Three-body Reaction Dynamics in Dissociative
Recombination
R. Thomas , F. Hellberg , M. Larsson , C. R. Vane† , P. Andersson , J. Pettersson ,
A. Petrignani‡ and W. J. van der Zande‡
Department of Physics, University of Stockholm, SCFAB, Stockholm, Sweden.
†
Oak Ridge National Laboratory, Oak Ridge, TN 37831-6377, USA.
Department of Chemistry, Göteborg University,SE-412 96 Göteborg,Sweden.
‡
FOM Instituut AMOLF, PO Box 41883, 1009 DB Amsterdam, The Netherlands.
Abstract. Dissociative recombination of molecular ions with electrons is the most important neutralising process in plasmas
cold enough to contain molecules. The basic features and principles describing recombination of diatomic molecular ions are
now reasonably well characterised and understood. However, recombination of polyatomic ions is much less well understood.
Over the last six years, experiments carried out at ion storage rings have shown that tri-atomic molecular ions tend to break
up into three atoms upon recombination with free electrons. The question of how this break-up occurs has started to be
investigated at ion storage rings using particle-imaging techniques. In this presentation,
used in these
the imagine technique
the use of
experiments will be discussed together with results obtained from studies of H2 O , NH2 and CH2 . Finally,
this technique to study the dissociative recombination of more complex polyatomic ions, for example D5 O2 , will also be
discussed.
INTRODUCTION
Dissociative recombination (DR) is a process in which
a molecular ion recombines with an electron and, to
achieve stability, the resulting molecule fragments into
neutral products. For plasmas containing molecular ions,
DR is the most important neutralising process. Thus,
DR is of great importance to the chemistry occurring
in such diverse regions as interstellar clouds, planetary
atmospheres and in semi-conductor etching.
In the last few years, many DR studies have been related to the break-up of polyatomic ions, mainly driven
by the situation that given the apparent simplicity of the
DR process, developing a general theory to predict product branching ratios for even the simplest polyatomic
ions, e.g. XH2 , has proven to be difficult. The earliest
models suggested that the reaction would predominantly
proceed via single bond breaking, i.e, the weakest X-H
bond, giving H + XH. However, all recent storage ring
studies for such ions show a propensity for three-body
break-up, i.e. X + H + H, see [1] for a comprehensive
list.
To try and understand the processes occurring in these
simple systems, the dynamics of the three-body fragmentation channel has been
investigated
in more detail.
So far, results on H2 O [2] and H3 [3] have been reported, and a review on the three-body dynamics also
published.[1] For H2 O , theoretical trajectory calculations were employed on the known potential energy surfaces, for an understanding of the experimental observations, and these were also recently published. [4]
Furthermore, the group at TSR looked at the dynamics in, and competition between, the two- and threebody fragmentation dynamics of H3 . Using a statistical
model to describe the process, they reached good agreement with those experimental data previously obtained at
CRYRING [5] as a function of reaction energy.[6]
It is interesting to examine similar systems in detail
to see if more insight can be obtained into the dynamical/statistical nature of the DR process. Furthermore, it
will be instructive to examine larger systems, especially
systems with molecular fragments, to see if information can be obtained on how much of the available reaction energy goes into internal excitation of the molecular
fragments.
DISSOCIATIVE RECOMBINATION IN
STORAGE RINGS
All the experiments discussed in this paper were carried
out at the heavy ion storage ring CRYRING located at
the Manne Sieghbahn Laboratory, which is part of the
University of Stockholm in Sweden. In the last 10 years
CP680, Application of Accelerators in Research and Industry: 17th Int'l. Conference, edited by J. L. Duggan and I. L. Morgan
© 2003 American Institute of Physics 0-7354-0149-7/03/$20.00
187
EXPERIMENTAL DISCUSSION
In this article, a selection of results from recent experi
mental work at CRYRING
on the DR of some XH2 sys
tems, i.e., H2 O , NH2 and CH2 , as well as the larger
system D (D2 O)2 will be discussed. At 0 eV collision
energy, the following three-body fragment channels, with
their associated kinetic energy release (KERn), are allowed:
H2 O
NH2
H2 C
D
D2 O 2
O(3 P)
e
O(1 D)
N(4 S)
e
N( D)
N(2 P)
C(3 P)
C(1 D)
D(2 S)
2
e
e
2H(2 S)
2H(2 S)
2H(2 S)
2H( S)
2H(2 S)
2H(2 S)
2H(2 S)
2
2(D O)
2
KER1 (1)
KER2 (2)
KER3 (3)
KER4 (4)
KER5 (5)
KER6 (6)
KER7 (7)
the following three main parameters to be obtained:
•
•
•
the branching ratio between the available channels,
determined from distributions of the measured distances between the particles
the intra-molecular angle between the two h atom
fragments and
the energy distributed to the two h atom fragments.
For D (D2 O)2 , preliminary results are discussed using
•
the 3 2-particle distance distributions between the
D, D2 O and D2 O.
RESULTS AND INTERPRETATION
Example
plots of the distance distributions
for H2 O and
NH2 , the angular distributions for H2 O and H2 C , and
the h atom energy distributions for H2 O and NH2 are
illustrated in Figures 1, 2 and 3, respectively. An example
plot
of the 2 particle distance distributions obtained from
D (D2 O)2 is shown in Figure 4.
3500
HN+2
H2O+
3000
2500
Counts /Arb. Units
or so, the use of such facilities for the study of DR processes has increased due to several experimentally desirable aspects, and these are discussed here. The high
beam energy, which is in the MeV range, is an advantage in several respects. The electron capture cross section in collisions between the stored ion beam and residual gas molecules is small at high beam energy, which
significantly reduces the background. Furthermore, the
good vacuum < 10 11 Torr in the experiment indicates
that the number of residual gas molecules is extremely
low, further adding to the long storage lifetime of the
ion beam. This time, which can be tens of seconds, allows metastable components to be removed and vibrationally excited levels to be relaxed. A review on the use
of merged-beams in atomic and molecular physics has
recently been published.[7]
2000
N(2D) limit
≈ 1.3 eV
O(3P) limit
≈ 3.0 eV
1
1500 O( D) limit
≈ 1.1 eV
N(4S) limit
≈ 3.7 eV
1000
500
0
0
5
10
15
20
25
30
Measured Projected KER /pixels
35
40
FIGURE
1. Measured Projected KER for the DR of H2 O
and NH2
KER8 (8)
It is noted that for (8) this is the available reaction energy and is available for internal excitation of the molecular fragments. Using time- and position-sensitive detectors (briefly, a stack of multi-channel plates (MCPs), a
phosphor screen and a photo-multiplier tube (PMT). See
[8, 9]) in a triple-coincidence experiment, monitoring the
position and arrival time of the fragments from the DR
reaction enables differentiation between the competing
channels. Further
to the standard detector arrangement,
used in the D (D2 O)2 study, a new technique was employed for the XH2 systems which enabled the heavier,
X, atom to be identified [2, 4]. In identifying the heavier
atom, the centre of mass (c.m.) is determined, allowing
188
Interpretation
For the XH2 study, due to the resolution of the detection system, it can be seen in Figure 1 that it is not
possible to completely separate the distance distributions
arising from the competing channels. Initial conclusions
from the angular
distributions shown in Figure 2 indicate
that H2 O preferentially dissociates from surfaces having an open geometry, with some preference for a closed
geometry, whilst H2 C from surfaces having an open geometry. Finally, though it is difficult to draw any conclusions from the h-atom energy
(Figure 3), it is
distribution
clearly seen seen that H2 O and NH2 are different.
To obtain accurate information on all the necessary parameters, as well as any understanding on the data ob-
70
tors of the fragment atoms were simulated, taking into
account the random orientation and location of the DR
event in the electron cooler. The main parameters in the
simulation are the available energy, the angle between the
fragments and the energy given to the fragments. Manipulation of these parameters, while simulating the same
“measurements" as were done in the laboratory, enables a
detailed comparison to be made between the initial/final
conditions in the simulation and the experimental observations. For example, Figure 5 shows the branching between the available channels for H2 O . The simulation
X = C(arbon)
X = O(xygen)
60
Counts (Arb. units)
50
40
30
20
10
0
0
20
40
60
80
100
120
140
160
180
H−c.m.−H Angle (Degrees)
3500
FIGURE 2. Measured Intra-molecular angle for XH2 , where
X = O and C
3000
2D Distributions
1
3
Sim. ( D) + ( P)
1
Sim ( D)
3
Sim ( P)
1
O ( D) limit
~ 1.1 eV
Counts (arb. units)
2500
2.5
+
HN
2
H O+
P(c.m.−H)/Distance /Arb. units
2
2
1.5
3
1500
3
O ( P) limit
~ 3.0 eV
1000
500
0
0
1
5
10
15
20
25
30
35
Measured TD (pixels)
0.5
FIGURE 5.
0
0
5
10
15
20
25
30
Distance /Pixels
FIGURE 3. Measured h-atom energy distribution for XH2 ,
where X = O and N
tained from the D (D2 O)2 study, a Monte Carlo simulation was written.
The main details on the simulation are discussed in [4]
(and see [10]). Briefly, the asymptotic momentum vec4
3
x 10
2.5
D2O <−> D2O
2
Counts
1
Ratio ( D):( P)
1:3.5(0.5)
2000
1.5
D <−> D2O
1
0.5
0
0
10
20
30
40
50
60
2 Particle Distances /Pixels
FIGURE
4. Measured 2 Particle distance distributions from
D (D2 O)2
189
Simulated Projected KER [TD] for DR of H2 O
allowed information and insight to be obtained on the
DR processes. The branching ratios between the competing
three-body fragmentation channels for each of the
XH2 systems studied is given in Table 1 For the angular
TABLE 1. Three-body branching for XH2 , where X = O,
N and C
Ion
H2 C H2 N H2 O
Ground state
C(3 P)
N(4 S)
O(3 P)
First Excited state
C(1 D)
N(2 D)
O(1 D)
Ratio
1:1
1:1
3.5:1
distributions, it was suggested that H2 O preferentially
dissociates from surfaces having an open geometry,
with
some preference for a closed geometry, with CH2 preferentially from surfaces having an open geometry. The
simulation matched the experimental observations best if
these conditions were true. Finally, for the energy distribution over the two h-atoms, Figure 6 plots data for three
different cases of the energy partition; ρ = 1 (equally distributed), ρ = 0.1 (one atom gets 90%), and where it is
randomised
ρ = c. Also plotted are data from H 2 O and
NH2 . It can be seen that the best match for H2 O is when
ρ = c and, for NH2 , when ρ = 1. For the larger system,
D (D2 O)2 , the simulation also allowed information to be
extracted on the internal excitation of the D2 O fragments.
Any internal excitation reduces the available kinetic energy, with the majority taken by the D atom. By simulating different available kinetic energies for the fragments,
9
P(c.m.−H)/Distance /Arb. units
detail than ever before. The information obtained from
these experiments indicates that in some cases a high degree of dynamics is occurring in the process, leading to a
scrambling of the available reaction energy and internal
molecular angles. The same techniques also allow us to
probe the internal excitation of any molecular fragments,
which is important in plasmas where any heating of the
plasma due to fragments of high kinetic energy is critical.
ρ=c
ρ = 1.0
ρ = 0.1
+
NH
2+
HO
8
7
2
6
5
4
3
2
1
0
0
3
6
9
12
15
18
21
24
27
ACKNOWLEDGMENTS
30
Distance /Pixels
FIGURE 6. Measured and simulated h-atom energy distribution for XH2 , where X = O and N
and comparing the resulting distributions with the experimental observations, it is possible to see if there is any
internal excitation of the D2 O’s. For example. Figure 7
plots data for such a set of simulations, where a good fit
to the observed data is obtained when there is internal
excitation of the D2 O’s up to about 3 eV, i.e. in some
cases there is only 2.1 eV given as kinetic energy to the
fragments.
The authors are grateful to the staff members of the
Manne Siegbahn Laboratory for their assistance and help
in the experiment. This work is supported by the following organisations: The Swedish Science Council, the
Swedish Foundation for International Cooperation in Research and Higher Education, the Fifth Framework Programme of the EU (HPRN-CT-2000-00142), Stcihting
voor Wetenschappelijk Onderzoek, and the U.S. Department of Energy, Offices of Basic Energy Sciences, Division of Chemical Sciences under Contract No. DEAC05-00OR22725 with UT-Battelle, LLC.
REFERENCES
4
3
x 10
Experimental data
Simulated data
2.5
D2O <−> D2O
1.
2.
Counts
2
1.5
D <−> D2O
1
0.5
0
0
10
20
30
40
50
60
Expt. and Sim. 2 Particle Distances /Pixels
FIGURE
7.
D (D2 O)2
Comparison
distance
distributions
from
COMMENTS AND CONCLUSIONS
DR is an extremely important process in cold, molecular
plasmas. An understanding of the process, specifically
the dominance of a high degree of fragmentation, is important to the chemistry of these regions, as the ionisation
balance and production of reactive free radical species is
critical for establishing chemical networks, A heavy ion
storage ring, which provides un-paralleled experimental
conditions, coupled with the latest in detector techniques,
enables us to study these and related systems in greater
190
M. Larsson and R. Thomas, PCCP, 3, 4471, 2001.
S. Datz, R. Thomas, S. Rosén, M. Larsson, A.M.
Derkatch, F. Hellberg, and W. van der Zande, Phys. Rev.
Lett. 85, 5555 2000.
3. D. Strasser, L. Lammich, S. Krohn, M. Lange, H. Kreckel,
J. Levin, D. Schwalm, Z. Vager, R. Wester, A. Wolf and
D. Zajfman, Phys. Rev. Lett., 86, 779 2001.
4. R. Thomas, S. Rosén, F. Hellberg, A. Derkatch, M.
Larsson, S. Datz, R. Dixon and Wim J. van der Zande,
Phys. Rev. A 66, 032715 2002.
5. S. Datz, G. Sundström, Ch. Biederman, L. Broström, H.
Danared, S. Mannervik, J. R. Mowat and M. Larsson,
Phys. Rev. Lett., 74, 896 1995.
6. D. Strasser, L. Lammich, S. Krohn, M. Lange, H. Kreckel,
J. Levin, D. Schwalm, Z. Vager, R. Wester, A. Wolf and
D. Zajfman, Phys. Rev. A, 65, 010702(R) 2001.
7. R.A. Phaneuf, C.C. Havener, G.H. Dunn, and A. Müller,
Rep. Progr. Phys. 62, 1143 1999.
8. Z. Amitay and D. Zajfman, Rev. Sci. Instrum., 68, 1387
1997
9. J. R. Peterson, A. Le Padellec, H. Danared, G. H. Dunn,
M. Larsson, Å Larson, R. Peverall, C. Strömholm, S.
Rosén, M. af Ugglas and W. J. van der Zande, J. Chem.
Phys., 108, 1978 1998
10. U. Müller and P. C. Cosby, Phys. Rev. A, 59, 3632 1999.