Laser photolysis of matrix-isolated methyl nitrate: Experimental and theoretical characterization of the infrared spectrum of imine peroxide (HNOO)

Article abstract of DOI:10.1021/ja9824223

Photodissociation of methyl nitrate (and its fully deuterated isotopomer) isolated in an argon matrix at 17 K has been investigated by 266 nm laser photolysis. Characterization of the product species has been performed by transmission Fourier transform infrared spectroscopy. There are two parallel initial reaction channels, one of which is formation of formaldehyde (H₂CO) and hydrogen nitryl (HNO₂). The second channel is formation of formaldehyde and a previously unreported species, imine peroxide (HNOO). Quantum chemical calculations of the structure, vibrational frequencies, and infrared band intensities have been performed at the B3LYP/6-311++G**, MP2(full)/6-311++G**, and QCISD/6-311++** levels of theory. The calculated vibrational frequency patterns allowed us to confirm the detection and infrared spectral assignments of imine peroxide for the first time.

Full text of DOI:10.1021/ja9824223

J. Am. Chem. Soc. 1998, 120, 12327-12333
12327
Laser Photolysis of Matrix-Isolated Methyl Nitrate: Experimental and
Theoretical Characterization of the Infrared Spectrum of Imine
Peroxide (HNOO)
Ping Ling, Alexander I. Boldyrev, Jack Simons, and Charles A. Wight*
Contribution from the Department of Chemistry, UniVersity of Utah, Salt Lake City, Utah 84112-0850
ReceiVed July 9, 1998
Abstract: Photodissociation of methyl nitrate (and its fully deuterated isotopomer) isolated in an argon matrix
at 17 K has been investigated by 266 nm laser photolysis. Characterization of the product species has been
performed by transmission Fourier transform infrared spectroscopy. There are two parallel initial reaction
channels, one of which is formation of formaldehyde (H2CO) and hydrogen nitryl (HNO2). The second channel
is formation of formaldehyde and a previously unreported species, imine peroxide (HNOO). Quantum chemical
calculations of the structure, vibrational frequencies, and infrared band intensities have been performed at the
B3LYP/6-311++G**, MP2(full)/6-311++G**, and QCISD/6-311++** levels of theory. The calculated
vibrational frequency patterns allowed us to confirm the detection and infrared spectral assignments of imine
peroxide for the first time.
Introduction
difficult to identify in the condensed phase. We therefore
initiated the current study in order to investigate the photo-
chemistry of the simplest organic nitrate ester, methyl nitrate,
at a high level of detail.
Because of their high oxygen content, nitrate esters of organic
compounds have been widely used as energetic materials. The
mechanism of thermal decomposition on the ground electronic
manifold of states has been previously investigated for several
Although there have been many previously reported studies
of the photochemistry of organic and inorganic nitro compounds,
1
-3
different types of nitrate esters.
However, little is known
1
0-14
nitroso compounds, and nitrites,
we were unable to find
about their photodissociation pathways from electronically
excited states. This subject is of current interest for energetic
materials because it has been postulated that electronic excitation
and predissociation of energetic materials may be important in
any studies of product branching or photochemical quantum
yields for nitrate esters. This is despite a growing interest in
the role of photochemistry of organic nitrate salts in atmospheric
1
5,16
12
chemistry.
Clark and Hester have reviewed much of the
the chemical mechanism of initiation and propagation of
literature on matrix isolation spectroscopic studies of nitrogen-
containing organic compounds, but nitrate esters were not
mentioned. We chose the rare gas matrix isolation technique
to study methyl nitrate because it provides a means of generating
and preserving sufficiently high concentrations of reactive
chemical intermediates for spectroscopic identification and
characterization. In fact, the photochemistry of nitromethane
was one of the pioneering works in the matrix isolation field
and was the first to experimentally identify the existence of
HNO. Since then, numerous species have been identified using
this technique. The objective of this paper is to report the
photochemistry of methyl nitrate isolated at low concentrations
in an argon matrix at 17 K. On the basis of an identification of
the reaction products and an analysis of their kinetic behavior,
we were able to determine the major initial photodissociation
channels under these experimental conditions.
detonation shock waves.4
-7
As more is learned about the
complex chemical and physical mechanisms of detonation,
through experimental studies and molecular dynamics calcula-
8
tions, the possible role of electronically excited states in the
process is expected to become clarified.
Recently, we reported an experimental study which revealed
that thermal and photochemical decompositions of poly(glycidyl
9
nitrate), or PGN, occur by distinctly different pathways. PGN
is a technologically important material that is used as an
energetic polymer binder in solid rocket propellant formulations.
However, the material and its chemistry are both quite complex,
involving formation of mixtures of several products that are
*
To whom correspondence should be addressed. E-mail: wight@
chemistry.chem.utah.edu.
(
1) Waring, C. E.; Krastine, G. J. Phys. Chem. 1970, 74, 999.
(2) Hiskey, M. A.; Brower, K. R.; Oxley, J. C. J. Phys. Chem. 1991, 95,
3
955.
(10) Brown, H. W.; Pimentel, G. C. J. Chem. Phys. 1958, 29, 883.
(11) Muller, R. P.; Russegger, P.; Huber, J. R. Chem. Phys. 1982, 70,
281.
(3) Chen, J. K.; Brill, T. B. Combust. Flame 1991, 85, 479.
4) Odiot, S. In Chemistry and Physics of Energetic Materials; Bulusu,
(
(12) Clark, R. J. H.; Hester, R. E. In Spectroscopy of Matrix Isolated
Species; John Wiley & Sons Ltd.: New York, 1989; pp 190-422 and many
references cited therein.
(13) Cheng, B. M.; Lee, J. W.; Lee, Y. P. J. Phys. Chem. 1991, 95,
2814.
(
7) Aduev, B. P.; Aluker, E. D.; Krechetov, A. G. Tech. Phys. Lett. 1996,
2, 236.
8) 8 Elert, M. L.; Deaven, D. M.; Brenner, D. W.; White, C. T. Phys.
ReV. B 1989, 39, 1453.
9) Ling, P.; Wight, C. A. J. Phys. Chem. B 1997, 101, 2126.
(14) Hjorth, J.; Ottobrini, G.; Restell, G. J. Phys. Chem. 1988, 92, 2669.
(15) Calvert, J. G.; Pitts, J. N., Jr. In Photochemistry; John Wiley: New
York, 1966; p 480.
(16) Barners, I.; Becker, K. H.; Starcke, J. J. Phys. Chem. 1991, 95,
9736.
2
(
(
1
0.1021/ja9824223 CCC: $15.00 © 1998 American Chemical Society
Published on Web 11/12/1998

12328 J. Am. Chem. Soc., Vol. 120, No. 47, 1998
Ling et al.
Experimental Section
Table 1. Photoproduct Infrared Bands (CH
66 nm)
3
ONO
2
:Ar ) 1:1500 at
2
Methyl nitrate and its fully deuterated isotopomer were prepared by
rel intensa
after 520 min
photolysis
possible
product
assignt
a
esterification of CH
3
OH (CD
3
OH) with nitric acid in the presence of
rel intens after
wavenumber
1
7
20 min photolysis
sulfuric acid as described previously. Argon matrix isolated methyl
nitrate was prepared by deposition of premixed methyl nitrate vapor
and argon gas onto a 17 K CsI window in a high-vacuum deposition
-1
(cm )
3
3
3
3
3
3
3
712.9
622.4
616.1
608.3
498.7
360.1
0.054
0
0.033
2.51
0.17
0.14
1.06
0.94
0.56
3.12
1.84
0.43
0.48
0.13
0.10
0.14
3.13
2.81
0.35
0.33
0.50
1.00
0.45
0.69
0.66
1.13
1.32
0.13
0.24
0.74
0.074
0.34
0.072
0.89
0.31
0.35
1.42
0.35
1.12
0.24
0.62
0.32
0.19
0.28
2
H O
O-H?
-
7
chamber. The base pressure of this chamber was maintained at 10
O-H?
Torr by a turbomolecular pump to avoid contamination of the samples
by impurity gases. The CsI window was cooled by a closed-cycle
helium refrigerator (CTI Cryogenics model 22), and the temperature
was maintained by a Lakeshore Cryotronics model 320 temperature
controller. Deposition rate was typically ∼10 (mol/cm )/min at the
sample window. After deposition, the sample temperature was quickly
raised to 30 K for about 1 min and then lowered to 17 K in order to
anneal the argon matrix to minimize multiple-site occupation of the
methyl nitrate molecules.
0
0
0
2.93
1.85
0
0.53
0
0.079
0.068
0.042
0.057
0
0
0
1.00
0
0
0.71
0
0
2
H O
HONO-trans
HONO-cis
HNOO
HNO2
HNO
287.7/3306
C
B
-
5
2
3137.1
3049.4
2831.4
2822.7
A
2
H CO
C-H?
C-H?
C-H?
2
2
2
2
1
1
1
1
1
731.3
715.1
344.5
151.9
876.4
762.4
747.8/1743.4
725.8
680.2
Methyl nitrate exhibits a weak n-π* absorption around 270 nm.
The photolysis was performed by using the fourth harmonic of a Nd:
YAG laser (Continuum model Surelite) at 266 nm. The pulse duration
was 7 ns, and a lens was used to expand the beam to approximately
CO
CO
2
NO
HC(dO)OH
HC(dO)NO?
H CO
2
HONO-trans
HONO-cis
HNO
2
HC(dO)NO?
HNO
2
2
3
.5 cm such that the average laser intensity was 10 mW/cm (at 10
Hz repetition rate) at the sample surface. The bulk temperature change
of the sample by absorption of the laser beam was negligible (<1 K).
Transmission FTIR spectra were recorded before photolysis and at
regular intervals during photolysis in order to monitor intensity changes
of the reactant and product peaks. Each spectrum was obtained using
a Mattson model RS/10000 FTIR spectrometer by averaging 32 scans
A
B
1625.4
1605.3/1603.2
1594.0/1591.4
1580.3/1575.5
1497.4
-
1
0
HNO
at 0.25 cm resolution.
1493.4
1381.6
1371.6
1310.3
1252.2
1183.7
1155.3
0.23
0.69
0.073
0.055
0.066
0.096
0.078
0
1.37
0.38
0
0.22
0
A
C
B
H
2
CO
HNOO
HNO
HONO-trans
2
H CO
Following the primary photolysis period at 266 nm, samples were
subjected to additional photolysis periods at 355 or 532 nm (third
and second harmonics of the same laser) in order to aid the identification
of reaction products.
The absorption coefficient of neat liquid methyl nitrate at 266 nm
was measured by a Hewlett-Packard UV-visible diode array spec-
trometer.
2
A
C-O?
C-O?
C-O?
HNO
2
HNOO
HONO-cis
HNO2
HONO-trans
HNOO
HNOO
HONO-trans,cis?
1116.2
9
8
8
23.6
43.2
68.7
B
C
Experimental Results
Photolysis of matrix-isolated methyl nitrate was conducted
using samples prepared at mole ratios in the range 1:1000 to
837.5
813.2
7
670.1
6
B
1:5000 (CH3ONO2:Ar). The IR spectra of matrix-isolated
90.7
0.33
0.19
0
C
C
methyl nitrate and its normal-coordinate analysis have been
studied by Lannon et al.18 At this dilution, the guest molecules
28.1
are well isolated and the concentration of dimers and higher
oligomers is negligible. The IR spectra show that CH3ONO2
molecules occupy multiple sites in the argon matrix, as indicated
by closely spaced multiple absorption peaks for many modes
in the spectrum.
a
CO at 1725.8 cm-1
.
Normalized to the intensity of H
2
establishing correlations among the bands and assignments to
the various products.
After 266 nm photolysis, numerous new bands appear. The
spectra show that reaction products also occupy multiple sites
in the argon lattice. The positions of these product bands of
photolyzed CH3ONO2/Ar and CD3ONO2/Ar are listed in Tables
(3) We compared the band positions with literature values
where available and otherwise with vibrational frequencies
determined using ab initio quantum chemical calculations of
suspected products. We performed experiments on two isoto-
pomers of methyl nitrate, CH3ONO2 and CD3ONO2, to check
the band assignments with the calculated values.
Tables 1 and 2 list the photolysis product infrared band
intensities from CH3ONO2 and CD3ONO2 after 20 and 520 min
exposure to 266 nm radiation. The relative intensity distribu-
tions of product bands of photolyzed CH3ONO2 are also shown
in Figure 1. Note that the spectra of the initial products (after
1
and 2, respectively. For simplicity, we have listed only the
absorption bands of dominant sites; however, if absorption bands
corresponding to two sites were comparable, we listed both.
To assign the new bands to corresponding products, we
employed an approach based on three criteria:
(1) The growth in intensity of all peaks during 266 nm
photolysis was analyzed to determine possible correlations in
the band intensities and to determine whether each peak exhibits
an induction period.
2
0 min photolysis) are fairly simple compared with spectra
recorded after a long periods of irradiation. The initial product
bands, which are labeled A, B, or C, all exhibit the same
qualitative behavior during 266 nm photolysis, as illustrated for
three representative bands in Figure 2. Note that none of the
bands of the initial products exhibit a noticeable induction
period; that is, the bands have their maximum rate of growth at
t ) 0.
(2) After 266 nm photolysis for several hours, we used
secondary photolysis periods at 355 and/or 532 nm to selectively
destroy the reaction products. Because different products
respond differently to these wavelengths, this technique aids in
(
17) Johnson, J. R. In Organic Syntheses; John Wiley Ltd.: New York,
939; Vol. XIX, p 64.
18) Lannon, J. A.; Harris, L. E.; Verderame, F. D.; Thomas, W. G.;
Lucia, E. A.; Koniers. S. J. Mol. Spectrosc. 1974, 50, 68.
1
In contrast, several secondary photolysis products were
observed, including H2O, CO2, CO, HNO, cis- and trans-
(

Laser Photolysis of Matrix-Isolated Methyl Nitrate
J. Am. Chem. Soc., Vol. 120, No. 47, 1998 12329
Table 2. Photoproduct Infrared Bands (CD
3
ONO
2
:Ar ) 1:1500 at
2
66 nm)
rel intensa
after 518 min
photolysis
possible
product
assignt
a
rel intens after
wavenumber
4
0 min photolysis
-
1
(cm )
2
766.5
756.3
739.6
637.0
443.5
342.2
334.1
315.3
259.4
214.8
152.7
139.9
118.2
094.5
875.7
719.0/1714.4
709.5/1705.5
704.3/1700.1
695.7
679.9
671.5
0
0.12
0.77
0.16
0.23
2.02
0.88
0.32
0.32
0.16
0.30
0.65
0.087
0.076
0.64
0.068
0.13
0.14
0.28
0.019
1.00
0.031
0.61
0.022
0.078
0.28
0.11
0.50
0.075
0.87
0.056
0.12
0.10
1.22
0.095
0.33
0.11
0.28
0.012
0.29
D-O?
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0.016
0
0
1.65
0.021
0
0.33
0
2
D O
D-O?
D O
2
C
B
DNOO
CO
DONO-trans
2
DNO
DNO
2
0
DONO-cis
CO
0.019
0
0
0.57
0
0
0
0
0
1.00
0
0.53
0
0
0
0
0.042
0.092
0.79
0
Figure 2. Kinetic behavior of the intensities of band sets A (represented
C-D?
-1
-1
by 1725.8 cm ), B (represented by 1605/1603 cm ), C (represented
C-D?
-1
-1
by 1381.6 cm ), and CH
3
ONO
2
(represented by 1654 cm ) during
A
2
D CO
2
66 nm photolysis. For clarity of presentation, the intensity of the CH -
3
NO
1
2
ONO band has been multiplied by /15.
DC(dO)OD
DC(dO)NO?
DC(dO)NO?
?
A
B
D
2
CO
DONO-trans
DNO
578.9/1575.8
569.5/1566.6
561.7/1558.4
541.4
196.4
179.3-1172.9
112.7
2
DC(dO)NO?
DC(dO)NO?
DNO
C-O?
C-O?
B
C
DNO
2
081.1/1079.2
036.8
DNOO
DONO-trans
DONO-cis?
?
9
9
8
8
8
6
6
6
5
92.1
71.3
95.5/892.2
34.3/832.2
23.4/822.0
58.7
36.3
18.9
88.1/585.7
0
0.014
0.97
0.11
0.24
0.096
0
B
B
C
C
DNO
DNO
2
2
Figure 3. Kinetic behavior of the intensities of some high-order
reaction products under 266 nm photolysis.
DNOO
DNOO
DONO-cis
DONO-trans
DNOO
0
0.33
C
a
CO at 1679.9 cm-1
.
Normalized to the intensity of D
2
Figure 4. Kinetic behavior of the intensities of band sets A, B, and C
-1
(
represented by bands observed at 1725.8, 1605/1603, and 1381.6 cm ,
respectively) during 355 nm photolysis. The sample was subjected to
66 nm photolysis for 520 min prior to 355 nm irradiation.
2
That is to say, the rates of production of these species at t ) 0
is essentially zero.
The infrared bands of the initial products can be ascribed to
three different species, which we have labeled A, B, and C, on
the basis of differences in their kinetic behavior upon secondary
photolysis at 355 nm (Figure 4). The bands associated with
initial product A can be readily assigned to formaldehyde, on
the basis of its strong characteristic CdO stretching frequency
Figure 1. IR spectra of product bands of CH
photolysis: (a) obtained after 20 min of photolysis; (b) obtained after
20 min of photolysis. The CH ONO bands have been subtracted from
the spectra.
3 2
ONO after 266 nm
5
3
2
-1
-1
HONO, HCOOH, and NO. The vibrational frequencies of all
of these species are well-known, and the products were easily
identified. All of the bands associated with these secondary
photolysis products exhibit a noticeable induction period at the
beginning of the photolysis period, as illustrated in Figure 3.
at 1725.8 cm for H2CO (or 1679.9 cm for D2CO). Matrix-
isolated H2CO (D2CO) has a characteristic CdO stretching
1
1
-1
frequency at 1742 (1697) cm . However, we expect that a
red shift should occur when the molecule exists as a van der
Waals complex with the partner photoproduct. For example,1
0,11

12330 J. Am. Chem. Soc., Vol. 120, No. 47, 1998
Ling et al.
the CdO stretching frequency of the complex CH2O‚HNO
CD2O‚DNO) is red-shifted relative to that of the isolated
(
-1
molecule by about 16 (12) cm . Therefore, the 1725.8 (1679.9)
-1
cm band observed in this experiment is assigned to a complex,
one partner of which is formaldehyde. Other characteristic
bands of formaldehyde were also observed and are listed in
Tables 1 and 2.
The infrared bands associated with initial product B are
assigned to hydrogen nitryl (C2V-HNO2) on the basis of a recent
1
9
IR assignment of this structure by Koch and Sodeau. They
generated this species by 248 nm photodissociation of nitric
acid isolated in an argon matrix. The band positions and relative
intensities of set B in our experiment are in good agreement
with their data, except for modest discrepancies in the positions
of the nitrogen out-of-plane bending mode and NO2 bending
mode. The most likely reason for these discrepancies is that
the hydrogen nitryl in their experiment was formed in close
proximity to an oxygen atom, whereas in our experiment the
partner product was a formaldehyde molecule.
Figure 5. Plot of ln(A/A
0
) vs σJ (dimensionless product of the methyl
nitrate absorption cross section and the cumulative laser fluence) for
determination of the photodestruction quantum yield Φ of methyl nitrate
from eq 2.
CO2, NO, HCOOH,20 H2O, and tentatively assigned HCONO.
It is important to note that HONO is not an initial product,
judged from its kinetic behavior shown in Figure 3.
For completeness, we determined the overall quantum yield
for photodestruction of CH3ONO2 at 266 nm. This was
accomplished by recording the integrated absorption intensity,
On the basis of the assignments of species A and B from
previous argon matrix studies, it is safe to conclude that one of
the photodissociation channels of methyl nitrate is generation
of H2CO and hydrogen nitryl:
-
1
A, of the nitro stretching band of methyl nitrate at 1654 cm
2
as a function of cumulative laser fluence, J (photons/cm ). For
optically thin samples, the quantum yield can be found from
the relationship
CH ONO + hν (266 nm) f
3
2
A
H CO + HNO (hydrogen nitryl) (1)
ln
) -ΦσJ
(2)
2
2
( )
A
0
There is no ready assignment for bands associated with initial
product C. Because these bands all appear to belong to a single
species and because the number and positions of the bands are
inconsistent with any known isomer of methyl nitrate, we
postulated that the two reaction channels likely had one product
in common. Therefore, it became necessary to determine the
quantum yield for formation of H2CO.
where Φ is the quantum yield and σ is the 266 nm absorption
cross section of CH3ONO2. The absorptivity of CH3ONO2 at
66 nm is too small to be measured directly in the argon matrix,
so it was estimated from the UV absorbance of neat methyl
nitrate, which was measured to be (0.023 ( 0.0016)/µm (base
0) at 266 nm. The density of methyl nitrate is 1.15 g/cm ,
2
3
1
-20
2
giving a molecular absorption cross section of 5.9 × 10 cm
We began by measuring the integrated infrared absorption
coefficient of the antisymmetric stretch mode of the nitro group
(
base e). A plot of ln(A/A0) was found to be linear in laser
fluence J, as shown in Figure 5. The slope of this plot yields
an estimated quantum yield Φ for photodestruction of methyl
nitrate to be 0.069 ( 0.014 at 266 nm, on the basis of an average
of five experiments.
-
1
of CH3ONO2 at 1654 cm and the CdO stretching mode of
H2CO at 1726 cm- by depositing known quantities of these
stable species in argon. The method for determining the
absolute quantity of each molecule in the argon matrix has been
described previously.9 The integrated absorption coefficients
for CH3ONO2 and CH2O are 150 and 28 km/mol (base e),
respectively.
1
Quantum Chemical Calculations
To aid the identification of the initial product C, whose
infrared bands have been quantified and tabulated in Tables 1
and 2, we performed ab initio quantum mechanical calculations
on isomers having the empirical formula HNO2. The five
isomers of nitrous acid are shown as follows:
On the basis of five different 266 nm photolysis experiments
on samples of CH3ONO2/Ar in the concentration range 1:1000
to 1:5000, we found that, at the initial stage of photolysis,
destruction of one CH3ONO2 molecule corresponds to genera-
tion of 1.07 ( 0.16 H2CO molecules. This result shows that
H2CO is always produced as an initial photolysis product and
that products B and C must be different isomers of HNO2 that
are formed in parallel competing reaction channels.
A control experiment was performed to rule out the possibility
that oxygen contamination might be contributing to production
of the unidentified product C. Oxygen was deliberately
introduced into a sample (CH3ONO2:O2:Ar ) 1:20:1500), and
photolysis of this sample at 266 nm gave initial products for
which the band positions and intensities were essentially
unchanged from those of the experiments without added oxygen.
It was judged unlikely that any other chemically active
contaminant could have been introduced without noticing it in
the infrared spectra obtained prior to photolysis.
Both cis and trans forms of nitrous acid (isomer 1) are well-
2
1-24
known
products in our experiments. Hydrogen nitryl (isomer 2) was
and were determined to be secondary photolysis
1
9
reported by Koch and Sodeau, and we have assigned the
product bands labeled B to this structure (vide supra). To our
knowledge, isomers 3-5 have not been observed experimentally
nor have their vibrational frequencies been calculated theoreti-
cally. The potential energies of the five isomers have been
(
20) Redington, R. L. J. Mol. Spectrosc. 1977, 65, 171.
(21) John, L. H.; Badger, R. M. J. Chem. Phys. 1951, 19, 1599.
Secondary or higher order products (shown in Tables 1 and
(22) McGraw, G. E.; Bernitt, D. L.; Hisatsane, I. C. J. Chem. Phys. 1966,
45, 1392.
23) McDonald, P. A.; Shirk, J. S. J. Chem. Phys. 1982, 77, 2355.
(24) Hall, R. J.; Pimentel, G. C. J. Chem. Phys. 1963, 38, 1889.
2
and Figure 3) are HONO (both trans and cis), HNO, CO,
(
(19) Koch, T. G.; Sodeau, J. R. J. Phys. Chem. 1995, 99, 10824.

Laser Photolysis of Matrix-Isolated Methyl Nitrate
J. Am. Chem. Soc., Vol. 120, No. 47, 1998 12331
experimental findings. These results give some idea of the
quality of the vibrational frequencies that can be expected using
the basis sets and electron correlation methods that we have
employed.
For isomers 2 and 4, where simple chemical formulas can
be drawn and where single-configuration-based correlation
methods are expected to perform well, our ab initio calculations
produce nearly identical geometries among the three methods.
As a result, we believe our calculated results to be reliable for
isomers 1, 2, and 4. However, when isomers 3 and 5 were
examined, we found significant variation in the computed
geometries and vibrational frequencies as the methods used to
treat electron correlation varied.
Both the trans and cis (singlet) isomers of 5 were found to
be minima at the MP2(full)/6-311++G** and B3LYP/6-
311++G** levels of theory, with the trans configuration being
more stable than the corresponding cis. However, the optimized
geometries are quite different at the two levels of theory (one
of the largest ranges of geometry being 0.017 nm in the O-O
distance). At the QCISD/6-311++G** level of theory, the cis
isomer is not predicted to be a minimum, but a first-order saddle
point with an optimal O-O distance of 0.211 nm, while the
trans isomer dissociates spontaneously into NO + OH (at large
O-O distances, the QCI displayed convergence difficulties).
The failure of the three correlation methods to agree on isomer
Figure 6. Optimized geometries for the HNO
-311++G** level of theory.
2
isomers at the QCISD/
6
theoretically studied by Fueno et al.25 whose calculations showed
that the energies increase in the order 1-2-5-3-4. They also
proposed that imine peroxide, isomer 3, might be formed by
reaction of ground-state NH with oxygen. Our calculations of
the energies, structures, and vibrational frequencies of the HNO2
isomers were undertaken with the intention of identifying
product bands labeled C as being due to one of the isomers
5
suggests that this isomer requires a multiconfigurational
treatment. One aspect of the structure for which all levels of
theory agree is the O-O distance. In the both cis and trans
isomers, this bond length is ca. 0.2 nm, which is considerably
longer than twice the covalent radius of an oxygen atom (0.074
3-5.
Quantum chemical calculations were performed on all five
2
6
nm). Therefore, it is reasonable to characterize isomer 5 as a
weakly bound complex of two radicals, HO + NO, rather than
a strongly bound molecule. This result is in agreement with
the MC-SCF and MRD-CI calculations performed by Fueno et
isomers of nitrous acid (isomer 1) employing analytical gradients
with polarized split-valence basis sets (6-311++G**) at three
sophisticated ab initio levels of theory, including the MP2 (full)
level (meaning all electrons were included in the correlation
calculations), the nonlocal density functional B3LYP level of
theory, and the quadratic configuration including singles and
doubles with approximate triples QCISD(T) method. All of
these calculations were carried out with the Gaussian 94
2
5
al., who found that HOON is extremely unstable and that the
barrier to decompose to HO + NO is negligible. We therefore
conclude that unknown isomer C found in the current experi-
ments cannot be the HOON isomer 5. Thus, it remains to
examine isomer 3 as a candidate for the unknown C.
3
1
program. The calculated optimized geometries are shown in
Figure 6, and the calculated (unscaled) harmonic frequencies
are collected in Tables 3-6 for isomers 1-4.
Both trans and cis singlet isomers of 3 were found to be
minima at the MP2(full)/6-311++G**, B3LYP/6-311++G**,
and QCISD/6-311++G+** levels of theory, with the cis
configuration being more stable than the corresponding trans.
However, the calculated optimized geometries and harmonic
frequencies again are different (although the differences are less
pronounced than those for isomer 5) among the various levels
of theory, again suggesting that structure 3 has appreciable
multiconfigurational character in its wave function. To further
consider the multiconfigurational nature of isomer 3’s electronic
state, we performed CASSCF(4,4)/6-31G**, CASSCF(6,6)/6-
31G**, and CASSCF(8,8)/6-31G** calculations for both iso-
mers of 3. We found the optimized geometries and harmonic
frequencies to vary considerably among the CASSCF(4,4)/6-
31G**, CASSCF(6,6)/6-31G**, and CASSCF(8,8)/6-31G**
data sets and to differ somewhat from those found at the QCISD/
For isomer 1, the optimized bond lengths and valence angles
are in excellent agreement with the experimentally known
structures,32 and our calculated harmonic frequencies shown in
Table 3 for both cis and trans nitrous acid are nearly the same
at all three levels of theory and agree well with the present
(
25) Fueno, T.; Yokoyama, K.; Takane, S. Theor. Chim. Acta 1992, 82,
99.
26) Benson, S. W. Thermochemical Kinetics; Wiley-Interscience: New
York, 1976; p 302.
27) Ho, P.; Bamford, D. J.; Buss, R. L.; Lee, Y. T. Moore, C. B. J.
Chem. Phys. 1982, 76, 3630.
2
(
(
(
(
28) Vasudev, R.; Zare, R. Z.; Dixon, R. J. Chem. Phys. 1983, 96, 399.
29) Hido, O. In Photochemistry of Small Molecules; John Wiley & Sons
Ltd.: New York, 1978; p 355.
30) Gardiner, W. C., Jr. In Combustion Chemistry; Springer-Verlag:
New York, 1984; p 167.
31) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.;
(
(
6
-311++G** level. QCISD theory is expected to perform
Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T. A.; Peterson, G.
A.; Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski,
V. G.; Ortiz, J. V.; Foresman, J. B.; Cioslowski, J.; Stefanov, B. B.;
Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.;
Wong, M. W.; Anders, J. L.; Replogle, E. S.; Gomperts, R.; Martin, R. L.;
Fox, D. J.; Binkley, J. S.; DeFrees, D. J.; Baker, J.; Stewart, J. J. P.; Head-
Gordon, M.; Gonzalez, C.; Pople J. A. Gaussian 94, Revision A.1; Gaussian
Inc.: Pittsburgh, PA, 1995).
reasonably well when multiconfigurational character is not
extreme which is certainly the case for isomer 3, where we found
the Hartree-Fock configurations to dominate (CHF ) 0.9014
for the trans isomer and CHF ) 0.9077 for the cis isomer) among
the 1764 configurations in the CASSCF(8,8) expansions. We
therefore believe that our results at the QCISD/6-311++G**
level of theory should be reliable.
(32) Fennigan, D. J.; Cox, A. P.; Brittain, A. H.; Smith, J. G. J. Chem.
Soc., Faraday Trans. 2 1972, 68, 548.

12332 J. Am. Chem. Soc., Vol. 120, No. 47, 1998
Ling et al.
Table 3. Vibrational Frequencies (cm-1) of Isomer 1, HONO (Nitrous Acid)
a
HONO/DONO-trans node
ref 21
ref 22
ref 23
ref 24
this expt
this calcn (singlet)
ν
1
ν
2
ν
3
ν
4
ν
5
ν
6
(a′)
(a′)
(a′)
(a′)
(a′)
(a′′)
3590
1696
1260
794
3512
1635
1421
801
3558/2620
1684/1682
1298/1030
815/769
625/618
583/444
3498/2334
1680/1671
1310/1037
813/...
3838/2795
1781/1773
1342/1069
858/802
658/645
543/410
1699
1265
791
593
702
628?/636
a
HONO/DONO-cis node
ref 21
ref 22
ref 23
ref 24
this expt
this calcn (singlet)
1
ν (a′)
2
ν (a′)
3
ν (a′)
4
ν (a′)
5
ν (a′)
6
ν (a′′)
3426
3410/2518
1633/1612
3360/2214
1625/- - -
.../992?
868/...
628?/618
3688/2683
1724/1711
1361/1115
920/889
664/607
665/508
1640
1261
853
1611
1292
856
857
721
865/828
608
658/522
a
At the QCISD/6-311+G* level of theory.
Table 4. Vibrational Frequencies (cm-1) of Isomer 2, HNO
Table 6. IR Absorption Frequencies (cm-1) of Isomer 3, HNOO
(Imine Peroxide)
2
(Hydrogen Nitryl)
HNO /DNO
2
2
this expt
(product B)
this calcna
(singlet)
this calcna
-
1
(cm
)
ref 19
trans-HNOO/
cis-HNOO/
possible assignt of
IR absorption of set C
ν
1
2
3
4
5
6
(a
1
1
1
)
)
)
3289/2437
1376/...
792/789
1016/890
1600/1595
3137/2315
1372/...
838/834
924/896
1605/1579
.../1113
3278/2403
1418/1396
800/792
1078/911
1665/1654
1561/1148
mode
DNOO
DNOO
ν
(a
(a
(b
(b
(b
ν
1
ν
2
ν
3
ν
4
ν
5
(a′)
(a′)
(a′)
(a′)
(a′)
3377/2469
1540/1391
1391/1199
841/816
660/635
794/599
3270/2387
1498/1429
1425/1177
899/894
632/577
875/699
3287/2443
(notobs.)
1382/1081
843/823
670/659 or 670/588
791/588 or 791/659
ν
ν
ν
ν
1
2
2
)
)
)
a
ν6(a′′)
At the QCISD/6-311+G* level of theory.
a
Table 5. Vibrational Frequencies (cm-1) of Isomer 4, HNO
At the QCISD/6-311+G* level of theory.
2
(Cyclic Form)
HNO
2
/DNO
2
this calculationa
(singlet)
and 6, respectively. We found that the set of product bands C
could not be matched to the calculated vibrational frequencies
of isomer 4, HNO2 (cyclic form), but can be fit well to either
the trans or cis forms of isomer 3, HNOO, although the
agreement with the trans form is somewhat better. Unfortu-
nately, we were unable to positively identify any bands in the
experiments corresponding to the ν2 vibration of HNOO (see
Table 6). This is possibly due to a weak absorption cross section
for this vibrational mode. We should stress that the calculated
frequencies are for isolated gas-phase HNOO, whereas the
observed frequency is for HNOO complexed by a cage partner
H CO molecule. The NH stretching frequency of CH O‚HNO
(cyclic) node
ν
1
ν
2
ν
3
ν
4
ν
5
ν
6
(a′)
(a′)
(a′)
(a′)
(a′′)
(a′′)
3352/2453
1463/1202
1203/1087
823/822
1259/938
812/812
a
At the QCISD/6-311+G* level of theory.
At our highest QCISD(T)/6-311++G(2df,2pd) level of
theory, the cis configuration of isomer 3 was found to be more
stable by 13.0 kJ/mol than the trans configuration. We
attempted to calculate the barrier connecting the trans and cis
configurations at the QCISD/6-311++G** level of theory.
However, when the HNOO dihedral angle approached 60°, we
met QCI convergence problems, and therefore, we were unable
determine the value of the trans-cis isomerization barrier,
although we can say that the barrier is at least 57 kJ/mol, the
last geometry where convergence was realized.
2
2
11
-1
is blue-shifted relative to that of HNO by 60-90 cm .
Because the calculations do not include effects of anharmonicity
or complexation, the calculated NH frequency could be over-
-1
estimated by 100 cm . Therefore, the agreement between
calculation and experiment is quite reasonable. As mentioned
earlier, our calculations showed that cis form is more stable
(about 13 kJ/mol below trans), and the barrier between cis and
trans is very high (>57 kJ/mol). The high barrier to intercon-
version means that any trans isomer will remain trans at the
temperatures of the argon matrix experiments. It is therefore
somewhat surprising that we observe only one of the two
isomers. Unfortunately, there is not enough difference in the
patterns of the calculated vibrational frequencies to determine
with certainty which of the two conformers is observed.
Discussion
1. Identification of Product C. Good agreement was found
between the experimentally observed and readily identified
infrared band positions and the calculated vibrational frequencies
for isomer 1, nitrous acid (Table 3), and for isomer 2, hydrogen
nitryl (Table 4). This gave us confidence that the quality of
the calculations was sufficient for assigning the structure of
product C and that any perturbations in frequencies due to the
argon matrix environment and the presence of a formaldehyde
reaction product would not be so severe as to make the
assignment ambiguous.
2
. Initial Photodissociation Channels of Methyl Nitrate
under 266 nm Radiation. On the basis of the infrared band
assignments and the measured quantum yield of CH2O, we have
demonstrated that photodissociation of methyl nitrate in an argon
matrix proceeds by two parallel channels, each of which forms
a thermodynamically unstable isomer of nitrous acid:
The calculated vibrational frequencies of isomer 4, HNO2
(cyclic form), and of isomer 3, HNOO, are listed in Tables 5

Laser Photolysis of Matrix-Isolated Methyl Nitrate
J. Am. Chem. Soc., Vol. 120, No. 47, 1998 12333
CH ONO + hν (266 nm) f
of CH2O could lead to photooxidation of the latter to HCOOH,
which is observed in this experiment. Reaction of H and NO
could result in HNO, which is also an observed secondary
product. Another high-order reaction product is probably
HCONO, which could be formed by reaction of HCO with NO.
CO2 can be formed via oxidation of CO by OH at very low
3
2
H CO + HNO (hydrogen nitryl) (1)
2
2
CH ONO + hν (266 nm) f
3
2
H CO + HNOO (imine peroxide) (3)
2
2
9,30
temperature.
In principle, it should be possible to determine the branching
ratio between the two channels by comparing the integrated band
intensities of products B and C corrected for differences in the
calculated absorption coefficients (Tables 4 and 6). However,
we found that the uncertainties in the calculated band intensities
Conclusion
There are two initial photodissociation channels of matrix-
isolated methyl nitrate under 266 nm radiation. One is H2CO
(as judged by comparison of experimental and calculated
+
HNO2 (hydrogen nitryl), and the other is H2CO + HNOO
intensity distributions for each isomer, nitrous acid, hydrogen
nitryl, and imine peroxide) were too large to make a quantitative
evaluation of the branching ratio. But, on a qualitative basis,
it appears that the contributions of the two channels (reactions
(
imine peroxide). Quantum chemical calculations of the
structures and vibrational frequencies of five isomers of nitrous
acid are presented. The ab initio frequencies were used to assign
experimentally observed compound C to HNOO (imine per-
oxide). This is the first report of the formation, detection,
characterization, and vibrational spectrum of imine peroxide.
1
and 2) are comparable.
HONO (both trans and cis) is not an initial product of 266
nm photodissociation of methyl nitrate (from their dynamic
behavior shown in Figure 3). It is likely the secondary
photochemical product of HNO2 (hydrogen nitryl) or HNOO.
Other products might be explained on the basis of the two
initial channels. Photodissociation of H2CO produces H, HCO,
Acknowledgment. The experimental part of this research
was supported by the Office of Naval Research, Program Officer
Dr. Richard S. Miller, under Contract N-00014-95-1339, and
the theoretical part was supported by NSF Grant CHE-9618904.
and CO.2 Photodissociation of HONO results in OH and NO.
7
28
Photodissociation of any of the HNO2 isomers in the presence
JA9824223