Laser velocity modulation spectroscopy of the 3 Δ(3d4s)–X 3 Φ(3d 2 ) visible system ofTiCl + and characterization of the spin–orbit structureC. Focsa, C. Dufour, B. Pinchemel, I. Hadj Bachir, and T. R. Huet Citation: The Journal of Chemical Physics 106, 9044 (1997); doi: 10.1063/1.474036 View online: http://dx.doi.org/10.1063/1.474036 View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/106/22?ver=pdfcov Published by the AIP Publishing Articles you may be interested in Photoelectron spectroscopic study of the Ee Jahn–Teller effect in the presence of a tunable spin–orbitinteraction. I. Photoionization dynamics of methyl iodide and rotational fine structure of CH3I+ and CD3I+ J. Chem. Phys. 134, 054308 (2011); 10.1063/1.3547548 High resolution study of spin-orbit mixing and the singlet-triplet gap in chlorocarbene: Stimulated emissionpumping spectroscopy of CH 35 Cl and CD 35 Cl J. Chem. Phys. 129, 104309 (2008); 10.1063/1.2977686 Perturbations in the pure rotational spectrum of CoCl (X 3 Φ i ): A submillimeter study J. Chem. Phys. 121, 8385 (2004); 10.1063/1.1795691 Further studies of 3d transition metal cyanides: The pure rotational spectrum of NiCN (X 2 Δ i ) J. Chem. Phys. 118, 6370 (2003); 10.1063/1.1557471 Laser velocity modulation spectroscopy of TiCl + : Observation of the A 3 Δ(3d 2 ) state and deperturbation of theX 3 Φ−A 3 Δ complex J. Chem. Phys. 107, 10365 (1997); 10.1063/1.475313

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

128.114.34.22 On: Tue, 02 Dec 2014 14:11:59

Laser velocity modulation spectroscopy of the 3D(3d4s )–X3F(3d 2) visiblesystem of TiCl 1 and characterization of the spin–orbit structure

C. Focsa, C. Dufour, and B. Pinchemela)Laboratoire de Dynamique Mole´culaire et Photonique, URA CNRS 779, Centre d’Etudes et de RecherchesLasers et Applications, Universite´ des Sciences et Technologies de Lille, 59 655 Villeneuve d’AscqCedex, France

I. Hadj Bachir and T. R. HuetLaboratoire de Spectroscopie Hertzienne, URA CNRS 249, Centre d’Etudes et de Recherches Lasers etApplications, Universite´ des Sciences et Technologies de Lille, 59 655 Villeneuve d’Ascq Cedex, France

~Received 13 January 1997; accepted 5 March 1997!

A single mode cw dye laser excitation along with velocity modulation detection was used to recordbetween 17 100 and 18 600 cm21 the absorption spectrum of the3D(3d4s) –X3F(3d2) visiblesystem of TiCl1 produced in an ac glow discharge with a gas mixture of He/TiCl4. The rotationalstructure of the~0,0! and~1,0! vibrational bands has been observed and fully analyzed for the mainisotopomer Ti35Cl1 as well as for Ti37Cl1. Beside the confirmation of the nature of the twoelectronic states, the observation of the forbidden3D2–

3F2 and 3D3–3F3 intercombination

transitions allowed the first experimental determination of the spin–orbit constants and a detailedcharacterization of both the3D and 3F states through a matricial fitting of the data. Molecularparameters have been derived. ©1997 American Institute of Physics.@S0021-9606~97!01622-X#

I. INTRODUCTION

The study of molecular ions has been proven to be ofgreat interest in astronomy and astrophysics. Especially, nu-merous positive ions are involved in the set of possible ki-netical reactions for modelising physical and chemical prop-erties, because ion-neutral reactions do not need activationenergy. The detection of H3

1 in the interstellar space is abeautiful illustration of the success of molecularastrophysics.1,2 Recent studies have shown that there are in-terstellar dusts which contain silicates or carbonates with me-tallic atoms like Fe or Mg.3 Also the observation of diatomicmolecules containing transition elements, like TiO, in emis-sion spectra of coolM -type stars4 strongly suggests that mo-lecular ions containing metallic atoms are present in the in-terstellar medium and that it could be possible to identifythem through their spectroscopic signature.5 Reactions ofions with neutral metal atoms are also important eventhrough simple charge transfer.6,7

The TiCl1 ion has been observed for the first time inemission spectra by Balfour and Chandrasekhar.8 They ob-served at low resolution a system in the yellow–green regionwith the typical structure of a triplet–triplet transition. Theytentatively assigned the spectrum to a3P–3D electronictransition. Only theDV521 allowed spin–orbit transitionsof the ~0,0! vibrational bands were rotationally analyzedleading to band origins and effective rotational parametersdetermination. The electronic assignment was revised byKaledin, McCord, and Heaven9 on the basis of calculationsperformed with ligand field theory models. The observed vis-

ible emission system was predicted to be due to a3D(3d4s) –X3F(3d2) transition. The experimental confir-mation was reported very recently by Kaledinet al.10 by us-ing the experimental setup developed for their study ofLaF1.11

At the same time, we have developed in Lille an experi-mental setup to produce molecular ions containing metallicatoms in order to study their laser absorption spectrum withthe velocity modulation technique. This technique12,13 hasbeen proven to be very efficient to characterize molecularions produced in electrical discharge with small concentra-tions and led to the observation in laboratory of numerousmolecular ions, often in the infrared region, and mainly inthe groups of Oka at Chicago and Saykally at Berkeley.

In this paper we report the experimental study of the~0,0! and ~1,0! vibrational bands of the3D(3d4s)–X3F(3d2) visible absorption system of TiCl1, for themain isotopomer Ti35Cl1, as well as the~0,0! and ~1,0!bands of Ti37Cl1. The analysis of the rotational structure, upto J values around 100, of the main sequences and of theforbidden intercombination bands was performed and mo-lecular constants have been obtained using a matrix model.The present work is the first extensive experimental charac-terization of the electronic, vibrational, and rotational struc-ture of TiCl1 leading to the determination of the relativepositions of the spin–orbit components of the two electronicstates.

II. EXPERIMENT

A schematic diagram of the experimental apparatus isshown in Fig. 1. A Coherent 899-29 ring cw dye laser run-ning either with R110 or R6G dyes and pumped by an Ar1

laser~Coherent CR20! was employed to record a spectrum ofTiCl1 in the region between 17 100 and 18 600 cm21. The

a!Author to whom correspondence should be addressed. Laboratoire de Dy-namique Mole´culaire et Photonique, URA CNRS 779, Baˆtiment P5, U.F.R.de Physique, Universite´ des Sciences et Technologies de Lille, F-59 655Villeneuve d’Ascq Cedex, France.

9044 J. Chem. Phys. 106 (22), 8 June 1997 0021-9606/97/106(22)/9044/6/$10.00 © 1997 American Institute of Physics This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

128.114.34.22 On: Tue, 02 Dec 2014 14:11:59

laser power was typically in the range 200–400 mW depend-ing on the spectral region. The TiCl1 ions are produced in awater-cooled glass cell~20 mm i.d., length 1 m! through anac glow discharge. A gas mixture of helium~99.999% pu-rity! at a pressure of 5–7 Torr and of TiCl4 vapor at a pres-sure of 0.1 Torr is introduced at both ends of the cell near theelectrodes, and is pumped through a central outlet by a rotarypump. The discharge is driven at 30 kHz with typical sinu-soidal current of 300 mA peak-to-peak and voltage of 1 kVpeak-to-peak. The signals of TiCl1 are found to be very sen-sitive to small variations of partial gas pressures and to therate flow. We observed that the experimental conditions areoptimized when the color of the discharge is blue–white, i.e.,when probably strong Ti atomic emission is present, andoutshines the pink color of a pure helium discharge. It waspossible to run the discharge in such stable conditions forhours. In addition to the spectrum of TiCl1, strong atomiclines belonging to Ti1 and Cl11 where also observed in thisspectral range.

The absorption spectrum was recorded using the velocitymodulation technique12,13 in order to eliminate signals fromneutral species. In addition a noise substraction technique14

was used in order to reduce the noise fluctuations from thedischarge and from the lasers. For this purpose15 two coun-terpropagating laser beams are sent into the cell, and detectedby two PIN photodiodes mounted on a low gain amplifierwith large bandwidth. Two passive RC filters are used forremoving the low frequency signals of the order of the kHz.Then the voltages are substracted, demodulated, and ampli-fied by a lock-in amplifier, at the frequency of the discharge,leading to a close first-derivative shape of the rotationallines. Most spectra were recorded with a time constant of300 ms and a sensitivity of 500mV to 20 mV. All the lineswere measured with the Coherent Autoscan software, cali-brated against the well-known iodine lines.16 The absoluteaccuracy is estimated to be better than 0.005 cm21. The rela-tive sharpness of the linewidth allowed us to take advantageof the high sensitivity and resolution of the experimentalmethod. From the study of the line profile17 we estimated

that the translational temperature in the discharge is around800 K.

III. ANALYSIS

A. Observed bands

In the present work the continuous recording of the ab-sorption spectrum of TiCl1 over 1500 cm21 (17 100–18 600 cm21) gives access to a large amount of informa-tion about the rovibrational structure of48Ti35Cl1 and48Ti37Cl1 isotopic species. The spectra of the ions involvingfour other weak isotopes of titanium, with relative abun-dances of 5%–8%, have not been studied in detail. Theirpresence is however observed, inducing a more or less con-gested background.

The good sensitivity of the experimental method ensuresa high signal to noise ratio especially for the 0–0 vibrationalbands. We first took advantage of this feature to acertain thelabeling of the electronic states involved in the transition.Figure 2 displays a small part of the spectrum of the reder0–0 vibrational transition of TiCl1 near the band origin(17 722 cm21). It was previously labeled as the3P2–

3D3

subsystem by Balfour and Chandrasekhar8 and renamed3D3–

3F4 by Kaledinet al.9 Our rotational assignment is in

agreement with the results presented in the Table III of Ref.8 but the obvious lack of theP(3) line on our spectrumconfirms both the theoretical9 and the experimental10 worksby Kaledin and co-workers. Indeed the first observed line isclearly theP(4) line linking the firstJ954 rotational levelof a low-lying 3F4 state to the firstJ853 rotational level ofan upper3D3 state.

Up to now no experimental information was available onthe spin–orbit coupling parameters of both the3F and 3Dstates involved in this system. Only theoretical predictionsbased on the ligand field theory were given in a recent paperby Kaledinet al.9 On our spectra a well resolved but weakband is observed at 17 985 cm21. The absence of aQ branchis typical of aDV50 transition and this band was thereforeassigned as the forbidden intercombination band linking the3D2 upper spin–orbit component to the3F2 lower one, asconfirmed later by the rotational analysis. In addition a weakblended band was observed at 17 916 cm21 and assigned asthe 3D3–

3F3 second intercombination transition. From theobservation of these two intercombination bands it has beenpossible to build an energy level diagram displaying the rela-tive position of the spin–orbit components for the two elec-tronic states~Fig. 3! and to handle the data through a matrixmodel.

B. Theoretical model

According to Brownet al.18 the effective Hamiltonianused for the analysis of the spin and rotational structure isgiven by the following expression:

FIG. 1. Schematic representation of the experimental setup.

9045Focsa et al.: The 3D(3d4s)–X3F(3d2) visible system of TiCl1

J. Chem. Phys., Vol. 106, No. 22, 8 June 1997 This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

128.114.34.22 On: Tue, 02 Dec 2014 14:11:59

H5T1BN22DN41 12 @A1ADN

2,LzSz!]1

1@l1lDN2,Sz

22 13 S

2#11 12 @g1gDN

2,N.S#1 , ~1!

where the terms can be identified by their coefficients, asfollows: T is the vibronic term value,B is the rotationalconstant,D is the centrifugal distortion constant for the ro-tational motion,A andl are the first- and second-order spin–orbit coupling constants,g is the spin-rotation interactionparameter, andAD , lD , and gD represent the centrifugaldistorsion corrections to the first- and second-order spin–orbit coupling, and to the spin-rotation parameter, respec-tively. The symbol@ ,#1 denotes an anticommutator. In theabsence of any resolvedL-doubling structure in the3D1 sub-state the corresponding Hamiltonian part has been omitted.We have used the formBN25B(J2S)2 for the rotationalenergy rather than the formBR25B(J2L2S)2. Indeed itwas shown by Brownet al.19 that the influence of the mass-dependence of isotopic substitution is more correctly ac-counted for by the first form when both rotational and vibra-tional isotopic shifts have to be determined. Brownet al.18

have pointed out that an indeterminacy exists among the mo-lecular constantsB, AD , lD , and g for Hund’s case~a!3D state, essentially because there are only three effectiveB values for the three spin–orbit components, but four pa-rameters to be determined for them. We therefore excludedthe spin-rotation parameterg as suggested by Brown andco-workers.18

C. Least-squares procedure

Eight bands~Fig. 3! have been simultaneously fitted inorder to determine the molecular parameters characterizingthe 3D and 3F states. We have used the three main se-quences of the 0–0 and 1–0 bands and the3D2–

3F2 and the3D3–

3F3 intercombination transitions of the 0–0 band. 1269rotational lines have been simultaneously fitted for the mainisotopomer48Ti35Cl1, and 769 lines for48Ti37Cl1. The set ofassigned lines has been placed in Physics Auxiliary Publica-tion Service~PAPS!.20 All the lines were weighted to theexperimental uncertainty (60.005 cm21). The overall stan-dard deviation is equal to 0.0045 cm21, within the experi-mental uncertainty.

The v850,1 vibrational levels of the3D state are verywell described by the matrix model suggesting that no closelying state is interacting with. On the contrary the lower3F state is not correctly accounted for by the model. Asalready observed by Balfour and Chandrasekhar,8 the evolu-tion of the effectiveB rotational parameters in this state doesnot follow a quasilinear evolution whenV increases, as it isobserved in the upper3D state. It turns out that the3F4

substate looks like being affected by a perturbation, as dis-cussed in the next section. We therefore used an effectivemodel by sharing the matrix of the3F state in two parts. Thefirst one is the 232 matrix associated with the3F3 and3F2 substates coupled by an off-diagonal matrix element andthe second one is a single term leading to a set of effective

FIG. 2. Example of spectrum for the3D(3d4s) –X 3F(3d2) visible system of TiCl1. The main sequence3D3–X3F4 between bothv50 vibrational levels

is displayed. The absence of theP(3) line confirms the reassignment of the electronic transition~see text!.

9046 Focsa et al.: The 3D(3d4s)–X3F(3d2) visible system of TiCl1

J. Chem. Phys., Vol. 106, No. 22, 8 June 1997 This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

128.114.34.22 On: Tue, 02 Dec 2014 14:11:59

parameters for the3F4 substate, as justified hereafter. In factwe assumed that the electronic term of the3F4 substate isnot influenced by the perturbation and therefore we decidedto keep the form of the constant term in the matrix elementof the 3F4 substate in order to determine a single set ofparametersT0 , A, andl for the 3F state. On the contrarythe introduction of an effectiveB09 parameter in the diagonalmatrix element of the3F4 substate leads to the indetermina-

tion of the second-orderlD constant. The molecular param-eters of both3D and 3F electronic states are listed inTable I.

The Table II gives a set of parameters obtained wheneach substate is described by the following expression:21

Tv1BeffJ~J11!2DeffJ2~J11!2, ~2!

whereTv is the vibronic term value,Beff is an effective ro-tational constant andDeff is an effective centrifugal distor-tion constant. Figure 3 has been built from the data listed inTable II which also includes the energies of the spin–orbitcomponents theoretically calculated by Kaledinet al.9 Theequilibrium parameters for the3D state are presented inTable III. From the data collected in Tables I–III, numerouscomparisons can be made between the parameters calculatedfor the two main isotopomers. As an example the ratio be-tween the equilibrium constantsBe for the

3D state is equalto 1.032 23 which has to be compared to the ratio of thereduced masses of the two isotopomers,r25m(48Ti37Cl1)/m(48Ti35Cl1)51.032 25.22 We also note that the first-orderand second-order spin–orbit constants are isotopically in-variant, as expected.

IV. DISCUSSION

The least-squares fitting of the data shows that the upper3D electronic state has been easily described by a matrixbuilt on the assumption of an isolated electronic state. If wecompare our results with those of Ref. 8 we can observe asurprising discrepancy on the value of the centrifugal distor-tion parametersD which are almost exactly twice larger thanours despite the fact that they include also lines up toJ5110 in a few branches. Also taking into account that thewave numbers listed in Table II of Ref. 8 are in good agree-ment with ours, this discrepancy can only be explained bythe factor of 2 appearing in the term 2(D81D9)J3 when theP andR branches of a same band are fitted simultaneously.21

The identification of thev850 and v851 vibrationallevels of the upper3D state gives access to the determinationof DG1/25ve822ve8xe5507.7 cm21 ~Table III!. A value ofve8 can be deduced from the well-known followingrelation:21

FIG. 3. The energy level diagram and the observed transitions of the3D(3d4s) –X 3F(3d2) visible system of TiCl1. The three spin–orbit sub-states lie at 0 cm21, 188.56 cm21, and 382.73 cm21 in the ground vibra-tional level of theX 3F state, at 17 873.18 cm21, 17 984.75 cm21, and18 104.66 cm21 for thev50 vibrational level of the3D(3d4s) state, and at18 381.06 cm21, 18 492.44 cm21, and 18 612.19 cm21 for the v51 vibra-tional level of the3D(3d4s) state. The effective energies are extrapolated tothe virtualJ50 rotational level.

TABLE I. Rotational and spin–orbit parameters~in cm21! derived from the analysis of the3D–X3F ~0,0! and~1,0! bands of Ti35Cl1 and Ti37Cl1 isotopomers~all uncertainties are 1s!.

Parameter

48Ti35Cl1 48Ti37Cl1

3D(v51) 3D(v50) 3F(v50) 3D(v51) 3D(v50) 3F(v50)

Tv 18 304.797~1! 17 797.096~1! 0a 18 296.633~1! 17 796.854~1! 0a

Av 58.137~1! 58.228~1! 64.136~1! 58.134~1! 58.224~1! 64.131~1!ADv3104 20.138(2) 20.135(2) 0.671~1! 20.089(8) 20.083(7) 0.639~2!

Bv 0.177 622~4! 0.178 478~4! 0.174 095~4!b 0.172 083~6! 0.172 899~6! 0.168 654~6!b

Dv3107 0.866~5! 0.863~5! 0.870~5!b 0.799~12! 0.793~11! 0.802~12!b

lv 2.276~1! 2.266~3! 1.579~1! 2.272~1! 2.263~1! 1.573~1!lDv3105 0.20~3! 0.34~2! 0a 0.53~9! 0.60~7! 0a

aFixed value in the fit.bThese entries are derived from a partial 232 matrix including the3F3 and

3F2 substates; the3F4(v950) component has been considered as an isolated and

perturbed state described by effective parameters, Ti35Cl1: B0950.174 007(4), D0950.795(5)31027; Ti37Cl1: B0950.168 576(7), D0950.715(12)31027.

9047Focsa et al.: The 3D(3d4s)–X3F(3d2) visible system of TiCl1

J. Chem. Phys., Vol. 106, No. 22, 8 June 1997 This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

128.114.34.22 On: Tue, 02 Dec 2014 14:11:59

ve85A4Be83

De8. ~3!

From the parameters listed in Table III we findve8'510 cm21. This is in good agreement with the experimen-tal result if we take into account the value estimated forve8xe'2 cm21 derived from the Pekeris’ expression:23

ae56SAvexeBe3

ve2Be2

veD . ~4!

The behavior of the ground electronic state seems to bemuch more puzzling. It turns out that its description with amatrix obtained from a Hamiltonian accounting for an iso-lated electronic state is not satisfactory. This is evidenced bythe value of the effectiveB rotational constant of the3F4

spin–orbit component, which is smaller than expected whenit is compared to the two otherB values. A possible pertur-bation by another vibrational level of the same electronicstate will be discussed hereafter.

When spectra of two isotopic species are observed it ispossible to determine an accurate value of (ve82ve9) fromthe energy shift between the origins of the 0–0 bands of thetwo isotopes which depends on the isotopic ratior,21

n02n0i512 ~ve82ve9!~12r21!. ~5!

Here all the parameters refer to48Ti35Cl1, exceptn0i whichis related to48Ti37Cl1. The derived value forve8–ve9 is30.5 cm21, and takingve8'510 cm21 we obtain an estima-tion of ve9'480 cm21. This value is compatible with theobservation of 1–1 bandheads by Balfour andChandrasekhar:8 the distances between their 1–1 bandheadsand our 1–0 ones are, respectively, 476 cm21, 483 cm21,and 484 cm21 for each of the three spin–orbit subsystems.Moreover, using Eq.~3! with B09 instead ofBe9 , a value ofve9'489 cm21 is obtained. Now we can see that the value ofve9 locates thev9(3F2)51 level around 475 cm21, close tothe v9(3F4)50 level located at 382 cm21. We might con-sider that an interaction occurs between these two vibrationallevels. In fact the direct interaction between the two substatesis not allowed because of the selection ruleDV561. Wetherefore have to assume that thev9(3F3)50, 1 levels are

also involved in the interaction as intermediate states be-tween v9(3F2)51 and v9(3F4)50 substates. The off-diagonal coupling matrix element is proportional to@J(J11)2V(V21)#1/2.24,25 If we apply the first order perturba-tion theory, the contribution of this term on the diagonalelement will be proportional toJ2, and this could explain theunexpectedB9 value of thev9(3F4)50 state which pre-vented the use of the matrix model. Despite the fact that theexpected positions of theDv521 sequences can be confi-dently determined from above considerations, we have notobserved any 0–1 vibrational bands. It could be due to thefact that the velocity modulation spectroscopy is an absorp-tion technique depending on the Boltzmann distribution forthe population of the vibrational levels. A high enough vi-brational temperature would not have been reached in ourdischarge. However theDv521 transitions seem to be in-trinsically very weak since they have not been observed inemission by Balfour and Chandrasekhar.8 The observation ofthe 0–1 vibrational bands will be necessary to characterizethe threev951 vibrational spin–orbit sublevels in order toclarify these interactions. Future efforts will be done into thisdirection.

V. CONCLUSION

The efficiency of the ac glow discharge in a gas mixtureof He/TiCl4 allowed the production of the TiCl1 ion insteady conditions over several hours. This method, associ-ated with the velocity modulation detection technique, al-

TABLE III. Equilibrium parameters~in cm21 unless quoted! for the3D stateof TiCl1 ~all uncertainties are 1s!.

Parameter 48Ti35Cl1 48Ti37Cl1

ve–2vexe 507.701~1! 499.779~1!Ae 58.274~1! 58.269~1!Be 0.178 906~8! 0.173 307~12!ae 0.000 856~8! 0.000 816~12!

De3107 0.861~10! 0.790~23!r e ~Å! 2.157 90~5! 2.157 96~7!

TABLE II. Effective parameters~in cm21! for the spin–orbit components of the3D andX 3F states of Ti35Cl1 and Ti37Cl1 ~all uncertainties are 1s!.

State

48Ti35Cl1 48Ti37Cl1Theoreticala

values forTvTv Beff Deff3107 Tv Beff Deff3107

3D3(v51) 18 612.186~1! 0.178 136~4! 0.919~7! 18 604.059~9! 0.172 564~16! 0.83~6!3D2(v51) 18 492.436~1! 0.177 661~4! 0.875~5! 18 484.282~5! 0.172 116~7! 0.83~2!3D1(v51) 18 381.064~1! 0.177 091~4! 0.856~6! 18 372.904~1! 0.171 569~7! 0.78~1!

3D3(v50) 18 104.662~1! 0.178 990~4! 0.894~6! 18 104.451~9! 0.173 383~16! 0.81~5! 180623D2(v50) 17 984.747~1! 0.178 520~4! 0.879~5! 17 984.517~1! 0.172 928~7! 0.80~1! 179373D1(v50) 17 873.179~1! 0.177 943~4! 0.854~5! 17 872.943~1! 0.172 381~7! 0.77~1! 17821

3F4(v50) 382.731~1! 0.174 013~4! 0.809~6! 382.766~9! 0.168 574~16! 0.71~5! 2853F3(v50) 188.558~1! 0.174 099~4! 0.880~5! 188.575~1! 0.168 653~7! 0.81~1! 1423F2(v50) 0 0.173 902~4! 0.887~5! 0 0.168 459~7! 0.80~1! 0

aFrom ligand field calculations by Kaledinet al. ~Ref. 9!.

9048 Focsa et al.: The 3D(3d4s)–X3F(3d2) visible system of TiCl1

J. Chem. Phys., Vol. 106, No. 22, 8 June 1997 This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

128.114.34.22 On: Tue, 02 Dec 2014 14:11:59

lowed to record well resolved spectra free from lines of neu-tral species. The relative measurement uncertainty is fourtimes and ten times better than those achieved by Kaledinet al.10 and by Balfour and Chandrasekhar,8 respectively.The high sensitivity of the detection was found to be essen-tial for observing the weak intercombination bands. Theiridentification and analysis allowed the determination of thespin–orbit coupling constants of the two electronic states.Also the present analysis clearly confirms the symmetry ofthe electronic states as determined both theoretically and ex-perimentally by Kaledin and co-workers.9,10

The agreement between our new experimental resultsand the theoretical calculations9 is a good test of the accu-racy of ligand field theory methods which are now widelyused for the determination of electronic energy level diagramof diatomic molecules. Indeed the experimental observationthat the3D state can be considered as free of interaction issupported by the theoretical work by Kaledinet al.9 who didnot calculate any electronic state closer than 3000 cm21 fromthe 3D state. On the contrary we believe that the3F statemight be in interaction with the three electronic states~3S,3P, and3D! resulting from the same 3d2 configuration andwhich are expected9 to lie in the first 2500 cm21 above the3F state. In addition and as explained before, interactionsbetween vibrational levels of the3F state is suspected.

Further experiments will be necessary to clarify theseinteractions. For this purpose it should be possible to use awavelength selected fluorescence excitation technique to evi-dence weak transitions overlapped by stronger ones, as suc-cessfully performed recently on NiF.26 A TiCl1 spectrumwas observed by Balfour and Chandrasekhar8 from the emis-sion of an ac discharge. It should be therefore possible torecord the fluorescence signal through a spectrometer to un-doubtly identify transitions sharing a same upper state and torecord weak bands free of overlapping bands. We note fromthe observation of strong Ti1 atomic lines that it might bepossible to form other Ti-containing ions by introduction of areactive gas into the discharge when there is no chemicalcompound able to directly create ions of interest.

ACKNOWLEDGMENTS

M. Lorenc is gratefully acknowledged for recording partof the spectrum. Financial support of the Centre d’Etudes etde Recherches Lasers et Applications is acknowledged. TheCentre d’Etudes et de Recherches Lasers et Applications issupported by the Ministe`re charge´ de la Recherche, the Re´-gion Nord-Pas de Calais and the Fonds Europe´en de De´vel-oppement Economique des Re´gions.

1T. Oka, inFrontiers in Laser Spectroscopy, Proc. Int. School of Physics‘‘Enrico Fermi,’’ Course CXX, edited by T. W. Ha¨nsch and M. Inguscio~North–Holland, New York, 1994!, pp. 61–87.

2T. R. Geballe and T. Oka, Nature384, 334 ~1996!.3Laboratory and Astronomical High Resolution Data, A.S.P. Conf. Series,edited by A. J. Sauval, R. Blomme, and N. Grevesse~San Francisco, CA,1995!, Vol. 81.

4H. Spinrad and R. F. Wing, Annu. Rev. Astron. Astrophys.7, 249~1969!.5S. R. Langhoff and C. W. Bauschlicher, Jr., Astrophys. J.375, 843~1991!.6V. G. Anicich and W. T. Huntress, Jr., Astrophys. J. Suppl. Ser.62, 553~1986!.

7V. G. Anicich, J. Phys. Chem. Ref. Data22, 1469~1993!.8W. J. Balfour and K. S. Chandrasekhar, J. Mol. Spectrosc.139, 245~1990!.

9L. A. Kaledin, J. E. McCord, and M. C. Heaven, J. Mol. Spectrosc.173,499 ~1995!.

10L. A. Kaledin, J. P. Parrish, and M. C. Heaven, in Proceedings of the 51thSymposium on Molecular Spectroscopy, Columbus, Ohio, 1996~unpub-lished!.

11L. A. Kaledin, A. L. Kaledin, and M. C. Heaven, J. Mol. Spectrosc.179,246 ~1996!.

12C. S. Gudeman, C. C. Martner, and R. J. Saykally, Phys. Rev. Lett.50,727 ~1983!.

13C. S. Gudeman and R. J. Saykally, Annu. Rev. Phys. Chem.35, 387~1984!.

14M. G. Bawendi, B. D. Rehfuss, and T. Oka, J. Chem. Phys.93, 6200~1990!.

15I. Hadj Bachir, T. R. Huet, and J. L. Destombes, J. Mol. Spectrosc.170,601 ~1995!.

16S. Gerstenkorn and P. Luc,Atlas du spectre de l’Iode~CNRS, Paris,1978!.

17D. Uy, C. M. Gabrys, M. F. Jagod, and T. Oka, J. Chem. Phys.100, 6267~1994!.

18J. M. Brown, E. A. Colbourn, J. K. G. Watson, and F. D. Wayne, J. Mol.Spectrosc.74, 294 ~1979!.

19J. M. Brown, A. S.-C. Cheung, and A. J. Merer, J. Mol. Spectrosc.124,475 ~1987!.

20See AIP Document No: PAPS JCPSA-106-9044-28 for 28 pages of tables.Order by PAPS number and journal reference from American Institute ofPhysics, Physics Auxiliary Publication Service, Carolyn Gehlbach, 500Sunnyside Boulevard, Woodbury, NY 11797-2999. Fax: 516-576-2223,e-mail: paps@aip.org. The price is $1.50 for each microfiche~98 pages! or$5.00 for photocopies of up to 30 pages, and $0.15 for each additionalpage over 30 pages. Airmail additional. Make checks payable to theAmerican Institute of Physics.

21G. Herzberg,Spectra of Diatomic Molecules, 2nd ed. ~Van Nostrand,Princeton, 1950!.

22C. H. Townes and A. L. Schawlow,Microwave Spectroscopy~Dover,New York, 1955!.

23C. L. Pekeris, Phys. Rev.45, 98 ~1934!24A. Jenouvier, B. Pascat, and H. Lefebvre-Brion, J. Mol. Spectrosc.45, 46

~1973!.25H. Lefebvre-Brion and R. W. Field,Perturbations in the Spectra of Di-atomic Molecules~Academic, Orlando, 1986!.

26C. Focsa, C. Dufour, and B. Pinchemel, J. Mol. Spectrosc.182, 65 ~1997!.

9049Focsa et al.: The 3D(3d4s)–X3F(3d2) visible system of TiCl1

J. Chem. Phys., Vol. 106, No. 22, 8 June 1997 This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

128.114.34.22 On: Tue, 02 Dec 2014 14:11:59