Work based on this was published as
  SPIE Proceedings, 1994, 2124, 418-425
  Laser Techniques for State-Selected and State-to-State Chemistyr II,
  John W. Hepburn, Ed.

Photodissociation spectroscopy of metal clusters:
a status report

Douglas Cameron, Susanne Haupt, Julian Kaller and Manfred M. Kappes
Institut für Physikalische Chemie, Universität Karlsruhe
76128 Karlsruhe, Germany


Abstract. Photodissociation spectroscopy has been applied to the characterization of the electronic and geometric structure of cold silver cluster cations (Agx+, x<22). Preliminary results are described.

Ag9 Dissociation

1. Introduction

        Characterization and rationalization of the optical properties of small bare metal clusters (neutral and charged) under rigorously particle size specific conditions can provide detailed insight into the systematics of cluster size dependent bonding1. This is of use in a number of areas ranging from basic (electron correlation and localization in finite systems2) to applied (catalytic properties of supported transition metal clusters3). Alkali clusters (primarily lithium and sodium) are probably the most extensively studied systems in this regard due to their comparative theoretical tractability. Consequently there have been numerous extensive studies in which observed spectra4 have been rationalized in terms of high level calculations describing electronic excitation from specific isomeric structures (and incorporating electron correlation to high order) 5. In a number of cases, notably for tetramer band spectra, it has been possible to assign cluster structures in this fashion6. Laser photodissociation spectroscopy has been the experimental method of choice in such studies, although femtosecond resonant two-photon ionization spectroscopy has recently been making significant inroads into the area7. Photoelectron and photodetachment spectroscopies have also been applied successfully to various aspects of the alkali cluster story8.

        Generally, spectroscopic probes of alkali clusters have suffered from the problem that the intense continuous sources "traditionally" used, produce clusters at comparatively high levels of internal excitation9. This coupled with a low excitation threshold for such weakly bound clusters to undergo large amplitude motion10, has so far made it impossible to obtain vibrationally resolved spectra beyond trimers. Furthermore the fluxionality issue has led researchers to interpret the band spectra of larger clusters cautiously with regard to specific cluster structures/isomers. In retrospect, this turns out to be warranted: more recent measurements on mass selected sodium cluster cations prepared using a new low temperature cold cathode discharge source, show strongly internal excitation dependent optical signatures11 (without yet resolving vibrations). Such results will open up a new round of experiment/theory interaction now requiring a description of cluster dynamics12. In spite of some elegant ionization potential measurements on sodium (and lithium) clusters at nominally liquid nitrogen temperatures, in which aggregates were prepared using a laser vaporization source13, it still remains for alkali clusters with more than three atoms to be generated at internal excitation levels low enough to resolve individual vibrations (let alone rotations).

        Laser vaporization can be much more conveniently applied to the generation of cold beams of the more refractory coinage metal clusters14--16. As a result, there have already been a number of gas-phase vibrationally resolved spectroscopic studies (R2PI and photodissociation) of neutral14,17 and charged18 coinage metal trimers and tetramers - primarily of silver and copper. While more difficult to describe theoretically than alkali clusters, detailed spectra of coinage metal clusters (in particular of the least d-electron perturbed coinage metal silver) are of interest as a stepping stone towards a quantitative treatment of d-electron bonding in transition metals, an area of intense activity in computational quantum chemistry19. In addition to the trimer and tetramer measurements, there also exists an extensive (vibrationally as yet unresolved) body of gas-phase data obtained for larger charged coinage metal clusters. These have been studied either as anions by photodetachment spectroscopy20 or as cations by photodissociation18,21. In the latter case, relatively hot silver cluster cations prepared using a continuous sputter source have been most extensively probed21. This data set is complementary to recently reported electronic absorption spectra obtained at 4 K for size selected neutral silver clusters deposited into inert gas matrices22. While vibrational structure cannot be resolved in the latter measurements either (due in this case to matrix effects), the electronic bandwidths are as much as an order of magnitude narrower than found in corresponding gas phase spectra of silver species having the same valence electron count. This holds out hope for systematic gas phase photodissociation measurements on cold silver cluster cations. Below we present some preliminary results from such a project.

2. Experimental

Figure 1. Schematic of the apparatus used to observe silver cluster ion photofragmentation. The pulsed valve at A controls He gas that carries silver, vaporized by an excimer laser near B, through the nozzle and into the source chamber as a molecular beam. This beam is skimmed and the ions in it are extracted by pulsed electric fields at C into a reflectron time of flight mass spectrometer. To photofragment only one size cluster, a mass gate at D that consists of wires normally at a few hundred volts and caged in a shielding metal box, is briefly grounded to permit the passage of the cluster size of interest. After passing through this gate the cluster packet is collimated and irradiated near E with the dissociation laser that passes through the reflecting fields F of the mass spectrometer and end window. Clusters that do not fragment and charged fragments then follow different trajectories within the reflecting field and reach G, the detector, at different times. The fluence of the dissociation laser is monitored with a photodiode placed after attenuating filters at H.

       Figure 1 shows a schematic of the experimental configuration used. Photodissociation of cluster cations was carried out in a two-stage molecular beam machine comprising a 1000 l sec-1 oil diffusion pumped source chamber and a 500 l sec-1 turbomolecular pumped detection chamber. Cluster ions were generated by laser vaporization into a helium pulse (20 bar stagnation pressure) using a source configuration modified slightly relative to that previously described15,23. The modifications consisted of mounting a standard pulsed valve external to the vacuum system and installing separate (flowing) liquid nitrogen cooling coils for source block and nozzle. Source block and nozzle were thermally insulated from the mounting tube connecting them to the pulsed valve and vacuum chamber by using quartz spacers. With this setup it was possible to run the source stably at (nominal) temperatures from 140 to 298 K as measured by embedding a thermocouple into the nozzle. For the experiments reported a 2mm diameter, 40mm length cylindrical equilibration channel terminating in a 10 degree conical nozzle was used. We have yet to systematically study the dependence of cluster internal excitation on source conditions. As in the previous ionization potential study, laser vaporization of silver was carried out using an XeCl (308nm) excimer laser15. In order to obtain high cluster cation yields it was found necessary to use higher irradiation fluxes (50-80 mJ per pulse focused onto 1 mm2) than optimal for neutral clusters.

        Ions generated were detected in a 1 m drift tube reflectron time-of-flight mass spectrometer (RETOF) mounted perpendicular to the cluster beam. The RETOF was equipped with a pulsed extraction field and a (pulsed) mass gate positioned approximately halfway down the drift tube. Cluster ions were extracted at delays for which maximum signal levels could be obtained. In order to reduce pump oil residuals generated by multiphoton processes the reflectron field was pulsed up from ground just prior to entrance of the irradiated mass selected ion packet. Mass selected cluster ions were photodissociated at predetermined points between mass gate and reflectron entrance by irradiation with a nanosecond pulsed laser. For the experiments described we made use of either 308 or 440 nm radiation provided by an excimer or excimer pumped dye laser. Laser light was directed collinear but counterpropagating to the mass selected ion packet. Charged fragments generated within the field free flight between irradiation point and reflectron could be subsequently mass analyzed in the reflectron field on the basis of their differing kinetic energies. For this purpose voltages on the two stage reflectron field were configured so as to allow for resolution of all possible fragment ion masses for the parent ions studied. Using the appropriate timing sequence it was possible to irradiate ion packets anywhere between mass gate and reflectron entrance - corresponding to an approximately 30 m sec wide time window for dissociation. Note that species dissociating within the reflectron field give rise to peak broadening as can be shown by ion trajectory calculations.

        In a typical experimental timing sequence (appropriate delays provided by two SRS digital delay generators) clusters were produced following valve opening at tA (see point A in figure 1) and laser irradiation at tB. They were then carried into the RETOF ion source where they were deflected and accelerated at tC, mass selected at tD, photodissociated at tE, reflected following pulsing of the turning field at tF and detected at tG. Typical values for times A-G (as indicated spatially in figure 1) were 0, 600, 1200, 1241, 1246, 1261 and 1327 m sec., respectively. Ion signal with and without dissociation laser was iteratively acquired on consecutive pulses using a PC based multichannel scaler having 5 nsec time resolution so as to maximize signal-to-noise. Typically 5000 pulses (+ 5000 reference pulses) were acquired at a repetition rate of 10 pps in order to obtain data at the signal to noise level shown in figure 2.

3. Results and Discussion

3.1 Dissociation channels and kinetics.

Figure 2. Photofragmentation of Ag9+, Ag11+ and Ag13+ with 308 nm light at high and low fluences.

        Figure 2 shows typical data obtained upon photodissociation of a number of size selected silver cluster cations at 308 nm immediately after mass selection. Two measurements are shown for each cluster size, corresponding to low (one-photon absorption) and high fluence (multiphoton absorption) irradiation conditions. Note the presence in some cases, of small amounts of water complexes with higher mass than the parent. This is a reflection of the high water affinity of silver cluster cations and the mass resolution of our present mass gate design. In the low irradiation fluence regime, the fragmentation channels observed for all sizes studied to date comprise: (i) neutral atom loss, (ii) neutral dimer loss or (iii) both atom and dimer loss - depending on parent cluster size selected. There was no indication of unimolecular dissociation of mass selected parents in the absence of irradiation (on the experimental time scale)24.

        Preliminary analysis of the photofragmentation yield as a function of time delay between mass gate and photodissociation pulse suggests that first step fragmentation following absorption of a 4.0 eV photon goes to completion within the "kinetic time window" of our experiment for the small clusters studied so far. We are in the process of performing studies of unimolecular dissociation rates following absorption of 4.0 eV and lower energy photons to put more accurate upper bounds on the respective dissociation energies 18,25. Atom and dimer loss is a phenomenon which has been observed in all s1-electron metal cluster photofragmentation studies sensitive to fragment mass 4,14,11,18. M and/or M2 loss can be rationalized in terms of the relative stabilities of metallic parents and fragments. Generally even valence electron counts are somewhat more stable on a per atom basis than their odd electron neighbours26. Of particular interest are those cluster sizes and excitation energies for which both atom and dimer loss occurs at comparable apparent rates measurable in our time window. Here one should be able to shift the rate ratios by varying cluster internal excitation. Ag9+ upon 308 nm irradiation fragments into roughly equal proportions of Ag7+ and Ag8+ suggesting comparable dissociation energies for the two processes. The highest level electronic structure calculations, which have so far been carried out for silver clusters over an extended size range predict Ag9+ to Ag8+ (+Ag) and Ag7+ (+Ag2) dissociation energies of 2.8 and 3.1 eV, respectively.19 The unimolecular dissociation of laser excited s1-electron metal clusters has been extensively modeled in terms of RRKM calculations using a simple model to describe the relevant vibrational frequencies (and assuming that electronic degrees of freedom can be ignored)25. It will be of interest to study the source temperature dependence of the respective rates modeled in terms of such a kinetic approach in future work.

3.2 Spectroscopic information

Figure 3.(click to enlarge) Fluence dependence of Ag11+ depletion and fragmentation from 308 nm light.

       Figure 3 shows typical fluence dependencies observed for parent and fragments upon Ag11+ photodissociation at 308 nm (see also figure 2). Note that smaller fragments onset at higher fluences than Ag9+, indicating that multiphoton processes lead to their formation. The corresponding fragment mass distributions are again consistent with (sequential) neutral atom and dimer loss. Over the one-photon fluence range, parent depletions or fragment signals may be used to obtain absorption cross sections. Within the size range studied, Ag12+ has the largest 308 nm depletion/fragmentation cross section. Overlap with the photodissociation data set obtained for silver cluster cations generated by sputtering is quite limited as yet. Where there is overlap, the numbers obtained in our experiment are consistent with those previous measurements given the likely lower internal excitation levels prevailing here 21. In contrast to the large cross sections determined for 308 nm irradiation, preliminary investigations carried out at 440 nm for Agx+, 8<x<12, indicate a general reduction in oscillator strength by at least an order of magnitude.

4. Outlook

        Future work will focus on: (i) fragmentation rate measurements for cluster sizes and excitation energies (including variable internal excitation27) within the kinetic window of our experiment and (ii) determination of near UV photodissociation spectra for a number of small silver cations as a function of internal excitation level.

5. Acknowledgements

        This research was supported by the Deutsche Forschungsgemeinschaft and the Fonds der Chemischen Industrie.


1. V. Bonacic-Koutecky, P. Fantucci and J. Koutecky, Chem. Rev., 91, 1035 (1991).

2. See for example: W. Ekardt, Phys. Rev. B29, 1558 (1984); D. Beck, Phys. Rev. B30, 6935 (1984); C. Yannouleas, R. Broglia, M. Brack and P. Bortignon, Phys. Rev. Lett., 63, 255 (1989); C. Gatti, S. Polezzo and P. Fantucci, Chem. Phys. Lett., 175, 645 (1990).

3. See for example: G. Schwab, Adv. Catal., 27, 1 (1978); J. Frost, Nature, 334, 577 (1988).

4. W. de Heer, K. Selby, V. Kresin, J. Masui, M. Vollmer, A. Chatelain and W. Knight, Phys. Rev. Lett., 59, 1805 (1987); J. Blanc, V. Bonacic-Koutecky, M. Broyer, J. Chevaleyre, P. Dugourd, J. Koutecky, C. Scheuch, J. Wolf and L. Wöste, J. Chem. Phys., 96, 1793 (1992); T. Martin and H. Fallgren, Chem. Phys. Lett., 168, 233 (1990); C. Brechignac, P. Cahuzac, F. Carlier and J. Leygnier, Chem. Phys. Lett., 164, 433 (1989); C. Wang, S. Pollack and M. Kappes, J. Chem. Phys., 93, 3787 (1990).

5. See for example: V. Bonacic-Koutecky, J. Gaus, M. Guest and J. Koutecky, 96, 7939 (1992).

6. See for example: C. Wang, T. Dahlseid, S. Pollack and M. Kappes, J. Chem. Phys., 96, 4918 (1992).

7. T. Baumert, R. Thalweiser, V. Weiß and G. Gerber, Z. Physik D26, 131 (1993).

8. See for example: K. McHugh, J. Eaton, G. Lee, H. Sarkas, L. Kidder, J. Snodgrass, M. Manaa and K. Bowen, J. Chem. Phys. 91, 3792 (1989).

9. U. Röthlisberger, M. Schär and E. Schumacher, Z. Phys. D, 13, 171 (1989); M. Kappes and S. Leutwyler in "Atomic and Molecular Beam Methods, Vol I", G. Scoles, Ed., p.380, Oxford University Press, Oxford (1988).

10. See for example: G. Delacretaz, E. Grant, R. Whetten, L. Wöste and J. Zwanziger, Phys. Rev. Lett., 56, 2598 (1986).

11. T. Reiners, W. Orlik, C. Ellert, M. Schmidt and H. Haberland, Chem. Phys. Lett., 215, 357 (1993).

12. R. Poteau and F. Spiegelmann, J. Chem. Phys., 98, 6540 (1993);U. Röthlisberger and W. Andreoni, J. Chem. Phys., 94, 8129 (1991); U. Röthlisberger, W. Andreoni and P. Gianozzi, J. Chem. Phys., 96, 1248 (1992).

13. M. Homer, J. Persson, E. Honea and R. Whetten, Z. Phys. D, 22, 441 (1991).

14. P. Cheng and M. Duncan, Chem. Phys. Lett., 152, 341 (1988); K. Laihing, P. Cheng and M. Duncan, Z. Phys. D, 13, 161 (1989).

15. G. Alameddin, J. Hunter, D. Cameron and M. Kappes, Chem. Phys. Lett., 192, 122 (1992).

16. M. Knickelbein, Chem. Phys. Lett., 192, 129 (1992).

17. M. Morse, Chem. Phys. Lett., 133, 8 (1987).

18. M. Jarrold and K. Creegan, Int. J. Mass Spec. Ion Proc., 102, 161 (1990); M. Jarrold et al., J. Chem. Soc. Faraday Trans., 86, 2537 (1990); M. Jarrold and K. Creegan, Chem. Phys. Lett., 166, 116 (1990).

19. V. Bonacic-Koutecky, L. Cespiva, P. Fantucci and J. Koutecky, J. Chem. Phys., 98, 7981 (1993).

20. K. Taylor, C. Pettiette-Hall, O. Cheshnovsky and R. Smalley, J. Chem. Phys., 96, 3319 (1992); G. Ganteför, M. Gausa, K. Meiwes-Broer and H. Lutz, J. Chem. Soc. Farad. Trans., 86, 2483 (1990).

21. P. Fayet and L. Wöste, Z. Physik D, 3, 176 (1986); J. Tiggesbäumker, L. Köller, H. Lutz and K. Meiwes-Broer, Chem. Phys. Lett., 190, 42 (1992);J. Tiggesbäumker, L. Koller, K. Meiwes-Broer and A. Liebsch, Phys. Rev. A48, R1749 (1993).

22. S. Federigo, W. Harbich and J. Buttet, Phys. Rev. B47, 10706 (1993).

23. G. Alameddin, Ph.D. thesis, Northwestern University, 1992.

24. See for example: C. Brechignac, P. Cahuzac, J. Leygnier and J. Weiner, J. Chem. Phys., 90, 1492 (1989).

25. U. Ray, M. Jarrold, J. Bower and J. Kraus, J. Chem. Phys., 91, 2912 (1989).

26. M. Brack, Rev. Mod. Phys., 65, 677 (1993); W. de Heer, ibid, 65, 611 (1993).

27. See for example: P. Hackett et al., J. Chem. Phys., 99, 4174 (1993).