Photodissociation measurements Tiz, Vi, Co,‘, and Co;
of bond dissociation
energies:
Larry M. Russon, Scott A. Heidecke, Michelle K. Birke, J. Conceicao, Michael D. Morse, and P. B. Armentrout Department of Chemistry, University of Utah, Salt Lake City, Utah 84112
(Received 26 October 1993; accepted14 December 1993) The bond dissociation energiesof Ti:, V: , Co:, and Co: have been measuredfrom the sudden onset of predissociation in the photodissociationspectra of these molecules, yielding values of Di(Tii)=2.435+0.002 eV, D~(V~)=3.140-+0.002 eV, D~(Co~)=2.7651!~0.001 eV, and D~(Co;)=2.086+0.002 eV. Thesevalues are in good agreementwith values previously determined from collision-induced dissociation experiments. General criteria for the interpretation of predissociationthresholdsas bond dissociationenergiesand periodic trendsin the bonding of the 3d transition metal diatomic neutrals and monocationsare discussed.
I. INTRODUCTION
Investigationsinto the bonding of small transition metal clusters have been carried out for many years now. Part of the interest in this field arises from the complex interplay of opposing forces which are at work in these difficult electronic systems. On one hand, d-orbital contributions to the bonding can strengthenthe bond. This effect lessensas one moves left to right across the periodic table becausethe radial size of the d orbitals decreasessignificantly relative to the radial size of the s orbitals. On the other hand, there is often an energetic price that must be paid to promote the constituent atoms to an electron configuration that diabatitally correlates to ground state molecules. This promotion energy may weaken the adiabatic bond energy relative to the diabatic bond energy. Significant studies of the electronic spectraof many neutral transition metal dimers have been performed by using resonant two-photon ionization (R2PI) spectroscopy.‘-5 From such studies molecular term symbols, vibrational and rotational constants,and bond lengths have been determined. Bond strengthshave beenmeasuredby the observationof the suddenonset of predissociationin R2PI spectra.5-7Knudsen cell mass spectrometry8p9 has also provided bond strength information for many such molecules. Together, all of this information contributes to a better understanding of the bonding in these complicated systems. With the advent of resonance-enhanced photodissociation (REPD) spectroscopy,10-‘similar 2 studies of transition metal clusters with nonzerochargeare now possible.As yet, relatively little rotationally resolved spectroscopyhas been performed, severely limiting the types of information that can be obtained.Nevertheless,the bond strengthsof charged specieshave been measuredby the observationof predissociation thresholds.‘0*““3Such thermodynamicinformation is also accessibleby collision-induced dissociation (CID)‘4-21 studies,although thesetwo methodsdo not always agree.For example, Lessen and co-workers” recorded a REPD spectrum of Cri and observeda predissociationthresholdat 2.13 eV, while Su et al. measured the bond strength as Di(Cr;)=1.30+0.06 eV by CID of Crz with Xe.14It is apparent from this difference that a predissociationthresholdis
not always an accurate representation of the true bond strength,a fact also demonstratedby the observationthat not all transition metal diatomics exhibit a sharp predissociation threshold.4 We report here the bond strengthsof Tit, Vl , Co:, and Co: measuredfrom the sudden onset of predissociationin REPD spectra.Thesevalues are in good agreementwith values obtainedby Armentrout and co-workers,‘5-18again using CID with Xe. From these studies and those performed on neutral species using R2P1, criteria for the assignment of bond strengthsto predissociationthresholdshave been determined. The experimental apparatusused in this study is described in detail in Sec. II, and results of the studies of Ti,f , V,‘, Co:, and Co: are presentedin Sec. III. In Sec. IV, we discuss criteria for the interpretation of a predissociation threshold as an accuratemeasureof the bond strength, and periodic trends in the bond strengths of homonuclear diatomic moleculesand monocationsin the 3d transition metal series are also examined.A brief summary is presentedin Sec. V. II. EXPERIMENT
This study was performed on a recently completed,jetcooled ion photodissociation apparatus (Fig. 1). The ion source is a laser vaporization, supersonicexpansionsimilar in design to that used in the R2PI apparatusof the Morse laboratories.22A metal target disk is rotated and translated against a stainless steel vaporization block23mounted on a magnetically actuated, double solenoid valve.24 Helium at -10 psig, seeded with a small amount of argon (cl%), passes through a 3A molecular sieve trap maintained at -78 “C (dry ice/isopropanolbath), and is used as the carrier gas.At the approximatepeak of the gas pulse, 248 nm radiation (KrF) from an excimer laser (LambdaPhysik, EMG 101 MSC, 20-40 mJ/pulse) is focused onto the metal disk by a 47 cm focal length lens. The resulting metal plasma is swept through a clustering region 2 mm in diameter and approximately 3 cm in length. Ion formation in this source is sufficient to preclude the necessity of any secondaryionization. The vapor then expandssupersonicallythrough a diverging
4747 J. Chem. Downloaded Phys. 100 (7),021Apr April2001 1994to 128.110.196.147. 0021-9606/94/100(7)/4747/9/$6.00 0 1994 Institute of Physics Redistribution subject to AIP copyright, see American http://ojps.aip.org/jcpo/jcpcr.jsp
Russon et al.: Bond strengths of Ti;, Vi, Co:,
Q
F
A
B
C
nu’ ?iYs%%E 10” Din.
4
pump
J=b
1 6” Difhtak F
FIG. 1. Schematic of the ion beam photodissociation spectrometer used in this study. Note that the two flight tubes are shortened to more easily illustrate the details. (A) gate valve, (B) 248 nm excimer laser radiation, (C) pulsed gas valve, (D) two-dimensional turning quadrupole, (E) WileyMcLaren acceleration assembly, (F) einzel lenses, (G) deflector plates, (H) dual microchannel plate detectors, (I) reflectron assembly, (J) tunable dye laser radiation.
nozzle into a low pressureregion (~10~~ Ton; Varian VHS10, 10 in. diffusion pump). The resulting cooling of the internal degrees of freedom is sufficient to produce cobalt atomic ions with several helium atoms attached(Fig. 2). The metal vapor is collimated by a 6.5 mm conical skimmer as it enters a differentially pumped extraction chamber (-2X10m6 Ton; 6 in. diffusion pump). Positive ions are extracted using a two-dimensional turning quadrupole.=In this device, four stainless steel, quarter-circle rods (7.5 cm long, 3.8 cm radius) are placed at the comers of a squarestainless steel box (11 cm square,7.5 cm high). Electrodesin opposite
FIG. 2. Mass spectrum of laser-vaporized cobalt entrained in helium carrier gas. For this spectrum, no argon was seeded into the helium. Iron and water are incidental impurities.
and Co;
comers are held at the same electrostatic potential while those in adjacent comers have opposite applied potentials, referenced to the potential of the surrounding box. Shim electrodes,positioned betweenthe box and the circular electrodes, help to generatethe requisite hyperbolic equipotential lines and carry potentials that are intermediatebetween that of the nearestrod electrode and the surrounding box. As the various species in the source beam have been acceleratedto nearly the full supersonicvelocity of the carrier gas, they will be traveling with nearly the same velocity. Because the species have different masses, however, each will have a characteristic translational energy. The quadrupole, being an electrostatic device, affects ions of different energy differently, making it a mass-sensitivedevice in this application. To overcomethis effect, the box surroundingthe rods carries a negativevoltage (- - 180 V), thus accelerating the positive ions as they approach. Because the ions start with only a few electron volts of energy,the addition of 180 eV greatly narrows the relative energy difference between ions of different mass. Defining the potential on the box as V,, the potentials on the circular electrodesare (l.O+-0.8)Vo and those on the shim electrodes are ( l.0?0.4)Vo. In this configuration, negative ions are repelled by the potential on the box, neutral speciespass through largely unaffected,and positive ions are turned 90”, pass through a single-element, electrostatic lens (maintained at -- 1350 V), and enter a Wiley-McLaren time-of-flight (TOF) accelerationregion.26 The TOF source consists of two steel tubes 10 cm in diameterand 5 cm long. They have thin metal plateswith 1.5 cm holes in the center attachedto one end and are assembled with their open ends facing each other, separatedby 1 mm. This assemblyconstitutesthe repeller and draw-out plates in the Wiley-McLaren scheme.A third piece of 10 cm diam tube, 3.8 cm long, is left open on both ends and is mounted 1 mm away from the second tube. This shield electrode is held at ground potential. There is no ground plate per se but the first element of an einzel lens, which is kept at ground potential, is about 2.5 cm away from the final element of the TOF assembly.The repeller and draw-out plates are kept at ground potential until the ions fill the intervening region, and are then pulsed to 900 and 750 V, respectively,using a homebuilt circuit with rise times of 800 and 670 ns, respectively. This voltage is maintainedfor about 8 ms, which is sufficient for all of the ions to be acceleratedout of this region.As the experiment is operatedat 10 Hz, there is no problem in returning the plates to ground potential before the next experimental cycle. Following acceleration, the ion beam passes through two einzel lenses2.5 and 30.0 cm downstreamfrom the TOF source. In each lens, the outer elementsare 0.8 cm thick and the center element is 1.6 cm thick. The outer elements are separatedfrom the inner element by 0.8 cm. The inside diameter of the outer elements is 3.2 cm and that of the inner element is 4.8 cm. The central element of both lenses is held at approximately - 1100 V. The combination of the Wiley-McLaren sourceand einzel lensesallows the ion beam to be brought to a longitudinal and radial focus in the spectroscopychamber, 1.73 m downstream from the center of the quadrupole.This chamber is pumpedby a 6 in. diffusion pump (Edwards 160M diffstak),
J. Chem. Phys., Vol. 100, No. 7, 1 April 1994 Downloaded 02 Apr 2001 to 128.110.196.147. Redistribution subject to AIP copyright, see http://ojps.aip.org/jcpo/jcpcr.jsp
Russon et al.: Bond strengths of Ti:, V,+, I&!,
and CO;
I lcol 1 which maintains a pressureof -9X 10m7Torr. Here the ions Ti; *Ti’ + Ti are irradiated with the output of a tunable dye laser (Lumon9Xits, HD-500) pumped by 308 nm excimer laser radiation (Lumonics, EX-700 running on XeCl). Typically the laser radiation counterpropagatesalong the axis of the ion beam. For higher resolution work it is necessaryto overcome the Doppler broadening inherent in the system. This is accomplished by directing the laser radiation across the ion beam path at right angles.In this arrangementthe residual Doppler width is below the laser linewidth of 0.03 cm-‘. As the ions exit the spectroscopychamber, they pass 19400 19800 through another einzel lens and horizontal and vertical deEnergy (cm-‘) flecting plates before entering a reflectron flight tube. The reflectron is mounted 2.9 m downstreamfrom the center of the quadrupole.The einzel lens is identical in design to those FIG. 3. The predissociation threshold of Ti: detected in a photodissociation previously describedand is also maintainedat about -1100 action spectrum. This spectrum was obtained by scanning the dye laser V. The stainlesssteel deflection plates are 2.5X3.2X0.6 cm using coumarin 500. The arrow marks where a congested spectrum rises with matching pairs mounted 6.3 cm apart. The voltage on abruptly out of background noise at 19 640 cm-‘, giving a bond strength of Db(Ti:)=2.435?0.002 eV. The line across the top of the arrow indicates the each set of plates is centered around 0 V. The first set of uncertainty. deflector plates is maintained at about ?20 V in order to changethe horizontal flight path of the ions for proper flight through a gridless reflectron assembly.The reflectron is censcanned.In each case, a congestedspectrum arose out of tered in the secondflight tube, offset from the first flight tube background noise apparently caused by collision-induced by 2.86 cm, so that the deflection angle is slight (1.6”). The dissociationof the metal cluster ions with residual helium or secondset of deflector plates changesthe vertical flight path pump oil molecules.The spectrawere calibrated by collectto adjust for any vertical m isalignmentof the apparatusand ing an I, absorptionspectrum,while simultaneouslycollectis usually operatedin the range of -C5 V. ing the experimentaldata, and comparing it to the I, atlas of The reflectron consists of a stack of 19 stainless steel Gerstenkomand Luc.~~Where the laser radiation was to the plates with an outside diameter of 13.3 cm, an inside diamblue of the Z2 atlas, a high-pressureH, cell was used to eter of 8.9 cm, and a separationof 5.0 m m . The first plate is Ramanshift the light back into the range of the atlas.At 500 6.35 m m thick and is shorted to ground. The 2nd through psig, H2 gives a precisely known Q(1) Raman shift of 17th plates are 0.5 m m thick and are connectedin series by 100 kfl resistors. The 17th through 19th plates are shorted 4155.162 cm-1.28As a final correction, the energy scalesof togetherwith a mesh on the 17th plate to prevent penetration the spectra were shifted to correct for the Doppler shift experiencedby the ions as they approachedthe source of the of ground potential into the reflectron. A potential of apradiation. This correction rangedfrom 1.68 cm-’ for Co; to proximately 400 V is applied to the second plate and ap3.33 cm-’ for V:. proximately 950 V is applied to plate 17 for reflection of the full ion beam. In order to separateundissociatedparent ions from fragment ions, the voltages in the reflectron, on the 111.RESULTS third einzel lens and the deflector plates are reducedby the ratio of the mass of the fragment to the mass of the parent A. The bond strength of Ti: (e.g., in the case of a homonuclear dimer, l/2). Fragment Figure 3 shows a strong, abrupt increasein the Ti+ fragions are then detected at the same time of flight as unfragment signal at 19 640 cm-‘. This is assignedas the predismented parents would be at full voltage on the reflectron. sociation thresholdwith the uncertainty of 2 15 cm-’ arising Reflected ions are detectedby a dual m icrochannelplate defrom the difficulty of precisely determining the threshold. tector (Galileo, 3025). Any ions that are not reflectedmay be Weak, broad features observed below the threshold are the detected by another dual m icrochannel plate detector result of periodic fluctuations in the parent ion signal that are (Hamamatsu,F2221-21) mounted directly behind the reflecassociatedwith the rotational period of the sample rotary tron. The signal is amplified (Pacific Instruments, model drive mechanism.These source fluctuations contribute to an 2A50 video amplifier, Xl00 gain, 150 MHz) and digitized uncertainty in the measured threshold. The threshold of by a 40 MHz digital oscilloscope (Markenrich Corp., 19 6402 15 cm-’ (2.435+0.002 eV) is in good agreement WAAGII) mounted in a 386-basedpersonalcomputer (&OS, with the CID bond strength measurement of 386-20/8). The data are then summed and stored for later Di(Til)=2.37+0.07 eV,t5 and we therefore assign this analysis. threshold as the bond strength of this molecule. Predissociationthresholdswere obtained by monitoring In a spectroscopicinvestigation of Ti2,t Doverstil et al. fragment ion signal intensity as a function of dye laser frereport a lower lim it to the Ti, bond strength as quency.The purposeof this study was to determinethe bond Di(Ti,)S1.349 eV. By using the thermochemicalcycle, strengthsof the moleculesinvestigated,so only thoseregions D;(M,) + IE(M) =D;(M;) + IE(M,), (1) of the spectrum where a threshold was expected were
J. Chem. Phys., Vol. 100, No. 7, 1 April 1994 Downloaded 02 Apr 2001 to 128.110.196.147. Redistribution subject to AIP copyright, see http://ojps.aip.org/jcpo/jcpcr.jsp
4750
Russon et al.: Bond strengths of Ti;, Vi, Co;,
FIG. 4. The predissociation threshold of V; detected in a photodissociation action spectrum. These data were collected while scanning the dye laser using exalite 398. The arrow indicates where the predissociation threshold arises at 25 326 cm-‘, giving a bond strength of 3.140-t-0.002 eV. The line across the top of the arrow indicates the uncertainty.
this lower limit on the dititanium bond strength,in combination with our measuredvalue of the Ti: bond dissociation energy and the ionization energy of titanium, IE(Ti) =6.828 12t0.000 04 eV,29provides a lower limit on the ionization energyof Ti, as IE(Ti.Jb5.742 eV This may be combined with the upper limit on the ionization energy of Ti2 measuredby Doverstil et al., IE(TiJ~6.125 eV, to provide IE(Ti2)=5.93t0.19 eV. The same thermochemical cycle then provides D”,(Ti&=1.54+0.19 eV. B. The bond strength
of Vz
The photodissociationspectrumof Vz shown in Fig. 4 is relatively weak, making it difficult to pinpoint the predissociation threshold. Nevertheless,an abrupt increasein the V+ fragment signal is evident at 25 326+15 cm-‘, providing D”,(Vl)=3.140+0.002 eV. This value is in excellent agreement with the CID measurementof Di(V,f)=3.13%0.12 eV.i6 Given the bond strengthof V, measuredfrom a predissociation threshold ’ R2PI spectrum, Di(V&=2.753?0.001 eV,6%d I~(V)=6.740~0.002 eV3’ Equation (1) provides an ionization energyfor divanadium of IE(V2)=6.353?0.003 eV. This compares very favorably with a recent direct measurementof IE(V,)=6.356?0.001 eV,31providing strong evidencethat predissociationoccurs at the thermochemical threshold in V2 and Vl. Employing a more recent,highly precise value for the ionization energyof atomic vanadium, IE(V)=54411.7+-0.1 cm-’ (6.746 19 +-0.00001 eV),32the ionization energy predicted for V2 using Eq. (1) and the bond energies of V2 and Vz is IE(V2)= 6.359? 0.002 eV, which remains within experimental error of the directly measuredvalue. C. The bond strength
of Co:
In the predissociation threshold of Co;,” there is an abrupt rise in the Co+ fragment ion signal at 22 300 cm-‘, and above this energy the spectrum is very congested, as shown elsewhere.”Although most of the featuresabove this energy are reproducible, no assignmentwas attempted be-
and Co;
FIG. 5. A photodissociation spectrum of Co; cohected by monitoring the Co; signal intensity while scanning the dye laser using a 7:3 mix of rhodamine 590 and rhodamine 610. The arrow marks the predissociation threshold at 16 82.5 cm-‘, giving a bond strength of 2.086ZO.002 eV. The line across the top of the arrow indicates the uncertainty.
causeof the complexity of the spectrum.The predissociation threshold was determinedto lie at 22 30025 cm-‘, providing Di(Col) =2.765 +O.OOl eV The lower uncertainty limit in this example reflects a sharper, more precisely defined threshold. This bond energy again compares very favorably with the CID value of Di(Coz)=2.75t0.10 eV.17 D. The bond strength
of Co;
To illustrate the extension of this technique to larger clusters, the dissociation energy of Co: was also investigated. Figure 5 displays the predissociation threshold for Co: dissociating to Co:+Co. The fragment Co: ion signal measuredin this experiment is slightly weaker than that for Vz becauseof the lower Co: parent ion signal. A second difficulty in recording this predissociation threshold results becauseit occurs in a spectral region where no single dye lases efficiently. Figure 5 was obtained by scanning light output from a 7:3 mixture of rhodamine 590 and rhodamine 610 laser dyes, and as a result only covers a limited range. The threshold and bond strength are assignedas 16825t 15 cm-‘=2.086+0.002 eV. This again agreeswell with the CID value of Di(Col-Co)=2.04?0.13 eV.‘* IV. DISCUSSION A. Assignment strengths
of predissociation
thresholds
as bond
In addition to the molecules presentedhere, the observation of predissociation thresholdsin R2PI experimentshave yielded the bond strengthsof TiV,6 TiCo,6 V2,6 VNi,6 Ni2,7 NiPt,2 Pt2,3 and AlNi.33 Other transition metal systems, NiPd4 and PdPt4for example, do not exhibit sharppredissociation thresholds.From this information, Spain and Morse6 proposedthat in order to interpret a predissociationthreshold as a measureof the true bond strength certain factors must hold. First, the molecule must have a large density of electronic states near the ground separatedatom limit, and second, this separated atom limit must generate repulsive curves. Table I lists the number of relativistic, adiabatic
J. Chem. Phys., Vol. subject 100, No.to7,AIP 1 April 1994 see http://ojps.aip.org/jcpo/jcpcr.jsp Downloaded 02 Apr 2001 to 128.110.196.147. Redistribution copyright,
4751
RUSSOn et al.: Bond strengths of Tii, V,‘, Co,‘, and Co; TABLE I. Number of adiabatic Hund’s case (c) potential energy curves evolving from separated atom limits 6 10 000 cm-’ above the ground separated atom limit and bond strengths of selected diatomic transition metal molecules. Number of curves
Di, eV
Molecule
Number of curves
Dh, eV
Ti; TiV TiCo
5114 2186 1388
2.43Sa 2.068b 2.401b
co; Ni, NiPd
3458 960 291
2.76Sa 2.042d ...
v; v2
4772 2711
3.140a 2.753b
NiPt PdPt
630 159
2.798e ...
VNi Cr:
1516 222
2.100b 1.30c
pt2
AlNi
392 123
3.14' 2.29g
Molecule
‘This work. bReference 6. ‘Reference 14. dReference 44.
[Hund’s case (c)] potential energy curves that arise within 10000 cm-’ of ground state atoms for selected transition metal diatomic neutrals and cations. Clearly the ions that are the subject of this report have a very large number of lowlying electronic states, thus fulfilling the first requirement stated above. From the ground states of Ti (4~~3d’,~F) and Ti+ (4~3d*,~F),~~ 588 curves are generated.Of those, 441 correlate diabatically to s~$(~Ads(~F)di(~F) attractive curves while the remaining 147 curves correlate to s~~sa~d:(“F)d~(~F) repulsive curves. Thus, Tii strictly follows both of the guidelines outlined above. In the case of V,’, however, all of the 350 curves evolving from the lowest separatedatom limit of V (4.~~3d~,~F)-i-V+(3d~,~D),~’correlate diabatically to s~~d~(~F)d~(~D) attractive curves. This diabatic correlation for both Tii and Vl is based on guidelines set forth by Armentrout and Simons35which consider the 4s bonding alone. It neglects any contributions of the d orbitals to the bond and is therefore most useful for predicting long range attractive or repulsive behavior. The more efficient mixing of wave functions among crossed curves as compared to nestedcurves might explain why the predissociationof Vl does not show as sharp a threshold as does Ti: . However, Co: has a nearly identical pattern of diabatic correlations as Vl. Ground state Co (4s23d7,4F) and Co+ ( 3d8,3F)3” combine to generate588 curves, all of which correlate to scridi(4F)dff(3F) attractive surfaces. Within the first 10 000 cm-’ above the ground state separated atom limit, there are fewer curves generatedthan for Vi (see Table I), yet the Co: threshold is the strongest of those presentedin this paper. Figure 6 displays a qualitative depiction of the predicted pattern of potential energy curve% evolving from the first several separatedatom limits of Co and Cot to form Co:. The number of curves is calculated for the various combinations of separatedatom terms; the curves themselves are drawn to approach the lowest spinorbit level of each separatedatom limit. The bands drawn representthe rich density of states formed from these atoms. These are drawn assuming no d-orbital interaction, as is appropriate for transition metal atoms from the right side of the 3d transition series. The second rule developed by Spain and Morse6 con-
eReference 2. ‘Reference 3. gReference 33.
cerned repulsive curves, and was introduced to account for the lack of strong predissociationthresholdsin the spectraof NiPd4 and PdPt.4Like Vz and Co:, theseare molecules that generateonly attractive curves from their lowest separated atom limits. As shown in Table I, NiPd and PdPt have a much lower density of states than Co: or V:. This occurs becausethe high stability of the ‘So(d”) ground term of palladium greatly limits the density of electronic states in NiPd and PdPt. The handicap of having no repulsive curves evolving from the ground separatedatom limit may apparently be overcome in the casesof Co: and V: becausethere is a sufficient density of electronic states. One might ask whether repulsive curves are required at all becausethe NiPd and PdPt molecules may have lacked a sufficient density of states to fulfill the first requirement.However, another molecule studied by R2P1, AlNi,33 has fewer electronic states than either NiPd or PdPt (see Table I), but neverthelessdisplays a sharp predissociation threshold, presumably because of repulsive curves evolving from the Al (s~~~,~P”)
o’Yd’d’ - 294 curves a’o’d’d’ - 980 CUW~S =y\
u’dd’d’ s “/
co - 588 curves
t
Co’
‘F(s’d’) + ‘D(sOd*) ‘F($d’)
+ ‘F(s’d’)
‘F(s’d*) t ‘F(sOd*) -
‘F(s’d’)
----w--‘F(s’d’)
t”F(s’d’)
+‘F(s’d’) *F(s’d’) + ‘F(r’d’)
-------
‘F(s’d’)
+ ‘F(sOd’)
$d’ ‘ d’ - 588 curves
FIG. 6. Qualitative potential energy curves for Co: illustrating the enormous density of states. The shaded bands represent the great number of potential curves deriving from each separated atom asymptote. The different shading patterns are to distinguish states with an even number of s electrons from states with an odd number of s electrons. The 4F(~1d8) +3F(sods) and the 4F(s2d7) +‘F(.s’d’) states are nearly, but not exactly, degenerate.
Downloaded 02 Apr 2001 to 128.110.196.147. Redistribution subject to AIP copyright, see http://ojps.aip.org/jcpo/jcpcr.jsp J. Chem. Phys., Vol. 100, No. 7, 1 April 1994
4752
Russon et a/.: Bond strengths of Ti:, V;, Co;,
+Ni(s1d9,3D) limit. Repulsive curves evolving from the ground state asymptotecan allow efficient predissociationin those systemswhere the density of statesis too low for predissociation among nestedcurves to occur to an appreciable degree. In the caseof Cr: ,I0 the predissociationthreshold is not as abrupt as those presentedhere and there are distinct features in the signal above the thresholdthat suggestthere may be Franck-Condon difficulties in either the excitation or the predissociation step. Additionally, assuming a “2,’ ground state with an sold” configuration evolving from ground state Cr(4s13d5,g7S)+Cr+(3d5,6S), the first strongly allowed optical transition in Cri is to a state that correlatesto the Cr(4s’3d5,‘S) +Cr+(3d5,6S) excited separated atom limit, 0.94 eV37 above the ground separatedatom limit. When the photodissociation threshold is corrected by this energy,it is in quantitative agreementwith the CID result.14 With these ideas in mind we revise and extend the Spain-Morse rules as follows. A predissociation threshold may be inferred to correspondto the bond dissociation energy if (1) The threshold is sharp and well defined, without evidenceof Franck-Condon difficulties in either the excitation or predissociationstep; (2) Dissociation can occur to the ground separatedatom limit while preservingthe good quantum numbers,such as a, g/u, and +; (3a) Either the ground separatedatom limit generatesa suitable number of molecular potential energy surfaces,some of which are repulsive; or (3b) The ground separatedatom limit generatesa sufficiently large number of attractive molecular potential energy surfaces to allow weaker predissociationprocessesto dominate. B. Electronic
configurations
7. 77:
Ti, is known by R2PI spectroscopy’to have a 3Ag,r ground state indicating that the dug orbital is mixed with the sag orbital to the extent that it lies higher in energy than the d rr,, orbitals, resulting in an s uid rrzd aid 8: configuration. The removal of one electron in order to form the ion leads to either a 22+(sa~dv~du~) or a 2Ag(su~dn~dS~) ground state. Knig&’ has performed electron spin resonancespectroscopy on a sample believed to contain Til but no spectrum was observed,making “2,’ an unlikely assignmentfor the ground state. This is consistent with the idea that in the transition metal cations the d orbitals are stabilized relative to the s orbital, enhancing the interaction between the su and du molecular orbitals to the point that the dag orbital lies abovethe d Sg orbital in energy.Assuming a ‘ha,, ground state for Ti: , the ground level will possessfi=3/2. Allowed electric dipole transitions can then access ungerude states with R=1/2, 312, or 512, all of which may be produced by combining a ground state titanium atom (3Fz) with a ground state titanium ion (4F3n).As a result, Til can dissociate to ground state atoms while preserving its value of fl and g/u symmetry. The 3F(4s23d2) and 4F(4s3d2),34 ground states of Ti and Ti+, respectively,combine diabatically to form Til with a su~su~d~(3F)d~(3F) configuration. On the other hand, combination of a ground state titanium atom, 3F(4s23d2), with the 4F(3d3), excited state of Ti+, which
and Co;
lies only 907.96 cm-’ higher than the ground state of Tic, should diabatically form Til without an su: antibonding electron, resulting in a stronger bond with an suidi(3F)di(4F) configuration and more d-orbital electrons available for additional bonding. The ground state of titanium dimer cation is thereforeexpectedto arise from this excited separatedatom limit.
2. v-l;
From R2PI data’the ground state of V, is known to be ‘2, arising from an sazdrr;fd4dc$ configuration. Because of su-da hybridization it is likely that the dug orbital lies betweenthe drr, and the d6, orbitals in energy.Removal of one electron to form the cation will most likely result in a “c, state, regardlessof the relative energiesof the do, and d Sg orbitals, as long as the dug orbital lies above the dr, orbitals, which is very probable.Van Zee and Weltner39have determinedthe ground state of the isoelectronic neutral molecule, TiV, to be 4c-, further supporting the 4Xg ground state assignmentfor Vl . From this state !J= l/2, and 3/2, are generated,which can be excited to s1~5/2, states under dipole selection rules. The ground separatedatom limit of V(4s23d3,4F3,2)+V+(3d4,5Do)30 yields Hund’s case (c) states of n=1/2, and 312,) allowing predissociation of a= l/2, and 312, stateswhile preservingthe values of R and g/u. The 0=5/2, states, which may be excited from the 4c,(3/2g) would be expectedto provide a secondpredissociation threshold in this experiment.Our inability to observe a second threshold may indicate complete cooling to a ground, 42;( l/2,) level. The ground state of vanadium atom is 4F(4s23 d3) and that for the vanadium atomic ion is 5D( 3d4).% These may combine to form an s$gd~(4F)d~(5D) bond, plus contributions from the d-orbital electrons.This is consistentwith the postulateof an suidrriduid 6, “C, ground state for Vz .
3. co,’
The ground state of Co: has been determinedfrom ESR spectroscopy@to be “2. Possible R values are l/2, 312, or 5/2. An electric dipole transition from the ground state can then accessstateswith fi=1/2, 312,512,or 712.All of these n values are generatedby the combination of a ground state cobalt atom (4s23d7,4F9,2) with a ground state cobalt ion (3d8,3F4) so there is no a-based restriction to prevent the states reachedby the optical excitation of ground state Co: from predissociating to ground state atoms. Becausecobalt occurs later in the transition metal series where the 3d orbitals are quite contracted, the 3d orbital contributions to the bonding in Co; should be greatly reduced as compared to Til and Vz . This effect may be mitigated to some extent by the contraction of the s orbital, which provides greater s-electron density between the nuclei for bonding. The 4F(4s23d7) and 3F(3d8)36 ground states of Co and Co+, res ectively, are perfectly set up for the formation of an s J gdi(4F)di(3F) bond by the donation of the two s electrons of Co into the empty s orbital of Co+.
J. Chem. Phys., Vol. subject 100, No. to7, AIP 1 April 1994 Downloaded 02 Apr 2001 to 128.110.196.147. Redistribution copyright, see http://ojps.aip.org/jcpo/jcpcr.jsp
Russon et al.: Bond strengths of Ti;, Vl, Co:,
C. Promotion
TABLE Ii. Bond strengths and promotion energies of diatomic transition metal ions and neutrals. All values in eV. Measured 0; ’
Molecule
2.435(2) 1.54(19) 3.140(2) 2.753(l)’ 1.30(6)h l&3(56)’ 31.4 0.3(3)k 2.74(10)’ 1.14(l)” 2.765(l)’ 0.7-1.4O 2.08(7)q 2.042(2)’ 1.84(8)’ 2.03(2)”
Tif -I$ V; V? C$ Cr2 Mni Mn2 Fe; Fez coy co2 Nif Ni, cu; cu2
4x
energies
In order to understand the bonding among transition metal dimers, it is important to consider the expectedextent of d-orbital interactions and the magnitudeof the promotion energiesinvolved. In many cases,the ground state configuration of the molecule does not correlate diabatically to ground state atoms. Such promotion effects often occur in the neutral specieswhere the atomic ground state configuration is usually s2dn. Notable exceptionsalong the 3d transition metal series are Cr(s’d’) and Cu(s’d”). Atoms with sldn+ 1 configurationsare expectedto form a strongly bound said nt Id”+’ molecular configuration. In the case of the monocations, the usual ground atomic configuration is sod”+’ which will combine readily with the usual ground state neutral atom (s’d”) to form an suid”+‘dn molecular configuration. Scandium, titanium, manganese,and iron are the exceptions among the 3d atomic cations, having sldn ground state configurations. In each of these metals, it is energeticallymore favorable to promote the ion to an s’d”+l configuration than to promote the neutral to an sldn+l configuration. The homonuclear dimer cation with an su2dn+l d n configuration is then formed from M[s’d”)+Mt(sod”“). For this discussion, we define the promotion energy as the energy required to promote each atom from the lowest J level of the ground state to the lowest J level of the lowest lying excited state having the appropriate electron configuration for formation of an suidn”dnfl molecular configuration for neutrals, or an suidn+*dn molecular configuration for ions. The diabatic bond strength will accordingly be defined as the measured,adiabatic bond
Epavb 0.1136 2X0.813d 2XO.2628
Diabatic 0; a2b 2.548(2) 3.17(19) 3.140(2) 3.277(l) l&3(56)
0.232m 2X0.859m 2XO.432p 2X0.025’
2.97(10) 2.86(l) 2.765(l) 1.6-2.3 2.08(7) 2.092(2) 2.03(2)
4753
and Co;
3
1
0 I
FIG. 7. strengths transition definition
I
I
I
I
I
I
l-i
V
Cr
Mn
Fe
Co
I Ni
I
I
Cl!
Measured (open symbols) and diabatic (filled symbols) bond of neutral (squares) and cationic (triangles) homonuclear diatomic metal molecules from the data listed in Table II. See the text for of diabatic bond strength.
strength plus the promotion energy.These concepts are discussedfurther in a review of CID data by Armentrout et ~1.~~ Table II lists the adiabatic and diabatic bond strengths and promotion energies,E, , of all the homonucleardiatomic neutrals and monocations in the 3d transition metal series from Ti to Cu. These data are also representedgraphically in Fig. 7. The diabatic bond strengthsof Crl, Mnzf , Mn2, and Cul are not representedbecausethey are not expected to have SU$YU: ground state configurations. Instead, so--u: and scr~su~ground configurationsare expectedfor Mnz and Mn2, respectively, and an suj ground configuration is expected for Crl14 and Cul . While this configuration for Crg has not been proven conclusively, it does correctly predict the observed predissociation threshold for Cri reported by Lessenet al.,” as previously discussed.14
D. Periodic trends
Reviewing Table II and Fig. 7 there are several observations concerningthe bonding that can be made. (A discussion including larger clusters in the 3d series,and Nb and Ta can be found in the review by Armentrout er a1.)41The diabatic bond strengths of Til and Ti, are lower than those for V$ and V,, suggestingan increasein the bonding among the d orbitals in the vanadium dimers, which place more d electrons in bonding orbitals than do the titanium dimers. The lower diabatic bond strength of titanium and vanadium cationic dimers compared to their neutral counterparts is the result of the removal of an electron from a bonding d molecular orbital. This phenomenonis reversedfor the iron and cobalt dimers, where the electron is removed from an antibonding d molecular orbital. The diabatic bond strength of Co2 plotted on Fig. 7, taken as the midpoint of the rangeof values listed in Table II, seemsanomalously low. The lower limit of this range is the lower limit of a bond strength for Co2 calculated by Shim
‘Uncertainties in the final digits are given in parentheses. ‘See the text for definition. This work. dReference 34. ‘See the text. ‘Reference 6. aReference 30. ‘Reference 14.
kReference 47. ‘Reference 19. “Reference 48. “Reference 49. ‘See the text. aReference 36. 4Reference 20. ‘Reference 44. ‘Reference 50.
‘Rcfcrcnce 45.
‘Reference 51.
and Gingerich.42The upper limit is derived from an upper
‘Reference 46.
“Reference 52.
limit to the ionization energy of Co2, IE(Cods6.42 eV
Downloaded 02 Apr 2001 to 128.110.196.147. AIP 1994 copyright, see http://ojps.aip.org/jcpo/jcpcr.jsp J. Chem. Redistribution Phys., Vol. 100, subject No. 7, 1toApril
4754
Russonet al.:
Bond strengths of Ti2+, Vi, Co;,
which was establishedby noting that CoZ is readily ionized with 193 nm ArF laser radiation.43The Co, bond strength is thus easily calculated through Eq. (l), using IE(Co)=7.864 eV.36Based on a comparison to similar molecules as shown in Fig. 7, it appearsthat the true bond strength of Co, is probably closer to the upper limit and that the true ionization energy of CoZ is probably close to the 6.42 eV upper limit. The diabatic bond strengths of the Fez, Fez, and Co: molecules are higher than those of Ni2, Nizf , and Cu,. Copper dimer neutral, having a full 3d subshell, is representative of a transition metal diatomic molecule with no d-orbital contributions to the bonding. As the nickel dimer species have similar bond strengths, they also appear to have no d-orbital bonding. The higher diabatic bond strengths for Fez, Fez, and Cot seem to indicate that d-orbital interactions may be occurring in these species.The diabatic bond strengthsof Fe, and Fe; are slightly higher than thoseof Co2 and Co;, consistent with both greater d-orbital interactions and higher d orbital bond orders in the iron species. V. SUMMARY
The sudden onset of predissociation in the resonanceenhancedphotodissociation spectra of Til, Vz, Co:, and Cot has been observed and is interpreted as providing the thermochemicalbond strengthof thesemolecules.Thesevalues, Di(Tiz)=2.435?0.002 eV, D~(V~)=3.140+0.002 eV, Di(Co,‘)=2.765+0.001 eV, and D~(Co~)=2.086+0.002 eV, are in good agreementwith values obtained from collisioninduced dissociation experiments.The presentmeasurements are the most precise currently available for these molecules. Combined with auxiliary data for Ti, and VZ, these results give D~(TiJ=1.54+0.19 eV, IE(Ti2)=5.93t0.19 eV, and IE(V,)=6.359?0.002 eV. This last result is in excellent agreement with a recent direct measurement of IE(V,)=6.356+0.001 eV.31Togetherwith the bond strengths of several neutral diatomic speciesmeasuredfrom predissociation thresholds in resonant two-photon ionization spectroscopy,the data presentedhere illustrate the general applicability of this technique to transition metal cluster cations that meet certain criteria which are discussedin detail. The addition of the triatomic metal cluster cation, Co;, demonstrates the extension of this technique to larger clusters. Comparisonsof diabatic bond strengthsfor homonucleardiatomic molecules and monocationsin the 3d transition metal series from Ti to Cu are made. Periodic trends in the metal dimer bond energies show that d-orbital interactions play major roles in the bonding of the early transition metal diatomics, with the importance of these roles decreasinglater in the series. ACKNOWLEDGMENTS
This research is funded by the Department of Energy, Office of Basic Energy Sciences(P.B.A.); the National Science Foundation under Grant No. CHE-9215193 (M.D.M.); and partial support is provided by the Donors of the Petroleum ResearchFund, administeredby the American Chemical Society (M.D.M.). Funds used to purchasethe excimer-
and Co;
pumped dye laser system employed in these experiments were provided by the National Science Foundation under Grant No. CHE-8917980 (P.B.A./M.D.M.). ‘M. Dover&l, B. Lingren, U. Sassenberg, C. A. Arrington, and M. D. Morse, J. Chem. Phys. 97, 7087 (1992). *S. Taylor, E. M. Spain, and M. D. Morse, J. Chem. Phys. 92,269s (1990). 3S. Taylor, G. Lemire, Y. M. Hamrick, 2. Fu, and M. D. Morse, J. Chem. Phys. 89, 5517 (1988). 4S. Taylor, E. M. Spain, and M. D. Morse, J. Chem. Phys. 92,271O (1990). ‘E. M. Spain, J. M. Behm, and M. D. Morse, J. Chem. Phys. 96, 2511 (1992). 6E. M. Spain and M. D. Morse, J. Phys. Chem. 96, 2479 (1992). 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. Chem. Phys. SO,5400 (1984). *A. Kant and S.-S. Lin, I. Chem. Phys. 51, 1644 (1969). 9S. K. Gupta, B. M. Nappi, and K. A. Gingerich, Inorg. Chem. 20, 966 (1981). “D. E. Lessen, R. L. Asher, and P. J. Brucat, Chem. Phys. L&t. 182, 412 (1991). “L M Russon, S. A. Heidecke, M. K. Birke, J. Conceicao, P. B. Armentr&t,‘and M. D. Morse, Chem. Phys. Lett. 204, 235 (1993). ‘*K. F. Willey, C. S. Yeh, D. L. Robbins, and M. A. Duncan, Chem. Phys. Lett. 192, 179 (1992); C. S. Yeh, K. F. Willey, D. L. Robbins, and M. A. Duncan, ibid. 196, 233 (1992). “R. L. Hettich and B. S. Freiser, J. Am. Chem. Sot. 109, 3537 (1987). 14C.-X. Su, D.A. Hales, and P. B. Armentrout, Chem. Phys. Lett. 201, 199 (1993); C.-X. Su and P. B. Armentrout, J. Chem. Phys. 99, 6506 (1993). ‘jL. Lian, C.-X. Su, and P. B. Armentrout, J. Chem. Phys. 97, 4084 (1992). 16C.-X. Su, D. A. Hales, and P. B. Armentrout, J. Chem. Phys. 99, 6613 (1993). 17D. A. Hales and P. B. Annentrout, J. Cluster Sci. 1, 127 (1990). ‘*D.A. Hales, C.-X. Su, L. Lian, and P. B. Armentrout, J. Chem. Phys. 100, 1049 (1994). ‘9S . K . Lob, D.A. Hales, L. Lian, and P. B. Armentrout, J. Chem. Phys. 90, 5466 (1989); S. K. Loh, L. Lian, D. A. Hales, and P. B. Armentrout, J. Phys. Chem. 92, 4009 (1988); L. Lian, C.-X. Su, and P. B. Armentrout, J. Chem. Phys. 97, 4072 (1992). ML. Lian, C.-X. Su, and P. B. Armentrout, Chem. Phys. L&t. 180, 168 (1991); J. Chem. Phys. 96, 7542 (1985). *IL. Lian, R. H. Schultz, and P. B. Armentrout (unpublished). “Z. Fu, G. Bemire, Y. M. Hamrick, S. Taylor, J.-C. Shui, and M. D. Morse, J. Chem. Phys. 88, 3524 (1988). =Similar in design to that described in S. C. O’Brien, Y. Liu, Q. Zhang, J. R. Heath, F. K. Tittel, R. F. Curl, and R. E. Smalley, J. Chem. Phys. 84, 4074 (1986). “Similar in design to that described in J. B. Hopkins, P. R. R. LangridgeSmith, M. D. Morse, and R. E. Smalley, J. Chem. Phys. 78, 1627 (1983). =H. D. Zeman, Rev. Sci. Instrum. 48, 1079 (1977). %W. C. Wiley and I. H. McLaren, Rev. Sci. Instrum. 26, 1150 (1955). 27S. Gerstenkom and P. Luc, Atlas du Spectre d’Absorption de la Mol&ule d’lode (CNRS, Paris, 1978); S. Gerstenkom and P. Luc, Rev. Phys. Appl. 14, 791 (1979). =D. J. CIouthier and J. Karolczak, Rev. Sci. Instrum. 61, 1607 (1990). 29J E. Sohl, Y. Zhu, and R. D. Knight, J. Opt. Sot. Am. B 7, 9 (1990). MJ: Sugar and C. Corliss, J. Phys. Chem. Ref. Data 7, 1191 (1978). 31B. Simard, A. M. James, P. Kowalczyk, R. Foumier, and P. A. Hackett, Proc. SPIE (in press). 32B Simard, A. M. James, and P. KowaIczyk (private communication). 33J ‘M. Behm, C. A. Anington, and M. D. Morse, J. Chem. Phys. 99,6409 (i993). “C. Corliss and J. Sugar, J. Phys. Chem. Ref. Data 8, 1 (1979). “P. B. Armentrout and J. Simons, J. Am. Chem. Sot. 114, 8627 (1992). 36C. Corliss and J. Sugar, J. Phys. Chem. Ref. Data 10, 197 (1981). 37J. Sugar and C. Corliss, J. Phys. Chem. Ref. Data 6, 317 (1977). 38L. B. Knight, Jr. (private communication). 39R. J. Van Zee and W. Weltner, Jr., Chem. Phys. Lett. 107, 173 (1984). 40R. J. Van Zee. Y. M. Hamrick. S. Li. and W. Weltner.I Jr..I Chem. Phvs. . ’ ’ Lett. 195, 214’(1992). 4LP. B . Armentrout, D. A. Hales, and L. Lian, Adv. Metal Semicond. Clus42rrrs,h~~t~n”,e,sl~~h,
J. Chem. Phys. 78, 5693 (1983).
J. Chem. Phys., Vol. 100, No. 7, 1 April 1994 Downloaded 02 Apr 2001 to 128.110.196.147. Redistribution subject to AIP copyright, see http://ojps.aip.org/jcpo/jcpcr.jsp
Russon
et a/.: Bond strengths of Tii, Vi, Co,‘, and Co,+
4755
48C. Corliss and J. Sugar, J. Phys. Chem. Ref. Data 11, 135 (1982). “Ionization energies from E. A. Rohhing, D. M. Cox, A. Kaldor, and K. H. Johnson, J. Chem. Phys. 81, 3846 (1984). “C. Corliss and J. Sugar, J. Phys. Chem. Ref. Data 10, 197 (1981). “A. D. Sappey, J. E. Harrington, and J. C. Weisshaar, J. Chem. Phys. 91, 3854 (1989). 52M. D. Morse, Adv. Metal Semicond. Clusters 1, 83 (1993).
432. Fu and M. D. Morse (unpublished data). uJ. C. Pinegar, J. D. Langenberg, and M. D. Morse (unpublished). 4sK. Hilpert and K. Ruthardt, Ber. Bunsenges. Phys. Chem. 91, 724 (1987). &M. E Jarrold, A. J. Illies, and M. T. Bowers, J. Am. Chem. Sot. 107, 7339 (1985). 47T L. Haslett, M. Moskovits, and A. L. Weitzman, J. Mol. Spectrosc. 135, 2;9 (1989).
Downloaded 02 Apr 2001 to 128.110.196.147. Redistribution subject to AIP copyright, see http://ojps.aip.org/jcpo/jcpcr.jsp .I Chem
Phvn
Vnl
IlXl
Nn
7 1 Aoril
1994
