Z. Phys. D - Atoms, Molecules and Clusters 20, 193-196 (1991)
Atoms,Molecules and Clusters f~ Phy~kD
© Springer-Verlag 1991
Fragmentation dynamics of ammonia cluster ions after single photon ionisation E. Kaiser, J. de Vries, H. Steger, C. Menzel, W. Kamke, and I.V. Hertel Fakult/it ffir Physik und Freiburger Materialforschungs-Zentrum,Albert-Ludwigs-Universit/it,Hermann-Herder-Strasse 3, W-7800 Freiburg, Federal Republic of Germany Received 10 September 1990; accepted in final form 15 October t990
Abstract. A reflecting time of flight mass spectrometer (RETOF) is used to study unimolecular and collision induced fragmentation of ammonia cluster ions. Synchrotron radiation from the BESSY electron storage ring is used in a range of photon energies from 9.08 up to 17.7 eV for single photon ionisation of neutral clusters in a supersonic beam. The threshold photoelectron photoion coincidence technique (TPEPICO) is used to define the energy initially deposited into the cluster ions. Metastable unimolecular decay (gs range) is studied using the RETOF's capacity for energy analysis. Under collision free conditions the by far most prominent metastable process is the evaporation of one neutral NH3 monomer from protonated clusters ( N H 3 ) , - z N H ~ . Abundance of homogeneous vs. protonated cluster ions and of metastable fragments are reported as a function of photon energy and cluster size up to n = 10. PACS: 36.40. + d; 35.20.Wg; 33.80.Eh
1. Introduction Ammonia clusters are of central interest as prototype systems for studying the transition from isolated ions to solvation in the liquid phase. In order to understand the structure and energetics as well as the fragmentation dynamics following ionisation much theoretical (e.g. [I-3]) and experimental work has been devoted to this system such as thermochemical studies using high pressure mass spectrometry (e.g. [4, 5]), supersonic beam experiments with electron impact ionisation (e.g. [6, 7]), multi-photon ionisation [8-10] and single photon ionisation by synchrotron radiation [11, 12]. Characteristic in mass spectra of ammonia cluster ions is the high abundance of the protonated (NH3),_ 2NH~as compared to the homogeneous species (NH3) + . This is attributed to fast intracluster proton transfer with subsequent fragmentation:
(NH3) + -~ (NH3),-2NH2NH~ ._+
(NH3),-2NH,+ + NH2
(1)
Usually in the ionisation process the ion clusters are left with ample excess energy so that they will suffer further fragmentation, evaporating one or successively several neutral NH3 monomers 1
x
(NH3),-2NH,~ -*. • • ~-~( N H a ) , - z - ~ N H ~ + xNH 3
(2)
as has been demonstrated particular]ly clearly very recently [7]. Under certain conditions this chain of successive fragmentations may constitute an 'evaporative ensemble' [13] exhibiting a broad range of fragmentation rate constants. By mass spectroscopic techniques with energy analysis of the fragments (e.g. by using a RETOF) one may study part of the kinetics of these processes on the kts time scale, so called metastable processes, and obtain information on the dynamics in this time domain as well as on cluster energetics as recently demonstrated for ammonia clusters using electron impact [5] and multiphoton-ionisation [10]. There the clusters are created with an initially high amount of internal energy. In contrast, single photon ionisation using the TPEPICO technique provides a limitation of this excess energy by selecting only such processes for which electrons of essentially zero kinetic energy are detected. Consequently the internal energy is given by the photon energy minus the true adiabatic ionisation potential. Thus, TPEPICO offers a promising alternative for defining the initial conditions of the fragmenting ensemble of cluster ions which allows us to change this excess energy in a controlled way by changing the photon energy and thus gives hope for a more detailed understanding of the fragmentation dynamics. The present paper reports first experimental results from such a study, in the energy range from 9.08 eV to 17.7 eV, in which we mass resolve both the initial and final products of the fast decay process according to (I). We also study the unimolecular fragmentation according to (2) occurring in a metastable time window and give
194 a first discussion of the most prominent features of the results.
2. Experimental The details of the experimental setup have been described elsewhere [14] (see also the accompanying paper [15]) except for the R E T O F part which has been added to mass resolve the homogeneous clusters and to study metastable processes. Briefly, a cluster beam is produced by expanding pure ammonia from a stagnation pressure of typically 2 bar at a temperature of 260 K through a 70 gm nozzle into high vacuum. The beam enters the ionisation chamber through a conical skimmer and is intersected at right angles by focussed synchrotron radiation in the center of the extraction field. Wavelength selection is done by a near normal incidence monochromator. Second order transmission at wavelengths above 105 nm is blocked by a LiF-filter. Threshold photoelectrons at essentially zero kinetic energy are selected by a steradiancy analyzer. In order to provide sufficient energy resolution for the electrons and high detection efficiency for the corresponding cluster ions the two-stage ion source of Witey-McLaren type [16] is driven by pulsed voltages [15]. The new R E T O F setup is completed by a two-stage ion mirror comprising a first decelerating field (20 mm, ,-~330 V/cm) and a subsequent reflecting field (100 mm, ~ 130 V/cm). Finally the reflected ions are detected by a stack of two multi-channel plates assembled in a 'chevron' configuration. The total length of the field free regions is approximately 760 ram. This 'reflectron' time of flight mass spectrometer [17] provides improved mass resolution with respect to the Wiley-McLaren setup by 2nd order spatial focussing Our new instrument exhibits a mass resolution m/Am "~ 4 0 0 . . . 600. Also, the R E T O F has the capability for separating prompt ions and product ions resulting from metastable fragmentations. The parent ion's toss of mass due to the fragmentation is accompanied by a proportional loss of kinetic energy. If the process occurs in the field free region in front of the reftectron the fragments appear at a shorter flight time with respect to corresponding stable ions and show up as isolated distinct peaks in the mass spectra. Fragmentations that occur in the ion source produce a broad background signal in an interval that ranges from the flight time of the prompt ion, with a mass equal to the metastable fragment, to the flight time of the metastable fragment, whereas fragmentations in the reflector contribute to the signal at flight times that range from the metastable peak to the corresponding stable peak. The high mass resolution of our R E T O F allows us to operate it in a mode reflecting both parents and fragments simultaneously. This optimises energy resolution as well as collection time and is thus particularly well suited for the intrinsically low signal T P E P I C O setup. In order to relate the metastable peaks to a distinct fragmentation process we model the R E T O F mass spectrum by use of a computer simulation. The most relevant parameters such as flight times, setup voltages, source ion distribution and an appropriate decay law (for simplicity taken to be a single exponential) are taken into account.
3. Results and discussion For seven discrete photon energies ranging from 9.08 eV (136.5 nm) below the appearance potential of the (NH3)~ up to 17.7 eV (70.0 nm) R E T O F spectra in a time interval (8-28.48 ~ts) comprising the flight times of the ammonia monomer up to the (NH3)~-o were recorded (for measurements of the appearance potentials, photon efficiency curves and T P E P I C O efficiencies see [11, 12]). Figure la shows a typical mass spectrum obtained at 9.39eV photon energy. It displays the characteristic features of ammonia cluster ion mass spectra: weakly abundant homogeneous species (NH3),+ (labeled n), dominating species ( N H 3 ) , - 2 N H 2 (labeled n') resulting from the fast
140
S' 4'
120
3' 6'
"E 100
so dO
7~
= so C
4o 20
= 31l " sll
~7'
*8'
t9' =10'1 11'
•
a
o
f ' l l l l 10 12
J .Ik~ 16 18 20 flight time/us
14
. 0~r_~L
22
tJ
24
11
26
i
28
10 51
experimental
5
S C *5'
4t
*6'
C "¢S
~= =:
o 4 2
~
j
"41
*s'
i
0 16
4 ,~ /5,*s~~,'sa simulated
i 17
i
i
18 flight time/us
19
20
Fig. 1. a TPEPICO-TOF-spectrum of ammonia cluster ions at a photon energy of 9.39 eV 0- = 132 nm). The labels n, n' and *n' are explained in the text. NH f (n = 2) is missing due to its appearance potential of 9.54 eV. b Determination of the metastabte fragmentation channel. The upper measured spectrum is a detail of Fig. la. Peaks are labeled accordingly. The lower spectrum is a computer simulation of the spectrum displayed in Fig. la. The simulated metastable peaks *ha, *nb, *n' correspond to the unimolecular reactions displayed in (3), (4) and (5) respectively and art labeled according to equations (3a)-(5a)
195
ion-molecule reaction with subsequent fragmentation according to (1) and/or (2) and small metastable peaks (labeled *n') which appear at flight times in between two subsequent protonated cluster ions. Note that throughout this work the label n refers to a hypothetical (NH3), neutral cluster from which it may have arisen (disregarding possible alternative preceding fragmentation steps in the ion). The following processes have to be considered and are included in the simulation of the mass spectra: 1. A homogeneous parent ion ejects an NH2 radical: (NH3) + ~ (NH3),_ z + NH~ + NH 2 n ~ *na. + NH2
(3) (3a)
2. A homogeneous parent ion evaporates an NH 3 neutral: (NH3) + --->(NH3),,+_, + NH3 n ~ *nb+
NH3
(4) (4a)
3. A protonated parent ion evaporates an NH3-unit: (NH3),_2NH,~ --* (NH3),-3NH + + NH3
(5)
n' --+ *n' + NH3
(5a)
The upper part of Figl lb displays a detail of Fig. la whereas the lower part shows the corresponding computer simulation. It is evident that the measured rectastable peaks result from the process (5) with perhaps a very small indication of process (3). For cluster sizes n > 6 product ions resulting from processes according to (3) and (5) cannot be separated any longer. However, metastable evaporation of a NH3-unit from a homogeneous ion (4) can be clearly excluded for all observed metastable peaks. Comparing in Fig. l b the experimentally determined and the simulated peak shapes (using a single exponential decay) of the fast fragments (5' and 6') clearly illustrates that the assumption of a monoexponential decay is not justified: The p r o m p t signal is created on a much shorter time scale than the extrapolation of the metastable signal towards time zero would predict. A single exponential would be expected if the initial internal energy distribution of the cluster ion was sufficiently narrow as one would expect after TPEPICO and only one NH 2 ejection according to (1). Thus, even at the moderate photon energies used in the present study, only slightly above the appearance potentials, we obviously have to treat the fragmenting ensemble as a whole, using R R K M or phase space theory. This will be subject to future work. Thus, for qualitative discussion and later quantitative evaluation we define the metastable abundance ratio [(NH3)n_3NH2 ]m~tat;1/{[(NH3)n_3NH2 ]me~a+ [(NH3),_ 2NH~ ] } by normalizing the metastable fragment abundance to the sum of fragment and parent ion signal and to the time spent in the field free region. Figure 2 displays this quantity for different photon energies showing its evolution with cluster size. Similarly as in the muttiphoton work of Wei et al. [10] the metastable abundance ratio increases with the cluster size. No spectacular 'magic number' behaviour is seen in Fig. 2. Also, the dependence of metastability on photon energy yields no obvious structure which again suggests that these
patterns are the result of a sequence of subsequent fragmentation processes with an almost total loss of memory for the initial preparation of the cluster ions. The metastable abundance ratios have been determined as a function of background pressure at 9.54 eV. Except for the NH~ we find a nonvanishing true unimolecular fraction (extrapolation to zero pressure) as well as a clear collision induced enhancement proportional to the pressure. Evaporation of a (NH3)z neutral, which is not observed under normal operating conditions (p ~ 2.10-v mbar), also shows a linear relationship with pressure. This process was also observed in the MPI studies of Echt et al. [9]. We see, however, in contrast to these studies, a more or less general increase of the cross sections with cluster size. Figure 3 displays the homogeneous abundance ratios [(NHa)+]/{[(NH3) +] + [(NH3),_2NH+]} for the prompt cluster ion signals defined by normalizing the homogeneous parent ion signal to the sum of it and the protonated fragment signal arising from the reaction (t). Notice the large homogeneous abundance ratio for the dimer which reduces at higher photon energies and becomes comparable to the larger clusters at tl.3 eV, approximately 2 eV above its AP. Notice a similar effect for the trimer. Here, for the first time, we obviously see a memory for the initial preparation process. The most striking feature is the pronounced minimum at the 'magic' number n = 6 (closure of the first solvation shell of 4 NH3 around NH~-) at all photon energies. This could be dis-
15
15
10
i0
5
5
7~ ~ 0 "~15
5
6
7
8
910
©
0 15
iO
5
6
7
8
9
5
6
7
8
9 10
:s <10
10 5
0
5
6
7
8
9 10
-~10
0 15 I0
©
5 .3
.~
0
5
6
7
8
9 10
0
5 6 7 8 910 cluster:" size n
I0
0
5 6 7 8 9 10 cluster size n
Fig. 2. M e t a s t a b l e a b u n d a n c e ratios for seven discrete p h o t o n energies as a f u n c t i o n of cluster size. D a t a labeled ( * ) are t a k e n w i t h o u t blocking the m o n o c h r o m a t o r s 2nd o r d e r t r a n s m i s s i o n by a LiFfilter. T h e s h a d e d a r e a s indicate the statistical error. T h e time w i n d o w (in Its) for o b s e r v a t i o n o f m e t a s t a b t e decay (to, to + tl ) is for n = 5: (0.9, 9.9), n = 6: (1.0, t 1.1), n = 7: (1.0, 12.1), n = 8: (1.t, 13.0), n = 9: (1.2, 13.9) a n d n = 10: (1.2, 14.7)
196 a b u n d a n c e of the homogeneous dimer appears to be essentially due to fragmentation of larger clusters.
0,3 =
4. Conclusion 234567891Q
2 3 4 5 6 7 8 9 1 0 L~
c~
2345678910 m
2
3
4
5
0,3 [
o
6
?
8
91(I
11.3eV(*)
The present work displays interesting features of the fragmentation processes subsequent to single p r o t o n ionisation of a m m o n i a clusters on a p r o m p t as well as on a metastable time scale. Before final conclusions on details of the fragmentation dynamics can be drawn, further evaluation and statistical modelling of the data is necessary.
0,2
g
This work has been funded by the Bundesminister ffir Forschung und Technologic (BMFT) under contract number 05 472 AB B 5.
0, I © 2 3 4 5 6 7 8 910
2
3
4
5
6
?
8
910
c l u s t e r size n °'31
17.7eV(*)
0.1
c l u s t e r size n
Fig. 3. Homogeneous abundance ratios [(NH3)+]/{[(NH3)+]+ [(NH3),.2NH+]} as a function of cluster size at seven discrete photon energies. Bars marked by an arrow reach the value indicated to the right of the top. The shaded areas indicate the statistical error. Data labeled (*) are taken without LiF-filter
cussed in terms of enhanced stability [1,5, 10] of ( N H 3 ) 4 N H ~ or alternatively as being due to the particular low a b u n d a n c e of the corresponding (NH3)~- parent. Closer inspection of our mass spectra (e.g. Fig. la) seems to indicate the latter to be more important. The strong abundance of the h o m o g e n e o u s dimer together with theoretical and energetic considerations [3] provides a puzzle. It seems to indicate that reaction (4) which is not observed on the metastable time scale plays nevertheless a significant role in the initial time steps of the fragmentation chain. This is also supported by most recent experimental data [18] where the b e a m expansion conditions are varied over a wide range: the observed
References 1. Hirao, K., F@kawa, T., Konishi, H, Yamabe, S.: Chem. Phys. Lett. 104, 184 (1984) 2. Tomoda, S.: Chem. Phys. 110, 431 (1986) 3. Greer, J.C., Ahlrichs, R., Hertel, I.V.: (to be published) 4. Arshadi, M.R., Futrell, J.H.: J. Phys. Chem. 78, 1482 (1974) 5. Lifshitz, C., Louge, F.: J. Phys. Chem. 93, 6625 (1989) 6. Stephan, K., Futrell, J., Peterson, K., Castleman, A.W., Jr. Wagner, H., Djuric, N., M/irk, T.: Int. J. Mass Spectrom. Ion Phys. 44, 167 (1982) 7. Buck, U., Lauenstein, C.: J. Chem. Phys. 92, 4250 (1990) 8. Shinohara, H, Nishi, N.: Chem. Phys. Lett. 87, 561 (1982) 9. Echt, O., Dao, P.D., Morgan, S., Castleman, A.W. Jr.: J. Chem. Phys. 82, 4076 (1985) 10. Wei, S., Tzeng, W.B., Castleman, A.W.: J. Chem. Phys. 93, 2506 (1990) 11. Kamke, W., Kamke, B., Wang, Z., Herrmann, R., Hertel, LV.: In: Physics and chemistry of small clusters. Jena, P., Rao, B.K., Khanna, S.N. (eds.). New York: Plenum Press 1987 12. Kamke, W., Herrmann, R., Wang, Z, Hertel, I.V.: Z. Phys. D - Atoms, Molecules and Clusters 10, 491 (1988) 13. Klots, C.E.: J. Phys. Chem. 92, 5864 (1988) 14. Kamke, W., de Vries, J, KrauB, J., Kaiser, E., Kamke, B., Hertel, I.V.: Z. Phys. D - Atoms, Molecules and Clusters 14, 339 (1989) 15. KrauB, J., de Vries, J., Kaiser, E., Kamke, B., Kamke, W.: Z. Phys. D - Atoms, Molecules and Clusters (this issue) t6. Wiley, W.C, McLaren, I.H.: Rev. Sei. Instrum. 26, 1150 (1955) 17. Mamyrin, B.A., Karataev, Vd., Zagulin, V.A.: Soy. Phys. -JETP 37, 45 (1973) 18. Kamke, W., de Vries, J., Menzel, C., Steger, H., Hertel, I.V.: (to be published)