Dissociation pathways in low energy (0-2 eV) electron attachment to Cl2O

Article abstract of DOI:10.1016/S0009-2614(01)00814-4

Dissociative electron attachment (DA) to ClOCl is studied in a high resolution crossed beam experiment. Two complementary ion pairs, Cl^-/ClO^- and O^-/Cl₂^-, are observed. The Cl^-/ClO^- pair arises from a simple Cl-OCl bond cleavage with the electron sitting on either of the two fragments. The O^-/Cl₂^- pair is formed by a concerted reaction with the expulsion of O^- (or O) and formation of Cl₂ (or Cl₂^-). Ab initio calculations indicate that in low energy electron attachment an electronically excited state of the precursor anion (ClOCl^-* (²B₂)) is involved.

Full text of DOI:10.1016/S0009-2614(01)00814-4

31 August 2001
Chemical Physics Letters 344 ꢀ2001) 471±478
www.elsevier.com/locate/cplett
Dissociation pathways in low energy ꢀ0±2 eV) electron
attachment to Cl2O
Wolfgang Sailer a, Petra Tegeder b, Michael Probst a, Herwig Drexel a,
Verena Grill a, Paul Scheier a, Nigel J. Mason b, Eugen Illenberger c,*
,
a,d

Tilmann D. Mark
a
Institut fur Ionenphysik, Leopold Franzens Universitat Technikerstrasse 25, A - 6020 Innsbruck, Austria
Department of Physics and Astronomy, University College London, Gower Street, London WC1E 6BT, UK
b
c
Institut fur Chemie-Physikalische und Theoretische Chemie, Freie Universitat Berlin, Takustrasse 3, D-14195 Berlin, Germany
Department of Plasmaphysics, Comenius University, 84248 Bratislava, Slovak Republic
d
Received1 May 2001; in ®nal form 2 July 2001
Abstract
Dissociative electron attachment ꢀDA) to ClOCl is studied in a high resolution crossed beam experiment. Two
complementary ion pairs, ClÀ=ClOÀ andO À=Cl2À, are observed. The ClÀ=ClOÀ pair arises from a simple Cl±OCl bond
cleavage with the electron sitting on either of the two fragments. The OÀ=Cl2À pair is formedby a concertedreaction
with the expulsion of OÀ ꢀor O) andformation of Cl 2 ꢀor ClÀ2 ). Ab initio calculations indicate that in low energy electron
ÀÃ
2
attachment an electronically excitedstate of the precursor anion ꢀClOCl
Elsevier Science B.V.
ꢀ B2†) is involved. Ó 2001 Publishedby
1. Introduction
source of ClO ꢀeither through direct photolysis [3]
or reactions with oxygen [4] or halogen radicals
[5,6]) in laboratory experiments andClO is central
to the catalytic loss processes of ozone in the
stratosphere. Dichlorine monoxide in the structure
Cl±O±Cl is non-linear with an angle /ꢀCl±O±
The chlorine oxides ꢀClxOy† have been exten-
sively studied in the last few years due to their
important role in reactions leading to ozone de-
struction in the Earth's stratosphere [1,2]. Al-
though dichlorine monoxide, Cl2O, plays only a
minor role in atmospheric chemistry, studies of its
chemical behaviour as well as its characterization
ꢀgeometry, electronic structure, etc.) is of general
interest for comparison with other more relevant
atmospheric chlorine oxides. Cl2O is also usedas a
Cl) ˆ 110.9° andhas
a
bondlength
RꢀCl±
O) ˆ 169.6 pm [7].
In this Letter we present results of a study of
electron attachment to ClOCl using a high reso-
lution crossedbeam experiment with mass spec-
trometric
detection
of
the
ions.
comparatively weak bondidssociation energy
The
DꢀCl±OCl) ˆ 1.48 eV [8] leads to the dissociative
attachment ꢀDA) channel ClÀ ‡ ClO ꢀandalso its
complement Cl ‡ ClOÀ) being considerably exo-
thermic ꢀsee below). It is also interesting to
* Corresponding author. Fax: +30-5838-6612.
E-mail address: iln@chemie.fu-berlin.de ꢀE. Illenberger).
0009-2614/01/$ - see front matter Ó 2001 Publishedby Elsevier Science B.V.
PII: S 0 0 0 9 - 2 6 1 4 ꢀ 0 1 ) 0 0 8 1 4 - 4

472
W. Sailer et al. / Chemical Physics Letters 344 62001) 471±478
compare the present results with the previously
studied OClO ꢀchlorine dioxide) [9]. For this
compoundthe corresponding numbers are /ꢀO±
Cl±O) ˆ 117.4°, RꢀO±Cl) ˆ 147 pm [10] and DꢀO±
ClO) ˆ 2.56 eV [9]. Both chlorine oxides also exist
in their respective isomeric forms ClOO and
OClCl. While in chlorine dioxide the two isomers
are very close in energy ꢀwith ClOO more stable
by about 0.013 eV [11]), the compoundClOCl is
predicted to be more stable by about 0.75 eV than
its isomeric form ClClO [12] which is not a stable
species in the gas phase.
The photo-absorption spectrum of ClOCl in
the range 650 < k < 220 nm ꢀ1.9±5.6 eV) exhibits
three broadandunstructuredbands [13±15] which
are assignedto the excitation of valence electronic
states with dissociative character [16]. Additional
strong features in the vacuum ultraviolet have
been revealedby experiments using synchrotron
radiation [17]: a strong continuous band with a
peak at 171 nm ꢀ7.25 eV) andseveral vibration-
ally resolvedbansd in the energy region being
foundbetween 7.5 and10 eV. The continuous
bandis interpretedas excitation to a dissociative
valence excitedstate while the higher energy
structured bands are classi®ed into several Ryd-
berg states.
the ionic products. The modi®ed TEM has re-
cently been described in detail [22] where it has
been shown that the very high energy resolution
ꢀseveral meV) at energies close to zero eV re-
portedearlier [23,24] quickly edtoriates to 100
meV with increasing electron energy. Using elec-
tron trajectory calculations, a modi®ed version of
a TEM has been designed to correct for these
de®ciencies. With the new monochromator, an
energy resolution of 50 meV can routinely be
achieved independent of the electron energy.
However, for the present experiments, where only
small amounts of sample were available, the in-
strument was operatedat an FWHM of ꢁ90
meV. In addition, introduction of the reactive
dichlorine monoxide causes some shift in the en-
ergy scale due to adsorption of ClOCl molecules
on spectrometer surfaces. This shift can be sta-
bilizedby ¯owing the gas through the instrument
prior to collecting data. A small amount of SF6 is
added to the sample throughout the experiments
to monitor the energy resolution andthe shift of
the energy scale via the narrow SFÀ6 feature close
to zero energy [25].
The ion draw out ®eld in the collision region is
very small ꢀ<0.1 VcmÀ1) which makes the Inns-
bruck device particularly sensitive for studying
processes at very low electron energies. However,
the instrument discriminates, to some degree,
against ions produced with higher translational
energy.
Early studies of Cl2O photolysis [18,19] showed
that the Cl±OCl bond®ssion is the odminant
fragmentation pathway. More recent studies using
laser excitation of jet cooledCl O molecules and
2
applying fragment translational spectroscopy and
ion imaging techniques [8,20,21] revealeda much
more detailed picture on the complex photo-dis-
sociation dynamics ꢀkinetic energy release and
angular distribution of fragments) including the
dissociation pathway leading to Cl2 ‡ O in addi-
tion to the usually dominant Cl±OCl bond
breakage.
Dichlorine monoxide samples were prepared
using the University College mobile ClOCl gener-
ator. A stream of chlorine gas passes through a
reaction column packedwith mercuric oxide at
room temperature. Gaseous dichlorine monoxide
is formedaccording to the reaction
2Cl2ꢀg† ‡ 2HgO ! HgCl2 Á HgO ‡ Cl2Oꢀg† ꢀ1†
The e‚uent gas shows the presence of a Cl2 im-
purity as can be seen on the ClÀ channel in the
observedDA spectrum ꢀsee discussion below) and
from the previous absorption experiments [17].
From a scaling procedure used in these absorption
experiments the sample purity was estimatedto be
90%. It is well known that this standardized pro-
duction methodology does not produce any other
impurities e.g. OClO.
2. Experimental
The DA spectrometer usedin these experi-
ments consists of a molecular beam system, a
high resolution trochoidal electron monochro-
mator ꢀTEM) anda quadrupole mass ®lter with a
pulse counting system for analyzing anddetecting

W. Sailer et al. / Chemical Physics Letters 344 62001) 471±478
473
OÀ ˆ 32:100:45:<1 ꢀpeak values)) di€er slightly
from those reportedin the previous work
ꢀ100:25:25:<1) [26]. The higher ClÀ intensity can
be attributedto the higher detection eciency of
energetic ions in the previous device. An analysis
of the kinetic energy release [26] showeda mean
kinetic energy of 0.2 eV for ClÀ at thresholdꢀ0 eV)
increasing with electron energy. Although no ki-
netic energy analysis was performedfor the di-
atomic anions we expect less translational energy
due to the smaller exothermicity of those channels
ꢀEqs. ꢀ2a)±ꢀ2c)) andalso from momentum con-
servation.
3. Results and discussion
Figs. 1 and2 show the negative ions formedby
electron impact with ClOCl in the energy range up
to 2 eV. Above that energy only the ClÀ channel
revealeda small signal which may arise from the Cl 2
impuritiy. In addition, a very weak anion signal of
the stoichiometric composition of the parent anion
close to zero eV is observedꢀnot shown here).
The present results are in goodagreement with
those of a previous ꢀunpublished) study of electron
attachment to ClOCl using an electron beam with
a broader energy width ꢀ0.2 eV FWHM) and an
ion extraction ®eldof 2 V cm À1 [26]. The presently
obtainedrelative intensities ꢀCl À: ClOÀ: Cl2À:
The present experiments have been performed
at a pressure of 5 Â 10À6 mbar ꢀas readon an
À
Fig. 1. Formation of the complementary ions ClÀ andClO
from ClOCl as a function of the electron energy. The insert
shows the two contributions, i.e., s-wave near zero eV and
Gaussian curve at 145 meV ꢀdeconvoluted data), from which
the experimental data can perfectly be synthesized ꢀsee the text).
Fig. 2. Formation of the complementary ions OÀ andCl 2À from
ClOCl as a function of the electroÀn energy. For comparison the
À
For comparison the SFÀ6 yieldꢀnormalizedto the Cl
height) from a small admixture of SF6 is also shown.
peak
SFÀ6 yieldꢀnormalizedto the O
admixture of SF6 is also shown.
peak height) from a small

474
W. Sailer et al. / Chemical Physics Letters 344 62001) 471±478
eÀ ‡ ClOCl ! ClÀ ‡ ClO ꢀDH0 ˆ À2:13 eV†
eÀ ‡ ClOCl ! ClOÀ ‡ Cl ꢀDH0 ˆ À0:80 eV†
eÀ ‡ ClOCl ! OÀ ‡ Cl2 ꢀDH0 ˆ ‡0:32 eV†
eÀ ‡ ClOCl ! Cl2À ‡ O ꢀDH0 ˆ À0:60 eV†
ꢀ2a†
ꢀ2b†
ꢀ2c†
ꢀ2d†
ionization gauge on one of the chamber ¯anges)
andthe count rates are given in absolute num-
bers. The pressure in the reaction volume is 2±3
orders of magnitude higher as can be estimated
from the pumping speedandthe diameter of the
capillary forming the e€usive molecular beam.
The absolute value for the attachment rate ka of
thermal electrons to ClOCl is available from a
previous electron cyclotron resonance ꢀECR)
study as ka ˆ 1:5 Æ 0:5  10À10 cm3 sÀ1 [27]. The
rate constant, ka, andcross-section, r, are con-
with the thermodynamic thresholds ꢀDH0) being
calculatedfrom the relevant electron anities
eV,
ꢀEAꢀClO) ˆ 2.28
EAꢀCl2† ˆ 2:38
eV,
EAꢀO) ˆ 1.46 eV andEAꢀCl) ˆ 3.61 eV [28]) and
bonddissociation energies ꢀ DꢀCl±OCl) ˆ 1.48 eV
[8], DꢀCl±Cl) ˆ 2.52 eV, DꢀCl±O) ˆ 2.82 eV [28]).
The EA values are known with spectroscopic ac-
curacy but are given here to only two digits. Note
that the two reaction pairs ꢀ2a)/ꢀ2b) andꢀ2c)/ꢀ2d)
are complementary with respect to the negative
charge with the ®rst pair corresponding to a simple
Cl±OCl bondcleavage andthe secondto the ex-
pulsion of OÀ ꢀor O) andformation of Cl 2 ꢀor ClÀ2 ).
The cross-section for Àthe expulsion of neutral O
nectedby
Z
ka ˆ f ꢀt†trꢀt†dt;
where f ꢀt† is the normalizedvelocity distribution
of the electrons at the correct gas temperature. As
a simple approximation we can use the averaged
quantities yielding
andformation of Cl
is of the same order of
2
ka ˆ hrihti
magnitude as the simple Cl±OCl bond cleavage,
i.e., formation of ClÀ or ClOÀ. This is in contrast to
photo-dissociation studies [20] which showed that
the formation of O ‡ Cl2 was only observedat 193
nm although the energetic thresholdis close to 700
nm. Excitation at 193 nm creates highly excitedCl 2
fragments with the majority undergoing sponta-
neous dissociation into atomic Cl fragments.
It can be seen in Figs. 1 and2 that the width of
the peaks andalso the peak shapes di€er between
andassume an average velocity from a standard
Maxwell±Boltzmann distribution ꢀhti ꢁ 107 cm sÀ1
at 300 K). We then derive a mean cross-section of
hri ꢁ 1:5Â10À17 cm2 for the attachment of thermal
electrons at 300 K. An estimate of the absolute
cross-section from the present experiment by
comparing with the SFÀ6 intensity was not possible
since the mixing ratio between ClOCl andSF
could not be determined under the ¯ow conditions
of ClOCl production.
6
À
the fragment ions; the ClÀ andClO
ion yields
As is evident from Figs. 1 and 2 electron at-
tachment to ClOCl is almost exclusively a dis-
sociative process producing fragment ions at low
energies. We can therefore conclude that in the
energy range below ꢁ0.3 eV the DA cross-sec-
tion ꢀunder single collision conditions) is in the
range of 10À17 cm2. This is more than one-order
of magnitude larger than the cross-section for
photolysis in the visible andUV: the total ab-
sorption cross-section ꢀwhich can be viewedas an
upper limit for the cross-section of photo-disso-
ciation) being in the range between 10À20 and
10À18 cm2 for photon energies between 2 and
6.5 eV [16].
seem to have some speci®c structure. To discuss
andinterpret the shape of these anion resonances
andtheir origin we needto recall how DA pro-
ceeds in a polyatomic molecule.
In the simplest case of DA to a diatomic mole-
cule via a single repulsive potential energy curve
without exciting any internal degrees of freedom in
the product fragments and with the dissociation
asymptote below the Franck±Condon regionÀꢀÃe.g.,
as in the oxygen system eÀ ‡ O2 ! O2
!
OÀ ‡ O) the ion yieldhas a simple Gaussian shape
ꢀre¯ection principle). The peak is eventually
modi®ed and red-shifted by an energy dependent
autodetachment rate ꢀsurvival probability shift)
which is the competing channel to dissociation.
Consequently the DA cross-section, rDA, is the
result of a two-step process characterizedby the
At low energies ꢀ<0.5 eV) the fragment ions
observedꢀCl À, ClOÀ, OÀ andCl À2 ) can be assigned
to one of the four following DA reactions:

W. Sailer et al. / Chemical Physics Letters 344 62001) 471±478
475
electron capture cross-section, r0, andthe disso-
ciation probability, viz,
rDA ˆ r0  P;
ꢀ3†
where P expresses the probability that the ionic
system survives dissociation with respect to au-
todetachment ꢀ0 < P < 1). In a polyatomic mole-
cule we can still describe the cross-section by Eq.
ꢀ3) but the situation becomes more complex with
respect to both steps: ꢀi) the increasing density of
electronic states can result in energetically over-
lapping resonances accessible via Franck±Condon
transitions andꢀii) P represents ꢀin addition to the
competitive channel of autodetachment) the
probability of decomposing into the particular
channel under observation ꢀwith respect to the
other energetically available channels) including
multiple fragmentation. P also contains the prob-
ability for creating the fragments in internally ex-
citedstates. Thus structures in the ion yieldcurve
can arise from di€erent overlapping precursor
states and by the opening of additional dissocia-
tion channels.
We have carriedout ab initio coupledcluster
ꢀCCSDꢀT)) calculations [29] with the aug-cc-
pVTZ basis set [30] on both the neutral andan-
ionic compoundin an attempt to gain some in-
formation on the structure andstability of the
anion andon the character of the lowest unoc-
cupiedmolecular orbitals ꢀLUMOs) involvedin
electron capture. For the neutral molecule the
calculation reproduces the experimental bond
length andCl±OCl angle within 1%. In the re-
laxedanion the bondlength is increasedby about
13% to 193.3 pm andthe angle by 18% to /ꢀCl±
O±Cl) ˆ 131°. The energy di€erence between the
neutral andthe relaxedanion ꢀby de®nition the
adiabatic electron anity) is predicted as
EAꢀClOCl) ˆ 2.07 eV. At this level of calculation
the energy of the lowest dissociation channel
ClÀ ‡ ClO is at )2.04 eV indicating a very shal-
low potential minimum ꢀ30 meV) for the ground-
state anion, see Fig. 3 ꢀthe asymptotic energy
obtainedfrom the literature thermochemical data
is at )2.13 eV). The vertical attachment energy
ꢀVAE) is calculatedto be )1.14 eV, i.e., the en-
ergy of the anion at the geometry of the neutral
ꢀR0 in Fig. 3) is 1.14 eV below that of the neutral
Fig. 3. Schematic two dimensional potential energy curve il-
lustrating dissociative electron attachment to ClOCl. The dot-
tedcurves represent the two low lying states of the anion ꢀsee
the text). The distance refers to the simple Cl±OCl bond ®ssion
ꢀchannels ꢀ2a) andꢀ2b)). The thermoydnamic limits of the
concertedreaction ꢀchannels O ‡ ClÀ2 andO ‡ Cl2) is also
À
shown.
andhence this anion state is not accessible by free
electrons.
We note that a very recent theoretical treatment
of electron collisions with ClOCl using the R ma-
trix method[31] yieldeda vertical detachment en-
ergy of )0.443 eV in qualitative agreement with
the present results but apparently foundno bound
ground-state anion.
In terms of molecular orbitals ꢀMOs) the
ground-state anion can be approximately de-
scribedby the addition of the extra electron into
the LUMO 1. Since this state is not accessible via a
Franck±Condon transition by free electrons the
higher lying unoccupiedMOs must be involvedin
electron attachment.
Our calculations indicate a series of relatively
closely spacedꢀ1 eV or less) virtual MOs which
are plottedin Fig. 4. LUMO 1 ꢀa symmetry; Fig.
1
4a) andLUMO 2 ꢀb symmetry; Fig. 4b) ꢀC2v)
2
are both Cl±O antibonding with considerable
density toward the Cl atoms while LUMO 3 ꢀFig.
4c) is extended and has no speci®c valence char-
acter.
On the basis of the one electron picture andour
ab initio calculations we conclude that the ®rst
virtual MO cannot be involvedin electron at-
tachment andinterpret the experimental observa-
tions in the following way:

476
W. Sailer et al. / Chemical Physics Letters 344 62001) 471±478
The unresolvedstructures ꢀseparatedby about
À
110 meV) in the comparatively broadClO
yieldcurve may arise from dissociation into the
ion
2
two spin±orbit states of Clꢀ2P3=2) andClꢀ P1=2
which are separatedby 109 meV. Although this
)
reaction is considerably exothermic, di€erent
probabilities P for the dissociation into the dif-
ferent spin±orbit states may result in some
structure in the overall DA cross-section. From
the shape of the ion yieldcurve one wouldthen
expect a larger population of the spin±orbit
groundstate. In photo-idssociation at 235 nm
ꢀ5.3 eV), 90% of the total Cl intensity consists of
ground-state Cl ꢀJ ˆ 3/2) [8]. An alternative ex-
planation for the structures is a contribution
from s-wave attachment close to zero eV ꢀhaving
an EÀ1=2 and EÀ1 energy dependence [25,32]).
Convolution of a cross-section containing a low
energy s-wave contribution ꢀpeak at zero eV)
anda Gaussian curve ꢀmaximum at 145 meV)
with the SFÀ6 curve ꢀthe monochromator func-
tion) in fact perfectly reproduces the experi-
mental distribution ꢀsee insert of Fig. 1). Such s-
wave contributions producing peaks at zero en-
ergy can also arise from transitions involving
vibrationally excitedstates [33]. They can even
arise in the case of very poor Franck±Condon
overlap where substantial attractive long-
range electron±molecule interactions are present
[34].
Fig. 4. Isodensity plot of the lowest three ꢀa±c) unoccupied
MOs in ClOCl. Isosurfaces are plottedat values of
‡0:02 bohrÀ3 ꢀdark shading) and of À0:02 bohrÀ3 ꢀlight shad-
ing).
It is interesting to note that the two channels
ꢀ2a) andꢀ2b) associatedwith a simple bond
cleavage have their peak maximum at some
distinct energy above the maximum of the SF6
ion yield, i.e., above zero. In terms of potential
energy curves this is due to a curve crossing at
some position to the right of the minimum of
the neutral con®guration ꢀFig. 3) which in turn
indicates some activation energy. A small acti-
vation energy ꢀ0.02 eV) for the attachment of
thermal electrons in fact was observedin the
early ECR work [27]. Hence we have a situa-
tion of direct bond breaking of considerable
exothermicity, albeit with a small activation
energy.
Dissociative electron attachment proceeds
through a single precursor state characterizedby
accommodation of the electron into the second
2
virtual MO representing the B2 shape resonance
locatedbetween 0 and ꢁ0.7 eV in the Franck±
Condon region. This resonance dissociates into the
observedDA channels ꢀindicatedin Fig. 3 by the
arrow).
À
The structure observednear 0.6 eV in the Cl
ion yieldcan be interpretedby the opening of the
multiple fragmentation channel ClÀ ‡ O ‡ Cl.
Available thermodynamical data indicates an en-
ergetic thresholdfor this channel at 0.69 eV which
is somewhat above the observedstructure ꢀsee also
discussion on the OÀ ion below).
The complementary channels ꢀ2c) andꢀ2)d
À
correspondto the expulsion of the O ion or
neutral O andthe formation of a new Cl±Cl bond

W. Sailer et al. / Chemical Physics Letters 344 62001) 471±478
477
References
in the neutral andionic compound, respectively.
The OÀ ion yieldcurve peaks at 0.15 eV while the
thermodynamic limit for reaction ꢀ2d) is at 0.32
eV. We note that in the previous unpublished
work [26], the OÀ resonance maximum was also
locatedat 0.15 eV which may be an indication
that the Cl±O andCl±OCl bonddissociation en-
ergies usedto calculate the thresholdare some-
what too high. This wouldalso place the limit for
the three particle dissociation ꢀsee above) at a
lower energy in accordance with the experimental
observation. The low intensity of the OÀ channel
not only arises from its comparatively unfavour-
able energy ꢀit is the only endothermic DA
channel), but also from the fact that in the initial
step of electron attachment to the parent mole-
cule ClOCl the additional charge is largely lo-
calizedtowarsd the Cl atoms which makes
dissociation into OÀ unlikely.
[1] R.P. Wayne, G. Poulet, P. Biggs, J.P. Burrows, R.A. Cox,
P.J. Crutzen, G.D. Hayman, M.E. Jenkin, G. Le Bras,
G.K. Moortgat, U. Platt, R.N. Schindler, Atmos. Envion.
29 ꢀ1995) 2677.
[2] J.W. Waters, L. Froidevaux, W.G. Read, G.L. Manney,
L.S. Elson, D.A. Flower, R. Jarnot, R.S. Harwood, Nature
362 ꢀ1993) 597.
[3] V. Handwerk, R. Zellner, Ber. Bunsenges. Phys. Chem. 88
ꢀ1984) 405.
[4] T. Ellermann, K. Jonsson, A. Lund, P. Pagsberg, Acta
Chem. Scand. 49 ꢀ1995) 28.
[5] J. Brunning, L.J. Stief, J. Chem. Phys. 85 ꢀ1986) 2591.
[6] S.P. Sander, R.R. Friedl, J. Phys. Chem. 93 ꢀ1989) 4794.
[7] M. Sugie, M. Ayabe, H. Takeo, C. Matsumura, J. Mol.
Spectrosc. 352/353 ꢀ1996) 259.
[8] Y. Tanaka, M. Kawasaki, Y. Matsumi, H. Fujiwara, T.
Ishiwata, L.J. Rogers, R.N. Dixon, M.N.R. Ashfold, J.
Chem. Phys. 109 ꢀ1998) 1315.

[9] G. Senn, H. Drexel, G. Marston, N.J. Mason, T. Mark, M.
Meinke, C. Schmale, P. Tegeder, E. Illenberger, J. Phys. B:
At. Mol. Opt. Phys. 32 ꢀ1999) 3615.
Finally, at higher electron energies the
ClÀ yield shows additional weak features near 3
and5.5 eV which were also observedin the
earlier work [26]. Since the ClÀ signal from Cl2
also exhibits features in that energy region
[26,35] the question remains open whether this
signal originates from Cl2 impurities or from
ClOCl.
We ®nally mention the observation of a very
weak signal due to the stoichiometric composi-
tion of Cl2OÀ near zero eV, also previously ob-
served[26]. This compoundis most likely the
result of a secondary reaction. If it has the
structure of the target molecule, ClOClÀ, it pos-
sesses a very weak binding energy ꢀFig. 3). An
alternative possibility is an ion±molecule complex
of the form O Á ClÀ2 .
[10] K. Miyazaki, M. Tanoura, K. Tanaka, T. Tanaka, J. Mol.
Spectrosc. 116 ꢀ1986) 435.
[11] H.F. Davis, Y.T. Lee, J. Chem. Phys. 105 ꢀ1996) 8142.
[12] B.T. Luke, J. Mol. Struct. 332 ꢀ1995) 283.
[13] C.L. Lin, J. Chem. Eng. Data 21 ꢀ1976) 411.
[14] L.T. Molina, M.J. Molina, J. Phys. Chem. 82 ꢀ1978) 2410.
[15] H.D. Knauth, H. Alberti, H. Klausen, J. Phys. Chem. 83
ꢀ1979) 1604.
[16] S.L. Nickolaisen, C.E. Miller, S.P. Sander, M.R. Hand,
I.H. Williams, J.S. Francisco, J. Chem. Phys. 104 ꢀ1996)
2857.
[17] F. Motte-Tollet, M.-P. Ska, G.M. Marston, I.C. Walker,
M.R.F. Siggel, J.M. Gingell, L. Kaminski, K. Brown, N.J.
Mason, Chem. Phys. Lett. 275 ꢀ1997) 298.
[18] H.J. Schumacher, R.V. Townsend, Z. Phys. Chem. Abt. B
20 ꢀ1933) 375.
[19] F.H.C. Edgecomb, R.G.W. Norrish, B.A. Thrush, Proc. R.
Soc. London, Ser. A 243 ꢀ1957) 24.
[20] C.M. Nelson, T.A. Moore, M. Okumura, T.K. Minton, J.
Chem. Phys. 100 ꢀ1994) 8055.
[21] T.A. Moore, M. Okumura, T.K. Minton, J. Chem. Phys.
107 ꢀ1997) 3237.

[22] V. Grill, H. Drexel, W. Sailer, M. Lezius, T.D. Mark, Int.
Acknowledgements
J. Mass Spectrom. 205 ꢀ2001) 209.
[23] S. Matejcik, A. Kiendler, P. Stamp¯i, A. Stamatovic, T.D.
Work partially supportedby the FWF, Wien,
the EU ꢀ5th framework program), Brussels, the
UK Engineering andPhysical Sciences Research
Council andthe Deutsche Forschungsgemeins-
chaft. One of us P.T. thanks the EU for a Marie
Curie fellowship ꢀEuropean Community pro-
gramme Human Potential under contract number
HPMF-CT-1999-00247).

Mark, Phys. Rev. Lett. 77 ꢀ1996) 3771.
[24] S. Matejcik, G. Senn, P. Scheier, A. Kiendler, A. Stama-

tovic, T.D. Mark, J. Chem. Phys. 107 ꢀ1997) 8955.
[25] D. Klar, M.-W. Ruf, H. Hotop, Aust. J. Phys. 45 ꢀ1992)
263.
[26] M. Meinke, Diplomarbeit, Fachbereich Chemie, Freie

Universitat Berlin, 1990.
[27] D. Wecker, A.A. Christodoulides, R.N. Schindler, Int. J.
Mass Spectrom. Ion Phys. 38 ꢀ1981) 391.

478
W. Sailer et al. / Chemical Physics Letters 344 62001) 471±478
[28] D.R. Lide ꢀEd.), CRC Handbook of Chemistry and
Physics, 78th edn., CRC Press, Boca Raton, 1997.
[29] G.D. Purvis, R.J. Bartlett, J. Chem. Phys. 76 ꢀ1982)
1910.
[32] D. Klar, M.-W. Ruf, H. Hotop, Int. J. Mass Spectrom. 205
ꢀ2001) 93.
[33] I. Hahndorf, E. Illenberger, L. Lehr, J. Manz, Chem. Phys.
Lett. 231 ꢀ1994) 460.
[30] D.E. Woon, T.H. Dunning Jr., J. Chem. Phys. 98 ꢀ1993)
1358.
[34] A. Schramm, I.I. Fabrikant, J.M. Weber, E. Leber, M.-W.
Ruf, H. Hotop, J. Phys. B 32 ꢀ1999) 2153.
[35] M.V. Kurepa, D.S. Belic, J. Phys. B: At. Mol. Opt. Phys.
11 ꢀ1978) 3719.
[31] K.L. Baluja, N.J. Mason, L.A. Morgan, J. Tennyson, J.
Phys. B: At. Mol. Opt. Phys. in press.