Photodestruction of NO2- using time resolved multicoincidence detection photofragment spectroscopy

Article abstract of DOI:10.1016/j.chemphys.2004.01.005

We present an experiment on the photodestruction of the NO₂ ^- anion at 266 nm. We have quantified the competition between photodetachment and photodissociation and have identified the nature of the photodissociation process from the photofragment angular distribution. This study involves a novel technique; time resolved multicoincident detection photofragment spectroscopy. A three-dimensional double exposure CCD camera is employed. The system provides the position (x,y) and the relative arrival time (t) of all fragments. In the case of photodissociation events, NO and O ^- fragments are detected for each event. The detection of multiple events per laser shot is made possible by using center-of-mass selection.

Full text of DOI:10.1016/j.chemphys.2004.01.005

Chemical Physics 300 (2004) 133–141
www.elsevier.com/locate/chemphys
ꢀ
Photodestruction of NO using time resolved
multicoincidence detection photofragment spectroscopy
2
a
Laura Dinu , Wim J. van der Zande
a,b,
*
a
FOM-Institute for Atomic and Molecular Physics, Kruislaan 407, 1098 SJ Amsterdam, The Netherlands
b
Department of Molecular and Laser Physics, University of Nijmegen, Toernooiveld 1, 6525 ED Nijmegen, The Netherlands
Received 6 October 2003; accepted 2 January 2004
Abstract
ꢀ
We present an experiment on the photodestruction of the NO2 anion at 266 nm. We have quantified the competition between
photodetachment and photodissociation and have identified the nature of the photodissociation process from the photofragment
angular distribution. This study involves a novel technique; time resolved multicoincident detection photofragment spectroscopy. A
three-dimensional double exposure CCD camera is employed. The system provides the position (x; y) and the relative arrival time (t)
ꢀ
of all fragments. In the case of photodissociation events, NO and O fragments are detected for each event. The detection of
multiple events per laser shot is made possible by using center-of-mass selection.
ꢀ
2004 Elsevier B.V. All rights reserved.
1
. Introduction
multianode employing capacitive charge division for
position determination. Such a position-sensitive ele-
ment can only handle one fragment at the time. Charge
division-based detectors (resistive anodes, wedge-and-
strip anodes, delay line anodes) have the advantage of
good temporal and spatial resolution, but they do not
allow simultaneous detection. Alternatively, multiele-
ment segmented anodes were made for multihit detec-
tion, often with a limited spatial resolution or dynamic
range [2–4]. Recently, a number of groups have used
delay line anodes together with microchannelplates.
These systems combine high spatial resolution with a
relatively short dead time for multihit detection (few
nanoseconds) [5,6]. Zajfman and co-workers have in-
troduced a hybrid system by combining the micro-
channelplate detector with a phosphor screen, such that
each arriving particle gives a short flash. These short
flashes are recorded on a high spatial resolution CCD
camera, whereas with the aid of a beam splitter, a
number of photomultiplier tubes (PMTs) also observe
the same flashes. By working close to saturation, the
PMT obtains a high time-resolution with crude position
resolution. The advantage is that the combination of
PMT and CCD camera allows multihit detection with
good time- and position-resolution and essentially no
Molecular dynamics has witnessed large progress
because of detector developments in the last decades.
The study of dissociation, ionization and detachment
processes requires special detectors, since optical spectra
become devoid of characteristic structures. A multihit
detector with the capability of sorting the fragments
resulting from multiple dissociation events per laser shot
is the ideal tool for such studies by allowing the largest
possible counting rate when using relatively low repeti-
tion rate lasers. The ideal three-dimensional (x; y; t) de-
tector measures the arrival time and the position of a
large number of particles with good time and position
resolution and no recovery times. In the last two decades
several time- and position-sensitive detectors have been
developed, which were more or less suitable for multi-
particle detection, but rarely for multievent detection.
Around 1982, de Bruijn and Los [1] presented a time-
and position-sensitive detector suitable for detection in
coincidence of two particles. Each particle is detected on
a microchannelplate (MCP) equipped with separate
*
Corresponding author.
0
301-0104/$ - see front matter ꢀ 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.chemphys.2004.01.005

134
L. Dinu, W.J. van der Zande / Chemical Physics 300 (2004) 133–141
dead time. The drawback of this setup is the limitation
in the maximum number of particles that can be de-
tected [7]. A similar approach for the time measurement
has been made by drawing a number of gold strips using
vapor deposition on the back of the MCPs to function
as independent timing devices [8].
on the photodetachment channel for this anion. When
photodestruction was studied, no distinction was made
between photodissociation and photodetachment. A
number of experiments have concentrated on the exis-
tence of possible peroxide isomers with energies close to
that of the anion ground state.
A new concept in the three-dimensional (3D) detec-
tion was introduced by Strasser et al. [9]. Inside the
vacuum, they used a combination of a MCP and a
phosphor screen. The back of the phosphor screen was
imaged onto two correlated cameras, taking two images
of the same event with a fixed time delay. The timing
determination uses the fact that the time profile of the
phosphorescence is highly reproducible. From the in-
tensity ratio of the each spot in each of the two images,
the arrival time of the particles can be retrieved. This
system allows multihit 3D detection with no dead time
and very good time and position resolution. The limi-
tation of this system is in the short dynamic range of the
timing measurement (limited by the characteristic decay
time of the phosphor screen). Instead of employing two
CCD cameras, the same principle was used in a single
double exposure CCD camera (LaVision, modified Im-
ager3 VGA). The camera retrieves the ðx; yÞ coordinates
of the spots (ꢁ80 lm resolution) and the arrival time of
the particles (ꢁ1 ns resolution) [10].
In this paper, we present two new aspects of our ex-
periment. First, we describe in detail the use of a double
exposure CCD camera in a fast beam experiment for the
coincidence detection of fragments resulting from mul-
tiple dissociation events. From the (x; y; t) data the full
3D dissociation kinematics is deduced. In the second
part, we present new results on the photodestruction of
ꢀ
NO . Here, we determine for the first time the compe-
2
tition between the two destruction channels at a photon
energy where both pathways are fully open. Moreover,
we provide kinetic energy release distribution and
photofragment anisotropy information for the photo-
dissociation channel.
2. Method
2.1. Experimental setup
ꢀ
The nitrogen dioxide anion, NO , is produced in a
2
We present the use of this detection system in a fast
hollow cathode discharge ion source filled with O2 and
N2 gas in proportions of 5:1. The discharge is main-
tained at 600 V and 20 mA. The negative species are
ꢀ
beam experiment on the photodestruction of NO . In
2
general, the photodestruction of molecular anions has
not been studied in detail because of experimental
difficulties. Photodestruction is the sum of two com-
peting processes: photodetachment and photodissocia-
tion. With sufficiently energetic photons, or in weakly
bound systems, also dissociative photodetachment is
possible [11,12]. By using the same MCP detector for the
detection of all molecular products, i.e., fragments from
photodissociation as well as neutral molecular products
from photodetachment, absolute branching ratios can
be determined. The branching between the different
processes is important under many conditions in very
dilute plasmas, such as our own atmosphere. The
products of photodetachment, free electrons and neu-
trals have different reactivity and affect the plasma in a
very different way than the photodissociation products.
ꢀ
extracted at 5 kV and the NO ions are mass selected by
a Wien filter. The continuous molecular ion beam of
2
ꢀ
NO is introduced into a fast beam setup (see Fig. 1). By
2
ꢀ
The photodestruction of NO2 plays a role as interme-
diate in the chemistry of the D-region of the ionosphere
[
13]. Nitrogen dioxide has a very large electron affinity
EA ꢁ 2:275 eV [14]. The anion and the neutral molecule
ꢀ
have similar N–O internuclear separations (1.15 A for
ꢀ
ꢀ
NO and 1.19 A for NO2) [14]. The attached electron
2
induces a small bond length contraction. The bond an-
Fig. 1. (a) Scheme of the experimental setup. A continuous molecular
ꢀ
ion beam of NO2 is crossed with a 25 Hz laser pulse (266 nm). Pho-
todissociation and photodetachment occur leading to NO2, NO and
ꢀ
1
gle for the ground state NO2 (X A1), h ¼ 117:5ꢁ is
smaller than the bond angle of the neutral ground state
h ¼ 135:1ꢁ [15]. The dissociation energy is ꢁ4 eV [19],
which is more than that of the neutral parent molecule
by about 0.8 eV. Most experiments have concentrated
ꢀ
O
fragments. The use of the beam dump is optional as well as the use
of deflection plates after the laser interaction region. The Imager CCD
camera measures the position (x; y) and the arrival time t of the frag-
ments on the detector. (b) Schematic of the dissociation process. The
direction of dissociation is indicated by ~v .

L. Dinu, W.J. van der Zande / Chemical Physics 300 (2004) 133–141
135
ꢂ
ꢃ
ð3Þ
means of four deflection plates the beam is guided
through the apertures of the apparatus. Two Einzel
lenses are used to focus the beam through the exit ap-
erture of the ion source region and to reduce the radial
size of the beam in the region of interaction with the
E
0
m
1
m
2
ꢀ
ꢁ
m
1
ꢀ m
2
2
v
0
Dt
L
2
2
0
2
KER ¼ L2
D þ v Dt
1 þ 2
m þ m
;
2
ðm
1
þ m Þ
2
1
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2
2
ꢀ
ðDyÞ þ ðv
0
DtÞ
v0Dt
laser. The NO2 anion beam is intersected at 90ꢁ by ul-
h ¼ arctan
;
U ¼ arctan Dy
;
ð4Þ
traviolet laser radiation. The laser system used for this
experiment was a 50 Hz Spectra Physics Nd-YAG laser.
The 532 nm output of the Nd-YAG laser is frequency
doubled in a BBO non-linear crystal and the energy of
the UV laser pulse, measured after the interaction region
is ꢁ10 mJ. The polarization of the laser is chosen par-
allel to the detector surface, along the y-axis (see Fig. 1).
The two processes we investigate are photodetachment
Dx
where m and m are the masses of the two fragments NO
1
q
2
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ꢀ
2
2
and O , D ¼ ðDxÞ þ ðDyÞ is the distance between
them on the detector, Dt is the arrival time difference of the
fragments, L ¼ 1:25 m is the distance between the reaction
0
region and the detector, E is the kinetic energy of the fast
beam (5 keV), and v
0
is the translational velocity of the
5
ꢀ
NO anions (ꢁ1.5 · 10 m/s). The angle h is defined with
ꢀ
2
and photodissociation of NO2
respect to the laser polarization axis (see Fig. 1(b)).
The camera provides the master clock of the experi-
ment and it controls the firing of the laser and the gating
of the MCPs. The acquisition rate is limited by the
camera to 25 Hz.
ꢀ
ꢀ
NO þ hm ! NO
2
þ e
ð1Þ
ð2Þ
2
ꢀ
!
NO þ O
The photodetachment signal consists of neutral NO2
molecules flying in the same direction as the initial
ꢀ
NO beam. The photodissociation signal, on the other
ꢀ
2.2. Multicoincident detection
2
hand, consists of NO and O fragments flying away
from the parent beam direction with velocities related
to the kinetic energy released in the dissociation pro-
cess. Other possible dissociation channels such as
This detection system providing (x; y; t) coordinates of
the particles allows detection of multiple dissociation
events. The data is taken and analyzed for each laser
shot separately. The maximum number of fragments
detected per laser shot is in principle only limited by the
finite resolution of the camera chip, 640 · 480 pixels and
the size of the spots, 3 · 3 pixels. In case of a coincident
measurement, when the fragments must be correlated as
coming from the same parent molecule, the maximum
number of particles that can be resolved depends
strongly on the experimental conditions, namely the size
of the parent molecular beam.
The correlation is performed by center-of-mass se-
lection. In the case of a homonuclear diatomic molecule,
the two fragments have identical mass and the center-of-
mass calculation is straightforward. The calculation is
performed for each combination of two particles. Only
those pairs having their center-of-mass inside a confined
region defined by the size of the parent beam are con-
sidered true coincidences. In principle, the narrower the
molecular beam is, the higher is the number of dissoci-
ation events that can be resolved. In this particular ex-
periment where the fragments have different mass (30
and 16 amu) the nature of the fragments must be iden-
tified prior to center-of-mass calculation.
ꢀ
N + O are energetically closed. After a time-of-flight
2
of ꢁ8.47 ls the fragments reach the detector consisting
of two MCPs and a P46 phosphor screen. For a good
signal-to-noise ratio, the detector is gated with respect
to the firing of the laser. The voltage on the front of
the MCP is raised to the optimum value (2 kV differ-
ence over the MCP stack) when the laser induced
fragments are expected to arrive. The detection gate is
open for 400 ns. Between laser shots, the voltage on the
front of the MCP is decreased by 800 V to a value that
does not permit particle detection.
Outside the vacuum system a so-called 3D double
exposure CCD camera (LaVision, modified Imager3
VGA) records images of the phosphor screen. The
principle of the 3D camera is described in detail else-
where [9,10]. In brief, the camera records the image of
the phosphor screen and the intensity of the spots
corresponding to individual particles. Due to the fast
interline transfer of the CCD chip, the camera is ca-
pable of taking two successive images of the same
phosphor decay for each single laser pulse. Subse-
quently, the arrival time can be found for each particle
from the ratio of intensities measured in both images
During a coincident measurement, we can detect all
photodissociation products, neutrals and ions. For each
laser shot up to 10 fragments are recorded. This number
is determined by experimental conditions (molecular
beam intensity, laser intensity). We avoided the use of
any electric field after the interaction with the laser such
that the negatively charged fragments can travel undis-
turbed towards the detector. We placed a circular beam
[
16]. The detection system retrieves the (x; y) coordi-
nates of the spots with ꢁ80 lm resolution and the
arrival time of the particles with ꢁ1 ns resolution. This
information is used to reconstruct the orientation of
the molecule at the moment of dissociation and to
calculate the kinetic energy released (KER) in the
process

136
L. Dinu, W.J. van der Zande / Chemical Physics 300 (2004) 133–141
dump (/ ¼ 1 mm) after the laser interaction region to
stop the parent beam, which otherwise would saturate
the detector in the central region. We detect only frag-
ments resulting from photodissociation processes which
have enough kinetic energy (>0.5 eV) and a sufficient
velocity component in the (x; y) plane to escape the beam
dump. The raw (x; y) data, as it is recorded by the
camera over 180,000 laser shots at a photon energy of
dump. Around the shadow of the beam dump, two main
rings can be distinguished, the outer limit being shown
by the two circles D1 (/ ¼ 300 pixels) and D2 (/ ¼ 560
pixels). The fragments detected inside the D1 circle are
mostly fragments having the high mass (NO fragments),
while those detected outside this ring must have a lighter
mass (O fragments).
We can also determine which one of the fragments
carries the charge. For this purpose, we have applied an
electric field after the laser interaction region to assure
that the only fragments that reach the detector are
neutral species. We do not use any beam dump since the
ionic parent beam is also deflected. The raw (x; y) data
integrated over 30,000 laser shots at a photon energy of
4.66 eV is shown in Fig. 2. In the middle of the detector
there are no counts due to the presence of the beam
500
400
300
200
100
0
D
1
4
.66 eV is shown in Fig. 3. The histograms of particles
coordinates show a narrow strong peak in the middle of
the image, corresponding to the NO2 photodetachment
signal. The size of the peak, / ’ 30 pixels, indicates the
size of the parent beam on the detector DCM (/ ’ 3
mm). Most of the fragments resulting from a dissocia-
tion process and having a velocity component in the
plane of the detector are confined inside the ring D1.
From the size of the ring, it is clear that neutral fragment
has a higher mass than the ionic fragment. We conclude
that the charge is taken by the oxygen fragment and we
ꢀ
D
2
identify the photodissociation products as NO and O
fragments.
0
100
200
300
400
500
600
x (pixels)
For a coincidence measurement, we analyze the im-
ages taken at each laser shot separately. Fig. 4 is a se-
lected example of an image taken during a single laser
shot. Ten particles are detected in this image. From the
(x; y) position of the fragments, we assign them as being
ꢀ
Fig. 2. Photodissociation signal of NO2 at 266 nm. The two rings D1
and D2 indicate the presence of two different masses. The ratio D2=D1
is 1.85, a value in good agreement with a mass ratio of 30/16. The
central arrow shows the direction of the laser polarization.
1
0000
100
Beam dump
cut-off
8
6
0
0
1
000
1
00
40
2
0
0
1
0
0
50 100 150 200 250 300
Radial distance (pixels)
4
00
00
00
00
0
3
2
1
0
100
200
300
400
500
600
10
100
1000 10000
x(pixels)
ꢀ
Fig. 3. Photodestruction signal of NO2 at 266 nm. Only the neutral fragments are detected. The central peak represents NO2 photodetachment signal
while the rest of the counts, distributed around the central peak are NO fragments from photodissociation. From the x and y histograms, the
branching between the two processes can be determined. The upper right plot shows the estimated radial distribution of the neutral NO calculated
with respect to the center of the parent beam.

L. Dinu, W.J. van der Zande / Chemical Physics 300 (2004) 133–141
137
ꢀ
center-of-mass position for every pair of NO and O . If
the center-of-mass lies inside the DCM ring, the pair is a
true coincidence and their center-of-mass is plotted. Out
of 10 fragments detected, eight were successfully corre-
lated, while two of them were not. The two uncorrelated
fragments are very likely to originate from a dissociation
event of which the other fragment has not been detected.
It is important to stress here that since the detection
efficiency of the MCPs is ꢁ50%, only in 25% of the cases
both fragments are detected. In two-body break-up
studies, the possibility to detect unlimited number of
particles increases the detection rate considerably.
In Fig. 5, we present the center-of-mass distributions
for two data sets taken at very different count rates, one
corresponding to a low molecular beam intensity and
another one corresponding to a high molecular beam
intensity. The data has been taken in coincidence by
allowing all fragments, ions and neutrals to reach the
detector (similar to the data presented in Fig. 2) while
blocking the central beam with the flag. From the his-
togram of number of particles detected per laser shot,
the low count rate data set (see Fig. 5(a)) is characterized
by a detection of mostly two particles per laser shot
Fig. 4. Sample of the images taken during laser shots. Ten fragments
are detected; from their position we identify the nature of each frag-
ꢀ
ꢀ
ment: NO (d) and O (j). For each pair of NO and O the center of
mass is calculated. We plot only those center of mass that lie inside the
inner ring DCM. Uncorrelated fragments are represented by (s).
ꢀ
NO, if they lie inside D1 ring, or as O fragments if they
are found in between the D1 ring and the D2 ring. Once
the mass of the fragment is know, we calculate the
2
104
.5 104
104
5000
4
3
2
1
000
000
000
000
0
1
1
5000
0
2
3
4
5
6
7
8
9
10
2
3
4
5
6
7
8
9 10
number of particles per lasershot
number of particles per lasershot
3
3
2
2
1
1
500
000
500
000
500
000
1200
1000
XCM histogram
XCM histogram
800
600
400
200
0
5
00
0
0
100 200 300 400 500 600
0
100 200 300 400 500 600
xCM(pixels)
xCM(pixels)
2
1
1
000
500
000
1200
000
00
600
YCMhistogram
YCMhistogram
1
8
4
2
00
00
0
5
00
0
0
100 200 300 400 500
yCM(pixels)
0
100 200 300 400 500
(
a)
(b)
yCM(pixels)
Fig. 5. Center-of-mass determination for two different data sets: (a) low ion beam intensity; (b) high ion beam intensity. The position histograms of
ꢀ
the center-of-mass for each combination of NO and O particles are shown. The fragments resulting from the same molecule appear in the central
peaks.

138
L. Dinu, W.J. van der Zande / Chemical Physics 300 (2004) 133–141
700
600
500
while for the high count rate data set (see Fig. 5(b)) up
to 10 particles are detected. The mass of each fragment
is identified and the center-of-mass position for all
NO
NO
ꢀ
combinations of NO and O fragments are calculated.
The results are plotted in xCM and yCM histograms. The
central peaks are a direct measure of the size of the
molecular parent beam at the detector. In Fig. 5(b),
the width of the center-of-mass peak is larger than the
one seen in Fig. 5(a). This reflects the fact the size of the
ion source exit aperture that defines the maximum size
of the parent beam was increased from 400 to 1150 lm.
The shoulders visible around 180 and 450 pixels in
Fig. 5(b) are due to false coincidences obtained by cal-
400
-
-
O
O
300
200
100
0
0
100
200
300
400
500
600
ꢀ
culating the center-of-mass for pairs of NO and O re-
x (pixels)
sulting from different parent molecules. These false
coincidences are easily avoided by center-of-mass se-
lection. The condition for a coincidence to be real is that
both xCM and yCM belong to the center-of-mass peak.
Of course, it is possible that fragments resulting from
different molecules give the correct center-of-mass po-
sition. These events cannot be differentiated from the
real events and they contribute to the background. We
estimate ꢁ5% of the pairs are accidental coincidences. In
the case of low count rate data less than 2% of the events
are accidental coincidences. It is obvious that the nar-
rower the center-of-mass peak is, the higher is the ac-
curacy in the determination of real coincidences.
In Fig. 6, we compare the kinetic energy release
spectra for the two data sets. From the low beam in-
tensity data set, we analyze only those laser shots where
two particles are detected. We compare this KER
spectrum with the one obtained from analyzing the high
beam intensity data set by selecting only the laser shots
where at least eight particles are detected. The two
spectra show no qualitative difference proving that by
allowing higher count rate, the quality of the data is
preserved. The cut-off around 500 meV is due to the flag
Fig. 7. Histograms of x-coordinates of the fragments before correction
ꢀ
(
grey curve) and after the 10 pixels correction of the O x-coordinates
black curve).
(
(r ꢁ 85 pixels) which intercepts also fragments with
small recoil velocities.
For coincidence detection, we avoided the use of any
ꢀ
electric field. Since the O fragment is negatively
charged, its trajectory is slightly perturbed by residual
fields and the EarthÕs magnetic field. The observed dis-
placement along the x-axis is approximately 10 pixels
(ꢁ1 mm on the phosphor screen) and 15 pixels along y.
ꢀ
Since most of the O fragments are isolated from the
NO fragments, it is possible to correct the coordinates of
ꢀ
the O . For illustration, in Fig. 7, we present two his-
tograms of all x-coordinates of all fragments detected
during a coincidence measurement (data set presented in
Fig. 2), before and after correction. The two inner peaks
correspond to the NO signal, while the outer peaks are
ꢀ
O
fragments. It can be seen that, the two projected
rings have their center displaced by ꢁ10 pixels. The
KER distribution is generated after the correction of the
ꢀ
(
x; y) coordinates for the O .
1.0
0.8
0.6
0.4
0.2
0.0
Beam
dump
cut
ꢀ
3
. Photodestruction of NO2
3.1. Branching between photodetachment and photodisso-
ciation
ꢀ
The photodestruction of the NO2 takes place via
photodetachment and photodissociation processes. For
measuring the branching between the two processes we
0
200
400
600
800
1000
1200
1400
1600
1800
2000
detect only the neutral NO molecules and NO frag-
2
kinetic energy release (meV)
ments as it is shown in Fig. 3. In order to have an
accurate estimate of the branching, we lowered the
molecular beam intensity such that mostly one frag-
ment is detected per laser shot (92% of the cases). This
is necessary since the photodetachment signal is con-
Fig. 6. Kinetic energy release spectrum. No significant difference can be
observed by analyzing separately laser shots with maximum two par-
ticles detected (black curve) and laser shots contributing with more
than eight particles (grey curve).

L. Dinu, W.J. van der Zande / Chemical Physics 300 (2004) 133–141
139
ꢀ
fined in a much smaller region on the detector than the
photodissociation signal. By lowering the count rate to
one fragment, we avoid overlapped photodetachment
counts that cannot be resolved. This experiment is
nearly background free and all signal detected is related
to the presence of the laser. The particles detected in
the center of the detector, confined in a circular region
KER decreases if the molecular (NO) and atomic (O )
fragments end up internally excited. In the molecular
fragment, energy can be stored in the form of rotational
and vibrational energy. As will be described below, en-
ergy stored in fragment rotation influences the photo-
fragment anisotropy. During the recoil, if the force is
ꢀ
along the NO bond direction in the bent NO geometry,
2
(
/ ¼ 30 pixels) are neutral NO from photodetachment
a significant fraction of available energy has to end up in
rotation of the NO fragment. Another reason for rota-
tional excitation is a possible large change in binding
angle in going from the molecular ground state to the
dissociative state. Vibrational energy may be added by
the recoil force or by the contraction or extension of the
2
process, while the particles distributed around it are
NO fragments from the photodissociation process. We
found that 7.5% of the excited NO2 decay via photo-
ꢀ
dissociation. This result is the most important new
physical result of this work. We note, that the photon
energy (4.66 eV) is not so much larger than the binding
energy (3.96 eV). The nature of the electronic excited
state is unknown other than that the transition occurs
along a parallel transition (see below). In view of the
limited excess energy, photoexcitation may well take
place via quasi-bound states with a certain vibrational
lifetime, which may then undergo enhanced photode-
tachment. We note that this experiment does not pro-
vide the information whether the photodetachment
signal is due to a direct photodetachment process or the
result of photoexcitation of the anion to an excited
state that undergoes autodetachment. A photoelectron
energy spectrum would be more revealing for this dis-
tinction than the present experiment, which only mea-
sures the neutral fragment yield. The photoelectron
spectrum is sensitive to the changing overlap between
the anion ground state and the anion excited state with
the different neutral product states. In the case of
photodetachment of the excited molecular state, at
higher photon energies and larger recoil velocities, the
photodissociation branching may well be significantly
larger.
ꢀ
NO-bond in going from the NO2 anion to the dissoci-
ation products.
We note that the coincidence KER spectrum of Fig. 6
lacks signal below 500 meV. As mentioned before, this is
due to the presence of the beam dump. The broad radial
distribution of the NO fragments detected in the absence
of the beam dump (see Fig. 3) suggests that most events
will escape the flag and reach the detector. Still, there are
events appearing at radial distances smaller than 85
pixels (KERpos < 500 meV), that in a coincidence de-
tection are stopped by the beam dump. These low KER
values suggest that the NO fragment can convert a big
part of the energy from the dissociation process into
internal energy. Busch and Wilson [17] developed a
model to predict the energy partitioning in the photo-
dissociation of neutral NO2. Taking into account the
ꢀ
smaller bond angle of NO , we find that 24% of the E
would appear as rotational energy of the NO fragment
2
avl
and 76% as KER.
It is more important to realize that surprisingly large
KER values are observed in the spectrum, which implies
ꢀ
contribution of rovibrationally excited NO2 anions in
the negative ion beam, indicating a non-thermal process
in the ion source. The absence of clear structure in the
KER spectrum, makes it impossible to say anything
about the internal energy distribution in the initial ion
beam. It is noteworthy that no autodetachment signal
was observed in this experiment. The autodetachment
channel would point at a part of the ion beam with a
significant internal energy of more than 2.5 eV. It has
been mentioned in the literature [18,19] that in the hol-
low cathode discharge ion source filled with O2 and N2,
vibrationally excited states may also be produced. In
these reports, an onset of the photodetachment signal
was observed at about 1.8 eV, indicating internal ener-
gies exceeding 0.5 eV. Our observation of KER values
around 1 eV are in agreement with even larger internal
energies as we expect some 25% of the excess energy at
least to end up in NO rotation.
3.2. Kinetic energy release spectrum and angular distri-
bution
In Fig. 6, we present the kinetic energy release spec-
ꢀ
trum for the photodissociation of NO induced by a 266
2
nm (4.66 eV) photon. The kinetic energy release for each
dissociation event is calculated according to Eq. (3).
Since the dissociation energy is ꢁ4 eV [19], assuming
that no energy is stored as internal energy (neither in the
parent molecule nor in the photodissociation products),
the expected KER is ꢁ0.66 eV. The experimental data
shows a large distribution of KER centered on 0.8 eV.
ꢀ
The conservation of energy during dissociation of NO ,
taking into account the internal energies reads
2
ꢀ
NO
2
Eint þ hv ꢀ D ¼ E ¼ Eint þ KER:
ð5Þ
0
avl
The product angular distribution (see Fig. 8) is de-
scribed by
Eq. (5) expresses that the final observed kinetic energy
release (KER) of the fragments increases with the in-
ternal energy of the anion and with photon energy. The
IðhÞ ꢁ ½1 þ bP ðcos hÞꢂ;
2
ð6Þ

140
L. Dinu, W.J. van der Zande / Chemical Physics 300 (2004) 133–141
5
4
3
2
1
00
00
00
00
00
0
petition between the torque related to the bending mo-
tion and the axial recoil force determines the final
anisotropy. The effect of the torque may be expressed in
an angle w, which weighs the radial kinetic energy re-
lease. The anisotropy parameter is written as
b ¼ 2P ½cosða þ wÞꢂ [20]. If a transition takes place to-
2
wards a state with a more linear geometry than the
ground state geometry, the combination of forces along
the rupturing bond and in radial direction may in fact
lead to an increase in the anisotropy parameter. In our
case our anisotropy parameter would imply a value of
w ¼ ꢀ5ꢁ. In general, however, large torques result in a
further reduction of the anisotropy parameter. In fact,
the shape of the excited state potential determines the
lifetime of the excited state and the distribution of the
forces over bending and recoil motion. This result in a
correlation between the fragment rotational energy and
the observed anisotropy parameter. As in diatomics, a
lifetime of the excited complex, that is comparable to the
rotational period of the parent anion, also reduces the
anisotropy parameter.
β =1.4
0
20
40
60
80
100
120
140
160
θ (degree)
Fig. 8. The angular distribution of the fragments in the detector plane
with respect to the laser polarization (see Eq. (4)). Fitting the histo-
gram (thick curve) the anisotropy parameter b ¼ 1:4 is obtained.
In our experiments, we find that the kinetic energy
release spectrum contains dominantly values from 500
meV to more than 1 eV. The magnitude of the observed
KER values implies the necessity of considerable exci-
tation energy in the parent anion. It is unlikely that most
of the available energy ends up in rotational energy of
the NO fragment. The magnitude of the anisotropy
parameter indicates a short lifetime of the excited state.
Finally, the value of b ¼ 1:4 suggests that the excited
anion state may have a more linear geometry than its
parent anion state. It is of interest to note that the
possible presence of an autodetachment channel in the
excited state may affect the anisotropy parameter. For
example, long-lived anionic excited states, which would
reduce the anisotropy parameter, may more prone to the
autodetachment process and do not contribute to the
photofragment channel.
where h is angle defined by the orientation of the mol-
ecule during dissociation with respect to the laser po-
larization and
P
2
ðcos hÞ is the second Legendre
polynomial. The histogram of the angular orientations
of the fragments shows a large anisotropy parameter
b ¼ 1:4 ꢃ 0:1. Clearly, the electronic transition is a
parallel transition. The predictive value of the anisot-
ropy parameter in triatomics is smaller than in the case
of diatomic molecules. For diatomics, a parallel transi-
tion yields an anisotropy parameter of b ¼ 2. If the
excited state has a finite lifetime larger than the rota-
tional period, the anisotropy parameter reduces to
b ¼ 1=2. In triatomics, other phenomena influence the
resulting anisotropy parameter. For molecules with a
ꢀ
bent geometry, such as NO , the dipole moment is not
2
parallel to the bond that breaks. In the case of axial
recoil, implying a repulsive force along the bond direc-
tion, the maximum anisotropy parameter is reduced to
The anion of nitrogen dioxide is a species that is
important in the upper atmosphere. This molecular
anion is more stable than its parent neutral, an unusual
situation. The excess electron is captured in a bonding
orbital. In view of its stability, photodestruction will be
an important channel for removal of these species in the
ionosphere. To the best of our knowledge the photo-
b ꢁ 2P ðcos aÞ;
ð7Þ
2
with a the angle between the bond axis and the transi-
tion dipole moment. For a bond angle of 135ꢁ as in the
2
case of neutral NO , this yields an anisotropy parameter
of 1.54, close to the value reported in [21]. We note that
this does not imply that the NO fragment is without
rotational excitation. In fact, axial recoil, as defined
above, results in rotational excitation of the fragment
depending on the molecular geometry. For the present
ꢀ
destruction of the NO2 has not been determined in
much detail. This work provides the branching of the
two pathways (7.5% photodissociation versus 92.5%
photodetachment). As important, this research shows
that coincident detection of molecular fragments from
photodissociation in order to obtain high-resolution
KER information is possible also when using relatively
low repetition rate lasers. The combination of a detector
that can determine the arrival position and arrival time
of many particles and a setup with well-defined beam
properties ensures this possibility.
ꢀ
case with an equilibrium angle of the NO of 117.5ꢁ, an
2
anisotropy parameter results of, b ¼ 1:2, which is in fact
smaller than the measured value in this work. In the case
that the transition occurs between states with different
equilibrium geometries, in particular the equilibrium
angle, the excited state molecule will be formed with a
considerable amount of bending excitation. The com-

L. Dinu, W.J. van der Zande / Chemical Physics 300 (2004) 133–141
141
Acknowledgements
C.L. Cocke, H. Schmidt-Bocking, Nucl. Instrum. 149 (1999)
90.
4
[
[
7] Z. Amitay, D. Zajfman, Rev. Sci. Instrum. 68 (1997) 1387.
8] S. Ros ꢁe n, R. Peverall, J. ter Horst, G. Sundstr o€ m, J. Semaniak, O.
Sundqvist, M. Larsson, M. de Wilde, W.J. van der Zande,
Hyperfine Interact. 115 (1998) 201.
The work described in this paper is part of the re-
search program of the Foundation for Fundamental
Research on Matter (Stichting voor Fundamenteel
Onderzoek der Materie) and it was financially supported
by the Netherlands Organization for Scientific Research
[9] D. Strasser, X. Urbain, H.B. Pedersen, N. Altstein, O. Heber, R.
Wester, K.G. Bhushan, D. Zajfman, Rev. Sci. Instrum. 71 (2000)
3
092.
(
Nederlandse Organisatie voor Wetenschappelijk Ond-
[
10] L. Dinu, A.T.J.B. Eppink, F. Rosca-Pruna, H.L. Offerhaus, W.J.
van der Zande, M.J.J. Vrakking, Rev. Sci. Instrum. 73 (2002)
4206.
erzoek) through the FOM-program Molecular Atmo-
spheric Processes. We thank M. Vrakking for critical
reading of the manuscript.
[11] K.A. Hanold, A.K. Luong, R.E. Continetti, J. Chem. Phys. 109
1998) 9215.
(
[
[
[
12] R. Li, R.E. Continetti, J. Phys. Chem. A 106 (2002) 1183.
13] L. Thomas, Radio Sci. 9 (1974) 12.
14] S.B. Woo, E.M. Helmy, P.H. Mauk, A.P. Paszek, Phys. Rev. A 24
References
(
1981) 1380.
[
[
1] D.P. de Bruijn, J. Los, Rev. Sci. Instrum. 53 (1982) 1020.
2] J. Becker, K. Beckord, U. Werner, H.O. Lutz, Nucl. Instrum.
Methods A 337 (1994) 409.
[15] C. Meredith, R.D. Davy, G.E. Quelch, H.F. Schaefer, J. Chem.
Phys. 94 (1991) 1317.
[16] R. Bobbenkamp, Ph.D. Thesis, Bielefeld, 2002.
[17] G.E. Busch, K.R. Wilson, J. Chem. Phys. 56 (1972) 3626.
[18] E. Herbst, T.A. Patterson, W.C. Lineberger, J. Chem. Phys. 61
(1974) 1300.
[
3] K. Beckord, J. Becker, U. Werner, H.O. Lutz, J. Phys. B 27 (1994)
L585.
[
4] M. Lavoll ꢁe e, Rev. Sci. Intrum. 70 (1999) 2968.
5] M. Braun, M. Beckert, U. M u€ ller, Rev. Sci. Instrum. 71 (2000)
[
[19] B.A. Huber, P.C. Cosby, J.R. Peterson, J.T. Moseley, J. Chem.
Phys. 66 (1977) 4520.
4535.
[
6] I. Ali, R. Dorner, O. Jagutzki, S. Nuttgens, V. Mergel, L.
Spielberger, K. Khayyat, T. Vogt, H. Brauning, K. Ullmann,
R. Moshammer, J. Ullrich, S. Hagmann, K.O. Groeneveld,
[20] G.E. Busch, K.R. Wilson, J. Chem. Phys. 56 (1972) 3638.
[21] A.V. Demyanenko, V. Dribinski, H. Reisler, M. Meyer, C.X.W.
Qian, J. Chem. Phys. 111 (1999) 7383.