Resonant two-photon ionization spectroscopy of jet-cooled Pt?

The gas phase optical spectrum of jet-cooled Pt? has been investigated over the range of 11 300 to 26 300 cm -1 using resonant two-photon ionization spectroscopy in combination with time-of-flight mass spectrometry. Numerous vibronic bands are observed. Analysis of the data gives the location of some 26 excited electronic states, which are characterized by the frequencies of their origin bands, vibrational frequencies, and anharmonicities. Variation of the second color in a two-color resonant two-photon ionization scheme has determined the ionization threshold of Pt? to be 8.68 ?0.02 eV. The observation of the onset of predissociation, characterized by a sharp drop in excited state lifetime, places the dissociation energy of Pt? at 3.14?0.02 eV. In combination with the Pt atomic ionization potential of 8.8?0.2 eV, these results give the bond strength of Pt?+ as D0(Pt--Pt+) = 3.26 ? 0.24 eV. The strength of the chemical bond in Pt?, as compared to Au?, demonstrates that there are significant 5d contributions to the chemical bonding in Pt?.

Resonant two-photon ionization spectroscopy of jet-cooled Pt2 Scott Taylor, George W. Lemire, Yoon Mi Hamrick, Zhenwen Fu, and Michael D. Morse Department of Chemistry, University of Utah, Salt Lake City, Utah 84112 (Received 2 June 1988; accepted 25 July 1988) The gas phase optical spectrum of jet-cooled Pt2 has been investigated over the range of 11 300 to 26 300 cm -I using resonant two-photon ionization spectroscopy in combination with time­of- flight mass spectrometry. Numerous vibronic bands are observed. Analysis of the data gives the location of some 26 excited electronic states, which are characterized by the frequencies of their origin bands, vibrational frequencies, and anharmonicities. Variation of the second color in a two-color resonant two-photon ionization scheme has determined the ionization threshold ofPt2 to be 8.68 ± 0.02 eV. The observation of the onset of pre dissociation, characterized by a sharp drop in excited state lifetime, places the dissociation energy ofPt2 at 3.14 ± 0.02 eV. In combination with the Pt atomic ionization potential of 8.8 ± 0.2 eV, these results give the bond strength ofPt2+ as Do(Pt - Pt+) = 3.26 ± 0.24 eV. The strength of the chemical bond in Pt2, as compared to Au2, demonstrates that there are significant 5d contributions to the chemical bonding in Pt2. I. INTRODUCTION Transition metal clusters form a class of chemical com­pounds which are of interest from both the experimental and theoretical points of view. The recent increase of interest in these species derives in part from advances in experimental methods, which have enabled ligand-free transition metal clusters to be produced in the gas phase, where they can be cooled to low internal energies. Once such species have been formed and cooled, their electronic and molecular architec­ture can be probed in detail using various elegant spectrosco­pies, including laser-induced fluorescence (LIP), 1-4 reso­nant two-photon ionization spectroscopy (R2PI), 5-7 resonant two-photon dissociation spectroscopy (R2PD), 8 and photoelectron spectroscopy (PES).9-11 Alternatively, the reactivity of cold transition metal clusters can be investi­gated, potentially providing insights into the mechanisms of heterogeneous catalysis on transition metal surfaces. 12-16 In other studies, cold transition metal cluster ions can be accel­erated to well-defined kinetic energies and collided with a variety of reaction partners, thereby providing much-needed thermochemical and kinetic information. 17-20 Advances in theoretical methods, and the availability of supercomputers have also contributed greatly to the in­creased interest and understanding of transition metal clus­ters. Most of the sophisticated calculations of small transi­tion metal molecules reported today would have been impossible a mere decade ago. As a result of these rapid improvements in theoretical methodology and computer power, and the simultaneous advances in experimental tech­niques, our knowledge and understanding oftransition met­al clusters is growing steadily. In view of the intrinsic diffi­culties associated with both the experimental and theoretical approaches to the complicated transition metal clusters, both disciplines are necessary for progress in understanding these interesting yet problematic molecules. The current state of our knowledge of the transition metal molecules has been reviewed in several recent publications? 1-27 In this paper we report the results of spectroscopic in­vestigations of the platinum dimer, Pt2. Along with its con­geners Pd2 and Ni2, this molecule is expected to possess a single sigma bond, with the remaining valence electrons ar­ranged in two nd 9 cores. Among this series of molecules, the strength of the chemical bond and the splittings between the various angular momentum couplings of the nd 9 cores are sensitive measures of the detailed electronic structure. In contrast to the nickel dimer, d-orbital participation in the chemical bonding of Pd2 and Pt2 is expected to be of some importance because the nd contraction is less severe in the 4d and 5d transition metals than it is in the 3d metals.22 On the other hand, the low bond strength of Pd2, as measured by Knudsen effusion mass spectrometry,28.29 has been rationa­lized22 as arising from the high promotion energy (1.90 eV)30 required to excite two ground state (4d 10/SO) Pd atoms to the 4d 95s1 configuration, which is required for for­mation of a sigma bond. In addition to these promotion ener­gy and nd contraction effects, chemical bonding in the Ni2 series is influenced by relativistic effects and spin-orbit in­teraction, which are both quite large in Pt2. Finally, correla­tion and exchange effects are quite important for an accurate description of the electronic structure in the transition metal dimers. The theoretician who investigates the electronic struc­ture and chemical bonding in Ni2 , Pd2 , and Pt2 faces a severe challenge: in order to accurately calculate these open d-shell transition metal dimers he must include all of the effects mentioned in the previous paragraph. Moreover, all of these important effects must be calculated to a comparable level of accuracy, since the balance between them may decide the electronic structure of the molecule. One motivation for in­vestigating Ni2,7 Ptz (and eventually Pdz) is to provide the necessary data for testing the various theoretical methods which are applied to these molecules. Section II of this paper deals with the experimental aspects of this study, which was conducted using resonant J. Chem. Phys. 89 (9), 1 November 1988 0021-9606/88/215517-07$02.10 © 1988 American Institute of PhySiCS 5517 Downloaded 19 Dec 2001 to 128.110.196.147. Redistribution subject to AlP license or copyright, see http://ojps.aip.org/jcpo/jcpcr.js,, 5518 Taylor et a/.: Spectroscopy of Pt2 two-photon ionization spectroscopy applied to the jet­cooled platinum dimer. Section III presents our findings and relates them to previous work. In Sec. IV the results are discussed in relation to the electronic structure of Pt2, and a summary of our conclusions is given in Sec. V. II. EXPERIMENTAL The molecular beam apparatus employed in our labora­tory is similar to that used in previous investigations. 16 ,31 A pulsed supersonic beam of platinum clusters is produced by laser vaporization of a platinum disk (Johnson Matthey Ae­sar, 0,5 mm thick, 99,99% purity) located in the throat ofa pulsed supersonic nozzle using the second harmonic light (532 nm, 100 mJ/pulse) ofa Q-switched Nd:YAG laser (Quantel, model 580-10), The vaporization laser is timed to fire at the peak density of helium carrier gas, which is pulsed over the target by a homemade magnetically operated dou­ble solenoid valve, The high pressure of helium (8 atm in the reservoir behind the pulsed valve) in the 2 mm diameter nozzle throat insures enough three-body collisions to quench the plasma and form the platinum clusters, Cluster formation, up to Pt 12 , is facilitated by attaching a face plate and extender to the nozzle, In order to prevent the drilling of a hole through the platinum disk, the disk is continuously rotated and translated with a mechanized assembly similar to that described in Ref, 32, Following cluster formation, the clusters are entrained in the carrier gas and expand freely into a vacuum chamber maintained at 1 X 10-3 Torr. The resulting supersonic ex­pansion then cools the internal degrees of freedom of the nascent clusters, The supersonic jet is collimated by a conical skimmer and subsequently enters a second vacuum chamber, which houses the ionization region of a homemade reflectron-type time-of-f1ight mass spectrometer. Ionization of the clusters is accomplished by resonant two-photon ioni­zation, following which the ions are accelerated up a 0,9 m flight tube, A reflecting electric field assembly is mounted at the apex of this flight tube, thereby reflecting the ions 18° down a second flight tube, A microchannel plate detector (Galileo, model 3025-B) located at the end of the second flight tube collects the signal, which is then amplified (Paci­fic Instruments model 2A50), digitized by a transient re­corder (Transiac, model 2001), and signal averaged by a DEC 11/73 microcomputer, The entire experiment operates under the control of this microcomputer. Resonant two-photon ionization spectra ofPtz were ob­tained using a dye laser (Molectron, model DL-II) pumped by the second or third harmonic of a Q-switched Nd:YAG laser (Quantel, model 581-C) to electronically excite the platinum dimer. Following excitation, photoionization was achieved by subjecting the excited molecules to the output of an excimer laser (Questek, model 2420) operating on either the ArF (193 nm, 6.42 eV) or Fz (157 nm, 7.89 eV) transi­tions, In certain regions of the spectrum, frequency sum­ming was required to generate the appropriate excitation wavelength. This was accomplished by combining the fun­damental of the Nd:YAG laser, 1064 nm, with the funda­mental of the appropriate dye in a 58° KD*P Type II fre­quency summing crystal. The crystal was angle tuned using a homemade autotracking device. In the course of this inves­tigation the dye laser was operated using all of the following dyes: stilbene 420, coumarins 440, 460, 480, 500, and 540A, fluorescein 548, rhodamines 590, 610, and 640, DCM, and LDS dyes 698, 750, 751, 821, and 867 (all obtained from Exciton), III. RESULTS A. Vibronic spectra of Pt2 Our investigations of the jet-cooled platinum dimer re­veal a large number ofvibronic bands throughout the visible, extending into the near infrared and ultraviolet (II 300- 26 300 cm - I ), In this respect the spectrum is reminiscent of that obtained for diatomic nickeV which also possesses a very dense optical spectrum. In contrast to previous observa­tions reported for Ni2, however, the vibronic spectrum ofPt:> displays recognizable, well-behaved vibrational progres­sions over most of the range studied, By carefully examining the spectrum ofPt2, paying particular attention to the inten­sity patterns and isotopic shifts of the vibronic features, we have been able to identify 26 distinct band systems which are characterized in Table I by the frequency of the origin band voo, the vibrational frequency of the upper state w~, and its anharmonicity w;x;, Figure I displays a small part of the optical spectrum of Pt2, taken from the low-frequency por­tion of our scan. In this spectrum specific bands from Sys­tems I-III are identified. It is evident from Fig. I that these progressions are quite regular, and this is borne out by the small error limits derived from the least-squares fits, which are listed in Table I. In addition to locating some 26 distinct band systems, we have attempted to find hot bands, which would have en­abled us to deduce the vibrational frequency of the ground electronic state. Despite careful examination of our spectra, no assignable hot bands were observed, Evidently, the excit­ed vibrational levels of Pt2 are effectively quenched in these experiments. Additional evidence of this is provided in Sec. III B below. Having identified 26 distinct band systems in Ptl , a question naturally arises: Do all of these band systems arise from the ground electronic state, or are metastable elec­tronic states ofPtz present in our molecular beam? Although one cannot be absolutely definite on this point, we believe that all of the band systems listed in Table I originate from the ground electronic state ofPtz. For these experiments our source was configured with a face plate and extender, as described in Sec. II. With this assembly, the helium and nas­cent clusters flow downstream from the point of vaporiza­tion through a 2 mm diameter channel 15 mm in length. At the end of this channel, the gases enter a channel 4 mm in diameter, 10 mm in length. Finally, this channel expands to 6 mm in diameter for another 5 mm in length, following which a final orifice of about 1.5 mm diameter is encoun­tered. The original purpose of this arrangement was to intro­duce turbulence and lengthen the amount oftime prior to the final expansion into vacuum, so that Pt2 could be generated in greater quantity. In addition, however, this design is quite effective in quenching the internal degrees of freedom ofPt2 . J. Chem. Phys., Vol. 89, No.9, 1 November 1988 Downloaded 19 Dec 2001 to 128.110.196.147. Redistribution subject to AlP license or copyright, see http://ojps.aip.org/jcpo/jcpcr.jsp Taylor et al.: Spectroscopy of Pt2 5519 TABLE I. Band systems of Pt2 . Band system v(K'( observed)a.h v(X,(fitted)' liJ: w:x;, (6)d 11426.4 11 426.58 ± 0.35 191.81 ± 0.39 0.79 ± 0.06 II (5) 11 649.4 11649.35 ± 0.21 187.96 ± 0.31 -0.13 ± 0.06 III (4) 11767.8 11 767.83 ± 0.28 170.69 ± 0.60 0.27 ± 0.15 IV (4) 11816.3 11816.45 ± 0.78 193.98 ± 1.65 1.95 ±0.40 V (4) 12507.8 12507.89 ± 0.37 185.91 ± 0.78 - 0.48 ± 0.19 VI (4) 12921.4 12921.19±0.72 194.61 ± 1.51 0.12 ± 0.37 VII (4) 13 089.2 13 089.41 ± 0.72 182.94 ± 1.74 0.25 ± 0.43 VIII (5) 13 255.5 13255.35 ± 0.29 187.21 ± 0.43 0.21 ± 0.08 IX (4) 13 408.6 13 408.47 ± 0.39 186.86 ± 0.82 0.60 ± 0.20 X (5) 13 516.9 13516.45 ± 0.79 178.38 ± 1.15 0.95 ± 0.22 XI (7) 14041.7 14040.80 ± 0.88 159.90 ± 0.80 - 1.09 ± 0.11 XII (5) 14501.7 14501.62 ± 1.19 146.73 ± 1.74 0.16 ± 0.34 XIII (4) 15297.1 15296.94 ± 0.78 163.52 ± 1.65 - 1.50 ± 0.40 XIV (4) 15367.4 15367.23 ± 0.52 145.68 ± 1.10 1.20 ± 0.27 XV (4) 15386.2 15386.20 ± 0.04 153.06 ± 0.09 1.10 ± 0.02 XVI (4) 15954.5 15954.45 ± 0.13 157.12±0.28 0.15 ± 0.07 XVII (4) 16003.4 16003.61 ± 1.11 146.63 ± 2.34 - 1.18 ± 0.57 XVIII (4) 17913.6 17 913.66 ± 1.47 171.33 ± 1.64 - 0.08 ± 0.27 XIX (4) 18345.9 18345.82 ± 0.20 191.93 ± 0.41 2.42 ± 0.10 XX (4) 19751.6 19751.81 ± 0.81 151.71 ± 1.69 - 0.58 ± 0.41 XXI (4) 20318.0 20317.96 ± 0.02 148.63 ± 0.05 0.37 ± 0.01 XXII (4) 20374.5 20374.20 ± 1.46 163.79 ± 3.07 0.08 ± 0.75 XXIII (8) 20954.0 20953.11 ± 0.98 167.49 ± 0.74 2.79 ± 0.09 XXIV (3)' 21015.7 21015.7 145.20 1.65 XXV (5) 22730.8 22730.56 ± 0.66 140.78 ± 0.96 -0.47 ±0.19 XXVI (6) 22762.7 22 762.85 ± 0.82 140.33 ± 0.91 1.08 ± 0.15 a All numerical values are in wave numbers. b Absolute band frequencies are measured using the 5d "6S(4 F) 6p (5 D ~ ) ~ 5d 96sc'D3) Pt atomic transition at 30 157.0 em - I as an internal calibration, and are expected to be accurate to within ± 5 em - I. 'The band origin V,)(j, fundamental frequency w;, and the anharmonicity w;.x;. of the upper electronic states were obtained by fitting a linear least-squares treatment of v = Voo + w;v + w;x; ( - v - v2). dQuantities in parentheses are the number of band members identified for that system. 'Since only three bands were observed for System XXIV, V,lO ' w;, and w;x; were uniquely determined, and no least-squares error estimate can be given. The uncertainty in the reported spectroscopic constants is expected to be comparable to that found for the other band systems. The case of sonic flow provides a lower bound on the amount of time spent in this high-pressure region of about 30 f.ls. In this period of time, a Pt2 molecule can be expected to suffer some 8 X 105 hard-sphere collisions with helium (assuming a Ptz Vibranic Spectrum System I 800 10-0 /1-0 12-0 13-0 14-0 5-0 I ~stom II l- 10-0 /1-0 12-0 13-0 4-0 I 0 c Slstem III \ 0-0 \ 1-0 \ 2-0 \ 3-0 (J) U; 600 I- - c 0 400 I- - + +'-"' Q. 200 l. ! J Il J ~I d o 11400 11600 11800 12000 12200 12400 Wavenumber (cm -1) FIG. I. Resonant two-photon ionization spectrum of Ptz, obtained using the dyes LDS 867 and 821 to excite the molecule and the F2 excimer laser transition (157 nm) as the ionization photon. The three lowest frequency band systems which we have observed are indicated in the figure, and their v.:., w;, and w;x; values are given in Table I. Members of band systems IV and V are also present in the figure, but these have not been labeled. local pressure of 2 atm). This is evidently sufficient to cool the vibrational motion ofPt2 quite effectively, and is expect­ed to cool the electronic degrees of freedom as well. Further cooling then occurs during the supersonic expansion into vacuum. Prior to this work matrix isolation studies of diatomic platinum had been conducted by Pellin,33 and by Jansson and Scullman.34 Pellin has observed numerous vibronic pro­gressions in the 600-850 nm region by dispersed fluores­cence studies of matrix isolated Pt2 . 33 Jansson and Scullman have identified a particularly simple single progression in an annealed argon matrix with Voo = 11 247.8 cm - I, w; = 218.1 cm- I , and w;x; = 0.9 cm- I .34 This progression appears to be quite isolated in the matrix study,34 whereas our data show numerous vibronic systems in this spectral region. Moreover, the value of w; observed in the matrix investigation is 12% higher than that obtained for any of the band systems we have observed. A matrix shift of the magni­tude is quite uncommon for vibrational frequencies, except for vibrational motions of weakly bound van der Waals mol­ecules. 35 Consequently, we do not believe that any of our band systems correlate with that observed in the matrix. The availability of near-infrared laser dyes has limited our inves­tigation to laser frequencies above 11 300 cm - I; we suspect the intense band system observed by Jansson and Scullman lies to the red of this limit in the gas phase. J. Chem. Phys., Vol. 89, No. 9,1 November 1988 Downloaded 19 Dec 2001 to 128.110.196.147. Redistribution subject to AlP license or copyright, see hUp:/Iojps.aip.org/jcpo/jcpcr.jsp 5520 Taylor et a/.: Spectroscopy of Pt2 B. Dissociation energy of Pt2 The ground electronic state of atomic platinum is 5d 9CZDs/2 )6s, 3D3,3° so the lowest dissociation threshold of Pt2 corresponds to production of two 3 D3 platinum atoms. From this limit 28 distinct molecular electronic states arise [in Hund's case (c), which is certainly appropriate for Pt2 l.36 Some of these electronic states spin-pair both 6s elec­trons in a sigma-bonding orbital and are strongly bound, possibly correlating to the ground state. Other states arising from the 3 D3 + 3 D3 separated atom limit couple the 6s elec­trons to a spin of 1, resulting in repulsive potential energy curves. In addition to these 28 electronic states correlating to the 3 D3 + 3 D3 separated atom limit, 42 electronic states cor­relate to 3D3 + 3D2 separated atoms (0.096 eV above 3 D3 + 3 D3 atoms), 72 electronic states correlate to 3 D3 + 3 F4 separated atoms (0.102 e V above 3 D3 + 3 D3 atoms), and 15 electronic states correlate to 3 D2 + 3 D2 separated atoms (0.192 e V above 3 D3 + 3 D3 atoms). 36 As one approaches the dissociation limit of Pt2, the number of distinct electronic states becomes very large indeed. With 157 molecular states present within 0.2 eV of the lowest dissociation threshold, it is likely that perturbations between these states will be significant. Moreover, as soon as the lowest dissociation limit is exceeded, perturbations between these states will lead inevitably to predissociation. Thus for Pt2, predissociation is likely to occur as soon as the lowest dissociation limit is exceeded. The occurrence of pre­dissociation above a certain excited state energy will mani­fest itself by an abrupt drop in excited state lifetime. In the case of Nil the excited state lifetime drops from about 10 I1S to IOns over a range of energy of about 20 cm ~ I; this has allowed an extraordinarily precise determination of the bond strength of the dinickel molecule.7 Using time-delayed resonant two-photon ionization techniques we have measured the lifetimes of many excited states of Pt2, particularly toward the high-frequency end of our scans. Figure 2 displays a typical lifetime plot for the 23 391 cm - I absorption band of Pt2 , which has been least- PI, lifelime (23,391 em- I band) 8000 -;;; 6000 "00 '" .":: 4000 + c:. 2000 0 -100 0 100 200 300 Dye-Exeimer Time Delay/(IO nanoseconds) FIG. 2. Lifetime of the upper state of the S--O band of System XXVI ofPt" measured by time-delayed resonant two-photon ionization methods. In this plot the dye laser excites the Pt, molecule at t = 0, and the ArF excimer laser is fired after the indicated delay. The solid curve is a least-squares fit to the data, which gives a lifetime of SIO ± 6 ns for this state of Pt,. Pt Dimer Lifetimes vs Wavenumbcrs 1000,-------------------------------, 100 or. s (l) f .5 ~ 10 H ;J £ l+-------,-------~----------------~ 24700 25000 25300 25600 25900 Wavenumber (em-I) FIG. 3. Lifetimes of successive vibronic levels near the predissociation threshold of Pt" measured by time-delayed resonant two-photon ioniza­tion. At the predissociation threshold the measured lifetime drops from 214.S ± SO.8 ns at 25313.3 cm- I to 30.0 ± 4.8 ns at 25359.0 em-I. To the blue of this latter frequency, no transitions are observed with lifetimes greater than 30 ns, while no transitions to the red of this limit have lifetimes below 140 ns. For this figure, short lifetimes were least-squares fitted to a convolution of a Gaussian laser time profile and an exponential molecular decay function. This was not necessary for the long lifetimes ( T > lOOns). squares fit to an exponential decay with a lifetime of 51 0 ± 6 ns. Figure 3 shows the measured lifetimes for the upper states of more than 20 transitions plotted as a function of excitation energy. An abrupt drop in excited state lifetime is clearly observed at 25 336.2 ± 23 cm ~ I. All bands to the blue of this limit exhibit lifetimes below 30 ns, while all bands to the red possess lifetimes longer than 140 ns. Although this abrupt drop in lifetime is not as dramatic as that observed for Niz, it nevertheless presents strong evidence for a predisso­ciation threshold at this energy. Accordingly, we assign Do(Pt-Pt) as 3.14 ± 0.02 eV. The observation of a sharp predissociation threshold provides additional evidence that the diplatinum molecules probed in these experiments are vibrationally and electroni­cally cold. If the molecular beam contained a significant population of vibration ally or electronically excited Pt2 mol­ecules, these species would exhibit a predissociation thresh­old shifted to lower frequencies by the amount of internal energy present. If this were the case we would expect to ob­serve an energy range containing both long-lived and short­lived bands. There is no evidence in Fig. 3 of such a range, providing confirmation that the observed band systems all originate from the vibration less level of the ground elec­tronic state. In previous work, Pt2 has been investigated by Knudsen effusion mass spectrometry. Although an early investiga­tion37 failed to detect Pt2 , Gupta, Nappi, and Gingerich were subsequently successful in detecting Pt2 in equilibrium with atomic platinum over the temperature range 2259-2736 K. 3K A second-law evaluation provided DO (Pt2 ) = 3.71 ± 0.61 eV, while a third-law estimate gave DO(Pt2 ) = 3.71 ± 0.16 eV.38 Both estimates exceed our measurement by 0.57 eV, although our result is within the error limits of the second­law evaluation. As has been discussed elsewhere,22 the third­law method relies on the statistical mechanical expression J. Chem. Phys., Vol. 89, No.9, 1 November 1988 Downloaded 19 Dec 2001 to 128.110.196.147. Redistribution subject to AlP license or copyright, see http://ojps.aip.org/jcpo/jcpcr.jsp Taylor et al.: Spectroscopy of Pt2 5521 for the equilibrium constant of the reaction Pt2 <22Pt, in combination with measured values of this equilibrium con­stant, to derive a value of Do (Pt2 ). To apply this method the partition function of the diatomic molecule must be calculat­ed, and this requires knowledge of the bond length, vibra­tional frequency, and electronic degeneracy of all low-lying electronic states. Gupta, Nappi, and Gingerich assumed only one low-lying electronic state, a 1 ~ ground state with re = 2.34 A. and We = 259.4 cm - J for their third-law esti-mate. 38 We have repeated their calculation using the low­lying electronic states reported in Table VI of Balasubra­manian's paper.36 In combination with the experimental values of Kp for the reaction Pt2<22pe8 this gives DO(Pt2 ) 3.10 ± 0.06 e V if a symmetry number of 1 is used for Pt2, or DO(Pt2 ) = 3.26 ± 0.06 eV if a symmetry number of 2 is adopted. Rigorously, a symmetry number of2 is only appro­priate for truly homonuclear dimers, such as J95Pt2, while a symmetry number of 1 is appropriate for heteronuclear spe­cies such as J95pe96Pt. Since 29.17% of all Pt2 molecules are truly homonuclear, a weighted average of these results is probably most appropriate, giving DO(Pt2 ) = 3.15 ± 0.06 eV. This is in excellent agreement with out result, which gives DO(Pt2 ) = 3.14 ± 0.02 eV. Both values are in fair agreement with empirical estimates by Krasnov39 (3.43 ± 0.48 eV) and Miedema and Gingerich40 (2.88 ± 0.26 eV) as well. C. Ionization potential of Pt2 In the resonant two-photon ionization spectroscopy ex­periment, as conducted here, the first photon is absorbed from a scanning dye laser, and the second photon is absorbed from a fixed-frequency excimer laser, thereby photoionizing the molecule of interest. As the dye laser is scanned to lower energies, eventually the sum of the two-photon energies will drop below the energy required to ionize the molecule, and no more transitions will be observed. Of course, the failure to Pt2 Vibronic Spectra 400 r--,------,------,------,------,--, ;;; c co Vi 300 c 200 .8 +N ,;: lOO ArF Photo ionization F 2 Photoionization 1 7900 18100 18300 18500 18700 Wavenumber (em'!) FIG. 4. Determination of the ionization threshold ofPt2. In the upper spec­trum, radiation from an ArF excimer laser (6.42 eV) is employed as the ionization photon in the R2PI scheme. The lower spectrum is obtained us­ing F2 excimer radiation (7.9 eV) as the ionization photon. The last band observed using the ArF radiation occurs at 18 202.0 em - " placing the ioni­zation potential ofPt2 at 8.68 ± 0.02 eV. Both spectra were obtained using coumarin 540A, pumped by the third harmonic ofa Nd:Y AG laser, to gen­erate the resonant photon in the R2PI process. observe transitions does not prove that the sum of the two photon energies is below the ionization limit; it is possible that there are simply no absorptions in this spectral region. By repeating the scan with a higher energy ionization laser, however, one may confirm that transitions are indeed pres­ent, but cannot be observed with the less energetic ionization laser. Figure 4 displays the results of just such an experiment. The upper scan shows the vibronic spectrum of Pt2 in the 17 800-18 800 cm - J range, obtained using ArF excimer ra­diation (193 nm, 6.42 eV) as the ionization laser. The lower scan shows the same spectrum obtained using F 2 radiation (157 nm, 7.90 eV) as the second color. Vibronic transitions are apparent throughout the region using the F2 laser, but these become weaker and eventually disappear as the dye laser is scanned to the red using ArF radiation. Presumably the decrease in intensity arises because there are poor Franck-Condon factors for production of ground state, v = 0 ions from the intermediate states accessed in this ener­gy range. Based on the last observed transition, located at 18 202.0 cm - J, we assign the ionization potential of Pt2 as 8.68 ± 0.02 eV. Having measured both the bond strength and ionization potential ofPt2, it is a simple matter to complete the thermo­dynamic cycle and obtain the bond strength ofPtt , through the relation Do(Pt-Pt+) = Do(Pt-Pt) + IP(Pt) - IP(Pt2 )· (3.1 ) In addition to our measured values of Do(Pt-Pt) and IP (Pt2 ), we also need the ionization potential of atomic plati­num to complete this thermodynamic cycle. In Moore's ta­bles30 the atomic ionization potential of platinum is given as 9.0 eV, but this number is derived by fitting the 5d 96seD3 ) and 5d 97 s( 3 D3 ) levels to the Rydberg formula, and correct­ing by the percent error expected, based on other elements which are more precisely known. Since the 5d 96s (3 D3 ) level is the ground state of the platinum atom, the basis for this extrapolation is tenuous at best. In other work, an electron impact study finds an appearance potential of Pt+ to be 8.82 ± 0.04 eV, however, this shifts to 8.61 eV when correc­tions are applied to account for electron impact processes which produce excited states of Pt+ .41 In view of these re­sults, we adopt the value IP (Pt) = 8.80 ± 0.20 eV, from which we derive Do(Pt-Pt+) = 3.26 ± 0.24 eV. This pre­sents the interesting and unusual case of a dimer for which the ionization potential is more accurately known than is that of the constituent atoms. IV. DISCUSSION A major unresolved question in the chemical bonding between transition metal atoms concerns the degree of d­electron participation in the bonding. Little is experimental­ly known about such effects, but the theoretical expectations are clear22: (1) d-electron participation in the chemical bonding will become less significant as one increases the atomic number within a transition series, because the addi­tional nuclear charge causes the d orbitals to contract. The outer s orbital is shielded more effectively from this in- J. Chem. Phys., Vol. 89, No.9, 1 November 1988 Downloaded 19 Dec 2001 to 128.110.196.147. Redistribution subject to AlP license or copyright, see http://ojps.aip.org/jcpo/jcpcr.jsp 5522 Taylor et at.: Spectroscopy of Pt, creased nuclear charge, and is less susceptible to contraction. (2) d-electron participation is expected to be more signifi­cant in the 4d and Sd transition metal series, because these orbitals suffer less contraction (relative to the Ss and 6s, respectively) than does the 3d orbital, in part due to relativ­istic effects. As a result of these effects, the 3d contribution to the chemical bonding in Ni2 is essentially nil. The bond strength of Ni2(2.068 ± 0.01 eV) 7 is nearly identical to that of Cu2(2.01 ± 0.08 eV).22 Likewise, Ni2 and CU2 have almost identical bond lengths (2.20 7 and 2.219S ).,42,43 respective­ly). The situation is quite different among the Sd transition metal dimers, however. Our measured bond strength of Pt2(3.14 ± 0.02 eV) is substantially stronger than that mea­sured for Au2(2.29 ± 0.02 eV),44 indicating significant par­ticipation of the Sd electrons in the chemical bonding ofPt2 • In an early ab initio investigation of Ni2, Pd2, and Pt2, Basch, Cohen, and Topiol found little or no d orbital contri­bution to the bonding in Ni2 and Pd2, but substantial Sd contributions in Pt2.45 These authors used a relativistic effec­tive core potential to limit the number of electrons actively considered, and then applied a multiconfiguration self-con­sistent- field treatment to several representative states arising from the (6sag )2Sd ~ Sd ~ electronic configuration of Pt2 • Correlation of the bonding (a) electrons was built into the calculation, but other correlations among the open d-shell electrons were omitted. Spin-orbit effects were also ignored. These limits to the calculation cause a drastic underestimate of the Pt2 bond strength, predicting Do (Pt2 ) = 0.93 e V for a 1 r state which places both d-shell holes in 0 orbitals. 45 g More recently, Balasubramanian has investigated Pt2 using a relativistic core potential method to describe the [Xe] 4P4 cores, along with a complete active space self-con­sistent- field treatment of the remaining 20 electrons (d 9sl _ d 9SI). 36 Following the CASSCF calculation, additional cor­relation effects were included by a first-order configuration interaction treatment, after which the spin-orbit interaction was introduced. This monumental effort represents the most detailed theoretical investigation of an open Sd-shell transi­tion metal dimer reported to date. Despite this effort, a bond strength of2.31 eV is obtained ignoring spin-orbit coupling. With spin-orbit coupling included this drops to 1.97 eV, and completely rearranges the energy ordering among the elec­tronic states. Evidently, considerable additional electronic correlation is required to bring the calculated value of DO(Pt2 ) into agreement with experiment. Although the high value of DO(Pt2) provides the stron­gest evidence of significant d-orbital participation in the chemical bonding of Pt2, additional support for this idea may be found in the vibronic spectrum itself. If the d orbitals were contracted so tightly that they did not take part in the chemical bonding, one would expect very slight differences in the orbital energies of the da, d1T, and do orbitals. On the other hand, these orbital energies would be expected to be widely separated if there were significant d-d interactions between the two metal atoms. Diatomic nickel conforms to the former expectation, resulting in an extremely congested vibronic spectrum with spectral features occurring with an average spacing of 10 cm -1.7 In Pt2, however, the average spacing between vibronic features is perhaps 40 or SO cm - I, resulting in a spectrum which is interpretable. The decreased density of electronic states in Pt2 arises in part because the d­orbital splittings in this species are significantly greater than in Nil; in addition, the greater spin-orbit splitting in Pt2 contributes to a decreased electronic state density as well. V. CONCLUSION A resonant two-photon ionization spectroscopic inves­tigation of jet-cooled Pt2 has revealed numerous vibronic bands throughout the visible, extending into the near in­frared and ultraviolet regions of the spectrum. Although the spectrum is dense, 26 vibronic progressions have been identi­fied and characterized by voo, w;, and w;x;. The onset of predissociation, inferred from an abrupt drop in excited state lifetime, has been used to determine Do(Ptz) = 3.14 ± 0.02 eV. This is consistent with high-temperature Knudsen effu­sion studies, provided the partition function ofPt1 is reevalu­ated using ab initio results for the low-lying electronic states of the dimer. In addition to these measurements, a comparison of the optical spectra generated using ArF and Fz excimer radi­ation as the ionization photon has permitted a measurement of the ionization potential of Pt2, which is determined to be 8.68 ± 0.02 eV. In combination with the measured bond strength of Pt2 and the ionization potential of atomic plati­num, this places the bond strength ofPt2+ at 3.26 ± 0.24 eV. Finally, we note that the bond strength of Pt2 is signifi­cantly greater than that of Au2, demonstrating the impor­tance of Sd contributions to the chemical bond in this spe­cies. Attempts are currently in progress to extend this inves­tigation to the congeneric Pd2 molecule, which is expected to be much less strongly bound. In addition, efforts are current­ly under way to construct a single-mode, transform-limited pulsed dye laser, which should enable the investigation ofthe observed spectra with rotational resolution, ACKNOWLEDGMENTS We are grateful to Professor Edward M. Eyring for the gift of some platinum foil, which was used in this work. We also gratefully acknowledge research support from the Na­tional Science Foundation, under Grant No. CHE-8S- 21OS0. IJ. L. Gole. J, H, English, and V. E, Bondybey, J. Phys, Chern. 86, 2560 (1982), 2D. R, Preuss, S, A. Pace, and J, L. Gole, J, Chern, Phys. 71, 3553 (1979), 'V, E. Bondybey, G, P. Schwartz, and J, H, English, J, Chern, Phys. 78, II (1983). 4E. A, Rohlfing and J. J. Valentini, J, Chern, Phys. 84, 6560 (1986). 'D, E. Powers, S, G, Hansen, M. E, Geusic, A, C. Puiu, J. B. Hopkins, T. G, Dietz, M, A. Duncan, p, R. R, Langridge-Smith, and R, E, Smalley, J. Chern, Phys, 86, 2556 (1982). 6D. E. Powers, S, G. Hansen, M. E. Geusic, D, L. Michalopoulos, and R. E. Smalley, J, Chern, Phys. 78, 2866 (1983). J. Chem. Phys., Vol. 89, No.9, 1 November 1988 Downloaded 19 Dec 2001 to 128.110.196.147. Redistribution subject to AlP license or copyright, see http://ojps.aip.org/jcpo/jcpcr.jsp Taylor et al.: Spectroscopy of Pt2 5523 7M. D. Morse, G. P. Hansen, P. R. R. Langridge-Smith, L.-S. Zheng, M. E. Geusic, D. L. Michalopoulos, and R. E. Smalley, J. Chern. Phys. 80, 5400 (1984). 8p. J. Brucat, L.-S. Zheng, e. L. Pettiette, S. Yang, and R. E. Smalley, J. Chern. Phys. 84, 3078 (1986). 9 A. D. Sappey, J. E. Harrington, and J. e. Weisshaar, J. Chern. Phys. 88, 5243 (1988). IOD. G. Leopold, J. Ho, and W. e. Lineberger, J. Chern. Phys. 86, 1715 (1987) . 110. Cheshnovsky, P. J. Brucat, S. Yang, e. L. Pettiette, M. J. Craycraft, and R. E. Smalley, in Physics and Chemistry of Small Clusters, edited by P. lena, B. K. Rao, and S. N. Khanna (Plenum, New York, 1987), NATO ASI Series B: Physics,Vol. 158, p. I. 12D. J. Trevor, R. L. Whetten, O. M. Cox, and A. Kaldor, J. Am. Chern. Soc. 107, 518 (1985). 13S. C. Richtsrneier, E. K. Parks, K. Liu, L. G. Pobo, and S. J. Riley, J. Chern. Phys. 82, 3659 (1985). 14M. O. Morse, M. E. Geusic, J. R. Heath, and R. E. Smalley, J. Chern. Phys. 83, 2293 (1985). I'R. J. St. Pierre and M. A. EI.-Sayed, J. Phys. Chern. 91, 763 (1987). lOy' Hamrick, S. Taylor, G. W. Lemire, Z.-W. Fu, J.-e. Shui, and M. O. Morse, 1. Chern. Phys. 88, 4095 (1988). 17K. Ervin, S. K. Loh, N. Aristov, and P. B. Armentrout, J. Phys. Chern. 87, 3593 (1983). 18p. B. Armentrout, S. K. Loh, and K. M. Ervin, J. Am. Chern. Soc. 106, 1161 (1984). 19M. F. Jarrold, A. J. IIIies, and M. T. Bowers, J. Am. Chern. Soc. 107,7339 (1985). 2°0. B. Jacobson and B. S. Freiser, J. Am. Chern. Soc. 107,1581 (1985). 21W. Weltner, Jr. and R. J. Van Zee, Annu. Rev. Phys. Chern. 35, 291 (1984). "M. O. Morse, Chern. Rev. 86, 1049 (1986). 230. R. Salahub, in Ab Initio Methods in Quantum Chemistry II, edited by K. P. Lawley (Wiley, New York, 1987); Adv. Chern. Phys. 69, 447 ( 1987). 24S. P. Walch and C. W. Bauschlicher, Jr., in Comparison of Ab Initio Quan­tum Chemistry with Experiment, edited by R. J. Bartlett (Reidel, Dor­drecht, 1985), p. 17. 2>W. Weltner, Jr. and R. J. Van Zee, in Comparison of Ab Initio Quantum Chemistry with Experiment, edited by R. J. Bartlett (Reidel, Oordrecht, 1985), p. I. 260. J. Trevor and A. Kaldor, in High Energy Processes in Organometallic Chemistry, edited by K. S. Suslick, ACS Syrnp. Ser. 333 (American Chemical Society, Washington, O.e., 1987), p. 43. 27A. Kaldor, O. M. Cox, and M. R. Zakin, Adv. Chern. Phys. (in press). 28S._S. Lin, B. Strauss, and A. Kant, J. Chern. Phys. 51, 2282 (1969). 291. Shim and K. A. Gingerich, J. Chern. Phys. 80, 5107 ( 1984). "'e. E. Moore, Natl. Bur. Stand. Circ. No. 467 (1971), Vol. III. 'IZ._W. Fu, G. W. Lemire, Y. Hamrick, S. Taylor, J.-e. Shui, and M. O. Morse, J. Chern. Phys. 88, 3524 (1988). "S. e. O'Brien, Y. Liu, Q. Zhang, 1. R. Heath, F. K. Tittel, R. F. Curl, and R. E. Smalley, J. Chern. Phys. 84, 4074 (1986). "M. J. PelIin (unpublished results). .HK. Jansson and R. J. SculIrnan, J. Mol. Spectrosc. 61, 299 (1976). "M. E. Jacox, J. Mol. Spectrosc. 113, 286 ( 1985). 'oK. Balasubrarnanian, J. Chern. Phys. 87, 6573 (1987). '71. H. Norman, H. G. Staley, and W. E. Bell, J. Phys. Chern. 71, 3686 ( 1967). "S. K. Gupta, B. M. Nappi, and K. A. Gingerich, Inorg. Chern. 20, 966 (1981 ). >OR. S. Krasnov, Teplofiz. Vys. Temp. 13,441 (1975). 4°A. R. Miedema and K. A. Gingerich, J. Phys. B 12, 2081 (1979). 41E. G. Rauh and R. J. Ackermann, J. Chern. Phys. 70, 1004 (1979). 42N. Aslund, R. F. Barrow, W. G. Richards, and O. N. Travis, Ark. Fys. 30, 171 (1965). 4'T. V. R. Rao and S. V. J. Lakshrnan, J. Quant. Spectrosc. Radiat. Transfer 11, 1157 (1971). 441. Kordis, K. A. Gingerich, and R. J. Seyse, J. Chern. Phys. 61, 5114 ( 1974). 45H. Basch, O. Cohen, and S. Topiol, Isr. J. Chern. 19,233 (1980). J. Chem. Phys., Vol. 89, No.9, 1 November 1988 Downloaded 19 Dec 2001 to 128.110.196.147. Redistribution subject to AlP license or copyright, see http://ojps.aip.org/jcpo/jcpcr.jsp