Photochemistry of chloropicrin in cryogenic matrices

Article abstract of DOI:10.1016/S0009-2614(02)01495-1

The photolysis of chloropicrin (CCl₃NO₂) was investigated in Ar and N₂ cryogenic matrices. The extent of reaction was monitored using FT-IR spectroscopy. Phosgene and nitrosyl chloride were the observed photoproducts at all wavelengths investigated (220, 251, 313, 365, and 405 nm). When the photolysis was performed with 220, 251, or 313 nm light, two additional bands were also observed. These bands have been assigned to CCl₃ONO. Chloropicrin was also photolyzed in the presence of O₂ and ¹⁸O₂. ¹⁸O-labeled photoproducts were not detected in cryogenic matrices.

Full text of DOI:10.1016/S0009-2614(02)01495-1

Chemical Physics Letters 365 (2002) 473–479
www.elsevier.com/locate/cplett
Photochemistry of chloropicrin in cryogenic matrices
a,
a
b
Elisabeth A. Wade *, Kristina E. Reak , Bradley F. Parsons ,
c
Thomas P. Clemes , Karen A. Singmaster
c
a
Department of Chemistry and Physics, Mills College, 5000 MacArthur Blvd.,
Oakland, CA 94613, USA
b
Combustion Research Facility, Sandia National Laboratories, Livermore, CA, USA
c
Department of Chemistry, San Jose State University, San Jose, CA, USA
Received 19 July 2002; in final form 27 August 2002
Abstract
The photolysis of chloropicrin (CCl3NO2) was investigated in Ar and N2 cryogenic matrices. The extent of reaction
was monitored using FT-IR spectroscopy. Phosgene and nitrosyl chloride were the observed photoproducts at all
wavelengths investigated (220, 251, 313, 365, and 405 nm). When the photolysis was performed with 220, 251, or 313
nm light, two additional bands were also observed. These bands have been assigned to CCl3ONO. Chloropicrin was
also photolyzed in the presence of O2 and 18O2. 18O-labeled photoproducts were not detected in cryogenic matrices.
Ó 2002 Elsevier Science B.V. All rights reserved.
1. Introduction
sequent evaporation from the soil [3,4]. Once in
the troposphere, chemicals are removed by some
combination of photodissociation, chemical re-
moval (e.g., reaction with trace species such as
OH, O3, or NO3), and physical removal through
dry deposition or rainout [5,6]. For most volatile
organic compounds, chemical removal is the most
important pathway [7]. Chloropicrin is unusual in
that direct photolysis appears to be the only sig-
nificant atmospheric removal pathway [8–10].
The ultraviolet absorption spectrum of chloro-
picrin has two bands, a weak band centered at 275
nm and a strong band centered at 200 nm [11].
Only the weak band is of interest to photochem-
istry in the troposphere, since the actinic solar
spectrum has little intensity below 295 nm. The
spectrum of chloropicrin is very similar to that of
Chloropicrin (trichloronitromethane, CCl3NO2)
is commonly used in agriculture as a soil fumigant,
either alone or in conjunction with other fumi-
gants such as methyl bromide. In 1990, 1610
metric tons of chloropicrin were used in California
alone, particularly in the strawberry industry [1].
In the future, chloropicrin is likely to become more
widely used as methyl bromide, the most widely
used soil fumigant, is to be phased out by 2005 [2].
Chloropicrin has a high-vapor pressure, 23.8
torr at 298 K, and can therefore enter the tropo-
sphere, either during application or through sub-
* Corresponding author. Fax: +1-510-430-3314.
E-mail address: ewade@mills.edu (E.A. Wade).
0009-2614/02/$ - see front matter Ó 2002 Elsevier Science B.V. All rights reserved.
PII: S0009-2614(02)01495-1

474
E.A. Wade et al. / Chemical Physics Letters 365 (2002) 473–479
CH3NO2 ! CH2O þ HNO
its unchlorinated equivalent, nitromethane, so, as
in nitromethane, the weak band should be a
pꢀ n transition, involving promotion of a non-
bonding electron of one of the O atoms [12].
Moilanen et al. photolyzed chloropicrin in the
gas phase using a 275-W RS Sunlamp, and found
that nitrosyl chloride (NOCl) and phosgene
(Cl2CO) were produced when chloropicrin was
photolyzed in air [13]. They observed no reaction
in nitrogen. Additionally, when the reaction was
performed with 18O2, all of the photoproducts
were found to be labeled [13]. Based on these re-
sults, they proposed that chloropicrin reacts with
oxygen to form a trioxalone intermediate that then
dissociates to produce NOCl and Cl2CO. They
further suggested that O2 catalyzes this process
[13].
CH3NO2 ! CH3NO þ O
Since, the spectroscopy of chloropicrin is compa-
rable to that of nitromethane, it is reasonable to
assume that its photochemistry may be compara-
ble as well.
This study focuses on the photodissociation of
chloropicrin, with and without oxygen present. In
cryogenic matrices, the primary photoproducts
can be detected, since these primary photoprod-
ucts will not be able to travel and undergo sec-
ondary reactions. In addition, intermediates,
which are too short lived to be detected in the gas
phase, may be trapped in cryogenic matrices. In
particular, we hoped to be able to trap the tri-
oxolane intermediate suggested by Moilanen et al.
[13].
2. Experimental
Chloropicrin (98%) was obtained from Fluka
Chemika, and purified by several freeze-pump-
thaw cycles. The primary contaminant is believed
to be dissolved Cl2, due to photodissociation of
chloropicrin [11]. This Cl2 appears as a yellow tint
that disappears after freezing and pumping out the
sample. O2 (99.99%) was obtained from Matheson
Gas Products, and 18O2 (99.2% 18O) was obtained
from Isotech Incorporated. Mixtures of chloro-
picrin/Ar (1/200), chloropicrin/N2 (1/200), chloro-
picrin/O2/Ar (1/2/200), and chloropicrin/18O2=Ar
(1/2/200 and 1/50/150) were deposited at a flow
rate of 1.0–2.0 mmol/h onto a cold CsI window.
The window was cooled to 11 K by an Air Prod-
ucts two stage closed cycle helium refrigerator
(APD Cryogenics HC-2). Deposition and photol-
ysis were carried out at 11 K. A medium pressure
mercury arc lamp (Bausch and Lomb 200 W Hg or
Oriel 66033 200 W Hg) was used as the photolysis
source. Assorted Oriel and Corning filters were
used to limit the lamp photolysis wavelengths.
Infrared spectra were recorded from 400 to
4000 cmꢁ1 at 11 K by an FT-IR spectrometer
(Mattson Nova Cygni II) at a resolution of
0:50 cmꢁ1 (200 scans) with a frequency accuracy of
More recently, Carter et al. [10] have studied
the reaction of chloropicrin with various organic
pollutants, in order to determine if it is a signifi-
cant generator of tropospheric ozone. They found
that the quantum yield for production of NOx and
Cl atoms from chloropicrin in the 300–360 nm
region was 87 Æ 13% [10]. They used their data to
test kinetic models of several mechanisms for
chloropicrin photolysis, and found that their re-
sults were more consistent with the first step of the
photodissociation being cleavage of the C–N bond
[10]:
CCl3NO2 ! CCl3 þ NO2
The mechanism suggested by Carter et al. would
be consistent with the primary dissociation path-
way of nitromethane photolyzed via the pꢀ n
transition. The major photodissociation pathway
for nitromethane is also cleavage of the C–N bond,
although two minor competing channels, molecu-
lar elimination and oxygen atom elimination, have
been observed [14–16]:
Æ0:1 cmꢁ1
.

E.A. Wade et al. / Chemical Physics Letters 365 (2002) 473–479
475
Table 1
Summary of Absorbance bands for chloropicrin/Ar matrix
following photolysis
3. Results
Chloropicrin has a moderately complex infra-
Wavenumber
(cmꢁ1
Identification
red spectrum, as shown in Fig. 1. In a N2 matrix,
its ten strongest bands, in order of decreasing in-
tegrated area, are observed at 1619, 903, 865, 1314,
)
577
674
NOCl
Chloropicrin
Chloropicrin
Cl3CONO
Chloropicrin
Phosgene
674, 712, 1354, 845, 3726, and 2917 cmꢁ1
.
711
772
Chloropicrin/Ar was photolyzed at several dif-
ferent wavelengths. Interference filters were used at
220 (Oriel 53320), 251 (53340), 313 (Oriel 56410),
365(Oriel 56430), and 405 nm (Oriel 56440). Ad-
ditionally, a long-pass filter (k P 300 nm, Corning)
was used. Table 1 summarizes the observed bands.
In all cases, the same photoproducts were detected:
phosgene (Cl2CO), which has major bands at 1837
and 849 cmꢁ1 and minor bands at 1673 and
1020 cmꢁ1 [17], and NOCl, which has major bands
at 1805 and 577 cmꢁ1 [18]. A typical difference
spectrum is shown in Fig. 2. In this spectrum, the
negative peaks are due to the loss of chloropicrin
and the positive peaks are due to photolysis
products. The strong band at 1828 cmꢁ1 is due to
the overlap of the bands of phosgene and NOCl.
Production of phosgene and NOCl, and con-
sumption of chloropicrin, follow nearly first order
kinetics, as shown in Fig. 3. However, at very short
times, there is a deviation from first order kinetics,
which suggests that phosgene and NOCl are not
the primary photoproducts of this reaction.
843
849
861
903
Chloropicrin
Chloropicrin
Phosgene
1020
1088
1310
1352
Cl3CONO
Chloropicrin
Chloropicrin
Chloropicrin
Phosgene
1614, 1617
1673, 1685
1805, 1820
1827, 1837
2142
NOCl
Phosgene
CO
Chloropicrin
Chloropicrin
2644
2906
Fig. 2. Difference spectrum of chloropicrin/Ar matrix following
photolysis for 150 min at 313 nm.
An intermediate, which absorbs at 1088 and
772 cmꢁ1, was also detected. The kinetic behavior
of the 1088 cmꢁ1 band is shown in Fig. 4. The band
at 1088 cmꢁ1 is in the spectral region typical for a
C–O or N–O stretch and the band at 772 cmꢁ1 is
Fig. 1. Matrix isolation spectrum of chloropicrin in N2 (1:200).

476
E.A. Wade et al. / Chemical Physics Letters 365 (2002) 473–479
photodissociation was less efficient, either because
less UV was absorbed by chloropicrin (at 365 or
405 nm, where chloropicrin does not absorb
strongly) or because more UV was absorbed by the
filter (with the long-pass filter).
Chloropicrin/O2/Ar and chloropicrin/18O2=Ar
were photolyzed using the 313 nm filter and the
long-pass filter. The infrared bands observed were
identical to those observed in the chloropicrin/Ar
matrices, including the intermediate at 1088 cmꢁ1
.
Even when the concentration of 18O2 was ex-
tremely high (chloropicrin/18O2=Ar:1/50/150), no
labeled product bands were observed. At very high
18O2 concentrations, the intermediate was not ob-
served. However, this is probably because, in that
matrix, all of the bands were broadened so that the
intermediate band, which is very small, was wa-
shed out. The kinetics of the chloropicrin/O2=Ar
matrices are consistent with the kinetics of the
chloropicrin/Ar matrices, given the limits of un-
certainty. Based on these experiments, there is no
experimental evidence of a trioxolane intermedi-
ate, as suggested by Moilanen et al. [13], in these
cryogenic matrices.
Fig. 3. Kinetic behavior of chloropicrin, phosgene, and NOCl
following 313 nm photolysis of cryogenic matrix. Closed circles
are chloropicrin, closed diamonds are phosgene, and open
squares are NOCl. The lines are provided to guide the eye.
In the chloropicrin/N2 matrix, the same bands
were observed as in chloropicrin/Ar, although, as
expected, they are somewhat shifted in frequency
when compared to the chloropicrin/Ar matrices.
For example, the major band of phosgene is ob-
served at 1847 cmꢁ1 rather than 1840 cmꢁ1, and
the intermediate bands are observed at 766 and
1097 cmꢁ1. It was thought that the chloropicrin/
N2 matrix might allow for better observation of
the intermediate bands by stabilizing vibrationally
excited species. However, the increase in the ab-
sorbance of the intermediate was no more than
10%, which is within the uncertainty of experi-
ments performed with lamp/filter combinations.
The kinetics were also the same as in the chloro-
picrin/Ar matrices, within this uncertainty.
Fig. 4. Kinetic behavior of cryogenic matrix intermediate at
1088 cmꢁ1 when photolyzed at 313 nm. The line is given to
guide the eye.
consistent with a C–Cl stretch. Other strong bands
of the intermediate species could easily be obscured
by chloropicrin or other product bands, and could
not be isolated. This intermediate was detected
when chloropicrin was photolyzed with 220, 251
and 313 nm light. Somewhat less of the interme-
diate was observed at shorter wavelengths, which
may indicate that the intermediate is also photo-
active. The intermediate was not detected when the
photolysis was performed at 365, 405 nm, or with
the long-pass filter. However, with these filters,
4. Discussion
It is clear that there is some mechanism by
which chloropicrin can photolyze in the absence of
oxygen, contrary to the observations reported by
Moilanen et al. [13]. Chemistry could occur via a

E.A. Wade et al. / Chemical Physics Letters 365 (2002) 473–479
477
the gas phase. The CCl3ONO would then either
thermally dissociate or photolyze to give either the
same radicals or phosgene and NOCl. It is likely
that the intermediate is photochemically active,
because it was not stabilized by the N2 matrix and
its detection was strongly dependent on the pho-
tolysis wavelength. This mechanism is consistent
with the photolysis of nitromethane in Ar matrices
[19,20], where nitromethane initially photolyzes to
produce NO2 and CH3, which then undergo cage
recombination to produce CH3ONO. CH3ONO
then photolyzes to produce HNO and H2CO
[19,20].
one-step molecular elimination process, as in ni-
tromethane [14–16], which in the case of chloro-
picrin would directly give the final products:
CCl3NO2 ! Cl2CO þ NOCl
However, such a one-step process would not ac-
count for the observed intermediates in the cryo-
genic matrices. Thus, such a one-step mechanism
cannot account for all of the photolyzed chloro-
picrin.
Another likely mechanism for chloropicrin
photolysis is cleavage of the C–N bond, as sug-
gested by Carter et al. [10].
In order to further support the identity of this
intermediate, ab initio calculations were per-
formed using GAUSSIAN 98 at B3LYP/6-31g [24].
The predicted vibrational frequencies and infrared
intensities are consistent with the observed inter-
mediate spectrum and are summarized in Table 2.
Only five bands, at 617, 691, 791, 1096, and 1829
cmꢁ1, are predicted to have significant infrared
intensity and to be in the detection region of our
instrument. Two of these bands, at 791 and
1096 cmꢁ1, are consistent with the observed in-
termediate bands at 772 and 1088 cmꢁ1. The band
at 1096 cmꢁ1 is predicted to be the strongest in-
frared band. A band at 1829 cmꢁ1 is predicted to
have nearly the same infrared intensity as the
1096 cmꢁ1 band, but would overlap with and be
CCl3NO2 ! CCl3 þ NO2
The fragments can then react to produce phosgene
and NOCl.
CCl3 þ NO2 ! Cl2CO þ NOCl
It is reasonable that the NO2 and CCl3 radicals
themselves would not be observed, since their vi-
brational bands overlap with strong bands in
chloropicrin. The strongest vibrational band of
CCl3 is at 898 cmꢁ1 [21], in a region where chlo-
ropicrin has several very strong bands. Two vi-
brational bands of NO2, at 2900 and 1610 cmꢁ1
[22,23], are similarly overlapped with the nitro-
group bands of chloropicrin at 2917 and 1619 cmꢁ1
.
The other vibrational band, at 749 cmꢁ1 [22,23], is
weaker than the other two, and is in the congested
700–900 cmꢁ1 region.
Table 2
Calculated ab initio (B3LYP/6-31g) [24] spectrum of Cl3CONO
The intermediate observed in the cryogenic
matrices cannot be absolutely identified, but the
two observed bands are consistent with a C–Cl
stretch and either a C–O or N–O stretch. Based on
these two bands, it is likely that the intermediate is
CCl3ONO. This could be produced as a primary
photoproduct through rearrangement, but is more
likely produced following the cleavage of the C–N
bond to produce CCl3 and NO2.
Frequency
(cmꢁ1
Intensity
(arb. unitsa)
)
57.09
111.06
116.22
216.00
223.80
111.07
328.15
371.89
398.72
567.83
617.21
691.04
791.92
1096.08
1828.76
0.003
0.001
0.005
0.032
0.001
0.070
0.084
0.008
0.034
0.032
0.321
0.636
0.674
1.000
0.910
CCl3NO2 ! CCl3 þ NO2
CCl3 þ NO2 ! CCl3ONO
In the matrix, these two radicals would be trapped
close to one another, and could rearrange slightly
in order to form CCl3ONO. This explains why no
reaction with oxygen was observed in the matrix,
even though there was a reaction with oxygen in
a Intensities normalized to intensity of 1096 cmꢁ1 band.

478
E.A. Wade et al. / Chemical Physics Letters 365 (2002) 473–479
obscured by strong phosgene bands at 1827 and
1837 cmꢁ1. Similarly, the 691 cmꢁ1 band is pre-
dicted to have nearly the same intensity as the
791 cmꢁ1 band, but is very close to a strong
chloropicrin band at 674 cmꢁ1 and would likely
not be observed. The fifth predicted band, at
617 cmꢁ1, is not observed, but is predicted to be
the weakest of these five bands with less than one-
third the intensity of the 1096 cmꢁ1 band. Given
the small observed signal in the most intense fea-
ture of the intermediate spectrum, our failure to
observe this band is not sufficient to disprove our
assignment of the intermediate as Cl3CONO.
partial support by a California State University
Research Fund Grant. Dr. Wade would also like
to thank the Clark Teacher/Scholar Fellowship
for allowing her the opportunity to initiate this
research at SJSU. T. Clemes would like to thank
the Lottery Funds program at SJSU for partial
financial support. K. Reak would like to thank
the Mills College Women’s Studies Program for
financial support. Prof. C.B. Moore and Stacie
Cowan provided many helpful discussions which
led directly to this work. The equipment used in
this work was funded through grants from NSF,
Research Corporation, the William M. Keck
Foundation, and San Jose State University.
5. Conclusion
References
Chloropicrin undergoes photolysis to produce
phosgene and nitrosyl chloride in cryogenic ma-
trices. These products are observed at a wide range
of photolysis wavelengths, and whether or not
oxygen is present. Furthermore, when 18O2 is
present, 18O-labeled photoproducts are not ob-
served in cryogenic matrices.
Based on these observations, the following
mechanism is proposed. Chloropicrin initially un-
dergoes cleavage of the C–N bond to produce NO2
and CCl3 radicals. In the matrix, these radicals are
kept close to each other, and can recombine to
form Cl3CONO, which itself undergoes photolysis.
CCl3 and NO2 eventually react with each other to
form phosgene and NOCl. It is also possible that
some fraction of the chloropicrin photolyzes to
produce phosgene and NOCl directly.
This mechanism would be consistent with
many, but not all, of the previous gas phase ex-
periments [10,11,13]. In particular, we have ob-
served that chloropicrin will dissociate in the
absence of oxygen in cryogenic matrices, contrary
to what was reported by Moilanen et al. [13].
Further, gas-phase studies will be required to
completely explain these results and predict the
atmospheric behavior of chloropicrin.
[1] State of California, Department of Pesticide Regulation,
Summary of Pesticide Use Report Data, 1994.
[2] P. Morse, Chem. Eng. News 76 (1998), Oct 26 (*look up p.
#).
[3] M. Leistra, S.J.H. Crum, Water, Air, Soil Pollut. 50 (1990)
109.
[4] P. Howard, Handbook of Environmental Fate and Expo-
sure Data for Organic Chemistry 3 (1990) 126.
[5] J. Heicklen, Atmos. Environ. 16 (1982) 821.
[6] J.H. Seinfeld, Atmospheric Chemistry and Physics of Air
Pollution, Wiley, New York, 1986.
[7] R. Atkinson, Atmos. Environ. 24A (1990) 1.
[8] R. Atkinson, J. Phys. Chem. Ref. Data, Monograph 2
(1994).
[9] R. Atkinson, W.P.L. Carter, Chem. Rec. 1984 (1984)
437.
[10] W.P.L. Carter, D. Luo, I.L. Malkina, Atmos. Eviron. 31
(1997) 1425.
[11] T.D. Allston, M.L. Fedyk, G.A. Takacs, Chem. Phys. Lett.
60 (1978) 100.
[12] N.S. Bayliss, E.G. McRae, J. Phys. Chem. 58 (1954)
1006.
[13] K.W. Moilanen, D.G. Crosby, J.R. Humphrey, J.W. Giles,
Tetrahedron 34 (1978) 3345.
[14] K. Honda, H. Mikuni, M. Takahasi, M. Bull. Chem. Soc.
Japan 45 (1972) 3534.
[15] S.J. Paszyc, Photochemistry 2 (1973) 183.
[16] W.D. Taylor, T.D. Allston, M.J. Moscato, G.B. Fazekas,
R. Kozlowski, G.A. Takacs, Int. J. Chem. Kinet. 12 (1980)
231.
[17] E. Catalano, K.S. Pitzer, J. Am. Chem. Soc. 80 (1958)
1054.
Acknowledgements
[18] D.E. Tevault, R.R. Smardzewski, J. Phys. Chem. 82 (1978)
375.
We gratefully acknowledge partial support
from the Petroleum Research Fund, as well as
[19] M.E. Jacox, J. Phys. Chem. 88 (1984) 3373.
[20] H.W. Brown, G.C. Pimentel, J. Chem. Phys. 29 (1958) 883.

E.A. Wade et al. / Chemical Physics Letters 365 (2002) 473–479
479
[21] E.E. Rogers, S. Abramowitz, M.E. Jacox, D.E. Milligan,
J. Chem. Phys. 52 (1970) 2198.
[23] R.V. St. Louis, B. Crawford, J. Chem. Phys. 42 (1965)
847.
[22] W.G. Fateley, H.A. Bent, B.J. Crawford, Chem. Phys. 31
(1959) 204.
[24] M.J. Frisch et al., GAUSSIAN 98, Revision A.6, Gaussian,
Pittsburgh, PA, 1998.