Return to Home Page

Also see earlier Communication:

Photofragmentation of Mass-Resolved Carbon Cluster Ions: Observation of a "Magic" Neutral Fragment __ M.E. Geusic , T. J. McIlrath, M. F. Jarrold, L. A. Bloomfield , R.R. Freeman and W. L. Brown__ J. Chem. Phys.84(4), 2421 (15 Feb 1986)

 

Photodissociation of Carbon Cluster Cations

M. E. Geusic, M. F. Jarrold, T. J. Mcllrath, R. R. Freeman and W. L. Brown
*RV in Department 11133; University of Maryland

 

INTRODUCTION

Carbon clusters are objects of increasing theoretical and experimental interest. This interest stems from the fundamental information they provide on the evolution of physical and chemical properties of carbon as it evolves from molecular to bulk material. In addition, carbon clusters play a significant role in processes ranging from catalysis and combustion to interstellar opacity. Although C2 and C3 are now fairly well characterized, little is known about the larger clusters. Kaldor and coworkers1 recently reported some results for the larger clusters, although their work was mainly concerned with the distribution of carbon clusters produced by laser evaporation. Bloomfield et. al.2 demonstrated the production of Cn- and Cn+while Geusic et. al.3 reported preliminary results on the photofragmentation patterns of Cn+. More recently Smalley and coworkers4and Cox and coworkers5 have focused on the stability of C60 clusters.

In this paper we report the details of a study of the photodissociation of mass selected carbon cluster ions. Data on dissociation thresholds, branching ratios and photofragmentation cross sections for carbon cluster cations (with between 3 and 20 atoms) have been obtained using both 351 nm and 248 nm light. In the next section we discuss the experimental techniques for isolation of specific carbon cluster ions and their subsequent photofragmentation. The results of photofragmentation measurements for carbon cluster ions are presented in the third section along with a discussion of the strengths and weaknesses of the photodissociation technique, particularly with regard to bracketing dissociation energies and errors introduced by loss of product ions from the beam due to the energy of fragmentation. The interpretation of the results, comparison with model calculations and sources of error are discussed in the last section.

EXPERIMENTAL METHODS

A schematic diagram of the experimental apparatus used in these cluster ion photofragmentation studies is shown in Fig. 1. The apparatus is designed to run on a pulsed basis, each pulse producing a new burst of clusters by laser vaporization.6 A cycle begins with the production of ions within the throat of a supersonic nozzle. A pulsed valve is used to produce a 100-400 microsecond (FWHM) pulse of helium. The helium pulse flows down a 0.1 cm diameter tube and passes over a rotating sample target. A Q-switched Nd:YAG laser (7 ns pulse length), operating at 532 nm and focused to a 0.1 -0.05 cm diameter spot on the target rod ( approximately 3 mJ per pulse ), is triggered coincidently with the arrival of the helium pulse. The laser vaporizes the sample, producing a plasma which is entrained within the helium carrier gas. An exit channel, 0.5 cm in diameter and of variable length, allows a substantial number of collisions to occur prior to expansion. The clusters ( anions, cations and neutrals) expand along with the helium carrier gas into the main chamber. This expansion produces significant translational cooling, and presumably some internal cooling of the clusters as well.7

The cluster beam proceeds through a skimmer into a differentially pumped inner chamber which houses the three plate acceleration region.8 As the ions enter this pulsed acceleration region, a power supply applies 2.25 kV and 1.85 kV for 3 usec to the back and middle plates respectively. The cations are turned 900, accelerated, and time focussed as they travel through vertical and horizontal deflection plates, an Einzel lens, and enter the mass separator region 68 cm from the acceleration plates. The mass selector consists of two parallel plates, one at ground and one at 100 volts. Just before the cluster ion packet of the mass of interest enters this region, the 100 volt side is switched to ground; it is returned to 100 volts just after the cluster packet leaves. Thus a mass selected packet is allowed to pass into the fragmentation region while all clusters with other masses are deflected off the beam axis and stopped.

The selected packet (containing approximately 103 ions) next approaches the photofragmentation region where it is decelerated to around 1000 eV and then enters the field free drift region. Here it is irradiated by a pulsed excimer laser at right angles to the cluster ion beam direction. The excimer laser pulse length is 10 ns. During this short time period the ions do not move significantly.Some fraction of the clusters in the laser beam dissociate and the remaining parent cluster ions and their charged fragments drift into the reacceleration region where they gain an additional 5 keV of energy. After traversing 150 cm through the second time-of-flight region, the mass-separated ions are detected by 3 cm diameter dual micro-channel plates. The signal from the channel plates is amplified and recorded on a LeCroy 200 megasample/second transient recorder. Once digitized, the signal is stored and processed by a computer. The overall experimental sequence is run at a repetition rate of 10 Hz.

The signals from the microchannel plates are normalized for each cycle to a constant total charge to compensate for fluctuations in the number of parent clusters produced by the source. The fluctuations are typically around + 25%. Data recorded with fluctuations of greater than + 75% of the average at the first ten sequencies recorded are discarded. Data are accumulated in the computer with the photodissociating laser alternately on and off. The normalized laser-off signal is su btracted from the normalized laser-on signal. The final data are derived from many (up to 5000) such pairs. For measurements as a function of fragmentation laser intensity the laser intensity is varied manually by inserting a large number of quartz plates into the beam. The intensity of the fragmentation laser is monitored using a UV diode just before the laser beam enters the vacuum system. The signal from the UV diode is fed to a box car averager and then read by the computer through a camac A to D converter. The result of the laser-on/laser-off cycle is then accumulated in the bin corresponding to the photodissociating laser power recorded by the UV diode. Data are recorded using 100 of these bins to cover the laser power range studied. The laser power range is calibrated using a power meter.

RESULTS

1. Initial Ion Distribution

Figure 2 shows the distribution of cluster ions taken directly from the source with the pulsed mass isolator turned off. The distribution reflects both the stability of various clusters and the details of the cluster formation process. Prominent peaks in the mass spectrum ("magic numbers") can be seen with a periodicity of four at n=7,11,15,19 and 23 but, because of the complexity of the creation process, the relationship of the observed distribution to the stability of various clusters is difficult to understand. This problem is most severe when attempts are made to study the neutral cluster distribution by laser photoionization. In this case the resultant charged cluster distribution is very different from that shown in Fig. 2 and most likely reflects competition between photoionization and photofragmentation of the neutrals as well as the processes involved in formation of the original clusters or photoionization cross sections. This issue will be considered later in discussing ionization and fragmentation rates. In any event, the measurements here largely sidestep this issue because a given cluster size is first selected, and experiments are performed on it. The results presented here are independent of the initial ion distribution.

2. Photodissociation Branching Ratios

The ability to select cluster ions of a specific size and subsequently photodissociate and mass analyze the charged fragments provides the opportunity to determine accurately the branching ratios for photofragmentation into various ion fragments. An example of a time-of-flight mass spectrum of the photofragments from the photodissociation of C18+ is shown in Fig. 3. The fragmentation was achieved with 248 nm light and a photodissociating laser fluence of approximately 12 mJ cm-2. The main peaks in the product ion mass spectrum are due to C10+,C11+, C12+, and C15+. The C18+ peak is recorded here as a negative signal because of the background subtraction technique used. This negative C18+ signal gives a direct measure of the total depletion of the C18+.

The interpretation of spectra such as those shown in Fig. 3 must be done cautiously due to the high incident fragmentation laser powers necessary to obtain useable signals. The products of photofragmentation include direct products following single photon absorption, direct products following multiphoton absorption and sequential, multi-step fragmentation. Possible examples of the two types of multiphoton absorption can be given for fragmentation of C18+ into C12+where the direct process is represented by

C18+---hv--> C18+*---hv--> C18+** -----> C12+ + C3+ C3

and the sequential process is represented by

C18+---hv--> C18+* -----> C15+ + C3 ---hv--> C15+*-----> C12+ + C3

In both of the above cases the production of C12+ will depend quadratically on the incident fluence, if the fluence is sufficiently low. Figure 4 shows the dependence of the production of C15+ and C12+ on the 248 nm fluence. The production is seen to have a linear dependence consistent with direct production from the parent C15+ ion. However, the production of C12+has a faster than linear dependence on intensity, presumably reflecting one of the two processes discussed above. It is not possible, with the data available from our experiment, to determine which of the two possible non-linear processes discussed above is operating for C18+ .

The identification of a faster than linear intensity dependence unambiguously identifies the presence of non-linear absorption processes. However, and this is the crucial point, the presence of a linear dependence at a given fluence is not sufficient to identify a linear process. For example, in the case of C12+ fragmentation from C18+, the production of C12+appears to be linear above a fluence of approximately 6 mJ cm-2 . This linear dependence may be intrinsic, or it may result from saturation of one of the two steps in a multistep absorption process. If a non-linear process involves two steps with greatly differing cross-sections, then at sufficiently high fluences the step with the larger cross section will saturate and the process will appear completely linear , proceeding at a rate limited by the step with the smaller cross section. In this case, without knowledge of the cross sections, it is never certain how low a fluence is required to avoid saturation. On the other hand, if the fragmentation depends on a resonant multiphoton transition, as opposed to a multiple step transition, then the quadratic rise in transition rate with incident fluence can be at least partially offset by depletion of the number of parent ions, giving rise to a curve that appear linear over a limited range. In this case, restriction of observations to a limited range of fluences can produce an apparently linear curve even when the underlying processes are non-linear.

Because of the small signals to background associated with minor fragment clusters, it is particularly difficult to establish whether the fluence dependence is linear for these products. We have restricted our branching ratio studies to the lowest feasible fluence and generally limited our assignment of direct product species to those representing more than 10% of the total fragmentation.

The effect of sequential fragmentation is clearly seen in Fig. 4 where fragmentation of the C15+ products by the incident radiation depresses the fluence dependence of C15+ at high fluences. In some cases the amount of depletion of the first generation daughter matches the depletion expected from the directly measured photodissociation cross section of this cluster ion as a parent. This supports the two step sequential absorption process. However, in other cases ( e.g., C15+ in Fig. 4) cross sections several times larger than those measured for smaller parent clusters are required to account for the depletion of the product species. These larger cross sections could arise because the intermediate product ions are produced in excited states.

 In general, photoexcitation in our experiment does not exactly match the threshold for dissociation, thus, there will be excess energy that is carried off by the fragments. This excess energy will appear as either internal energy of the product fragments or kinetic energy distributed among the fragments. The effect of product kinetic energy could be a significant loss of ions from the detected beam. Consider a parent ion, initially moving along the time-of-flight (TOF) axis towards the detector, which dissociates to generate a product ion with a velocity component normal to the TOF axis. If W is the half angle of the detector subtended at the fragmentation region, ET the translational energy of the product fragment associated with the velocity normal to the TOF axis and EL the energy of the product fragment associated with motion along the TOF axis, then the requirement for the fragment to reach the detector is ET<ELW2. The value of EL is given by EL= (m/n) E0 where m and n are the masses of the product and parent cations and E0 is the kinetic energy of the parent cation. If a parent ion of mass n fragments to produce a product ion of mass m, then the energy associated with the product fragment motion transverse to the TOF axis is ET= (n-m/n) Ecsin2Q where Ec is the center-of-mass (COM) fragmentation energy and Q is the angle of fragmentation relative to the TOF axis as measured in the COM frame. The requirement for detection of the Cm+ fragment is Ecsin2Q < (m/n-m) E0W2 and no products at all are lost if EcQ < (m/n-m) E0W2. For large parent cations the fragmentation usually involves a heavy charged product and a light neutral product so that (m/n-m) >>1 and loss of charged fragments from the beam is insignificant. For n = 18, m = 15, E0=5x103 eV and W = 10-2, there is no loss unless Ec > 2.5 eV. On the other hand, for smaller parent cations loss can be a problem. For n = 6 and m = 3, , loss begins for Ec = 0.5 eV.

In our experiment the fragment loss problem was minimized by fragmenting at the focus of an Einzel lens system. Therefore, even products with significant transverse velocity are collimated before entering the second time-of-flight tube. We have looked for an effect of loss from the beam by varying the energy in the final time-of-flight tube, EL. A variation from 6 keV to 2.5 keV showed no effect on the total number of charges reaching the detector (to within ~ 10%) even for fragmentation of C5+ to C3+. Thus, for our apparatus, charge conservation in the beam is a valid assumption, due to effective ion optics and the fact that some fraction of the excess energy probably remains as internal energy in the fragment ions, rather than appearing entirely as product ion kinetic energy.

The above considerations have led us to conclude that our apparatus  records most of the product ions. Figure 5 shows a histogram of the photofragments with 248nm light recorded with a laser fluence of 2 mJ cm-2. At this laser fluence most of the products with intensities greater than 10% in Figure 5 show a linear laser fluence dependence at higher laser fluences. It is clear from the results shown in Fig. 5 that the loss of a C3 unit dominates the fragmentation of nearly all carbon cluster cations, while for some of the larger cluster ions, the loss of C5 is also an important fragmentation pathway.

The photofragmentation branching ratios measured with 248 nm and 351 nm light are tabulated in Fig. 6 where all the product ions detected at a laser fluence of ~2 mJ cm-2 are recorded. The photofragment branching rations at 351 nm are, for the most part, similar to those obtained at 248 nm, differences of less than 5% of the total are not significant. Qualitatively, similar results were obtained using a doubled Nd:YAG laser (532 nm), but these results will not be discussed in detail here because poor beam quality from the Nd:YAG laser made fluence dependence studies unreliable.

3. Photofragmentation Thresholds

The laser fluence dependence of photodissociation can be used to bracket dissociation energies of the cluster cations by establishing whether the photodissociation process is one photon (i.e. depletion of the parent ion is linear  with laser fluence) or multiphoton (i.e., depletion of the parent ion is quadratic or greater with laser fluence.) For these studies it is necessary that the photodissociating laser beam be spatially homogeneous in order to reduce saturation effects. For this purpose we used a portion of an excimer laser beam having an uniform spatial intensity pattern. In the interpretation of the results of these studies we assume that the cluster ions represent a single isomer and that their internal energy is negligible compared with the photon energy. Figure 7 shows the depletion curves for C5+ and C6+ as a function of 248nm light intensity. Our interpretation of these curves is that photofragmentation of C6+ is single photon while C5+ is multiphoton. Thus, in this analysis 4.98 e V is above the dissociation energy of C6+ and below that of C5+. This method has the same problem as that encountered in studying branching ratios. If the process were a sequential two-photon process, with a large difference between the cross sections for the two steps, the depletion would be limited by the smaller cross section. Useable amounts of fragmentation would not occur until the easier step became saturated, producing experimental curves which appear linear (i.e., one photon). For the data reported here this effect was reduced by working with the minimum laser fluences allowed by signal-to-noise considerations.

 The photofragmentation results for carbon cluster cations with 248 nm and 351 nm light are summarized in Table 1. Photodissociation of cluster cations with n=4 and n=6-20 at 248nm (4.98 eV ), appears to be a one photon process, for n=5 it appears to be a multiphoton process. Photodissociation at 351 nm {3.53 eV) of clusters with n=6-20 again appears one photon. We were unable to obtain sufficient depletion of C3+ with either 248 nm or 351 nm light to determine the laser fluence dependence. The dissociation energy of C3+, as we discuss below, is known to be greater than 5 eV so photodissociation with either 248 nm or 351 nm light must be a multiphoton process. Similarly, we were unable to obtain sufficient depletion to determine reliably the laser fluence dependence for C4+ with 351nm light; again this most likely reflects multiphoton absorption. The dissociation energies deduced from these measurements are summarized in Table II.

4. Photodissociation Cross Sections

Photodissociation cross sections were derived from measurements of the slope of the depletion curves of the parent cluster cation as a function of laser fluence. Measurements were made under conditions of low depletion (generally less than 10%) to reduce any effects of multiphoton processes. Cross sections were determined from the laser power, laser beam spot size and the depletion. At 248 nm for the larger clusters with large cross sections we were able to deplete >80% of the parent ions at high laser fluences which indicates that there is good overlap between the laser beam and the cluster packet, at least for the larger clusters. The overlap for the smaller clusters may not be as good (we were able not to deplete such large fractions) which would introduce a systematic error into our measurements. In any case we expect the relative values to be accurate for nearby masses.

The depletion was measured as the ratio of the residual parent ions to the total charges (parent plus product), this assumes that the total charge in the beam is always conserved during a photofragmentation run. As discussed above, large fragment kinetic energies can in principal, result in a loss of charge from the beam. In that case the effect of normalization would be to reduce the apparent depletion and the resulting derived cross section. The assumption of charge conservation appears to be valid in our experiment because the depletion results were found to be independent of changes of over a factor of 2 in the kinetic energy of the cluster ions in the final time-of-flight region as discussed in section 2 above.

The relative photofragmentation cross sections are shown in Fig. 8. The cross sections derived from our data are for one photon processes, for clusters for which photodissociation is multiphoton, the one photon cross sections shown in Fig. 8 are reported as zero. The cross sections measured with 248 nm light are shown in Fig. 8a and those measured with 351 nm light are shown in Fig.8b. The absolute cross sections can be obtained by multiplying the relative values by 10-17cm2. The relative error for adjacent masses is estimated to be + 30% and the uncertainty in the absolute cross section is approximately a factor of 2.

5. Competition Between Ionization and Dissociation

The photodissociation cross sections measured for the carbon cluster ions with 248 nm light have absolute values in the 10-18 cm2 to 10-17 cm 2 range. Since photoionization cross sections for most species are also in this range and since multiphoton rates are much lower than single photon rates at our intensities, if carbon cluster ions were generated by photoionization of neutral species, substantial photodissociation would also be expected to occur. Figure(9a) shows the distribution of positive cluster ions produced directly at the source, while (9b) shows positive ions produced by 193 nm photoionization of neutral clusters from the source. The difference in the positive cluster ion distribution in (9a) and (9b) probably reflects the result of photodissociating simultaneously with the ionization. For carbon clusters with 30 or fewer atoms, the two spectra show similar features: there are "magic numbers at 11,15,19 and 23, although the relative prominence of these peaks is greatly enhanced in (9b ). In this region, both spectra have qualitative features that are similar to those obtained by Furstenau and Hillenkamp9 who used laser vaporization of thin carbon foils in a vacuum. Spectrum (9b) is nearly identical to that reported by Rohlfing et al.1 for 193 nm ionization of neutral carbon clusters.The important differences between (9a) and (9b) are for clusters greater than 30 atoms. In the direct Cn+ production (9a) there is a progressive fall off of intensity with increasing cluster size, with only a weak even-odd alternation in abundance. However, when neutral clusters are ionized (9b ), only even clusters appear in the spectrum. Figure 9c shows the spectrum obtained when C60+ is isolated by the pulsed mass isolator and fragmented by a large fluence of 532 nm radiation. This spectrum shows the unfragmented portion of the original  C60+ beam and the daughter fragments. The important observation is that this spectrum contains several features that are qualitatively similar to trace (9b ), for cluster fragments with more than 30 atoms only even fragments appear, while for cluster fragments with fewer than 30 atoms both even and odd fragments are present, with "magic numbers" again at 7,11,15 and 19. While trace (9c) contains a larger fraction of cluster ions with 3- 7 atoms than does trace (9b ), it is apparent that much of the contrast between prominent and weak peaks observed in trace (9b), including the absence of odd peaks above clusters with 30 atoms, may well be a consequence of fragmentation. Rohlfing et al.1 discuss the possible role of fragmentation due to the ionization laser in their Cn+ spectra and speculate that fragmentation of larger cluster ions contributes to the spectrum of lighter clusters. Our result shows explicitly that fragmentation of large carbon ions can indeed produce a Cn+ spectrum with features similar to one obtained by ionizing a distribution of neutral carbon clusters.

DISCUSSION

Perhaps the most striking result from these studies is the dominance of the neutral C3 loss channel in the photofragmentation of carbon cluster ions. These results differ from those for silicon10 and germanium11 studied previously in this laboratory. The photofragment branching ratios for silicon and germanium cluster ions are quite similar to each other and are characterized by preferential loss of a single atom for initial cluster ions with less than seven atoms and formation of a Si6+ or Ge6+species for initial cluster ions with more than six atoms.

The special nature of the C3 loss channel clearly has its origin in the dissociation mechanism. One possibility is that we are observing direct photodissociation where the building blocks of the cluster, C3 are removed.The persistence of this loss channel for cluster sizes from n = 4 to 20, a range over which theoretical models 12-14 predict significant changes in the morphology of the ion, make this explanation unlikely. The dominance of the C3 loss channel more likely reflects the fact that it is the lowest energy pathway, the C3 species being known to be very stable. In this case the photodissociation of the carbon cluster ions would not be direct but would involve a transition to a bound excited electronic state followed by internal conversion to the ground electronic state in a highly excited vibrational state. Dissociation then occurs statistically (i.e., by a unimolecular reaction) on the ground state potential surface to give preferentially the lowest energy products. It is interesting to note that the photofragmentation branching ratios vary little with the wavelength of the photodissociating light. Although unimolecular branching ratios are not generally independent of energy, the branching ratios would be even more likely to show a strong energy dependence if photodissociation occurred by a direct mechanism.

Recent work by O'Keefe and coworkers16 supports the conclusions discussed above. They investigated the low energy collision induced dissociation of carbon cluster ions, C5+-C15+, and observed fragmentation patterns quite similar to those observed from photodissociation. For low energy collision induced dissocation the excitation is deposited primarily as vibrational energy. So the observation that collision induced dissocation and photodissociation result in similar product branching ratios supports the conclusion that photodissocation occurs by excitation to a bound state followed by internal conversion and dissocation on the ground state potential surface.

An anomaly in the photofragmentation branching ratios occurs for n=5 where the only products observed are C3++ C2. The charge is on the C3 product. This observation suggests that the ionization energy of C2 is larger than that of C3, making it energetically favorable to place the charge on the C3 fragment. By the same argument, ionization energy of the fragments larger than C3 must be smaller than that of C3 for the neutral C3 loss channel to be so dominant in fragmentation of the larger clusters.

Our data on dissociation energies, shown in Table 1, suggests that C6+ -C20+ have binding energies less than 3.53 eV,C5+ is bound by greater than 4.98 eV and C4+ is bound by less than 4.98 eV. The binding energies thus oscillate quite substantially for small clusters, but for n>6 they drop to less than 3.53 eV. We noted earlier that apparently linear multiphoton processes could lead to an underestimation of the real dissociation energy. A second source of possible error that would also lead to an underestimate of the binding energies is internal energy in the initial clusters. We feel that the use of free expansion in the He carrier gas in the cluster source reduces the kinetic temperature of the clusters and should reduce internal temperatures9 below room temperature, but there are at present no experimental checks on this assumption. Another source of potential error, in this case leading to an overestimate of cluster binding energies, is the possibility of activation barriers along the reaction coordinate for dissociation. For the smaller carbon cluster ions studied here, the fact that we observe no loss of ions from the beam indicates that the center-of-mass energy is low which suggests that there are no substantial activation barriers.

There exists little other experimental data on the binding energies of carbon cluster ions to compare with our data. For a rough comparison we have combined the measured heat of formation of some of the smaller neutral clusters with approximate ionization energies and derived some values for the dissociation energies of th.e cluster ions with n=3 to 7. These derived values are not very reliable, mainly due to uncertainties in the ionization energies. It would be optimistic to consider them accurate to within leV. They are shown for comparison with our measurements in Table 2 and there is a correlation between the two data sets. However, some recent detailed ab initio calculations by Raghavachari and Binkley10 are in partial conflict with our results. The calculations do predict that C3 loss is the lowest energy dissocation pathway but the calculated dissociation energies are generally smaller than our derived limits. However these calculations do not accurately reproduce the dissociation18 energy of C2+ (calculated: 4.7 eV; measured: 5.32[18]) which is well known and we expect the errors for the larger clusters to be larger.

The most striking features of the photodissociation cross sections (Fig. 9) are the dramatic variation of the photodissociation cross sections near n = 10 and the large increase in cross sections above n = 16 for 248nm radiation. In an earlier communication3 we noted that the drop in cross section at n=10 might reflect a shift in the absorption due to a change in structure from linear chains to monocyclic rings. There have been several calculations of the structure of carbon clusters and some predict this change in structure to occur at around n=10, however the reported calculations11 are generally low level. Some recent higher level calcuations10,12 suggest that the transition to a cyclic structure occurs for cluster sizes less than n=10, at least for the even numbered clusters. A definitive answer to the question of the origin of the cross section variations requires measurements of the photodissociation spectra to fully map out the shifts in the absorption which must be occurring.

CONCLUSIONS

In this paper we have reported the results of a detailed study of the photodissociation of carbon cluster ions containing between 3 and 20 atoms. The most striking result from these studies is the dominance of the C3 loss channel in the photofragmentation. This fragmentation pathway probably reflects the stability of the C3 species. The photofragmentation cross sections were observed to change dramatically with cluster size. Limits on the dissociation energies of the carbon cluster ions derived from the laser fluence dependences of photodissociation have been given. There are clear possibilities for misleading results which cannot be ruled out experimentally, but this method of determining dissociation energies is currently the only available experimental method for large clusters.