On the initial steps in the decomposition of energetic materials from excited electronic states

Article abstract of DOI:10.1063/1.1315609

To study optical spectroscopy for nonvolatile and fragile large molecules, matrix assisted laser desorption (MALD) has been successfully combined with a molecular beam supersonic expansion technique. Sample preparation methods and proper experimental parameters are established to obtain constant sample introduction for more than 6 h. LD laser beam intensity does not disturb the molecular beam dynamics in the nozzle. This setup is applied to investigate the photodissociation of an explosive, RDX.

Full text of DOI:10.1063/1.1315609

On the initial steps in the decomposition of energetic materials from excited electronic
states
H.-S. Im and E. R. Bernstein
Citation: The Journal of Chemical Physics 113, 7911 (2000); doi: 10.1063/1.1315609
View online: http://dx.doi.org/10.1063/1.1315609
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/113/18?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Vibrational and electronic excitations in fluorinated ethene cations from the ground up
J. Chem. Phys. 138, 124301 (2013); 10.1063/1.4795428
Decomposition of excited electronic state s-tetrazine and its energetic derivatives
J. Chem. Phys. 134, 024318 (2011); 10.1063/1.3523649
A new method for determining absorption cross sections out of initially excited vibrational states
J. Chem. Phys. 130, 054303 (2009); 10.1063/1.3067923
Photodissociation of Mg(CH 2 =CHCN) n + : Excited electronic states of n=1 and 2 and intracluster electron
transfer for n=3 and 4
J. Chem. Phys. 118, 5456 (2003); 10.1063/1.1541020
Investigations into the fragmentation and ionization of highly excited La @ C 82
J. Chem. Phys. 117, 9811 (2002); 10.1063/1.1513465
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.198.228.112 On: Sun, 04 May 2014 07:58:26

JOURNAL OF CHEMICAL PHYSICS
VOLUME 113, NUMBER 18
8 NOVEMBER 2000
On the initial steps in the decomposition of energetic materials
from excited electronic states
H.-S. Im and E. R. Bernstein
Department of Chemistry, Colorado State University, Fort Collins, Colorado 80523-1872
͑
Received 19 June 2000; accepted 16 August 2000͒
Decomposition studies of hexahydro-1,3,5-trinitro-1,3,5-triazine (RDX-C H N O , see Fig. 1͒
3
6
6
6
isolated in the gas phase and cooled in a supersonic expansion are reported for the excited electronic
state near 225 nm. The RDX is handled safely and effectively through matrix-assisted laser
desorption ͑MALD͒ of a thin film of RDX/R6G laser dye ͑1:1͒ adsorbed on an aluminum oxide
coating on an aluminum drum. The aluminum oxide coating is generated by plasma electrolytic
oxidation of aluminum. The combination of MALD and supersonic molecular beam techniques
generates intact and cold RDX molecules isolated in the gas phase. Two basic conclusions are
reached in this study. ͑1͒ Photodissociation of RDX at Ϸ225 nm generates NO as an initial product.
͑
2͒ Nascent NO thus generated is vibrationally hot (Tvibϳ1800 K͒ and rotationally cold (Trot
ϳ20 K͒. © 2000 American Institute of Physics. ͓S0021-9606͑00͒01142-9͔
I. INTRODUCTION
ated with a shock wave or a spark into a solid sample. In
both cases, very large electric fields (Ϸ10 V/cm͒ are gen-
8
Hexahydro-1,3,5-trinitro-1,3,5-triazine
(RDX-
erated as crystal planes sheer or as the spark propagates.
Such events in solids generate molecules in highly excited
electronic states. Clearly, the decomposition of solid ener-
getic materials under shock, spark, laser, or plasma ignition
must include contributions from both ground and excited
electronic state species. Excitation in the UV can markedly
reduce the power requirements for detonation of some sec-
ondary explosives. Therefore, elucidation of the initial steps
of RDX decomposition from excited electronic states is an
important goal to pursue.
C H N O , see Fig. 1͒ is an energetic material with broad
3
6
6
6
applications as an explosive and a fuel. An energetic material
is any compound with a C:H:N:O ratio close to 1:1:1:1. This
classification includes both cyclic and noncyclic species such
as HMX (C H N O ), TNT (C H N O ), TNAZ
4
8
8
8
7
5
3
6
(
(
C H N O ), ADN (NH N͑NO ) ), NA ͑H NNO ), DMNA
3 4 3 6 4 2 2 2 2
1
(CH ) NNO ), and many others. Past studies of energetic
3
2
2
material decomposition reactions have focused mainly on the
ground electronic state in the condensed phase with the ini-
tiation stimulus being such events as thermal pulses, shocks,
sparks, and various laser pulses. Experimental, theoretical,
A number of experimental and theoretical studies have
been conducted to investigate the physicochemical properties
2
and simulation studies have all been pursued. Elucidation of
3
of RDX. To date, the majority of decomposition studies of
the detailed fundamental steps in the initiation and propaga-
tion phases of energetic material decomposition reactions is
central for understanding, controlling, and enhancing the per-
formance of these materials as fuels and explosives. The ef-
ficient use of these materials for combustion and explosion
also presents a number of environmental issues. Addition-
ally, such an in-depth and molecular understanding of the
chemistry of energetic materials should lead to suggestions
concerning their performance enhancement and even new
candidate molecules for synthesis.
energetic materials are for the ground electronic state. In
addition, a variety of prototype, model energetic molecules
has been examined to explore the kinetics and reaction
mechanisms of the decomposition processes.4 Different
mechanisms have been suggested for the decomposition of
RDX, depending on the initiation processes, system phase,
and the physical condition of RDX. Nitramine ͑N–N͒ bond
5
6
rupture, the concerted decomposition into CH NNO , in-
2
2
5
͑a͒
ternal ring formation and HONO generation,
heterogeneous mechanisms have all been proposed.
Most of the RDX decomposition studies in condensed
and other
7
Unimolecular fragmentation and pathways and energy
partitioning amongst product species and degrees of freedom
depend sensitively on the state of the reactant molecule.
Even for similar systems such as nitro-containing aromatics,
alkanes, nitramines, saturated heterocyclics, etc., initial prod-
ucts of the dissociation process can very considerably; for
example, the first step in dissociation can yield HONO, NO2,
NO, NNO , CH NNO , . . . , depending on reactant energy
phase have suggested that C–N bond rupture is its primary
8
reaction step to give NO, N O, H , CO, CH O, and CO .
2
2
2
2
One study suggests that C–H bond breakage is the rate-
9
determining step for RDX decomposition. An x-ray photo-
electron spectroscopy study of shock-induced decomposition
of powered RDX shows that N–N bond scission occurs in
such samples.10 When decomposition is studied at higher
temperatures producing gas phase molecules, two other
mechanisms predominate. The first pathway involves an NO2
2
2
2
content, partitioning ͑electronic, vibrational, rotation͒, and
state. These issues become particularly compelling for the
rapid decomposition of highly energetic molecules such as
RDX.
stripping mechanism in which all three NO groups are re-
2
The decomposition of energetic materials can be initi-
moved before the triazine fragment decomposes to form
0021-9606/2000/113(18)/7911/8/$17.00
7911
© 2000 American Institute of Physics
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
4.198.228.112 On: Sun, 04 May 2014 07:58:26
8

7912
J. Chem. Phys., Vol. 113, No. 18, 8 November 2000
H.-S. Im and E. R. Bernstein
bined with molecular beam techniques. The sample prepara-
tion method and the proper experimental parameters are es-
tablished to obtain constant and stable sample introduction
for more than
(
6
h. The desorption laser beam
ϳ106 W/cm ) does not disturb the molecular beam dynam-
2
ics from the nozzle. This setup is applied to the dissociation
study of an explosive, RDX. In our studies thus far, we fol-
low the initial NO product of RDX dissociation from excited
state potential energy surfaces and characterize its rotational,
vibrational, and electronic states. The basic conclusions we
reach are that ͑1͒ product NO is vibrationally hot ͑ϳ1800 K͒
but is rotationally cold ͑ϳ20 K͒ in the dissociation process at
Ϫ1
Ϸ44 200 cm of RDX excitation energy, and ͑2͒ intact and
cold RDX can be obtained by the combination of matrix-
assisted laser desorption ͑MALD͒ and molecular beam tech-
niques. By comparing the NO signals for photofragmented
model systems, CH NO , C H NO , C H (NO ) , NO ,
FIG. 1. Molecular structures of RDX and other energetic materials.
3
2
6
5
2
6
4
2 2
2
and HONO, the dynamics of RDX in excited electronic
states can be investigated to determine that NO is fragmented
directly from the photolysis of RDX using UV photons
HCN and, presumably, H .5 The second pathway, which
͑b͒
2
clearly has been identified as a concerted reaction in molecu-
Ϫ1
6
above 38 650 cm in energy.
lar beam studies, is a symmetric triple fission to form
CH NNO fragments that then further decompose to form
2
2
CH O and N O. From the study of 248 nm photodecompo-
2
2
II. EXPERIMENT
sition of RDX, another interesting initial step has been pro-
posed through the observation of OH emission.5 The most
likely mechanism for OH production is unimolecular decom-
position of RDX via a five-membered ring formation
(a)
In order to obtain a supersonic molecular beam of non-
volatile, thermally unstable solid samples like RDX, a laser
desorption ͑LD͒ approach is adapted and combined with the
supersonic jet technique. The desorbed molecules must be
efficiently entrained in the high-pressure throat region of a
pulsed supersonic expansion to produce sufficient jet cooling
to give well-resolved rovibronic spectra. In our experiment,
this is accomplished using a desorption nozzle assembly
similar to that described previously by Smalley.13 In this
assembly a LD extension block is attached to a pulsed super-
sonic nozzle ͑Jordan Company, PSV͒.
The desorbed molecules are released and entrained into
supersonic expansion within the extension block. The exten-
sion block consists of a stepper motor and gears, a shaft for
a sample holder, a sample holder, a flow channel, a laser
beam inlet hole, and a laser beam alignment hole. At the
center of the extension block, a flow channel of 2 mm diam
and about 6 cm long is aligned with the nozzle orifice ͑2 mm
diam͒. This flow channel does not change the supersonic
nozzle performance significantly, although the final expan-
sion pressure is reduced by a rough factor of 5–10. The
rotational distribution of small molecules is thus not as nar-
row as would be expected from an unmodified nozzle; the
rotational temperature of NO gas increases to Ϸ11 K from
Ϸ2 K with the addition of the laser ablation/desorption ex-
tension block to the normal supersonic nozzle. For large
molecules, this effect is negligible.
͑CHONN͒. The immediate precursor of OH in this mecha-
nism is probably the primary fragment HONO. Depending
on the experimental conditions, eventually all the above pos-
sible six pathways can be suggested as an initial step for the
decomposition of RDX. Since these studies have been con-
ducted with solid RDX or highly heated RDX ͑mostly above
200 °C͒, the behavior of single RDX molecules is difficult to
extract. Solid state chemical reactions or thermal decompo-
sition reactions could control and mask the behavior of
single RDX molecules.
Through studies of intact, isolated RDX molecules, one
can access such fundamental information as unimolecular
decomposition pathways, decomposition rates, and energy
distributions of the initial species in the reaction/
fragmentation processes. This information will serve as a
reference point from which to consider and analyze con-
densed phase behavior and various applications under stan-
dard ignition conditions. In order to study the decomposition
of an individual RDX molecule from excited electronic
states, intact RDX molecules should be generated and en-
trained in a supersonic molecular beam. Under these condi-
tions, one can remove the effect of intermolecular interac-
tions on decomposition. Matrix-assisted laser desorption
ionization ͑MALDI͒ has proved to be a viable technique to
obtain intact molecular ions of medium size molecules in the
The ablation laser beam inlet channel of 1 mm diam is
perpendicular to the gas/sample flow channel. The sample
holder is made of Al and has a right circular cylindrical
shape. The dimensions of the Al drum are 38.79 mm and
27.03 mm for the outer diameter and the inner diameters,
respectively, with a 20 mm height. The sample drum is in-
stalled in the extension block on a shaft that can be simulta-
neously rotated and translated through the combination of a
1
1
gas phase. In addition, intact neutral molecules can be pro-
duced with this technique for nonvolatile and fragile large
molecules ͑MALD process͒. The combination of MALD and
supersonic molecular beam techniques enables optical spec-
troscopy studies for individual molecules that cannot be in-
vestigated otherwise.12
In the present work, MALD has been successfully com-
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.198.228.112 On: Sun, 04 May 2014 07:58:26

J. Chem. Phys., Vol. 113, No. 18, 8 November 2000
Decomposition of RDX
7913
stepper motor and gears. The whole sample holder setup is
designed for the sample surface to be tangential to the gas
flow channel.
The use of laser desorption as a source of molecules for
laser spectroscopy requires the production of a stable supply
of sample vapor pulses for extended periods of time. This
condition can be achieved by preparing a porous and smooth
surface for even deposition of a sample. In our experiments,
such a surface is obtained using an aluminum oxide thin film
coated on an aluminum drum. The aluminum oxide thin film
is generated by plasma electrolytic oxidation.14 An Al drum
as an anode is immersed into the electrolytic solution, and dc
voltage of 120–150 V is applied to maintain a current of
1.0–1.5 A. The electrolytic solution is composed of NaF,
Na B O , (NH ) HPO , and NaH PO with the concentra-
2
4
7
4 2
4
2
4
tion of 0.6, 0.1, 0.1, and 0.5 M, respectively. Under these
conditions, a plasma is formed on the surface of the Al drum
and oxidizes it. A ␥-Al O film about 10 ␮m thick can be
FIG. 2. NO distribution in the molecular beam of our MALD supersonic
nozzle with and without LD laser.
2
3
produced on the Al drum surface within 10 min. The coated
surface is smooth with a grain size of less than 10 ␮m.
ecules. In this experiment, the probe laser is applied to in-
duce photodissociation of sample molecules and to ionize
photodissociated NO via its A(vЈϭ0,1)� X(vЉϭ0–3) and
Ϫ5
Equimolar quantities (4ϫ10 mol͒ of the sample and a
matrix substance are dissolved in an appropriate solvent,
mainly acetone in this experiment. The laser dye R6G is
chosen as a matrix substance because of its absorption coef-
ficient at 532 nm. A mixture of RDX or 1,4-dinitobenzene
and R6G at 1:1 molar concentration is sprayed on the pre-
pared drum. An air atomizing spray nozzle ͑Spraying System
Co.͒ with siphon pressure of 20 psig is used to deposit the
RDX/R6G sample on the drum surface. During the spraying,
the drum is heated with a halogen lamp to help solvent dry
and rotated with a clock motor for homogeneous coating.
The prepared sample is placed in the extension block
and irradiated with a 532 nm laser beam focused with a 1000
mm focal length lens. The desorption spot is approximately
ϩ
I� A transitions. The NO is then detected, mass selec-
tively, in the TOFMS. If necessary, the laser beam is focused
by a lens of 360 mm focal length. In order to obtain the
proper UV photons for this process, a dye laser output is
doubled and then mixed with the fundamental output of a
Nd:YAG laser. The frequency doubled output ͑532 nm͒ of
the fundamental ͑1.064 mm͒ Nd:YAG laser ͑Spectra Phys-
ics, GCR-3͒ is used for pumping the dye laser. The four dyes
required to cover the NO A� X transitions are a mixture of
R610 and R590 for ͑0–0͒ and ͑1–1͒, R640 for ͑0–1͒, DCM
for ͑0–2͒, and LDs 689 for ͑1–3͒ and ͑0–3͒. Energy of the
UV laser output is 0.2–0.5 mJ/pulse depending on the exact
wavelength of interest. This implies that the laser intensity
becomes 1.3–2.1ϫ108 W/cm as the laser beam is focused
with a lens of 360 mm focal length.
2
.75 cm downstream from the nozzle orifice. The laser beam
radiant energy right after the lens is set to be about 300
J/pulse by putting an orifice of 1–3 mm in front of the lens.
2
␮
The laser intensity is determined by dividing the laser beam
radiant energy measured with an Ophir PE10-SH pyroelec-
tric probe by the laser beam pulse width ͑ϳ7 ns͒ and the
laser beam area at the sample holder surface. The size of the
laser beam spot on the sample holder surface is adjusted to
be 1.0 mm diam at the sample holder surface. This yields the
laser intensity of about 5ϫ106 W/cm . As depicted in Fig.
Timing of the pulsed valve and the probe laser relative to
the desorption laser is controlled through time delay genera-
tors ͑SRS, DG535͒. The NO ions detected by the micro-
channel plate ͑MCP͒ in TOFMS are recorded on a PC
through an ADC card ͑Analog Devices, RTI-800͒ and a box-
car averager ͑SRS, SR250͒. For velocity distribution mea-
surements, the chamber working pressure and pulsed valve
timing are kept constant and unchanged as the timing of the
probe laser is swept. The detailed experimental scheme is
depicted in Fig. 3.
ϩ
2
2, NO gas distribution in the supersonic beam with the LD
laser on is almost same as that with the LD laser off. The
pulse shape and width of 50 ␮s FWHM is typical of this kind
of commercial pulsed valve.
In order to estimate the vibrational temperature of pho-
tofragmented NO from RDX, a Boltzmann distribution cal-
Desorbed, undissociated molecular vapor is entrained in
a supersonic molecular beam of 10% Ar/He. By rotating and
translating the sample drum, fresh sample can be supplied
continuously for each laser shot. With the given size drum,
fresh sample can be supplied for more than 6 h continuously.
About 3.25 cm downstream from the desorption spot the
desorbed molecules entrained in the carrier gas flow into a
vacuum chamber and are carried by ensuing molecular beam
through a 2 mm skimmer into the ion source of a Willey–
McLaren-type time-of-flight mass spectrometer ͑TOFMS͒.
Another laser is used to probe the desorbed sample mol-
culation is applied using the ratios of each vibronic transition
0
intensity to that of the 0 transition. To estimate the rota-
0
0
tional temperature of NO, the rotation spectrum of the 00
transition is simulated with the estimated temperature using
diatomic molecular rotational level calculation software de-
veloped at Sandia National Laboratory.15
RDX is received from Lawrence Livermore Laboratory
and used without any further treatment. R6G dye is pur-
chased from Eastman Kodak Co. and is used as received. For
one experiment, 10 mg of RDX and 25 mg of R6G are
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.198.228.112 On: Sun, 04 May 2014 07:58:26

7914
J. Chem. Phys., Vol. 113, No. 18, 8 November 2000
H.-S. Im and E. R. Bernstein
FIG. 3. Schematic drawing of experimental setup for the present work ͑DG
ϭdelay generator͒.
dissolved in acetone and sprayed. Other samples are exam-
ined to determine and prove the source of NO. These
samples are NO gas, NO gas, nitromethane ͑NM͒, nitroben-
2
zene ͑NBZ͒, and 1,4-dinitrobenzene ͑DNBZ͒. All are pur-
chased from Aldrich and used without any further treatment.
For the NO gas experiment, a significant amount of oxygen
2
gas ͑10%͒ is included in the sample vessel in order to sup-
press the formation of NO as a contaminant. The mixture of
NO ͑0.5%͒ and O ͑10%͒ is prepared in He gas premixed
FIG. 4. NO spectra A (vЈϭ0)� X (vЉϭ0,1,2,3). Each transition and its
relative intensity to the NO ͑0–0͒ transition are indicated in the figure. NO
is generated from the photodissociation of RDX.
2
2
with 10% Ar. This mixture is delivered through PVC tubing
to the pulsed valve and then expanded into the vacuum
chamber with a total pressure of 100 psig through the above
described nozzle with and without the extension block for
MALD. In order to generate enough vapor pressure, NBZ is
heated to 40 °C. The laser desorption technique is also used
for hexamethylenetetramine ͑HMT͒, tyramine, and DNBZ to
test the process. For all the samples, He gas is premixed with
ϭ0):(vЉϭ1):(vЉϭ2):(vЉϭ3)ϭ1:0.22:0.05:0.02. These values
are quite different from a Boltzmann population distribution
of Tϭ300 K and below. These observed ratios can be well
fitted to a Boltzmann distribution with a temperature of about
1800 K. This fitting is represented in Fig. 7. For NO gas
itself, the transitions from vЉϭ2,3 cannot be observed. In
addition, the molecular beam distribution of NO from the
laser desorbed RDX is different from that of NO gas. As
displayed in Fig. 8, only a distribution width of 10 ␮s
FWHM can be produced for NO from the laser desorbed
RDX. A 50 ␮s FWHM distribution is detected for NO gas,
as is typically for this kind of commercial nozzle. These
experimental results clearly indicated that the detected NO
molecules are not from NO contamination, but are related to
the generation/photolysis of RDX.
Since the laser desorption process is employed in the
present experiment, however, the above results do not distin-
guish, in principle, whether the photolysis of RDX occurs at
the ionization region of TOFMS or at the LD sampling re-
gion within the extension block. If fragmented NO is pro-
duced in the latter instance, it would be entrained in the
backing gas and expanded to the ionization region. The NO
spectrum obtained through this process should be nearly the
same as that obtained from NO gas from a supersonic nozzle.
If the photolysis of RDX would happen at the ionization
region of TOFMS, on the other hand, it would provide dif-
ferent NO molecules from these provided by NO gas mo-
lecular supersonic jet expansion. In this case, the spectrum of
1
1
0% Ar and used as the expansion gas at a total pressure of
00 psig.
III. RESULTS AND DISCUSSION
Spectra of NO excited from X(vЉϭ0,1,2,3) to A(vЈ
ϭ0) are obtained from laser desorbed RDX as presented in
Fig. 4. All the spectra show almost the same pattern except
for intensity; each spectrum shows one strong peak, which
can be assigned as the (Q ϩP ) band for each vibrational
1
1
21
1
6
transition, and other lower intensity rotational peaks that
Ϫ1
are spread within 100 cm about the (Q ϩP ) band. Such
1
1
21
a pattern can also be observed for the other transitions, like
the ͑1–1͒ and ͑1–3͒ transitions, for example, as shown in
Fig. 5. From the simulation of the NO ͑0–0͒ rotational spec-
trum, the rotational temperature of NO from RDX can be
estimated as T ϭϳ20 K ͑see Fig. 6͒. NO gas, under the
rot
same expansion conditions, is found to have a rotational tem-
perature of ϳ11 K.
A strongly wavelength-dependent signature is obtained
for the NO A� X vibronic transitions due to the populations
of vЉϭ0,1,2,3. In an attempt to obtain A� X ͑0–4͒ spec-
trum, no significant signal is observed. The ratio of the in-
tensities of the (Q ϩP ) band of each vibration are (vЉ
1
1
21
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.198.228.112 On: Sun, 04 May 2014 07:58:26

J. Chem. Phys., Vol. 113, No. 18, 8 November 2000
Decomposition of RDX
7915
FIG. 7. Boltzmann distribution calculation for determining the vibrational
temperature of fragmented NO from photolysis of RDX. Ratios of the
(
Q11ϩP21) band intensity of each vibrational transition to that of ͑0–0͒ are
used to determine Tvibϳ1800 K.
generated from the photolysis of RDX at the ionization re-
gion. The logic behind this conclusion is as follows: ͑1͒ NO
from RDX MALD has Tvibϳ1800 K, so the vibrational cool-
ing of NO is not very good if the NO comes from the nozzle;
͑2͒ the rotational cooling of NO from RDX MALD is very
good as Trot ͑0–1͒ϳ20 K; ͑3͒ but nozzle cooling of gaseous
NO gives Trot ͑0–1͒ϳ100 K and Tvibϳ200 K; and ͑4͒ there-
fore, assuming NO from RDX MALD comes from the
nozzle and not the TOFMS ion source presents a logical
inconsistency with hot vibrational levels and cold rotational
levels from the same nozzle. Apparently, rotational and vi-
brational cooling of gas phase NO expanded from our nozzle
are comparable for the higher vibrational level X (vЉϭ1),
so the only way that NO from RDX MALD could have
Tvibϳ100 Trot is if this temperature disparity exists in the
dissociation process. The following velocity slip experiment
confirms the present interpretation and proves that NO is
generated at the TOFMS ionization region.
FIG. 5. NO/RDX spectra of the ͑1–1͒ and ͑1–3͒ transitions. These spectra
look very similar to the other transitions obtained from nascent NO derived
from RDX photodissociation ͑see Fig. 4͒.
fragmented NO should be substantially different from that of
NO gas from a supersonic nozzle. In Fig. 9, the spectra at the
͑0–1͒ transition are clearly not the same; the rotational tem-
perature of NO from the desorption of RDX is much lower
than that found for NO gas from a supersonic expansion
nozzle. This result implies that the fragmented NO is
FIG. 6. Simulation of the NO rotational spectrum at the ͑0–0͒ transition
obtained by RDX photolysis. The dotted line is the simulated spectrum fit to
the experimental spectrum ͑solid line͒. The rotational temperature is deter-
mined to be 20 K.
FIG. 8. A comparison between the detected distributions of NO TOFMS
intensity from two different sources: ͑1͒ Gas phase NO expanded through a
supersonic nozzle ͑triangles͒; and ͑2͒ gas phase NO derived from MALD-
generated RDX fragmented at the TOFMS ion source region ͑squares͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
4.198.228.112 On: Sun, 04 May 2014 07:58:26
8

7916
J. Chem. Phys., Vol. 113, No. 18, 8 November 2000
H.-S. Im and E. R. Bernstein
FIG. 9. NO spectra of the ͑0–1͒ transition from NO fragmented from RDX
photolysis ͑bottom͒ and from NO gas itself ͑top͒. NO spectrum from RDX
fragmented NO displays a much colder rotational temperature than NO gas
itself.
FIG. 10. Velocity distribution differences among laser desorbed HMT,
RDX, and R6G. Equimolar quantities of HMT and RDX samples are coated
together with the matrix molecule, R6G. Signals from the parent molecule
for HMT, NO for RDX, and C for R6G have been used to determine the
3
velocity distribution data.
In a molecular beam, the velocity distribution of a
sample molecule is known to be dependent upon its molecu-
lar mass; light molecules travel faster than heavy ones. This
timing shift of the velocity distribution ͑the velocity slip͒ in
the molecular beam has been used to match photofragmented
species to their parent molecules.17 The velocity distribution
of the photofragmented NO from laser desorbed RDX shows
a different arrival time compared to that of NO gas itself or
other vapor samples examined. The velocity distributions of
gas phase molecules are almost matched to the gas pulse
shape of the commercial pulsed valve, the FWHM of the
distribution is about 50 ␮s; however, that of the photofrag-
mented NO is about 10 ␮s as seen in Fig. 8. This value is
exactly the same as the velocity distribution of laser des-
orbed hexamethylenetetramine ͑HMT͒. Also, the photofrag-
its excited electronic states to be investigated.
Even though the above results determine that the source
of NO is the photolysis of isolated, gas phase RDX, they do
not reveal the initial reaction pathway that produces NO.
When the photolysis of RDX occurs, it could produce sev-
eral possible fragmented products as the initial step in the
dissociation to generate NO. Since these processes could
happen within several hundred femtoseconds, one cannot
distinguish each step, if a multistep process occurs, using a
ns laser. From previous studies on the dissociation of RDX
by thermal or laser photolysis,5
–10
one can choose several
species as possible candidates for a primary product; these
include NO , HONO, or CH NNO . In order to investigate
2
2
2
the initial step of the photodissociation of RDX, several
mented C from R6G, which is the matrix compound used,
model systems, such as NO , NO, NBZ, and DNBZ, are
3
2
shows exactly the same velocity distribution; the FWHM is
about 10 ␮s. Even though the pattern of each velocity distri-
bution looks the same, a velocity slip is observed. The maxi-
mum intensity position of HMT ͑mass 146͒ is about 2 ␮s
earlier than that of RDX ͑mass 222͒, while that of R6G
examined. NBZ and DNBZ are chosen for a potential inter-
mediate source of HONO; previous studies1 show that
HONO can be an intermediate for NO production through
five-membered ring formation in these molecules. The NO
spectra obtained from the photolysis of these model systems
can be compared to that from the photolysis of RDX. This
comparison will serve as a guide to determine potential, ap-
propriate intermediates, if they exist, because the final NO
state will be determined from the species found to generate
NO directly. Different pathways will result in different inter-
mediates for RDX*→→→NO reaction, if they exist.
8
͑mass 479͒ is about 4 ␮s later, as depicted in Fig. 10. One
should note that this measurement has been conducted with
NO for RDX, C for R6G, and the parent molecule for HMT
3
at the ionization region of the TOFMS. These observations
make clear that the photofragmented NO species has come
from a molecule that is heavier than HMT and that is lighter
than the parent molecule for C , R6G. In this experiment,
In Fig. 11, the NO spectrum of the ͑0–1͒ transition from
the photolysis of RDX is compared with those from NO2,
NM, NBZ, and DNBZ photodissociation. The MALD tech-
nique is applied for DNBZ. Each NO spectrum is obviously
different from that of fragmented NO from RDX. Photolysis
of all the model systems produces highly excited rotational
spectra of fragmented NO. NM is well known to have the
intermediate NO ͑that is, CH NO ϩh␯→NO , NO ϩh␯
3
such a molecule must be RDX.
Therefore, these results ͓that is, the velocity distribution,
the spectrum of NO A� X at ͑0–1͒, and the velocity slip
data͔ lead to the conclusion that the photofragmented NO
species from RDX is generated at the ionization region of the
TOFMS. This fact indicates that, using a MALD technique,
intact gas phase RDX can be generated from the solid and
entrained in the supersonic beam. This allows the initial fun-
damental steps in the decomposition of isolated RDX from
2
3
2
2
2
→NOϩO); at the first step of the dissociation, NO produc-
2
1
9
tion is a major pathway for the CH NO molecule. This
3
2
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
84.198.228.112 On: Sun, 04 May 2014 07:58:26

J. Chem. Phys., Vol. 113, No. 18, 8 November 2000
Decomposition of RDX
7917
We will begin studies of HONO photodissociation in the
near future to determine photofragmented NO Tvib and Trot
from cold HONO. In order to determine the exact pathway of
RDX dissociation in excited electronic states, time-resolved
measurements using ultrafast, fs probing will be conducted
in the near future. This method will allow both fragment
species and RDX itself to be accessed by mass spectroscopy.
IV. CONCLUSIONS
In order to study optical spectroscopy for nonvolatile
and fragile large molecules, MALD has been successfully
combined with a molecular beam supersonic expansion tech-
nique. Sample preparation methods and proper experimental
parameters are established to obtain constant sample intro-
duction for more than 6 h. LD laser beam intensity
ϳ106 W/cm ) does not disturb the molecular beam dynam-
2
(
ics in the nozzle. In the present study, this setup is applied to
investigate photodissociation of an explosive, RDX. The
present experimental results lead us to conclude the follow-
ing:
͑
1͒ Intact and cold RDX can be obtained by the combination
of matrix-assisted laser desorption ͑MALD͒ and molecu-
lar beam techniques;
͑2͒ Photolysis of RDX produces vibrationally hotter (Tvib
ϭϳ1800 K͒, but rotationally colder (T ϭϳ20 K͒ pho-
rot
tofragmented NO compared to that from NO gas itself or
FIG. 11. NO spectra of the ͑0–1͒ transition from various model samples
͑
1
from the bottom to top͒: RDX photolysis, NO , CH NO , C H NO , and
other NO -compared molecules;
2
3
2
6
5
2
2
,4-(NO ) C H ). Each molecule shows its own pattern of NO rotational
2 2 6 4
͑3͒ By comparing the NO signal among the model systems,
the dynamics of RDX in excited states can be investi-
gated to determine that NO is probably fragmented di-
rectly from RDX using UV photons of higher than 258
transitions. This means that each molecule has its own unique pathway for
NO generation.
Ϫ1
nm ͑above 38 650 cm ͒ and that NO and CH NNO
2
2
2
fact is supported by our results; the fragmented NO spectrum
of NM resembles very closely that of NO ͑see Fig. 11͒. The
are not intermediates in this process.
2
NO spectra from NBZ and DNBZ are different from those
RDX is a true energetic material and apparently has a
distinct dissociation pathway that is different from those of
the model systems compared to it in this investigation.
generated by RDX and NO and, in fact, from one another.
2
These differences indicate that the dissociation pathways of
these molecules are much more complicated than the others,
and do not just produce NO or NO during the dissociation.
At this point, however, one cannot be certain that HONO is
the intermediate for NBZ and/or DNBZ dissociation. Addi-
2
ACKNOWLEDGMENTS
This research is supported by the USARO. We wish to
thank Dr. R. Simpson for supplying the RDX for this work.
We thank Dr. G. J. Stueber for making available to us the
technique of plasma electrolytic oxidation and pointing out
to us Ref. 14.
tionally, for NO and NM photodissociation, the transition
2
from vЉϭ2 can be observed very easily, while no transition
from vЉϭ3 can be observed. A� X transitions up to vЉϭ3
are reported for NBZ, but the NO spectrum from NBZ does
not resemble any of the other spectra. The transition from
vЉϭ3 is observed for RDX ͑see Figs. 4 and 5͒. These model
systems should have their own unique pathways through
which excited state dissociation occurs. One can confirm
1
͑a͒ T. B. Brill, T. P. Russell, W. C. Tao, and R. B. Wardle, Decomposi-
tion, Combustion, and Detonation Chemistry of Energetic Materials, MRS
Symposium Proceedings ͑Material Research Society, New York, 1996͒,
Vol. 418; ͑b͒ G. F. Adams and R. W. Shaw, Jr., Annu. Rev. Phys. Chem.
43, 311 ͑1992͒; ͑c͒ G. A. Olah and D. R. Squire, Chemistry of Energetic
Materials ͑Academic, San Diego, 1991͒; ͑d͒ Chemistry and Physics of
Energetic Materials, edited by S. N. Bulusu ͑Kluwer Academic, The Ned-
erlands, 1990͒.
from these results that NO or CH –N–NO should not be
2
2
2
an intermediate species in the photodissociation of excited
electronic state RDX. Vibrationally hot, but rotationally cold
NO molecules may well be produced directly from the dis-
2
V. Swayambunathan, S. Singh, and R. C. Sausa, Appl. Opt. 38, 6447
Ϫ1
sociation of RDX using UV photons ͑about 38 650 cm
,
͑
1999͒, and references therein.
3
258.7 nm͒, but one cannot completely dismiss the possibility
͑a͒ K. J. Smit, J. Energ. Mater. 9, 81 ͑1991͒; ͑b͒ S. N. E. Bulusu, Chem-
istry and Physics of Energetic Materials, NATO ASI Series C ͑Kluwer,
Dordrecht, 1990͒, Vol. 309; ͑c͒ Y. Oyumi and T. B. Brill, Combust. Flame
of the HONO molecule as an intermediary, because the
NBZ, DNBZϩh␯→HONO studies were performed at high
temperature and pressure. The NO thus generated, however,
62, 225 ͑1985͒; ͑d͒ F. I. Dubovitskii and B. L. Korsunskii, Russ. Chem.
Rev. 50, 958 ͑1981͒; ͑e͒ J. D. Cosgrove and A. J. Owen, Combust. Flame
22, 13 ͑1974͒; ͑f͒ F. C. Fauch and A. J. Fanelli, J. Phys. Chem. 73, 1604
1
8(a)
from HONO at the vЉϭ1 state is rotationally very hot.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
4.198.228.112 On: Sun, 04 May 2014 07:58:26
8

7918
J. Chem. Phys., Vol. 113, No. 18, 8 November 2000
H.-S. Im and E. R. Bernstein
͑
1969͒; ͑g͒ A. J. B. Robertson, Trans. Faraday Soc. 45, 85 ͑1949͒.
J. Chem. Phys. 92, 7625 ͑1990͒; ͑e͒ J. Phys. Chem. 94, 4394 ͑1990͒; ͑f͒ J.
R. Cable, M. J. Turbergen, and D. H. Levy, J. Am. Chem. Soc. 111, 9032
͑1989͒; ͑g͒ 110, 7349 ͑1988͒; ͑h͒ L. Li and D. M. Lubman, Appl. Spec-
trosc. 42, 418 ͑1988͒.
4
Y.-X. Zhang and S. H. Bauer, Int. J. Chem. Kinet. 31, 655 ͑1999͒, and
references therein.
5
͑a͒ C. Capellos, P. Papagiannakopoulos, and Y.-L. Liang, Chem. Phys.
1
3
Lett. 164, 533 ͑1989͒; ͑b͒ C. Capellos, S. Lee, S. Bulusu, and L. Gams,
Advances in Chemical Reaction Dynamics ͑Reidel, Dordrecht, 1986͒, pp.
͑
a͒ R. E. Smalley, Laser Chem. 2, 167 ͑1983͒; ͑b͒ M. Foltin, G. J. Stueber,
and E. R. Bernstein, J. Chem. Phys. 111, 9577 ͑1999͒.
3
98–404; ͑c͒ H. Zuckermann, G. D. Greenblatt, and Y. Hass, J. Phys.
Chem. 91, 5159 ͑1987͒; ͑d͒ M. D. Pace and W. B. Moniz, J. Magn. Reson.
7, 510 ͑1982͒.
14
͑
8
a͒ G. P. Wirtz, S. Brown, and W. M. Kriven, Mater. Manuf. Processes 6,
7 ͑1991͒; ͑b͒ P. Kurze, W. Krysmann, J. Schreckenbach, Th. Schwarz,
4
and K. Rabending, Cryst. Res. Technol. 22, 53 ͑1987͒; ͑c͒ W. Krysmann,
P. Kurze, K.-H. Dittrich, and H. G. Schneider, ibid. 19, 973 ͑1984͒; ͑d͒
K.-H. Dittrich, W. Krysmann, P. Kurze, and H. G. Schneider, ibid. 19, 93
6
7
͑a͒ X. Zhao, E. J. Hinsta, and Y. T. Lee, J. Chem. Phys. 88, 801 ͑1988͒;
͑b͒ T. Sewell and D. L. Thompson, J. Phys. Chem. 95, 6223 ͑1991͒.
͑a͒ R. Behrens, Jr. and S. Bulusu, J. Phys. Chem. 96, 8877 ͑1992͒; ͑b͒ 96,
͑
1984͒.
8
1
891 ͑1992͒; ͑c͒ T. B. Brill, P. E. Gongwer, and G. K. Williams, ibid. 98,
2242 ͑1994͒.
1
1
5
6
W. G. Breiland, Sandia National Laboratory, Albuquerque, NM, 1992.
͑a͒ M. Hippler and J. Pfab, Chem. Phys. Lett. 243, 500 ͑1995͒; ͑b͒ G.
Herzberg, Spectra of Diatomic Molecules ͑Van Nostrand, New York,
1950͒, p. 257.
8
9
͑a͒ J. Alix and S. Collins, Can. J. Chem. 69, 1535 ͑1991͒; ͑b͒ M. Farber
and R. Srivastava, Chem. Phys. Lett. 64, 30 ͑1979͒.
S. Bulusu, D. Weinstein, J. Autura, and R. Veliky, J. Phys. Chem. 90,
4
1
1
7
8
͑
a͒ Q. Y. Shang, P. O. Moreno, S. Li, and E. R. Bernstein, J. Chem. Phys.
121 ͑1986͒.
1
1
0
1
98, 1876 ͑1993͒; ͑b͒ S. Li and E. R. Bernstein, ibid. 97, 792 ͑1992͒.
͑a͒ C. I. Ning and J. Pfab, J. Phys. Chem. A 101, 6008 ͑1997͒; ͑b͒ S. W.
Novicki and R. Vasudev, J. Chem. Phys. 95, 7269 ͑1991͒; ͑c͒ Chem. Phys.
Lett. 176, 118 ͑1991͒; ͑d͒ J. H. Shan, S. J. Wategaonkar, and R. Vasudev,
Chem. Phys. Lett. 160, 614 ͑1989͒; ͑e͒ S. J. Wategaonkar, J. H. Shan, and
R. Vasudev, Chem. Phys. 139, 283 ͑1989͒; ͑f͒ R. N. Dixon and H. Rieley,
ibid. 137, 307 ͑1989͒; ͑g͒ J. Chem. Phys. 91, 2308 ͑1989͒.
͑a͒ Y.-X. Zhang and S. H. Bauer, J. Phys. Chem. B 101, 8717 ͑1997͒; ͑b͒
D. S. Y. Hsu and M. D. Lin, J. Energ. Mater. 3, 95 ͑1985͒; ͑c͒ A. Perche,
J. C. Tricot, and M. J. Lucquin, J. Chem. Res. 304, 3219 ͑1979͒; ͑d͒ K.
Galnzer and J. Troe, Helv. Chim. Acta. 55, 2884 ͑1972͒.
F. J. Owens and J. Sharmar, J. Appl. Phys. 51, 1494 ͑1979͒.
͑
͑
a͒ A. Meffert and J. Grotemeyer, Ber. Bunsenges. Phys. Chem. 102, 459
1998͒; ͑b͒ K. W. D. Ledingham and R. P. Singhal, Int. J. Mass Spectrom.
Ion Processes 163, 149 ͑1997͒, and references therein; ͑c͒ R. J. Conzemius
and J. M. Cepellen, Int. J. Mass Spectrom. Ion Phys. 34, 197 ͑1980͒, and
references therein.
1
2
͑a͒ F. L. Plows and A. C. Jones, J. Mol. Spectrosc. 194, 163 ͑1999͒; ͑b͒ A.
19
C. Jones, M. J. Dale, G. A. Keenan, and P. R. R. Langridge-Smith, Chem.
Phys. Lett. 219, 174 ͑1994͒; ͑c͒ M. J. Dale, A. C. Jones, P. R. R.
Langridge-Smith, K. F. Costello, and P. Cummins, Anal. Chem. 65, 174
͑
1993͒; ͑d͒ G. Meijer, M. S. de Vries, H. E. Hunziker, and H. R. Wendt,
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
4.198.228.112 On: Sun, 04 May 2014 07:58:26
8