Nitromethane decomposition under high static pressure

Article abstract of DOI:10.1021/jp1035508

The room-temperature pressure-induced reaction of nitromethane has been studied by means of infrared spectroscopy in conjunction with ab initio molecular dynamics simulations. The evolution of the IR spectrum during the reaction has been monitored at 32.2 and 35.5 GPa performing the measurements in a diamond anvil cell. The simulations allowed the characterization of the onset of the high-pressure reaction, showing that its mechanism has a complex bimolecular character and involves the formation of the aci-ion of nitromethane. The growth of a three-dimensional disordered polymer has been evidenced both in the experiments and in the simulations. On decompression of the sample, after the reaction, a continuous evolution of the product is observed with a decomposition into smaller molecules. This behavior has been confirmed by the simulations and represents an important novelty in the scene of the known high-pressure reactions of molecular systems. The major reaction product on decompression is N-methylformamide, the smallest molecule containing the peptide bond. The high-pressure reaction of crystalline nitromethane under irradiation at 458 nm was also experimentally studied. The reaction threshold pressure is significantly lowered by the electronic excitation through two-photon absorption, and methanol, not detected in the purely pressure-induced reaction, is formed. The presence of ammonium carbonate is also observed.

Full text of DOI:10.1021/jp1035508

9
420
J. Phys. Chem. B 2010, 114, 9420–9428
Nitromethane Decomposition under High Static Pressure
,
†
†,‡
‡
†,‡
Margherita Citroni,* Roberto Bini, Marco Pagliai, Gianni Cardini, and
†
,‡
Vincenzo Schettino
European Laboratory for Nonlinear Spectroscopy (LENS), Via Nello Carrara 1, 50019 Sesto Fiorentino, Italia,
and Dipartimento di Chimica, UniVersit a` di Firenze, Via della Lastruccia 3, 50019 Sesto Fiorentino, Italia
ReceiVed: April 20, 2010; ReVised Manuscript ReceiVed: June 8, 2010
The room-temperature pressure-induced reaction of nitromethane has been studied by means of infrared
spectroscopy in conjunction with ab initio molecular dynamics simulations. The evolution of the IR spectrum
during the reaction has been monitored at 32.2 and 35.5 GPa performing the measurements in a diamond
anvil cell. The simulations allowed the characterization of the onset of the high-pressure reaction, showing
that its mechanism has a complex bimolecular character and involves the formation of the aci-ion of
nitromethane. The growth of a three-dimensional disordered polymer has been evidenced both in the
experiments and in the simulations. On decompression of the sample, after the reaction, a continuous evolution
of the product is observed with a decomposition into smaller molecules. This behavior has been confirmed
by the simulations and represents an important novelty in the scene of the known high-pressure reactions of
molecular systems. The major reaction product on decompression is N-methylformamide, the smallest molecule
containing the peptide bond. The high-pressure reaction of crystalline nitromethane under irradiation at 458
nm was also experimentally studied. The reaction threshold pressure is significantly lowered by the electronic
excitation through two-photon absorption, and methanol, not detected in the purely pressure-induced reaction,
is formed. The presence of ammonium carbonate is also observed.
Introduction
Nitromethane, a liquid-insensitive high explosive, is the
prototype of the nitro-organic compounds, a family of molecules
containing the NO group bonded to an organic moiety through
2
a C-N bond. These are energetic molecules that, when subjected
to heat or shock waves, undergo a fast and highly exothermic
chemical reaction leading to the formation of gaseous com-
1
pounds. The molecular mechanism of detonation is not well
understood to date, though its knowledge and control would be
of interest in the design of novel insensitive and high energetic
materials, which is now mostly performed on a phenomenologi-
cal basis. In particular, understanding the mechanism of the
detonation initiation, i.e., how many molecules and which
chemical bonds are involved in the first steps of the reaction,
would be of great importance in the production of safe and
efficient explosives or propellants. In this respect, understanding
the mechanism of the high-pressure and high-temperature
reactivity of nitromethane would be of primary importance in
view of the simplicity of its molecular structure.
-
2 2
Figure 1. Electronic structure of the aci-ion CH NO as extracted
from the molecular dynamics simulation discussed in the text. The
Maximally Localized Wannier Centers (MLWC) are reported in green.
The H atoms are indicated in white, the C atoms in gray, the N atoms
in blue, and the O atoms in red.
In the gas phase, thermal decomposition of nitromethane
involves as the first step the intramolecular homolytic cleavage
2,3
of the C-N bond, whereas in condensed phases experimental
4-6
7,8
data and theoretical predictions suggest that the tempera-
ture- or pressure-induced chemical reactions involve two or more
molecules in the initial step, even if they ultimately result in a
decomposition into small gaseous molecules. In particular, time-
resolved Raman measurements under stepwise loading5 re-
vealed an intensification of intermolecular interactions involving
the H atoms, evidenced by a broadening of the C-H symmetric
stretching mode and by a softening of the C-N stretching mode
in the first steps of the reaction. The role of intermolecular
interactions involving the migration of an H atom or an H ion
is supported by the evidence that nitromethane is sensitized by
small quantities of amines both in static compression experi-
ments and under shock waves. Sensitization likely occurs
through the formation of the aci-ion of nitromethane (Figure
), consequent to the abstraction of an H atom by the basic
amine molecule. Engelke et al. provided experimental
evidence that the aci-ion is also present in pure liquid ni-
+
,6
9
10
*
To whom correspondence should be addressed. E-mail: margherita@
1
lens.unifi.it.
11
12
†
LENS.
Dipartimento di Chimica.
‡
1
0.1021/jp1035508  2010 American Chemical Society
Published on Web 07/07/2010

CH3NO2 Decomposition under High Static Pressure
J. Phys. Chem. B, Vol. 114, No. 29, 2010 9421
tromethane at high pressure (2 GPa and 81 °C), in a concentra-
tion 1000 times larger than that at room conditions. It was also
demonstrated that irradiation with UV light increases the room
13
pressure concentration of the aci-ion. Car-Parrinello molecular
dynamics simulation of the reactivity of hot (3000-4000 K)
8
and dense (V/V
0
) 0.68-0.61) liquid nitromethane showed that
in these extreme conditions the first step of the reaction is an
intermolecular proton abstraction leading to the formation of
+
-
3 2 2 2
CH NO H and the aci-ion CH NO , with a concerted mech-
anism involving at least three nitromethane molecules.
The decomposition reaction of nitromethane under high static
4
9,14,15
pressure has been studied in the liquid and solid phase
from room temperature to 350 °C. There is consensus on the
strong increase of the reaction rate with increasing temperature
or pressure, the latter observation indicating a negative activation
volume and thus suggesting a bimolecular reaction mechanism.
However, the reaction product, composed of a solid and a fluid
phase, has not been characterized. The lack of a Raman signal
suggested that the product of the reaction is a solid amorphous
compound.14 On the basis of the infrared and mass spectra,4
the solid product of the high-pressure thermal decomposition
Figure 2. Image of the crystal structure of nitromethane at 15 GPa
resulting from the simulation. The solid lines indicate the edges of the
simulation cell, as described in the text. The H atoms are indicated in
white, the C atoms in gray, the N atoms in blue, and the O atoms in
red. The dotted lines indicate the shortest intermolecular O · · ·H
distances (e2.3 Å).
4,9
liquid nitromethane, g99% from Aldrich, was loaded for each
experiment. The initial sample diameter was 100 µm. In most
experiments, the sample was either loaded in a gasket about 35
µm thick or deposited as a 10 µm thick layer on a KBr pellet
previously formed in the DAC.
(
at temperatures above 100 °C and pressures from 2 to 7 GPa)
9
has been suggested to be primarily ammonium oxalate or a
mixture of ammonium formate and water. On the other side,
4
the high-pressure (28-50 GPa) room-temperature reaction
1
5
product has been reported to be a dark yellow liquid
amorphous polymer containing a variety of functional groups
as suggested by the infrared spectrum.
The IR spectra were recorded with a Bruker IFS-120HR FTIR
1
7
spectrometer, modified to allow measurements in the DAC,
-1
in the mid-IR region 600-5000 cm with a spectral resolution
In the present work, the pressure-induced reaction of ni-
tromethane at room temperature has been studied in full detail
with a combined experimental and theoretical approach. The
time evolution of the reaction has been monitored by infrared
spectroscopy obtaining definite evidence of a bimolecular
reaction mechanism. This is fully supported by the results of
ab initio molecular dynamics simulations in the Car-Parrinello
approach based on a recently proposed crystal structure deter-
mination of nitromethane up to the high-pressure reaction
threshold of 32 GPa.16 The connection of the reaction mecha-
nism with the details of the crystal structure is evidenced in a
solid state reactivity point of view. Particular attention has been
devoted to the characterization of the reaction product evidenc-
ing some unique features of the nitromethane reaction. At high
pressure, an extended three-dimensional disordered compound
is obtained as attested by both experimental and computational
results. The disordered compound cannot be quenched at room
pressure. On decreasing the pressure, the polymeric array
decomposes giving rise to a number of small volatile molecules.
The nonvolatile product of the reaction can be identified as
mainly composed by N-methylformamide. The experimental
behavior monitored by infrared spectroscopy is nicely repro-
duced in the molecular dynamics simulation, and the resulting
molecules can be at least partly identified. The effect of laser
irradiation on the high-pressure reaction of nitromethane has
also been studied showing that, as reported for other molecular
crystals, two-photon absorption significantly lowers the high-
pressure reaction threshold. The spectrum is similar to that of
the pure pressure-induced reaction with important additional
features. On decreasing the pressure, besides formamide the
presence of methanol and ammonium carbonate can be identified.
-
1
of 1 cm . The pressure was determined by the ruby fluores-
cence technique, using as the excitation source the 2nd harmonic
of a cw Nd:Yag laser (532 nm) with a power lower than 0.5
mW to avoid unintended photochemical effects. The photo-
chemical behavior of nitromethane at high pressure was probed
+
by irradiation with 100 mW of the 458 nm Ar laser line,
focused in such a way to irradiate the sample uniformly.
Computational Details
A series of ab initio molecular dynamics simulations of the
nitromethane reaction under high pressure have been performed
1
8,19
with the Car-Parrinello method (CPMD).
All the calcula-
tions have been carried out using the BLYP2
0,21
exchange-
correlation functional along with a 80 Ry cutoff of the plane
waves expansion of the Kohn-Sham orbitals. Norm conser-
2
2
ving Martins-Troullier pseudopotentials with Kleinman-
Bylander23 decomposition have been adopted for all atoms, and
the equations of motion have been integrated with a time step
of 1 au (0.02419 fs) for a time ranging from 40 000 to 180 000
steps at each volume except for the last simulation at 750 K,
where the time step has been increased gradually until 4 au
(0.0968 fs) after the system did not show any further reactivity.
Preliminary calculations on a sample with 16 molecules (i.e.,
with two unit cells in the a and b directions) confirmed the
16
results reported by Citroni et al. showing that, despite the poor
representation of the van der Waals interactions, the present
approach reasonably reproduces the crystal structure and the
vibrational spectra of nitromethane at pressures below the
reaction threshold. The minor relevance of the correction for
the van der Waals interactions, in this case, has been discussed
by Conroy et al.24 For the purpose of the present work, a major
point is that the adopted approach nicely reproduces the network
of hydrogen bonds (Figure 2) that is a key feature of the crystal
Experimental Methods
A membrane diamond anvil cell (DAC), equipped with IIa-
type diamonds, was used to generate high pressure. Rhenium
gaskets were used for all of the experiments. A fresh sample of
16
structure at high pressure and for the onset of the high-pressure
chemical reaction, as will be discussed in the following.

9
422 J. Phys. Chem. B, Vol. 114, No. 29, 2010
Citroni et al.
Figure 3. FTIR spectra of a ∼10 µm thick sample of nitromethane, deposited on a KBr pellet, at 32 GPa as a function of time. The absorbances
are calculated using a reference spectrum made on a KBr pellet in the DAC.
Simulations of the chemical reaction of nitromethane were
carried out on a sample of 16 molecules in a simulation box of
GPa, to detect any spectral change ascribable to a chemical
transformation. At 32.3 GPa, after an induction time of about
3 h, the first spectral modifications due to a chemical reaction
were observed. The pressure was maintained between 32.3 and
32.1 GPa for the whole duration of the reaction. The FTIR
spectrum was monitored until no more changes occurred with
time, which was after about 680 h (see Figure 3). Another thin
2
unit cells in the a and b directions and 1 unit cell in the c
direction. The initial unit cell parameters were chosen as the
experimental values at 1.1 GPa. The pressure was gradually
increased stepwise to give starting configurations for constant-
volume simulations at different densities. The final products
were kept for about 180 000 steps (∼4 ps) at 300 K without
observing any further reaction. The reverse procedure was
followed to enlarge the simulation box to reach a final density
comparable to that experimentally observed at ambient pressure
and 300 K. At this temperature, the reaction evolution on
decreasing the pressure is very slow, and to increase the reaction
rate the temperature was gradually raised by velocity scaling
to 750 K, continuing the simulation until no further reaction
was observed for about 50 ps, on a simulation of more than
CH
3
NO
2
sample (∼5 µm) deposited on a KBr pellet was
pressurized, in steps of 1-3 GPa, up to 35.5 GPa. At this
pressure, the transformation was completed in 120 h. Different
experiments on pure nitromethane samples were carried out to
check if the presence of the KBr pellet affected the reaction
pressure threshold and evolution or the recovered products. No
significant differences could be observed.
The reaction is characterized by the slow growth of very
strong and broad absorptions in the regions 900-1700 and
1
00 ps.
-
1
-1
2
500-3500 cm , with maxima at about 1200 and 3200 cm ,
To understand the changes of the electronic structure and the
and by a corresponding decrease of intensity of the nitromethane
absorptions (Figure 3). When the spontaneous evolution of the
spectrum is completed, a small amount of nitromethane is still
present. The broad and almost structureless spectrum is clear
evidence that the reaction product is an amorphous extended
polymeric array. The assignment of the observed spectral
features is not straightforward considering also, in the regions
of lower intensity absorptions, the overlap with the interference
fringes between the diamond faces delimiting the sample.
However, it can be said with confidence that the structureless
initial steps of the reaction pathway, selected configurations have
been characterized by calculating the Maximally Localized
Wannier functions.2
5-27
These are obtained by minimizing the
spread of the Wannier functions (w
n
) in the real space
N
2
2
S )
(〈w |r |w 〉 - 〈w |r|w 〉 )
(1)
∑
n
n
n
n
n)1
Maximally Localized Wannier functions have also been used
to identify and characterize the species obtained during the
decompression. To correctly select the chemical species present
in the sample, the final configurations have been inspected
applying periodic boundary conditions.
-
1
absorption centered at ∼3200 cm arises from strongly
associated -OH, -NH, or -NOH groups. In this region,
features to be assigned to saturated -CH bonds can also be
-
1
recognized. The broad band centered at ∼1200 cm likely
arises from vibrations of CO, CC, CN, and NO single bonds. It
is remarkable that the overall shape of the band is reminiscent
of the broad feature observed in the same region for the high-
pressure decomposition product of furane and assigned to CC
and CO single bonds.28 In the present case, the band is
considerably broader, as expected. Weak features in the
Results and Discussion
High-Pressure Reaction. In all of the experiments, the
pressure was increased in steps of 1-2 GPa, and the FTIR
spectrum was measured at each step after pressure stabilization.
The evolution of the absorption frequencies, intensities, and
bandwidths on compression reproduces exactly the behavior
previously found and reported in ref 16. To precisely locate
the reaction threshold pressure, the FTIR spectrum of a thin
layer (∼10 µm) of nitromethane deposited on a KBr pellet was
monitored for more than 24 h at all the pressure steps above 21
-
1
spectrum observed at ∼1735 and ∼1600 cm can likely be
associated with CO or NO double bonds, as is also suggested
by the further reaction evolution on decreasing the pressure, as
will be discussed in the next section.

CH3NO2 Decomposition under High Static Pressure
J. Phys. Chem. B, Vol. 114, No. 29, 2010 9423
These experimental observations are, at least qualitatively,
nicely reproduced in the molecular dynamics simulation. In fact,
on decreasing the pressure, the high-pressure three-dimensional
array decomposes and at room pressure gives a number of small
molecules (including water, CO
2
, formaldehyde, formic acid,
-
dihydroxylamine, a NH
3
OH complex) together with larger
molecules. These latter seem to be unstable, and in fact, they
further decompose on heating at 500 and 700 K. A list of the
molecules observed in the simulation on decompression is
shown in Figures 1 and 2 of the Supporting Information.
The spectrum of the reaction product after decreasing the
pressure and the cell opening (Figure 6) closely resembles, but
is not exactly equal to, the spectrum of the product reported by
Pruzan et al.15 from experiments carried out at room pressure
and 50 GPa. Comparison of the spectrum of the recovered
sample after decompression and cell opening with the spectrum
of a sample of N-methylformamide compressed to 4.7 GPa and
decompressed (Figure 6) allows us with some confidence to
identify this residual reaction product as N-methylformamide.
In fact, the position and relative intensity of the most intense
Figure 4. Simulation cell at the end of the compression. The H atoms
are indicated in white, the C atoms in gray, the N atoms in blue, the O
atoms in red, and the MLWC in green.
-
1
-1
bands, the amide I (1670 cm ), amide II (1540 cm ), and the
-
1
The results of the molecular dynamics simulation are of
considerable help to figure out the possible structure of the high-
pressure reaction product. In the simulation, a reaction is
band at 1390 cm , to be likely assigned as the C-H bending
mode, match very nicely, apart from the significant broadening
of the bands of the product as a consequence of the residual
observed at V/V
0.67). This is due to the need of accelerating the reaction in
the simulation and to the already mentioned poor representation
0
< 0.43, much smaller than experimental (V/V
0
stress. This is remarkable considering that the amide bands are
generally sensitive to the environment.3
0,31
Unfortunately, the
<
weakness of the spectrum in the lower frequency region and
the overlap with the interference fringes prevents us from
gathering further evidence for this identification. However, the
29
of the van der Waals interactions. Figure 4 shows the structure
of the simulation cell at the end of the reaction. It can be seen
that a complex three-dimensional array of chemical bonds is
formed with terminal -COH, -CNH, -NOH, -CNOH, and
-
1
comparison of the spectrum in the 3200 cm region is in good
agreement with identification of the product as N-methylfor-
mamide. This result is of considerable interest since the product
contains the prototype peptide bond. The basic binding unit of
biochemistry is thus formed in a high-pressure reaction from
an apparently unrelated molecule.
-
CH bonds that nicely account for the high-frequency part of
the experimental spectrum. In the inside of the array, we find
CNO, CON, NCO, OCO, and ONO sequences that all give rise
-
1
to absorption frequencies in the region around 1200 cm . As
a whole, the molecular dynamics simulation indicates that the
complex absorption spectrum does not result from the overlap
of spectra of different reaction products but from a single
amorphous three-dimensional array.
The molecular dynamics simulation cannot account for the
details of the formation of this reaction product since the
conditions (temperature and pressure) of the simulation differ
from experiments. However, it is rewarding that both formic
acid and ammonia are observed on decompression. It can also
be noted that the formation of N-hydroxyformamide is observed
in the molecular dynamics simulation. It is quite possible that
the peptide bond is produced by condensation of some of the
small molecules formed from the decomposition of the high-
pressure polymeric array.
Reaction on Decreasing the Pressure. The high-pressure
reaction product described in the previous section is metastable
and cannot be quenched at ambient pressure. In fact, on
decreasing the pressure, the infrared spectrum changes substan-
tially as can be seen from Figure 5. In particular, the prominent
-1
broad feature at ∼1200 cm decreases in intensity and
disappears at pressures below 15 GPa. At the same time, a strong
Reaction Mechanism. The time evolution of the intensity
of the nitromethane IR absorption bands during the reaction
has been used for a kinetic analysis. The absorption band due
to the ν12 mode of nitromethane, at 32.3 and 35.5 GPa, was
fitted to a pseudo-Voigt profile, and the integrated intensity y(t),
proportional to the nitromethane concentration, is reported as a
function of time in Figure 7. The band was chosen because it
is isolated, always in scale and with a sufficient intensity. The
-1
band at ∼1700 cm develops below 20 GPa to finally become
the strongest band in the spectrum. Only minor changes, on
the contrary, are observed on decompression in the region
-1
around ∼3200 cm , apart from the complete disappearance of
the residual nitromethane peaks. This indicates the persistence
of strongly associated hydrogen bonded groups. When the
pressure in the membrane is completely released (with a residual
pressure on the sample of 1.8 GPa), the sample appeared
transparent in transmission on observation with the microscope
with inclusion of some gas bubbles. Actually, a peculiar feature
of the spectrum of the decompressed sample is the appearance
of a sharp peak at 2350 cm- clearly assigned to carbon dioxide.
The formation of carbon dioxide has also been reported in the
3
2
time evolution of y(t) was fitted to the Avrami equation,
a
rate law describing a process of nucleation, i.e., formation of
reactive sites and diffusion-controlled nuclei growth inside a
material.
1
n
y(t) ) y - (y - y )[1 - exp[-k(t + t ) ]]
(2)
0
0
∞
0
thermal decomposition (20-50 GPa, 100-160 °C) of ni-
9
tromethane. Opening the cell in a N
2
atmosphere causes in most
In eq 2, y
intensity when the reaction evolution is completed; t
the induction time; n is a parameter related to the spatial
0
is the integrated area at time t
0
; y
∞
is the integrated
cases a complete loss of the product. However, occasionally
on cell opening a part of the sample was left in the cell, probably
adsorbed on the KBr pellet, and the resulting infrared spectrum
is reported in Figure 6.
0
represents
3
2
distribution of the nuclei growth; and k is the rate constant.
The results of the fit are reported in Table 1. The value n ) 2

9
424 J. Phys. Chem. B, Vol. 114, No. 29, 2010
Citroni et al.
Figure 5. FTIR spectra of a ∼5 µm thick sample of nitromethane, deposited on a KBr pellet, on decompression after the spontaneous reaction at
5.5 GPa. The absorbances are calculated using a reference spectrum made on a KBr pellet in the DAC.
3
Figure 6. Trace a: FTIR spectrum of the reaction product before the cell opening after decompression, with a residual pressure of 1.80 GPa (∼5
µm thick sample deposited on a KBr pellet). Trace b: FTIR spectrum of the reaction product at room pressure after cell opening (∼5 µm thick
sample deposited on a KBr pellet). Trace c: FTIR spectrum of N-methylformamide in the DAC at room pressure after compression to 4.70 GPa (∼5
µm thick sample deposited on a KBr pellet). The absorbances are calculated using a reference spectrum made on a KBr pellet in the DAC.
implies a three-dimensional nuclei growth with a nucleation rate
decreasing with time. The positive value of t indicates an
• The H+ ion is attracted by an oxygen of a neighboring
+
0
nitromethane molecule and forms a CH
3
NOOH ion (panel b).
induction time for the reaction. Our data show that the induction
time is much smaller at 35.5 than at 32.2 GPa. Even though we
have only two data points, the marked increase of the reaction
rate with pressure can be taken as a further proof that the
activation volume of the reaction is negative, and thus the
reaction proceeds through at least a bimolecular mechanism.
This is supported by the results of the molecular dynamics
simulation.
The positive charge is localized on the nitrogen atom.
+
•
The bond between the H and the deformed nitro group is
+
not very stable, and the H ion moves close to the nitro group
of another molecule (panel c). A positive charge is present on
the nitrogen and a negative charge on one of the O atoms of
+
the nitro group of the nitromethane that has lost the H ion.
•
The two charges are very close to each other, and a covalent
+
N-O bond forms; the H is attracted back by the oxygen atom
forming an OH group (panel d).
The first stage of the reaction mechanism has been investi-
gated by the Maximally Localized Wannier Centers (MLWC)2
and can be described in four steps (see Figure 8):
5-27
This reaction mechanism onset is quite reasonable if the
crystal structure is taken into account. In fact, the crystal
structure of nitromethane is stabilized above 11 GPa by the
presence of N-O· · ·H hydrogen bonds extending along the three
crystallographic directions (especially a and b).16 At 15 GPa,
the minimum O · · ·H distance is 2.16 Å, but several inequivalent
•
A nitromethane molecule deformed by the neighboring
-
molecules (panel a) loses a proton and forms the CH
panel b) with a double bond localized between the carbon and
the nitrogen atom.
2 2
NO ion
(

CH3NO2 Decomposition under High Static Pressure
J. Phys. Chem. B, Vol. 114, No. 29, 2010 9425
Figure 7. Time decay of the integrated area of the mode ν12 of nitromethane during the reaction at 32.2 GPa (left panel) and at 35.5 GPa (right
panel). Circles: experimental data. Lines: fit to the Avrami equation.
TABLE 1: Kinetic Parameters Obtained from the Fit to the
Avrami Equation
reaction P (GPa)
y
0
y
∞
n
k (h-1)
t
0
(h)
ꢀ2
-
-
6
4
3
3
2.2
5.5
18.4 3.0 2.0 3.5 × 10
16.0 2.4 1.9 9.2 × 10
179
15
0.04
0.04
O· · ·H distances have very similar values, with ∠OHC angles
ranging from 122° to 140°, creating an extended network of
intermolecular H-bonds. The lack of phase transitions up to the
reaction pressure, and the volume contraction (V/V
0
is 0.65 at
1
6
1
5 GPa and 0.60 at 27 GPa ) that involves the three crystal-
lographic axes in a comparable amount, make it very reasonable
+
that the first step of the reaction involves an H migration and
that later the reaction is propagated along nearly any direction
in the crystal. Consistently, the reaction kinetics for the
formation of this extended compound can be described by an
Avrami equation, deriving from a model of growth of a new
phase inside a material, where a fast nucleation is followed by
a diffusion-controlled nuclei growth in the three directions.
Photoassisted Pressure-Induced Reaction. The UV absorp-
Figure 8. Mechanism of the reaction onset as obtained by the
simulation. The H atoms are indicated in white, the C atoms in gray,
the N atoms in blue, the O atoms in red, and the MLWC in green. (a)
33,34
tion spectrum of nitromethane
consists of a strong absorption
centered at 198 nm (extending from 165 to 225 nm) assigned
to a πfπ* transition and a much weaker absorption centered
at 270 nm (extending from 250 to 355 nm) mainly due to a
σfπ* transition, overlapped to a forbidden nfπ* transition.
The involved molecular orbitals are all localized on the nitro
+
The C-H bond starts to break. (b) The H ion migrates to the O atom
-
+
of the neighbor molecule. The CH NO and CH NOOH ions are
2
2
3
+
formed. (c) A third molecule attracts the H , while the N atom of the
CH NOOH ion and an O atom of the aci-ion start forming a new
3
+
N-O bond. (d) The new N-O bond is formed, and the formed species
contains a terminal OH group.
33
group (including the σ orbital of the C-N bond), and,
assuming a C2V symmetry, the three transitions mentioned
˜
1
correspond to excitations from the fundamental X A
B , B , and A
1 2 2
1
state to
states, respectively. The band gap in the
nitromethane crystal has been calculated to be 3.28 eV (378
each step. Irradiation cycles of 3-4 h with 100 mW of 514 or
458 nm laser light were performed at 12 GPa, and no changes
were observed in the FTIR spectrum before and after irradiation.
At 15.0 GPa, irradiation with 100 mW at 458 nm induced an
extremely slow chemical transformation. The 458 nm line is
absorbed through a two-photon process, presumably matching
the πfπ* transition. Hence, we performed irradiation cycles
at 18.5 GPa until the transformation reached a stationary state.
The evolution of the infrared spectrum during the irradiation
(Figure 9) is qualitatively similar to that observed in the purely
pressure-induced reaction. The nitromethane peaks at 710 and
1
1
1
3
5
36
nm) or 3.61 eV (344 nm) at room pressure and to decrease
by about 0.43
5,37
or 0.23 eV, depending on the calculation
36
method, in the pressure range 0-20 GPa. A significant lowering
of the reaction threshold pressure by excitation to the first
electronic states has been reported in several crystalline mo-
lecular systems38 and is due to the attainment of highly reactive
3
9,40
structures after decay to the ground electronic state.
A 35 µm thick sample of pure nitromethane was used to probe
the effects of laser irradiation on the high-pressure reactivity.
After loading, the pressure was raised in steps of 2 GPa, waiting
for pressure stabilization and monitoring the FTIR spectrum at
-
1
1120 cm do not disappear completely. Other nitromethane
peaks are only barely visible probably because they are heavily
overlapped with the product bands. Also, in this case the reaction

9
426 J. Phys. Chem. B, Vol. 114, No. 29, 2010
Citroni et al.
Figure 9. FTIR spectra of a 35 µm thick sample of nitromethane at 18.5 GPa as a function of the irradiation time (irradiation performed with 100
mW at 458 nm).
Figure 10. FTIR spectra of nitromethane on decompression after the photoassisted reaction.
1
is characterized by the intensification of the IR absorptions in
nitromethane bands at 710 and 1030 cm- are unchanged on
decreasing the pressure: their relative intensity is the same as
in the initial spectrum of the experiment, and this confirms their
-1
the regions 900-1800 and 2600-3500 cm . However, some
significant differences can be noted. The CdO stretching band
at 1700 cm- assigned to an amidic group is already evident in
the spectrum. This can be due to the lower pressure of this
experiment. In fact, in the pure pressure experiment, on
decreasing the pressure this peak is already clearly seen at 14.6
GPa (see Figure 5). The major difference is the prominence of
1
-1
assignment to nitromethane. The peaks at 1370 and 1030 cm
that represent the most evident difference with respect to the
purely pressure-induced reaction do not change substantially
upon unloading. In conclusion, apart from the residual bands
of nitromethane, the spectrum of the sample after unloading
the pressure looks like a superposition of the spectrum of
N-methylformamide and that of other species that can be most
likely identified as methanol, whose most intense band falls
around 1030 cm , and ammonium carbonate, assuming that
the three bands of this compound below 1500 cm
broadened and overlapped with the N-methylformamide band
falling in this same region (see Figure 10). The formation of
an ammonium salt is not really surprising. In fact, Brasch
already reported on the possible formation of ammonium oxalate
in the thermal decomposition of nitromethane. Piermarini et al.
reported that the pressure decomposition of nitromethane above
400 K produces ammonium formate even though the limited
region of the spectrum does not allow for this identification with
-1
two peaks at 1370 and at 1030 cm . In the high-frequency
region of the spectrum, an enhancement of the intensity around
1
3
200 cm- is observed.
-
1
The effect of decreasing the pressure on the infrared spectrum
-
1 41
is shown in Figure 10. On complete unloading of the membrane
pressure, a residual pressure of 4.74 GPa is still present on the
sample, whereas the cell opening caused a complete loss of the
sample, as happened in all the experiments where no KBr pellet
was used. Also in this case an overall loss of intensity of the
are
9
-1
4
broad band centered at 1200 cm is observed, while the amidic
band at 1700 cm- increases to finally become the most intense
band of the spectrum. At low pressure, the antisymmetric
stretching mode of carbon dioxide is observed again. The
1

CH3NO2 Decomposition under High Static Pressure
J. Phys. Chem. B, Vol. 114, No. 29, 2010 9427
certainty. On the other side, experiments on H
2
O/CO
2
/NH
3
apply further to the reaction stages on pressure unloading when
the small molecules forming from the product decomposition
likely acquire an increased mobility.
mixtures resulted in the formation of ammonium bicarbonate
4
2
under shock synthesis or on the formation of ammonium
43
carbamate in thermal processes. On closer observation, it seems
that the spectral features assigned to ammonium carbonate are
also weakly present in the product of the pure pressure reaction
and are washed out on cell opening (Figure 6). The formation
of methanol is consistent with the known photochemistry of
nitromethane. In fact, photodissociation of gas-phase ni-
tromethane causes primarily the omolythic cleavage of the C-N
bond, through excitation of either the πfπ* or the σfπ*
The recovered products, once volatile molecules are elimi-
nated, include N-methylformamide and ammonium carbonate.
The recovery of N-methylformamide, a molecule containing the
prototype peptide bond, is a remarkable result since formamide
is considered a possible key molecule for the formation of
complex biochemical systems.57 The formation of formamide
has been reported from reactions of elementary molecules
containing H, C, O, and N atoms that are indeed the atomic
constituents of nitromethane. N-Methylformamide does not form
directly from nitromethane but most likely from the small
molecules arising from the decomposition of the high-pressure
product. The interesting point is that N-methylformamide forms
in the absence of any catalytic support exploiting the increased
density at high pressure. On the other side, the formation of an
ammonium salt has been reported by several authors both from
nitromethane decomposition and from catalyzed reactions of
simple molecules. The presence of methanol as a product only
in the photoassisted reaction is perfectly consistent with the
known photochemistry of nitroalkanes that photodissociate
through the C-N bond cleavage. This represents a new reaction
channel competing in the crystal with the one observed in the
purely pressure-induced reaction.
44
transition. The C-N bond breaking was identified as the main
primary process also in the nitromethane photodissociation in
45
46
the liquid phase and in the Ar matrix at 14 K, as well as in
higher nitroalkanes.47 Our finding of the possible methanol
formation at high pressure, occurring only in the photoassisted
reaction, strongly suggests that the C-N bond photodissociation
also occurs in the bulk crystalline phase of nitromethane.
Conclusions
The high-pressure chemical reaction of nitromethane shows
some peculiar features compared to other molecular crystals
4
8,49
2
investigated so far. A number of crystals, including N ,
50
51
CO
2
,
and formic acid, polymerize at high pressure and
transform in 3-dimensional arrays that cannot be quenched at
ambient conditions. In these systems, the reaction is reversible,
and at ambient pressure the original molecular system is fully
Acknowledgment. This work was supported by the European
Union FP7 G.A.No 228334-LASERLAB EUROPE, by the
Italian Ministero dell’Universit a` e della Ricerca Scientifica e
Tecnologica (MURST), and by “Firenze Hydrolab” through a
grant by Ente Cassa di Risparmio di Firenze. We would like to
thank the CINECA supercomputer center for a generous
allocation of computer time.
5
2
recovered. Small unsaturated hydrocarbons like acetylene,
ethylene,53 and butadiene, on the contrary, polymerize com-
pletely at high pressures, and the polymer can be recovered.
Depending on the reaction conditions, and in particular on
photophysical activation, conformationally or crystalline pure
polymers can be recovered. A still different behavior is observed
54
54
53
5
5
28
for aromatics like benzene and furan which are unable to
react completely at high pressures giving an admixture of an
amorphous material and the reactant. On releasing the pressure,
the residual reactant further transforms into the amorphous
tridimensional polymer. As described, nitromethane fully trans-
forms at high pressure in a tridimensional array which is
intrinsically unstable but does not revert to the original reactant.
The high-pressure product decomposes on decreasing the
pressure in a number of small volatile molecules, only in part
identified, that are lost at zero pressure and on cell opening. As
a whole, the chemical decomposition of nitromethane under
hydrostatic pressure is not dramatically dissimilar from that
under shock loading since small volatile molecules are ultimately
obtained in both cases. However, under hydrostatic pressure the
reaction proceeds smoothly and in two quite distinct stages.
Therefore, the idea that studying the nitromethane under static
pressure can give information on the reaction mechanism that
is useful also under shock loading is fully supported by the
experiments.
Supporting Information Available: Graphic showing some
of the molecules isolated at the end of the simulations at 300 K
and graphic showing some of the molecules isolated at the end
of the simulations at 750 K. This material is available free of
charge via the Internet at http://pubs.acs.org.
References and Notes
(1) Manaa, M. R.; Fried, M. E.; Melius, C. F.; Elstner, M.; Frauenheim,
Th. J Phys. Chem. A 2002, 106, 9024–9029.
(
2) Crawforth, C. G.; Waddington, D. J. Trans. Faraday Soc. 1969,
5, 1334–1349.
3) Wang, J.; Brower, K. R.; Naud, D. L. J. Org. Chem. 1997, 62,
9048–9054.
(4) Piermarini, G. J.; Block, S.; Miller, P. J. J. Phys. Chem. 1989, 93,
57–462.
5) Pangilinan, G. I.; Gupta, Y. M. J. Phys. Chem. 1994, 98, 4522–
529.
6) Winey, J. M.; Gupta, Y. M. J. Phys. Chem. B 1997, 101, 10733–
10743.
6
(
4
4
(
(
(
7) Bardo, R. D. Proceedings of the Eighth Symposyum (Int.) on
Detonation, NSWC MP 86-194, NSWC White Oak, Silver Spring, MD
From kinetic analysis at two different pressures, further
evidence has been gathered in favor of a bimolecular reaction
mechanism in the first stage of the reaction. This is also
evidenced by the ab initio molecular dynamics simulation
2
0903-5000, Albuquerque, NM, July 15, 1985.
(
8) Manaa, M. R.; Reed, E. J.; Fried, L. E.; Galli, G.; Gygi, F. J. Chem.
Phys. 2004, 120, 10146–10153.
9) Brasch, J. W. J. Phys. Chem. 1980, 84, 2084–2085.
(
+
(10) Constantinou, C. P.; Winey, J. M.; Gupta, Y. M. J. Phys. Chem.
994, 98, 7767–7776.
showing that the reaction is triggered by an intermolecular H
1
+
-
3 2 2
transfer to form a CH NOOH species and the aci-ion CH NO .
(
11) Engelke, R.; Earl, W. L.; Rohlfing, C. M. J. Chem. Phys. 1986,
This reaction step nicely correlates with the high-pressure crystal
structure of nitromethane dominated by a tridimensional network
of fairly strong hydrogen bonds. This is a nice demonstration
84, 142–146.
(12) Engelke, R.; Schiferl, D.; Storm, C. B.; Earl, W. L. J. Phys. Chem.
988, 92, 6815–6819.
1
5
6
(13) Engelke, R.; Earl, W. L.; Rohlfing, C. M. J. Phys. Chem. 1986,
0, 545–547.
of the relevance of the topochemical principle in solid state
reactivity, allowing for reactions occurring with minimum
atomic displacements. The topochemical principle does not
9
(14) Courtecuisse, S.; Cansell, F.; Fabre, D.; Petitet, J.-P. J. Chem. Phys.
1995, 102, 968–974.

9
428 J. Phys. Chem. B, Vol. 114, No. 29, 2010
Citroni et al.
(
15) Pruzan, Ph.; Canny, B.; Power, C.; Chervin, J. C. Proceedings of
(37) Margetis, D.; Kaxiras, E.; Elstner, M.; Frauenheim, Th.; Manaa,
M. R. J. Chem. Phys. 2002, 117, 788–799.
the XVII International Conference on Raman Spectoscopy (ICORS 2000);
Zhang, S.-L., Zhu, B.-f., Ed.; John Wiley and Sons: New York, 2000; p
(38) Bini, R. Acc. Chem. Res. 2004, 37, 95–101.
1
42.
16) Citroni, M.; Datchi, F.; Bini, R.; Di Vaira, M.; Pruzan, Ph.; Canny,
B.; Schettino, V. J. Phys. Chem. B 2008, 112, 1095–1103.
17) Bini, R.; Ballerini, R.; Pratesi, G.; Jodl, H. J. ReV. Sci. Instrum.
997, 68, 3154–3160.
(39) Citroni, M.; Bini, R.; Foggi, P.; Schettino, V. Proc. Natl. Acad.
Sci. 2008, 105, 7658–7663.
(40) Citroni, M.; Costantini, B.; Bini, R.; Schettino, V. J. Phys. Chem.
B 2009, 113, 13526–13535.
(41) Peeters, Z.; Hudson, R. L.; Moore, M. H.; Lewis, A. The formation
and stability of carbonic acid on outer solar system bodies; 2009; http://
ntrs.nasa.gov, document ID: 20090033103.
(
(
1
(
(
18) Car, R.; Parrinello, M. Phys. ReV. Lett. 1985, 55, 2471–2474.
19) Hutter, J.; Alavi, A.; Deutch, T.; Bernasconi, M.; Goedecker, St.;
(
42) Price, M. C.; Burchell, M. J.; Miljkovic, K.; Kearsley, A. T.; Cole,
Marx, D.; Tuckerman, M.; Parrinello, M. CPMD; MPI f u¨ r Festk o¨ rperfor-
schung and IBM Zurich Research Laboratory: Stuttgart, 1995-1999.
M. J. Shock synthesis of organics from simple ice mixtures, 41 Lunar and
Planetary Science Conference, The Woodlands, TX, March 1-5, 2010.
(
(
(
(
20) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652.
(
43) Bossa, J. B.; Theul e´ , P.; Duvernay, M.; Borget, F.; Chiavassa, T.
21) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785–789.
22) Troullier, N.; Martins, J. L. Phys. ReV. B 1991, 43, 1993–2006.
23) Kleinman, L.; Bylander, D. M. Phys. ReV. Lett. 1982, 48, 1425–
Astron. Astrophys. 2008, 492, 719–724.
(
44) Wade, E. A.; Reak, K. E.; Li, S. L. J. Phys. Chem. A 2006, 110,
405–4412, and references therein.
45) Jarosiewicz, M.; Szychlinski, J.; Piszczek, L. J. Photochem. 1985,
9, 343–351.
46) Jacox, M. E. J. Phys. Chem. 1984, 88, 3373–3379.
(47) Ross, P. L.; Van Bramen, S. E.; Johnston, M. V. Appl. Spectrosc.
4
1
428.
(
(
24) Conroy, M. W.; Oleynik, I. I.; Zybin, S. V.; White, C. T. J. Phys.
2
Chem. A 2009, 113, 3610–3614.
(
(
(
25) Mazari, N.; Vanderbilt, D. Phys. ReV. B 1997, 56, 12848–12865.
26) Silvestrelli, P. L.; Mazari, N.; Vanderbilt, D.; Parrinello, M. Solid
1996, 50, 608–613.
(48) Eremets, M. I.; Hemley, R. J.; Mao, H. K.; Gregoryanz, E. A.
Nature 2001, 411, 170–174.
(49) Eremets, M. I.; Gavriliuk, A. G.; Trojan, I. A.; Dzivenko, D. A.;
Boehler, R. Nat. Mater. 2004, 3, 558–563.
State Commun. 1998, 107, 7–11.
(
27) Silvestrelli, P. L. Phys. ReV. B 1999, 59, 9703–9706.
(28) Ceppatelli, M.; Santoro, M.; Bini, R.; Schettino, V. J. Chem. Phys.
2
1
1
003, 118, 1499–1506.
29) Neumann, M. A.; Perrin, M. A. J. Phys. Chem. B 2005, 109, 15531–
5541.
(
(50) Santoro, M.; Gorelli, F. A.; Bini, R.; Ruocco, G.; Scandolo, S.;
Crichton, W. Nature 2006, 441, 857–860.
(
51) Goncharov, A. F.; Manaa, M. R.; Zaug, J. M.; Gee, R. H.; Fried,
L. E.; Montgomery, W. B. Phys. ReV. Lett. 2005, 94, 065505.
52) Ceppatelli, M.; Santoro, M.; Bini, R.; Schettino, V. J. Chem. Phys.
000, 113, 5991–6000.
53) Chelazzi, D.; Ceppatelli, M.; Santoro, M.; Bini, R.; Schettino, V.
J. Phys. Chem. B 2005, 109, 21658–21663.
54) Citroni, M.; Ceppatelli, M.; Bini, R.; Schettino, V. J. Chem. Phys.
003, 118, 1815–1820.
55) Ciabini, L.; Santoro, M.; Bini, R.; Schettino, V. J. Chem. Phys.
(
30) Albrecht, M.; Rice, C. A.; Suhm, M. A. J. Phys. Chem. A 2008,
12, 7530–7542.
(
(
31) Gaigeot, M. P.; Vuilleumier, R.; Sprik, M.; Borgis, D. J. Chem.
Theory Comput. 2005, 1, 772–789.
32) Allnat, A. R.; Jacobs, P. W. M. Can. J. Chem. 1968, 46, 111. Jacobs,
P. W. M. J. Phys. Chem. B 1997, 101, 10086–10093.
33) Walker, I. C.; Fluendy, M. A. D. Int. J. Mass Spectrom. 2001, 205,
71–182.
2
(
(
(
(
2
2
1
1
1
1
(
(
34) Flicker, W. M.; Mosher, O. A.; Kuppermann, A. J. Chem. Phys.
002, 116, 2928–2935.
(
(
980, 72, 2788–2794.
56) Cohen, M. D.; Schmidt, G. M. J. J. Chem. Soc. 1964, 1996–2000.
(
35) Reed, E. J.; Jannopulos, J. D.; Fried, L. E. Phys. ReV. B 2000, 62,
6500–16509.
36) Liu, H.; Zhao, J.; Wei, D.; Gong, Z. J. Chem. Phys. 2006, 124,
24501–124510.
57) Saladino, R.; Crestini, C.; Ciciriello, F.; Costanzo, G.; Di Mauro,
E. Chem. BiodiVersity 2007, 4, 694–720.
(
JP1035508