Interaction of an aluminum atom with a closed subshell metal atom: spectroscopic analysis of AIZn

Resonant two-photon ionization spectroscopy has been employed to investigate diatomic AlZn produced by laser vaporization of a 1:2 Al:Zn alloy target disk in a supersonic expansion of helium. Several discrete transitions are reported in the energy range from 18 400 to 19 100 cm-?. Most of these are assigned as members of the B ?II.?x ?II system, although an isolated band has been observed and assigned as the 2-0 band of the A ?' =0.5?X ?II 1/2 system. A pair of strongly mixed levels are identified as resulting from a homogeneous spin-orbit perturbation between the A ?' =0.5, v' =3 and the B ?II1l2, v' = 1 levels, and the perturbation matrix element has been deduced to be 8.11 cm-? for 27 Al64Zn, 8.23 cm-? for 27 Al66Zn. The ground state has been unambiguously identified as a 2rrr state with a bond length of 2.6957?0.0004 ?. Comparisons to the results of the preceding article on the spectroscopy of AlCa are also provided, along with a discussion of the chemical bonding in AlZn in relation to AlCa, AlAr, and AIKr.

The Journal of AlP Chemical Physics Interaction of an aluminum atom with a closed subshell metal atom: Spectroscopic analysis of AIZn Jane M. Behm, Thorsten Blume, and Michael D. Morse Citation: J. Chem. Phys. 101,5454 (1994); doi: 10.1063/1.467334 View online: http://dx.doi.org/1 0.1 063/1.467334 View Table of Contents: http://jcp.aip.org/resource/1/JCPSA6/v101/i7 Published by the American Institute of Physics. Additional information on J. Chern. Phys. Journal Homepage: http://jcp.aip.org/ Journal Information: http://jcp.aip.org/about/abouCthejournal Top downloads: http://jcp.aip.org/features/most_downloaded Information for Authors: http://jcp.aip.org/authors ADVERTISEMENT ,. AIPAdvances Su m·t OW Explore AlP's new open-access journal Article-level metrics now available • Join the conversation! Rate & comment on articles Downloaded 29 Feb 2012 to 155.97.11.184. Redistribution subject to AlP license or copyright; see htlp:lljcp.aip.org/abouUrights_and_permissions Downloaded 29 Feb 2012 to 155.97.11.184. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/about/rights_and_permissions Interaction of an aluminum atom with a closed subshell metal atom: Spectroscopic analysis of AIZn Jane M. Behm,a) Thorsten Blume,b) and Michael D. Morse Department of Chemistry, University of Utah, Salt Lake City, Utah 84112 (Received 25 March 1994; accepted 24 June 1994) Resonant two-photon ionization spectroscopy has been employed to investigate diatomic AlZn produced by laser vaporization of a 1:2 Al:Zn alloy target disk in a supersonic expansion of helium. Several discrete transitions are reported in the energy range from 18 400 to 19 100 cm -I. Most of these are assigned as members of the B 2rr.-x 2rr system, although an isolated band has been observed and assigned as the 2-0 band of the A !1' = 0 . 5 .-X 2 rr 112 system. A pair of strongly mixed levels are identified as resulting from a homogeneous spin-orbit perturbation between the A !1=0.5, v' =3 and the B 2rr1l2, v' = 1 levels, and the perturbation matrix element has been deduced to be 8.11 cm -I for 27 Al64Zn, 8.23 cm -I for 27 Al66Zn. The ground state has been unambiguously identified as a 2rrr state with a bond length of 2.6957±0.0004 A. Comparisons to the results of the preceding article on the spectroscopy of AlCa are also provided, along with a discussion of the chemical bonding in AlZn in relation to AlCa, AlAr, and AIKr. I. INTRODUCTION The spectroscopic analysis of AIZn presented in this pa­per provides another link in the spectroscopic investigation of the transition metal aluminides, which we began with studies of AICu (Ref. 1) and AINi,2 and have now continued with studies of AlCa (Ref. 3) and AlZn. As stated more ex­plicitly in the preceding paper on the spectroscopy of AlCa,3 the purpose of this series of studies is to investigate the chemical bonding which results between a simple p-block main group element, aluminum, and the 3d series of transi­tion metal atoms. Although the bonding in AlZn and AICa is expected to arise from the same basic type of interaction (a 4s~a,zn atom interacting with a 3pl, atom), the spectra re­corded for the two diatomic molecules are quite different in their details. As will be shown, the ground state bonding in the two diatomic molecules is indeed similar, with the favorable p'TT orientation of the Al atom resulting in 2rrr ground states for both molecules. The differences between the spectra re­corded for the two molecules therefore reflect differences in the accessible excited electronic states. A glance at the ex­cited states of atomic calcium shows that the 4s13d l , 1.3D and 4s 14 pi, 1.3 pO states all lie in the range of 15 000-24 000 cm-I above'the 4s2 , IS ground state, while in zinc the first excited state is 3d104s14pl, 3pO, lying approximately 32500 cm-I above ground state atoms (the isoconfigura­tional 3d104s14pl, IpO state lies even higher in energy, at 46745 cm-I).4 Thus it may be expected that the spectrum of AICa will display a greater number of band systems extend­ing further to the red than does the spectrum of AlZn. This is precisely what is found in the present study; only two dis­tinct excited states of AlZn have been identified, and spectra have only been found in the limited region from 18 400 to 19 100 cm - '. although the regions from 17 500 cm -I to ')Kodak Fellow. b)Present address: Braunschweig Universitat, Germany. nearly 24 000 cm -I and from 26 000 to 29 800 cm -I were carefully searched. From the spectroscopic analysis of AlZn we hope to ob­tain as much information as possible, despite the absence of any prior experimental or theoretical studies on this mol­ecule. Apart from discovering any periodic trends which may exist, the study of the transition metal aluminides is aimed at affording insights into the forces which exist between atoms of these elements, which may in turn prove beneficial in understanding the properties of the bulk alloys. Section II presents a description of the experimental methods employed in this study while Sec. III presents the results obtained for AlZn. In Sec. IV these are interpreted and discussed in relation to the preceding results on AICa. Section V then concludes the article with a summary of our most important findings. II. EXPERIMENT Diatomic AlZn was spectroscopically investigated in a resonant two-photon ionization apparatus that is described in detail in the preceding paper.3 For brevity's sake, only the experimental aspects unique to AIZn will be described here. The 1:2 molar ratio AI:Zn alloy was prepared by heating a mixture of the pure elements in an evacuated quartz tube using a natural gas flame. The metals melted at approxi­mately 600°C, creating a homogeneous metal alloy which was subsequently tUrned on a lathe to produce a flat surface. Although this sample disk was more homogeneous than the AICa metal target disk employed in the previous study,3 it was also somewhat harder, and the AIZn molecular beam was more difficult to produce. A sufficient number of AIZn molecules was generated by using a higher vaporization laser fluence and, more importantly, higher backing pressures of helium (180 psi vs 80 psi). The excitation photon in the two-photon process was supplied by either frequency dou­bling (as described in the preceding paper? the fundamental light produced by a dye laser pumped by the second har­monic radiation (532 nm) of a Nd:YAG laser, or by using the 5454 J. Chem. Phys. 101 (7), 1 October 1994 0021-9606194/101 (7)/5454/1 01$6.00 © 1994 American Institute of Physics Downloaded 29 Feb 2012 to 155.97.11.184. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/about/rights_and_permissions Behm, Blume, and Morse: Spectroscopy of AIZn 5455 fundamental radiation produced by the dye laser pumped by the third harmonic of the Nd:YAG laser (355 nm). The sec­ond, ionizing photon was supplied in all scans by a fixed frequency excimer laser operating on KrF (248 nm, 5.0 eV). High resolution analyses and excited state lifetimes were obtained as described for AICa.3 Since all of the transitions observed in AIZn fell within the limits of the 12 atlas,5 Raman shifting in H2 was not necessary for calibration. The optical spectra of 27 AI64Zn (48.89% natural abundance), 27 Al66Zn (27.81%), 27AI67Zn (4.11%), and 27Al68Zn (18.56%) were recorded by specifically monitoring the ion signal at masses 91,93,94, and 95 as a function of dye laser frequency. III. RESULTS A. Low resolution spectrum Diatomic AIZn was investigated in two energy regions using low resolution (=0.8 cm-I ) resonant two-photon ionization (R2PI) spectroscopy. The ultraviolet energy regime was initially chosen for investigation because the first excited separated atom limit [Al(3s24s l ,2 S 112) +Zn(3dI04s2 , I So)] lies more than 25 000 cm -I above the ground state separated atom limit.4 To investigate this region, the LDS dyes were scanned from 13 000 cm -I to nearly 14 900 cm -I, and were frequency­doubled using an angle-tuned, servo-tracked KD*P doubling crystal to create ultraviolet light in the 26 000-29 800 cm- I range. The AIZn molecules were subjected to both the fun­damental and the ultraviolet radiation, but no electronic tran­sitions were observed. The energy region from 17500 to nearly 24000 cm-I was also scanned, using the fundamental radiation from cou­marin 540A, 500, 480, 460, 440, and stilbene 420 laser dyes for excitation. Several vibronic bands were located in the range from 18 400 to 19 100 cm -I, while base line prevailed elsewhere. Using a rather low vaporization laser fluence a cold spectrum is obtained, as depicted in the lower portion of Fig. I for the predominant isotopic species, 27 AI64Zn. Al­though vibrational progressions are not immediately obvious in this spectrum, the measurement of excited state lifetimes and the rotational analyses of the individual bands demon­strates that the red-shaded bands at 18506, 18891, and 19 083 cm -I correspond to the 0-0, 2-0, and 3-0 bands of a 201/2-201/2 system, which we designate as the B-X sys­tem. Measured upper state lifetimes of these bands are 197:t5 ns (v I =0) and 165:t44 ns (v I =2). Assuming that the decay is governed by fluorescence back to the ground state, this corresponds to a fairly large absorption oscillator strength of 1=0.02, indicative of a spin-allowed transition. The v '=1 level of the B 201/2 state appears to be strongly perturbed by another state, which gains intensity through this interaction, resulting in two strong bands originating from the v"=O level of the X 201/2 state observed at 18690 and 18 707 cm -I. Excluding the 1-0 band from a vibronic fit and including only the 0-0, 2-0, and 3-0 bands which were mea­sured in high resolution, values of w; = 193.54 cm-I and w;x; = 0.28 cm- I are obtained for the B 201/2 state of 27 AI64Zn. In addition to these bands originating from the X 2IT 112 AO=O.S - X2II1f:l: System AO= 1.5 - XlTI312 System (1) ~-'3-17 r- B1 ll311 - XlII3fl System ro:r -- II-I 12-1 B'll,n - X'II,n System . -,~ ~'O~~----~I~I~~----~'2~-O~----"3~ AIl=O.5 - X'll,n System '2-07 '3-07 18300 18500 18700 18900 191vv Frequency (em") FIG. 1. Low resolution resonant two-photon ionization spectra of the A 11=O.5<-X 2II 1I2, B 2II 1I2<-X 2II 1I2, and B 2II312<-X 2II3/2 band systems of jet-cooled 27 AI64Zn, recorded using Coumarin 500 and 540A laser dyes for excitation in conjunction with KrF excimer radiation for photoionization. The upper portion of the figure was collected at higher vaporization powers than the lower portion resulting in a hotter, more con­gested spectrum. state, an additional set of weaker features appear at 18 606, 18802, and 18993 cm-I . These are considerably enhanced in the upper panel of Fig. 1, which displays the low­resolution spectrum collected under conditions employing a higher vaporization laser fluence, resulting in a warmer spec­trum. On the basis of rotationally resolved work reported below, these red-shaded bands are assigned as the 0-0, 1-0, and 2-0 bands of the analogous B 203/2-X 203/2 sub­system. Finally, a strongly blue-shaded feature at 18424 c~ -I in the cold spectrum is assigned as accessing an en­tirely different A' =0.5 excited state located at lower ener­gies, termed the A 0.=0.5 state. The enhancement of the B 20312 - X 2rr 312 bands in the warmer spectrum displayed in the upper portion of Fig. 1 allowed the 0-0, 1-0, and 2-0 bands to be rotationally re­solved, allowing fitted values of w; = 20 1. 9 6 cm -I and w;x; = 2.84 cm-I for the B 203/2 state to be determined. The warmer spectrum also allows hot bands (v"= 1) associ­ated with the B 20 112,3/2-X 20 112 ,312 system to be ob­served, permitting ~G';/2 to be fairly well estimated as 153.4 :to.7 cm- I and 154.5:t2.3 cm- I for the X 201/2 and X 20312 states, respectively. In addition, this study finds the v'-O bands of the B 203/2-X 20312 system to lie on average 101 cm -I higher in frequency than the corresponding bands of the B 20 112-X 2rr 1l2 system, implying an increase in the spin-orbit constant, A, by 101 cm-1 upon electronic excita­tion. Assuming that the ground state spin-orbit constant is similar to that of AICa (65.2 cm-I),3 the excited state spin­orbit constant is approximately 166 cm -I. This large increase in the spin-orbit constant is evidence for a significant con- J. Chern. Phys., Vol. 101, No.7, 1 October 1994 Downloaded 29 Feb 2012 to 155.97.11.184. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/about/rights_and_permissions 5456 Behm, Blume, and Morse: Spectroscopy of AIZn TABLE I. Vibronic bands of 27AI64zn.a System Band Observed frequency (cm-I ) Isotope shift (cm-1)b A 0=0.5<-X 2IIl/2 2-0 18424.2945b -2.8535(38)C,d A 0= 1.5<-X ZII3n(?) 2-0(?) 18634.37 -3,92 0-0 18506.0350b -0.2650(40r,d 2-0 18891.4039b -1.803 7(50)c,d 3-0 19083.237 1b -2,7034 (40)< 0-1 18352.66 0.31 2-1 18739.23 -1.19 3-1 18928.53 -2,16 0-0 18605.8926b -0,5068(46)c.d 1-0 18802.1741b -0,997 9(19)C,d 2-0 18992.7734b -1.85 0-1 18446.88 0.00 1-1 18650.81 -0.13 2-1 18839.78 -1.66 Perturbed C <-X 2 II 112 c-o 18690.2710b -2.9379(37)C,d C-1 18532,30 -1.60 C-2 18383.34 -1.07 Perturbed D <-X 2 II liZ D-O 18 706.533 5b -2.1854(40)C,d D-l 18549.09 -1.60 Lifetime (ns) 197(5)d 165(44)d 247(17)d 301(32)d 259(23)d 'Vibronic bands were fitted to the formula P = Poo + w;v' - w;x;(v'z + v') - v" ~G~12 for v"=O, 1. For the B<-X systems the three bands measured in high resolution were fitted to extract w; and w;x;, giving the values provided in Table III, while all six bands were fitted to extract ~G';/Z . bIsotope shifts are defined as v(Z7 AI66Zn)_v(z7 AJ64Zn). "Measured in high resolution as a band origin, with absolute calibration based on the lz atlas. ~uantities in parentheses represent I u error limits, and correspond to the last digit(s) in the reported values. tribution in the B 2IT state from the 4s 14 p I configuration of Zn (which has an atomic spin-orbit parameter, ~4p, of 386 cm- I).4.6 Finally, it is believed that these conditions have allowed a band of the An' = 1.5 <-X 2IT312 system to be observed at 18 634 em -I. Although this has not been confirmed through rotational analysis because of the weak intensity of the feature, no other reasonable assignment of this undoubt­edly blue-shaded feature is apparent. A listing of all observed vibronic bands is provided in Table I along with their respec­tive isotope shifts and measured excited state lifetimes. shift [v(27 AI66zn)- v(27 A164Zn)= -0.265 em -I] which is characteristic of origin bands. As was expected from the red­shading of this feature in the low resolution spectrum (see the upper panel of Fig. 1), the band exhibits a band head in the R branch. In addition, a weak Q branch is present along with an extended series of P lines. Splitting of the higher P lines is evident in the spectrum, and close inspection reveals that the higher R lines are split as well. This is due to A-doubling, which in principle can exist in both the upper and lower states. The rotational analysis of six bands involving the v"=O level of the X 2ITII2 state and three bands involving the v"=O level of the X 2IT312 state has allowed precise values of B~ = 0.12247 ± 0.000 06 and 0.122 06±0.OOO04 to be de­termined for the X 2IT\12 and X 2IT312 components, respec­tively, for the predominant isotopic modification, 27 AI64Zn. Averaging and inverting these values then results in a ground state bond length of 27 AI64zn of r~ (X 2 IT) = 2.6957 ± 0.0004A. B. Rotationally resolved spectra of the B 2II 1I2<-X 2II112 band system The unperturbed 0-0, 2-0, and 3-0 bands of the B 2IT lI2<-X 2ITII2 band system had sufficient intensity to allow rotationally resolved spectra to be collected and ana­lyzed. Figure 2 displays a high resolution (0.04 em-I) scan over the 0-0 band of the B 2IT 1I2<-X 2ITII2 system, which is similar in structure to the 2-0 and 3-0 bands. The vibrational assignment of the band is confirmed by the small isotope Given that the only reasonable alternatives for the 600 -;;; c:: .~ en + 400 N'" ~ ::; 200 18500 P(J) 15II.5 " " " " lO"S nUl 6I.5 I I I I 1I.5 R(J) mnlll1llJfJlll 0.5 '1'1 Q(o.') 18506 18508 Wavenumber (em") FIG. 2. Rotationally resolved scan over the 0-0 band of the B ZIIlIz<-X ZIIII2 system. The slight bandhead in the R branch indicates that the bond lengthens upon electronic excitation. J. Chem. Phys., Vol. 101, No.7, 1 October 1994 Downloaded 29 Feb 2012 to 155.97.11.184. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/about/rights_and_permissions Behm, Blume, and Morse: Spectroscopy of AIZn 5457 ground state are the 2~,+, 211 112, and 2113/2 states that derive from the ground states of the atoms, the assignment of the system as a 211 1/2 ........ X 211112 transition follows without ques­tion. Assuming that the spin-rotation constants, y, of the upper and lower 2I+ states are small, a 2I+(b) ........ 2!,+(b) assignment may be ruled out by the presence of a Q branch and the requirement of half-integer rotational quantum num­bers in fitting the P, Q, and R structure of the band. Like­wise, the observation of only three main branches instead of the four main branches expected for a 211(a) ........ 2I+(b) or a 2I (b) ........ 211 (a) system rules out the possibility of a transi­tion between 2ll(a) and 2I(b) states. At low rotational lev­els the possibility of a 211( b) state in a heavy molecule such as AIZn may be dismissed because the expected spin-orbit splitting, A, is much larger than the expected rotational con­stant, B, so the only remaining possible assignments for the transition are 2ll 1l2 ........ X 211 112, 211312 ........ X 2113/2, 26.3/2 ........ X 2111/2' and 26.S12+-X 2113/2, All of these are expected to have P, Q, and R branches, but the observation of R(0.5) and P(l.5) in the spectrum implies 0'=0"=0.5, providing a definite as­signment of the band system as 2111/2 ........ X 2111/2. Having determined the nature of the states involved in the transition, the observed rotational lines were then fit to the standard expression7 v= vo+B~,J'(1' + l)+p~,(1' + 112)/2 - B~J" (J" + I) ± p~(1" + 112 )/2, (3.1) where the upper sign corresponds to e levels and the lower to J levels. In the case of AIZn it is impossible to establish the absolute parity of the observed levels with certainty, so the labels a and b are employed instead of e and J. The fit of the data to Eq. (3.1) confirmed that the transition is indeed 0"=0.5+-0"=0.5 and that the bond does lengthen upon ex­citation < r b > r~). The 2-0 band is very similar in appearance to the 0-0 band. In contrast, the data for the 3-0 band was collected under higher temperature source conditions. As a result of the greater population of the higher J levels, the R branch band head and A-doubling are more obvious. Rotational con­stants and band origins of all rotationally analyzed bands are given in Table II for all of the isotopic species resolved. The corresponding rotational line positions are available from the Physics Auxiliary Publication Service (PAPS) of the Ameri­can Institute of Physics8 or from the author (M.D.M.). From a linear fit of the Bb, B~, and B~ constants (excluding the rotational constant of the perturbed 1-0 band), B;<B 2ll l /2) 0.11661 ± 0.000 15 cm- I and a;(B 2111/2) = 0.000 32 ± 0.00007 cm- I were deter­mined for the 27 AI64Zn species. Inverting the value of B;(B 2111/2), a value of r;(B 2ll1/2) = 2.7603 ± 0.00 17 A is obtained. This is longer than the ground state bond length of r~(X 2ll) = 2.6957 ± 0.0004 A, consis­tent with the development of a head in the R branch. C. Rotationally resolved spectra of the B 203/2+-X 203/2 band system The 0-0 band of the B 2113/2+-X 2113/2 band system was rotationally resolved under both hot and cold vaporiza-tion conditions. The colder of the two spectra had an intense Q branch, but the overall lack of intensity in the Rand P branches prevented an accurate analysis. In contrast, the warmer spectrum displayed in Fig. 3 has sufficient assign­able rotational lines for an excellent fit to be obtained. In addition, the Q branch is much weaker than in the cold spec­trum, consistent with the rapid decline in intensity of Q lines with increasing J which is expected for an 0' =312 ........ 0"=312 transition. No A-doubling is evident in the spectrum, even in the highest J lines observed. This is consistent with the fact that in a 2113/2 state the A-doubling is proportional to (1-1/2)(1+ 112)(1+3/2) and is typically much smaller for low J than is found in the corresponding 2111/2 state. (In Van Vleck's pure precession model,7 for ex­ample, the A-doubling in a 2113/2 state is smaller than that found in the 2111/2 state by a factor proportional to B2 / A 2, which of the order of 10-6 in AIZn.) Accordingly, the spec­trum was fitted to the formula (3.2) Similar spectra were recorded for the 1-0 and 2-0 bands of the B 211 3/2 ........ X 2113/2 system, leading to the rotational constants and band origins listed in Table II for the 27 AI64Zn and 27 Al66Zn isotopic modifications. From a linear fit of the Bb, B;, and B~ constants B;(B 211 3/2) = 0.11928 ± 0.000 09 cm-I and a;(B 2113d 0.001 69 ± 0.000 06 cm-1 were determined for the 27 AI64Zn species. Inverting the value of B;(B 2ll3d, a value of r;(B 211 3/2) = 2.7292 ± 0.0010 A is obtained. It is somewhat surprising that the B; values for the B 2111/2 and B 2113/2 states differ so significantly (0.11661 ±O.OOO 15 vs 0.119 28±0.000 09 em-I, respectively). This is unexpected because the spin-orbit splitting of the B 211 state is about 101 em -I greater than that of the ground state (where the B~ values for the X 211112 and X 2II3/2 components are more similar, at 0.12247 ±O.OOO 06 and 0.12206 ±O.OOO 04 em-I), making spin-uncoupling interactions be­tween the B 2111/2 and B 2IT3/2 states too small to account for the differences in B; values. This suggests that one or both of the B 2IT components is perturbed by another state. Addi­tional evidence for such an effect is the difference in vibra­tional constants found for the B 2111/2 and B 2113/2 subs tates, with w; = 193.54, w;x; = 0.28 cm-I obtained for the formerandw; = 20I.96,w;X; = 2.84 cm- I found for the latter. Likewise, the lifetime of the excited B 2113/2 V '=0 level has been measured as 247 ± 17 ns. Although this is not outrageously different from the lifetimes of the B 2111/2 V '=0 and v'=2 levels (l97±5 and 165±44 ns, respectively), the difference appears to be real, again suggesting that one or both components of the B 211 state are perturbed by another state. D. Rotationally resolved spectrum of the A !l=O.5 ........ X 2n1/2 band system A weak blue-shaded band was found at 18 424 cm -I and the lifetime of the upper state of this band was measured to be nearly a microsecond, much longer than that of the B 2n 112 state. This implies that the absorption oscillator J. Chern. Phys., Vol. 101, No.7, 1 October 1994 Downloaded 29 Feb 2012 to 155.97.11.184. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/about/rights_and_permissions 5458 Behm, Blume, and Morse: Spectroscopy of AIZn TABLE n. Fitted rotational constants for 27 Al64zn and 27 AI66zn." Band system Band 27 Al64Zn 27 AI66Zn B 2IIlt2<-X ZIIIt2 0-0 1'0= 18506.0350(28) 1'0= 18505.7700(29) Bo=0.11647(12) Bo=0.11611(15) B~=0.122 44(11) B~=0.121 65(14) Po = -0.00746(51) Po = -0.014 13(58) p~ = 0 (fixed) p~ = 0 (fixed) 2-0 1'0= 18891.4039(32) 1'0= 18889.600 2(38) B~=0.115 72(16) B~=0.113 95(22) B;=0.12257(17) B~=0.121 03(22) p~ = -0.00783(119) p~ = o (fixed) 3-0 1'0= 19 083.2371(27) 1'0= 19080.5337(29) B~=0.115 62(23) B~=0.114 36(18) B~=0.123 06(20) B~=0.121 71(16) p~ = -0.00433(34) p~ = -0.00437(41) p~ = 0 (fixed) p3 = o (fixed) B 2II312<-X 2II312 0-0 1'0= 18605.892 6(28) 1'0= 18605.3858(37) Bo=0.11877(7) Bo=0.11856(13) B~=0.121 94(7) B~=0.121 15(12) 1-0 1'0=18802.1741(15) 1'0=18 801.176 2(11) B;=0.11640(5) B;=0.11522(4) B~=0.122 10(5) B~=0.120 94(4) 2-0 1'0= 18 992.773 4(19) B~=0.115 88(12) B3=0.12218(13) A 0.5<-X 2IIlI2 2-0 1'0= 18424.2945(28) 1'0= 18 421.441 0(26) B~=0.141 43(20) B~ =0.140 80(18) B~=0.121 43(23) B~=0.121 06(21) p~ = -0.041 34(77) p~ = -0.039 96(63) p~ = o (fixed) p~ = o (fixed) Perturbed pair 18690 cm- I band 1'0= 18690.271 0(26) 1'0 = 18 687.333 1(27) of bands B'=0.12987(17) B'=0.13136(31) Bo=0.12236(16) B3=0.12177(41) p'=-0.02849(61) P' = -0.029 40(77) p~ = o (fixed) Po = o (fixed) 18707 cm- I band 1'0= 18706.5335(26) 1'0= 18704.348 1(31) B' =0.127 10(20) B'=0.12382(20) B~=0.122 74(17) Bo=0.121 16(25) P' = -0.02293(5) P' = -0.01898(59) p~ = o (fixed) p~ = o (fixed) "All constants reported in wave numbers (em-I) with lcr error limits in parentheses. strength of this transition is at least an order of magnitude reduced from that of the B 2rr 112 state, indicating that the transition may be spin-forbidden. The blue-shaded nature of the band signifies that the bond shortens upon electronic ex- 800 - 1 600 0; ~ 400 ~ 200 Wavenumber (em I) FIG. 3. Rotationally resolved spectrum of the 0-0 band of the B 2Tl 312<-X 2II312 band system. This spectrum was collected under hotter source conditions than that of Fig. 2 and thus exhibits many more rotational lines. citation, a conclusion that is substantiated by the presence of a bandhead in the P branch in the rotationally resolved spec­trum displayed in Fig. 4. In addition, A-doubling is immedi­ately apparent in this band, where it is evident in the first R line and in the large separation between the heads of the two P branches. The presence of R(0.5) and P(1.5) again identi­fies the transition as an 0 1 = 1I2~0" = 112 band, and a least squares fit of the observed rotational lines to Eq. (3.1) pro­vides the rotational constants listed in Table II for the 27 Al64Zn and 27 Al66Zn species. In addition, the upper state rotational constant, B ~, may be inverted to provide r ~ = 2.5064 ± 0.0018 A, which represents a considerable shortening of the bond relative to the ground state [r~(X 2rr=2.6957±0.0004 A)]. Evidently this band repre­sents excitation to a new 0 1 =0.5 state, which we designate as the A 0=0.5 state. Further, as will be proven in Sec. III G below, the measured isotope shift e7 A166zn_ 27 AI64Zn) of J. Chern. Phys., Vol. 101, No.7, 1 October 1994 Downloaded 29 Feb 2012 to 155.97.11.184. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/about/rights_and_permissions Behm, Blume, and Morse: Spectroscopy of AIZn 5459 -r---y-----,----,---,----,----, 800 - .~ 600 V) 400 - 200 - 0- 18423 18425 18427 Wavenumber (em·l ) FIG. 4. Rotationally resolved spectrum of the 2-0 band of the A 0=0.5.-X ~n112 band system. The sudden band head in the P-branch indicates that the bond shortens substantially upon electronic excitation, and a fit of the rotational lines provides r~ = 2.5064 :t 0.0018 A. The A-doubling of the rotational lines is apparent at very low values of J. -2.85 cm- J suggests a tentative assignment of the 18424 cm- 1 feature as the 2-0 band of the A 0=0.5--X 2n 1l2 system. E. Rotationally resolved spectra of the perturbed 18690 and 18707 cm-1 bands The measured lifetimes of the features at 18 690 and 18707 cm- 1 are found to be 301±32 and 259±23 ns, re­spectively. Given that the 1-0 band of the B 2n \/2<-X 2n 112 system is expected in this region, and its lifetime would be expected to be approximately 200 ns, the appearance of two bands in this region with lengthened upper state lifetimes suggests that the B 20112, v' = 1 level is perturbed strongly by another state which has a poorer radiative coupling to the ground state. In such a coupling mechanism the dark state gains oscillator strength, leading to a second vibronic feature in the spectrum. An obvious candidate for such a state is another vibronic level of the A 0=0.5 state. To investigate this possibility further, the two bands at 18 690 and 18 707 cm - J were examined in high resolution. A rotationally resolved scan over the 18690 cm- J band is displayed in Fig. 5, revealing a nearly symmetric rotational structure. Lambda doubling is also evident in both the P and R branches, even at low J -values. The obvious presence of R(0.5) and PO.5) immediately identifies the band as an 0' = 112<-0"= 112 band, as expected. A similar rotationally resolved scan over the 18707 cm - J band is displayed in Fig. 6, again showing a nearly symmetric rotational structure without a strong tendency to form a head in either the P or the R branch. Lambda dou­bling is again evident in the P and R branches, even at low J -values. In its general appearance the 18 707 cm -\ band is quite similar to the 18 690 cm -\ band. Both bands have been analyzed as 0' =0.5--0" =0.5 transitions, with the resulting rotational constants reported in Table II. The upper state rotational constants (B') and A-doubling parameters (P') for both bands are nearly iden- P(J) Q(n R(J) 600 lUll 1 11111 II I 1111111111111111 55 I S II S 4 . .5 7.S 1400 oJfl + ~ 18694 Wavenumber (em·l ) FIG. 5. Rotationally resolved scan over the 18690 cm- I band of 27 AI64Zn. As demonstrated in the text, the upper state of this transition is an approxi­mately 54:46 mixture of the A 0=0.5, v'=3 and B 2rr l12, v'=1 states. These states are mixed by a homogeneous spin-orbit perturbation with a perturbation matrix element of approximately 8 cm-I . tical and are almost precisely an average of the correspond­ing values of the B 201/2 v' =0 and the A 0=0.5 v' =2 states. This strongly suggests that vibrational levels of these two states are nearly completely mixed by a homogeneous perturbation, which is probably due to the spin-orbit inter­action. Furthermore, when one state perturbs another, the en­ergy of one of the states is shifted by an amount that is equal and opposite in direction to the shift of the other state. Using this knowledge it is possible to predict the unperturbed fre­quency of the A 0=0.5<-X 20112 3-0 band, from which a value of w; of =279 cm -1 may be estimated for the A 0 =0.5 state. This deperturbation analysis is discussed further in Sec. III G below. F. A-doubling A proper analysis of the A-doubling in the upper and lower states observed in AIZn requires careful thought be­cause our experimental resolution (0.04 cm-I ) makes it dif­ficult to determine the A-doubling parameters accurately. A 400,-,----.----,----,----,----,---, Q(U.l) P(I) R(J) IN 1111111 I n 1\11111 \I I ~ \ LS 1\111111111111 o.~ ~ ~ 6. ~ 300 - " 200- r ,~ o -1~87~04~~~~1~87-06---J--~18~7~08-~--~18J710 Wavenumber (em I) FIG. 6. Rotationally resolved scan of the 18 707 cm -I band. This band also results from the perturbation between the A 0=0.5, v I =3 and B 2n1l2• v I = 1 states, which are almost completely mixed by a homogeneous spin­orbit perturbation. J. Chern. Phys., Vol. 101, No.7, 1 October 1994 Downloaded 29 Feb 2012 to 155.97.11.184. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/about/rights_and_permissions 5460 Behm, Blume, and Morse: Spectroscopy of AIZn major problem in this regard is the fact that the splittings in the individual lines were only measured reliably in the P branch for the B 2IT1/2<--X 2ITI/2 system (due to the band head in the R branch), and only in the R branch for the A 0=0.5<--X 2ITII2 system (due to the band head in the P branch). In such cases it is difficult if not impossible to in­dependently determine P I and p", because the splitting in the R or P branch grows linearly in i", and is proportional to (p' -p")J". Thus, despite our observation of A-doubling in all of the observed bands involving 0= 112 states, it is only in the perturbed 18690 and 18707 cm- I bands, where A-doubling is cleanly observed in both the Rand P branches, that reliable values of p I and p" may be derived. Using Eq. (3.1) to analyze these perturbed bands demon­strated that the A-doubling parameter, Po' of the X 2II112 ground state is small and poorly determined, with fitted val­ues of -0.00077(80) and -0.00270(391) cm-I for the 18 690 and 18 707 cm -I bands of 27 AI64Zn, and -0.00245(162) and -0.00581(518) cm- I for the corre­sponding bands of 27 Al66Zn (Here the 1a error limits are given in parentheses.) Given these small and indeterminate fitted values of Po, the choice was made to constrain Po to zero in all of the band analyses reported here. It is perhaps not surprising that p" is small; the only state close in energy to the X 2IT ground state (and therefore the major cause of A-doubling in the X 2IT ground state) is the 2~+ state deriv­ing from ground state atoms. The small value of Po obtained here suggests a rather large splitting between the X 2IT state and the 2~ + state, at least at the ground state internuclear separation of 2.6957 A. Holding Po equal to zero led to a negligible deterioration in the quality of the fit of the rota­tional lines, and reduced the uncertainties in the reported upper state A-doubling parameters significantly. Another difficulty in this analysis is the determination of the absolute parity of the A-doublet levels. In the case of AlZn, the major contributor to the A-doubling of the ground X 2ITr state is its interaction with the 2l: + state arising from the same ground separated atom limit. Since this 2l: + state lies above the X 2IIr state, and we expect the matrix ele­ments a+=( 1Tlat+la) and b=( 1Tlt+la) to have the same sign, one would expect the A-doubling parameter Po to be negative, with I levels lying below e levels.7 If the magni­tude of this parameter had been well-determined in the fits of the rotational bands, this would have allowed a determina­tion of the proper e,f labeling scheme. Unfortunately this is impossible at the present level of resolution, so the more conservative a,b labeling scheme is used instead. With the value of Po constrained to be zero, fits of the observed bands yield fairly small A-doubling parameters for the vibronic levels of ·the B 2II112 state of p~ =-0.00746(51) and p~ -0.00433(34) cm-I for 27AI64Zn and p~ -0.014 13(58), p~ =-0.00783(119), and p~ -0.00437(41) cm- I for 27 AI66Zn. The fitted values of p~ (obtained holding Po = 0) for the perturbed 18690 and 18707 cm-I bands are much larger in absolute magnitude, giving -0.02849(61) and -0.02293(55) cm-I for 27AI64Zn, and -0.02940(77) and -0.01898(59) cm -I for 27 A166Zn, respectively. Finally, the fitted value of p~ obtained from the rotationally resolved scan over the 2-0 band of the A 0=0.5<--X 2II112 system at 18 634 cm -I is even larger in magnitude, with p ~ =- 0.04134(77) for 27AI64Zn and p~ = -0.03996(63) cm -I for 27 A166Zn. Again, 1 a error limits are provided in parentheses. Although these values are undoubtedly not accurate be­cause of experimental difficulties in resolving all of the lines and our constraint of Po to zero, we feel confident in the trend that is established. The A 0=0.5 state of AlZn displays a larger A-doubling than does the B 2ITII2 state, which in tum displays a larger A-doubling than does the ground state. Moreover, the A-doubling of the perturbed pair of levels at 18690 and 18707 cm-1 is intermediate in magnitude to that found for the A 0=0.5, u:=2 level and the various levels of the B 2IT 112 state. This supports our interpretation of these levels as arising from a perturbation between the u I = I level of the B 2II112 state and the u I =3 level of the A 0==0.5 state. Furthermore, if we estimate the unperturbed value of p; (B 2IT 112) by interpolating between the values given above, values of -0.00642 and -0.010 98 cm- I are ob­tained for 27 AI64Zn and 27 A166Zn, respectively. If these are averaged with the corresponding value for the u I =3 level of the A 0:=0.5 state (estimated as equal to that of the A 0 =0.5, u I =2 level) we can obtain estimates of the average of p I values for the perturbed pair of states. The fact that these averaged values (-0.02388 and -0.02547 cm- I for 27 AI64Zn and 27 A166Zn, respectively) are in close agreement with the average of the p t values actually found for the two perturbed bands (-0.02571 and -0.02419 cm- I for 27 AI64Zn and 27 A166Zn, respectively) substantiates the assign­ment of the perturbation as A 0=0.5, u' = 3 - B 2II 112 , U ' := 1. Furthermore, it also establishes that the sign of p t (which is undetermined in our present study because of the difficulty in establishing the absolute parity of the levels) is the same in both the A 0=0.5 and B 2III/2 states. Thus in both states either the I levels lie below the e levels or vice versa, but in any event the ordering of the ell levels is the same in both states. G. Deperturbation analysis of the 18 690 and 18 707 cm-1 bands The deperturbation of the 18 690 and 18 707 cm -I bands is greatly facilitated by the high resolution analyses of the B 2IT1/2<--X 2II1/2 0-0, 2-0, and 3-0 bands for both the 27 AI64Zn and the 27 AI66Zn isotopic modifications. By fitting the band origins of these bands to the formula (3.3) one may obtain accurate values of voo' w;, and w;x; , which in tum may be used to predict the frequency at which the 1-0 band of the B 2 II 112 <--X 2 IT 112 transition would lie if the perturbation were absent. Through this procedure the 18 706.534 (18 704.348) cm -I band for 27 Al64Zn CZ7 A166Zn) is found to lie 7.531 cm-I (6.335 cm-I) above the predicted frequency of the B 2 II 112 <--X 2 IT 112 1-0 band. This further implies that the 18690.271 (18687.333) cm-I band has been shifted lower in frequency as a result of the perturba­tion, assuming that only two states are involved. This finally allows the energies of the deperturbed levels to be fixed as J. Chern. Phys., Vol. 101, No.7, 1 October 1994 Downloaded 29 Feb 2012 to 155.97.11.184. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/about/rights_and_permissions Behm, Blume, and Morse: Spectroscopy of AIZn 5461 18720 18710 i3 ~ r 18700 18690 18680 AO'.O.S, v'=3 Slate ------'" \ ---', '\ -------' .... \"'\ ",'--- ... ,------- "'\~~.---- Deperturbed FIG. 7. Plot of the effect of the homogeneous spin-orbit perturbation be­tween the A 0=0.5, v' =3 and the B 2nl12, v' = 1 states illustrating the effect on the measured isotope shifts. The 27 AI64Zn energy levels are drawn as solid lines while the 27 AI66Zn energy levels are depicted with dashed lines. Both are measured relative to the X 2n1l2• v"=O level of the corre­sponding species. The measured energy levels are in the central column and the deperturbed levels are displayed in the left and right columns. The per­turbation causes the measured isotope shift for the 18 690 em -I band to be reduced and that of the 18 707 cm -I band to be increased as compared to their unperturbed values. 18697.802 (for the A 11=0.5, v' state) and 18699.003 cm- I (for the B 2nw, v'=1 state) above the v"=O level of the X 2nw state for 27 AI64Zn (and 18693.668 and 18698.013 cm - I, respecti vel y, above the v" = 0 level of the X 2n 112 state for 27 AI66zn). A pictorial representation of both the actual and deperturbed levels is displayed in Fig. 7. These deperturbed energies are satisfying because the resulting isotope shifts [11(27 AI64Zn) - v(27 A166Zn)] for the v'-O bands of the B 2n l12+-x 2n ll2 system now fall in line as 0-0, 0.265 cm- I ; 1-0, 0.990 cm- I (deperturbed); 2-0, 1.804 cm- I ; 3-0, 2.703 cm- I . Assuming that the perturber is a vibrational level of the A 11=0.5 state, it is also satisfying that the isotope shifts of the band at 18 424.294 cm -I and the second deperturbed band now correspond to 2.854 and 4.134 cm -I, respectively. Assuming that these two bands corre­spond to sequential vibrational levels of the A 11=0.5 state, the isotope shift equation relating isotopic species i and j, v~'oo- v{,'oo= [(v' + 1I2)W;i- w~/2][ 1-(t-tJ t-tY12 ] - [(v' + 1I2)2w;ix;i- w~ix~/4] X[I-(t-tJt-tj)] (3.4) may be used to establish the vibrational numbering of the A 11=0.5 state unambiguously. Using estimated values of w; = 279 cm- 1 (based on the difference in frequency be­tween the A 11=0.5, v' =2 and the deperturbed A 11==0.5, v'=3 levels), an estimated value of w;x; = 0.8 cm- I , w~ == 155 cm -I (as determined from a vibrational hot band analysis of the B 2n 1l2+-X 2n ll2 band system), and w~x; = I cm-1 (as a reasonable guess), we find expected isotope shifts [11(27 AI66Zn) - 11(27 AI64Zn)] for bands of the A 11=0.5.-X 2n l12 band system of -0.280 cm-1 (0-0 band), -1.522 cm- I (I-O band), -2.750 cm-1 (2-0 band), -3.964 cm- I (3-0 band), and -5.164 cm- I (4-0 band). Clearly, the observed bands, with isotope shifts of -2.854 and -4.134 cm- I (deperturbed) correspond to the 2-0 and 3-0 bands of the A 11==0.5+-X 2n ll2 system, respectively, and the perturbation is between the A 11=0.5, v' =3 and the B 2n 112, v' = 1 levels. Knowing the deperturbed energy levels as well as the final observed levels, one may calculate the magnitude of the perturbation matrix element which is responsible for their mixing. This may be done independently for 27 AI64Zn and 27 A166Zn, yielding values of 8.109 and 8.226 cm -1 for H 12' respectively. The close agreement between these two values provides validation for the deperturbation method, particu­larly in the use of the two-state model. This perturbation matrix element, in tum, has two components, a vibrational factor (approximated by the Franck-Condon overlap of the A 11=0.5 v' =3 and B 2nw, v' = 1 vibrational wave func­tions) and an electronic matrix element. An approximate Franck-Condon calculation provides a value of the vibra­tional factor (B 2n 112,V' = llA 11 = 0.5,v' = 3) of 0.25, im­plying that the electronic matrix element coupling the B 2nll2 and A 11=0.5 states is roughly 33 cm- I . Combined with the long lifetime of the A 0=0.5, v' =2 level (which may be limited by predissociation rather than fluorescence), our inability to observe other vibrational levels of the A 11 ==0.5 state strongly suggests that the A 11=0.5 state is pri­marily quartet (S = 3/2) in character. Spin-orbit selection rules governing the coupling of a 2n state to a quartet state then require that the A 11=0.5 state must be a 4I112, 4n1l2, or 4AW state.7 Having deduced the energies of the deperturbed energy levels and the magnitude of the perturbation matrix element, it now becomes possible to determine the degree to which the A 0=0.5, v' =3 and B 2nl /2, v' = 1 states are mixed in the upper levels of the perturbed bands. For the 27 AI64Zn isotopic species the upper state of the 18 690 cm - I band is found to have 54% A 11==0.5 character and 46% B 2nl/2 character while for the 27 Al66Zn isotopic modification the upper state of the 18687 cm-1 band has 63% A 11=0.5 char­acter and 37% B 2nl/2 character. These percentages are ex­actly reversed when one considers the higher frequency, 18 707 cm -I band, as is required for the two state perturba­tion problem. The fact that the A and B states are less thor­oughly mixed in the 27 Al66Zn species is a simple result of the fact that the deperturbed levels are further apart in energy in this species, as is shown in Fig. 7. The nearly 50-50 mix of A and B state character nicely explains why the fitted rotational constants (both B' and p ') of the perturbed excited states lie so close to the average of the A 11=0.5, v' =2 and B 2n1/2, v' =0 values. In fact, using the percent A and B character listed above and the estimated B values of the deperturbed A 11=0.5, v' ==3 and B 2n1/2, v' == 1 levels [taken as equal to the measured value of B;(A 11 = 0.5) and the average of the Bo(B 2n 1l2) and B;(B 2n 1l2) values, respectively], it is possible to predict the B values expected for the upper states of the 18 690 and 18 707 cm-l bands. This procedure gives predictions of 0.12972 and 0.12785 cm-l for the upper state of the 18690 and 18707 cm-1 bands in 27AI64Zn (compared to measured values of 0.12987 and 0.127 10 cm- I , respectively) and 0.13133 and 0.12481 cm- I in 27A166Zn (compared to mea- J. Chern. Phys., Vol. 101, No.7, 1 October 1994 Downloaded 29 Feb 2012 to 155.97.11.184. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/about/rights_and_permissions 5462 Behm, Blume, and Morse: Spectroscopy of AIZn TABLE Ill. Electronic states of 27AJMzn.a State To (em-I) We (em-I) w.x. (em-I) B. (em-I) a e (em-I) r. (A) T (ns) B 2TI312 18605.8926(28)+ x 201.96 2.84 0.11928 (9) 0.00169(6) 2.729(1) 247 (17) B 2TII12 18 506.035 0(28) 193.54 0.28 0.II66I(15) 0.000 32(7) 2.760(2) 197 (5) A 0=0.5 -17870b AGs12=273.5 B2==0.14143 (20) 2.506 (2)C 977 (244) X 2TI312 x AGI12 ==154.5 (2.3) Bo==0.12206 (4) 2.6957 (4)d X 2TII12 0.000 AG 112= 153.4 (0.7) Bo=0.12247 (6) alu error limits are provided in parentheses and correspond to the last digit(s) in the corresponding values. bCaleulated assuming w.=279 em-I and w.x.=0.8 em-I. "This value is r 2, not r •. dThis value is an average of the ro values obtained for the X 2TII12 and X 2TI312 states. sured values of 0.13136 and 0.12382 cm-I, respectively). The close agreement of these results substantiates the valid­ity of the deperturbation analysis. With the completion of this analysis, the known electronic states of Z7 AI64Zn are summarized in Table III. H. Ionization energy of AIZn In using resonant two-photon ionization to probe the spectroscopy of AlZn the A O=0.5 ....... X zIT 1I2 2-0 transi­tion at 18424 cm-1 (2.284 eV) was observed using KrF excimer radiation (4.997 eV) for photoionization. This places the ionization energy of AlZn below 7.281 eV. In tum, the fact that KrF radiation is suitable for the second, ionizing photon sets the lower limit on the ionization energy of AlZn as 4.997<IE(AIZn). Studies with ArF excimer radiation have further indicated that AlZn is one photon ionized at this wavelength, thereby loosely bracketing the ionization energy of AlZn as 4.997<IE(AlZn)<6.42 eY. IV. DISCUSSION The present gas phase analysis has unambiguously de­termined the ground state of AlZn as X 2IT 112, with a ground state electronic configuration of 3 si!4s~n 'IT!, 2IT 112, where the 'IT orbital is primarily 3 p 'IT AI in character, although some 4P'ITzn character is undoubtedly present as well. As was found from the theoretical studies reported for AICa in the preceding paper,3 it is likely that there is some partial trans­fer of 4szn electron density into the empty 3pa AI orbital, although electronegativity differences make this dative bond­ing interaction much less favorable in AlZn than in AICa. For this reason we believe AlZn is probably less strongly bound than is AICa. This is consistent with the greater difficulty in preparing a molecular beam of AlZn as compared to AICa. In comparing AIZn to AICa one is struck by the substan­tial difference in bond lengths of these isovalent molecules. The bond length of ground state AlZn (2.6957±0.0004 A) is 0.452 A shorter than that of AICa (3.1479±0.001O A),3 a result that is directly related to the decrease in atomic radius in going from calcium (rCa= 1.97 A) to zinc (rZn= 1.33 A).9 This decrease in atomic radius results from the substantial contraction of the 4s orbital as one moves across the 3d series, which may be quantified by the radial expectation values (r4s) calculated in a numerical Dirac-Fock method by Desclaux to be (r 4S) Ca = 2.22 A and (r 4s >Zn = 1.51 A.1O Like the closed subshell Zn and Ca atoms, the rare gases Ar and Kr also combine with Al to form complexes with 2IT1/2 ground statesY-!3 The rare gas species, however, have considerably longer ground state bond lengths because the bonding in these species is dominated by dispersion forces. As a result the AlAr and AIKr molecules may be classified as van der Waals complexes. The equilibrium internuclear sepa­ration of AIAr,!2 for example, has been experimentally deter­mined as r~ = 3.79 ± 0.01 A while that of AIKr (Ref. 13) is r~ = 3.84 ± 0.01 A. We may use these values to test for the existence of chemical bonding in AlZn. Assuming additivity of bond lengths in van der Waals complexes, the bond lengths of Arz (r~ = 3.80 A),!4 AlAr (r~ = 3.79 A),12 and ZnAr(r~ = 4.18 A),15 may be used to establish van der Waals radii of 1.90, 1.89, and 2.28 A for Ar, AI, and Zn, respectively. If the AlZn molecule were bonded primarily by van der Waals forces one would then expect a bond length of roughly 4.17 A. The large discrepancy be­tween this value and the measured bond length of 2.69 A demonstrates convincingly that there is indeed some type of chemical bonding occurring in diatomic AlZn. This probably results from a weak a donation from the filled 4s orbital of zinc into the empty 3pa orbital of aluminum, with concomi­tant weak back-donation of 3p'IT electrons from aluminum into the 4p'IT orbital of zinc. The nature of the chemical bonding in the excited B 2IT state may be considered by first establishing the excited separated atom limits that can generate zIT states. The lowest limits capable of generating such states are the A13s23d1 , zV+Zn4sz, IS limit at 32436 cm-I , the Al3s23pl, 2pO+Zn 4S14pl, 3pO limit at =32500 cm-I , and the Al3sz4pl, zpo+Zn 4s2, IS limit at =32950 cm-I . Given that the B 2IT state of AlZn correlates to one of these sepa­rated atom limits, the observation of the B - X 0-0 band at 18 506 cm -I implies that the bond strength of the excited B 2IT state exceeds that of the X zIT ground state by at least 13 930 cm-I (32436-18506 cm- I ). Given this fact, it seems most likely that the B state correlates to the Al3sz3pl, 2p o+Zn4s14pl, 3pO limit, which opens the 4s subshell of zinc permitting strong a-bonding to occur. The A13s24pl, 2p O+Zn 4s2, IS limit may certainly be excluded from consideration, because the zIT state resulting from this limit corresponds simply to promotion of the weakly bond­ing 'IT electron to a Rydberg 'IT orbital which is certainly nonbonding in character. Likewise it is difficult to imagine how at least 13930 cm- I (1.73 eV) of additional bonding could be obtained from an Al3sz3dl, 2V+Zn 4s2 , IS inter­action, since the 3d'ITAI orbital would be at best only slightly J. Chern. Phys., Vol. 101, No.7, 1 October 1994 Downloaded 29 Feb 2012 to 155.97.11.184. Redistribution subject to AIP license or copyright; see http://jcp.aip.org/about/rights_and_permissions Behm, Blume, and Morse: Spectroscopy of AIZn 5463 more capable of a bonding interaction with the empty 4p'lT orbital of zinc than the 3 p 'IT Al orbital. On this basis we cor­relate the B 20 state with the A13s23pl, 2pO+Zn 4S14pl, 3 pO separated atom limit. Low-lying quartet states that are candidates for the A 0 =0.5 state of AIZn must derive from either the A13s 13p2, 4P+Zn 4s2, IS limit at =29070 cm- I or from the Al 3s23pl, 2pO+Zn 4S14pl, 3pO limit at =32500 cm- I. The former limit leads to 4II and 4I- states while the latter gives 4I + (2), 4I -, 4II(2), and 4~ states, all of which have the requisite symmetry to undergo spin-orbit perturbation with the B 2II state. Again, the A state must be strongly bound, with a bond strength at least 11 000 cm- I (29070-18000) (1.36 e V) greater than the ground state. As in the B 20 state, this greatly enhanced bond strength arises because a hole has been opened in a filled s orbital. In this case it could be either the 3s Al or the 4sZn orbital that has been opened, in either case resulting in greatly improved O'-bonding. v. CONCLUSION The present spectroscopic investigation of AIZn provides the first study, spectroscopic or otherwise, to be performed on this diatomic species. High resolution resonant two­photon ionization spectroscopy has allowed the ground state of the molecule to be unambiguously determined as X 20112, similar to the ground states of the diatomic aluminide spe­cies, AICa,3 AIAr,"·l2 and AIKr. 12,13 These studies have con­firmed the preferential orientation of the Al 3 p electron to be 3 p 'IT in the presence of another filled subshell atom. In the present study of AIZn two electronic states, labeled the A 0 =0.5 and B 20 states, have been characterized and a homo­geneous spin-orbit perturbation between two vibrational lev­els of these states has been analyzed. A weighted least squares value for the ground state bond length of AIZn has been determined as r~(X 2II) = 2.6957 ± 0.0004 A. The bonding in AIZn appears to be substantially differ­ent than that proposed for AICu (Ref. 1) and AINi,2 where covalent interaction between the 3pO' electron of aluminum and a lone 4s electron of the transition metal leads to a strong two-electron if bond. In contrast, it is rather similar to the bonding in AlCa.J It is hoped that knowledge gained through the study of AIZn and AlCa (Ref. 3) will allow us to rationalize the bonding interactions which occur in other transition metal aluminides, particularly species such as AISc and AIMn. The scandium and manganese transition metal partners in these diatomic transition metal aluminides pos­sess large 3d;.-4s promotion energies which may obstruct the formation of a two electron 3pO' AI-4sO'M if bond, lead-ing to a preferential p'lT orientation as the aluminum atom approaches the scandium or manganese atom. Further studies are currently underway to test this hypothesis. ACKNOWLEDGMENTS We thank Professor William H. Breckenridge for the use of the intracavity etalon employed in the high resolution studies and for a number of stimulating conversations re­garding AIZn and its comparison to AlAr and AIKr. We also thank Caleb A. Arrington for his expert preparation of the AIZn alloy required for these studies and John G. Kaup for his help in the evaluation of Franck -Condon factors in the B 20, v' = 1 ~ A 0=0.5, v' =3 perturbation matrix element. Research support from the National Science Foundation un­der Grant No. CHE-9215193 is gratefully acknowledged. Acknowledgement is also made to the Eastman Kodak Com­pany for a fellowship, and to the donors of the Petroleum Research Fund, administered by the American Chemical So­ciety, for partial support of this research. 1 J. M. Behm, C. A. Arrington, J. D. Langenberg, and M. D. Morse, J. Chern. Phys. 99, 6394 (1993). 21. M. Behm, C. A. Arrington, and M. D. Morse, J. Chern. Phys. 99, 6409 (1993). 3 J. M. Behm, M. D. Morse, A. L Boldyrev, and 1. Simons, J. Chern. Phys. 101, 5441 (1994). 4C. E. Moore, Natl. Bur. Stand. Circ. 467 (1971). 5 S. Gerstenkorn and P. Luc, Atlas du Spectre d'Absorption de la Molecule d'/ode (CNRS, Paris, 1978); S. Gerstenkorn and P. Luc, Rev. Phys. Appl. 14,791 (1979). 6E. U. Condon and G. H. Shonley, The Theory of Atomic Spectra (Cam­bridge University, Cambridge, 1970). 7H. Lefebvre-Brion and R. W. Field, Perturbations in the Spectra of Di­atomic Molecules (Academic, Orlando, 1986). 8 See AlP document No. PAPS JCPSA-lO 1-5454-3 for 3 pages of absolute line positions. Order by PAPS number and journal reference from Ameri­can Institute of Physics, Physics Auxiliary Publication Service, Carolyn Gehlbach, 500 Sunnyside Boulevard, Woodbury, New York 11797-2999. 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 additional page over 30 pages. Airmail additional. Make checks payable to the American Institute of Physics. 9R. W. G. Wyckoff, Crystal Structures, 2nd ed. (Interscience, New York, 1963). 10 J. P. Desclaux, At. Data Nucl. Data 12, 311 (1973). 11 C. L. Callender, S. A. Mitchell, and P. A. Hackett, J. Chern. Phys. 90, 5252 (1989). 12M. J. McQuaid, J. L. Gole, and M. C. Heaven, J. Chern. Phys. 92, 2733 (1990). 13Z._w. Fu, S. Massick, J. G. Kaup, O. B. d'Azy, and W. H. Breckenridge, J. Chern. Phys. 97, 1683 (1992). 14D. E. Woon, J. Chern. Phys. 100, 2838 (1994). 15 I. A. Wallace, R. R. Bennett, and W. H. Breckenridge, Chern. Phys. Lett. 153, 127 (1988). J. Chern. Phys., Vol. 101, No.7, 1 October 1994