Formation of Methyl Nitrite and Methyl Nitrate during Plasma Treatment of Diesel Exhaust

Article abstract of DOI:10.1021/es034126i

CH₃ONO and CH₃ONO₂ were identified as products of the nonthermal plasma treatment of simulated diesel exhaust using FTIR spectroscopy. CH₃ONO and CH₃ONO₂ were formed in the nonthermal plasma treatment of simulated diesel exhaust. The source of CH₃ONO_x was the association reaction of CH₃O radicals with NO_x. The source of CH₃O radicals was oxidation of propylene, which produces CH₃ radicals, which in turn, were oxidized to CH₃O radicals. At 180°C, a significant fraction (60-85%1 of the CH₃ONO decomposed upon flowing through the experimental apparatus. CH₃ONO was formed during the passage of simulated diesel exhaust through a nonthermal plasma. If the plasma-treated exhaust is not passed over a catalyst, or if the catalyst is deactivated, CH₃ONO will be emitted into the atmosphere. CH₃ONO is not a desirable vehicle emission.

Full text of DOI:10.1021/es034126i

Environ. Sci. Technol. 2003, 37, 4242-4245
controlling NOx emissions from diesel vehicles requires an
Formation of Methyl Nitrite and
Methyl Nitrate during Plasma
Treatment of Diesel Exhaust
understanding of the gas-phase chemistry in such devices.
Methyl nitrate, CH3ONO2, is produced during the nonthermal
plasma treatment of simulated diesel exhaust (10). CH3ONO2
is believed to be formed via the association reaction between
CH3O radicals and NO2 (10):
T . J . W A L L I N G T O N , * J . W . H O A R D ,
M . P . S U L B A E K A N D E R S E N , A N D
M . D . H U R L E Y
CH3O + NO2 + M f CH3ONO2 + M
(1)
If reaction 1 is indeed the source of CH3ONO2, then there
should also be significant formation of methyl nitrite, CH3-
ONO, by the analogous reaction with NO:
Ford Motor Company, Mail Drop SRL 3083,
Dearborn, Michigan 48121-2053
Y. N A K A N O A N D M . K A W A S A K I
CH3O + NO + M f CH3ONO + M
(2)
Department of Molecular Engineering and Graduate
School of Global Environmental Studies, Kyoto University,
Kyoto 606-8501, Japan
CH3ONO photolyzes rapidly in the sunlit atmosphere,
releasing NOx and generating OH radicals which can promote
the formation of urban smog. An understanding of the
possible formation of CH3ONO in vehicle exhaust is clearly
of interest. While its presence has been predicted by kinetic
models (11, 12), the observation of CH3ONO in plasma
treatment systems has not been reported. As part of a
collaborative research effort in our laboratories to elucidate
the chemistry in nonthermal plasma devices, an investigation
of the formation of CH3ONO was undertaken. The aim of the
present work was to determine if, and how, CH3ONO is
formed in such systems.
FTIR spectroscopy was used to identify CH ONO and
3
CH ONO as products of the nonthermal plasma treatment
3
2
of simulated diesel exhaust. This is the first observation
of CH ONO formation in such systems. The yield of CH -
3
3
ONO relative to CH ONO scaled linearly with the average [NO]/
3
2
[NO ] ratio in the system. A plot of [CH ONO]/[CH ONO ]
2
3
3
2
versus [NO]/[NO ] gives a slope of 1.81 ( 0.30. This result
2
is indistinguishable from the literature value of the rate
Experimental Section
constant ratio k(CH O + NO)/k(CH O + NO ) ) (2.6 × 10-11)/
3
3
2
(1.5 × 10-11) ) 1.73 ( 0.37. The experimental observations
The experimental system used in the present work is
described in detail elsewhere (10). Figure 1 is a schematic of
the apparatus. NO, NO2, CO, CO2, O2, C3H6, and C3H8 gases
can be mixed in N2 carrier. Liquid water can be injected into
the gas in heated lines. For the experiments here, a simplified
gas blend was used, consisting of C3H6, O2, NO, and NO2.
Previous experiments have shown little or no impact of the
other components listed above on the chemistry occurring
in the system (9, 10), and their presence (especially H2O)
complicates the FTIR measurements. The test plasma was
in an oven which was maintained at either 30 or 180 °C. The
effluent was mixed with extra N2 at a ratio of 2.5:1 dilution,
to match flow rates to the analytical equipment, and passed
through measurement instrumentation. The principal ana-
lytical instrument is a Mattson Nova Cygni FTIR spectrometer
operated at room temperature with a spectral resolution of
0.25 cm-1 and equipped with a long-path-length (20.7 m)
Foxboro sampling cell. A Horiba FIA-220 flame ionization
detector, a MPA-220 magnetopneumatic oxygen analyzer, a
CLA-220 chemiluminescent NOx analyzer, and AIA-220 NDIR
CO and CO2 analyzers were also used.
The present study is an extension of our previous work
in which FTIR spectroscopy was used to identify CH3ONO2
formation (10). In our previous work we did not observe the
presence of CH3ONO. The IR features of CH3ONO are
significantly weaker than those of CH3ONO2 in the 700-900
cm-1 region of the spectrum where the FTIR analysis is
performed. CH3ONO may have been present in the previous
experiments, but we would not have been able to detect it.
To improve the detection sensitivity in the present experi-
ments, three changes to the experimental protocol were
made. First, we employed 512 (rather than 32) co-added
interferograms, giving IR spectra with 4-fold better S/ N at
the cost of a 16-fold (768 s versus 48 s) increase in data
acquisition time. Second, instead of diluting the sample gas
with N2 10:1 prior to passing through the FTIR cell to minimize
exposure of the moisture-sensitive optical components to
suggest that reactions of CH O radicals with NO and
3
NO are the sources of CH ONO and CH ONO in such
2
3
3
2
systems. The linear relationship between the yields of CH -
3
ONO and CH ONO provides a means of estimating the
3
2
yield of these compounds during nonthermal plasma treatment
of diesel exhaust.
Introduction
Concerns about global climate change have increased the
pressure on automobile manufacturers to increase vehicle
fuel efficiency. Diesel vehicles have superior fuel economy
and approximately 20-30% lower CO2 emissions than their
gasoline counterparts. Unfortunately, NOx emissions from
diesel engines are difficult to control to proposed future
emissions standards. Modern spark-ignition engines operate
at stoichiometry, and NOx emissions are controlled by a three-
way catalytic converter. In contrast, diesel engines operate
fuel lean, and the exhaust contains substantial amounts of
O2 (typically 6-10%). The reduction of NOx by a three-way
catalyst in such oxidizing environments is difficult (1).
Nonthermal plasma discharge combined with a down-
stream catalyst is a technology under evaluation for use in
the removal of NOx from diesel exhaust (2-8). The details of
the gas-phase chemistry occurring in the plasma discharge
of exhaust gas are unclear. It is believed that O(3P) atoms
and, to a lesser extent, OH radicals are the dominant species
responsible for initiating hydrocarbon oxidation in such
systems (9). However, the subsequent chemistry occurring
during nonthermal plasma treatment of diesel exhaust is
not fully understood. Optimization of the efficiency of plasma
exhaust treatment devices and assessment of their utility in
* Corresponding author phone: (313) 390-5574; fax: (313) 594-
2923; e-mail: twalling@ford.com.
9
4 2 4 2 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 18, 2003
10.1021/es034126i CCC: $25.00
2003 Am erican Chem ical Society
Published on Web 08/08/2003

FIGURE 1. Schematic of the test lab configuration (FTIR, Fourier transform infrared spectrometer; FID, flame ionization detector for
hydrocarbons; CLA, chemiluminescence NO analyzer). A schematic of the plasma is shown in the oven.
x
H2O vapor, the exhaust gas was diluted only 2.5:1. Third, a
simplified gas blend was used as mentioned above. The
detection limits for CH3ONO and CH3ONO2 in the present
experiments were 2.0 and 0.3 ppm, respectively.
TABLE 1. CH ONO Survival through the Experimental Apparatus
3
under Various Conditionsa
plasma
[CH ONO]0, [NO]0, [O ] , [C H ] , temp, discharge, CH ONO
3
2 0
3
6 0
3
In the tests reported here, the gas composition was 500-
740 ppm C3H6, 8% O2, and 0-420 ppm both NO and NO2 in
1 atm of N2 diluent. A sketch of the dielectric barrier discharge
device used to generate a plasma is shown (in the oven) in
Figure 1. It consists of two 24 × 100 mm sheets of alumina.
Each sheet has a conductive electrode embedded in it. The
two sheets are bonded together with a ceramic adhesive,
leaving a gap of 1.3 mm between the sheets (13). The assembly
is mounted in a quartz tube such that all the gas must flow
through the gap between sheets. The two electrodes are
connected to a Trek 10/ 10 high-voltage amplifier driven by
an HP33120A function generator. For these tests, the excita-
tion wave was a 250 Hz triangle wave with peak voltage
between 5 and 6 kV, adjusted so that the energy density is
30 J of energy/ L of gas flow (typical of that which may be
used in passenger car applications) when the plasma is turned
on. The gas residence time in the oven was approximately
10 s.
Methyl nitrite, CH3ONO, was synthesized by the dropwise
addition of concentrated H2SO4 to a saturated solution of
NaNO2 in methanol using the method of Taylor et al. (14).
The CH3ONO sample was devoid of any detectable impurities
using FTIR analysis. A small sample (4-5 cm3) was transferred
to an evacuated metal gas cylinder (size 300) and pressurized
to 200 psi with pure nitrogen gas. The final concentration of
CH3ONO in the steel cylinder was measured to be 601 ppm
by FTIR. Control experiments were performed to determine
the stability of CH3ONO/ N2 mixtures; there was no observable
(<1%) change in the CH3ONO concentration after standing
for 2 months.
ppm
ppm
%
ppm
°C
J/L
loss, %
20
20
20
20
20
20
0
0
260
0
0
260
0
0
8
0
0
8
0
0
510
0
0
510
30
30
30
180
180
180
0
30
0
0
30
0
6
12
11
60
85
75
a
All experim ents were perform ed in 1 atm of N2 diluent.
in this table, but discussed below in the main experiment).
With the gas lines held at 90-120 °C and the oven at 180 °C,
CH3ONO loss was 60% in N2. This increased to 85% when the
plasma was on. Adding NO, O2, and C3H6 without the plasma
caused a 75% loss of CH3ONO. These control experiments
show that when the apparatus is maintained at 30 °C, there
is only a minor loss of CH3ONO on passage through the
system. With the oven at 180 °C and transfer lines at 90-120
°C there is a substantial loss of CH3ONO on flowing through
the system. Although not shown here, the primary decom-
position products are HCHO and NO. In subsequent experi-
ments at 30 °C the observed concentration of CH3ONO was
corrected by multiplication by 1/ 0.89 to account for de-
composition losses. As discussed previously (10), there are
no discernible losses of CH3ONO2 on passage through the
experimental system. No corrections were applied to the CH3-
ONO2 data.
Figure 2 shows spectra taken before (A) and after (B)
passage of a mixture of 8% O2, 739 ppm C3H6, 419 ppm NO,
and 371 ppm NO2, in 1 bar of N2 diluent, through the plasma
treatment device. After passage through the plasma the FTIR
analysis gave 347 ppm C3H6, 356 ppm NO, and 393 ppm NO2.
Subtraction of features attributable to C3H6, NO2, HNO2, and
HNO3 from panel B in Figure 2 gives the product spectrum
shown in panel C. Comparison of panel C with reference
spectra for CH3ONO and CH3ONO2 clearly shows the
formation of these species. The concentrations of CH3ONO
and CH3ONO2 in panel C were 8.7 and 4.6 ppm, respectively.
Comparison with a reference spectrum of propylene oxide
shows the formation of a 15 ppm concentration of this species.
Propylene oxide is formed by reaction of O atoms with
propene and demonstrates the importance of O atom
chemistry in the system (9). The goal of the present
experiments was to investigate the formation of CH3ONO. A
detailed analysis of the propene oxidation products was not
performed.
Results
CH3ONO is a reactive compound which is prone to decom-
position (particularly at elevated temperatures (15)). The first
question was whether it would survive the flow through the
apparatus consisting of stainless steel lines, glass lines,
stainless steel couplings, and the ceramic plasma device
which was located in an oven maintained at either 30 or 180
°C. To answer this question, control experiments were
performed in which mixtures containing 20 ppm CH3ONO
diluted in N2 flowed through the plasma treatment system
under various conditions.
Table 1 shows the results of these control experiments.
First, 20 ppm CH3ONO in N2 flowed through the system with
the oven maintained at 30 °C and the plasma off. There was
a 6% difference between the nominal composition and the
FTIR measured value. Turning the plasma discharge on (30
J/ L) increased the difference to 12%. Adding NO, O2, and
propylene to the flow with the plasma off caused an 11% loss
of CH3ONO. When exposed to the plasma discharge, the CH3-
ONO concentration in this gas blend increased (not shown
It is important to consider potential gas-phase losses of
CH3ONO and CH3ONO2 in the system. O(3P) atoms, and to
a lesser extent OH radicals, are the dominant species
responsible for initiation of the oxidation of organic com-
pounds in nonthermal plasma of simulated diesel exhaust
9
VOL. 37, NO. 18, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 4 2 4 3

FIGURE 3. [CH ONO]/[CH ONO ] ratio present in plasma-treated
3
3
2
exhaust versus the average [NO]/[NO ] ratio in the gas passing
2
through the plasma for experiments conducted at 30 °C.
FIGURE 2. IR spectra of a mixture of [O ] ) 8%, [NO] ) 419 ppm,
2
[NO ] ) 371 ppm, and [C H ] ) 738 ppm in 1 bar of N2 diluent at 30
2
3 6
°C with the plasma turned off (A) and on (B) with 30 J/L of energy
deposition. Panel C is the product spectrum obtained by stripping
IR features attributable to C H , NO , HNO , and HNO from panel
3
6
2
2
3
B. Panels D-F are reference spectra of CH ONO , CH ONO, and
3
2
3
propylene oxide (CH CHOCH , methyl oxirane).
3
2
(9). O(3P) atoms and OH radicals react at least 100 times
more rapidly with C3H6 than with either CH3ONO or CH3-
ONO2 (16). Given that the consumptions of C3H6 on passage
through the plasma were 5-60%, it can be concluded that
loss of CH3ONO or CH3ONO2 via reaction with O(3P) atoms
and OH radicals will be of negligible importance. Gas-phase
processes within the plasma are unlikely to be a significant
loss of either CH3ONO and CH3ONO2 in the present experi-
ments.
Experiments were performed at 30 °C using a variety of
initial [NO] and [NO2] in the range 0-420 ppm. Observed
concentrations of CH3ONO and CH3ONO2 in the plasma
exhaust were 7-13 and 5-9 ppm, respectively. Figure 3 shows
a plot of [CH3ONO]/ [CH3ONO2] observed in the plasma
exhaust versus the ratio of the average [NO] and [NO2] in the
gas stream passing through the plasma. The average [NO]
and [NO2] values were computed from the average concen-
tration values with the plasma on and off (all other parameters
held constant). As seen from Figure 3 there is a linear
relationship between [CH3ONO]/ [CH3ONO2] and [NO]/ [NO2].
The line through the data is a linear least-squares fit which
gives a slope of 1.81 ( 0.30. Quoted errors are 2 standard
deviations from the regression analysis.
FIGURE 4. IR spectra of a mixture of [O ] ) 8%, [NO] ) 240 ppm,
2
and [C H ] ) 537 ppm in 1 bar of N2 diluent at 180 °C with the plasma
3
6
turned off (A) and on (B) with 30 J/L of energy deposition (B). Panel
C is the product spectrum obtained by stripping IR features
attributable to C H , NO , HNO , and HNO from panel B. Panels D
3
6
2
2
3
and E are reference spectra of CH ONO and CH ONO.
3
2
3
guishable from the rate constant ratio k2/ k1, suggesting that
the origin of CH3ONOx is reaction of CH3O radicals with NOx.
The data presented in Figure 3 were obtained in experi-
ments conducted at 30 °C. CH3ONO formation was also
observed in experiments conducted at 180 °C. For example,
Figure 4 shows spectra taken when a mixture of 8% O2, 537
ppm C3H6, and 240 ppm NO in 1 bar of N2 diluent was passed
through the system with the oven at 180 °C and the plasma
treatment device turned off (A) and on (B). With the plasma
turned on the FTIR analysis gave 399 ppm C3H6, 98 ppm NO,
and 131 ppm NO2. Subtraction of features attributable to
C3H6, NO2, HNO2, and HNO3 from panel B in Figure 4 gives
the product spectrum shown in panel C. Comparison of panel
C with reference spectra for CH3ONO and CH3ONO2 clearly
shows the formation of these species. As discussed above,
control experiments established that there is significant loss
of CH3ONO on flowing through the heated experimental
system. The computation of suitable corrections to account
It is of interest to compare this result with the ratio of the
rate constants for the association reactions 1 and 2.
CH3O + NO2 + M f CH3ONO2 + M
CH3O + NO + M f CH3ONO + M
(1)
(2)
In 1 bar of N2 at 298 K, the IUPAC Committee recommends
k1 ) 1.5 × 10-11 and k2 ) 2.6 × 10-11 cm3 molecule-1 s-1 (17).
Assuming an uncertainty of 15% for these rate constants,
which seems reasonable on the basis of the agreement in the
literature data (17), we arrive at k2/ k1 ) 1.73 ( 0.37. The
slope of the line through the data in Figure 3 is indistin-
9
4 2 4 4 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 18, 2003

for decomposition of CH3ONO in the apparatus at high
temperature prior to reaching the FTIR detection apparatus
is problematical, and experiments at high temperature were
not pursued in detail. It is clear from Figure 4 that CH3ONO
is formed in the plasma at high temperatures. This finding
is reasonable given the observation of CH3ONO in the low-
temperature experiments and the fact that the rate constant
ratio k2/ k1 does not change substantially over the temperature
range 30-180 °C (17).
consideration for future commercial diesel exhaust treatment
systems which incorporate nonthermal plasma devices.
Acknowledgments
The work presented here was performed at Ford Research
Labs as part of Ford’s participation in CRADA PNL-183 “New
Technologies for Exhaust Emission Reduction”, a cooperative
project of the USCAR Low Emissions R&D Partnership, Pacific
Northwest National Lab, and the U.S. Department of Energy.
M.P.S.A. thanks the Royal Danish Academy of Sciences and
Letters and R. H. Lollegaard’s Fond for financial support. We
thank Ole John Nielsen (University of Copenhagen) for helpful
discussions.
Discussion
Two conclusions can be drawn from the present work. First,
CH3ONO and CH3ONO2 are formed in the nonthermal plasma
treatment of simulated diesel exhaust. Second, the source of
CH3ONOx is the association reaction of CH3O radicals with
NOx. The source of CH3O radicals in the present study is
oxidation of propene, which produces CH3 radicals which,
in turn, are oxidized to CH3O radicals. Essentially all
hydrocarbon fuels contain substantial amounts of chemicals
containing CH3 groups, and it should be expected that CH3-
ONO and CH3ONO2 will be common products formed during
plasma treatment of exhaust from vehicles powered using
a variety of fuels.
The goal of the present work was to determine if, and
how, CH3ONO is formed during nonthermal plasma treat-
ment of diesel exhaust. We show that CH3ONO is formed in
such systems and that reaction of CH3O radicals with NO is
the likely source. At elevated temperature (180 °C) it was
observed that a significant fraction (60-85%) of the CH3-
ONO decomposed upon flowing through the experimental
apparatus. It seems likely that thermal decomposition of CH3-
ONO would also be significant in real world applications. In
real world applications it is envisioned that, after passage
through the plasma treatment device, the exhaust would flow
over a catalyst designed to facilitate reduction of NO2 into
N2 (9). Given the reactive nature of CH3ONO, it seems very
likely that most, if not all, of the CH3ONO produced in the
plasma would be removed by reactions on the catalyst
surface. A detailed study of the fate of CH3ONO on passage
over a catalyst was not within the scope of the present study.
The present work demonstrates unambiguously that CH3-
ONO is formed during the passage of simulated diesel exhaust
through a nonthermal plasma. If the plasma-treated exhaust
is not passed over a catalyst, or if the catalyst is deactivated,
then it is likely that CH3ONO will be emitted into the
atmosphere. CH3ONO is not a desirable vehicle emission.
The control of CH3ONO emissions should be a design
Literature Cited
(1) Heck, R. M.; Farrauto, R. J. Catalytic Air Pollution Control; John
Wiley & Sons: New York, 2002 (ISBN 0-471-43624-0).
(2) Penetrante, B. M.; Brusasco, R. M.; Pitz, W. J.; Vogtlin, G. E.;
Kung, M. C.; Kung, M. H.; Wan, C. Z.; Voss, K. E. SAE Tech. Pap.
Ser. 1998, No. 982508.
(3) Hoard J. W.; Balmer, M. L. SAE Tech. Pap. Ser. 1998, No. 982429.
(4) Balmer, M. L.; Tonkyn, R.; Kim, A.; Yoon, S.; Jiminez, D.; Orlando,
T. M.; Barlow, S. E. SAE Tech. Pap. Ser. 1998, No. 982511.
(5) Fisher, G. B.; DiMaggio, C. L.; Yezerets, A.; Kung, M. C.; Kung,
H. H.; Baskaran, S.; Frye, J. G..; Smith, M. R.; Herling, D. R.;
LeBarge, W. J.; Kupe, J. SAE Tech. Pap. Ser. 2000, No. 2000-01-
2965.
(6) Schmieg, S.; Cho, B.; Oh, S. SAE Tech. Pap. Ser. 2001, No. 2001-
01-3565.
(7) Hoard, J. W.; Panov, A. SAE Tech. Pap. Ser. 2001, No. 2001-01-
3512.
(8) Panov, A.; Tonkyn, R. G.; Balmer, M. L.; Peden, C. H. F.; Malkin,
A.; Hoard, J. W. SAE Tech. Pap. Ser. 2001, No. 2001-01-3513.
(9) Hoard, J.; Wallington, T. J.; Bretz, R. L.; Malkin, A., Dorai. R.;
Kushner, M. J. Int. J. Chem. Kinet. 2003, 35, 231.
(10) Hoard, J. W.; Wallington, T. J.; Ball, J. C.; Hurley, M. D.; Wodzisz,
K. Environ. Sci. Technol. 1999, 33, 3427.
(11) Dorai, R.; Kushner, M. J. J. Appl. Phys. 2000, 88, 3739.
(12) Dorai, R.; Kushner, M. J. J. Phys. D 2001, 34, 574.
(13) Novak, R. SAE Tech. Pap. Ser. 2003, No. 2003-01-1189.
(14) Taylor, W. D.; Allston, T. D.; Moscato, M. J.; Fazekas, G. B.;
Kozlowski, R.; Takacs, G. A. Int. J. Chem. Kinet. 1980, 12, 231.
(15) Selle, H. Z. Elektrochem. 1957, 61, 672.
(16) Westley, F.; Herron, J.; Frizzell, D. NIST Standard Reference
Database 17-2Q98; NIST: Gaithersburg, MD, 1998.
(17) Atkinson, R.; Baulch, D. L.; Cox, R. A.; Hampson, R. F., Jr.; Kerr,
J. A.; Rossi, M. J.; Troe, J. J. Phys. Chem. Ref. Data 1999, 28, 191.
Received for review February 12, 2003. Revised manuscript
received June 18, 2003. Accepted June 24, 2003.
ES034126I
9
VOL. 37, NO. 18, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 4 2 4 5