Z. Phys. D 40, 78–80 (1997)

ZEITSCHRIFT ¨ PHYSIK D FUR

c Springer-Verlag 1997

Fragmentation of Na(NH3)+n cluster ions C.P. Schulz1 , J. H¨ohndorf1 , P. Brockhaus1 , I.V. Hertel1,2 1 2

Max-Born-Institut, Rudower Chaussee 6, D-12489 Berlin, Germany Fachbereich Physik, Freie Universit¨at Berlin, Arnimallee 14, D-14195 Berlin, Germany

Received: 4 July 1996

Abstract. Photoionization and -fragmentation of Na(NH3 )n clusters by 170 fs and 8 ns laser pulses are studied with photon energies of 2.98 eV to 3.46 eV. In the reflectron timeof-flight mass spectra a strong metastable loss of NH3 is observed independent of the laser pulse length. From the fragmentation rate constants the internal energy of the cluster ions prior to the fragmentation process is determined by an RRK approach. PACS: 36.40.+d; 33.80.Eh; 82.40.Dm

1 Introduction Photoionization of clusters is usually accompanied by subsequent fragmentation of the ionic clusters. On the one hand, this causes a well known identification problem when studying neutral clusters. On the other hand, it allows interesting insights into the dynamics and structure of the ions. It has been argued that fragmentation may be suppressed by femtosecond laser pulses. While this is true for the fragmentation through intermediate excited states in the neutral system there is, however, no a priori reason to assume this for the final ionic channel. The latter fragments on a timescale significantly longer than the laser pulses involved and so called metastable decay occurs even on a µs time scale where fragmentation channels can be distinguished with relative ease. From the fragmentation rate constants information on the stability of cluster ions can be obtained and internal energies can be estimated (e.g. [1, 2]). Here we present results for the metastable fragmentation after one-photon ionization of Na(NH3 )n clusters: a model system for solvated metal atoms in polar liquids. Specifically, we address the question whether the laser pulse length and the photon energy influences the observed fragmentation pattern. 2 Experiment The clusters are produced in a pick-up source [3, 4]. Briefly, a pulsed supersonic beam of neat ammonia (stagnation pressure 2 bar) is expanded into an effusive sodium beam. The

neutral cluster beam enters the detection chamber through a 1.5 mm skimmer. Further downstream the clusters are ionized with pulsed laser light. The clusters may be ionized with laser pulses in two different time domains. A Nd:YAG laser (Spectra Physics GCR-270) provides pulses of 8 ns at 416 nm (2. 98 eV, Raman shifted THG) and 355 nm (3. 46 eV, THG). Alternatively, a regenerative amplifier (Quantronix Model 4810) seeded by a Ti:Sapphire (Spectra Physics Tsunami) is used to produce pulses of 170 fs at 410 nm (3.02 eV, SHG). In the present experiment a reflectron time-of-flight (RETOF) arrangement [5] is used. The photoions are accelerated to 1600 eV by two adjustable electric fields in the direction parallel to the initial neutral cluster beam. After a field free 90 cm long drift region the ions are reflected by a grid-free electrostatic ion reflector which is tilted 2.5◦ against the beam axis. Behind a second field free drift region (50 cm) the ions are detected by a standard arrangement of two multi-channelplates. The RETOF enables us to discern metastable fragmentation processes which occur a few µs after ionisation when the clusterions are in first field free drift region. The fragmented ions will arrive at the detector earlier than the unfragmented parent ions. 3 Results and discussion The photon energies used allow single photon ionization for Na(NH3 )n for n ≥ 4 [4]. Fig. 1 shows two examples of the resulting mass spectra, using the 8 ns and the 170 fs, laser pulses to ionize the clusters, respectively. In both cases two series of mass peaks are observed: unfragmented Na(NH3 )n cluster ions up to n = 40 (dashed envelope) and fragmented cluster ions (dotted envelope) which result from a metastable fragmentation process Na(NH3 )+n → Na(NH3 )+n−1 + NH3 .

(1)

The overall shape of the mass spectrum of the unfragmented cluster ions appears to be roughly independent of the laser pulse length, except that larger cluster ions seem to be somewhat more abundant in the spectrum obtained with the lower photon energy femtosecond laser pulses (Fig. 1b). This points towards a suppression of prompt fragmentation

79

Fig. 2. Fragmentation rate constants for different photoionization energies and pulse lengths

Fig. 1. Mass spectra showing metastable fragmentation (dotted line) of Na(NH3 )+n after photoionization with 355 nm (3.49 eV) 8 ns pulses a and with 410 nm (3.02 eV) 170 fs pulses b

during the ionization, an effect of the energy rather than the pulse length as we shall see. In contrast, the lower abundance of the Na(NH3 )+16 when ionized with 3.02 eV (410 nm) is a known effect caused by the somewhat higher ionization potential of Na(NH3 )+16 , which has been attributed to the closing of the second solvation shell [4] . After ionization the Na(NH3 )+n cluster ions undergo a slow metastable fragmentation process. The acceleration fields and length of the field free region in our reflectron leads to an observation window for metastable fragmentation between 5 to 25 µs for n ∼ 10 up to 10 to 50 µs for n ∼ 40. Metastable fragmentation occurs for clusters with n > 5 independent of the laser pulse length and photon energy used. For larger clusters (n > 20) the mass peaks of the fragments are even stronger than the peaks of the parent ions. To allow a more quantitative comparison for the different ionizing laser pulses we have parametrized our data. For the sake of simplicity we assume a single exponential decay process [6]. The observed relative fragment to parent intensity If /Ip can be converted into a metastable rate constant k by the relation [1]:

If  = exp k tff − 1 exp k tref Ip

(2)

where the first factor describes the decay during the travel time tff in the field free region and the second exponential corrects for further fragmentation of the parent ions while passing the reflectron (tref ). Figure 2 shows the resulting rate constants k which increase relatively smoothly towards larger clusters and appears to decrease somewhat for n > 30. Consistently, for all three photons used n = 16 appears to be slighly less stable than its neighbors: filling the new third shell is more efficient than closing the second one. A slightly less bound ionic cluster thus explains the higher ionization potential of Na(NH3 )16 found earlier. Clearly, the different ionization pulses lead to different rate constants. Inspection of Fig. 2 reveals that the difference can mainly be attributed to a wavelength effect: The two measurements with more than 4 orders of magnitude different laser pulse lengths (8 ns vs.170 fs) but nearly the same photon energy (2.98 eV vs. 3.02 eV) give rates which differ much less than the rates for the two different photon energies at 8 ns (2.98 eV and 3.49 eV). It can be concluded at this point, that the metastable fragmentation rate of Na(NH3 )+n cluster ions does not depend on the pulse length but on the photon energy of the ionizing laser (similar to what has been observed for C60 [2]).The energy dependence also shows that we do not deal with a simple evaporative ensemble [7] which would have lost all memory of its preparation process. Thus, we may proceed one step further and try to estimate the rate constants in a simplified model by using the Rice-Ramsberger-Kassel (RRK) formula for the microcanonical decay rate constants [8]: s−1  ∗ E −D ∗ (3) k(E , D) = ν E∗ If we assume a single metastable decay process we can extract the internal energy E ∗ of the parent ion from a comparison of the rate constants from (3) with the experimental values. As a first approximation for the dissociation energy we take the asymptotic value D = 0.4 eV for the loss of one ammonia molecule from large Na(NH3 )+n cluster ions [4]. The parameter ν denotes a typical vibrational frequency in the fragmentation coordinate. We use the calculated value (200 cm−1 = 6×1012 s−1 ) for (NH3 )2 from Muguet et al. [9]. The number of vibrational degrees of freedom s = 3n − 6

80

magnitude higher (cf. Eq. 3). Interestingly the energy difference Eexc − E ∗ shows no size dependence. Within our line of arguments we attribute this energy difference to kinetic energy of the photoelectron, i.e. photoionization of these systems appears to lead more or less into the ionic ground state. Note, however, that presently we cannot exclude multiple fragmentation - outside our observable time window - during which the system could loose much of its initial excess energy. On the other hand, since the rates depend on the ionizing photon energy, we cannot treat the system as an evaporative ensemble. 4 Conclusions Fig. 3. Comparison of the internal energy E ∗ as calculated from the rate constants by (3) with the excess energy Eexc of the ionization process (4). The open symbols denote the different ionization pulses (see Fig. 2). The solid line presents a linear fit to the calculated internal energy with a slope of 62 meV/molecule. The broken lines indicate Eexc for photon energies 3.0 eV (−·−) and 3.49 eV (···) with a cluster size independent temperature of 200 K

is determined by the number of ammonia molecules n in the cluster since one may safely assume the internal vibrational modes of NH3 to be frozen. Figure 3 shows E ∗ as a function of the cluster size n with nearly identical values for all three ionizing laser pulses. The differences seen in the rates, Fig. 2, are hardly discernible here due to the exponential nature of (3). The so determined internal energy E ∗ of the parent cluster ion rises linearly with cluster size with a slope of ∼ 62 meV per molecule. Thus, it appears that we can explain the metastable fragmentation from an average cluster-ion temperature T ∼ 242 K. A comparison with the neutral cluster temperature is not trivial since the latter cannot be determined easily. In earlier experiments we have estimated it to be ∼ 200 K by using the sodium dimer as a probe for which the vibrational temperature was determined spectroscopically [3]. Thus, if our line of argumentation is correct, only very little additional excess energy is transfered to the nuclear motion of the cluster in the ionization process. The total excess energy Eexc available in the photoionization process given by Eexc (n) = [hν − IP (n)] + (3n − 6)kb T

(4)

is also shown in Fig. 3. Here hν denotes the photon energy (we only show curves for 3 eV and 3.49 eV), IP the ionization potential, and (3n−6)kb T gives the thermal (van der Waals) vibrational energy of the neutral clusters (kb : Boltzmann constant). The ionization potentials are known up to n = 20 [4] and for larger n are assumed to decrease with the cluster radius towards the bulk value 1.5 eV [10]. We see that the available internal energy Eexc is drastically higher than E ∗ determined from the metastable decay. To put it differently: if all the available excitation energy would be left in the cluster ion the fragementation rates would be orders of

Metastable fragmentation is observed in Na(NH3 )+n clusters after photoionization with ns and fs laser pulses. The metastable fragmentation rates do not depend on the laser pulse lengths but increase with photon energy and cluster size. Internal energies of the cluster ions prior to the metastable decay have been extracted from the fragmentation rate constants by the RRK model. To fully understand the dynamics more information is needed on the energy partitioning between internal excitation and kinetic energy of the photoelectron. Clusterion - photoelectron coincidence experiments to determine this as function of cluster size are on the way. The authors thank Dr. F. Noack for his effort in the femtosecond laser experiment. This work has been financially supported by the Deutsche Forschungsgemeinschaft through Sonderforschungsbereich 337.

References 1. Campbell, E.E.B., Ulmer, G., Busmann, H.-G., Hertel, I.V.: Chem. Phys. Lett. 175, 505 (1990) 2. Hohmann, H., Ehlich, R., Furrer, S., Kittelmann, O., Ringling, J., Campbell, E.E.B.: Z. Phys. D 33, 143 (1995) 3. Schulz, C.P., Haugst¨atter, R., Tittes, H.U., Hertel, I.V.: Z. Phys. D 10, 279 (1988) 4. Nitsch, C., Schulz, C.P., Gerber, A., Zimmermann-Edling, W., Hertel, I.V.: Z. Phys. D 22, 651 (1992) 5. Mamyrin, B.A., Karataev, V.I., Shmikk, D.V., Zagulin, V.A.: Sov.Phys.-JETP 37, 45 (1973) 6. In our experimental time window only single ammonia losses were observed due to the large binding energy of 0.4 eV 7. Klots, C. E.: J. Phys. Chem.92, 5864 (1988) 8. Frost, W.: Theory of unimolecular reactions, p. 311. New York: Academic Press 1973 9. Muguet, F.F., Robinson, G.W., Bassez-Muguet, M.P.: J. Chem. Phys. 102, 2655 (1995) 10. H¨asing, J.: Ann. Physik 5. Folge 37, 509 (1940)

This article was processed by the author using the LaTEX style file pljour2 from Springer-Verlag.