Electronic transitions of jet-cooled SiC2, Si2Cn (n=1-3), Si3Cn (n=1,2), and SiC6H4 between 250 and 710 nm
aa r X i v : . [ phy s i c s . a t m - c l u s ] M a r Electronic transitions of jet-cooled SiC , Si C n ( n = 1 − ), Si C n ( n = 1 , ), and SiC H between 250 and 710 nm published in Astrophys. J. 801 (2015) 119 M. Steglich, J. P. Maier
Department of Chemistry, University of Basel, Klingelbergstrasse 80, CH-4056 Basel, Switzerland [email protected], [email protected]
ABSTRACT
Electronic transitions of the title molecules were measured between 250 and 710 nm using amass-resolved 1+1’ resonant two-photon ionization technique at a resolution of 0.1 nm. Calcula-tions at the B3LYP/aug-cc-pVQZ level of theory support the analyses. Because of their spectralproperties, SiC , linear Si C , Si C, and SiC H are interesting target species for astronomicalsearches in the visible spectral region. Of special relevance is the Si–C –Si chain, which featuresa prominent band at 516.4 nm of a strong transition ( f = 0 . H at 445.3 nm were also investigated at higher resolution (0.002 nm). Subject headings: dust, extinction — ISM: lines and bands — ISM: molecules — methods: laboratory— molecular data
1. Introduction
Silicon and carbon are among the most abun-dant refractory elements in the universe. Theyplay an important role in astrochemical pro-cesses and are present in more than 70% of allinter- and circumstellar molecules. Already inthe 1920s, Merrill and Sanford observed strongabsorption bands in the 420–550 nm wavelengthrange of cool carbon stars (Merrill 1926; Sanford1926; Sarre et al. 2000), which were later as-cribed to an Si–C–C molecule by Kleman (1956).It took almost three decades before the gasphase study of Michalopoulos et al. (1984) re-vealed the correct triangular geometry of thismolecule. More recently, the electronic emission(Lloyd Evans et al. 2000) and rotational spectra(e.g. Thaddeus et al. 1984; Gottlieb et al. 1989;M¨uller et al. 2012) were detected in circumstellarregions. The Merrill-Sanford bands have evenbeen found outside our galaxy (Morgan et al.2004). Other molecules containing both siliconand carbon were identified by radio astronomyin circumstellar shells (McCarthy et al. 2003). These are SiC (Cernicharo et al. 1989), linearSiNC and SiCN (Gu´elin et al. 2000, 2004), rhom-boidal SiC (Apponi et al. 1999b), and linear SiC (Ohishi et al. 1989). Spectroscopic detections inthe laboratory preceded the astronomical observa-tions (Apponi et al. 1999a; McCarthy et al. 1999,2001; Kokkin et al. 2011).Several silicon-carbon clusters have been thesubject of further laboratory investigations: vi-brational modes of SiCH x ( x = 3 − H,SiC n ( n = 2 , , , , C n ( n = 1 − C n ( n = 1 ,
2) could be identified by IR ma-trix isolation spectroscopy (Kafafi et al. 1983;Shepherd & Graham 1985; PresillaM´arquez et al.1990, 1995, 1996, 1997; PresillaM´arquez & Graham1991, 1992, 1994; Withey & Graham 1992; Han et al.1997; Ding et al. 1999, 2000; Kaiser & Osamura2005; Lˆe et al. 2014a,b). Gas phase IR spec-tra of SiC , Si C , and Si n C m ( n + m = 6)have been obtained by Van Orden et al. (1994,1995), Thorwirth et al. (2011), and Savoca et al.(2013). Cations of SiC n ( n = 3 −
69) and an-ions of Si n C m (1 ≤ n ≤
7, 1 ≤ m ≤ n C m and SiC H , with bond lengths in ˚A.ity experiments (Fye & Jarrold 1997) and pho-toelectron spectroscopy (Nakajima et al. 1995;Duan et al. 2002), respectively. Photodissocia-tion studies were performed on Si n C m cations( n = 3 − m = 1 ,
2; Ticknor & Duncan2005) and ionization energies were determined forSiC H x ( x = 0 −
2) (Kaiser et al. 2012). Elec-tronic absorption spectra were measured mainlyfor small molecules in the past, such as SiC, SiCH,and SiC2 (Bernath et al. 1988; Brazier et al. 1989;Ebben et al. 1991; Butenhoff & Rohlfing 1991b;Grutter et al. 1997; Smith et al. 2000; Kleman1956; Weltner & McLeod 1964; Bondybey 1982;Michalopoulos et al. 1984; Butenhoff & Rohlfing1991a). Studies on larger systems were real-ized only recently. Stanton et al. (2005) ana-lyzed the e C B ← e X A transition of Si C, andKokkin et al. (2014) investigated the optical spec-tra of silicon-terminated carbon chains, SiC n H( n = 3 − , Si C n ( n = 1 − C n ( n = 1 , H be-tween 250 and 710 nm.
2. Methods
Optical absorption spectra were measured in-directly by applying a resonant two-color two-photon ionization scheme (R2C2PI) and moni-toring the ions (as a function of scanning laserwavelength) with a linear time-of-flight mass spec-trometer (TOF-MS). Mixed silicon-carbon clus-ters were created by laser ablation (532 nm, 5ns, ≈ ) of a crystalline Si rod in anHe atmosphere containing 2% acetylene (C H ).In later experiments, SiC H was also producedby discharging a gas mixture of He and phenyl-silane (SiC H ), the latter being vaporized in abubbler. The molecular beam source contains apulsed valve (5 bar backing pressure) creating asupersonic jet, which was skimmed 40 mm down-stream to generate a collimated beam. A +300V potential was applied to the skimmer to re-move ions before neutrals entered the ionizationregion of the TOF-MS. The spectral scans wererealized by counter-propagating the radiation ofthe laser into the molecular beam. Broad rangescans between 250 and 710 nm were conductedwith an optical parametric laser (OPL; 5 −
10 ns,20 Hz, 0.1 nm bandwidth). A dye laser ( ≈ excimer laser(157 nm), crossing the molecular beam perpendic-ularly. The jitter of both excimer lasers was about40 ns, those of the other lasers 5 −
10 ns. Absorp-tion bands with shorter excited state lifetimes aredifficult to observe in R2C2PI using two differentlaser sources.Quantum-chemical calculations were conductedto assist the spectral analysis. Geometries andvibrations of ground and excited states as wellas vertical transition and ionization energieswere predicted with density functional theory(DFT) and time-dependent DFT (TDDFT) as2mplemented in the Gaussian09 software package(Frisch et al. 2013). The B3LYP functional (Becke1988; Lee et al. 1988) was applied in conjunctionwith the aug-cc-pVQZ basis set (Dunning 1989;Woon & Dunning 1993). Vibrational progressionsand rotational profiles were simulated with Gaus-sian09 and PGOPHER (Western 2010).
3. Results & Discussion
The theoretical ground state geometries of theobserved molecules are shown in Fig. 1. Thelowest Si n C m isomers were recently predictedwith the same method of theory (among others;Duan et al. 2010). Several more species appearedin the mass spectra of the laser ablation products,including Si n up to n = 4 and hydrogenated clus-ters, such as Si n C m H ( n = 2 , m = 1 , H was deemed to be the linearmolecule Si C . Calculations and further exper-iments, discussed in Sect. 3.7, led to the moreconclusive assignment.Fig. 2 displays the low-resolution spectra be-tween 250 and 710 nm. Spectral assignments toelectronic transitions are summarized in Table 1.A discussion follows in the subsections below. After its first detection in the laboratory,SiC was assumed to be linear by analogy withC (Kleman 1956; Bondybey 1982) until theR2C2PI study of the S (B ) ← S (A ) transition by Michalopoulos et al. (1984) revealed a triangu-lar ground state geometry. The S ← S transitionis the only electronic one that has been stud-ied between 410 and 550 nm in the laboratoryand in carbon stars (Butenhoff & Rohlfing 1991a;Sarre et al. 2000). The measurements presentedhere extend further into the UV, probing two ad-ditional transitions, assigned as S (B ) ← S (A )with an apparent origin band at 393.0 nm andS (B ) ← S (A ) with an origin band at 292.2 nm.Vibrational assignments are denoted in Fig. 3and given in Table 2. Two singlet states (S (A )and S (B )) are predicted in between. The corre- or e A B ← e X A S (B ) S (A )S (B ) S (A ) n0 n0 n0 (* 1 of S S )* n0 (cm -1 ) S (B ) S (A ) Fig. 3.— Vibrational assignments in the low-resolution (0.1 nm) spectrum of SiC .sponding transitions from the ground state haveoscillator strengths equal or close to zero.The three normal modes of SiC correspond toa C–C stretching vibration, ν (a ) = 1746.0 cm − in the ground state, a stretch between the Si atomand the C moiety, ν (a ) = 840.6 cm − , and anantisymmetric bending-type vibration, ν (b ) =196.4 cm − (Butenhoff & Rohlfing 1991a). TheS ← S spectrum features excitations to all threevibrational modes. The assignments in Table 2follow those of Butenhoff & Rohlfing (1991a) andSarre et al. (2000). In addition, the 1 , 1 ,and 1 bands have been identified.The vibrational pattern of the S ← S system isdominated by almost evenly spaced bands whichstart at 393.0 nm and have a separation of 5083
50 300 350 400 450 500 550 600 650 700ionization with 7.9 eV
52 amu: SiC
68 amu: Si C x3 5.8 eV
80 amu: Si C
92 amu: Si C
96 amu: Si C x7 7.9 eV6.4 eV
108 amu: Si C
250 300 350 400 450 500 550 600 650 7005.8 eV
104 amu: SiC H (nm) Fig. 2.— Low-resolution (0.1 nm) R2C2PI spectra. The energies of the ionizing photons are indicated.4able 1: Observed and calculated transitions and ionization energies.observed origin theory (B3LYP/aug-cc-pVQZ)molecule band (nm) symmetry assignment position (nm) a f IE (eV) b SiC v S (B ) ← S (A ) 458 (485) 0.018 9.88 (9.73)393.0 S (B ) ← S (A ) 368 0.009292.2 S (B ) ← S (A ) 253 0.036Si C 379.7 C v S (B ) ← S (A ) 379 (451) 0.033 9.19S (B ) ← S (A ) 373 4 · − c S (A ) ← S (A ) 344 0.023300 c S (B ) ← S (A ) 270 0.007S (B ) ← S (A ) 266 0.072S (A ) ← S (A ) 253 0.002Si C ∞ h e C Σ u ← e X Σ g
465 (486) 0.254 7.58 (7.48)D h S (B g ) ← S (A g ) 492 0.0 d h S (B g ) ← S (A g ) 468 0.0 d c D h S (B u ) ← S (A g ) 360 0.049D ∞ h e E Π u ← e X Σ g
385 0.015295 c D h S (B u ) ← S (A g ) 281 0.013D h S (B u ) ← S (A g ) 253 0.004D h S (B u ) ← S (A g ) 245 0.023D ∞ h e H Π u ← e X Σ g
291 0.040D ∞ h e I Σ u ← e X Σ g
280 0.830Si C ∞ h e F Σ u ← e X Σ g
292 (305) 1.766 8.22298.9 e G Π u ← e X Σ g
279 0.105Si C 680 c C v S (B ) ← S (A ) 663 0.002 8.12 (7.77)401.3 S (B ) ← S (A ) 401 (404) 0.028338.5 S (B ) ← S (A ) 330 0.017S (B ) ← S (A ) 320 0.019Si C c C v S (B ) ← S (A ) 404 0.122 7.43 (7.26)353 c S (B ) ← S (A ) 343 0.005290 c S (B ) ← S (A ) 285 0.005S (A ) ← S (A ) 275 0.019SiC H v S (B ) ← S (A ) 410 (465) 0.012 8.47 (8.24) e a vertical excitation energy (adiabatic value in brackets if calculated) b vertical ionization energy (adiabatic in brackets) c onset of broad absorption profile d unlikely assignment (see the text) e observed adiabatic IP = (8 . ± .
02) eV (see Fig. 3). λ ∆ ν band (nm) a (cm − ) b S (B ) 0 (B ) 0 (B ) 0 a band maxima ( ± b approx. band centers cm − (between the first and second bands) to 473cm − (between fourth and fifth). This progressionis attributed to the excitation of ν . Since 2 is themost intense band, a substantial displacement be-tween the Si atom and C unit is suggested. After n = 5, bands in the 2 n series become broad andare hardly discernible.Three narrow bands separated by 1372 and1356 cm − are observed in the S ← S transition.The first one, presumably the origin band, appearsat 292.2 nm. A weak and broad fourth band seemsto follow the progression at the edge of the scan-ning range where the laser power was quite low.The spacing between the bands is only 100 cm − lower than the energy of ν in the S state. Theprogression is therefore assigned as 1 n .The ground state of SiC has a highly an-harmonic potential energy surface, in contrast tothe first excited state, which is quite harmonicat the triangular geometry (Butenhoff & Rohlfing1991a). Rather harmonic potentials are also ob-served for the S and S states as demonstrated bythe almost evenly spaced 1 n and 2 n bands. Spec-tral scans at higher resolution could provide a fur-ther insight to the assignments. C The only experimental data on Si C are IRspectra of mixed carbon-silicon vapors trapped inAr matrices (Kafafi et al. 1983; PresillaM´arquez & Graham1991). Absorptions at 839.5 cm − and 1188.4cm − were ascribed to the symmetric Si–Sistretching fundamental ν (a ) and the antisym-metric Si–C stretching vibration ν (b ), respec-tively. The third mode ν (a ), a symmetric bend-ing, was inferred indirectly from an observed com-bination band to be at 166.4 cm − .Three broad absorption profiles with onsetsat about 395, 350, and 300 nm are apparent inthe Si C spectrum (Fig. 2). According to theTDDFT calculations, they can be caused by sev-eral electronic transitions with notable oscillatorstrengths, probably overlapping each other (fromS up to S ; see Table 1). The first singlet transi-tion to an A state is symmetry-forbidden.The profile extending from 395 to 350 nm, as-cribed to S (B ) ← S (A ), features a partially re-solved vibrational progression with a slightly de-creasing band separation of (550 ±
50) cm − to6490 ±
50) cm − (Fig. 4). The excitation of ν (b )in S ← S is dipole forbidden and the energy of ν (a ) is considerably lower. The progression cantherefore be assigned to 1 n . Hot bands on the redside of the profile are vaguely perceptible.Calculations suggest that the molecule is lin-ear (D ∞ h ) in the S state. The symmetric Si–Si stretching vibration ν (a ) is now better de-scribed as a symmetric Si–C–Si stretch ν ( σ g ).The computed energy of this mode is 510 cm − (B3LYP/aug-cc-pVQZ), in quite good agreementwith the observed 1 n progression. The symmetricbending mode ν (a ) becomes ν ( π g ), calculatedat 193 cm − . Because of the geometry change tothe linear form, this mode should be considerablypopulated upon S ← S excitation. Finally, theantisymmetric Si–C–Si stretching vibration ν ( σ u )is predicted at 999 cm − in the excited state.Line widths in the S ← S bands could be asbroad as 500 cm − , corresponding to a natural life-time on the order of 10 fs. In view of the compar-atively large and irregular jitter of the ionizationlaser (40 ns, pulse width of about 5 ns), an excitedstate with such a short lifetime should have beenonly hardly detectable unless a relaxation to an-other long-lived state above the ground state takesplace. Other broadening mechanisms caused bythe energetic overlap of several transitions mighthave to be considered as well.The appearance of the Si C electronic spectrumdisplaying only broad bands in the UV impedes adetection by UV-visible astronomy. An observa-tion is more likely in the radio range (calculateddipole moment of 0.71 Debye). C So far, Si C was only observed by IR spec-troscopy in Ar matrices. Two vibrational modesbelonging to the rhombic isomer (D h ) wereidentified (PresillaM´arquez et al. 1995). At theB3LYP/aug-cc-pVQZ level of theory, the linearD ∞ h structure is essentially isoenergetic (only 0.01eV higher than D h ). The electronic spectrumpresented in Fig. 2 exhibits two broad features inthe UV and four narrow bands around 515 nm.The first UV feature extends from about 365 to320 nm. Two maxima separated by ca. 890 cm − can be recognized. Substantial bandwidths pre-vent a further analysis. The second UV feature between 295 and 260 nm consists of several broadbands with irregular separation. Upon compari-son with TDDFT results, it is found that both UVsystems can be caused by the D h as well as theD ∞ h isomers (see Table 1). An overlap of severalelectronic transitions, especially within the secondabsorption system, is possible.A detailed view of the green bands and a high-resolution scan of the strongest one at 516.4 nmare presented in Fig. 5. The rhombic isomer is pre-dicted to have two symmetry-forbidden transitionsat 492 and 468 nm. The observed bands couldarise from vibrational excitations with ungerade symmetry. However, the calculated vertical ion-ization energy (IE) of the D h structure is 9.01 eV(adiabatic: 8.60 eV), 0.8 eV (0.4 eV) higher thanthe sum of both the photon energies provided atthese wavelengths, making this assignment ratherunlikely. The predicted IE of the linear molecule,on the other hand, is sufficiently low for a two-color absorption (vertical: 7.58 eV, adiabatic: 7.48eV). The first allowed transition e C Σ u ← e X Σ g is calculated at 465 nm (adiabatic: 486 nm) andhas a fairly high oscillator strength f = 0 . Σ u ← Σ g transition. The rotational constantsin the ground and excited states B ′ v and B ′′ v werecalculated by DFT and TDDFT, respectively, andkept fixed during the fitting procedure. Centrifu-gal distortion (D ′ v , D ′′ v ) was also considered, butit has an almost negligible influence on the overallprofile. The spin-spin ( λ ′ , λ ′′ ) and spin-rotation( γ ′ , γ ′′ ) coupling constants were allowed to float.All calculated and fitted values are summarizedin Table 3. The rotational temperature is esti-mated at 80 ±
40 K. Lorentzian line widths are 0.5cm − . It is found that the spin-spin coupling ismost important to simulate the separation of thethree narrower features correctly. In particular, λ has to be negative in the ground and positive inthe excited state. The difference can be explainedqualitatively. Spin-spin interaction is usually largewhen electronic Π states are nearby (Herzberg1950). Three Π states (with low f -values from7 S (B ) S (A ) (cm -1 )1 n0 Fig. 4.— Vibrational progression in the S ← S spectrum of Si C. (cm -1 ) Fig. 5.— Detailed view of the green absorptionfeature of Si C . Bottom: low resolution (0.1 nm).Top: high-resolution scan (0.002 nm; black curve)and rotational profile fit assuming the D ∞ h isomer( T rot = 80 K; red trace). e X Σ g ) are predicted in close vicinity to e C Σ u .In addition to the strong band at 516.4 nm(19359 cm − ), three weaker bands at 19442,19490, and 19703 cm − are recognizable in thelow-resolution spectrum. Franck-Condon simula-tions predict a vibrational progression (includinghot bands), which differs from the observed one.A vibronic interaction with one of the nearby Πstates is possible. Additional high-resolution andin-depth theoretical studies are needed for furtherclarification.Linear and rhombic Si C molecules have nopermanent dipole moment. An astronomical de-tection of the D ∞ h isomer is now possible in thevisible with the data presented here. The 516.4 nmband is close to three recently discovered DIBs be-tween 517 and 518 nm (Hobbs et al. 2008). Theclosest of these three, however, is 0.6 nm away. C The linear Si C C C Si molecule hasbeen studied by anion photoelectron spectroscopyand IR spectroscopy in matrix and gas phase(Van Orden et al. 1994; PresillaM´arquez & Graham1994; Nakajima et al. 1995; Duan et al. 2002;Thorwirth et al. 2011). The electronic spectrumas presented in Fig. 2 is dominated by twoabsorption systems starting at about 335 and300 nm. They can be assigned to the first al-lowed transitions, ˜ F Σ u ← ˜ X Σ g ( f = 1 .
77) and˜ G Π u ← ˜ X Σ g ( f = 0 . e F Σ u ← e X Σ g transition displays severalbroad bands of about 150 cm − widths (Fig. 6).The vibrational structure can hardly be rational-ized by a Franck-Condon approach. The strongestbands are likely caused by 5 n and 4 n progres-sions. The corresponding modes are computedat ν ′′ ( σ g ) = 437 cm − , ν ′′ ( σ u ) = 1453 cm − ,and ν ′′ ( σ u ) = 721 cm − in the excited state. Inthe ground state, they are ν ′ ( σ g ) = 471 cm − , ν ′ ( σ u ) = 2036 cm − , and ν ′ ( σ u ) = 920 cm − .Assignments are based on a Franck-Condon-Herzberg-Teller simulation with the Gaussian09software, which accounts for a linear varia-tion of the dipole moment during the transition(Frisch et al. 2013). An excess of absorption is ob-served on the blue side of the spectrum ( > − ). A mixing with the nearby e G Π u state8able 3: Spectroscopic constants (in cm − ) of lin-ear Si C obtained from the 0 band of e C Σ u ← e X Σ g . calculation a fitorigin 20567 19358.6 ± ′ v ′′ v λ ′ -0.80 ± λ ′′ ± γ ′ -0.100 ± γ ′′ -0.115 ± ′ v . × − D ′′ v . × − a B3LYP/aug-cc-pVQZ n0 n0 FCHT (cm -1 ) n0 Fig. 6.— e F Σ u ← e X Σ g transition of Si C com-pared to the Franck-Condon and Herzberg-Tellersimulation ( T vib = 0 K). is possible. Lacking a permanent dipole momentand strong visible absorptions, this molecule mightprove difficult to detect in space. C Photodissociation experiments suggest thatSi C is a particularly stable cluster (Ticknor & Duncan2005). Infrared spectroscopic studies in Ar matri-ces found excellent agreement with calculated vi-brational frequencies of the rhomboidal isomer(PresillaM´arquez & Graham 1992). The bandsystem around 400 nm (Fig. 2) has already beenobserved by R2C2PI and analyzed (Stanton et al.2005). It was assigned as e C B ← e X A basedon equation-of-motion coupled-cluster calcula-tions. At the B3LYP/aug-cc-pVQZ level of the-ory, this transition appears as e D B ← e X A (or S (B ) ← S (A )) and is labeled accordinglyhere. In addition, the S (B ) ← S (A ) andS (B ) ← S (A ) transitions were observed around630 and 330 nm. Fig. 7 is a detailed view ofthe three absorption systems along with Franck-Condon simulations of the vibrational progres-sions. Band assignments are listed in Table 4. Anextensive spectroscopic discussion is presented inanother paper (Reilly et al. 2014).The S ← S transition in the red displays abroad (1336 cm − ) profile of almost perfect Gaus-sian shape. According to the calculations, S isa transition state that leads to a different isomer.The fully optimized geometry differs substantiallyfrom that of the ground state. It has a C v sym-metry with a triangular Si C ring and an exocyclicSi atom (see also Stanton et al. 2005). Because ofthe huge geometry difference, the Franck-Condonapproach is actually unsuitable for the predictionof the pattern. Nevertheless, multiple excitationsof vibrational modes during the electronic transi-tion can be expected and, in addition to lifetimebroadening of the individual bands, likely lead tothe observed unstructured profile.The optimized geometry of S is similar to theground state structure. Thus the most intense fea-ture of S ← S is the origin band at 401.3 nm. Ex-citations involving the symmetric deformation vi-brations ν (a ) and ν (a ) dominate the rest ofthe spectrum. Weaker features are due to thelow-frequency ν (b ) out-of-plane umbrella mode,hot bands, as well as combination bands. It was9 S (B ) S (A )S (B ) S (A ) n0 n0 n0 n0 (cm -1 ) S (B ) S (A ) Fig. 7.— Electronic transitions of Si C (black)in comparison to vibrational patterns calculatedwith the Franck-Condon approach ( T vib = 300 K;red). Table 4: Si C electronic absorptions (Fig. 7). λ ∆ ν band (nm) (cm − ) a S (B ) – 625 –S (B ) 2 (B ) 3 a ±
10 cm − ∼ ps) internal conversionto a long-lived triplet state causes the observedband widths, which are about 10 to 20 cm − (Stanton et al. 2005).The band widths in the S ← S electronic spec-trum are on the order of 100 cm − indicating aneven shorter lifetime. The origin band is locatedat 338.5 nm. Excitations of the symmetric breath-ing vibration ν (a ) are mainly responsible for theobserved progression. The 1 band is the strongestfeature of this transition.An optical detection of Si C in space might beaccomplished by searching for its visible and near-UV transitions. A detection of pure rotationallines is less likely. The calculated dipole momentis only 0.09 Debye. C Three vibrational modes of Si C have beenidentified in Ar matrices, confirming the pentago-nal C v structure of the ground state (PresillaM´arquez et al.1996). The observed frequencies, ν (a ) = 681.1cm − , ν (b ) = 956.7 cm − , and ν (b ) = 597.8cm − , agree well with calculated values applyingthe B3LYP/aug-cc-pVQZ method (683, 987 and599 cm − ). The other totally symmetric modesare computed as ν (a ) = 1501 cm − , ν (a ) =468 cm − , and ν (a ) = 168 cm − .The main feature of the electronic spectrum isa broad ( ∼ − ) structure, which can beassigned to the strong S (B ) ← S (A ) transition.The theoretical oscillator strength is f = 0 . (B ) ← S (A ), iscomparatively weak ( f = 0 . (A ) ← S (A ), at 597 nm issymmetry-forbidden. Another broad and unstruc-tured absorption system starting at ca. 290 nmcan probably be attributed to the S (B ) ← S (A )or S (A ) ← S (A ) transition.A considerable geometry change is predictedupon S ← S excitation. The excited state has C s symmetry. Excitations of all four a vibrationalmodes can be expected, which is confirmed by aFranck-Condon simulation (Fig. 8). For example,excitations of six and more quanta of the lowest-frequency mode a contribute distinct intensity tothe vibrational progression. Hot bands are impor-tant at vibrational temperatures typical for molec- ular beam conditions ( T vib = 200–1000 K). Theonset of the absorption profile depends heavily on T vib (see Fig. 8). The exact position of the originband is unknown. No clear vibrational structureis resolved in the experimental spectrum due tospectral congestion and lifetime broadening. Bandwidths are probably around 100 cm − , which iscomparable to the near-UV transitions of Si C.The weak S (B ) ← S (A ) transition ( f = 0 . − ), wherea faint shoulder is recognizable.The calculated dipole moment of Si C is 1.0Debye. Considering the broad absorption in theUV, a search for this molecule might prove moresuccessful in the radio range. H On mass 104, a simple spectrum was measuredconsisting of two bands, a strong one at 445.3 nmand one at 444.0 nm, which is about five timesweaker (Figs. 2 and 9). The plotted spectra wereobtained by using the OPL at 215 nm (5.8 eV) asan ionization laser. Scans with the ArF excimerlaser (6.4 eV) revealed no additional bands furtherto the red. Considering the other species emergingfrom the ablation source, it is tempting to assignthe spectrum to the linear Si–C –Si triplet, thelowest energy isomer of Si C . The benzene-likeSiC H , however, has the same mass and its calcu-lated spectral properties are in better agreementwith the measurements as outlined in the follow-ing.The first allowed transition of Si C is pre-Table 5: Spectroscopic constants (in cm − ) ofSiC H in S and S .calculation a fit b origin 21514 22448.4 ± ′ ′ ′ ′′ ± ′′ ± ′′ ± ν ′′ (b )- ν ′ (b ) 80.8 66.9 ± a B3LYP/aug-cc-pVQZ b obtained by fitting the origin band of S (B ) ← S (A ) S (B ) S (A ) (cm -1 )
500 K
Fig. 8.— S ← S spectrum of Si C and Franck-Condon simulation at two different temperatures. simulation 50 K (cm -1 )discharge in Helaser ablation in 10%Ar/Hesimulation 10 K0 ~ Fig. 9.— High-resolution (0.002 nm) scans ofthe mass 104 spectrum compared to calculatedabsorption profiles of the S (B ) ← S (A ) originband of SiC H . dicted at 575 nm and is very strong ( f = 0 . u ← Σ g symmetry.Its origin band would feature neither the strongQ branch nor the spin-orbit split necessary to ex-plain the measured band profile and the presenceof the weak 444.0 nm band. Excited state vi-brations were calculated and compared to thosein the ground state. A vibration or hot band ofSi C cannot explain the weak band either. Thecomputed S (B ) ← S (A ) transition of SiC H ( f = 0 . f = 0 . (B ) ← S (A ) ispredicted at 284 nm where the laser power of theOPL is about ten times lower than at 445 nm, butthe real transition might be located further in theUV, outside of the scan range, and therefore itcould have escaped detection.The same spectrum was also observed by dis-charging phenylsilane (SiC H ) evaporated atroom temperature in He atmosphere (5 bar backpressure on pulse valve). Fragments of the parentmolecule were the main species that appearedin the mass spectra, i.e., SiC H should havebeen much more abundant than Si C in theseexperiments, confirming the above assignment.A simulation of the band profile based on calcu-lated SiC H geometries in the ground and excitedstates is in good agreement with the observed pro-file (see Fig. 9). The band at 444.0 nm can beexplained by the 18 hot band. The low-frequencymode ν (b ) is calculated at 153 cm − in S and 234 cm − in S . Table 5 lists the theoreti-cal and deduced spectroscopic constants. The hotband intensity could be lowered by adding 10% ofAr to the supersonic expansion. This decreasedthe vibrational temperature from (150 ±
10) Kto (120 ±
10) K and the rotational temperaturefrom (70 ±
10) K to (50 ±
5) K, as concluded froma Franck-Condon analysis and rotational profilefits. The overall calculated vibrational patterndoes not agree with the observation of only twobands. Predissociation can be excluded. The min-imum energy to remove a hydrogen (or silicon) iscalculated at 4.5 eV (4.7 eV). An interaction withthe nearby dark state S (A ) is possible.A two-color photoionization threshold scan was12onducted to determine the IE of the absorbingspecies. The band intensity was monitored whilekeeping the wavelength of the first OPL at 445.3nm (2.78 eV) and scanning the wavelength of theionization laser, a second OPL, between 210 and224 nm (5.90 – 5.54 eV). The IE of (8 . ± . H (vertical IE = 8.47 eV,adiabatic IE = 8.24 eV). At the same time, the cal-culated value of Si C differs by more than 1 eV(vertical IE = 7.29 eV). By measuring the 445.3nm band intensity while varying the delay betweenscan and ionization laser the excited state lifetimewas determined to be τ = (8 ±
1) ns.This is the first spectroscopic study of SiC H .Aside from its absorptions in the visible, it hasa strong dipole moment, calculated at 1.65 De-bye. Therefore, SiC H could be easier to find inspace than benzene, which was discovered only byinfrared observations of a protoplanetary nebula(Cernicharo et al. 2001).
4. Summary
Several silicon-carbon molecules have been pre-pared by laser ablation and studied by R2C2PIspectroscopy between 250 and 710 nm. New bandsand transitions were found for SiC and Si C. Themolecules Si C, Si C , Si C , and Si C were in-vestigated for the first time by electronic spec-troscopy. This is also the first experimental studyof the linear Si C isomer and the benzene-likeSiC H radical. Aside from SiC , which has al-ready been found in space, linear Si C , Si C,and SiC H are interesting target molecules forastronomical searches in the visible due to theirspectral properties. A band at 516.4 nm with arather unique shape due to spin-spin splitting wasfound for linear Si C . The calculated oscillatorstrength of the corresponding transition is fairlyhigh ( f = 0 . H could be easier to detectthan benzene because of its transition at 445.3 nmand its strong dipole moment. The simplicity ofthe electronic spectrum could also hint at the DIBproblem. Like SiC H , the carriers of these enig-matic bands probably do not feature complicatedvibrational progressions and strong UV bands asmost DIBs do not correlate with each other andnone have been found below 400 nm. This work has been funded by the SwissNational Science Foundation (Project 200020-140316/1). REFERENCES
Apponi, A.J., McCarthy, M. C., Gottlieb, C. A., &Thaddeus, P. 1999, J. Chem. Phys., 111, 3911Apponi, A.J., McCarthy, M. C., Gottlieb, C. A.,& Thaddeus, P. 1999, ApJ, 516, L103Becke, A. D. 1988, Phys. Rev. A, 38, 3098Bernath, P. F., Rogers, S. A., O’Brien, L. C., &Brazier, C. R. 1988, Phys. Rev. Lett., 60, 197Bondybey, V. E. 1982, J. Phys. Chem., 86, 3396Brazier, C. R., O’Brien, L. C., & Bernath, P. F.1989, J. Chem. Phys., 91, 7384Butenhoff, T. J., & Rohlfing, E. A. 1991a,J. Chem. Phys., 95, 1Butenhoff, T. J., & Rohlfing, E. A. 1991b,J. Chem. Phys., 95, 3939Cernicharo, J., Gottlieb, C. A., Guelin, M., Thad-deus, P., & Vrtilek, J. M., D. 1989, ApJ, 341,L25Cernicharo, J., Heras, A. M., Tielens, A. G. G. M.,Pardo, J. R., Herpin, F., Gu´elin, M., & Waters,L. B. F. M. 2001, ApJ, 546, L123Ding, X. D., Wang, S. L., Rittby, C. M. L., &Graham, W. R. M. 1999, J. Chem. Phys., 110,11214Ding, X. D., Wang, S. L., Rittby, C. M. L., &Graham, W. R. M. 2000, J. Phys. Chem. A,104, 3712Duan, X., Burggraf, L. W., Weeks, D. E., Davico,G. E., Schwartz, R. L., & Lineberger, W. C.2002, J. Chem. Phys., 116, 3601Duan, X., Wei, J., Burggraf, L., & Weeks, D. 2010,Comput. Mater. Sci., 47, 630Dunning, T. H. 1989, J. Chem. Phys., 90, 1007Ebben, M., Drabbels, M., & ter Meulen, J. J. 1991,Chem. Phys. Lett., 176, 40413risch, M. J., Trucks, G. W., Schlegel, H. B., et al.2013, Gaussian 09, Revision D.01 (Wallingford,CT: Gaussian, Inc.)Fye, J. L., & Jarrold, M. F. 1997, J. Phys. Chem.A, 101, 1836Gottlieb, C. A., Vrt´ılek, J. M., & Thaddeus, P.1989, ApJ, 343, L29Grutter, M., Freivogel, P., & Maier, J. P. 1997, J.Phys. Chem. A, 101, 275Gu´elin, M., Muller, S., Cernicharo, J., Apponi, A.J., McCarthy, M. C., Gottlieb, C. A., & Thad-deus, P. 2000, ˚a, 363, L9Gu´elin, M., Muller, S., Cernicharo, J., McCarthy,M. C., & Thaddeus, P. 2004, ˚a, 426, L49Han, D., Rittby, C. M. L., & Graham, W. R. M.1997, J. Chem. Phys., 106, 6222Herzberg, G. 1950, Molecular Spectra and Molecu-lar Structure Vol. 1, D. van Nostrand CompanyInc., Toronto, CanadaHobbs, M., York, D. G., Snow, T. P., Oka, T.,Thorburn, J. A., Bishof, M., Friedman, S. D.,McCall, B. J., Rachford, B., Sonnentrucker, P.,& Welty, D. E. 2008, ApJ, 680, 1256Kafafi, Z. H., Hauge, R. H., Fredin, L., & Mar-grave, J. L. 1983, J. Phys. Chem., 87, 797Kaiser, R. I., & Osamura, Y. 2005, ApJ, 630, 1217Kaiser, R. I., Krishtal, S. P., Mebel, A. M.,Kostko, O., & Ahmed, M. 2012, ApJ, 761, 178Kleman, B. 1956, ApJ, 123, 162Kokkin, D. L., Br¨unken, S., Young, K. H., Patel,N. A., Gottlieb, C. A., Thaddeus, P., & Mc-Carthy, M. C. 2011, ApJS, 196, 17Kokkin,D. L., Reilly, N. J., Fortenberry, R. C.,Crawford, T. D., & McCarthy, M. C. 2014,J. Chem. Phys., 141, 044310Lˆe, T. H., Rittby, C. M. L., & Graham W. R. M.2014a, J. Chem. Phys., 140, 064314Lˆe, T. H., Rittby, C. M. L., & Graham W. R. M.2014, J. Chem. Phys., 141, 044315 Lee, C., Yang, W., & Parr, R. G. 1988, Phys. Rev.B, 37, 785Lloyd Evans, T., Hurst, M. E., & Sarre, P. J. 2000,MNRAS, 319, 111McCarthy, M. C., Apponi, A. J., Gottlieb, C. A.,& Thaddeus, P. 2001, J. Chem. Phys., 115, 870McCarthy, M. C., Apponi, A. J., & Thaddeus, P.1999, J. Chem. Phys., 111, 7175McCarthy, M. C., Gottlieb, C. A., & Thaddeus,P. 2003, Molecular Physics, 101, 697Merrill, P. W. 1926, Publ. Astron. Soc. Pac., 38,175Michalopoulos, D. L., Geusic, M. E., Langridge-Smith, P. R. R., & Smalley, R. E. 1984,J. Chem. Phys., 80, 3556Morgan, D. H., Hatzidimitriou, D., & Cannon, R.D. 2004, MNRAS, 355, 1196M¨uller, H. S. P. , Cernicharo, J., Ag´undez,M., Decin, L., Encrenaz, P., Pearson, J. C.,Teyssier, D., & Waters, L. B. F. M. 2012, J.Mol. Spectrosc., 271, 50Nakajima, A., Taguwa, T., Nakao, K., Gomei,M., Kishi, R., Iwata, S., & Kaya, K. 1995,J. Chem. Phys., 103, 2050Ohishi, M., Kaifu, N., Kawaguchi, K., Murakami,A., Saito, S., Yamamoto, S., Ishikawa, S.-I., Fu-jita, Y., Shiratori, Y., Irvine, W. M. 1989, ApJ,345, L83PresillaM´arquez, J. D., & Graham, W. R. M. 1991,J. Chem. Phys., 95, 5612PresillaM´arquez, J. D., Gay, S. C., Rittby, C. M.L., & Graham, W. R. M. 1995, J. Chem. Phys.,102, 6354PresillaM´arquez, J. D., & Graham, W. R. M. 1994,J. Chem. Phys., 100, 181PresillaM´arquez, S. C., Rittby, C. M. L., & Gra-ham, W. R. M. 1997, J. Chem. Phys., 106, 8367PresillaM´arquez, J. D., & Graham, W. R. M. 1992,J. Chem. Phys., 96, 6509PresillaM´arquez, S. C., Rittby, C. M. L., & Gra-ham, W. R. M. 1996, J. Chem. Phys., 104, 281814resillaM´arquez, S. C., Graham, W. R. M., &Shepherd, R. A. 1990, J. Chem. Phys., 93, 5224Reilly, N. J., Steglich, M., Kokkin, D. L., Maier,J. P., Stanton, J. F., & McCarthy, M. C. 2014,J. Mol. Spectrosc., accepted
Sanford, R. F. 1926, Publ. Astron. Soc. Pac., 38,177Sarre, P. J., Hurst, M. E., & Lloyd Evans, T. 2000,MNRAS, 319, 103Savoca, M., Lagutschenkow, A., Langer, J., Hard-ing, D. J., Fielicke, A., & Dopfer, O. 2013, J.Phys. Chem. A, 117, 1158Shepherd, R. A., & Graham, W. R. M. 1985,J. Chem. Phys., 82, 4788Smith, T. C., Li, H., Clouthier, D. J., Kingston,C. T., & Merer, A. J. 2000, J. Chem. Phys.,112, 3662Stanton, J. F., Dudek, J., Theul´e, P., Gupta,H., McCarthy, M. C., & Thaddeus, P. 2005,J. Chem. Phys., 122, 124314Thaddeus, P., Cummins, S. E., & Linke, R. A.1984, ApJ, 283, L45Thorwirth, S., Krieg, J., Lutter, V., Keppeler, I.,Schlemmer, S., Harding, M. E., V´azquez, J., &Giesen T. F. 2011, J. Mol. Spec., 270, 75Ticknor, B. W., & Duncan, M. A. 2005, Chem.Phys. Lett., 405, 214Van Orden, Giesen, T. F., A., Provencal, R.A., Hwang, H. J. & Saykally, R. J. 1994,J. Chem. Phys., 101, 10237Van Orden, A., Provencal, R. A., Giesen, T. F.,& Saykally, R. J. 1995, Chem. Phys. Lett., 237,77Weltner, W., & McLeod, D. 1964, J. Chem. Phys..41, 235Western, C. M. 2010, PGOPHER, ver-sion 7.1.108 (University of Bristol,http://pgopher.chm.bris.ac.uk)Withey, P. A., & Graham, W. R. M. 1992,J. Chem. Phys., 96, 4068 Woon, D. E., & Dunning, T. H. 1993,J. Chem. Phys., 98, 1358