core

Select a Journal or Book, Acc. Chem. .... Various monocations of H+, H2+, H3+, CHn+ (n = 0−4), SiCHn+ (n = 0−5), SiC2Hn+ (n = 0−7), and SiC3H9+ ...
1 downloads 0 Views 398KB Size
J. Phys. Chem. 1996, 100, 523-531

523

Dissociative Multiple Ionization Following Valence and Si:2p Core Level Photoexcitation of HSi(CH3)3 in the Range 38-133 eV Bong Hyun Boo,*,† Sang Yeon Lee,‡ Hackjin Kim,| and Inosuke Koyano§ Department of Chemistry, Chungnam National UniVersity, Taejon 305-764, Korea, Center for Molecular Science, 373-1 Kusung-dong Yusung-gu, Taejon 305-701, Korea, and Department of Material Science, Himeji Institute of Technology, 1479-1 Kanaji, Kamigohri, Hyogo 678-12, Japan ReceiVed: July 24, 1995; In Final Form: October 11, 1995X

Dissociation processes of HSi(CH3)3 have been investigated in the valence and Si:2p core-level excitation/ photoionization by photoelectron-photoion coincidence (PEPICO) and photoion-photoion coincidence (PIPICO) techniques together with synchrotron radiation. Various monocations of H+, H2+, H3+, CHn+ (n ) 0-4), SiCHn+ (n ) 0-5), SiC2Hn+ (n ) 0-7), and SiC3H9+ are observed in the whole energy range. Partial ion yield and PIPICO spectra were measured as a function of the incident photon energy in the range 65133 eV. Ab initio calculations are performed to predict the dissociation pathways and their thermochemistry and to provide estimates of the term values of the core-excited states. The variation of the dissociation pattern with the photon energy is discussed in conjunction with the relevant electronic states.

Introduction The interpretation of the photoionization mass spectra of silicon containing compounds provides information on the electronic and charge states of the precursor ions. Also the simultaneous detection of two or more ions formed Via the absorption of the single VUV photon leads to insight into the dissociation pathways and repulsive states as well as identification of the charge states of the parent ions. In recent years, dissociative ionization processes following core-level photoexcitation of silicon compounds with Td symmetry such as SiH4,1 SiF4,2,3 SiCl4,4,5 and SiMe4 (Me ) CH3)4,6 have been a topic of greater interest. HSiMe3 can be considered a model compound containing two kinds of ligands for which much information is available on the site-selective dissociation upon the Si:2p core excitation. The contrast among the dissociation patterns following direct valence shell excitation and core excitation explains the characteristic energy dissipation processes of the core-excited HSiMe3 molecule. The titled compound was studied before by electron impact ionization7 and far ultraviolet8 and Si:2p photoelectron9 spectroscopy. The Si:2p3/2 binding energy has been reported as 106.21 eV.9 The value is close to those of SiH4 and SiMe4, 107.31 and 106.04, respectively,9 reflecting the similar electronegativities of H and C atoms. In the present study, we report the dissociative single and multiple photoionizations of HSiMe3 investigated by the use of synchrotron radiation, photoelectron-photoion coincidence (PEPICO), and photoion-photoion coincidence (PIPICO) methods, and the ab initio results on the dissociation pathways of the singly, doubly, and triply ionized precursors and their thermochemistry. We also report the discrete resonance energies estimated Via the equivalent ionic core virtual orbital model (EICVOM) suggested by Schwarz10 and compare the energies with the peak positions in the photonion mass spectra. By applying this model, the core excitation spectra of SF6 and SiF4 are well understood.10 The Gaussian-2 theory11 (G2 theory) is also applied to estimate the thermochemistry of ionic neutral fragments and to †

Chungnam National University and Center for Molecular Science. Center for Molecular Science. | Chungnam National University. § Himeji Institute of Technology. X Abstract published in AdVance ACS Abstracts, December 15, 1995. ‡

0022-3654/96/20100-0523$12.00/0

elucidate the energetics of the plausible reaction pathways. The G2 theory is known to be a significant improvement over the G1 theory, since it eliminates a number of deficiencies present in the G1 theory.11 Experimental and Theoretical Section The present experiments were carried out by using a timeof-flight (TOF) mass spectrometer, coupled to a constantdeviation grazing incidence monochromator installed at the BL3A2 beam line of the ultraviolet synchrotron orbital radiation (UVSOR) facility in Okazaki. The design and construction of the apparatus have been described in detail elsewhere12,13 and briefly described here. The time-of-flight mass spectrometer employs a drift tube of adjustable length for a facile detection of metastable ions or to obtain better resolved mass spectra. Flight-path lengths of 20 and 30 cm were used in the present study. In addition, the mass detection angle can be varied by rotating the mass spectrometer in the plane perpendicular to the incident photon beam direction. In this study, a quasi-magic angle of 55° between the TOF axis and the polarization direction of the synchrotron beam is used for the isotropic collection of fragment ions. The TOF spectrometer is operated by the timecorrelated ion counting technique in two different modes: a photoelectron-photoion coincidence (PEPICO) mode and a photoion-photoion coincidence (PIPICO) mode. To operate the PEPICO mode, the start pulses for a time-to-amplitude converter (TAC) are provided by the photoelectron signals, sampled from the collision chamber in the opposite direction to the ion flight direction and perpendicular to the incident photon beam, and the stop pulses are generated by the photoion signals. Variation of the total PIPICO intensity with photon energy is obtained by recording PIPICO count rates and photon count rates simultaneously while the photon wavelength is scanned and then by dividing the recorded PIPICO count rates by the recorded photon count rates. When we measure the PIPICO count rates, the coincidence time range (gate width in TAC) was set to be 0-5 µs, since the TOF difference between any pairs of ions formed from the titled molecule falls in this time range. A thin optical filter of aluminum was used in the energy range 38-68 eV for elimination of higher order radiation. The slit © 1996 American Chemical Society

524 J. Phys. Chem., Vol. 100, No. 2, 1996

Boo et al.

TABLE 1: Electronic Configuration of HSiMe3 Determined by the HF/6-31G* Calculation orbital involved

symmetry and no. of electrons occupied

Si:1s C:1s Si:2s Si:2p valence MO

(1a1)2 (1e)4, (2a1)2 (3a1)2 (4a1)2, (2e)4 (5a1)2, (3e)4, (6a1)2, (7a1)2, (4e)4, (5e)4, (8a1)2, (9a1)2, (6e)4

TABLE 2: Molecular Parameters of HSiMe3 Optimized in C3W Point Groupa Si-C Si-H C-H C-Si-C H-C-H a

expb

HF/6-31G*

MP2/6-31G*

1.868 1.489 1.095 110.2 107.9

1.892 1.484 1.087 110.4 107.5

1.883 1.494 1.094 110.2 107.7

Bond length in Å and bond angle in degree. b Reference 15.

width of the monochromator was 500 µm to give an optical resolution of 0.1 and 0.2 nm in the regions above and below 50 eV, respectively. The background pressure of the main chamber was ≈3 × 10-9 Torr. When the trimethylsilane gas was introduced to the ionization chamber, the pressure in the main chamber was maintained at ≈6 × 10-7 Torr. The trimethylsilane gas was purchased from the Tori Chemical Research Institute Ltd. with a purity of 99.9999+ wt/wt % and was used without further purification. The TOF mass spectra confirmed its purity. The molecular geometries are optimized at HF/6-31G* and MP2/6-31G* level of theory by using the Gaussian 92 program suite14 at SERI of KIST in Korea. The term values of the coreexcited state, i.e., the energy difference between the excited state and the corresponding ionization limit, are estimated by the EICVOM calculation.10 The HF/6-31++G** calculations were performed on HPMe3+ at the experimental geometry15 of HSiMe3, and the negative values of the resulting unoccupied orbital energies are taken as the term values of HSiMe3. The peak maxima in the total photoionization efficiency curve are theoretically evaluated by combining the computed term values, the ionization limit for Si:2p3/2, 106.21 eV,9 and the difference between the binding energies of Si:2p3/2 and Si:2p1/2, 0.617 ( 0.005 eV.9 The G2 theory is applied to estimate the energies for various ionic fragmentations of HSiMe3. The theoretical procedures of the G2 theory have been detailed in the literature.11 Results and Discussion In the independent particle description, the ground-state electronic configuration of HSiMe3 can be written with orbital ordering displayed in Table 1, with the aid of our MP2/6-31G* calculation. The first valence ionization energy of HSiMe3, 10.8 eV, has been measured from the He I photoelectron spectrum.8 The molecular parameters of HSiMe3 are also determined from the MP2/6-31G* calculation and are found to be in excellent agreement with the data (Table 2) derived from the microwave spectra reported previously.15 The microwave spectroscopic result indicates that each methyl group is staggered to the Si-H bond and the adjacent Si-C bonds. The MP2/6-31G* calculation is found to be more reliable than the HF/6-31G* calculation. We present in Figure 1 the total photoionization efficiency curve of HSiMe3 in the range 65-133 eV. The overall peak shape is similar to the spectra of other silicon-containing compounds such as SiF4,2,3 SiCl4,4,5 SiMe4,4,16-18 Me3SiOMe,17

Figure 1. Total photoionization efficiency curve of HSiMe3 in the range 65-133 eV.

Figure 2. Total photoionization efficiency curve of HSiMe3 in the range 97-110 eV.

and Me3SiSiMe3.16 Figure 2 shows the total photoionization efficiency curve obtained in the Si:2p excitation region (97110 eV). The peaks are labeled as A to D and A′ to D′. The transition from the Si:2p3/2 component to some orbital is labeled with a letter and that from the Si:2p1/2 component to the same orbital is designated with its prime. Since, as shown in Table 3, the peak shapes and photon energy positions of HSiMe3 are quite similar to those of MeOSiMe3, the peaks in the total ion yield spectra shown in Figure 2 are labeled with the same letter and its prime designated for the Si L- and K-edge X-rayabsorption near-edge spectra of MeOSiMe3.17 The HF calculation using the 6-311++G** basis set was performed on HPMe3+ to estimate the orbital energies of the core-ionized HSiMe3 by using the equivalent core approximation method based on EICVOM.19 When the calculated orbital

Dissociative Ionization of Trimethylsilane

J. Phys. Chem., Vol. 100, No. 2, 1996 525

TABLE 3: Peak Position in the Ion Yield Spectrum of HSiMe3 in the Si:2p Excitation Region energy, eV

a

peak

HSiMe3

MeOSiMe3a

SiMe4a

A A′ B B′ C C′ Si:2p IP D D′

103.64 104.25 104.65 105.26 105.39 106.00 106.21 106.35 106.96

103.52 104.12 104.78 105.34 105.39 105.95 106.3(1) 106.38

102.82 103.48 103.38 103.98 104.00 105.15 105.86 104.74

Reference 17.

TABLE 4: Energies of Various Electronic States of HPMe3+ Determined by the HF/6-311++G** Calculation peak assignment A B C D

(2p3/2 f a1) (2p3/2 f a1) (2p3/2 f e) (2p3/2 f a1) (2p3/2 f e) (2p3/2 f e) (2p3/2 f e) (2p3/2 f a1)

energy, eV 103.85 104.36 104.39 105.14 105.28 105.45 106.01 106.21

peak assignment A′ B′ C′ D′

(2p1/2 f a1) (2p1/2 f a1) (2p1/2 f e) (2p1/2 f a1) (2p1/2 f e) (2p1/2 f e) (2p1/2 f e) (2p1/2 f a1)

energy, eV 104.47 104.98 105.01 105.76 105.90 106.07 106.63 106.83

energies having the similar values are grouped together as exhibited in Table 4, the theoretical energies are quite close to the experimental peak values indicated in Table 3. As shown in Figures 1 and 2, the first peak starts to rise at 101.9 eV and then reaches a maximum at ≈106 eV. The obvious increase of the ion intensity with increasing energy in the range above 100 eV is due to the opening of the Si:2p photoexcitation and ionization channels. Figure 3 shows the variation of the total PIPICO intensity (Itot-PIPICO/Itot-photon) with photon energy. The energy dependence of the total PIPICO intensity is a little different from that of the total photoion intensity. The peak seen in the range 102-111 eV in Figure 3 is due to the double or multiple ionizations Via the multipleresonant and “normal” Auger processes around the Si:2p ionization edge. The peaks observed around 121 eV in Figures 1 and 3 are assigned to shape resonance coupled to the Si:2p continuum above the threshold. In the photoabsorption spectrum of SiF4, this shape resonance is also observed above the Si:2p ionization threshold.10 A typical example of TOF mass spectrum of HSiMe3 taken by excitation at 110 eV in the PEPICO mode is shown in Figure 4. At this energy, various monocations of H+, H2+, H3+, CHn+ (n)0-4), SiCHn+ (n)0-5), SiC2Hn+ (n)0-7), and SiC3H9+ are observed. The formation of these ions is explained Via fragmentation of the singly or multiply charged precursor ions. Unfortunately, some of doubly charged species cannot be separated since, for example, the m/z values of Si2+, SiC22+, SiC2H22+, and SiC2H42+ are the same for those of CH2+, C2H2+, C2H3+, and C2H4+, respectively. It is reported that with substitution of the fluorine atoms for methyl groups bonded to Si in SiMenF4-n (n ) 0-4), progressively more doubly charged ions disappear in the photoion mass spectra of the molecules in the core excitation range.20 In the photoion mass spectra of SiMe2F2, SiMe3F, and SiMe4 in the range 100-800, 102-800, 102-190 eV, respectively, peaks due to F2+, Si2+, SiF2+, and SiF22+ are not observed. However, these ions are observed in the spectra of SiF4. Among the doubly charged species, only the SiF22+ ion is observed in the photoion mass spectrum of SiMeF3. Similar spectral behaviors

Figure 3. Total photoion-photoion coincidence efficiency curve of HSiMe3 in the range 65-133 eV.

Figure 4. PEPICO (mass) spectrum of HSiMe3 measured at a photon energy of 110 eV. Flight path length is 30.0 cm. TOF axis is 55° with respect to the direction of photon beam polarization. (a) Whole mass spectrum; (b) scale-expanded mass spectra.

may be observed in the photoionization of HSiMe3 in view of the structural similarity of HSiMe3 to FSiMe3 and SiMe4. It is shown in the literature that the degree of the multiple ionization of silicon compounds is dependent on the electron

526 J. Phys. Chem., Vol. 100, No. 2, 1996

Boo et al.

Figure 5. PEPICO (mass) spectra of HSiMe3 taken by excitations at photon energies of 70, 100, and 115 eV.

donating properties of the ligands surrounding the silicon center.2,3 Fluorine or chlorine lone pairs with the symmetries of e and t1 do not contribute directly to the silicon spectrum. But the presence of a Si:2p hole increases the valence electron population in the a1 and t2 orbitals Via the back-donation of the fluorine or chlorine lone pair electrons to Si:2p. On the basis of the Mulliken population analysis, the total electron population on the silicon atom increases.21 This effect can increase the rate for the Auger processes giving rise to the double and/or triple vacancies.22 It is shown in the literature that triple ionization of SiF4 is found to be relatively efficient.2,3 In our recent study of the dissociative ionization of SiCl4 in the vicinity of the Si:2p edge, doubly charged species such as Si2+ and Cl2+ are observed, but the efficiencies of the formation of the doubly charged species are very low.5 In the case of HSiMe3, the valence electron population around the silicon center is quite low due probably to the absence of the lone pair electron in the CH3 group and the H atom. Therefore, the Auger rate in HSiMe3 may be low in comparison with those of SiF4 and SiCl4. It is of higher interest to clarify why the Si:2p photoexcitation/ ionization will lead to a doubly and multiply ionized species. Herein emphasis is given on the double ionization following the Si:2p core level excitation, process 1.

HSiMe3* f HSiMe32+ + 2e

(1)

Double-resonant Auger processes and two-step autoionization could explain the formation of the doubly charged species below the Si:2p ionization edge.23 Hayaishi et al. argued that twostep autoionization seems more likely than shake-off as the reason for the formation of Ar2+ ions below the Ar:2p ionization potential.24

Ar 2p-1nd f Ar+ 3p-2(1D)n′d + e1

(2)

Ar+ 3p-2(1D)n′d f Ar+ 3p-2 + e2

(3)

Morin et al. proposed that the double-resonant Auger process is more important in the photoionization of Si(CH3)4 than the direct double ionization below the Si:2p edge.6 Souza et al. proposed in photoelectron and Auger spectroscopy of SiH4 that a very fast dissociation pathway involving process (4) occurs in addition to autoionization and resonant Auger processes in the vicinity of the Si:2p edge.1

Figure 6. Partial photoion yield spectra of HSiMe3 in the range 65133 eV. The spectral intensities (Iphotoion/Itot-photon) indicated in (a)-(d) are presented on the same relative intensity scale.

SiH4* f SiH3* + H f SiH3+* + e (or SiH3+ + e) (4) On the basis of the argument by Morin et al.6 and the structural similarity of HSiMe3 to SiMe4, the double ionization of HSiMe3 below the Si:2p ionization limit is believed to occur Via the double-resonant Auger processes. The variation of the partial ion yields with energy is discussed in conjunction with the relevant electronic states. H+ Channel. We show in Figure 5 the photoionization mass spectra taken at various photon energies. Fragment ions such as H+ and CHn+ (n ) 0-3) are more abundant at 115 eV, the energy for the Si:2p photoexcitation and/or ionization. We present in Figure 6 the energy dependence of the individual ion yields (Iphotoion/Itot-photon). Notice that, at a given energy, the

Dissociative Ionization of Trimethylsilane value for Iphotoion/Itot-photon is derived by multiplying the value for Itot-photoion/Itot-photon by the value for the ion branching ratio displayed in Figure 7 and that the sum of the intensities for all the ions indicated in Figure 6 equals the total photoion intensities displayed in Figure 1. We show in Figure 6 the individual ion yield spectra only in the range 65-133 eV due to the limited photon energy range covered by one grating of the constant deviation monochromator at the BL3A2 beam line. The H+ ion intensity starts to rise from about 100 eV and then reaches a maximum at 105 eV. The energy dependence of the ion intensity is almost the same as observed in that of the total ion intensity. Around this energy, the intensity of the ion pairs H+-CHn+ (n)0-3) and H+-SiCnHp+ (n)0-2, p)0-3) is also greatly enhanced. These features are accounted for by dissociative double and/or triple ionizations following the Si:2p core excitation/ionization as shown in Scheme 1.

SCHEME 1 HSiMe3 + hν f HSiMe3* (Si:2p excitation to antibonding or Rydberg orbitals) HSiMe3*f HSiMe32+ + 2e f HSiMe33+ + 3e HSiMe32+ f H+ + SiCnHp+ (n)0-2, p)0-3) + np +

+

HSiMe3 f H + CHn (n)0-3) + 3+

SiCmHp+ (m)0-2, p)0-3) + np +

f H + SiCnHp

2+

(n)0-2, p)0-3) + np

where np denotes neutral products. Direct evidence of the triple and/or multiple ionizations could not be presented since some uncertainties exist on the charge states of some fragment ions and on the ionic fragmentation patterns. However, the formation of the ion pairs H+-CHn+ (n)0-3) implies at least the occurrence of triple ionization since the electronegativities of H and C atoms are larger than that of Si, and thus ionic silicon species are most likey formed along with the ion pairs H+CHn+ (n)0-3) as the undetected ionic fragments. Actually in our PIPICO data, the ion pairs H+-SiCmHp+ (m)0-2, p)03) and CHn+ (n)0-3)-SiCmHp+ (m)0-2, p)0-3) are observed in Figure 8. Therefore, the detection of the ion pair H+CHn+ (n)0-3) would support the possibility of triple ionization of HSiMe3 upon VUV absorption. We exhibit in Figure 8 a typical PIPICO spectrum taken at 110 eV. Here we observe several dissociation channels involving the H+ ion.

HSiMe3n+ (n)2,3) f H+ + SiHn+ (n)0-3) + other product (5) HSiMe3n+ (n)2,3) f +

+

H + SiCHn (n)0-3) + other product (6) n+

HSiMe3

(n)2,3) f H+ + SiC2Hn+ (n)0-3) + other product (7)

But these channels are not observed only in the Si:2p excitation region. As seen in Figure 9, below the Si:2p edge, processes (5)-(7) also occur.

J. Phys. Chem., Vol. 100, No. 2, 1996 527 In the partial PIPICO spectra displayed as a function of photon energy (Figure 10), we also observe distinct features in the range 100-106.21 eV (the Si:2p edge) corresponding to reactions 5-7. Thus this energy could correspond to the Si:2p core excitation below the threshold. H2+ Channel. The energy dependence of the ion intensity for the H2+ ion is quite similar to that of H+ as seen in Figure 6a. In this study, its ionic counterpart in the dissociation of the precursor HSiMe32+ ion cannot be identified due to the low PIPICO intensity. H3+ Channel. This ion is observed, but its intensity is quite low. Thus the variation of the ion intensity with energy was not measured. CHn+ (n)0-4) Channel. The ion intensity of CH3+ begins to increase after ≈100 eV and reaches a maximum at ≈107 eV (Figure 6b). Similar features are also observed in the ion intensities for CHn+ (n)0-2) in the Si:2p excitation region. These observations above 100 eV are presumably due to the ion pair formation according to process (8).

HSiMe3n+ (n)2,3) f CHn+ (n ) 0-2) + H+ + other product (8) The channel leading to the formation of the ion pair is clearly observed in Figure 8a,b. In the dissociative multiple ionization following Si:2p excitation, the major contributions to the formation of CH3+ are from processes (9)-(11).

HSiMe3n+ (n)2-3) f CH3+ + SiCH3+ + other product (9) HSiMe3n+ (n)2,3) f CH3+ + SiC2H6+ + other product (10) HSiMe3n+ (n)2,3) f CH3+ + SiHn+ (n ) 0, 1, 3) + other product (11) It is quite interesting that as seen in Figure 6a,b, the photon energy position for the maximum of the ion intensity for H+ is ≈105 eV, being a little lower than that for CH3+, ≈107 eV. The photon energy positions for the maximum ion intensity of CHn+ (n ) 0-2) are observed to be ≈105 eV, the value being the same for that of H+. One can easily see one shoulder ≈107 eV in the ion yield spectrum for H+ (Figure 6a). This feature could explain that process (6) also contributes the formation of H+ ion in the Si:2p excitation region. It is noticed that process (6) occurs abundantly at ≈107 eV (Figure 10). It is shown that a double-resonant Auger process has also been observed below (105 eV) and a “normal” Auger process occurs above (106.8 eV) edge in SiMe4.6 These two energies in SiMe4 are quite close to the photon energy positions for the maxima of the total photoionization efficiency in HSiMe3 (Figure 2). It is also reported that at a higher energy of 106.8 eV, the PIPICO counts corresponding to process 12 are larger than those observed at a lower energy of 105 eV.6

SiMe42+ f CH3+ + SiMen+ (n ) 0-3) + np

(12)

Here, the most probable case is n ) 2, 3. Similar behavior is also observed in the PIPICO spectrum of HSiMe3 (Figure 3). The PIPICO counts at 105 eV (below the edge) are much less than those at 107 eV. The decay channel of the resonance below the edge leading to a single ionization and to an ion in an excited configuration would explain the low PIPICO counts at 105 eV (Figure 3). Morin et al. suggest that the single hole state in SiMe4 (the inner valence, 4a1 hole in Td symmetry of

528 J. Phys. Chem., Vol. 100, No. 2, 1996

Boo et al.

Figure 7. Ratios of integrated intensities of ion peaks in TOF mass spectrum to total photoion intensity (Iphotoion/Itot-photoion) in HSiMe3 as a function of photon energy.

SiMe4) decays also into double (or multiple) ionization continua.6 Around 107 eV, the “normal” Auger processes are presumed to prevail to give rise to the high ion-pair yield above the edge. The energy dependence of the ion yield and PIPICO intensity implies that the “normal” Auger process seems to weaken the Si-C bond relatively, whereas the resonant Auger process seems to weaken the Si-H bond. As shown in Figure 10, 107 eV represents the energy that can produce ion pairs such as CH3+-SiHn+ (n ) 0-3) and CH3+-SiCH3+ in the highest yields, while 105 eV refers to the energy which can form ion pairs such as H+-SiHn+ (n ) 0, 1) and H+-SiC2Hn+ (n ) 0-3) abundantly. But the reason for these discrete fragmentation patterns is not elucidated. We also observed a small quantity of CH4+ in the whole energy range examined. The formation of the ion is theoretically evaluated and discussed below. C2Hn+ (n)1-3) or SiC2Hn2+ (n)0, 2) Channels. The mass charge ratios for C2Hn+ (n)2, 3) are the same for SiC2Hn2+ (n)0, 2), respectively. The peak shapes of SiC2Hn+ (n ) 0-3) seen in Figure 4 are very similar to those of H+-SiC2Hn+ (n ) 0-3) (Figure 8a,c). This indicates that the SiC2Hn+ ions may be formed mostly Via process (13).

SiC2Hn2+ (n)1-4) f H+ + SiC2Hn-1+ + np

(13)

The most probable case is n ) 2. Thus the doubly charged SiC2Hn2+ (n ) 0, 2) ions may contribute the observed mass charge ratios (m/z) of 26 and 27, respectively. On the other hand, if the C2Hn+ species is responsible for the ionic species, how is it possible to form the C-C bond in the dissociation of the charged precursor? Groenewold and Gross observed that Me3Si+ yields predominat extrusion of C2H4 upon collisionactivated dissociation (CAD) (process (14)).25

Me3Si+ f Si(CH3)H2+ + C2H4

(14)

Figure 8. Photoion-photoion coincidence (PIPICO) spectra of HSiMe3, (b) and (c) scale-expanded mass spectra. The spectrum was measured under the same condition for the mass spectra shown in Figure 4.

Bakhtiar et al. indicate in the recent Fourier transform mass spectrometric study that unimolecular dissociation of Me3Si+ leads to ethene extrusion concomitant with the formation of H2SiCH3+ through a reaction intermediate such as (CH3CH2)(CH3)SiH+.26,27

[Me3Si+]* a (CH3CH2)(CH3)SiH+ f H2SiCH3+ + C2H4 (15) These workers also report that isomerization of (CH3CH2)SiH2+ to the thermodynamically more stable Me2SiH+, a process similar to reaction 15, efficiently occurs by collision-activation in the gas phase.26

Dissociative Ionization of Trimethylsilane

J. Phys. Chem., Vol. 100, No. 2, 1996 529

[Me3Si3+]* f [(CH3CH2)(CH3)SiH3+]* f H+ + C2Hn+ (n)1-3) + HSiCH3+ + np (17) The ion pair H+-HSiCH3+ + H -H2SiCH3+ in Figure 8c. SiHn+ (n)0-3) Channels.

Figure 9. Photoion-photoion coincidence (PIPICO) spectra of HSiMe3 measured at various photon energies.

is clearly observed along with

These ions are formed in the whole energy range examined. The variation of the ion intensity with photon energy exhibits a small feature in the core excitation region (Figure 6d). This indicates that the excess internal energy due to core excitation is deposited in the silicon center, giving rise to nearly exclusive cleavage of the Si-C bonds and Si-H bond formation for n ) 2 and 3. The Si+ and SiH+ ions are observed in higher yields than the SiH2+ and SiH3+ ions. The low yield of SiH2+ may be accounted for by the relatively weak bond energy, D0(SiH+H) ) 207.1 ( 9.2 kJ/mol, as compared with the bond energies D0(Si+-H) ) 316.7 ( 7.2 kJ/mol and D0(SiH2+-H) ) 381.2 ( 11.0 kJ/mol, values derived from literature thermochemical data.28 The observation of SiH3+ indicates the occurrence of hydrogen shift from carbon to silicon in H3SiCH2+ prior to the ionic fragmentation as shown in reaction 18 and/or of direct 1,1-elimination from H2SiCH3+ giving rise to H3Si+ + CH2.

H2SiCH3+ f H3SiCH2+ f H3Si+ + CH2

(18)

It is noticed that the H2SiCH3+ ion is actually observed as seen in Figure 4b since the ion intensity for 30SiCH3+ contributes only 23% of the total intensity for the TOF of 3.23 µs. One can invoke that isomerization occurs, converting H2SiCH32+ to H3SiCH22+, which then dissociates into H3Si+ and CH2+. But our MP2 calculation indicates that only weakly bound, van der Waals complexes are found as the local minimum in the potential energy surfaces of the electronical ground-state [H5,C,Si]2+. This theoretical consideration may not be applied to the potential energy surface of the electronically excited [H5,C,Si]2+ species. In the Si:2p excitation region, the major contribution to the formation of SiH3+ comes from process (19) as shown in Figure 8a,b.

HSiMe3n+ (n) 2, 3) f CH3+ + SiH3+ + other product (19) Figure 10. Partial PIPICO yield spectra of HSiMe3 in the range 65133 eV. The spectral intensities (IPIPICO/Itot-photon) are presented on the same relative intensity scale.

Process (15) is believed to occur as the minor process along with process (16) in the photoionization of HSiMe3, the argument based on the detection of H2SiCH3+ and C2Hn+ (n)1-3) and on the mass spectrometric studies.25-27 +

The intensity of the PIPICO channel for H+-SiH3+ (Figure 8a,c) is quite small, indicating that the formation of SiH3+ arises largely from the PIPICO channel of CH3+-SiH3+ (Figure 8b). As shown in Figure 8, the major processes leading to the detection of the Si+ and SiH+ ions are reactions 20 and 21,

HSiMe3n+ (n)2, 3) f CH3+ + Si+ + other products f H+ + Si+ + other products

(20) (21)

+

[Me3Si ]* a [(CH3CH2)(CH3)SiH ]* f H2SiCH3 + C2H4+ (16) Notice that Me3Si+ could be formed in the highly excited state to sufficiently overcome the barrier to the isomerization of Me3Si+ to (CH3CH2)(CH3)SiH+ and to the 1,2-elimination to yield H2SiCH3 + C2H4+ (process (16)). The observation of the ion pairs H+-C2Hn+ (n)1-3) in the PIPICO channels indicates that multiply charged silicon ions such as Me3Si3+ may undergo similar rearrangement to yield ion pairs, for example, as shown in process (17).

and reactions 22 and 23, respectively.

HSiMe3n+ (n)2, 3) f CH3+ + SiH+ + other products (22) (23) f H+ + SiH+ + other products SiCHn+ (n)0-5) Channel. Among the ions, SiCH3+ is formed in a relatively higher yield in the core excitation region (Figure 4b). SiCH4+ and SiCH5+ are also observed, although their yields are relatively small. A recent higher level ab initio calculation indicates that H2SiCH3+ is the only minimum energy structure on the [H5,C,Si]+ potential energy surface and H3SiCH2+ is the saddle point29 and is higher in energy by about

530 J. Phys. Chem., Vol. 100, No. 2, 1996

Boo et al.

SCHEME 2 +

HSiMe3 + hν f HSiMe3 + e (valence ionization) HSiMe3+ f HSi(CH3)2+ + CH3 40 kcal/mol than H2SiCH3+.30 This experimental and theoretical consideration indicates that hydrogen migration from carbon to silicon occurs as discussed above and previously.26 As discussed above, H2SiCH3+ represents the [H5,C,Si]+ species and is formed Via the extrusion of ethene as shown in processes (16) and (17). SiC2Hn+ (n)0-7) Channel. At 38 eV, the lowest energy examined, the major product is SiC2H7+, formed probably Via Scheme 2. At low energies, below about 45 eV, the most predominant process in the dissociative single photoionization of HSiMe3 is the Si-C bond rupture forming HSi(CH3)2+. As the energy is increased, the efficiencies for the formation of SiC2Hn+ (n)07) are reduced due to competition with the other dissociation processes. The ab initio calculation based on the G2 theory indicates that the two lowest energy pathways of HSiMe3 upon the single ionization involve processes (24) and (25).

HSiMe3 f H + SiMe3+ + e - 10.33 eV

(24)

HSiMe3 f CH3 + HSiMe2+ + e - 10.68 eV

(25)

HSiMe32+ f H+ + SiC3H9+ - 23.94 eV

(31)

where the endothermicity of E0 ) 23.94 eV is derived by the G2 calculation. Instead we observe process (30) as the major dissociation channel. The G2 calculation shows that process (32) is energetically more favorable than process (31).

HSiMe32+ f CH3+ + SiC2H7+ - 20.09 eV

(32)

Interestingly, the energy dependence of the partial ion yield of SiC3H9+ is almost invariant as shown in Figure 6d. This implies that the ion intensity for SiC3H9+ measured in our photoion mass spectra mostly comes from the nonresonant valence ionization of HSiMe3, since the core excitation would give rise to the Coulomb explosion leading to the dissociation of SiC3H9+ into small fragment ions due to the excess internal energy upon photoabsorption. Conclusions

The MP2/6-31G* calculations were performed to predict the dissociation pathways of several precursor ions such as HSiMe3+, HSiMe32+, and HSiMe33+. The optimized geometry of the ground-state [HSiMe3]+ is found to be a weakly bound, van der Waals complex, [HMe2Si‚‚‚Me]+. Interestingly, the MP2/6-31G* calculation predicts that the optimized structure of the HSiMe32+ ion in the ground state is sensitive to its spin state involving two different structures as shown in reactions 26 and 27.

HSiMe3 f [Me2Si‚‚‚CH4]2+ (singlet) + 2e

(26)

HSiMe3 f [CH3‚‚‚MeSiH‚‚‚CH3]2+ (triplet) + 2e

(27)

Process (26) would explain the formation of CH4+, although the formation of CH4+ in the photoionization of HSiMe3 may not correctly reflect the theoretical pathways, since the doubly charged precursor ions might be formed in electronically excited states upon VUV photoabsorption. But the dissociation efficiency is not high as represented in the partial ion yield spectra (Figure 6c). The calculation also indicates that triple ejection of electrons from HSiMe3 allows the molecule to be an unbound, quartet state as shown in process (28).

HSiMe3 f [SiH‚‚‚3CH3]3+ + 3e

This explanation also supports from the peak shape analysis of H+-SiC2Hn+ (n)0-3) discussed above in process (13). SiC3H9+ Channel. Among the possible product ions containing three carbon atoms, this ion is the only one formed. In the whole energy range examined, only one possible pathway for the ion pair formation, process (31), is not observed.

(28)

The PIPICO data shown in Figure 8a,b indicate that SiC2Hn+ (n)0-7) ions are formed mostly by process (29) or (30).

HSiMe3n+ (n)2, 3) f H+ + SiC2Hn+ (n)0-3) + other products (29) HSiMe3n+ (n)2, 3) f CH3+ + SiC2Hn+ (n)5-7) + other products (30)

The present mass spectrometric studies led to the observation of various fragmentation patterns of singly and doubly charged HSiMe3 as a function of the photon energy. At discrete resonance energies below the ionization edge, the photoionization efficiency is greatly enhanced by the core-hole decay process. The discrete resonance energies of HSiMe3 below the Si:2p ionization edge were esimated by using the equivalent core approximation method by performing the HF calculation on HPMe3+. The PIPICO efficiency measurements show that Si:2p core excitation is the key process in the multiple ionization processes. Our ab initio calculation based on the MP2 and MP4 perturbation theories and the G2 thoery predicts the structures and energies and also provides plausible dissociation patways of the charged precursor. Acknowledgment. This work was partially supported by the Pohang Light Source (PLS) (1992) in Pohang, Korea, through a visiting scientist research program, and by the Center for Molecular Science (CMS) in Taejon, Korea, and from a grant by the Basic Science Research Institute Program, 1995-96, Project No. BSRI-95-3432, Ministry of Education, Korea. S.Y.L. is grateful to C.M.S. for a postdoctor fellowship. B.H.B. thanks Dr. Masuoka for his help in conducting the experiment. We are grateful to one of the referees for his helpful comments on the interpretation of the results. The members of the UVSOR facility at the Institute for Molecular Science in Okazaki, Japan are greatly acknowledged for their valuable help. References and Notes (1) de Souza, G. G. B.; Morin, P.; Nenner, I. Phys. ReV. A 1986, 34, 4770. (2) Imamura, T.; Brion, C. E.; Koyano, I.; Ibuki, T.; Masuoka, T. J. Chem. Phys. 1991, 94, 4936. (3) Lablanquie, P.; Souza, A. C. A.; de Souza, G. G. B.; Morin, P.; Nenner, I. J. Chem. Phys. 1989, 90, 7078. (4) Nagaoka, S.; Ohshita. J. Ishikawa, M.; Masuoka, T.; Koyano, I. J. Phys. Chem. 1993, 97, 1488. (5) Boo, B. H.; Park, S. M.; Koyano, I. J. Phys. Chem. 1995, 99, 13362. (6) Morin, P.; de Souza, G. G. B.; Nenner, I.; Lablanquie, P. Phys. ReV. Lett. 1986, 56, 131. (7) Potzinger, P.; Ritter, A.; Krause, J. Z. Naturforsch. 1975, 30a, 347.

Dissociative Ionization of Trimethylsilane (8) Roberge, R.; Sandorfy, C.; Matthews, J. I.; Strausz, O. P. J. Chem. Phys. 1978, 69, 5105. (9) Bozek, J. D.; Bancroft, G. M.; Tan, K. H. Phys. ReV. A 1991, 43, 3597. (10) Friedrich, H.; Pittel, B.; Rabe, P.; Schwarz, W. H. E.; Sonntag, B. J. Phys. B: At. Mol. Opt. Phys. 1980, 13, 25. (11) Curtiss, L. A.; Raghavachari, K.; Trucks, G. W.; Pople, J. A. J. Chem. Phys. 1991, 94, 7221. (12) Masuoka, T; Horigome, T.; Koyano, I. ReV. Sci. Instrum. 1989, 60, 2179. (13) Ishiguro, E.; Suzuki, M.; Yamazaki, M.; Nakamura, E.; Sakai, K.; Matsudo, O.; Mitsutani, N.; Fukui, K.; Watanabe, M. ReV. Sci. Instrum. 1989, 60, 2105. (14) Gaussian 92, Revision D.2, Frisch, M. J.; Trucks, G. W.; HeadGordon, M.; Gill, P. M. W.; Wong, M. W.; Foresman, J. B.; Johnson, B. G.; Schlegel, H. B.; Robb, M. A.; Replogle, E. S.; Gomperts, R.; Andres, J. L.; Raghavachari, K.; Binkley, J. S.; Gonzalez, C.; Martin, R. L.; Fox, D. J.; Defrees, D. J.; Baker, J.; Stewart, J. J. P.; Pople, J. A. Gaussian, Inc., Pittsburgh, PA, 1992. (15) Pierce, L.; Petersen, D. H. J. Chem. Phys. 1960, 33, 907. (16) Simon, M.; Lebrun, T.; Martins, R.; de Souza, G. G. B.; Nenner, I.; Lavollee, M.; Morin, P. J. Phys. Chem. 1993, 97, 5228. (17) Sutherland, D. G. J.; Kasrai, M.; Bancroft, G. M.; Liu, Z. F.; Tan, K. H. Phys. ReV. B 1993, 48, 14989.

J. Phys. Chem., Vol. 100, No. 2, 1996 531 (18) Sodhi, R. N. S.; Daviel, S.; Brion, C. E.; de Souza, G. G. B. J. Electron. Spectrosc. Relat. Phenom. 1985, 35, 45. (19) Adams, P. B. J. Electron Spectrosc. Relat. Phenom. 1994, 69, 117. (20) Bozek, J. D.; Tan, K. H.; Bancroft, G. M. Chem. Phys. 1991, 158, 171. (21) Larkins, F. P.; McColl, J.; Chelkowska E. Z. J. Electron. Spectrosc. Relat. Phenom. 1994, 67, 275. (22) McColl, J.; Larkins, F. P. Chem. Phys. Lett. 1992, 196, 343. (23) de Souza, G. G. B.; Morin, P.; Nenner, I. J. Chem. Phys. 1989, 90, 7071. (24) Hayaishi, T.; Murakami, E.; Yagishita, A.; Foike, F.; Morioka, Y.; Hansen J. E. J. Phys. B: At. Mol. Opt. Phys. 1988, 21, 3203. (25) Groenewold, G. S.; Gross, M. L. J. Organomet. Chem. 1982, 235, 165. (26) Bakhtiar, R.; Holznagel, C. M.; Jacobson, D. B. Organometallics 1993, 12, 621. (27) Bakhtiar, R.; Holznagel, C. M.; Jacobson, D. B. Organometallics 1993, 12, 880. (28) Boo, B. H.; Armentrout, P. B. J. Am. Chem. Soc. 1987, 109, 3549. (29) Gordon, M. S.; Pederson, L. A.; Bakhtiar, R.; Jacobson, D. B. J. Phys. Chem. 1995, 99, 148. (30) Hopkinson, A. C.; Lien, M. H. J. Org. Chem. 1981, 46, 998.

JP952088M