THRESHOLD PHOTO-IONISATION AND DENSITY FUNCTIONAL THEORY STUDIES OF METAL-CARBIDE CLUSTERS

Transcription

1 THRESHOLD PHOTO-IONISATION AND DENSITY FUNCTIONAL THEORY STUDIES OF METAL-CARBIDE CLUSTERS Viktoras Dryza A thesis submitted in total fulfillment of the requirements for the degree of Doctor of Philosophy November, 2008 Department of Chemistry The University of Adelaide

2 This work contains no material which has been accepted for the award of any other degree or diploma in any university or other tertiary institution and, to the best of my knowledge and belief, contains no material previously published or written by another person, except where due reference has been made in the text. I give consent to this copy of my thesis, when deposited in the University Library, being made available in all forms of media, now or hereafter known. Viktoras Dryza November, II

3 Dedicated to my Dad. III

4 Serenity now, serenity now! - Frank Costanza (Seinfeld, Season 9/Episode 3) IV

5 Abstract Neutral gas-phase metal-carbide clusters are generated by laser ablation and are detected in the constructed time-of-flight mass-spectrometer by laser ionisation. Photo-ionisation efficiency (PIE) experiments are performed on the metal-carbide clusters to determine their ionisation potentials (IPs). Complimentary density functional theory (DFT) calculations are performed on the energetically favorable structural isomers of the metalcarbide clusters. Comparison between the calculated IPs of the isomers and the experimental IP allows the carrier of the observed ionisation onset for a metal-carbide cluster to be assigned. The niobium-carbide clusters Nb 3 C y (y = 0 4), Nb 4 C y (y = 0 6) and Nb 5 C y (y = 0 6) are examined by PIE experiments and DFT calculations. The IPs of the niobium-carbide clusters are found to be either left reasonably unchanged from the IPs of the bare metal clusters or moderately reduced. The clusters Nb 3 C 2, Nb 4 C 4, Nb 5 C 2 and Nb 5 C 3 display the largest IP reductions for their corresponding cluster series. The structures assigned to the IPs of the Nb 3 C y (y = 1 3) clusters are based on the carbon atoms attaching to the niobium faces and/or niobium-niobium edges of the triangular Nb 3 cluster. However, for Nb 3 C 4 the ionisation onset is assigned to a low-lying isomer, which contains a molecular C 2 unit, rather than the lowest energy isomer, a niobium atom deficient face-centred cubic (fcc) nanocrystal structure. The structures assigned to the IPs of the Nb 4 C y (y = 1 4) clusters are based on the carbon atoms attaching in turn to the niobium faces of the tetrahedral Nb 4 cluster, developing a fcc nanocrystal structure for Nb 4 C 4. For Nb 4 C 3 two ionisation onsets are observed; one weak onset at low energy and another more intense onset at high energy. It is proposed that the two onsets are due to ionisation from both a metastable 3 A 1 state and the ground 1 A 1 state of the lowest energy isomer. The ionisation onsets of Nb 4 C 5 and Nb 4 C 6 are also proposed to originate from metastable triplet states of the lowest energy isomers, with the transitions from the ground singlet states calculated to be greater than the highest V

6 achievable photon energy in the laboratory. The structures of Nb 4 C 5 and Nb 4 C 6 have one and two carbon atoms in a fcc nanocrystal substituted with molecular C 2 units, respectively. The structures assigned to the IPs of the Nb 5 C y (y = 1 6) clusters are based on the underlying Nb 5 cluster being in either a prolate or oblate trigonal bipyramid geometry; the former has six niobium faces available for carbon addition, while the latter has two niobium butterfly motifs and two niobium faces available for carbon addition. Both the structures of Nb 5 C 5 and Nb 5 C 6 have the underlying Nb 5 cluster in the oblate trigonal bipyramid geometry and contain one and two molecular C 2 units, respectively. The tantalum-carbide clusters Ta 3 C y (y = 0 3), Ta 4 C y (y = 0 4) and Ta 5 C y (y = 0 6) are examined by PIE experiments and DFT calculations. The IPs of the tantalum-carbide clusters in each series show trends that are very similar to the corresponding iso-valent niobium-carbide cluster series, although the IP reductions upon carbon addition are smaller for the former. For the vast majority of tantalum-carbide clusters, the same structural isomer is assigned to the ionisation onset as that assigned for the corresponding niobium-carbide cluster. Bimetallic tantalum-zirconium-carbide clusters are generated using a constructed double ablation cluster source. The Ta 3 ZrC y (y = 0 4) clusters are examined by PIE experiments and DFT calculations. The IP trend for the Ta 3 ZrC y cluster series is reasonably similar to that of the Ta 4 C y cluster series, although the IP reductions upon carbon addition are greater for the former. The structures assigned to the IPs of the Ta 3 ZrC y (y = 1 4) clusters are based on the carbon atoms attaching in turn to the metal faces of the tetrahedral Ta 3 Zr cluster. In summary, the work presented in this thesis demonstrates that the structures of metalcarbide clusters can be inferred by the determination of their IPs through PIE experiments in combination with DFT calculations on candidate structural isomers. VI

7 Acknowledgments First and foremost I would like to thank my supervisor Dr. Greg Metha for giving me this opportunity. Professionally his helpful advice, extensive knowledge and forward thinking attitude to research have been invaluable. Personally he has been a great friend and mentor. I will always remember our stay in Munich; getting lost in a dark snowing forest and throwing snowballs at each other on the way home after a late night out. Also, the many laughing fits at our mutual appreciation of Borat. Thanks must also go to my co-supervisor Prof. Mark Buntine for his smart-arse comments and the occasional helpful one. I am also totally indebted to our post-doc Dr. Jason Gascooke, whose ability to solve problems which seem unsolvable to me is truly amazing. He has also been a great drinking buddy and is the co-inventor of beer-table tennis. Thanks to Matt Addicoat for his help with the computational calculations. The support in the mechanical workshop from Peter Apoefis has been fantastic in converting my caveman-like drawings into reality. I have also had many a fun time in the lab with my partner in crime Olivia Maselli; whether it be dancing around like a fool, discussing our small overlap in musical tastes or laughing at her stories. Also thanks to all the other Honours and PhD students I have shared the lab with. Huge thanks must go to my extremely supportive family (Mum, Dad and Kristina) for the Sunday lunches and for still believing I have some intelligence, despite observing my constant stupidity over the years. Last, but definitely not least, I must thank my beautiful girlfriend Milena. She has always believed in me and has the magical ability to make me happy whenever I m with her (even when my experiments are not working). VII

8 Publications The following publications were produced from the work presented in this thesis: Ionization Potentials of Tantalum-Carbide Clusters: An Experimental and Density Functional Theory Study Dryza, V.; Addicoat, M.A.; Gascooke, J.R.; Buntine, M.A.; Metha, G.F. J. Phys. Chem. A 2005, 109, Threshold Photo-ionization and Density Functional Theory Studies of the Niobium- Carbide Clusters Nb 3 C n (n = 1 4) and Nb 4 C n (n = 1 6) Dryza, V.; Addicoat, M.A.; Gascooke, J.R.; Buntine, M.A.; Metha, G.F. J. Phys. Chem. A 2008, 112, Onset of Carbon-Carbon Bonding in the Nb 5 C y (y = 0 6) Clusters: A Threshold Photoionisation and Density Functional Theory Study Dryza, V.; Gascooke, J.R.; Buntine, M.A.; Metha, G.F. PCCP 2008, (in press). The following publications are currently in preparation from the work presented in this thesis: Onset of Carbon-Carbon Bonding in the Ta 5 C y (y = 0 6) Clusters: A Threshold Photoionization and Density Functional Theory Study Dryza, V.; Gascooke, J.R.; Buntine, M.A.; Metha, G.F. (in preparation). Threshold Photo-ionization and Density Functional Theory Studies of Bimetallic-Carbide Clusters: Ta 3 ZrC y (y = 0 4) Dryza, V.; Gascooke, J.R.; Buntine, M.A.; Metha, G.F. (in preparation). VIII

9 Abbreviations AO C DFT EA ECP ev FC FCC FEL FWHM GTO HOMO IP IR LUMO MO MPD MPI MRCI Nb NBO PES PFI-ZEKE PIE REMPI SPI Ta TOF-MS ZPE Zr Atomic Orbital Carbon Density Functional Theory Electron Affinity Effective Core Potential Electron Volt Frank-Condon Face-centred Cubic Free Electron Laser Full Width at Half Maximum Gaussian Type Orbital Highest Occupied Molecular Orbital Ionisation Potential Infrared Lowest Unoccupied Molecular Orbital Molecular Orbital Multi-photon Dissociation Multi-photon Ionisation Multi-Reference Configuration Interaction Niobium Natural Bond Order Potential Energy Surface Pulsed Field Ionisation Zero Electron Kinetic Energy Photo-ionisation Efficiency Resonance Enhanced Multi-photon Ionisation Single-photon Ionisation Tantalum Time-of-flight Mass-spectrometer Zero-point Energies Zirconium IX

10 Table of Contents Chapter 1: Introduction to Metal-Carbide Clusters Generation and Distribution of Metal-Carbide Clusters The Structures of the Ti 8 C 12 Metcar and Ti 14 C 13 Nanocrystal Potential Applications of Metal-Carbide Clusters Scope of this Thesis Why Determine the Ionisation Potentials of Metal-Carbide Clusters? References 10 Chapter 2: Experimental Approach Laser Ablation Laser Ablation Theory Experimental Design and Operation Photo-ionisation Efficiency (PIE) Photo-ionisation Theory PIE Theory PIE of Transition Metal Containing Clusters Experimental Procedure for PIE Experiments References 31 Chapter 3: Computational Approach Introduction to Density Functional Theory (DFT) Basis Sets Computational Method and Procedure for Metal-Carbide Clusters References 40 Chapter 4: The Time-of-Flight Mass-Spectrometer The Constructed TOF-MS Apparatus References 47 X

11 Chapter 5: Niobium-Carbide Clusters Introduction to Niobium-Carbide and Tantalum-Carbide Clusters Photo-ionisation Efficiency Experiments DFT Calculated Isomers The Nb 3 C y (y = 0 4) Cluster Series The Nb 4 C y (y = 0 6) Cluster Series The Nb 5 C y (y = 0 6) Cluster Series Comparison between Calculated and Experimental IPs The Nb 3 C y (y = 0 4) Cluster Series The Nb 4 C y (y = 0 6) Cluster Series The Nb 5 C y (y = 0 6) Cluster Series Discussion on Niobium-Carbide Clusters Comparison to Previous Spectroscopic Data Onset of Carbon-Carbon Bonding Low-Lying Isomers and Metastable Electronic States IP trends, Molecular Orbitals and HOMO-LUMO gaps Thermodynamic Cycle Summary and Future Directions References 109 Chapter 6: Tantalum-Carbide Clusters Photo-ionisation Efficiency Experiments Comparison between Calculated and Experimental IPs The Ta 3 C y (y = 0 3) Cluster Series The Ta 4 C y (y = 0 4) Cluster Series The Ta 5 C y (y = 0 6) Cluster Series Discussion on Tantalum-Carbide Clusters Comparison to the Niobium-Carbide Clusters Summary References 130 XI

12 Chapter 7: Tantalum-Zirconium-Carbide Clusters Introduction to Bimetallic-Carbide Clusters Double Ablation Cluster Source Design Generation of Tantalum-Zirconium-Carbide Clusters Photo-ionisation Efficiency Experiments DFT Calculated Isomers The Ta 3 ZrC y (y = 0 4) Cluster Series Comparison between Calculated and Experimental IPs Discussion on Tantalum-Zirconium-Carbide Clusters Relative Energetics of the Bimetallic-Carbide Cluster Isomers Electronic Structure of the Bimetallic-Carbide Clusters Comparison of IP Trends for the Ta 3 ZrC y and Ta 4 C y Clusters Summary References 158 Appendix I 159 Appendix II 161 Appendix III 166 Appendix IV 189 Appendix V 193 Appendix VI 203 Appendix VII 205 XII

13 Chapter 6: Tantalum-Carbide Clusters Chapter 6: Tantalum-Carbide Clusters 6.1 Photo-ionisation Efficiency Experiments A mass spectrum of tantalum-carbide clusters ionised at 210 nm, under the SPI conditions used to conduct a PIE scan, is shown in Figure 6-1. Compared to the corresponding mass spectrum of niobium-carbide clusters (see Section 5.2), the mass distribution of tantalum-carbide clusters shows a lower amount of clustering for the Ta atoms. Ta 7 C y Ta 6 C y Intensity (Arb. Units) TaC y Ta 4 C y Ta 5 C y Ta 3 C y Mass (M/Z) Figure 6-1: Mass spectrum of ionised neutral Ta x C y clusters obtained by singlephoton ionisation (210 nm). Parts a e of Figure 6-2 show a portion of the mass spectra of tantalum-carbide clusters following ionisation at five different wavelengths; 262, 245, 230, 220 and 210 nm, collected under otherwise identical conditions. In the spectrum recorded at 210 nm (Figure 6-2e), clusters containing Ta 3 appear with 0 3 C atoms attached, clusters containing Ta 4 appear with 0 4 C atoms attached and clusters containing Ta 5 appear with 0 6 C atoms attached. The spectrum recorded at 262 nm (Figure 6-2a) shows only baseline levels, as none of the species are ionised with one photon. Following ionisation at 245 nm (Figure 6-2b) both Ta 5 C 2 and Ta 5 C 3 increase in intensity. At 230 nm (Figure 6-2c) both Ta 3 C 2 and 112

14 Chapter 6: Tantalum-Carbide Clusters Ta 4 C 4 have dramatically increased in intensity, with the clusters Ta 5 C 4, Ta 5 C 5 and Ta 5 C 6 also appearing. By the wavelength 230 nm the clusters Ta 3, Ta 3 C 3, Ta 4 C 2, Ta 5 and Ta 5 C have appeared (Figure 6-2d). Finally back to 210 nm (Figure 6-2e), the remaining clusters Ta 3 C, Ta 4, Ta 4 C and Ta 4 C 3 have increased from baseline intensity. PIE spectra are recorded by monitoring the signal of each cluster as a function of photon energy. PIE spectra for the Ta 3 C y (y = 0 3), Ta 4 C y (y = 0 4) and Ta 5 C y (y = 0 6) clusters are shown in parts a d of Figure 6-3, a e of Figure 6-4 and a g of Figure 6-5, respectively. Nearly all clusters show a dramatic rise from their baseline levels. Only Ta 4 C 4 shows a gradual onset of ionisation displaying slight structure, suggesting a notable geometry change between the neutral and cation. For all spectra two lines are fitted; one to the baseline and one to the linear rise of signal, and their intersection defined as the IP (with an estimated error of ±0.05 ev). This procedure was previously described in Section The determined IPs for all the tantalum-carbide clusters considered in this study are displayed in the PIE spectra and are also given later in the second column of Table 6-1, Table 6-2 and Table 6-3. No tantalumcarbide clusters have previously had their IPs examined. As a check, the ionisation energies extracted for the bare tantalum clusters Ta 3, Ta 4 and Ta 5 are found to be in good agreement with those previously determined. 1 Unlike the iso-valent bare niobium clusters, where the IPs decrease with size, the Ta 4 cluster has the highest IP, followed by Ta 3 and then Ta 5. In the Ta 3 C y series the addition of one and three C atoms results in very little change in the IP, relative to that of Ta 3 ; an increase and decrease of 0.08 and 0.08 ev, respectively. The addition of two C atoms results in an intermediate IP reduction of 0.35 ev. For the Ta 4 C y series, addition of one and three C atoms to Ta 4 results in very slight decreases in the IP, by 0.01 and 0.05 ev respectively. However, addition of two C atoms results in an intermediate IP reduction ( 0.25 ev), while addition of four C atoms results in a large IP reduction ( 0.68 ev). In the Ta 5 C y series addition of one C atom to Ta 5 essentially leaves the IP unchanged (an increase of 0.03 ev), yet addition of two and three C atoms results in large reductions in the IP of 0.64 ev and 0.60 ev, respectively. Addition of four C atoms causes an intermediate IP reduction ( 0.27), while five and six C atoms cause smaller IP reductions ( 0.15 and 0.13 ev, respectively). 113

15 Chapter 6: Tantalum-Carbide Clusters (a) : Ta 3 IP = 5.59 ev Cluster Signal Intensity (Arb. Units) a) 262nm b) 245nm c) 230nm Cluster Signal Intensity (Arb. Units) (b) : Ta C 3 IP = 5.67 ev (c) : Ta C 3 2 IP = 5.24 ev (d) : Ta C 3 3 IP = 5.51 ev d) 220nm Photon Energy (ev) Ta 3 C 2 Ta 4 C 2 Ta 5 C 2 e) 210nm Figure 6-3: Photo-ionisation efficiency spectra for the Ta 3 C y (y = 0 3) clusters. The determined IPs are also displayed Mass (M/Z) Figure 6-2: Mass spectra of Ta 3 C y, Ta 4 C y and Ta 5 C y clusters at four different ionisation wavelengths: (a) 262 nm, (b) 245 nm, (c) 230 nm, (d) 220 nm and (e) 210 nm. 114

16 Chapter 6: Tantalum-Carbide Clusters (a) : Ta 4 (a) : Ta 5 IP = 5.82 ev IP = 5.41 ev Cluster Signal Intensity (Arb. Units) (b) : Ta 4 C IP = 5.81 ev (c) : Ta 4 C 2 IP = 5.56 ev (d) : Ta 4 C 3 IP = 5.77 ev (e) : Ta 4 C 4 IP = 5.14 ev Cluster Signal Intensity (Arb. Units) (b) : Ta C 5 IP = 5.44 ev (c) : Ta C 5 2 IP = 4.77 ev (d) : Ta C 5 3 IP = 4.81 ev (e) : Ta C 5 4 IP = 5.14 ev Photon Energy (ev) (f) : Ta 5 C 5 IP = 5.26 ev Figure 6-4: Photo-ionisation efficiency spectra for the Ta 4 C y (y = 0 4) clusters. The determined IPs are also displayed. (g) : Ta 5 C 6 IP = 5.28 ev Photon Energy (ev) Figure 6-5: Photo-ionisation efficiency spectra for the Ta 5 C y (y = 0 6) clusters. The determined IPs are also displayed. 115

17 Chapter 6: Tantalum-Carbide Clusters 6.2 Comparison between Calculated and Experimental IPs For each tantalum-carbide cluster species DFT calculations are only performed on the low energy structural isomers examined with the extended basis set for the analogous niobium-carbide cluster (see Section 5.3). However, the tantalum-carbide cluster calculations are only conducted with the SDD basis set, not the extended basis set. Note that the isomer labels given (e.g. A, B, C, etc) reflect the energy ordering of the isomers. Therefore, an assigned isomer label can refer to a different structural isomer for a tantalum-carbide cluster species, compared to the niobium-carbide cluster species, if structural isomer energy reordering occurs between the iso-valent clusters. All the calculated isomers considered for the neutral Ta x C y clusters are shown in Figure 6-6 (Ta 3 C y, y = 0 3), Figure 6-8 (Ta 4 C y, y = 0 4), Figure 6-10 (Ta 5 C y, y = 0 4) and Figure 6-11 (Ta 5 C y, y = 5 and 6). The relative energies ( E in ev) for each isomer of the neutral species are also given in the figures. It is important to note that similar geometric minima for each isomer are also identified on the cationic surface. All details (i.e. geometric and energy information) for both neutral and cationic isomers examined are contained in Appendix IV and V. Identical to the procedure for the niobium-carbide clusters (Section 5.3), adiabatic ionisation transitions are considered for the lowest energy isomer, as well as the lowlying isomers, for each tantalum-carbide cluster species. Similarly, only ionisation energies including ZPE are considered for discussion. As observed for the Nb x C y clusters, the absolute calculated IPs for all the Ta x C y isomers are higher than the experimental values. Identical to the procedure for the Nb x C y clusters, a linear offset is applied to the IPs of the Ta x C y isomers to generate an offset value, IP. These ionisation energies are listed for the Ta 3 C y (y = 0 3), Ta 4 C y (y = 0 4) and Ta 5 C y (y = 0 6) clusters in Table 6-1, Table 6-2 and Table 6-3, respectively. In addition, calculated IP s for each isomer and experimental IPs for each of the Ta 3 C y (y = 0 3), Ta 4 C y (y = 0 4) and Ta 5 C y (y = 0 6) cluster series are also shown in Figure 6-7, Figure 6-9 and Figure 6-12, respectively. 116

18 Chapter 6: Tantalum-Carbide Clusters The Ta 3 C y (y = 0 3) Cluster Series For the Ta 3 C y cluster series a large offset of ev is applied so that the calculated and experimental IPs of Ta 3 overlap. The resultant IP is the value shown in final column of Table 6-1, while Figure 6-7 shows the calculated IP s of the isomers ( ) for each Ta 3 C y species, as well as the experimental IP ( ). Figure 6-6: Structures of calculated isomers for the neutral Ta 3 C y (y = 0 3) clusters. Written beneath each isomer are the relative energies ( E in ev) calculated using the SDD basis set. For Ta 3 an obtuse triangle of high spin multiplicity was found to be the lowest energy structure (XX A [ 4 A 2, C 2v ]), consistent with the proposed non-equilateral triangular structure from Raman studies of Ta 3 deposited in an argon matrix. 2 For Ta 3 C, isomer XXI A [ 2 A, C s ] has the C atom bound to a Ta face of the triangular Ta 3 cluster, whereas in isomer XXI B [ 2 A 1, C 2v ] the C atom is bound across a Ta-Ta edge. Both isomers are essentially equal in energy (XXI B E = ev) but have quite different calculated IP s; XXI A has a deviation of ev between its IP 117

19 Chapter 6: Tantalum-Carbide Clusters and the experimental value, whereas XXI B has a larger deviation of ev. Therefore, isomer XXI A is assigned to the experimental IP. Cluster Expt. IP Isomer Calc. Transition Calc. IP (exc. ZPE) Calc. IP (inc. ZPE) Calc. IP Ta XX A 3 B 1 4 A Ta 3 C 5.67 XXI A XXI B 3 A 2 A A 1 2 A Ta 3 C XXII A 1 A 1 2 A Ta 3 C XXIII A XXIII B 1 A 2 A A 2 A Table 6-1: List of experimental ionisation potentials (reported in ev) observed for Ta 3 C y (y = 0 3) clusters. Also listed are calculated transitions and ionisation potentials: excluding ZPE, including ZPE, and offset IP (i.e. IP ). 6.5 B Experimental IP (ev) Ta 3 C y Expt. IP B A A Calculated Offset IP (ev) Ta 3 C y Calc. IP A No. Carbon atoms (y) Figure 6-7: Graph showing experimental IP values for the Ta 3 C y clusters as a function of y. Also shown on the same scale are the offset values, IP, calculated using DFT. The letters (A, B, etc.) denote the isomers for that particular cluster (see text for details). 118

20 Chapter 6: Tantalum-Carbide Clusters The only isomer considered for Ta 3 C 2, XXII A [ 2 A, C s ], is that where both C atoms are bound to opposite Ta faces of the triangular Ta 3 cluster. Its calculated IP is in excellent agreement with the experimental value, with a deviation of ev. The isomer XXIII A [ 2 A, C s ] of Ta 3 C 3 is much lower in energy than XXIII B [ 2 A, C 1 ] ( E = ev) and also has a calculated IP only ev higher in energy than the experimental IP and so is assigned as the carrier of the observed ionisation onset. Isomer XXIII A has one C atom bound to a Ta face and the remaining two C atoms bound across separate Ta-Ta edges of the triangular Ta 3 cluster The Ta 4 C y (y = 0 4) Clusters Series For the Ta 4 C y cluster series an offset of ev is applied so that the calculated and experimental IPs of Ta 4 overlap. The resultant IP is the value shown in final column of Table 6-2. Figure 6-9 shows the calculated IP s of the isomers ( ) for each Ta 4 C y species, as well as the experimental IP ( ). The lowest energy structure for Ta 4 is an ideal tetrahedron (XXIV A [ 1 A 1, T d ]). The two isomers of Ta 4 C have the C atom either bound to a Ta face (XXV A [ 1 A, C s ]) or Ta-Ta edge (XXV B [ 1 A 1, C 2v ]) of the tetrahedral Ta 4 cluster. As both isomers calculated IP s are in excellent agreement with the experimental value (0.039 and ev higher in energy, respectively), XXV A is assigned to ionisation onset as it is ev lower in energy. This is not a definitive assignment though, due to the small E of XXV B. The two isomers of Ta 4 C 2 are separated by ev; XXVI A [ 3 A 2, C 2v ] being lower in energy than XXVI B [ 1 A 1, C 2v ]. Isomer XXVI A is assigned as the carrier of the experimental IP as the deviation between its IP and the experimental value ( ev) is much smaller than that of XXVI B ( ev). XXVI A has both C atoms bound to separate Ta faces of the tetrahedral Ta 4 cluster and has a high spin multiplicity. 119

21 Chapter 6: Tantalum-Carbide Clusters The IP of Ta 4 C 3 XXVII A [ 1 A 1, C 3v ] is in excellent agreement with the experimental IP (a deviation of ev). This is the only isomer considered for Ta 4 C 3 and has the C atoms bound to three of the four available Ta faces of the tetrahedral Ta 4 cluster; i.e. a C deficient nanocrystal. Figure 6-8: Structures of calculated isomers for the neutral Ta 4 C y (y = 0 4) clusters. Written beneath each isomer are the relative energies ( E in ev) calculated using the SDD basis set. The nanocrystal isomer of Ta 4 C 4 XXVIII A [ 3 A, C s ] has a calculated IP only ev higher in energy than the experimental IP and is consequently assigned as the carrier of the ionisation onset. This is consistent with the structure proposed for Ta 4 C 4 based on its IR-REMPI spectrum. 3 Note that this isomer has a triplet electronic state, which is lower in energy (0.686 ev) than its singlet state (see Appendix IV). 120

22 Chapter 6: Tantalum-Carbide Clusters Cluster Expt. IP Isomer Calc. Transition Calc. IP (exc. ZPE) Calc. IP (inc. ZPE) Calc. IP Ta XXIV A 2 A 1 1 A Ta 4 C 5.81 XXV A XXV B 2 A 1 A A 1 A Ta 4 C XXVI A XXVI B 2 A 3 A B 2 1 A Ta 4 C XXVII A 2 A 1 1 A Ta 4 C XXVIII A 2 B 1 3 A Table 6-2: List of experimental ionisation potentials (reported in ev) observed for Ta 4 C y (y = 0 4) clusters. Also listed are calculated transitions and ionisation potentials: excluding ZPE, including ZPE, and offset IP (i.e. IP ). 6.0 B Experimental IP (ev) 5.5 Ta 3 C y Expt. IP B A A A A Calculated Offset IP (ev) Ta 3 C y Calc. IP No. Carbon atoms (y) Figure 6-9: Graph showing experimental IP values for the Ta 4 C y clusters as a function of y. Also shown on the same scale are the offset values, IP, calculated using DFT. The letters (A, B, etc.) denote the isomers for that particular cluster (see text for details). 121

23 Chapter 6: Tantalum-Carbide Clusters The Ta 5 C y (y = 0 6) Cluster Series For the Ta 5 C y cluster series an offset of ev is applied so that the calculated and experimental IPs of Ta 5 overlap. The resultant IP is the value shown in final column of Table 6-3, while Figure 6-12 shows the calculated IP s of the isomers ( ) for each Ta 5 C y species, as well as the experimental IP ( ). Figure 6-10: Structures of calculated isomers for the neutral Ta 5 C y (y = 0 4) clusters. Written beneath each isomer are the relative energies ( E in ev) calculated using the SDD basis set. 122

24 Chapter 6: Tantalum-Carbide Clusters Figure 6-11: Structures of calculated isomers for the neutral Ta 5 C y (y = 5 and 6) clusters. Written beneath each isomer are the relative energies ( E in ev) calculated using the SDD basis set. The lowest energy structure for Ta 5 is a distorted trigonal bipyramid (XXIX A, [ 2 B 1, C 2v ]). The two Ta 5 C isomers XXX A [ 2 A, C s ] and XXX B [ 2 A, C 1 ] are very close in energy, with the former being ev lower in energy. XXX A has the C atom bound across a Ta butterfly motif of the oblate trigonal bipyramid Ta 5 cluster, while XXX B has the C atom bound to a Ta face of the trigonal bipyramid Ta 5 cluster. A definitive assignment of the experimental IP is not possible as both the calculated IP s of XXX A and XXX B are in excellent agreement with the experimental value; deviations of and ev, respectively. The lowest energy isomer (XXX A) is therefore assigned to the ionisation onset. For Ta 5 C 2, isomer XXXI A [ 2 B 1, C 2v ] is ev lower in energy than XXXI B [ 2 A, C s ]. Isomer XXXI A has the two C atom bound across separate Ta butterfly motifs of the oblate trigonal bipyramid Ta 5 cluster and is assigned to the experimental IP as its 123

25 Chapter 6: Tantalum-Carbide Clusters calculated IP is only ev higher, whereas the IP of XXXI B has a much higher deviation ( ev). Cluster Expt. IP Isomer Calc. Transition Calc. IP (exc. ZPE) Calc. IP (inc. ZPE) Calc. IP Ta XXIX A 3 A 1 2 B Ta 5 C 5.44 XXX A XXX B 1 A 2 A A 2 A Ta 5 C XXXI A XXXI B 1 A 1 2 B A 2 A Ta 5 C XXXII A 1 A 1 2 A Ta 5 C XXXIII A XXXIII B XXXIII C XXXIII D 3 A 2 A A 2 A A 1 2 B A 2 B Ta 5 C XXXIV A XXXIV B XXXIV C XXXIV D 1 A 2 A A 2 A A 2 A A 2 A Ta 5 C XXXV A XXXV B 1 A 1 2 A A 1 2 B Table 6-3: List of experimental ionisation potentials (reported in ev) observed for Ta 5 C y (y = 0 6) clusters. Also listed are calculated transitions and ionisation potentials: excluding ZPE, including ZPE, and offset IP (i.e. IP ). The only isomer considered for Ta 5 C 3, XXXII A [ 2 A, C s ], has a calculated IP in excellent agreement with the experimental IP (with a deviation of ev). This isomer has the C atoms bound to all three of the available Ta faces which share a common axial Ta atom of the trigonal bipyramid Ta 5 cluster. This isomer can also be considered as a substituted nanocrystal, where a C atom is replaced with a Ta atom. This is consistent with the structure proposed for Ta 5 C 3 based on its IR-REMPI spectrum

26 Chapter 6: Tantalum-Carbide Clusters 6.0 B A B Experimental IP (ev) B A Ta C Expt. IP 5 y Ta C Calc. IP 5 y B A A C B D D A C A Calculated Offset IP (ev) No. Carbon atoms (y) Figure 6-12: Graph showing experimental IP values for the Ta 5 C y clusters as a function of y. Also shown on the same scale are the offset values, IP, calculated using DFT. The letters (A, B, etc.) denote the isomers for that particular cluster (see text for details). Four isomers are examined for Ta 5 C 4 ; XXXIII A [ 2 A, C s ], XXXIII B [ 2 A, C s ], XXXIII C [ 2 B, C 2 ] and XXXIII D [ 2 A 1, C 2v ]. The lowest energy isomer (XXXIII A) has the C atoms bound to four of the Ta faces of the trigonal bipyramid Ta 5 cluster, where three of the Ta faces share a common axial Ta atom. However, this isomer does not have a calculated IP which agrees with the ionisation onset, being ev higher in energy than the experimental value. Of the remaining isomers only XXXIII B and XXXIII D have IP s in good agreement with the experimental IP (see Figure 6-12) with deviations of and ev, respectively. The observed ionisation onset is assigned to XXXIII B, as it is much lower in energy than XXXIII D; E = and ev, respectively. The structure of XXXIII B can be considered as a Ta atom bound above a C atom of a nanocrystal. Four isomers are also examined for Ta 5 C 5 ; XXXIV A [ 2 A, C s ], XXXIV B [ 2 A, C 1 ], XXXIV C [ 2 A, C s ] and XXXIV D [ 2 B 1, C 2v ]. Isomers XXXIV B and XXXIV D are discounted as their IP s are not in reasonable agreement with the experimental value; deviations of and ev, respectively. Both isomers XXXIV A and 125

27 Chapter 6: Tantalum-Carbide Clusters XXXIV C have calculated IP s in excellent agreement with the experimental value (see Figure 6-12), being only and 0.39 ev higher in energy, respectively. Therefore, the calculated IP s cannot distinguish between XXXIV A, which has all C atoms bound to separate Ta faces of the trigonal bipyramid Ta 5 cluster, and XXXIV C, which contains a C 2 unit. As XXXIV A is calculated to be the lowest energy isomer by ev, it is assigned to the experimental ionisation onset. The lowest energy isomer of Nb 5 C 6 (XXXV A [ 2 A 1, C 2v ]) has an acetylide C 2 unit bound across each of the two Ta butterfly motifs of the oblate trigonal bipyramid Ta 5 cluster, with the remaining two C atoms bound to the two outer Ta faces. Isomer XXXV A is assigned to the ionisation onset as its calculated IP is in much better agreement with the experimental IP (a deviation of ev) compared to the higher energy isomer XXXV B, which has an IP deviation of ev. 126

28 Chapter 6: Tantalum-Carbide Clusters 6.3 Discussion on Tantalum-Carbide Clusters Comparison to the Niobium-Carbide Clusters The extracted IPs for the tantalum-carbide clusters in each series show trends that are very similar to the iso-valent niobium-carbide cluster series, although the IP reductions upon addition of C atoms are greater for the latter. Furthermore, there are other differences to note between the tantalum-carbide and niobium-carbide clusters. An ionisation onset is observed for Ta 3 C but not for Nb 3 C, since it was beyond the range of our experiment (> 5.91 ev). This is attributed to the fact that the in-plane structure (i.e. M-M edge bound C atom) is the global minimum for Nb 3 C and has a much higher IP than the out-of-plane isomer (i.e. M face bound C atom). Conversely, the lowest energy isomer of Ta 3 C is the out-of-plane structure. Additionally, ionisation onsets are not observed for Ta 3 C 4, Ta 4 C 5 and Ta 4 C 6. It is not clear if these species are simply not generated under our experimental conditions or whether there are other factors at play. For example, it may be that for Ta 3 C 4 the lowlying structural isomer assigned to the ionisation onset of Nb 3 C 4, which has a C 2 unit and a much lower IP than the global minimum, is not generated. While for Ta 4 C 5 and Ta 4 C 6 the higher spin states may not be metastable, as proposed for Nb 4 C 5 and Nb 4 C 6. Apart from Ta 5 C and Ta 5 C 5, all other Ta x C y species identified have the same structural isomer assigned to their ionisation onsets as those assigned for the corresponding Nb x C y species. However, for both these metal-carbide cluster species in the Ta and Nb cases, two isomers are identified which both have IPs in excellent agreement with the experimental values. Therefore, the assignments are made to the lowest energy isomer, with the energy ordering of the two structural isomers swapping between Ta and Nb. In the example of M 5 C 5 (for equivalent basis sets), the isomer containing a C 2 unit is significantly higher in energy relative to the isomer containing no C-C bonding for Ta 5 C 5, yet for Nb 5 C 5 the two isomers are essentially equal in energy. Therefore, Ta-C bonding appears to be slightly more energetically favorable than Nb-C bonding, 127

29 Chapter 6: Tantalum-Carbide Clusters relative to C-C bonding. This point is consistent with the fact that the Ta 8 C 12 Metcar (which is expected to contain 6 C 2 units) has not been generated. 4 This latter observation was explained to be because of slightly stronger Ta-Ta bonding, relative to Nb-Nb bonding; i.e. high C:Ta ratios for Ta x C y clusters impose detrimental disruptions to Ta-Ta bonding. 5,6 128

30 Chapter 6: Tantalum-Carbide Clusters 6.4 Summary In summary, this chapter has shown that the isomeric structures of tantalum-carbide clusters can be inferred by the determination of their IPs by PIE experiments in combination with DFT calculations on candidate isomers. The IP trends in each tantalum-carbide cluster series are very similar to the iso-valent niobium-carbide cluster series, although the IP reductions upon addition of C atoms are greater for the latter. For the vast majority of tantalum-carbide cluster species, the same structural isomer is assigned to the ionisation onset as those assigned for the corresponding niobiumcarbide cluster species. For the Ta 3 C y and Ta 4 C y cluster series these isomers correspond to fragments of the nanocrystal structure for Ta 4 C 4. The only cluster which has an isomer containing C-C bonding assigned to its ionisation onset is Ta 5 C 6, with the structure containing two C 2 units. The next chapter will describe the generation of novel bimetallic tantalum-zirconiumcarbide clusters, which are examined though essentially identical PIE experiments and DFT calculations. The IP measurements and structural assignments made in this chapter for the Ta 4 C y clusters will serve as a comparison for the effects occurring upon substitution of a Ta atom with a Zr atom to generate Ta 3 ZrC y clusters. 129

31 Chapter 6: Tantalum-Carbide Clusters 6.5 References (1) Collings, B. A.; Rayner, D. M.; Hackett, P. A. Int. J. Mass. Spec. Ion Proc. 1993, 125, 207. (2) Fang, L.; Shen, X.; Chen, X.; Lombardi, J. R. Chem. Phys. Lett. 2000, 332, 299. (3) van Heijnsbergen, D.; Fielicke, A.; Meijer, G.; von Helden, G. Phys. Rev. Lett. 2002, 89, (4) Wei, S.; Guo, B. C.; Purnell, J.; Buzza, S.; Castleman Jr, A. W. Science 1992, 256, 818. (5) Cartier, S. F.; May, B. D.; Castleman Jr, A. W. J. Phys. Chem. 1996, 100, (6) Armentrout, P. B.; Hales, D. A.; Lian, L. Collision-Induced Dissociation of Transition-Metal Cluster Ions. In Cluster Reactions; Duncan, M. A., Ed.; Jai Press Inc.: Greenwich, CT, 1994; Vol. 2; pp

32 Chapter 7: Tantalum-Zirconium-Carbide Clusters Chapter 7: Tantalum-Zirconium-Carbide Clusters The following chapter is concerned with the development of a double ablation cluster source to generate bimetallic-carbide clusters of suitable composition for experimental PIE studies. As will be shown, mixing of tantalum (Group 5) and zirconium (Group 4) transition metals to generate bimetallic-carbide clusters is achieved with this cluster source. Bimetallic-carbide clusters containing three tantalum atoms and one zirconium atom (i.e. Ta 3 ZrC y ) are investigated through a combination of PIE experiments and DFT calculations on their energetically favourable isomers. 7.1 Introduction to Bimetallic-Carbide Clusters Bimetallic-carbide clusters are interesting systems to study because not only their size, but also their composition, can be systematically varied. Constructing bimetalliccarbide clusters of suitable composition may tailor new cluster properties, which are not possessed by any monometallic-carbide clusters. Extensive studies have been performed on the properties and reactivity of bare bimetallic clusters; notably by the separate research groups of Kaya, 1-3 Knickelbein, 4,5 Silverans and Lievens, 6-8 and Wang. 9,10 However, experimental studies on bimetallic-carbide clusters have been limited to those performed by Castleman and co-workers on the Metcar species. Their initial studies focused on the generation of the titanium Metcar, where titanium atoms were substituted with various alternative metal atoms; i.e. Ti 8 x M x C 12, where M = Zr, Hf, Y, Nb, Mo, Ta or W. 11,12 Zirconium (which is in the same group as titanium) displayed the greatest degree of substitution, with up to five atoms incorporated into the bimetallic Metcar. Further studies examined the delayed ionisation and delayed atomic ion emission of the Ti 8 x M x C 12 clusters, where M = Zr or Nb and x = 0 4, 13 and also determined the IPs by PIE experiments of the Ti 8 x Zr x C 12 clusters (where x = 0 4 and 8). 14 These latter experiments showed that the IP of the titanium Metcar (4.40 ev) is higher than that of the zirconium Metcar (3.95 ev), with the IP of the bimetallic Metcar clusters decreasing continuously from the pure titanium Metcar 131

33 Chapter 7: Tantalum-Zirconium-Carbide Clusters towards that of the pure zirconium Metcar, as the number of substituting zirconium atoms increased. Poblet, Rohmer, Benard and co-workers used DFT to investigate the bimetallic Metcar clusters generated experimentally by Castleman and co-workers. 15 For the Ti 8 x Zr x C 12 (x = 0 5) clusters, two isomers were considered for each stoichiometry; i.e. the metal with the lowest number of atoms present would have all its atoms located in either the inner tetrahedron positions or the larger outer tetrahedron positions, but never a combination of the two. The difference in energy between the two isomers for each cluster was found to be very small (never exceeding ev). The similar energy for substitution in either of the two different metallic positions in the titanium Metcar explained the regular statistical distribution of the peaks in the mass spectrum of Ti 8 x Zr x C 12 (x = 0 5), which had initially been interpreted as evidence that all eight of the metal atoms in Ti 8 C 12 were located in symmetry equivalent positions. 11 Of interest to the present study are investigations performed on metal-oxide cluster anions. IR-MPD spectra were obtained for vanadium-oxide cluster anions of the form (V 2 O 5 ) x. 16 In comparison with DFT calculations, the structures were shown to be based on polyhedral cages. Further studies performed on bimetallic vanadiumtitanium-oxide cluster anions of stoichiometry (V 2 O 5 ) x (VTiO 5 ) (x = 1 3) showed that they also have the same structures as the corresponding pure vanadium-oxide cluster anions and those predicted for the iso-electronic neutral vanadium-oxide clusters. 17 Additional investigations on the V 4 x Ti x O 10 (x = 1 4) clusters showed that these species also possessed the same polyhedral caged structure as V 4 O 10. Overall this work showed that substitution of a Group 5 transition metal atom with a Group 4 transition metal atom does not change the structure of the metal-oxide clusters in this size range. This is relevant to the current study as a similar substitution regime is performed, albeit for metal-carbide clusters. 132

34 Chapter 7: Tantalum-Zirconium-Carbide Clusters 7.2 Double Ablation Cluster Source Design Bimetallic clusters can be generated by either ablation of an alloy target with varying molar composition or by mixing the ablated products from two separate metal targets. The ablation of separate metal targets is the approach chosen for our double ablation cluster source. This technique is much more flexible as any combination of elements can in principle be used and the power and timing of the separate ablation lasers can be varied individually to optimise generation of the desired bimetallic cluster. The double ablation source design (Figure 7-1) is similar to that originally proposed by Kaya and co-workers 1 and is based on our previous single ablation source (see Section 2.1.2), except that two 2 mm diameter rods are utilised instead of one 5 mm diameter rod. In the double ablation cap the two metal rods are displaced by a distance of 5 mm along the gas channel. The two rods are situated on opposite sides of the gas channel. Therefore, the two rods are ablated from opposite sides of the ablation cap. Both rods in the double ablation source are rotated and translated by the same screw mechanism motor in conjunction with a gear mechanism. The gas channel in the double ablation cap has a 3 mm internal diameter, increased from the 2 mm diameter in the single ablation setup. Figure 7-1: Double ablation source used to generate bimetallic-carbide clusters shown in the (a) side-on and (b) isometric orientation. Dimensions given in mm. 133

35 Chapter 7: Tantalum-Zirconium-Carbide Clusters The condensation tube (10 mm in length) is also slightly modified, with a screw thread tapped into the end of the inner channel so that further tube attachments of alternating lengths can be connected and interchanged easily. The internal diameter of the double ablation condensation tube and attachments is maintained at 2 mm. In the current experiment a condensation tube attachment of 5 mm is connected, making the total condensation tube length 15 mm (identical to that in the single ablation experiments). To improve the amount of generated clusters reaching the ionisation region the length of the hollow extension cylinder was increased to 15.2 cm, making the output of the condensation tube now only ~10 cm away from the skimmer. The two rods are ablated with focussed 532 nm laser pulses originating from separate Nd:YAG laser systems. The relative timing of each ablation laser is controlled with the digital pulse generator, with the actual separation time between ablation events confirmed with a fast photodiode. 134

36 Chapter 7: Tantalum-Zirconium-Carbide Clusters 7.3 Generation of Tantalum-Zirconium-Carbide Clusters Bimetallic-carbide clusters are chosen to be constructed from tantalum (Group 5) and zirconium (Group 4) transition metals. Generating tantalum-zirconium-carbide clusters containing one or more Zr atoms, with a total number of 3 5 metal atoms, are desired as the IPs and structures of tantalum-carbide clusters containing 3 5 tantalum atoms were determined in Chapter 6. The primary difficulty in generating bimetallic-carbide clusters composed of Ta and Zr atoms is that the monometallic-carbide clusters may overlap in mass with the bimetallic-carbide clusters, rendering the PIE technique unable to be applied. This overlap is primarily because the mass of a Zr atom (isotope pattern: 90 Zr 51.45%, 91 Zr 11.22%, 92 Zr 17.15%, 94 Zr 17.38% and 96 Zr 2.80 %) is approximately half that of a Ta atom ( 181 Ta 99.99%). For example, Ta x Zr z C y clusters containing one Zr atom will overlap with Zr x C y clusters containing an odd number of Zr atoms, while Ta x Zr z C y clusters containing two Zr atoms will overlap with Zr x C y clusters containing an even number of Zr atoms and Ta x C y clusters. This overlap is likely to be occur under SPI conditions as the Zr x C y and Ta x Zr z C y clusters are expected to have IPs in a similar energy range to that of the Ta x C y clusters, as the IPs of the Zr atom (6.63 ev) and Zr 3 cluster (5.22 ev) 18 are lower than that of the corresponding Ta atom (7.89 ev) and Ta 3 cluster (5.59 ev), respectively. In consideration of the desired bimetallic-carbide clusters to be generated and the above points, tantalum-zirconium-carbide clusters which only contain one Zr atom are the most practical candidates, as the only source of mass interference will be large zirconium-carbide clusters containing an odd number of Zr atoms (e.g. the Zr 7 C y clusters will interfere with the Ta 3 ZrC y clusters). Figure 7-2 shows the mass spectrum of Zr x C y clusters generated from the ablation of the Zr rod, under essentially identical conditions to those described for generation of Nb x C y and Ta x C y clusters (see Section 2.1.2), although now a slightly higher ablation power is used (~8 mj). This mass spectrum is obtained with both the Zr and Ta rods in the double ablation source, yet only the Zr rod is ablated. The mass distribution is 135

37 Chapter 7: Tantalum-Zirconium-Carbide Clusters similar to that observed for the Nb x C y and Ta x C y clusters (see Sections 5.2 and 6.1), although the present spectrum is significantly congested due to the isotope pattern of the Zr atom. Clusters containing oxygen are also present due to an oxide layer on the Zr rod. Overall from this mass spectrum it can be seen that large Zr x C y clusters are readily generated and may overlap with the desired Ta x ZrC y clusters. Intensity (Arb. Units) ZrC y Zr 2 C y Zr 3 C y Zr4 C y Zr 6 C y Zr C 10 y Zr C 9 y Zr C 8 y Zr C 7 y Zr 5 C y Mass (M/Z) Figure 7-2: Mass spectrum of ionised neutral Zr x C y clusters obtained by singlephoton ionisation (220 nm). Initial generation of Ta x ZrC y clusters is approached with the Ta rod in the first position (i.e. closest to the pulsed nozzle) and the Zr rod in the second position of the double ablation cap. As Ta rich Ta x Zr z C y clusters are desired, the Ta rod is ablated first (~ 6 mj), on the leading edge of the gas pulse. The Zr rod, located 5 mm away, is then ablated after a time delay with a lower laser power (~5 mj) than applied to obtain the above mass spectrum of Zr x C y clusters. This time delay (~ 6 µs) is employed so that the ablated Ta and Zr products (the former being more abundant) are overlapped to facilitate mixing. Figure 7-3 shows the mass spectrum of the clusters generated by the procedure described above. Ionisation is conducted at 220 nm, under SPI conditions. As can be seen it is very similar in appearance to the addition of the two mass spectra of the separately ablated Zr and Ta (see Section 6.1) rods. A threshold ablation power (~5 mj) is found for the ablation of the Zr rod, with large Zr x C y clusters rapidly appearing 136

38 Chapter 7: Tantalum-Zirconium-Carbide Clusters above this power, which makes controlling the concentration of Zr products difficult. Any Ta x ZrC y clusters generated by this procedure are not produced in significant concentration, as blocking the Ta ablation laser has no significant effect on peaks present in their expected mass regions, demonstrating that these peaks are primarily Zr x C y clusters. However, Ta x ZrC y clusters are generated in substantial concentration, without having interference from large Zr x C y clusters, by a procedure of changing the relative timing and power of the ablation lasers. Here the Zr rod is ablated ~7 µs earlier than the Ta rod, with both rods ablated with ~ 6 mj and the pulsed nozzle and Ta ablation laser still triggered at the same relative time as before. By applying this procedure it is found that the distribution of the clusters differs at various regions in the molecular beam. These different regions are investigated by changing the timing of the ionisation laser. Ionisation is similarly conducted at 220 nm, under SPI conditions. Parts a e of Figure 7-4 show mass spectra recorded at 10 µs firing intervals of the ionisation laser. Time zero is arbitrarily set to that used when recording PIE spectra. The mass spectrum at 30 µs (Figure 7-4a) primarily consists of only Zr x C y clusters. The next mass spectrum at 20 µs (Figure 7-4b) also primarily consists of Zr x C y clusters, but now a new large peak appears at mass 567, due to Ta 3 C 2. Next at 10 µs (Figure 7-4c) there is a decrease in Zr x C y clusters but an increase in Ta x C y clusters observed. Furthermore, the peaks in the mass region of the Ta x ZrC y (x = 3 5) cluster have become sharper, indicating that these peaks are now bimetallic-carbide clusters rather than Zr x C y clusters, which are broader due to the many combinations of the Zr atom isotopes. The next mass spectrum at 0 µs (Figure 7-4d) primarily consists of Ta x C y clusters and Ta x ZrC y (x = 3 5) clusters. Interestingly, bimetallic-carbide clusters of the form Ta 2 ZrC y are not observed with any great intensity, possibly as their IPs are greater than the photon energy available at 220 nm. Negligible amounts of Zr x C y clusters are now present, as evident by the absence of substantial low mass peaks. An expanded larger version of this 0 µs mass spectrum is also shown in (Figure 7-6). The last mass spectrum at +10 µs (Figure 7-4e) primarily consists of only Ta x C y clusters, with small amounts of Ta x ZrC y clusters still present. 137