United States
Environmental Protection
Agency
Environmental Sciences Research
Laboratory
Research Triangle Park NC 27711
EPA-600/2-80-024
January 1980
Research and Development
4>EPA
Kinetic Studies of
Simulated Polluted
Atmospheres
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-80-024
January 1980
KINETIC STUDIES OF SIMULATED POLLUTED ATMOSPHERES
BY
Jack G. Calvert
Department of Chemistry
The Ohio State University
Columbus, Ohio 43212
Grant Number R804348-01
Project Officer
Joseph J. Bufalini
Atmospheric Chemistry and Physics Division
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U. S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
-------
DISCLAIMER
This report has been reviewed by the Gas Kinetics and Photochemistry
Branch, U.S. Environmental Protection Agency, and approved for publication.
Approval does not signify that the contents necessarily reflect the views
and policies of the U.S. Environmental protection Agency, nor does mention
of trade names or commercial products constitute endorsement or recom-
mendation for use.
-------
ABSTEACT
This research was initiated to help quantify several key aspects of
the chemistry of the polluted atmosphere which remain ill-defined. Kinetic
and mechanistic studies were made related to several important atmospheric
contaminants: sulfur dioxide, formaldehyde,, nitrous acid, and the nitros-
amines.
Through the use of laser excitation, sulfur dioxide has been excited
at selected wavelengths to characterize the reactions and lifetimes of the
several common excited states of this species with various atmospheric com-
ponents. The reactions of sulfur dioxide with the methylperoxy radical were
also studied by flash spectroscopy. All of these data and other published
rate data were reviewed and evaluated. It was concluded that the homogeneous
oxidation of sulfur dioxide in the atmosphere is initiated largely by way of
three free radical intermediates; the hydroxyl radical (HO) is the most im-
portant of these. The hydroperoxy (H02) and methylperoxy (CHsOa) may con-
tribute significantly for certain highly polluted atmospheres.
The present studies have shown that the ubiquitous formaldehyde molecule
can be a major source of the important hydroperoxyl radical through the photo-
lysis process, CH20 + sunlight -» HCO + H, followed by, HCO + 02 -» H02 + CO and
H + 02 + M -» H02 + M. Our data allow a quantitative evaluation of the ap-
parent first order rate constants for the rates of H02 generation by this
process within the troposphere at various solar zenith angles.
The absolute extinction coefficients for nitrous acid were determined,
and estimates were made of the rates of hydroxyl radical generation in the
troposphere through the reaction, HONO + sunlight -> HO + NO.
Long-path Fourier transform infrared spectroscopy was employed in one
phase of the work to help evaluate the potential for nitrosamine formation
in the polluted atmosphere. It was found that the dimethylamino radical,
(CHs)2N, is 1.5 x 10~5 -times and 3-9 x 10~7 times less reactive toward
oxygen than nitric oxide and nitrogen dioxide, respectively. Thus alkyl
amino radicals formed by H-atom abstraction from amines by HO-radical attack
in the atmosphere have a reasonable chance of forming nitrosamines (R2n-W=0)
and nitramines (R2N-M)2) even though nitric oxide and nitrogen dioxide
impurities are at concentrations in the pphm range.
This report was submitted in fulfillment of grant number R804348-01 by
Ohio State University under the sponsorship of the U.S. Environmental Pro-
tection Agency. This report covers a period from January, 1976,to April,
1979, and work was completed as of March 1979.
111
-------
CONTENTS
Abstract ill
Figures vi
Tables xi
1. Introduction and Conclusions 1
2. Studies Related to Sulfur Dioxide Removal Mechanisms in
the Atmosphere 5
Kinetics of fluorescence decay of S02 excited in the
2662-3273A region 5
References 35
The mechanism of photochemical reactions of S02 with
isobutane excited at 3130A 37
References 60
A kinetic flash spectroscopic study of the CHs02-
CH302 and CH302-S02 reactions 63
References 77
Mechanism of the homogeneous oxidation of sulfur dioxide
in the troposphere 79
References 129
3. Studies Related to Formaldehyde Removal Mechanism in the
Atmosphere
Quantum effieiency^of the primary processes in CH20
photolysis at 3130A and 25°C
References 159
The wavelength dependence of the quantum efficiencies
of the primary processes in CH20 photolysis at 25°C 162
References 178
An unusual H2-forming reaction in the 3130A photolysis
of CH20-02 mixtures at 25°C 180
References 198
k. Studies Related to the Removal Mechanisms of Nitrogenous
Compounds in the Atmosphere 201
The near uv absorption spectrum of gaseous HOWO and
N203 201
References 2lU
The kinetics of dimethylamino radical reactions in
simulated atmospheres: the formation of dimethyl-
nitrosamine and dimethylnitramine 216
References 263
-------
FIGURES
Number
1. Comparison of the wavelength dependence of the trans-
mission of filters used. 7
2. Time resolved fluorescence decay curves of S02 excited at
3020 A. 9
3. Stern-Volmer plot of the long-lived component of S02
fluorescence excited at 3107 A. 11
k. Stern-Volmer plot of the short-lived component of S02
fluorescence excited at 3107 A. 11
5- Plot of the radiative lifetime of the long-lived fluores-
cing species versus excitation energy of S02. 13
6. Plot of the radiative lifetime of the short-lived fluores-
cing species versus excitation energy of S02. 13
7. Plot of the rate constant for the quenching of the L
state by S02 versus the excitation energy. 13
8. Plot of the rate constant for the quenching of the S state
by S02 versus the excitation energy. 13
9. Semilog plot of the intensity of the total fluorescence of
the S02 molecule versus time with excitation at 2998 A. 18
10. Plot of the calculated initial rate of decay of the L
species of S02 versus the excitation energy. 18
11. Plot of the apparent first-order rate constant for the dif-
fusional loss of the L state versus 1/Pqn • 20
12. Plot of the ratio I /I versus pressure: excitation wave-
length 3107 A; assuming S -» L conversion. 20
13- Plot of the ratio I /I versus pressure: excitation wave-
length 3107 A; assuming S ^ L interconversion. 2k
o / o
Ik. Plot of the ratio I /I versus pressure for excitation at
3211, 3225, 3275 A. 2k
15. Plot of the Brus and McDonald data for the ratio I /I
as a function of pressure of S02. 29
vi
-------
16. Semilog plot of the fluorescence intensity versus time for
S02 under collision free conditions. 30
17. Example of the pressure variation in the 3130 A photoly-
sis of S02 with isobutane; P = 15.0 Torr. 39
18. Light scattering observed in the photolysis of S02-iso-
butane mixtures. 40
19. Example of the pressure variation in the 3130 A photoly-
sis of S02 with isobutane; P0_ = 0.195 Torr. 40
bU2
20. Plot of quantum yields as a function of light intensity. 4l
21. Plot of quantum yields as a function of [EH]/[S02] ratio. 44
22. Plot of the function of the quantum yield, relation C,
versus the benzene pressure. 44
23. Plot of the function of the left-hand term in relation D
as a function of the [S02]/[RH] ratio. 47
24. Plot of P.(I /I ) as a function of P0_ . 51
j a o bu2
25. Plot of quantum yields as a function of P
5 Torr. ^1°' S°2 52
26. Plot of quantum yields as a function of P J^qn ~
2.5, 5.0, 10.0, and 15.0 Torr. °4Ml° b°2 55
27. Plot of the function of quantum yield, relation C, versus
benzene pressure. 56
28. Plot of quantum yields as a function of the C02 pressure. 58
29. Plot of the rate of increase of initial light scattering
as a function of the rate of pressure drop. 59
30. The measured initial absorbance of CH302 at 265 nm versus
azomethane pressure in the flash photolysis of Me2N2-02. 66
31. The gas phase extinction coefficients of CH302 as a
function of wavelength. 67
32. Plot of the rate constant function kis[S02] + k5[Me2I\[2]
versus P . 74
bO2
33« Comparison of the extinction coefficients of S02 within
the first allowed band, the "forbidden" band, and a
typical wavelength distribution of the flux of solar
quanta at ground level. 8l
vii
-------
34. P.elation between the rate constants for the triplet
quenching reactions with oxygen and E - E0 . 89
J-l >Jo
35. Relation between the rate constant for the triplet
quenching reactions with nitric oxide and E - E . 9°
36. The theoretical percentage of H02 in the Gas phase which
is complexed with H20 vapor as a function of the temper-
ature and relative humidity. 101
37. Variation of the apparent second order rate constant for
the reaction 57 with pressure of the added gases. 108
38. Plot of the reciprocal of the apparent second order rate
constant for the reaction 39 versus the reciprocal of the
pressure of the added gas. 109
39. Plot of the ratio of rate of 03 loss to rate of S03
formation versus [S02]-1. 119
40. Plot of the ratio of slope/intercept for the plots of
Figure 7 versus [H20]. 119
4l. The theoretical monthly average of the rate (% hr"1) of
S02 oxidation within the northern atmosphere as a
function of the month of the year. 123
42. The theoretical rate (% hr"1) of S02 oxidation by HO
(reaction 39) and H02 (reaction 31) at various elevations
within the troposphere and at various N latitudes. 123
43. The theoretical rate of attack of various free radical
species on S02 (% hr 1) for a simulated sunlight-
irradiated polluted atmosphere. 125
44. First-order plot of a function of the pressure in CH20
photodecomposition at 3130 A. 145
45. Effect of incident light intensity on the rate of
CH20 photodecomposition at 3130 A. 145
46. Plot of (0 - 02) - 1 versus [C4HB]/[CH20]. 153
ns
47. The previously published estimates of the primary quantum
yield of the free radical-forming process 1 in CH20
photolysis. 163
48. The wavelength dependence of the primary quantum yield
of the free radical-forming process 1 in CH20 photolysis. 171
Vlll
-------
^9. Variation with solar zenith angle of the apparent first
order rate constants for sunlight absorption by CH20
and the decomposition of CH20 by primary process 1. 176
50. Pressure-time profiles in the photooxidation of CH20. 183
51. Beer's law plots for W20s(g) at some representative wave-
lengths . 2.0k
52. The absorption cross section for N203(g) as a function
of wavelength. 208
53. The absorption cross section for HONO(g) as a function
of wavelength. 209
5^-- Estimated first order rate constants for HONO(g) photo-
lysis in sunlight. 213
55- Comparison of irradiance inside photolysis cell with solar
irradiance. 220
56. Gas handling system for introducing known amounts of
chemicals into the photolysis cell. 222
57- Optical paths in a standard White cell(4n pass). 223
58. Placement of images on ML in a standard White cell. 22^
59• Mirror system for the modified White cell. 225
60. Optical transfer system from interferometer to photo-
chemical cell. 227
6l. Optical diagram of interferometer. 228
62. Concentration/time profile for the reaction of dimethyl-
amine with nitrous acid. 230
63. • Plot of the disappearance of dimethylnitrosamine as a
function of photolysis time. 232
6k. Plot of the logarithm of dimethylnitrosamine concentration
as a function of time. 233
65. Plot of the reciprocal concentration of dimethylnitrosamine
as a function of time for the photolysis of DMN with 5-0
ppm NO in N2. 235
66. Initial data from the photolysis of DMN in N2. 236
IX
-------
67. Plot of in(Absorbance) for three low pressure photolyses
of 1T02. 239
66. Concentration-time profile for the photolysis of di-
me thy initrosamine in a mixture of 1+9 Torr 02 and 65!
Torr of N2.
69. Concentration-time profile for the photolysis of di-
me thy luitrosamine in a mixture of 1^0 Torr of 02 and
560 Torr of N2.
70. Concentration-time profile from the photolysis of di-
methylnitrosamine in a mixture of 730 Torr 02.
71. Concentration-time profile from photolysis of dimethy1-
nitrosamine with 5-^ PP31 NO in a mixture with 23 Torr
02 and 677 N2.
72. Concentration-time profile from the photolysis of di-
me thyInitrosamine with 5-5 PP01 NO in a mixture of 69
Torr 02 and 621 N2.
73- Concentration-time profile from the photolysis of di-
me thy Initrosamine with 5-5 ppm NO in a mixture with
138 Torr 02 and 562 Torr N2. 250
7^-. Concentration-time profile for the photolysis of di-
me thy Initrosamine with 5-0 ppm NO in 700 Torr 02. 251
75- Concentration-time profile from the photolysis of di-
methyInitrosamine with 25 ppm NO in a mixture of 1^5
Torr oxygen and 555 Torr N2. 252
76. Plot of functions B and B' as a function of [02]/[NO]
used to extimate k4/k3. 256
77. Plot of functions C and C1 as a function of [02]/[N02]
for the estimation of the rate constant ratio k5a/k3
and k5b/ks. "259
x
-------
TABLES
Number Page
1. Lifetime and quenching rate constants for long-lived and
short-lived species excited in S02. lh
2. Stern-Volmer parameters of the L and S components as a
function of the fluorescence wavelengths monitored. 15
3. Reciprocal lifetimes of the S and L components and I°/I °
as a function of the fluorescence wavelengths monitored. 16
k. The effect of pressure on the I °/I values and re-
ciprocal lifetimes of the L and S components of S02
fluorescence excited at 3107A. 25
5« The estimated fraction of the total fluorescence intensity
from the L species in S02 steady state irradiations. 32
6. Rate constants for the quenching by various atmospheric
gases of the L and S components of the S02 fluorescence
excited at 3130 and 2662A. 3!+
7. Quantum yields for the photolyses of S02 and isobutane
mixtures under low pressure conditions. 42
8. Quantum yields for the photoreaction of S02 with iso-
butane in the presence of benzene under low pressure con-
ditions . 43
9. Quantum yields for the photolysis of S02-isobutane
mixtures under medium pressure conditions. ^9
10. Quantum yields for the photolyses of S02-isobutane-C6H6
mixtures under medium pressure conditions. 5^4
11. Quantum yields for the photolyses of S02-isobutane-C02
mixtures under medium pressure conditions* 57
12. Estimates of the apparent rate constant for the reaction k. 68
13. Rate data for the reaction, CH302 + S02. 73
14. Extimated rates of S02(3B1) generation by solar radiation
in the lower troposphere. 85
XI
-------
1^. Percentage of 302(3B1)-quenching by various atmospheric
gases in air at 25°C and 1 a tan. 86
16. SGp^BjJ "chemical" quenching rate constants for various
atmospheric components and impurity species in the overall
reaction shown. 91
17. The theoretical rate of reaction of S02(3BX) reactions
with various impurity species and 02 in a hypothetical
sunlight-irradiated lower troposphere. 9^
18. Enthalpy changes and recommended rate constants for
potentially important reactions of ground state S02
and SOs molecules in the lower troposphere. 95
19. Experimental data from the photolysis of formaldehyde at
3130A and 25°C.
20. Experimental data from the photolysis of formaldehyde in
the presence of trimethylsilane.
21. Experimental data from the photolysis of formaldehyde
in the presence of isobutene. 152
22. The rates of H2 formation in formaldehyde photolysis with
small amounts of added isobutene. 153
23. Experimental data from the photolysis of formaldehyde in
the presence of nitric oxide. 156
24. The effect of isobutene on hydrogen formation in the
photolysis of formaldehyde at 25°C and with excitation
at 2890 to 3380A. 166
25. The effect of C02 on hydrogen formation in the photolysis
of formaldehyde at 25°C. 167
26. The wavelength dependence of the primary quantum effici-
ency of the photodecomposition of formaldehyde into H
and HCO. 169
27. Estimated apparent first order rate constants for
sunlight absorption by CH20 and for photodecomposition
by reaction 1 in sunlight in the lower atmosphere. 175
28. Experimental data from the photoJ^rsis of CH20-02 mixture
at 3130A and 25°C. 185
29. The effect of absorbed light intensity on the product
quantum yields in the photolysis of CH20-02 mixtures. 186
xxi
-------
30. Experimental data from the photolysis of CH20-02-C02
mixtures at 3130A and 25°C. 18?
31. Summary of the reaction mechanism and rate constants
employed in the simulation of the CH20, Q2, and C02
mixture photolyses. 194
32. Dinitrogen trioxide absorption cross sections. 205
33- Nitrous acid absorption cross sections. 207
3^-. Values of the rate constant for the photolysis of dimethyl-
nitrosamine obtained from the initial slope of the logarithm
of the dimethylnitrosamine concentration as a function of
time. 238
35• Extinction coefficients, quantum yields and values of the
light intensity within the cell used in the determination
of the photolysis rate of dimethylnitrosamine. 2*4-1
36. Calculated values of B and B' used in the estimation of
k4/k3. 254
37. Calculated values of C and Cf used in the estimation of
k5a/k3 and k51)/k3. 257
Xlll
-------
SECTION 1
INTRODUCTION MD CONCLUSIONS
OBJECTIVES
There are several aspects of the chemistry of the polluted atmosphere
which remain ill-defined. The research effort supported by this grant was
designed to elucidate some of the remaining key problem areas. (1) The
atmospheric removal and transformation mechanisms for sulfur dioxide remain
a major area of concern and interest. In spite of the rather general
concensus that sulfuric acid, ammonium sulfate, and ammonium bisulfate are
ultimate major sinks for atmospheric sulfur dioxide, the mechanism by which
the S02 is transformed to these and other products is not clear. The research
on this grant has provided some new insights into these processes and some
important information related to the S02 tropospheric reactions.
(2) In recent years it has become increasingly clear that the aldehydes
are significant participants in the chemical transformations which occur
within the troposphere. Thus the photodissociation of formaldehyde and the
higher aldehydes generates H-atoms and HCO and alkyl free radicals:
CH20 + hv -*- H + HCO
ECHO + hv -*• HCO + R
Each of these radicals can generate H02 radicals by reactions with oxygen:
H + 02 (+M) *~ H02 (+M)
HCO + 02 -*- H02 + CO
E + 02 ->- R02
R02 + NO -* • RO + N02
RO + 02 -*- H02 + R'CHO
The H02 and RC^ radicals can initiate the chain photooxidation reactions of
NO to N02:
H02 + NO >~N02 + HO
HO + RH -*- HgO + R
1
-------
R + 02 ->- R02
R02 + NO ->~ RO + N02
RO + 02 -*- R'CHO + H02
Since the [N02]/[NO] ratio is increased by the occurrence of these reactions,
03 builds up in the sunlight-irradiated atmosphere as the N02-NO-03 reaction
sequence occurs:
N02 + hv -»~ 0 + NO
0 + 02 (+M) ->- 03 (+M)
03 + NO -*- 02 + N02
Although it has been recognized for years that the aldehydes may play an
important role in the initiation of chain reactions responsible for photo-
chemical smog, the quantitative aspects of this thesis, so necessary for
accurate., realistic modelling efforts and control strategy development,
remain unevaluated. In particular, the quantum efficiency of the radical
formation from CH20 and RCHO, irradiated at the wavelengths of sunlight
•within the troposphere, have not been characterized well. In this work
several studies have been completed which help define these important
parameters.
(3) The nature of the chemical transformations among the nitrogen-
containing molecules such as NO, N02, HONO, HON02, H02N02, NHs, Rfflg, R2NH,
etc., are directly related to the NOX balance and ultimately to the 03
generation potential within the troposphere. The amines are of special
concern in our considerations of tropospheric chemistry by virtue of their
potential transformation to the highly carcinogenic compounds such as the
nitrosamines and the nitramines. Our work of the past year has provided
some new quantitative kinetic information which should allow a realistic
evaluation of the significance of nitrosamine and nitramine generation by
free radical reactions in the troposphere.
The following section A-2 provides a brief summary of our several
findings. In the section B, Parts I, II, III, the detailed reports on each
of the several aspects of our studies are given.
SUMMARY OF FINDINGS DURING THE PROJECT PERIOD
a) Several key points related to the tropospheric removal mechanism of
sulfur dioxide have been established in this work:
i) Through the use of laser excitation of S02 at selected wave-
lengths within the region, 2662-3273 A, we have defined the lifetimes and
quenching rate constants for the S02(1B1) and S02(1A2) excited states of
S02 with various atmospheric gases. The S02(XA2) state is quenched extremely
effectively by all of the molecules studied. For example N2 and 02 quench
-------
this molecule with rate constants: 1.9 x 1012 I/mole-sec, respectively.
Die S02(1B1) species is quenched less efficiently; for example, k = 3.5 x
1010 and 3.4 x 1010 I/mole-sec by N2 and 02, respectively. These data
suggest that S02 excited by sunlight within the troposphere will be quenched
to other longer lived states of S02 very quickly. The S02(3B1) state is a
dominant one which is clearly involved in S02 photochemistry.
ii) In a further quantitative study we have established the
mechanism of the reactions of the singlet [S02(1B1)j and triplet [S02(3B1)j
states of S02 with isobutane. The reaction in the absence of a large excess
of quencher molecules involves both the S02(1Bi) and the S02(3BX) species
with rate constants: 8.4 x 109 and 8.7 x 10s I/male- sec, respectively.
iii) The reaction of CHs02 with S02 was studied using kinetic
flash spectroscopy to monitor the CH302 radical directly. The results give
the rate constant for the reaction CHs02 + S02 -**• Products , k = (3.2 ± 0.7)
x 10s I/mole- sec. In the same study we derived the rate constant for the
reaction, CH302 + CHs02 -+- Products, k = (2.4 ± 0.1) x 10s I/mole-sec.
iv) All of our existing data related to S02 atmospheric transforma-
tion paths and that of the literature were reviewed carefully in another
phase of 'our work, and a quantitative evaluation has been made of the sig-
nificance of various pathways for S02 reaction in the troposphere. The
results suggest that the oxidation occurs largely by way of the reactions :
HO + S02 (+M) -*- HOS02 (+M) (a); H02 + SOS -*- HO + S03 (b); CHs02 + S02
GHsO + 80s (or CHsOOS02) (c). Through computer simulations of the chemistry
within both "clean" and polluted atmospheres, the theoretical rates of S02
conversion were estimated. Our present evaluation for the case of the
highly polluted atmosphere, suggests the three reactions (a), (b), and (c)
may all occur and with near equal rates. In these cases, the total of the
rate corresponds to an S02 removal as high as 4$/hr. The many remaining
uncertainties and alternatives in the mechanisms for S02 atmospheric removal
have been discussed in detail in an attempt to provide a reasonable basis
for the planning of the further studies.
b) In three detailed kinetic studies which were carried out on this
grant, we determined the quantum efficiency of the two photochemical
decomposition paths in formaldehyde in experiments at a variety of wave-
lengths :
CHpO + hv -*- H 4- HCO (I)
+ hv ->- H2 + CO (II)
The kinetics and the quantum yields of the products of the photolysis of
CH20-M), CI^O-MesSiH, CH20-iso-butene, and CH20-02, CH20-02-C02 mixtures have
been studied. The wavelength dependence of the absolute quantum yields of
the prijnary processes (l) and (ll) have been established. From these results
and solar irradiance estimates, the apparent first order rate constants for
CH20 dissociation by process (l) at various solar zenith angles (in parentheses
parentheses) are: 2.31 x 10-3(0°); 2.17 x 10-3(20°); 1.71 x 10~3(40°);
-------
0.92 x 10~3(60°); and 0.1? x 10~3 min"1(78°). In further studies of CH20-02
mixtures we discovered an unexpected reaction leading to the formation of H2,
CO, and HC02H in a chain reaction. From the unusual kinetics observed we
concluded that the reaction involves H02 and/ or HO-radical addition to CH20.
These considerations bear directly upon the mechanism of transformation of
formaldehyde to formic acid with the polluted troposphere.
c) Two detailed studies have been made in this grant period of the
atmospheric reactions of some important nitrogen-containing species.
i) The absolute extinction coefficients of nitrous acid vapor have
redetermined. The large disagreement between the previous literature values
has been resolved, and apparent first order rate constants for the reaction
(HI),
HONO + hv ->- HO + NO (ill)
in the sunlight-irradiated lower atmosphere have been derived for various
solar zenith angles (in parentheses): 0.089 (0°); 0.086 (20°); 0.077
0.05^ (60°); 0.017 min'M 78°).
ii) In the second phase of our work on nitrogen compounds of
atmospheric interest, we have used the long-path FT- IRS photolysis system to
study the reaction modes of atmospheric generation of the (CHs)2N radical
and the kinetics of its subsequent reaction with 02, NO, and N02: (CHs)2N +
02 ->- CH2=NCH3 + EOs (d); (CHs)2N + NO •*- (CHs)2NNO (e); (CHs)2N + N02
(CHs)2NN02 (f). The rate data give the following preliminary estimates of
the rate constant ratios near 25°C: k /k = (1.48 ± 0.072) x 10~6; k /k =
U, o LL i
(3.90 ± 0.28) x 10-7. The very large reactivity of the (CHs)2N radical with
NO and N02 compared with that with 02 suggests that the alkyl amino radicals
formed by HO abstraction from amines present in the atmosphere have a
reasonable chance of forming nitrosamines and nitramines even though the NO
and N02 impurities are at concentrations in the pphm range.
-------
SECTION 2
STUDIES RELATED TO SULFUR DIOXIDE REMOVAL MECHANISMS IN THE ATMOSPHERE
KINETICS OF FLUORESCENCE DECAY OF S02 EXCITED IN THE 2662-3273A REGION
Introduction
A thorough understanding of the photochemistry of S02 excited within the
first allowed electronic absorption region, 2500-3^00 A, has been impossible,
since the nature of the excited species and their mechanisms of population
and decay are poorly characterized today. The spectroscopic assignment of
this band of S02 has puzzled spectroscopists for many years. Both the 1B1
and the 1A2 states of S02jare predicted to lie in this energy region,"5 and
although only the ^-E^ -*- X 1Ai transition is formally allowed, the vibronical-
ly induced transition to the *A2 state may also occur. The Renner-Teller
coupling between the 1B1 and % ^A.^ states of S02, the Jahn-Teller interaction
between the ^Bj. and 1A2 states, and the spin-orbital coupling between the
triplet states (3AS, 3B";L, and B2) and singlet states (1B1 , 1A2), make the
singlet electronic transitions in the 2500-3^-00 A region extremely difficult
to analyze. It was generally believed that this absorption system belonged
to the allowed 1B1 • <- X 1A1 electronic transition.6"9 However Dixon and
Halle10 did a partial rotational analysis of the 33^-0 A band, and they sug-
gested that it corresponded to a vibronically induced *AQ -+• X 1A.1 electronic
transition. Hamada and Merer 11s12 carried out a much more detailed
rotational analysis of twelve bands of S1602 and two bands of S1802 in the
3000-3^-00 A region; they concluded that all of the obvious banded Ejtructure
in this wave length region could indeed be assigned to the 1A2 . X ^A^
transition and that its origin was near 35^0 A. Analyzable vibrational or
rotational structure which might be attributed to the ^^ -*- % 1A1 transition
has not been observed, but on the basis of the other less direct evidence at
hand, Hamada and Merer12 tentatively suggested that the band origin of the
1B1 -< X*^ transition lay between 3100 and 3160 A. Brand^ et aJL.13 have
rationallized the observed Zeeman effect within the 1A2 -4 X^Ai band system
in terms of coupling of the 1A2 state with a background of interacting
vibrational levels of the ground state and the low lying states of the
triplet manifold.
Recently Shaw, jet al., have carried out an analysis of the fluorescence
of S02 from single vibronic level excitation in the 3150-3080 A14 and the
3070-2930 A15 regions. The qualitative conclusions from the data support the
view that the S02 (1B1) excited state, with an origin near 3l6o A, provides
-------
the r.iajor component of the emitting levels. In time-resolved fluorescence
studies of 302 excited in the 2617-3350 A region, Brus and McDonald16 and
Butler and McDonald17 found double exponential decay of the excited S02•
long-lived (L) state and the short-lived (S) state which contribute to this
emission were attributed tentatively to the 1B1 and XA2 states of S02,
respectively. McDonald and co-workers concluded that the ^-Bj state of S02
could be populated even at wavelengths as short as 3350 A, in seeming con-
tradiction of the tentative band origin assignment from the data of Hamada
and Merer12 and Shaw, et al.l4»15 It was not possible to choose between
alternative mechanisms of independent generation and decay of the L and S
species and that involving generation of one from the other by way of first
order or second order, collisionally perturbed processes.
Heicklen and his co-workers18~21 have presented a great deal of indirect
evidence for the existence of some other ill-defined, non-emitting singlet
state of S02 which they designate as S03*. This hypothesis was formulated
in view of the observed kinetics of product formation resulting from excited
S02 interactions with various reactants as well as in S02 and biacetyl
emission studies in various mixtures using steady illumination. These workers
have concluded that excitation of S02 within the first allowed singlet band
leads directly to the S02* state; they speculated that its radiative lifetime
was about 6 x 10~7 sec, corresponding to that expected from simple theory
relating lifetime and integrated absorption data for the first excited singlet
band of S02.x '22 They further suggested that the emitting singlet,
presumably the long-lived state usually seen by previous workers, was formed
by internal conversion from the S02* species. However there is no corre-
spondence between the observed kinetic properties of the S02* state sug-
gested by Heicklen, e_t al., and the Stern-Volmer quenching parameters derived
for the emitting states of S02. Obviously firm conclusions concerning the
true nature of the excited singlet states involved in the photochemistry of
S02 are not now possible from the information at hand.
This study was initiated in an attempt to define better the kinetic
properties of the first excited singlet states of S02. We have excited S02
at discrete wavelengths in the range 2662-3273 A employing a quadrupled
neodymium laser and a powerful, frequency doubled, tunable dye laser. We
have studied the fluorescence decay kinetics of S02 at finite pressures and
under essentially collision free conditions, and in S02-added gas mixtures.
Some significant new evidence has been obtained which provides a clearer
understanding of the kinetic mechanisms which control the generation and
decay of the singlet emitting species in S02 excited at wavelengths in the
2662-3273 A range.
Experimental Section
The experimental apparatus used in most of these studies was similar
to that employed by Brus and McDonald.16 A flashlamp pumped dye laser,
Candela model ED-66, was used to excite S02 molecules at low pressures
(0.05-12 mTorr) with a light pulse covering a narrow band of wavelengths
(~1 A). With a Rhodamine-B (Eastman No. 4^53) solution (10~4M) in ethanol,
this laser produced approximately kOO kW of light at 6260 A, or approximately
150 kW of light at 6522 A with an optimum mixed solution of Rhodamine-6G
-------
(Eastman No. 1072^ and Cresylviolet (Eastman No. 11884) in ethanol. Kiton
Red (Exciton No. 620) in methanol was used in some experiments to obtain a
better output around 6^00 A. The pulse length of this laser was about 200
nsec. These laser pulses were frequency-doubled with a KDP crystal to give
the corresponding uv-output with a power of 3 to 20 kW. The dye laser
system was employed to secure pulses of near monochromatic ultraviolet
radiation over the wavelength range, 2975-3273 A. Wavelength determinations
were performed with a Jarrell-Ash 3.U m Ebert spectrograph.
In all of the lifetime studies at low pressures we used a cell of 22 S,.
volume, similar to that employed by Sackett, et al.23 Two cylindrical
sidearms with Brewster angle windows of Suprasil were fused directly to the
spherical cell body through quartz-Pyrex graded seals; these served as the
entrance and exit routes for the exciting laser light pulse. The uv laser
pulse passed an appropriate filter (Corning 7-51)-) to exclude the fundamental.
At 90° to the excitation path a third sidearm, 3«5 in. long, was fused to
the cell; it had a 2 in. diameter Suprasil window to which an RCA 7265
photomultiplier was positioned as closely as possible. Various cut-off
filters were inserted between the window and the photomultiplier; the
transmission of these is shown in Figure 1. The fluorescence signal from the
photomultiplier was monitored via a Tektronix model 770^ oscilloscope, and
stored on a Biomation model 6lO transient digitizer. With the help of the
interactive light pen display of a PDP-7 computer interfaced with a Datacraft
602^ computer for computations, the data tape could be transformed into
semi-logarithmic form, displayed on the screen, judged to be a double or
single exponential decay, and resolved by least squares computer fit ac-
cordingly.
100
CD
50-
300
400 500
Wavelength, nm
650
Figure 1. Comparison of the wavelength dependence of
the transmission of filters used in front of the
fluorescence intensity detector in this study, the
photomultiplier response, and a typical S02 fluorescence
envelope for S02 excited at 3133 A [12]. Curve k
represents the relative photomultiplier response.
-------
In the quenching experiments of the long-lived singlet with various
reactant gases at 2662 A, a separate laser system and cell were employed,
and a 20 nsec pulse of 2662 A light was generated using a neodymium laser
source. The oscillator was a 9 in. neodymium-doped glass rod, excited by
three linear xenon flash lamps placed symmetrically about the rod; these
provided an input of about 4000 J. Q-switching was accomplished with a
rotating prism (24,000 rprn), and the laser beam was polarized by an angled
Pyrex plate. A dielectric coated mirror completed the oscillator cavity.
The peak intensity of the beam from the oscillator was of the order of 10 MW
at a wavelength of 1.064 u. A ten-fold increase in peak intensity was
obtained by use of an amplification system. This consisted of a 12 in.
neodymium-doped glass rod which was excited by a 700 J linear xenon flash
lamp. The Q-switeh was used to time the flash of both the oscillator and
amplifier flash lamps; the amplifier flash was delayed with respect to that
of the oscillator lamps by an electronic network. Both oscillator and
amplifier were forced air cooled. The 100 MW 1.064 M- laser beam was twice
frequency doubled to produce the desired uv pulse. The first frequency
doubling was effected by a precisely oriented potassium dihydrogen phosphate
(KDP) crystal. Any scattered flash lamp ultraviolet light was filtered from
the beam by a Corning 7-5.7 filter, and the resultant beam contained about
1 MW of peak power at 5324 A, together with unchanged fundamental neodymium
frequency. The latter was removed by a Corning 4-96 filter which also
served as a beam splitter. The light from the beam splitter activated the
oscilloscope trace by means of a phototube (EGA 922) and allowed the
recording of the fluorescence from the reaction cell. For doubling the
green light, an ammonium dihydrogen phosphate (ADP) crystal was chosen to
minimize absorption of the ultraviolet frequency. The arrangement employed
generated about 10 k¥ peak power at 2662 A with a half-intensity peak
duration of 20 nsec. A Corning 7-54 filter placed before the sample cell,
removed the green radiation from the beam. The fourth harmonic of the
fundamental Nd centered at 2662 A was found to have a band width of 1.2 A
for our system. In these studies at relatively high pressures where only the
long-lived singlet emission was followed, the reaction cell was constructed
from a Pyrex tube, 25 mm in diameter and 88 cm long, with Brewster angle
windows of Suprasil sealed to the ends. The emission beam was filtered
with a Corning 7-60 filter and the intensity of the fluorescence was
monitored at right angles to the excitation beam.
The grease-free, mercury-free, high vacuum system employed was equipped
with Teflon and metal valves (Granville Phillips). Pressure was measured
with a MKS Baratron capacitance manometer (MKS-XH). The S02 reactant used
in this study was purchased from the Baker Chemical Co. The condensed gas
was distilled from trap to trap at -196°C, retaining only the middle third;
this was degassed many times. In the quenching experiments we used high
purity gases from Phillips Petroleum Company (isobutane, trans-2-butene,
cis-2-butene, methane), Fisher Scientific Company (Benzene), Matheson
Company (argon, oxygen, nitric oxide), and Airco (carbon monoxide, carbon
dioxide). All gases except nitrogen were distilled on the vacuum line and
only the middle fraction was retained.
-------
Results and Discussion
The Double Exponential Decay of S02 Fluorescence—
The time resolved emission from S02 excited within the first allowed
band shows a most interesting behavior first observed by Brus and
McDonald.16 Tupical fluorescence decay curves for S02 excited at 3020A,
0.97 mTorr pressure, and 27°C, are shown in Figure 2. Only the time scale
of the observations has been changed in the different experiments shown in
order to accent the nature of the short- and long-time decay regions and the
degree of the fit of the data to the kinetic equations employed. Clearly the
10 so
Time, usec
20 100
Time, usec
40 200
Time, jusec
100 500
Time, jjsec
Figure 2. Time resolved fluorescence decay curves
of S02 excited at 3020 A and 0.97 mTorr. The solid
lines are calculated assuming a double exponential
decay as outlined in the text.
fluorescence intensity does not follow a simple, single exponential decay.
However the results are described well using a double exponential decay
formula :
I = I -e
o/^s
*i°*
-t/n
(1)
or
-t/Tq -t/i
+ e
L
where Ta^dT"-,, respective^, are the total" intensity of the fluorescence
X fci-1-
-------
ana the relative fluorescence intensity at time t, and Ig° and IL° are the
initia^. intensities of the short-lived and long-lived components, respec-
tively. The solid curves shown in Figure 2 have been calculated from
equation (2) using a single set of parameters for the components of emission:
I/TS = 6.93 x 104; I/TL = 8.46 x 10g; (IS°/IL°) = 1-36. Obviously the locus
of the data points are reproduced well assuming the double exponential decay
of the fluorescence, in confirmation of the observations of Brus and
McDonald.16 Failure to observe this phenomenon in our previous study at
wavelength 2662 A24 was the result of the wavelength choice and the relatively
high pressure range employed (2-80 mlorr) in this study. Even at low pressures
of S02 the short-lived component of the fluorescence is not readily detected
using 2662 A excitation. This is confirmed by the fact that Brus and
McDonald did not attempt to derive Stern-Volmer quenching data related to the
short-lived species only in their experiments at 2617-4 and 2715*1 A.
All of the fluorescence intensity-time data observed in this work could
be fitted well in terms of relation (2) to derive rg, TL, and Ig°/IL° values.
It should be noted that a third exponential component to the fluorescence
decay is observed at very long times; this emission which originates from
the S02(3B1) state is the subject of a further study in a subsequent report.
The Lifetimes and Stern-Volmer Quenching Parameters for the Short-Lived
(S) and the Long-Lived (L) Components of the S0g Fluorescence-- The values of
I/T and I/TT determined at each excitation wavelength (2975-3273 A) were
b L
plotted versus the S0a pressure over the range 0.5 to 12 mTorr. Typical data
obtained in experiments with the excitation at 3107 A are shown for the L and
S components in Figures 3 and 4, respectively. It is seen that the emission
from the S component shows Stern-Volmer behavior down to the lowest pressure
at which data could be obtained. The data from the L component also shows
Stern-Volmer behavior down to a pressure of about 1 mTorr; below this pressure
the I/TT versus P plot shows upward curvature. We will consider this
J-l DUg
deviation later and merely note here that this appears to arise from dif-
fusional loss of the L species from the observation zone during their life-
times.
Stern-Volmer quenching constants for each of the emitting species were
derived from the linear portion of the 1/T versus pressure plots. These data
are summarized in Table 1, and they may be compared with those published by
Brus and McDonald16 in Figures 5-8. The error bars shown for both sets of
data correspond to ± twice the standard deviation. It is seen that the two
studies are in general agreement within the error limits for most of the
parameters determined. The one apparent difference between the two studies
is the lifetime of the L species excited near 31833 cm"1; see Figure 5. Thus
Brus and McDonald found a lifetime of 617 ± 103 M-sec for excitation at
314-0.5 A, while we observed a value of TL = 220 + 30 M-sec in the same general
wavelength region at 314-3 A. The effect may result from the selective
excitation of a peculiar, narrow region of S0£ states for which perturbations
of near lying states are less important; the band width of the laser light
employed by Brus and McDonald is somewhat narrower (~3 cnr1) than that which
10
-------
0 I I I I I I I I I
Figure 3. Stern-Volmer plot of the long-lived (L)
component of S02 fluorescence excited at 3107 A.
p mTorr
Figure 4. Stern-Volmer plot of the short-lived (S)
component of S02 fluorescence excited at 310? A.
11
-------
we employed (~10 cm"1). However this difference may in part result from the
choice of cut-off filter which was placed before theodetector inothe Brus
and McDonald experiments; they switched from a 3100 A to a 3550 A cut-off
filter in experiments at SlJ-J-0.5 A excitation and the longer wavelengths.
The influence of the apparent lifetime on the choice of fluorescence wave-
lengths monitored for excitation at a given wavelength can be seen in the
data of Tables 2 and 3. We found that the observed TL° values varied
from extremes of 22^ to 6k6 usec with a 3300 A or a 6000 A cut-off filter
in place.
In both the study of Brus and McDonald and the present work the
observation was made that the apparent lifetime of the L component increased
as the observed fluorescence was restricted to the longer wavelength region.
Since the emission intensity-time profile of the S species is obtained by
subtracting the intensity of the L state extrapolated from the data at long
times, the T , and I °/I ° values are somewhat dependent on the choice of
S- S L
filters employed; see Table 3- In general the Tr,° and the quenching rate
b
constants for the S and L species all decrease as the fluorescence is
observed at increasing wavelengths. As Brus and McDonald have pointed out
previously, the filter effect on the Stern-Volmer parameters indicates that
the L and S species created at a given excitation wavelength, must be a
group of rovibronic states with a range of TO° and T values. Therefore
o li
the lifetimes and quenching rate constants obtained from the Stern-Volmer
plots must be only an average value of this group of states. To overcome
the potential problem of enhancement of the fraction of the L or S fluorescence
envelope observed, we have employed a cut-off filter and photomultiplier
response which maximized the width of the band of wavelengths monitored in
each experiment. See Figure 1.
The dependence of the apparent values of Ia°/Ir° on the choice of cut-
fa j_i
off filter suggests that the emission spectrum of the S component is some-
what shifted to longer wavelengths than the L component. If one accepts the
premise that the S and L components arise from the 1A2 and i'B± states of S02,
respectively, then, the minor differences in emission profiles can be
rationalized in terms of the differences in the geometry of the ^A2, 1B1,
and X, ^-AX states of S02 predicted in theoretical studies.3"5 Thus Lindley5
has estimated that the S-0 bond distances in the ^-As and 1B1 states are 1.50
and 1.49 A, respectively, compared to l.ifO A in the ground state; the 0-S-O
angles in the ^A^, 1B1, and ground states, respectively, are 92.6°, 117.3°,
and 119°. The relatively large difference between the equilibrium 0-S-O
angles in the X 1A1 and the XA2 states is expected to enhance the transition
probability for radiative decay of the XA2 to the higher vibrational levels
of the ground state, and hence shift the fluorescence envelope to somewhat
longer wavelengths than the •'•B! emission.
Vibrational relaxation of the long-lived excited S02 molecules, first
observed by Sidebottom, et al.,24 and confirmed by Brus and McDonald,16 was
also observed in this work. The effect arises largely from the variation in
quenching rate constant and radiative lifetime with excitation energy; see
Table 1 and Figure 10. As an excited molecule is vibrationally relaxed to
12
-------
SOOp
L
0
30 35
Excitation -nergy , CTT.
35
Excitation Energy, cm"', xlO
39
Figure 5- Plot of the radiative life-
time of the long-lived (L) fluoresc-
ing species versus excitation energy
of S02. Closed circles - this work;
open circles -
Figure 7. Plot of the rate constant
k.4 for the quenching of the L state
by S02 versus the excitation energy.
Closed circles - this work; open
circles -
120
u
cs
(/I
a
>_^ 60
-
""
~*
I .
k |
i ,
A
If
1
>
(
• <
1 1
1
1
t
,f
(
|
) I
*
! i
31 33
Excitation Enengy. cm ,'
35
O
'•10
i
i
30 32 34
Excitation Energy, crrr, X10"
Figure 6. Plot of the radiative life-
time of the short-lived (S) fluoresc-
ing species versus excitation energy
of S02. Closed circles - this work;
open circles -
Figure 8. Plot of the rate constant
ks for the quenching of the S state
by S02 versus the excitation energy.
Closed circles - this work; open
circles - [l^f ].
13
-------
TABLE 1. LIFETIMES AND QUENCHING RATE CONSTANTS FOR THE LONG-
LIVED (L) AND SHORT-LIVED(S) SPECIES EXCITED IN S02 AT VARIOUS
WAVELENGTHS
0
X , A
exc
2662
2975
2998
3020
3046
3065
3107
3129
3143
3196
3211
3225
3235
2251
3273
TT°, sec, Ts° » sec«
Lx 104 x 105
2.0
2.4
3.0
1.8
3.0
3.2
2.2
2.2
2.0
2.5
1.4
1.5
1.1
1.8
± 0.1
± 0.2
t 0.6
± 0.1
± 0.4
± 0.2
± 0.3
± 0.3
± 0.3
± 0.4
± 0.3
'± 0.5
± 0.3
± 0.5
4.3 ±
1.7 ±
8.3 ±
4.0 ±
4.0 ±
3.2 ±
2.4 ±
4.1 ±
1.7 ±
1.9 ±
2.5 ±
2.7 ±
1.8 ±
2.8
0.6
10.6
3.8
1.0
2.2
0.7
2.8
0.8
0.8
2.2
1.6
1.6
. . . -1 -1
k^, -(..mole sec,
x 10
3.8
7.2
7.9
8.5
8.9
9.0
9.1
10.4
10.2
9.7
9.8
9.4
11.3
10.5
8.4
± 0.
± 0.
± 0.
± 0.
± 0.
± 0.
± 0.
± 0.
± 0.
± 0.
± 0.
1
1
1
2
1
1
1
2
1
2
5
± 0.2
± 0.
± 0.
± 0.
7
4
4
x
, -1 -1
.mole sec,
io~n
10.5 ±
7
10
8
7
7
7
6
5
4
6
8
6
.9 ±
.2 ±
.5 ±
.8 ±
.6 ±
.2 ±
.4 ±
.4 ±
.8 ±
.1 ±
.4 ±
.9 ±
1.1
0.7
1.0
1.3
0.5
1.4
0.6
0.8
1.1
0.7
2.2
1.6
1.3
Filter13
d
c
c
c
d
c
c
d
d
d
c
b
d
d
f
Ttenperature, 25 ± 2°C; error limits shown here and elsewhere throughout this work are
± twice the standard deviation. ^Transmission curves for the filters referenced
here are shown in Figure 1; these were placed before the fluorescence detector.
-------
TABLE 2. STERN-VOIMER PARAMETERS FOR TEE L AND S COMPONENTS
__ ct
AS A FUNCTION OF THE FLUORESCENCE WAVELENGTHS MONITORED
V- s
Lx ]
2.1 ±
2.8 ±
2.2 ±
2.9 ±
2.5 ±
3.1 ±
6.5 ±
sec.
2.5
2.5
0.3
0.8
1.2
1.0
4.2
T •. sec,
Sx 105
2.4
4.2
3.2
3.3
2.3
2.2
2.4
±
+
+
±
+
+
+
8.4
4.6
2.2
0.7
0.9
2.2
0.6
k., -t.mole^sec"1 k , ^.mole~1sec~1 Detector Filter13
x 10~i0 J x 10~U
10.1
10.0
10.4
7.9
7.7
7.8
6.6
± 0.5
± 0
± 0
± 0
± 0
± 0
± 0
.6
.2
.3
.6
.3
.3
9.5
9.2
7.6
8.4
7.5
7.9
5.9
+
+
±
+
+
+
+
4.3
0.4
1.4
0.2
0.5
0.4
0.5
(a)X & 3200 A
(b)X ^ 3300 A
(d> 3400 '* X £ 3800 A
(e)X ^ 3700 A
(h)X * 3800 A
tg)X * 3850 A
(j)X * 6000 A
Excitation wavelength, 3130 A; temperature, 25 ± 2'C; See Figure 1 for complete
transmission curves of the filters employed which are designated by the letters.
-------
TABLE 3. RECIPROCAL LIFETIMES OF THE S AND L COMPONENTS AND IS°/IL° AS A
FUNCTION OF THE FLUORESCENCE WAVELENGTHS MONITORED
-1
I/TL» sec ,
x 10~3
16.9 ± 0.3
16.5 ± 0.2
15.7 ± 0.3
15.4 ± 0.3
14.3 ± 0.3
14.5 ± 0.5
13.0 ± 0.3
12.6 ± 0.2
10.7 ± 0.1
11.4 ± 0.1
1/Tg,
X
10.9
10.6
12.4
13.9
8.9
9.4
12.1
11.4
13.3
12.4
sec ,
io-4
± 0.6
± 0.4
± 0.5
± 0.9
± 0.5
+ 0.7
± 0.8
± 1.2
± 0.1
± 0.2
V'
.1.08
1.14
1.16
1.58
0.97
0.95
1.42
1.40
2.22
'I ° T , psec/ Detector Filter0
L
div.
to
(b)X > 3300 A
± 0.07 10 "L (e)X > 3700 A
± 0.13 10
± 0.07 20
± 0.11 20 J
± 0.13 10 "V (i)X > 4800 A
± 0.17 20 J
± 0.04 10 T (j)X > 6000 A
20
Excitation wavelength, 3130 A; Pso2 * 2-96 nfforr in all runs except those with
filter (j) where 2.85 mTorr was used; bsweep time of the oscilloscopic trace
employed; see Figure 1 for complete transmission curves of the filters.
16
-------
the lower vibrational levels, its larger quenching rate constant and some-
what shorter radiative lifetime, increases its chance for decay; thus in
theory the slope of the Sn IL versus time plot should become increasingly more
negative with increasing time. At sufficiently high pressures (p > 8 mlorr)
where collisional deactivation is significant, this vibrational relaxation
phenomenon can be seen in the emission of SO-,, excited in the 2662-310? A
region; typical data from S02 excited at 2998 A are shown in Figure 9. The
phenomenon is almost undetectable at 8 mTorr for 3211 A excitation, where
the original vibrational excitation is near that of the completely relaxed
molecule. From the data of Figure 9 and similar decay curves from excitations
at other wavelengths, we can estimate the average internal energy removed
per collision. From the change in slope of the Urn. I versus time plot, and
the initial rates of decay of the L component anticipated for the S02(8.08
mTorr) excited at various energies (See Figure 10), we estimate that the
approximate loss in energy (kcal/mole) per collision is: 1.2 (2998 A);
0.8 (3020 A); 0.9 (3065 A); 1.3 (310? A). Thus among those excited S02
molecules which do not suffer electronic quenching on collision, about
1 kcal/mole per collision appears to be lost by S02(L) species excited in
the 2998 to 310? A region.
The Diffusional Loss of Excited L Molecules--
In their earlier study of the time dependence of the fluorescence of the
L and S states of S02, Brus and McDonald16 considered the kinetics of the
emission in terms of two contrasting mechanisms: (l) essentially independent
population and decay of the L and S states; (2) collisional conversion of one
state into the other (S L). They used the variation in the I °/I ° ratio
b Jj
as a function of pressure in an attempt to choose between these mechanism.
They felt that their data obtained at 3225,9! provided the most sensitive
test of the mechanism. From these data they concluded that an upper limit
of 15$ of the S state deactivating collisions gave L state molecules for ex-
citation at this wavelength; that is, the L and S states decayed essentially
independently of one another. Data from the shorter excitation wavelengths
were not accurate enough over a wide pressure range to provide a choice
between the mechanism alternatives. The main difficulty in the experimental
test of the mechanism lies in the necessary limit in the low pressure range
in which the experiments can be performed. Diffusional loss of the excited
molecules and the very low initial population of the states at low S02
pressures, limit the useful range of pressures. As we noted previously in
Figure k, the S state with its very short lifetime, follows Stern-Volmer
behavior down to the lowest limit of pressures which we could employ (0.05
mTorr). This is expected since essentially none of the S species is lost
from the observation zone by diffusion during their very short lifetimes.
However the L state shows a departure from Stern-Volmer behavior below about
1 mTorr: see Figure 3. Since the data below 1 mTorr are very important in
allowing a mechanism choice, we have attempted to understand the origin of
the upturn in the I/TT versus P0^ plot in this region. The increasing im-
L oU2
portance of L loss at the walls at low pressures may be the origin of this
effect. Low pressure deviations from the Stern-iVolmer relation for excited
S02(3B1) molecules were observed by Otsuka and Calvert,25 Strickler, et al.,26
and Briggs, et al.27 In view of the much shorter lifetime of the L component
of the excited S02 compared to that of the 3B1 state, one expects the dif-
/-
17
-------
I I I I I I I
20 50
Time, psec
80
Figure 9. Semilog plot of the intensity of the total
fluorescence of the S02 molecule versus time with ex-
citation at 2998 A and P0. = 8.06 mTorr; note the
bUg
downward curvature at long times which is thought to
result from vibrational relaxation of the L state.
'o
0
_J
£
T3
73
2
Q^
ON
X
__ N,
" "" o"" - - °
o
s 1 1 1 =L 1 1 1 L
34
Excitation Energy, cm-',
38
Figure 10. Plot of the calculated initial rate of
decay of the L species of S02 versus the excitation
energy; P = 8.08 mTorr. Closed circles - kinetic
QL>2
data from this work; open circles - data from
18
-------
fusional loss of the L species to become important in a ten-fold lower pressure
range than in the case of S02(3BX). Indeed this is the pressure range for
the L state deviations seen here. Theory requires that at sufficiently low
pressures the rate of diffusional loss of excited molecules from the
observation zone will be determined solely by their molecular velocities
and the geometry of the cell and the observation zone, provided that excited
molecules which reach the wall are deactivated effectively. At an S02
pressure of 0.9 mTorr and excitation at 310? A, an L molecule will undergo
about 1 collision during its lifetime, and its mean free path is about 4.8 cm.
Obviously below this pressure region we expect the molecular velocities of
the excited molecules to be a major factor in controlling the diffusional
loss rate. At somewhat higher pressures diffusion out of the observation
zone will be controlled largely by molecular diffusion and collisionally
controlled processes. In this latter case the kinetic theory the Brownian
motion can be applied.28 Ihe time t required for a molecule to diffuse a
distance x is given approximately by: t = 2f/D, where D is the pressure
dependent diffusion constant: D = a/p. Thus we would expect for these
conditions that the apparent first order constant for diffusion of excited
molecules out of the observation zone of average radius r~ will be given by:
k"diff ~ ^ V r2 ~ a-/??"2' Experimentally we can estimate k, . from the
measured lifetimes of the excited species at low pressures (T ) and that
anticipated in the absence of diffusional loss from an extrapolation of the
Stern- Volmer lifetime plot from the higher pressure region (T .. ): k =
(l/T ) - (l/T -, ). Thus if the molecular diffusion mechanism is operative
G2Cp GcUJLC
then we expect that (l/T ) - (l/T , ) to be a linear function of 1/p.
^ x ' exp' v ' calc' ' *
Such a plot is shown in Figure 11 for the data from the L state excited at
3107 A. It can be seen that the functional form of the highest pressure
regime (1/p -*~ 0) is near linear as anticipated from simple theory, but
the values became less dependent on the pressure as the pressure is reduced
(1/p •*- oo ) and collisionally controlled diffusion becomes less important.
The data of Figure 11 fit well the empirical equation: k^f-^360"1) =
5.0 x 103[1 - e-°-°6o/P(mTorr)]< If we interpret this equation in terms of
the simple diffusion model outlined, then the experimental limiting slope
of the k „ versus 1/p plot as l/p-^-0, is 300 sec~a rn'Rar"1. For gaseous S02
at 25°C we can estimate that D - 7-9 x 104/P (mTorr). Thus simple theory
suggests: 300 sec"1 mTorr'1 - 7-9 x 104-yr^ , from which r =* 16 cm, a
physically realistic result. At the low pressure limit (l/p->~oo), the
empirical experimental law gives k = 5.0 x 103 sec"1. Then according to
the simple theory, v/f ^ 5.0 x 103 see"1; taking v = 3-2 x 104 cm/sec,
r - 6 cm. This is in qualitative accord with the alternate estimate made
from the higher pressure limit and as good an agreement as one might expect
from the very approximate model assumed here. In S02(3B1) lifetime studies
carried out in this same system, the k^^ versus 1/p plot was linear as
expected for this longer lived species, and r ^ 10 cm was estimated.
Certainly the deviation of the L state lifetime data from Stern-Volmer
19
-------
V d
O
2 -
10
1/P mTorr
20
Figure 11. Plot of the apparent first-order rate
constant for the diffusional loss of the L state
[kdiff = (l/Texp) " (l/Tcalc)J V6rSUS the reciProcai
of the pressure of S02. The dashed line is the
calculated fit of the data by the empirical
equation given in the text.
Is/1?
2-
i
^
--^
. — --:::
x^--
...-.
i i i i
- - 3
-..-.:. :.:.--:•-. -_™5«
0 •
1 1 1
PSO2, mTorr
Figure 12. Plot of the ratio I °/I ° versus pressure:
b J_i
excitation wavelength 310? A. The curves shown are
the theoretical fits to the data assuming the S •*- L
conversion calculated from the nonequilibrium model
[eg. (2k)j. The parameters chosen to obtain the
various curves are outlined in the text.
20
-------
behavior at low pressures is consistent with the diffusional loss of the L
species from the observation zone, and it is reasonable to use the experjjnent-
aHy determined k • equation to estimate the diffusional loss term for L
at low pressures.
The Kinetics of the Formation and Decay of the L and S Components of S02
Fluorescence --
It seems highly probable that the two components of the S02 fluorescence
seen here involve two excited species as suggested by Brus and McDonald.16
It will be convenient to consider the results in terms of the following
general reaction scheme: ___
S02 + hv -*• S02(S) (Ia)
i
S02 + hv -*- S02(L) (Ip)
S02(S) + S02 •+- S02(L) + S02 (3a)
S02(S) + S02 ->- (2S02) (3b)
S02(L) + S02 -*- S02(S) + S02
S02(L) + S02 -*• (2S02)
S02(S) -*- S02 + hvc (5a)
S02(S) -*• S02(L)
S02(S) ,- (S02) (5c)
S02(L) -*- S02 + hvL (6a)
S02(L) •+- S02(S) (6b)
S02(L) -+- (S02) (6c)
In reactions (3b), (^b), (5c), and 6c above, (2S02) and (S02) symbolize
Stt2(X1A1), (3B!), (%) or other long-lived or non-emitting product
molecules. In terms of the above mechanism the following decay laws may be
written:
d[S]/dt = -kg[S] + kt/[Ll W
d[L]/dt = -kL[L] +kt[S] (8)
where kg = (^a + ^ [S02] + k?a + k^ + k^ (9)
kL = (Ha + kUb} [S°2] + k6a + k6b + k6c
21
-------
Solving the coupled differential equations (7) and (8) gives expressions for
the tir.'.e dependence of [S] and [L]:
[ S ] = Ae + Be
where C = (1/2) [ (k_ + k_) + u]
D J_j
| = (1/2) [(kg ^ - CO] (16)
andco=[(ks-kL)2 +tetktx]*
A = co-i {(1/2) (w + kg - kL)[S]0 - kt/[L]0) (17)
B = co-1 {(1/2) (co - k + k )[S]0 + k '[L] } (18)
D Ll u w
C = co-1 {-kt[S]0 + (1/2) (co - ks + kL) [L]0] (19)
D = co-1 (kt[S]0 + (1/2) (co + ks - kL) [L]Q] (20)
Clues concerning the relative significance of the various reactions in the
general mechanism outlined can be had from a consideration of the variation
of the I °/I ° ratio with pressure. Values of this ratio estimated for ex-
b L
periments at 3107 A are summarized in Table k . The plot of these data
shown in Figures 12 and 13 may be considered in view of the trends expected
for various mechanism choices.
Mechanism I--
The simplest possible mechanism choice, that favored by Brus and
McDonald^for their experiments at 3225-9A, involves the independent decay
of the S and L species; that is the rate constants k_ , k. , k_, , and kx-,
all are equal to zero. For this condition the fluorescence intensity and
the I_°/I ° values are described by the simple expressions (21) and (22):
b Lr
I = a [S]0e"kSt + p [LJoe'V (21)
IS°/IL° = (a/p) ([S]0/[L]0) (22)
Thus for this condition we expect I °/I ° to be independent of the pressure.
It is seen in Figures 12, 13, and 14 that the data for 3107, 3211, and 3225 j
22
-------
excitation do not follow this expectation. There is a strong increase in the
S 7iL ratl° at the iow Pressures. Only one set of data derived from our
experiments with 3273 A excitation appear to follow relation (22); note the
points marked by triangles in Figure 14. However in this case the results
are least accurate since the extinction coefficient of S02 is very low. and
the necessarily limited pressure range employed (1.8-12 mTorr) could not be
extended to the region in which I °/I ° is most sensitive to the pressure.
Clearly we can conclude from the present results that the L and S states of
S02_excited at 3107, 3211, and 3225 A do not decay independent of one another;
it isopossible that this mechanism is operative at excitation wavelength
3273 A. Thus it is necessary to consider these results in terms of alternative
mechanisms which involve S to L conversions. We will consider two such
mechanisms.
Mechanism II—
The least complex of the alternative L and S interconversion mechanisms
is one chich allows only the S L transformation; that is, k,/ = 0, k, / 0.
0 ~G
This non-equilibrium mechanism leads to the following relations for the
fluorescence intensity and the I °/I ° values:
k \ [S]0}e S + (k t k [S]0 + rL]0)e L (23)
*S " L S ~ L
/P) - kt/(kS - V V([L]o/[S]0 + kt/(kS " kL)}
We may test the compatibility of this mechanism using the data set for 3107 A
excitation (Table 4 ); in this case the strong absorption by S02 allowed
study of the fluorescence to very low pressures (0.05 mTorr) corresponding
to nearly collision-free conditions. The 3107 A data can be fitted well
employing the following experimentally determined rate constant expressions:
kfsec-1) = 2.51 x 104 + 4.19 x 104p (25)
b
kjsec-1) = 3.H x 103 + h.86 x 103P + 5-0 x 103[1 - e-°-060/p] (26)
Jj
where the pressure p is expressed in mTorr units. Equation (2k} contains
three unknown rate parameters: a/p, [L]o/[S]o, and k^. Since the decay of
both the S and L states was determined in the same experiment using a fixed
geometry of the cell and the detector, a/6 should be a constant which
depends only on the lifetimes of the emitting species, the detector wave-
length response, and the relative emission envelopes of the S and L states.
From the effect of filters on the apparent IS°AL° values, Brus and McDonald
have concluded that the emission envelopes for the two species are not
greatly different, and in this case, o/p = TL%g = 8-°8 (3107 A) is a good
approximation. However it is seen in Table 3 that the calculated IS°/IL°
23
-------
7 *^C — ' —
r
—
1 1
0
*— »
1 1 1 1 1
4
...
I
8
PS02 • mTorr
Figure 13. Plot of the ratio IS°/IL° versus pressure:
excitation wavelength 3107 A. The curves shown axe
the theoretical fits to the data assuming the S <-L
interconversion calculated from the equilibrium model
[eq. (31)]. The parameters chosen to obtain the
various curves are outlined in the text.
Ql I I I I I I I I I I I
Figure 14. Plot of the ratio I °/I ° versus pressure.
b Jj
Excitation wavelengths, open circles - 3211 A; closed
circles - 3225 A; triangles - 3275 A; the solid curves
are the theoretical fits to the 3211 - and 3225 A data
calculated from eq_. (24) for curves 2 and 4, and eq_.
(31) for curves 1 and 3; values of the parameters
used for curves 1, 2, 3, and 4, respectively, are
a/p = 8.26, 8.26, 3.2, 4.3: [L]Q/[s]p 0.50, 1.67,
0.03, 1.0; k = k, = 8.64 x 10s, 3.40 x 10s,
u.u.2, j-.u} Jv,, = r>.\^ — u.u'-h js. -LU ,
2.91 x 104, 2^56 x lo4, 2.56 x 104
sec
-i
mTorr
-------
TABLE k. THE EFFECT OF PRESSURE ON THE I °/I ° VALUES
AND RECIPROCAL LIFETIMES OF THE L AND S COMPONENTS OF
S02 FLUORESCENCE EXCITED AT 310?A
Pressure ,
mTorr
8.19
8.19
6.12
6.12
3.99
3.99
2.93
2.93
2.93
2.93
1.71
1.71
1.14
0.93
0.91
0.91
0.77
0.55
0.55
0.40
0.40
0.38
0.29
0.20
0.20
0.18
0.17
0.16
0.14
0.11
T, usec/
div.a
5
10
5
10
10
20
10
20
20
50
10
20
20
20
20
50
20
20
50
20
50
50
50
50
50
50
50
50
50
50
VT , sec" ,
X10"3
42.9
43.0
•32.7
32.8
22.6
22.6
17.1
17.3
17.4
17.5
11.7
11.2
8.54
7.88
8.00
8.29
7.01
6.15
6.19
6.30
5.41
5.50
5.33
6.08
5.60
4.94
5.41
5.34
5.70
5.46
1AS, sec"1,
x 10~4
33.6
—— «
28.5
........
19.8
____
16.5
(10.3)b
(10.9)
12.1
6.93
7.13
5.54
(4.73)
5.36
4.73
(3.94)
4.73
(3.00)
3.37
3.21
3.48
3.39
2.80
2.86
2.86
3.02
2.60
v/v
1.34
— »-.
1.58
-_.»
2.14
— _
2.00
(0.92)b
(0.88)
2.24
2.25
2.52
2.00
(1-72)
2.23
2.78
(2.08)
2.95
(2.84)
3.10
3.76
2.73
3.24
4.18
3.04
3.18
2.88
3.42
aSweep time of the oscilloscope; these data in parentheses were determined
using longer sweep times for which the S component data were highly compressed
on the time scale; they are not considered as accurate at the small T runs; the
runs at large T served as a check on the L component analysis at small T values.
-------
values may differ by a significant factor depending on the wavelength region
monitored. The emission of the S state appears to be shifted somewhat
toward, the longer wavelengths. In consideration of these results it appears
to us that a realistic choice of a/p for 3107 A excitation is in the range:
7-°/Tc° = 8.08 > a/P > 4.7. The choice of magnitude and kinetic form of
k in thTs~case is somewhat arbitrary. Theoretically29 one expects that non-
0
radiative unimolecular processes such as (5^), (5c), (6b), and (6c) will be
unimportant in a small molecule such as S02. In studies of S02 fluorescence
excited within the second allowed band, Hui and Rice30 have found that the
quantum yield of radiative decay of the S02(1'B2} state is near unity at ex-
citation energies less than the dissociation limit. The magnitude of the
fluorescence quantum yields at zero pressure of S02 excited within the first
allowed band have been the subject of some disagreement. lf 2 The failure in
the earlier work to recognize two emitting species, the very great efficiency
of the collisional quenching of the S species, and the very long lifetime of
the L species, together with the use of cells too small to avoid diffusional
loss of the L state with these unrecognized properties, all contributed to
this problem. However most workers and we now agree that the limiting quantum
yield of fluorescence at zero pressure is probably near unity, in accord with
the original suggestions of Mettee.31 Therefore it is reasonable to assume
that only second order reactions such as (3a) should be important in con-
verting S to L states, and we may take k equal to some fraction (F) of the
observed bimolecular cfienching rate constant for the S state; for the ex-
periments at 3107 A, k (sec"1) = k_ p = F x 4.19 x 104p (mTorr).
Using reasonable choices for the various parameters in equation (24)
does provide a good fit of the experimental I °/I ° variation with pressure.
O -U
In Figure 12 the effect of various choices of F on the fit can be seen.
Taking [L]O/[S]O = 2.3, and a/p = 8.08, F was varied; 0, 0.25, 0.5, 1.0 in
curves 1, 2, 3, and 5 respectively. Obviously the F = 1 curve best describes
the experimental results. The sensitivity of the theoretical curve shape
to the choice of the unknown variable [L]0/[S]O, taking f = 1, is also
shown in Figure 12; for the choices of a/p = 8.08, the value of [L]O/[S]O
was varied: 2.0, 2.3, and 2.6 in curves 4, 5s and 6, respectively. The
best match of the data is had for [L]O/[S]O =2.3. As we pointed out
previously, a/p may be somewhat less than T °/T °, and we may investigate the
D Ju
effect of this choice on the results. Taking the apparent lower extreme for
this ratio, a/p = 4-7, F = 1, and [L]O/[S]O = 1.0, gives a somewhat better
fit vcurve 7 of Figure 12) at higher pressures. It is evident that this
mechanism involving the efficient, second order S -» L conversion is con-
sistent with our data at 3107 A. For other excitation wavelengths where
fairly complete Ic /IT° versus pressure data were determined in this work
b J_i
(3215 and 3225 A), the mechanism II also fits well; see Figure 14-
The mechanism II appears to be unattractive in one sense; the physical
circumstances which would favor the conversion of S >- L, without the
reverse reaction occurring significantly, are hard to imagine. It is not
likely that the density of states available to the L species is much greater
26
-------
than that for the S species in the region of the potential energy surfaces
to which excitation occurs. In fact if we associate the L species with
S02( BjJ and the S species with SQ^1^), one might expect the density of
the states available to the S species to be somewhat greater since this
molecule is formed in a highly vibrationally excited state where anharmonicity
would create a close packing of the states. If there is no energy difference
between the states, then microscopic reversibility requires that L -*- S
conversion occur as well as S -^ L. In this light a test of the compatibility
of the more complete, "equilibrium" mechanism with our data is in order.
Mechanism III —
In this equilibrium mechanism we may make the reasonable assumption that
&t = k^ . It is important to recognize that under these conditions the ex-
perimentally determined decay constants for the L and S states no longer are
equal to the real decay constants; i.e. , ^ k and £ ± k_. From equations
(13) -(16) we obtained the ^following relationships:
5 + I = kg + kL (27)
- k)2 + Uk2]* (28)
Let t = £ - I , (29)
JL
Uien 4> = (i]*2 - te2)2 = k - k (30)
Now equation (20) reduces to (31):
- *)([S]0/2[L]0) - kt) + (-kt([S]0/[L]0)
- *)([S]o/2[L]0) + kJ + (kt([S]0/[L]0)
Again three parameters must be specified to test the fit of this equation to
the data: q/p, [S]O/[L]9, and k.. For various "reasonable" choices of these
parameters the fit to the 310? A data is shown in Figure 13. For curves 1,
2, and 3 we have assumed [L]O/[S]O = 1.6? and O/p = 8.08, but we have varied
the second order rate constant kt (mTorr'1 sec'1) : 3 x 103, k x 103, 5 x 103,
respectively. Note that the choice of kt = 5-0 x 103 mlorr-1 sec-1 gives an
excellent fit of the data.
It is instructive to observe the expected influence on the IS°/IL°
-pressure relationship of a significant first order component to the inter-
system crossing constant kt; i.e., k^ = kgb ^ 0. Taking k^ = k^ = 3.1 x
103 sec'1 and no second order component (k^ = k^a = 0) gives curve 5 in
Figure 13. With the same value for k^ and ^ but taking
5.0 x 103 mTorr"1 sec"1, curve U results. It is evident that both curves have
the incorrect form of I°/IT° variation with pressure and that first order
D J-J
2?
-------
interconversion of S and L at any significant rate appears to be an inap-
propriate choice..
Our only other extensive data sets for excitation at 3225 and 3211 A
are also described well by this equilibrium mechanism III. Obviously one
cannot differentiate between mechanisms II and III on the basis of the fit
to the I °/I ° -pressure data alone. However out results do exclude the
mechanism I in every case where reliable, low pressure results are available,
and they are incompatible with the significant occurrence of first-order
intersystem crossing between the L and S systems.
It is interesting to note that the [S]o/[Ljo values which we estimate
using either of the two S ^ L interactive mechanisms are relatively higher
for experiments with excitation at 3225 A than at other wavelengths which we
employed. Brus and McDonald observed a jump in I °/IL° to 19 in this wave-
length region (3226 A), which seems consistent with our present result.
However in their work the I °/I ° values appeared to be independent of the
b ij
pressure, so it is not clear that the two results apply to the same phenomenon;
the somewhat smaller excitation bandwidth in the Brus and McDonald ex-
periments may have led to a rather different set of states. In any case, our
observations for 3225 A excitation and Brus and McDonald1s results at 3226 A
both are consistent with the work of Hamada and Merer11 who reported that the
absorption band of S02 at 3226 A is less perturbed than its neighbors in the
spectrum.
Brus and McDonald's data for 31^0.5, 2981, 2961.6, and 2889.3 A did not
extend to pressures below 0.^ mlorr where the most sensitive test of the
mechanism can be made, but their data also can be fitted well to both the
mechanisms II and III outlined here, using reasonable choices for the
parameters; see Figure 15-
Mechanism IV--
One might question whether the present data could support equally well
the hypothesis that only one excited state of S02 is involved in the emission.
Presumably a variation in the radiative and nonradiative decay constants may
occur with change in vibrational excitation within the single state, in a
somewhat more dramatic fashion than we believe occurs in the L species.
Then a "step-ladder" collision dynamics scheme such as that outlined by
Freed and Heller34 may be used to rationalize the non-exponential behavior.
This case is approximated in terms of the present mechanism taking (3 = 0
and- CLJ0 - ° in equation (20):
IS°/IL° = (a, + kg - kL)/(u> - ks H- kL) (32)
If k and k ' involve only the bimolecular quenching terms then I..,o/ITo->~oo
"CO b Jj
as the pressure approaches zero. Thus at the very low pressures the col-
lisionally induced internal conversion becomes negligible and single ex-
ponential decay should be observed. However even at the lowest pressure at
which we could carry out the experiments successfully, 0.05 niTorr, the
fluorescence of S02 follows double exponential decay; see Figure 16. Further-
more the decay constant of the L component does depend on the excitation
28
-------
Figure 15. Plot of the Brus and McDonald
data for the ratio I °/I ° as a function of
D J_j
pressure of S02; excitation wavelengths,
hexagons -- 2961.6 A| squares — 2981 A;
triangles -- 2889.3 A; circles -- 31^0.5 A;
the solid curves are the theoretical fits
to the data calculated from eq^ (2k); the
dashed curves are the theoretical fits
calculated from eq. (31)j values for the
parameters used for the excitation wave-
lengths, 31^0.5, 2981, 2961.8, and 2889.3 A,
respectively, are: q/p k.O, 2.0, 2.7, 2.0
for mechanism II; 8.0, 15, 11+, 15 for
mechanism III; [L]O/[S]O = 1.15, 0.60, 1.0,
2.0 for mechanism II; 2.11, 5-71, 5-88, 12.5
for mechanism III; k- = 3.67 x 104, 5.90 x
104, 6.00 x 104, 5.90 x 104 sec"1 mTorr'1 for
mechanism II, and k = k, = 1+.27 x 103,
k.2k x 103, 3.1+8 x 103, 2.67 x 103 sec"1 mTorr'1
for mechanism III.
-------
-re.
0 250 500
Time, usec
Figure 16. Semilog plot of the fluorescence
intensity versus time for S02 under essentially
collision free conditions; excitation wave-
length 3107 A; P0~ =0.05 mTorr.
energy (see Table 1), in contrast to the expectations of the one state,
vibrational relaxation model. Thus the single state model seems inappro-
priate to explain the present double exponential decay data. However it
appeal's to us that the fluorescence behavior of the L state of S02 would be
an excellent example to test quantitatively the Freed and Heller model, in
that the properties of the L state appear to follow, at least qualitatively,
all of the predictions from their theory.
Ihe Nature of the S and L Species-- The tentative assignment of Brus
and McDonald of the S species to the SO^(1A2) state and L species to the
S02(1B1) state seems to be in accord with most of the experimental data
available today, including our own. Strong evidence for this is provided by
the observed enhancement of the [S]O/[L]O state population ratio on ex-
citation within the relatively unperturbed spectral region near 3226 A,
attributed to the 1A2 -4- X 1A1 transition. We would like to point out that
the tentative identification of the structure in the fluorescence emission
excited within the 3150-2930 A region 14sl5 as originating from the -"-Bj. -*•
X 1A1 transition, when coupled with the present results, also supports this
assignment indirectly. It may appear incongruous that absorption by S02 in
a structured region characteristic of a S02(1A2) -4- S02(X 1A1) transition
should yield fluorescence which on analysis appears to support the 1B1 state
as the emitter. However from the kinetic parameters derived in this work,
it can be seen that in the steady state experiments of Shaw, _et al.14s15
carried out a 2 and 1 mTorr pressure, the dominant intensity of the
fluorescence could originate from the L species for most excitation wavelengths
30
-------
they employed. From the kinetic mechanisms II and III presented, we would
expect the equations (33) and (3^-) to describe the ratio of the intensities
for the steady state condition (I /
Mechanism II: (IS/IL)SS = («/P) ([o]o/[ LJ0)kL/tks + k3a([S]0/[L]o)p) (33)
(Q/P)(([S]o/[L]0)(k,p + kj + k, p}
Mechanism III: (l) - - (3>0
Using equations (33) and (3^) and the appropriate kinetic data which we
obtained from the fitting of the I °/I °-pressure data and the Stern- Voljaer
quenching constants, we simulated the conditions employed by Shaw, et al.
We estimated the ratio (I /I ) and the fraction of the total fluorescence
b J_j S S
from the long-lived state [I /(lc + IT)] for the two seemingly realistic
L b Jj
mechanism choices II and III; see Table 5- One evident feature of these
data is the rather striking difference between the I /(I + I ) values
Ll O Ll
expected for the two mechanism models. The choice of mechanism II, invoking
only S ->- L conversion, leads to fractions in the range 0.9^- to 0.72 for
wavelengths in the 2889 to 3211 A range, while there is a near equal
fluorescence from the L and S components expected if the mechanism III, in-
volving S <2 L interconversion, is operative. Thus the observations of
Shaw, _et al., who observed only 1B1 ->- 3T ^-A.^ structure, coupled with the
present work, appear to favor indirectly the choice of mechanism II as well
as add further support for the L and S state assignments as •'•BI and ^Ag,
respectively. These considerations indicate that additional fluorescence
steady-state studies should be made with excitation near wavelength 3225 A,
where it appears that -"-Ap emission should be more intense than that of the
1B1 component if the interpretation outlined here is correct.
One important experimental result remains to cloud the assignment of
the L and S states. Both states can be readily populated with excitation
wavelengths as long as £273 A, while the tentative spectroscopic assignment
of the S02(1B1) -*• S02(X ^) transition is near 3150 A.12>14>15 If one is
to accept the L and S states as S02(^Bi) and S02(XA2), respectively, then
the origin of the transition 1B1 -4- X ^AX must be at some wavelength greater
than 3273 A, since the direct population of the L state at this wavelength
is too large to reflect absorption from vibrationally rich ground state
molecules.
The Quenching of the S and L States by Some Atmospheric Gases--
The quenching rate constants for the first excited singlet states of
S02 with various common atmospheric gases and pollutant molecules are of
interest to many atmospheric scientists and photochemists. These data for
excitation at 313QA are of special value since many photochemical experiments
have been carried out at this wavelength. We have determined rate constants
for both the L and S states of S02 excited at 3130A for quenching by some
important atmospheric gases through direct lifetime studies in gaseous S02
31
-------
TABLE 5. THE ESTIMATED FRACTION OF THE TOTAL FLUORESCENCE INTENSITY DERIVED
FROM THE L SPECIES IN S02 STEADY STATE IRRADIATIONS
Excitation Q
wavelength, A
2889.3
2961.6
2981
3020
3065
3107
3140.5
3211
3225
^he values
2 raTorr and
(24) of the
chosen for
data for I-
o
________ _TT
llu
b
Mechanism IT
0.94 (0.90)
0.88 (0.85)
0.88 (0.85)
0.92 (0.91)
0.85 (0.86)
0.74 (0.76)
0.80 (0.81)
0.72 (0.72)
0.44 (0.42)
/(I -t-i)] a---- Source of kinetic
ss data employed
Mechanism III
0.59 (0.60)
0.52 (0.53)
0.44 (0.45)
0.51 (0.58)
0.46 (0.49)
0.46 (0.43)
0.51 (0.47)
0.51 (0.45)
0.26 (0.23)
shown were calculated for steady state irradiations
1 mTorr pressure (in parentheses); ^estimated using
text; cestimated using equation (31) of the text;
a/S, [S]0/[L]0, and k
°/I_ ° versus pressure
Lt
Ref. 16
Ref. 16
Ref. 16
This work
This work
This work
Ref. 16
This work
This work
of SO at
equation
the parameters
were those providing the best fit of the
data.
-------
-quencher molecule mixtures. Hie quenching rate constants for the L state
•were also determined using 2662 A excitation. In this case the emission from
the S state was too weak to allow estimates of rate constants. All of these
data are summarized in Table 6 • Several features of these results should
be noted. The quenching rate constants for the S state are significantly
larger than the gas collision number for each of the quencher molecules.
Only a very minor perturbation is effective in quenching the short-lived
component, and this presumably leads to the L state. Thus it is highly
unlikely that the S state is involved in any chemical reactions.
The data of the last column of Table 6 show the ratio of the quenching
constant for the long-lived species to that of the short-lived species,
kj/k . Observe that this ratio is approximately constant for the great
variety of reactant molecules studies. For the potentially chemically
reactive gases such as 02, isobutane, and trans-2-butene, the ratio may be
somewhat larger than in the case of the chemically unreactive species; if
this difference is real, some chemical contribution to the observed total
quenching of the L species may occur in these cases. Another interesting
feature of these data is the regular enhancement of the quenching rate
constant for the L state with excitation at the longer wavelengths; ^•h.(ofifio}/
kki'qiqol ^s aPProxijna"t;ely constant, independent of the nature of the
quenching gas. This effect has been inferred from less direct measurements
of k./k/r by Mettee35 and Horowitz and Calvert36 from steady state fluorescence
studies. The effectiveness of the L state quenchers appears to be related to
the polarizability of the molecules as suggested in other studies of excited
S02 singlet and triplet quenching.35"40 Rresumably the excited S02 molecule
acts as an electrophilic reactant on collision with polarizable molecule to
form a highly polarized charge-transfer complex which may enhance intersystem
crossing or internal conversion of the excited S02 species. The constancy
of the k, , S/-2\Aii(OIOQI ra/tio may merely reflect the relative density of
states (X3^; Bls A2, B2) in the two regions of the L state potential
energy surface which correspond to these excitations. The strength of the
S02-quencher molecule encounter involving molecules of different polari-
zability may govern the extent of this enhancement of crossover at each
region of the potential energy surface.
The present work gives no new insight into the mysterious, non-radiative,
reactive singlets of S02 invoked by Heicklen, _et al. If as seems likely
the S state is indeed the S02(1A2) and the L state is the S02(1B1), then
these states are excluded as condidates for the S02* species since the
quenching properties do not correlate with those of S02*. It appears to us
that the only other potentially important excited singlet reactant which
remains a viable candidate for the Heicklen S02* ^species is a highly
vibrationally excited ground state molecule, S02(X ^-Ai). It is probable
that these species are formed in a large fraction of the quenching collisions,
and they could be highly reactive toward some reactants employed by re-
searchers. It is not clear that they could survive the usually rapid
vibrational relaxation processes to be important reactants with CO, C2H2,
C2F4, etc., in mixtures at the relatively high H2 and C02 concentrations which
were employed.
33
-------
IAHLE 6. RATE CONSTANTS FOR THE QUENCHING BY VARIOUS ATMOSPHERIC GASES OF
THE L(k L) AND S(k S) COMPONENTS OF THE S02 FLUORESCENCE EXCITED AT 3130
0. o q
AND 2662A
Quenching
gas
Nitrogen
Oxygen
Carbon
dioxide
Carbon .
monoxide
Argon
Isobutane
trans -2-
Butene
Sulfur
dioxide
Nitric oxide
Methane
Benzene
Biacetyl
Quenching rate constants, £.mole sec ,x 10
k L(2662) k L(3130) k S(3130)
q q q
0.12
0.12
0.27
0.17
0.11
0.45
0.69
0.38
0.28
0.23
0.80
0.81
± 0.01 0.35 ± 0.04 19 ± 18
± 0.01 0.34 ± 0.07 16 ± 14
± 0.02 0.91 ± 0.13 59 ± 8
± 0.01 0.46 ± 0.14 30 ± 21
± 0.01 0.27 ± 0.03 21 ± 11
± 0.02 1.41 ± 0.46 70 ± 56
± 0.04b 1.68 ± 0.98 51 ± 60
± 0.01 1.04 ± 0.02 76 ± 14
t 0.02
± 0.02
± 0.04
± 0.05
11
k L(2662) k L(3130)
q q
k L(3130) k S(3130)
q q
0.34 0.018
0.35 0.021
0.30 0.015
0.37 0.015
0.41 0.013
0.32 0.020
0.41 0.033
0.37 0.014
Temperature, 25±2°C; in 2662 A experiments, the cis-2-butene isomer was used.
-------
REFERENCES AMD NOTES
1. Present address, Chemistry Department, University of Alberta, Edmonton,
Alberta.
2. Present address, Chemistry Department, University College, Dublin,
Ireland.
3. I.E. Hillier and V.R. Saunders, Mol. Phys., 22, 193 (1971).
4. K.J. Chung, "Experimental and Theoretical Study of Sulfur Dioxide", Ph.D.
Thesis, The Ohio State University, Columbus, Ohio, 1974.
5. D.D. Lindley, "Ad Initio SCF Studies on the Ground and Excited States of
Sulfur Dioxide", Masters Thesis, The Ohio State University, Columbus,
Ohio, 1976.
6. G. Herzberg, "Electronic Spectra and Electronic Structure of Polyatomic
Molecules", D. Van Nostrand, Princeton, N.J., 1967? p. 605.
7. J.H. Clements, Phys. Rev., 1+7, 224 (1935).
8. N. Metropolis, Phys. Rev., 60, 295 (191*-!).
9. J.C.D. Brand and R. Nanes, J. Mol. Spectrosc.;, 46, 194 (1973).
10. R.N. Dixon and M. Halle, Chem. Phys. Lett., 22, 450 (1973).
11. Y. Hamada and A.J. Merer, Can. J. Phys., 52, 1443 (197*0 .
12. Y. Hamada and A.J. Merer, Can. J. Phys., 53, 2555 (1975).
13. J.C.D. Brand, J.L. Hardwick, D.R. Humphrey, Y. Hamada, and A.J. Merer,
Can. J. Phys., 54, 186 (1976).
14. R.J. Shaw, J.E. Kent, and M.F. O'Dwyer, Chem. Phys., 8, 155 (1976).
15. R.J. Shaw, J.E. Kent, and M.F. O'Dwyer, Chem. Phys., 8, 165 (1976).
16. L.E. Brus and J.R. McDonald, J. Chem. Phys., 6l, 97 (1974).
17
S. Butler and J.R. McDonald, Paper ¥C9, Thirty-First Symposium on
Molecular Spectroscopy, The Ohio State University, 1976.
18. E. Cehelnik, C.W. Spicer, and J. Heicklen, J. Amer. Chem. Soc., 93 »
5371 (1971).
19. L. Stockburger, III, S. Braslavsky, and J. Heicklen, J. Photochem.,
2, 15 (1973/74).
20. E. Cehelnik, J. Heicklen, S. Braslavsky, L. Stockburger, III, and E,
Mathias, J. Photochem., 2, 31 (1973/74).
35
-------
21. A.M. .tatta, E. Mathias, J. Heieklen, L. Stockburger, III, and S.
Eraslavsky, J. Rootochem., 2, 119 (1973/74).
22. J. Heicklen, "Atmospheric Chemistry", Academic Press, 1976, p.
23. P.B. Sackett and J.T. Yardley, J. Chem. Phys., 57, 152 (1972).
2k. H.W. Sidebottom, K. Otsuka, A. Horowitz, J.G. Calvert, B.R. Rabe, and
Z.K. Damon, Chem. Phys. Lett., 13, 337 (1972).
25. K. Otsuka and J.G. Calvert, J. Amer. Chem. Soc., 93, 258! (1971).
26. S.J. Strickler, J.P. Vikesland, and H.D. Bier, J. Chem. Phys., 60,
664 (1974).
27. J.P. Briggs, R.B. Caton, and M.J. Smith, Can. J. Chem., 53, 2133 (1975).
28. S.W. Benson, "The Foundations of Chemical Kinetics", McGraw-Hill Co.,
I960, pp. 128-130.
29. M. Bixon and J. Jortner, J. Chem. Phys., 50, 3284 (1969).
30. M.H. Hui and S.A. Rice, Chem. Phys. Lett., 17, 474 (1972).
31. H.D. Mettee, J. Chem. Phys., 49, 1784 (1968).
32. T.N. Rao, S.S. Collier, and J.G. Calvert, J. Amer. Chem. Soc., 91, 1609
(1969).
33. J.G. Calvert, Chem. Fhys. Lett., 20, 484 (1973).
34. K.F. Freed and D.F. Heller, J. Chem. Phys., 6l, 3942 (1974).
35. H.D. Mettee, J. Phys. Chem., 73, 1071 (1969).
36. A. Horowitz and J.G, Calvert, Int. J. Chem. Kinet., 4, 191 (1972).
37. H.W. Sidebottom, C.C. Badcock, J.G. Calvert, B.R. Rabe, and E.K. Damon,
J. Amer. Chem. Soc., 93, 3121 (1971).
38. F.B. Wampler, K. Otsuka, J.G. Calvert, and E.K. Damon, Int. J. Chem.
Kinet., 5, 669 (1973). ~
39. F.B. Wampler, J. Environ. Sci. Health, All (6), 397 (1976).
40. F.B. Wampler, Int. J. Chem. Kinet., 8, 687 (1976).
36
-------
THE MECHANISM OF PHOTOCHEMICAL REACTIONS OF S02 WITH ISOBUTAME EXCITED AT
3130A.
Introduction
The primary photophysical and photochemical processes in S02 excited into
the first allowed singlet band (2500-3^00 A) and the forbidden band (3^00-
ij-000 A) have been studied for about a decade. Since the OS-0 bond dis-
sociation energy is about 135 kcal/mole, corresponding to a wavelength
smaller than 2180 A,1 any photochemically induced reactions of S02 in the
excitation wavelength region 2500-^000 A are the results of interactions with
bound excited states of S02 molecules. Three emitting states of S02, one
short-lived singlet and one long-lived singlet, presumably the XA2 and -"-BI
states,2j3 and one triplet S02(3B1),4 have been observed in emission upon
excitation into the region 2500-4000 A.
Dainton and Ivin5'6 studied the photolysis of S02 in the presence of
several paraffinic and olefinic hydrocarbons. The principal product was the
sulfinic acid. A slight decrease in reaction rate with increasing temperature
was found between 15 and 100°C. Timmons7 then reexamined the photochemical
reactions of S0a with various alkanes. The negative temperature coefficient
reported by Dainton and Ivin was confirmed, but the product analysis showed
a number of important products. Since no evidence for the formation of
alkyl radicals was found, he supported the insertion mechanism for the S02 +
RH reactions, as suggested by Dainton and Ivin. He also argued that the
great variety of products arise from thermal and/or .photochemical decomposi-
tion of the sulfinic acid. Badcock, et al.,8 presented arguments questioning
this conclusion and suggested a primary reaction by excited S02 of H-atom
abstraction in irradiated S02-RH mixtures. A quantitative analysis of
products of the photochemical reactions of S02 with n-butane and isobutane
was performed by Penzhorn, et al.9»10 They concluded that the reaction of
excited S02 molecules with alkane involves H-atom abstraction rather than
insertion, and more than one triplet state of S02 molecules was thought to
be involved in the reaction mechanism.
The photochemical reaction mechanism of S02-added gas mixtures is
known to be complex. The results of these studies carried at low pressures
(P < 10 Torr)11'12 can be understood by invoking reactions of one singlet
and one triplet state, presumably the ^B^ and 3B! states. However, both the
Heicklen and the Calvert research groups have concluded that the low pressure
mechanism requires an expansion of some sort for the high pressure conditions
(P > 10 Torr).13"16 In addition to the previous states, one more unknown X
state was suggested by Calvert ,and co-workers,13'16 and three more states,
designated as S0£, SO^* and SOg, were proposed by Heicklen and co-workers.
14,15,17-25 A.J- ^g time of mos^ Of .y^ previous studies of the photochemistry
of S02 at high pressures, the pressure saturation effect of S02(3B;]_)
quenching 26~29 was unknown, the quenching rate constants determined at low
pressures were applied incorrectly to the high pressure experiments. It is
thus necessary to reexamine all the old kinetic reaction mechanisms of SOS
photochemical reactions.
37
-------
Penzhorn, e_t al.,9'10 have shown that a good mass balance (> 97%)
between the rate of press-ore drop and the rate of formation of the principal
products is obtained in the photolysis of S02-isobutane mixtures. We thus
carried out our studies on the S02-isobutane photoreaction by following the
pressure drop during short irradiation periods (generally less than 3
minutes) at excitation wavelength 3130 A. A new insight into the photo-
chemical reaction mechanism of the S02-alkane system is given here.
Experimental
Light from a high pressure point source mercury lamp (HBO 500 W/2 L2)
was filtered through the following solutions:30 NiS04 (0.178 M, 5 can.), K2Cr04
(5.0 x 10-4 M, 5 cm), potassium biphthalate (0.02^ M, 1 cm) and a Corning
glass filter 7-5^ (9863), 0.3 cm. A nearly uniform parallel beam through
the cell was obtained utilizing a condensing lens and light stops. The
incident light was monitored continuously during photolysis by reflecting a
small fraction of the incident light from a quartz plate onto an RCA 935
phototube (S-5 cathode). Radiation exiting the reaction cell was attenuated
with a uniform density filter and monitored with another RCA 935 phototube.
The light intensity measured by both phototubes showed the stability of lamp
to be about ± 5$> during the experimental period. The reaction chamber was a
5 cm long, ^.5 cm diameter quartz cell. The light intensity was calibrated
by potassium ferrioxalate actinometry.31 The linearity of the response of
the phototube with light intensity was also checked with actinometry. At
S02 pressures of less than 50 Torr, the absorption followed the Beer-Lambert
law with the decadic extinction coefficient, 1.59 x 10~3 Torr""1 cm"1 (0 < P
< 15 Torr). This checks reasonably-well with the literature values of 1.60
x 10"3,32 1.57 x 10"3,13 and 1.53 x 10"3 Torr"1 cm"1,33 for experiments at the
same excitation wavelength.
A grease-free vacuum line was used in all the photolytic experiments.
A mercury diffusion pump, separated from the main vacuum line by a liquid
nitrogen trap, kept the system at a pressure of 5 x 10~5 ~ 1.0 x 10~6 Torr.
Matheson Co. isobutane and J.T. Baker Chem. Co. S02 (anhydrous) were
used. Both gases were distilled from trap to trap at -196°C, retaining
only the middle third, which was degassed many times. Before each run, the
gases were degassed once again. Spectroquality benzene (Matheson Coleman
and Bell) was degassed and distilled under vacuum, and the middle third was
retained for use. These three gases were checked by mass spectrometry
(Du Pont Model 21-%1 B), and no impurity was observed. C02 (Matheson Co.)
was also purified by vacuum distillation.
The photoreaction of S02 with isobutane was followed by means of the
pressure drop in the reaction cell. The pressure was measured directly
by three different Baratron capacitance manometers (MKS 310AHS-1, 310AHS-100,
and 310BHS-1000), which were fused onto the reaction cell as closely as
possible. In the medium pressure case (1.2 < P < 120 Torr), the pressure
drop was quite linear with time during the irradiation period; i.e., the
least squares fit of the pressure drop vs. time was almost exactly equal to
the total pressure drop divided by the total irradiation time; see Figure 17.
However, even at very short reaction times (l ~ 3 minutes), scattered light
38
-------
P, Torn
59.55-
59.50
59.45
59.40
light off
/
"-••„..•.
-2
I I I I I
O 2 4
Time, min
Figure 18. Example of the pressure variation in the
3130 A photolysis of S0a (15.0 Torr) with isobutane
(hk.6 Torr).
in the cell was appreciable; see Figure 18. Since the pressure drop was
quite linear with time, it was assumed that this scatter did not influence
the absorption of light by S02, and the Beer-Lambert law was used to
calculate the intensity of light absorbed. Light scatter was not an ap-
preciable problem at low pressures (P < 1.2 Torr); see Figure 18. However,
the pressure drop was not linear with time; see Figure 19. The reason for
this non-linearity is not clear. To investigate the possibility that it
was caused by interference of the product, the photolysis was carried out
in two steps. After an initial radiation period, the contents of the cell
were condensed in a liquid nitrogen trap and then allowed to warm to room
temperature before radiation was continued. Removal of the solid and
liquid products in this way indicated that the observed decrease in quantum
yields was not due to the interference by these products. Because of the
non-linearity of the data, only data taken during the first minute of ir-
radiation (10 second interval) were utilized in the calculations of quantum
yields. High pressure data (P > 120 Torr) were not reported because of the
nonlinear pressure drop and the high magnitude of light scatter. Qualitatively,
however, they showed a leveling-off of the quantum yield at high con-
centrations of isobutane (P > 500 Torr).
Blank experiments were carried out for each experiment to ensure that
the pressure drop was due to reactions involving the isobutane.
39
-------
Time, sec
36O 240 12O O
360 240 120
Figure 18. Light scattering observed in the photo-
lysis; pressure of S02 and isobutane for curves A,
B, C, D, and E are 0.196, 0.751; 2.50, 17.20; 2.50,
51.70; 1^.99, 32.06; and 15.00, 707.7, respectively.
0.440 -
Figure 19. Example of the pressure variation in
the 3130 A photolysis of S02 (0.195 Torr) with
isobutane (0.251 Torr).
-------
In the medium pressure case, the gases in the cell after irradiation
were analyzed by mass spectrometry. Within the maximum sensitivity ob-
tainable (1-10 ppm), no gaseous products were detected. These results are
contrary to those reported by Penzhorn, et al.1Q
Results and Discussion
Low Pressure Kinetics of Excited S02 Molecules--
Penzhorn, et al,9 have shown that the photolysis of S02 with paraffins
does not involve a chain mechanism under high pressure conditions. This is
consistent with our observation that the quantum yield of product formation
is independent of the light intensity; see Figure 20.
§
0.3 =_
0.2
0.1
1 A
I0, quanta sec, x10~
Figure 20. Plot of quantum yields as a function of
light intensity; a test of the chain reaction
mechanism;. P = 15.0 ± 0.2; P. , , = 38.0 ±
S02 isobutane
0.3 Torr.
We have shown in our previous studies3*34 that two excited singlet
states, presumably the ^-Ag and 1B1 states, can be reached at excitation
wavelength 3130 A. However, the kinetic data3 and the quenching rate constants
of S02(XA2) and S02(1B1) molecules2?3 all indicate only the SOaC1!^)
molecules are chemically important for steady state experiments. Furthermore,
Butler and McDonald35 have shown that the SQ2(1A2) species are not the pre-
cursors of S02(3B1) molecules. Therefore, for simplicity, we will not con-
sider the S02(ZA2) molecules in the following discussion of kinetic scheme
of S02.
The low pressure quantum yields for the S02-isobutane photolyses are
tabulated in Tables 7 and 8. The reaction mechanism can be written as
follows:
S02 + hv
S02(1B1)
- S02(1B1)
S02 -+ S
S02
(I)
(la)
S04(1B1) + S02 -*~ (2S02)
S02(1B1) + isobutane •*- products
la
(2a)
-------
TABLE 7- QUANTUM YIELDS FOR THE PHOTOLYSES OF S0a
AMD ISOBUTANE MIXTURES UKDER LOW PRESSURE CONDITIONS
fP < 1.2 TORR)
[SCfe],
Torr
0.050
0.100
0.200
0.300
[BH],
Torr
0.160
0.258
0.314
0.405
0.462
0.515
0.907
0.069
0.096
0.135
0.231
0.307
0.423
0.522
0.535
0.589
0.591
0.688
0.734
0.822
0.881
0.924
0.071
0.161
0.252
0.297
0.376
0.404
0.498
0.500
0.512
0.526
0.540
0.612
0.683
0.8l4
0.111
0.180
0.273
0.282
0.430
0.502
0.632
[RH]/[S02]
3-2
5.16
6.28
8.10
9-24
10.3
18.14
0.69
0.96
1.35
2.31
3-07
4.23
5-22
5-35
5.89
5-91
6.88
7-34
8.22
8.81
9.24
0.36
0.81
1.26
1.48
1.88
2.02
2.49
2.50
2.56
2.63
2.70
3.06
3.42
4.07
0.37
0.60
0.91
0.94
1.43
1.67
2.11
J
0.251
0.236
0.277
0-313
6-319
0.313
0.306
0.125
0.161
0.185
0.240
0.239
0.257
0.291
0.286
0.320
0.291
0.283
0.309
0.327
0.335
0.326
0.109
0.158
0.201
0.212
0.235
0.201
0.242
0.243
0.245
0.252
0.248
0.245
0.264
0.267
0.102
0.128
0.162
0.187
0.211
0.198
0.231
a
0.265
0.287
0.296
0.302
0.306
0.308
0.319
0.151
0.179
0.207
0.246
0.263
0.279
0.288
0.289
0.292
0.293
0.298
0.300
0.303
0.305
0.306
0.099
0.164
0.201
0.214
0.232
0.237
0.250
0.250
0.252
0.254
0.255
0.263
0.269
0.277
0.102
0.139
0.174
0.177
0.211
0.223
0.240
*calc is calculated from relation (A) ty assuming kphAp = 0.273.
-------
TABLE 8. QUANTUM YIELDS FOR THE PHOTOREACTION OF S02
WITH ISOBUTANE IN THE PRESENCE OF BENZENE UNDER LOW
PRESSURE CONDITIONS
[S02],
Torr
0.200
0.199
0.200
0.200
0.200
0.200
0.200
0.200
0.200
0.200
[HH],
Torr
0.517
0.515
0.515
0.513
0.516
0.512
0.518
0.51T
0.524
0.5li
[C6Hs],
Torr
0.00000
0.00652
0.0118
0.0185
0.0215
0.0297
0.0406
0.0572
0.0721
0.0867
$
0.248
0.144
0.118
0.0901
0.0375
0.0779
0.0744
0.0644
0.0592
0.0553
43
-------
.3
I I
.2
.5 1
5 1O
30
Figure 21. Plot of quant-urn yields as a function
of [KH]/[S02] ratio. Experimental conditions:
Pcn = 0.05 (0), 0.10 (•), 0.20 (A), and 0.30
O(j£
Torr (•); curves A, B, and C are calculated from
relation (A) by assuming k /k = 0.291, 0.255,
and 0.273S respectively.
fei*
0 20 40
OH,., mTorr
Figure 22. plot of the function of the quantum
yield, relation (C), versus the benzene pressure.
Data are from the benzene-inhibited 3130 A photo-
lyses of S02-isobutane mixtures under low-pressure
conditions.
-------
(2b)
(2c)
(3a)
(3b)
SOpX3^) + isobutane •+• SOp^B].) + isobutane
302(1B1) + isobutane -»- (S02) + isobutane
-*• GOp + hvf
S02(1B1) -*- (SQa)
SQ2(3BX) + SOp ->~ S03 + SO '
S02(3B1) + S02 -+- (2S02)
S02(3B1) + isobutane ->- products (5a)
S02(3BX) + isobutane -*~ (S02) + isobutane (5b)
-*- S02 + hv? (6a)
-*• (SQs) (6b)
where (2S02) and (S02) symbolize S02(X ^-AjJ or other long-lived or non-
emitting product molecules, and "products" represents the various sulfinic
or sulfonic (solid or liquid phase) products. '10 In the pressure range
employed here, steps (3a), (3b), (6a), and (6b) are not important.3'4 Using
the steady state assumption, the above mechanism predicts:
2a
l [EH]
+
fk 1DW2J + k I k
lKla [RH] + K2b] K5a
{kl [EH] + k2} {
k4 [EH] '
^k5)
Q rn
+ «
(B)
S T
•where $ and $ represent the quantum yields contributed from the excited
singlet and triplet molecules, respectively; EH represents isobutane.
Eelation (A) predicts the quantum yield § is a function of the [S02J/
[EH] ratio. Figure 21 shows that the experimental data do support this con-
tention. Most of the rate constants in relation (A) have been determined
previous^: ^ = (1.04 ± 0.02) x 101X[3]; k2 = (l.kl ± 0.46) x 10X1[3];
k^ = (If. 21 ± 0.12) x 108[4]; k = (8.7k ± 0.17} x 108 i* mole'1 sec'1 [36].
Since the quantum yields for this reaction are near unity, we then assume
=* k
.
Ihere are still three unknown variables,
« , k , and k .
An estimate of the extent of the singlet excited S0g involvement in
the quantum yields in the photolysis of S0g-isobutane mixtures-- The photolyses
-------
of the S02-isobutane mixtures were inhibited by benzene addition in a series
of runs at fixed pressures of S02 (0.200 ± 0.001) and isobutane (O.|?l6 ±
0.007 Torr). These are summarized in Table 2. The rate constant for S02
(3BX) quenching by benzene is about 100 times that for isobutane.36 Note
in Table 2 the effectiveness of small quantities of benzene in quenching a
large part of quantum yields; for these conditions $ is reduced from 0.248
in the absence of benzene to 0.07^4- for benzene pressures greater than about
0.0^06 Torr. The concentration of the excited singlet fluorescence state of
S02 could not have been altered significantly by the benzene addition (less
than Idfj), so that the benzene-inhibited quantum yields can be used to
extract the contribution from the singlet state involved. The following
step is added to the reaction mechanism:
S02(3B1) + C6H6 ->-nonradiative products (7)
The total mechanism thus predicts
Here $°, $^ and °° refer to the measured quantum yields of products formed
in the S02-isobutane-C6H6 mixtures with P0_ = 0.200, P. = 0.516
o U2 iso DU"cane
"Torr, and with" benzene at 0, p Torr, and at a pressure which gives nearly
complete triplet quenching, respectively. Relation (c) is tested with the
experimental data in Figure .22. The best value of $°° extrapolated from the
data of Table 8 is 0.0^65. From the slope of Figure 22 (158 ± 33 Torr-i)
together with the measured rate constants k, and kj-, we estimate the quenching
rate constant of S02(3BX) for CSH6, k , to be (8.5 ± 1.8) x 1010 I. mole"1
sec"1. This checks reasonably well with the values determined previously in
the direct S02(3BX) lifetime studies: (8.1 ± 0.7) x 101O[37]; (8.8 ± 0.8)
x 1010 I. mole'1 see"1 [36].
From the value of $ = 0.0^65, relation (A), and the previously
reported rate constants for k., and k?, we estimate k? = Q.k x 109 i. mole"1
sec"1. This rate constant has not been extimated previously. Although
Badcock, est ai.,8 assumed that the triplet molecules of S02 are alone
responsible for the photochemistry observed in S02-hydrocarbon mixture photo-
lyses in both the first allowed and the "forbidden" absorption bands of S02,
the present work shows that the contribution from the singlet excited S02
molecules, although small, is not negligible, at least for the S02-isobutane
system.
An estimate of the intersystem crossing ratios for excited S02 "singlet
induced by collisions with S02 and isobutane --Relation (A) can be rearranged
to the more convenient form of (D):
-------
r0.738
[RH]
+ 1) (0.1432
kla ,k2bs [RH1
k2 + ( k2 > [SOp]
(D)
where * = (12.3 [S02]/[RH] + 16.7)
"1
(l)
Relation (D) has been tested in Figure 23, where the term on the left
side of relation (D) was plotted versus the ratio [RH]/[S02]. A very good
linear relationship between the variables is seen. From the slope of Figure
23 we estimate: ^^ ~ °'291 * 0.011. In order to optimize the sensitivity
of the quantum yields to [S02] and improve the accuracy of the intercept,
which determines the intersystem crossing ratio for S02, the data for [RH]/
[S02] > 5 were discarded. A least squares fit of these data gives a inter-
cept of 0.107 ± 0.027. Together with the previously reported rate constants
is extimated to be 0.1^5 ± 0.037. This checks reasonably
k and k,
JL i
i
well with the various literature values: 0.21 ± 0.0k [32]; 0.12 ± 0.01 [38];
0.20 ± 0.05 [38]; 0.1k +.0.02 [39]; O.Ik ± 0.01 [39]. In this case the
slope is somewhat smaller, 0.255 - 0.012, than that determined from total
sets of data, 0.291 ± 0.011. The theoretically calculated quantum yields
with the above reported rate constants were plotted in Figure 21 as curves A,
B, and C with kp /k = 0.291, 0.255, and an averaged value 0.273, respectively.
It is seen that the averaged value of k /k = 0.273 ± 0.018 gives the best
fit to the experimental data. Therefore, in the following discussion we will
assume kp, /k = 0.273 ± 0.018.
No estimates of this ratio have been reported
previously. The apparently higher intersystem crossing ratio for isobutane,
0.273 ± 0.018, is not surprising in view of the comparison with that for
alkenes previously reported at 3130 A: 0.51 ± 0.10 (cis-2-pentene) [37];
0.62 ± 0.12 (trans-2-pentene) [38]; 0.85 ± 0.37 (cis and trans-2-butene)
[32]; and 0.62 ± 0.05 (cis-CsHs) [39] •
&T
m
s*
CM -I
00 3
•».
?2
fr.
^ 0
• *^
**L±
I ! i i I I
10
Figure 23. Plot of the function of the left-
hand term in relation (D) as a function of
the [S02]/[isobutane] ratio.
-------
Kinetics of Excited S02 Molecules at High Pressures--
The quantum yields of the product formation for the photolyses of S02-
isobutane mixtures in the absence or presence of CeHg or C02 under medium
pressure conditions (1 < P < 120 Torr) are reported in Tables 9, 10 and 11.
In these Tables the initial pressure jump of the system after irradiation
(see Figure 1) and the rate of light scattering (see later discussion) are
also reported.
Thermal effect on the initial pressure jump-- Dainton and Ivin6 first
found that on irradiation of any mixture containing S02 a small but very
reproducible initial pressure rise was always observed during the first
seconds of the irradiation. This has been shown in Figure 7; it is apparently
due to the conversion of the absorbed light into heat. Dainton and Ivin also
found a smooth, almost linear, relation between the pressure jump and the
subsequent rate of pressure fall for given pressures of reactions, independent
of the method of variation of the incident light intensity and wavelength.
However, no theoretical argument to explain this effect was given by them.
We have found that this thermal effect is a very interesting and useful
phenomenon; it not only affords us a useful method to calibrate internally
the absorbed light intensity in the system, but also gives us a new way to
determine various physical properties such as thermal conductivities and
concentrations of the reactants in the system. A detailed analysis is given
elsewhere,34 and only a brief description of some of the results will be given
here to illustrate these observations.
It can be shown that for a pure S02 system the initial pressure jump,
P., is related to the pressure of S02, P, and the absorbed light intensity,
V by
P.. = 7PIa = 7PI0(la/I0) (F)
where 7 = d/T0Sb', 9 is a proportionality constant which relates the rate of
heat generation to the rate of absorbed light intensity; 5 is the "heat
transfer coefficient" per unit area;40 To is the temperature of the wall; £>
is the total area of the wall; IQ is the incident light intensity. Relation
(F) is tested by experimental data in Figure 24. An experimental 7 = (2.3 ±
0.3) x 10~19 sec/quanta is obtained. This checks reasonably well with the
theoretically calculated 7 value, 1.7 x 10~19sec/quanta.34
In the presence of added gases it can be shown that
Pj = (71P1 + 72P2 + 73P3) Jo (i ~ 10-°-°0903Pl + 0.0014 Pg/P^ (G)
(P-L < 15 Torr)
where 7^ = Q/(T<=S&±)'} P-^ Pg, and P are the initial partial pressures of
S02, isobutane, and C5Hs or C02, respectively. A good check of relation (G)
with experimental data in Tables 9> 10, and 11 is obtained. Theoretically,
7-/7- - 5./6. - A /A , where A. is the thermal conductivity of gas i.
J- jj ^J -L J _L _L —••
48
-------
TABLE 9 . QUANTUM YIELDS FOR THE PHOTOLYSES OF
S02-ISOBUTAHE MIXTURES UNDER MEDIUM PRESSURE
CONDITIONS (P
x 103
5.03
-
5.10
-
5.40
_
6.45
5.78
7.73
_
6.53
5.78
4.95
6.60
5.48
-
7.42
8.48
-
7.95
8.48
.
7-58
-
8.4
.
8.70
1.58
.
6.60
7.50
2.78
7.73
8.02
8.02
8.70
7.50
7.20
7-35
9.75
7.05
-
6.98
-
AP/At
mTorr/sec
0.400
-
0.433
-
0.479
.
0.507
0.463
-
0.538
-
0.513
0.381
0.429
0.412
0.466
-
0.539
0.498
-
0.526
0.540
-
0.556
-
0.563
-
0.682
0.350
-
0.558
0.569
0.557
0.685
0.667
0.644
0.795
0.637
0.851
0.696
0.835
0.679
-
0.679
-
(Continued)
49
-------
TABLE 9. (Continued)
[S02],
Torr
5.00
5.00
5.00
5.00
4.93
5.00
4.99
9-99
10.00
10.00
10.00
10.00
10.00
10.00
9.99
9.99
10.00
9.99
9.99
10.00
10.00
9-99
10.00
10.00
10.00
10.00
10.00
15.00
15.00
15.00
14.99
15.00
14.99
15.00
14.98
15.01
14.93
14.94
15.00
14.94
[RH],
Torr
79-90
87.55
89.48
97.70
102.27
102.13
109.31
9.86
15-43
21.62
25.07
34.94
39-40
45.19
49.24
55.01
58.57
58.64
64.61
69.86
70.69
73-97
79-67
81.24
89.41
96.35
102.16
14.85
19-91
27-22
32.06
40.15
45.51
53-27
59-46
66.88
76.78
85.25
95.15
100.69
5
0.442
0.462
0.462
0.488
0.499
0.453
0.530
0.191
0.234
0.269
0.287
0.323
0.323
0.346
0.367
0.365
0.381
0.383
0-394
0.397
O'.38l
0.42
0.399
0.399
0.444
0.438
0.428
0.208
0.240
0.258
0.288
0.313
0.322
0.341
0.358
0.363
0.366
0.378
0.405
0.394
va
mTorr
39
37
44
45
46
-
53
19-7
24.4
29-5
33-2
43
45
49-3
52.7
57
61.6
60.6
61.7
64.1
66.5
75
75-8
74.4
80.6
83.6
82.8
4l
-
55
58.6
70
73
81.3
87
88
100
108
113
121
Rscat>
x 103
8.55
9.60
9.52
8.52
11-32
-
8.48
8.7
11-55
13.12
15-15
16.80
16.95
18.45
18.52
19-12
18.38
21.52
22.88
19.28
19.42
19-95
19-72
21.75
19.12
21.30
20.4
11.1
14.2
16.2
19-1
15.6
22.8
23-4
22-3
24.1
24.1
24.4
27.2
26.3
£P/4t,
znlorr/sec
0.878
0.912
0.854
0.789
0.829
.
0.955
0.742
0.909
0.997
1.10
1-25
1.15
1-35
1.34
1.42
1.53
1.54
1.62
1.57
1.48
1.61
1.51
1-54
1.76
1.75
1.55
1.08
1.25
1.34
1.50
1.62
1.71
1.78
1.90
1.89
1-91
1.98
2.16
2.04
a
The nuabers reported here have been normalized to Io
3.07 x 10 quanta/sec.
-------
Experimentally we found that 7-Jy± = 1.99, 1.20, and 1.64 for isobutane,
C6H6, and C02, respectively. Thermal conductivity data41 show that A /A
i' '.
1.11t 1.19, and 1.71 for isobutane, CeRe, and C02, respectively.
Figure 24. Plot of P./(l /I ) as function of P
n a o
"SCt
w ._ — £-
The I0 values for lines A,B,C,D,E, and F are 3.79,
3.16, 2.26, 1.58, 1.13, and 0.69 x 10~16 quanta/
sec, respectively.
Evaluation of the high pressure photoreaction mechanism of S0g— The
deviation of the phosphorescence lifetimes of S02(aB1) molecules from the
simple Stern-Volmer quenching relation in experiments at high pressures
(P - 10-760 Torr) was recently reported by Rudolph and Strickler,26'27 Su,
et al.,28 and Calvert.29 This phenomenon has been described elsewhere and
need not be considered in detail here. We have shown that the results of
the photolyses of S02-isobutane mixtures carried out at low pressures (P < 1.2)
Torr) can be understood by invoking reactions of one singlet and one triplet
state, presumably the 1B1 and 3B! state. However, the quantum yields of
products from S02 photoexcited in isobutane mixtures at high pressures
(P > 10 Torr), are much higher than those anticipated from the low pressure
mechanism; this is shown in Figure 25. The dashed line in Figure 25 is
calculated from relation (A) of the low pressure mechanism with the rate
constants reported previously except kr = 2.6 x 10s + 2.0 x 104/{1 +
0.0133P (Torr)} Torr"1 sec"1 was used to take into account the known
b(J2
saturation effect.28 This phenomenon has been found also in many other
photolysis systems of S02-added gas mixtures.I3"25 m addition to the
previous low pressure steps, one more unknown X state was suggested,by the
Calvert group3-3'16*32, and three more states designated as S02, SQs*, and S02
were proposed by the Heicklen research group.I4>l5»l7~35 In our view the
high pressure mechanism of Heicklen has some serious shortcomings. First,
there is no correspondence between the observed kinetic properties of the
S0| state suggested by Heicklen, _et al., and the Stern-Volmer quenching
parameters determined either by Brus and McDonald2 or by us.3 Second, the
produced fractions, a, p, and 7, for the designated S02(3BX), S0|, and S02
excited molecules, defined by Heicklen and co-workers, are assumed to be
-------
three constants at a given excitation wavelength. The intersystem crossing
ratios from SOa^Bj) to SOg^Bj.) for various foreign gases are known to be
different.41 Thus, it seems unlikely that the CC value will be a constant
in the presence of a mixture of different added gases within a wide pressure
range (10~4-103 Torr) . The P and 7 values determined by Heicklen, _et al. ,
are always different for different systems at the same excitation wavelength.
For example, the 7 value was determined to be 0.019 by Kelly, _et al
21
_ but
0.092 by Partyml 1 1 er and Heicklen.25 Furthermore, Su and Calvert2 have
reexamined the data from the photochemical reactions carried out in the
presence of high pressures of CO42 and C2H2[22]. The postulate of Kelly,
_et al. ,22 that the photolysis of S02-C2H2-added gas mixtures excited within
the S02(3B1) -4- S02(X ^A^} band involves the reactions of three different
excited states of S02, namely the S02(3B1), S02 and SOj* states, appears to
be incorrect. We feel that the Heicklen mechanism is essentially an empirical
equation used to fit the experimental data; in our opinion it is not based
upon a realistic photochemical reaction mechanism. Therefore, we have not
used the Heicklen mechanism in our further discussions on the S02-isobutane
photolysis system.
.1 -
50
P. Torr
IS°-C4H10'
100
Figure 25. plot of quantum yields as a function of
P
isobutane
; P =5.0 Torr. Dashed curve - calculated
DU2
from relation (A); solid curve - calculated from rela-
tion (H).
In applying Calvert's photoreaction mechanism of S02 to the high
pressure conditions, one should also consider the saturation effect on the
phosphorescence lifetime of the S02(3B1) molecules. Now we are ready to
rationalize the data of the high pressure photolyses of S02-isobutane mixtures.
Photolyses of S02-isobutane mixtures at high pressures: In addition to
the reaction mechanism from (l) to (6b) suggested before, the following
steps should be considered:
S02(1B1) + S02 -+- X + S02
-------
S02(1B1) + isobutane -*- X + isobutane (2d)
S02(1B1) + M -*• S02(3BX) + M (8a)
S02(1B1) + M ^X + M (8b)
S02(1B1) + M -*• (S02) + M (8C)
X + S02 -** S02(3BX) + S02 (9a)
X + S02 -+- (2S62) (9b)
X + isobutane -*- S02(3B1) + isobutane (lOa)
X + isobutane -*- (S02) + isobutane (10b)
X + M -»- S02(3B1) + M (lla)
X + M -*- (S02) + M (lib)
X •>• (Spa) (12)
S02(3B1) + M -*- (S02) H- M (13)
Using the steady state approximation, the following relation can be derived:
k2jRH] k
+ kg[M])
(kla[S02] +k.
k2[RH] + kg[M]T
In our present experiments M represents the C02 molecule. All the rate con-
stants concerning the X state are unknown. ObviousJ^r, relation (H) is very
complicated and cannot be tested adequately with the very limited data now-
available. However, in the absence of C02 as quenching gas, a reasonably
good fit to the data in Table 9 was obtained by using a curve-fitting method.
Hae results are shown in Figures 25 (solid curve) and 26 (all curves). The
ratios of rate constants used for these fits were: k /k = 0.053}
k10/k12 = 0.0067, k9a/k9 = 0.005, k10a/k1Q = 0.7, k^/^ = 0.136, k2d/k2 -
0.65 [all from this work], and k^ = 2.6 x 10s + 2.0 x 104/(1 + 0.0133[S02]}
Torr'1 sec"1 [28]. All the other rate constants (or ratios) have been reported
previsouly in our discussion of the low pressure mechanism.
53
-------
TABLE 10. QUANTUM YIELDS FOR THE PHOTOLYSES OF
SOp-ISOBUTANE-C6H6 MIXTURES UNDER MEDIUM PRESSURE
CONDITIONS (P < 120 TORR)
Torr
I)? 0.000
0.000
0.022
0.054
0.101
0.153
0.247
0.359
0.450
0.545
0.551
0.653
0.800
0.997
1.202
l.4oo
1.600
1.800
1.960
2.197
2.387
3.130
3-727
4.492
.^.989
D
II). 0.000
0.000
0.152
0.447
0.719
o.84i
1.127
1.219
1.474
1.536
1.917
2.461
3.001
3.860
4.611
•
0.411
0.398
0.386
0.362
0.340
0.313
0.283
0.251
0.239
0.247
0.226
0.222
0.201
0.181
0.173
0.176
0.145
0.142
0.121
0.127
0.113
0.111
0.090
0.094
0.092
0.383
0.381
0.304
0.234
0.215
0.194
0.178
0.177
0.163
0.168
0.134
0.128
0.109
0.099
0.094
V
nfforr
31
32
28
31
31
27
32
27
30
28
31
29
32
31
31
31
28
28
30
30
30
32
35
33
35
60
62
59
59
61
59
63
63
62
64
60
66
64
62
64
Rscat
8.25
6.68
6.82
6.45
8.02
8.18
4.05
5.10
4.88
5.32
5.62
4.20
3.75
3.60
3.00
3.38
2.92
3.22
2.25
2.78
3.45
2.40
1.80
1.50
1.50
20.1
19.1
17.8
12.9
9-52
9-52
8.55
6.98
6.38
7.88
5.18
5-32
4.28
3^5
3.68
AP/A't,
mTorr/sec
0.851
0.696
0.666
0.700
0.618
0.624
0.451
0.409
0.486
0.499
0.450
0.446
0.408
0.361
0.318
0.375
0.302
0.285
0.250
0.260
0.238
0.234
0.178
0.187
0.190
1.54
1.54
1.24
0.941
0.830
0.764
0.696
0.695
0.613
0.658
0.531
0.506
0.434
0.401
0.372
aThe numbers reported here have been normalized to I = 3.07 x 101S
quanta/sec. °
kExperimental conditions for part I and II are: P_ = 5.02 + 0 13
*'* * "^ *'* * " '
-------
12O
P.
iso-C4H10-
Figure 26. Plot of quantum yields as a function of
isobutane* S02 ~ \u'* •?' v"J» • (&•)>
and 15.0 Torr (• ). Solid curves are calculated
from relation (H).
The photolyses of the S0s-isobutane mixtures were inhibited by
addition in a series of runs at fixed pressures of SOg and isobutane. These
have been summarized in Table 10 . Again it can be seen that the effectiveness
of small quantities of C6Hs in quenching a large part of quantum yields. For
our experimental condition the concentration of the excited singlet fluores-
cence state of SC>2 could not have been altered significantly by the CsKs
addition (less than 10$). The effect of addition of a small amount of C6H6
on the X state is not clear. We will assume that the C6H6-inhibited quantum
yields can be tested simply to extract the contribution from the singlet
state involved.13 Thus by adding the reaction step 7, we predict relation
(C) in the low pressure case can also be applied to the high pressure case.
This is tested in Figure 27. Relation (H) predicts ° and $00 to be O.ij-10
(0.373) and 0.056(0.053), respectively, for the case of Pan = 5-02 ± 0.13
DW2
(10.00 ± 0.01) and P. , , = 58.3 ± 0.6 (58. k ± 0.5) Torr. As we have
v ' isobutane ' ^ '
argued in the low pressure case, the two slopes of Figure 27, 1.86 ± 0.17
and 1.52 ± 0.10 Torr"1, together with the previously reported values of k, ,
S02>
Pisobutane>
= (9'9
. mole"1 sec"1, respectively. Within the error limit, these values are
in reasonable accord with the low pressure value, (8.5 + 1.8) x 1010, and the
literature values derived from lifetime studies: (8.1 ± 0.7) x 101D I. mole'1
sec'1 [ 36] .
55
-------
Figure 27. Plot of the function of quantum yield,
relation (C), versus benzene pressure. Pressure
of S02 and isobutane, respectively, are 5.0, 58.3
(• )j and 10.0, 58.4 Torr ( 0).
The quantum yield data for the photolyses of S02-isobutane-C02 mixtures
have been given in Table 11. Several rate constants were determined
previously: kg = (4.9 ± 0.7) x 10s [3], k^ = 2.6 x 103 + 2.0 x 104/U +
0.0133PS02 + 0.0045PC02} [28], k13 = 6.76 x 103/{1 + 0.0133P + 0.0045?
Torr"1 sec"1 [28], Here, for convenience, P
bU2
and P
^
are in units of Torr.
Three more unknown parameters in relation (H), kn , k , and k ., , need to
oa J.J.a J — Lo
be determined. Again, relation (H) is too complicated to be tested adequately
by the limited data now available. Figure 28. shows that a very good fit
can be obtained by assuming kn /kn = O.Op, ko-,/kn = 0.5, k, ,/k.,p = 0.025
Torr"1, and k.... /k,, = 0.64. Of course, these are not unique solutions.
However, the fit in Figure 28. indicates the relation (H) can be used to
rationalize all the data in this work.
The intersystem crossing ratio for COS) kn /kn = 0.05, determined
from this work checks fairly well with the previously reported values:
0.052 + O.OlU (2662 A);41 0.052 (3130 A).32 'kRhA8 = 0.5 determined in this
work is of comparable magnitude to that determined by Demergian and Calvert,32
0.76 ± 0.11. However, the k^A^ = °'°25 Torr"1 estimated here is a factor
of 5.6 higher than that reported previously [Ref. 32], 0.00^5 Torr"1. Since
the saturation effect of the phosphorescence lifetimes of the S02(3B1)
molecules had not been discovered and was not considered in the previous
data treatment,32 the values derived in the previous work at high pressures
must be incorrect.
-------
TABLE 11. QUANTUM YIELDS FOR THE PHOTOLYSES OF
S02-ISOBUTANE-C02 MIXTURES UNDER MEDIUM PRESSURE
CONDITIONS (P < 120 TORR)
[COa],
Torr
0.00
5.089
7.745
8.0T6
ll.JO
19.18
23-55
26.75
30.94
35-19
37. 94
41.52
46.41
53-57
57-98
58.76
66.46
*
0.269
0.264
0.259
0.251
0.259
0.246
0.242
0.251
0.243
0.235
0.242
0.244
0.232
0.227
0.222
0.226
0.224
V'
Bffiorr
29
35
38
40
42
46
50
57
55
62
62
64
71
-
81
8l
84
Rscat,
x 103
13-72
12.75
13.95
12.08
10.8
13.7
13.0
13-3
11.8
11.2
11.4
10.8
11.7
11.4
12.4
11.0
11.7
4P/At,
mlorr/sec
1.00
1.02
0.993
0.964
0.951
0.943
0.887
0.959
0.904
0.878
0.913
0.927
0.888
0.859
0.839
0.855
0.823
aExperimental conditions are: Pg^ = 10.00 ± 0.01; PisobutaEe = 21-58
* 0.17 Torr.
Jhe numbers 3
quanta/sec.
± 0.17 Torr.
The numbers reported here have been normalized to 1Q =3-07 x 10
i
57
-------
20
40
Torr
60
Figure 28. plot of quantum yields as a function
of the C02 pressure. The curve is calculated
from relation (H).
The nature of the X species --The nature of the ill-defined X species
remains uncertain. Presumably it could be an excited nonemitting singlet
or a triplet state of S02, one of its unstable isomers S-0-0 or cyclic S02,
or a high energy vibronic state of the ground electronic state. The
structure of the phosphorescence spectrum of S02 at high pressures (250 Torr)
was examined recently by Rudolph and Strickler.^7 The shape and position of
the spectrum were identical to those observed for emission from the 3'B1 state
at low pressures. Thus the X species does not emit significantly for A <
6500 A. It cannot be the 1A2 or ^B^ state because the kinetic properties of
this species show no correlation with those observed in our previous studies
of these states.3?34 it cannot be the 3A2 or 3B2 state either, because both
theoretical arguments4'3 and experimental evidence4'44'45 indicate that the
efficient unimolecular radiationless process such as step 12 are impossible.
Therefore, we feel that the X state may be one of the isomers (S-0-0, or
cyclic S02) or a high energy vibronic state of the ground electronic state
of the S02 molecule. Obviously, more studies on this state under high
pressure conditions are needed to further characterize this species.
Light scattering and aerosol formation-- Although no appreciable oily
product was visible in the photolyses carried out here, the solid and
liquid products, presumably the sulfinic and sulfonic acids,9'10 were capable
of being followed by the scattered light observed at the rear phototube in the
optical train. As has been shown previously in Figure 8., the scattering was
not appreciable under the low pressure conditions, but it could be measured
quite reproducibly under the high pressure conditions. The scattering had
about a 5-10 sec induction period and then started to increase for about
40-60 sec. It then reached the maximum value and fluctuated slowly around
the maximum. The rate of increase of the scattering, R , reported
in Tables 9,. 10, and 11 is defined to be the initial r!£e of increase of
intensity of scattered light divided by the initial transmitted light
intensity. The R was plotted versus the rate of pressure fall, AP/At,
in Figure 29. A roughly linear relationship was found. This indicates that
58
-------
the initial increase of scattered light is a result of the formation of the
product molecules. About 40-60 sec after the homogeneous nucleation46 starts
the new particle production ceases. Hie particles already present grow by
condensation mainly (there is some growth by coagulation) and decrease in
number through this process.46 The observed scattered light must depend on
both the particle size and particle number. Ihe condensation and coagulation
thus cause a slow fluctuation of the scattering around the maximum value.
30
20
ra
u
in
o: 10
1 2
, mTorr/ sec
Figure 29 . Plot of the rate of increase of initial
light scattering R j_9 as a function of the rate of
sea o
pressure drop.
Instead of measuring the scattered light, Penzhorn, ejb al.,9 measured
the aerosol mass concentration as a function of time for the photolyses of
S02-n-butane mixtures. A good correlation between the product and aerosol
formation was found. The detailed kinetic studies of the particle formation
processes for the photolyses of S02-C2H2 and S02-allene had been reported by
Luria, et al.47*48 They found that initially as the reaction mixture was
irradiated there was an induction period of about 45 sec, before which
particles with size larger than 25 A in diameter were not produced, although
CO was produced immediately. Mucleation occured suddenly, particles were
produced for a few sec, and then particle production ceased. The induction
period of about ^5 sec in their experiments is about the tijne of reaching
maximum scattering in our experiments. Since large particles (diameter >
25 A) were not formed during the first ^5 sec, the total number of small
particle then increased linearly with the time (because the rate of reaction
is constant). Thus a linear relationship between Rscat an<* AP/At found in
Figure 29 is not unexpected.
59
-------
1. J.G. Calvert and J.W. Pitts, Jr., "Photochemistry", John Wiley and Sons,
Inc., New York, 1966, p. 209.
2. L.E. Brus and J.R. McDonald, J. Chan. Phys., 6l, 97 (1971)-).
3. F. Su, J.W. Bottenheim, H.W. Sidebottom, J.G. Calvert, and E.K. Damon,
Int. J. Chem. Kinet, in press.
k. F. Su, J.W. Bottenheim, D.L. Thorsell, J.G. Calvert, and E.K. Damon,
Chem. Phys. Lett., in press.
5. F.S. Dainton and K.J. Ivin, Trans. Faraday Soc., 46, 374 (1950).
6. F.S. Dainton and K.J. Ivin, Irans. Faraday Soc., 46, 382 (1950).
7. R.B. Timmons, Photochem. Photobiol., 12, 219 (1970).
8. C.C. Badcock, H.W. Sidebottom, J.G. Calvert, G.W. Reinhardt, and E.K.
Damon, J. Amer. Chem. Soc., 93, 3115 (1971).
9. R.D. Penzhorn, W.G. Filby, K. Gunther, L. Stieglitz, Int. J. Chem. Kinet.,
Symposium No. 1, 6ll (1975).
10. R.D. Penzhorn, _et al., Conference at Atmonton, 1975-
11. A. Horowitz and J.G. Calvert, Int. J. Chem. Kinet., 4, 175 (1972).
12. A. Horowitz and J.G. Calvert, Int. J. Chem. Kinet., 4, 191 (1972).
13. F.B. Wampler, A. Horowitz, and J.G. Calvert, J. Amer. Chem. Soc.,
94, 5523 (1972).
14. E. Cehelnik, C.W. Spicer, and J. Heicklen, J. Amer. Chem. Soc., 93>
5371 (1971).
15. E. Cehelnik, J. Heicklen, S. Braslavsky, L. Stockburger III, and E.
Mathias, J. Photochem., 2, 31 (1973/7^).
16. A. Horowitz and J.G. Calvert, Int. J. Chem. Kinet., 5, 2*K3 (1973).
17. S.E. Braslavsky and J. Heicklen, J. Amer. Chem. Soc., 94, kQ6k (1972).
18. L. Stockburger, III, S. Braslavsky, and J. Heicklen, J. Photochem.3
2, 15 (1973/7*0.
19. A.M. Fatta, E. Mathias, J. Heicklen, L. Stockburger, III, and S.
Braslavsky, J. Fhotochem., 2, 119 (1973/7^).
20. M. Luria and J. Heicklen, Can. J. Chem., 52, 3^51 (19711).
60
-------
21. N. Kelly, J.F. Meagher, and J. Heieklen, J. Photochem., 5, 355 (19?6).
22. IT. Kelly, J.F. Meagher, and J. Heieklen, J. Photochem., 6, 157 (1976/77),
23. K. Partymiller, J.F. Meagher, and J. Heieklen. J. Photochem., 6, U05
(1977). ' '
24. M. Kelly and J. Heieklen, J. Photochem., 7, 123 (1977).
25. K. Partymiller and J. Heieklen, J. Photochem., in press.
26. R.N. Rudolph and S.J. Strickler, J. Amer. Chem. Soc., 99, 387! (1977).
27. R.N. Rudolph and S.J. Strickler, paper to be published; we thank the
authors for a preprint of this work.
28. F. Su, F.B. Wampler, J.W. Bottenheim, D.L. Thorsell, J.G. Calvert, and
E.K. Damon, Chem. Phys. Lett., in press.
29. F. Su and J.G. Calvert, Chem. Phys. Lett., in press.
30. J.G. Calvert and J.N. Pitts, Jr., "Photochemistry", John Wiley and Sons,
Inc., New York, 1965, p. 732.
31. J.G. Calvert and J.H. Pitts, Jr., "Photochemistry", John Wiley and Sons,
Inc., New York, 1965, pp. 783-786.
32. K.L. Demerjian and J.G. Calvert, Int. J. Chem. Kinet., 7, ^5 (1975).
33. K.J. Chung, "Experimental and Theoretical Study of Sulfur Dioxide",
Ph.D. Thesis, The Ohio State University, Columbus, Ohio, 197^.
3^. F. Su, "A Study on the Photochemistry of Sulfur Dioxide", Ph.D. Thesis,
The Ohio State University, Columbus, Ohio, 1977.
35• Private communication with Dr. J.R. McDonald.
36. F.B. Wampler, K. Otsuka, J.G. Calvert, and E.K. Damon, Int. J. Chem.
Kinet., 5, 669 (1973).
37. H.W. Sidebottom, C.C. Badcock, J.G. Calvert, B.B. Rabe, and E.K. Damon,
J. Amer. Chem. Soc., 93, 3121 (1971).
38. F.B. Wampler, Int. J. Chem. Kinet., 8, 9^5 (1976).
39. F.B. Wampler, Int. J. Chem. Kinet., 8, 183 (1976).
40. M.F.R. Mulcahy, "Gas Kinetics", John Wiley and Sons, Inc., N.Y., 1973,
pp. 214-219.
61
-------
4l. The thermal conductivities of S02, isobutane, C02, and Cglfe at 25°C are
estimated to be 22.15, 39.2, 37.9, and 26.3 cal cm^sec"1 ( C cm"1)'1, x
106; Handbook of Thermodynamic Tables and Charts, Kuzman Raenjevie,
McGraw-Hill Book Comp., 1976, pp. 303-306.
42. G.E. Jackson and J.G. Calvert, J. Amer. Chem. Soc., 93, 2593 (1971).
^3. M. Bixon and J. Jortner, J. Chem. Phys., 50, 3284 (1969).
kk. K.H. Hui and S.A. Rice, Chem. Fays. Lett., 17, 4?4 (1972).
45. H.D. Mettee, J. Chem. Phys., ^9, 1784 (1968).
46. J. Heicklen and M. Luria, Int. J. Chem. Kinet., Symposium No. 1, 567
(1975).
47. M. Luria, R.G. dePena, and J. Heicklen, J. Phys. Chem., 78, 325 (1974).
48. M. Luria, K.J. Olszyna, R.G. dePena, and J. Heicklen, J. Aerosol Sci.,
5, ^35 (1974).
62
-------
A KINETIC FLASH SPECTROSCOPIC STUDY OF THE CH302-CH302 and CH302-S02 REACTIONS.
Introduction
The a3Jiylperoxy radicals, R02, are a dominant reactive species generated
within the sunlight-irradiated, NOX-RH-polluted troposphere, and as such, they
are of special interest to the atmospheric chemist. Computer simulations of
the chemistry of the lower atmosphere suggest that organic peroxy radicals
are important in the oxidation of NO and possibly other impurity species
such as SOs.1'3 However there are very few direct determinations of the rate
constants for the gas phase reactions of the R02 radicals, and the suggested
role of their significance in the chemistry of the polluted troposphere has
been inferred largely from "reasonable" kinetic speculation.
In this work we report the first results of a kinetic flash spectro-
scopic study of some of the reactions of the simplest alkyl peroxy radical,
CH302. A first estimate of the rate constant for the CHs02-S02 reaction is
presented.
Experimental
The CHs02 radicals were generated by the flash photolysis of azomethane-
oxygen mixtures. The azomethane was prepared by the HgO oxidation of the
symmetrical dimethyl hydrazine (Merck) by the method of Renaud and Leiteh;4
it was purified, degassed, and dried by repeated vacuum distillation at low
temperatures. The S02 reactant gas (Matheson, anhydrous) was also further
purified by vacuum distillation in the vacuum line.
The xenon flash lamp and multiple pass cell employed in these ex-
periments was similar to that developed at the Physics Division of the
National Research Council (Canada) and was constructed from drawings
provided by them. The flash lamp, constructed of Vycor, operated at 20 kg
per flash with a width at half maximum intensity of about 50 M-sec. It
paralleled the quartz reaction vessel (220 cm in length, 6.3 cm i.d.) but
was separated from it by a Pyrex-plate filter (0.5 cm thickness) which
transmitted only light of A > 2900 A. This wavelength range of the exciting
light caused negligable excitation of S02 or CHs02 product of the flashed
azomethane-S02-02 mixture. A white surfaced (Kodak 6080 paint) reflector
surrounded the flash lamp and cell. The cell housed a multiple reflection,
White optical system which was usually adjusted for a path length in the
range of 800-l600 cm in most of this work. The actual effective path length
of the analytical beam was determined at each wavelength used for analysis
by employing acetone vapor as the calibration gas. The ultraviolet and
visible spectrum of the flash products was recorded photographically in a
few experiments using a 0.65 m Hilger spectrograph and a small argon flash
lamp which was fired with delay times varied from 0.6 to 2.0 msec following
the large photochemical flash. In all kinetic experiments a continuous,
450 w point source xenon arc was employed together with a Jarrell Ash mono-
chromator equipped with an RCA 1P-28 photomultiplier-oscilloscope com-
bination. In most experiments a wavelength of 265 nm was selected from the
63
-------
analytical beam to follow CH302 • Absorption due to reactants Me2N2 and S02
and the possible secondary product radical H02 were minimized with this
wavelength choice. The oscilloscope trace was photographed, and a kinetic
analysis of the measured trace was made using computer techniques described
in the next section.
The number of methyl radicals formed per flash, and the CH302 radicals
formed ultimately in the Me2N2-02 mixtures, were extimated from the nitrogen
yield from the flash photolysis of pure azomethane in the cell. The measured
pressure increase following the flash was not a good monitor of the extent
of azomethane decay since a small temperature increase (l-2°C) occurred
slowly following the flash, largely as a result of the dissipation of the
residual heat generated from the flash lamp itself following the discharge.
The amount of nitrogen formed was determined both by volume measurement and
by gas chromatography (Varian 5100TC) using He carrier gas and molecular
sieve columns. These measurements showed that 0.29$ of the azomethane was
decomposed per flash for the usual charging voltage employed.
Hie gaseous mixtures of Me2W2-02, Me2H2-02-N2, and Me2W2-02-S02 was
prepared using a capacitance manometer (Baratron) and other calibrated
pressure gauges (Wallace and Tiernan) . Uniform mixing of the components was
ensured through the operation of a mechanical, glass, circulating pump which
was attached in series with the cell.
The flash photolysis of the Me2N2~02-S02 mixtures with the P A above
bU2
0.5 Torr, resulted in aerosol formation following the flash, and the
resultant light scatter prevented quantitative measurements of the CH302
decay for these conditions. Thus all Me2N2-02-S02 mixture studies were
carried out at S02 pressures below 0.2 Torr.
Results and Discussion
The Ultraviolet Absorption Spectrum of the CH302 Radical -
The methylperoxy radical was generated in the filtered (A > 290 nm),
flash photolysis of azomethane-oxygen mixtures through the reactions (1) and
(2):
+ hv •*» 2CH3 + N2 (l)
CH3 + 02 (+M) •+- CH302 (+M) (2)
For our conditions of high 02 concentration, less than 1% of the CH3 radicals
formed ethane in reaction (3):
2CH3 •+• C2H6 (3)
The temperature of the flash mixture remained essentially constant (23 ± 2°C)
during the flash photolysis period.
-------
The extinction coefficient of CHs02 at 265 nm was determined in an ex-
tensive series of experiments utilizing mixtures of azomethane (1-5 Torr)
in 50 Torr of 02. The absorbance of CH302 was followed as a function of
time (£.t)s and an extrapolation of these data to the initial absorbance AQ
was made using the linear function, A^1 = A^ + (2k/el)t. Here 2k is the
experimentally observed second order rate constant for CHs02 decay, 1 is the
effective path length of the analytical beam at 265 nm, and e is the~~ex-
tinction coefficient for CH302 at 265 nm {e = [loglo (lo/l)]/[CH302]l }. The
AQ measurements derived from a series of 17 different flash experiments are
plotted versus azomethane pressure in Figure 12. The observed linear relation-
ship between the variables is consistent with the limiting form of the Beer-
Lambert absorption law and the small fractions of the incident flash lamp
light absorbed by the azomethane for our conditions. The slope of this plot
(equal in theory to 2fl/760 x 0.0821 x T), the measured fraction of the
azomethane photo-dissociated per flash (f = 0.0029), and the measured effective
path length of the analytical beam (1 = ^33 cm, acetone standard), gave
e(265 nm) = 530 ± 27 I/mole-cm for CHs02. Extinction coefficients at several
wavelengths were derived in a similar fashion from less extensive data and
are summarized in Figure 13 where they may be compared with other recent
estimates of e^ The general broadband structure of the CHs02 absorption
reported previously5'6 is confirmed. The absolute values of e are most
consistent with those reported by Hochanadel, et al. ,6 where the experimental
method was very similar to that employed here. We find A = 235 with e —
870 I/mole-cm. The estimates of Parkes, et al.5 are somewhat higher than
ours at most wavelengths; they were derived by molecular modulation spectro-
scopy in a less direct fashion from photolyses at much lower intensities and
radical concentrations.
The Rate Constant for the CHs02-CH302 Reaction --
Most of our kinetic studies were carried out using the wavelength 265 nm
for the analytical beam to follow CHs02 decay. Plots of the reciprocal of
the measured absorbance versus time were linear over a large portion of the
time period measured. However small positive deviations from linearity were
always observed at long times. The linear portion of the decay curves gave
extimates of the ratio 2k, f/el , from which values of k^', the apparent second
order rate constant for CH302 decay, were derived. These are summarized in
Table 30 for experiments over a
2CH302 -*- Products
range of azomethane, oxygen, and added nitrogen gas pressures. There appears
to be no significant trend in the observed rate constant for the runs with
P =50 Torr and with a ten-fold variation in azomethane pressure or with
02
6^5 Torr of added nitrogen gas. In experiments at the lowest gas pressures,
5 Torr of azomethane and 20 Torr of 02, values of k^, appear to be somewhat
higher, (3.08 ± 0.17) x 10s, than those observed in similar mixtures at
higher gas pressures. The origin of this effect is not clear. It is
possible that it arises from non-thermally equilibrated radicals present in
65
-------
Figure 12. ihe measured initial absorbance of CHs02
at 265 run versus azomethane pressure used, in the
flash photolysis of Me2N2-02 (50 Torr) mixtures;
path length, 633 cm.
66
-------
1.0
o
x:
E
|o5
E
1
200
220 240 26O
Wavelength, nm
280
Figure 13. The gas phase extinction coefficients of
CHs02 as a function of wavelength, _e = loglo(lo/l)/l
[CH302]; data are from this work (closed circles);
Parkes, et al.5 (triangles); and Hochanadel, et al.6
(open circles).
67
-------
TABLE 12. ESTIMATES OF THE APPARENT RATE CONSTANT
FOR THE REACTION k, 2CH302 PRODUCTS (k), DERIVED
IN THE FLASH PHOTOLYSIS OF AZOMETHANE-02-N2-
MIXTUBES'"
,a
p ,
torr
1.0
1.0
1.0
1.0
1.0
2.0
2.0
5.1
5.1
5.1
5.1
5.1
5.1
5.1
5.1
5.1
5.1
5.1
p. = 50 torr
°2
k., I/mole-sec,
x 10"8
2.38 1
2.62
2.56
2.28
2.64J
•2.50 ± 0.16
2-49 \2.45 ± 0.06
2.40 '
2.03-,
2.52
2.47
2.10
2.70
2.88
2.36
2.76
2.65
2.36
>2.53 ± 0.31
3.00 •
P ,
torr
5.0b
5.0b
1_
5.0b
5.0b
5.0°
5.0=
5.0°
5.0C
s.oc
10.0
10.0
10.0
10.0
10.0
10.0
10.0
k , , I/mole-sec,
x 10"8
3.24-j
2-97 I 3.08 ± 0.17
3.21J
2.89 '
2.60-
2.40
2.14
2.24
2.40 ,
2.79 >
2.72
2.96
2.59
2.31
2.29
2.74 '
. 2.36 ± 0.18
' 2.63 ± 0.25
except for runs labeled b; temperature, 23 ± 3°C;
CH 0 absorbanoe followed at 265 nm.
32
P0 = 20 torr
cp 2 = 50 torr
.
, p >* 645 torr.
68
-------
the system at the lowest total pressure. However this seems unlikely to us,
since a filtered flash was employed in this work and "hot" radical effects
arising fron. either azomethane photolysis within its second absorption region
(A < 2100 A) or from absorption by the CHs02 primary product were eliminated
here. There is no significant difference between kj , estimates derived from the
the experiments for which the total pressure of 02 and added nitrogen ranged
from 50 to 695 Torr. The average of the k^, values from these experiments
provides our best extimates of the apparent second order rate constant for
CH302 loss: k^,= (2.5 ± 0.3) x 10s I/mole-sec (23 ± 2°C). This is in good
agreement with the estimates of Hochanadel, et al.,6 k, / = (2.3 ± 0.3) x 10s
I/mole-sec (22 ± 2°C), Parkes, et al.,5a k^' = (1.3-2.6) x 10s, and Parkes,5*
k. t = (2.8 ± 0.7) x 10s I/mole-sec (room temperature). In view of the dif-
ferences in the absolute values of the extinction coefficients employed by
each group of workers, it is somewhat surprising that the agreement between
the results is this good.
Deviation from linearity of the plot of the reciprocal of the CHs02
absorbance versus time was observed in this work at long times; this effect
is also evident in the data of Figure it- by Hochanadel, et al.6 It seems
unlikely to us that the effect is attributable to product interference or
instrumental artifacts. It is conceivable that at long times where [CHs02]
is low, reactions other than the bimolecular decay reaction (k} may occur.
The possibility that the effect results from the reaction, CHs02 + CH2 0 -*-
CHs02H + HCO, is not favored for several reasons. First, the magnitude of
the deviation as seen in successive flashes of a given azomethane-02 mixture,
is the same within the experimental error j one would expect about twice the
deviation with the second flash if the reaction with CH20 were the origin of
the effect. Second, even if one assigns the rate constant for the CHs02-
CH20 reaction at the unreasonably high value of 107 //mole-sec, simulations
show that this reaction could account for only a small part of the total
effect seen. It appears to be more likely to us that the reaction (5) may
lead to the effect:
CH302 + CHsJWCHs -*- Product (A) (5)
If we accept the mechanism of CH302 decay through reactions (U) and (5), then
we expect the CH302 absorbance at time t, A^/ , to be given by relation (6):
1
= In
1
- bfc (6)
where b = [Me2N2]k[-. Estimates of k , were derived from the data using
relation (6). The function 2k, //el was determined from the slope of the
linear portion of the AT1 versus time curve. An initial choice of b was made,
and the term on the left of (6) was calculated for various times. The slope
of the least squares fit of the plot of this term versus time was used to
derive a new estimate of b, and successive iterations of this procedure gave
-------
the final irbest" estimate of b. From such procedures we derived the value
for the rate constant for the apparent first order loss of CH302, k5 =
(1.6 ± 0.6) x 105 1/rnole-sec. Farther evidence for this reaction is presented
in the next section.
Although the present data fit reasonably well the simple reaction scheme
outlined, one should question whether the observed second order rate constant
k. , really applies to the bimolecular CH302 reaction k alone. The question
arises since reactive CH30 radicals are formed a fraction of the time, and
ultimately H02 radicals are generated from these as well. Both radicals
can remove CH302 in theory. It has been suggested by Heicklen and co-workers7
that reaction (4) has three active product channels :
2CH302 -*• 2CH30 + 02
•+• CH20 + CH3OH + 0
02
They concluded that at 25°C about 43, 50, and 7% of the reaction occurs in
(4a), (4b), and (4c), respectively. From studies of the products of the
CD302 reactions they found 22, 60, and 18% of the reaction occurred by the
steps analogous to (4a), (4b), and (il-c), respectively. Parkes5* has
estimated that k, /(k, + k, + k, ) ^0.3^. Thus the reactive CH30 radicals
may be generated in 33 ± 10$ of the reactive collisions between CH30
radicals. It is necessary to investigate to what extent the subsequent CH30
reactions may influence the estimates of k. in our system.
We expect the additional reactions (7-15) to occur following reaction
CH30 -f CHs02-*~ CH20 + CHs02H (7)
CHsO + 02 -+- H02 + CH20 (8)
CHsO + CHsO •*• CHsOH + ClfeO (9)
•+- CH3OOCHs (10)'
CH302 + H02 -*- CHs02H + 02 (H)
CHsO + H02 •*- CHsOH + 02 (12)
H02 + H02 -»- H202 4- 02
CHsO + MeW2 ->- Product (B)
CHs02 + Product A (or B) -*~ Stable Products (15)
70
-------
The largest effect on the k^ that we can expect from the occurrence of this
sequence results for the peculiar circumstance that following GHsO formation
in reaction (^a), an additional CH30 radical is always destroyed in either
reaction 7 or 11. For this condition the measured second order rate constant
k^ for CH30 decay may be greater than the true k. by a factor of 1.33 ±
0.10. We may estimate more accurately the magnitude of the effect expected
using a computer simulation of the more complete reaction sequence. We have
employed the following reasonable rate constant choices (I/mole-sec): k =
(0.20) (k^kg)'* = 3.9 x 10s [7]; k9 = 1.5 x 1010; kQ - 3.7 x 105 [8]; k10 =
k.O x 109 * 5 ku = 1.0 x 108 or k.6 x 10s + ; k.^ = 1.8 x 101Q by analogy
with the rate constant for the reaction Cl + H02-*-HCl + 02 [12]; k,~ =
2.0 x 109 [13, IV]; k^ = 1.1 x 105, consistent with that suggested for the
analogous reactions;15 k = 8.2 x 104 (this work), in our simulations we have
assumed further that each radical product of reactions (5) and (14) reacts
at a non-limiting rate to remove another CH302 radical in (15) with k =
1.0 x 10T. The magnitude of k. was altered in the simulations so that the
apparent second order rate constant for CHs02 removal was equal to that
observed experimentally (k. = 2.5 x 10s I/mole-sec). Using the low and
high extimates for k,, (1.0 x 10s, 4.6 x 108), the simulations give k, =
2.3 x 10s and 2.2 x 10 , respectively. It is not clear which of these
estimates is the more correct. In view of the low quantum yields of alkyl
hydroperoxides seen in the recent R02 study of Alcock and Mile,16 the lower
estimate of k-,, may be more nearly correct, and the apparent second order
rate constant k. for CH302 loss may be only slightly larger than the true
k, . Thus we conclude 2.5 x 10s > k, > 2.3 x 10s I/mole-sec. This conclusion
is also consistent with the findings of Parkes [5b] that the observed values
for k. are independent of the pressure of 02 from 0.1 to 760 Torr in his
experiments at relatively low intensities and 1 atm total pressure. Since
the fraction of CH30 radicals which would react by (8) would change greatly
over this range of 02 pressures, either CH302 loss in (7) and (11) is
* Estimated from the measured rate constant for the reverse of reaction (10)
by Barker, et al.8 and the appropriate equilibrium constant for 2CH30
CHsOOCH3 estimated from AH° data of Batt, et al.,9 and S° estimates of
Benson.10
+These two estimates represent the range of values based upon the observed
R-^ _ TT/R_TT „ ratios in experiments at high [02] in the study of Dever and
CH302H CHpO
Calvert,11 0.10, and Weaver, et al.,7 O.lfl, and a steady state treatment of
the radical concentrations.
71
-------
relatively unimportant, or the decrease in the rate of (7) in experiments at
increasing [02] which must accompany the increased importance of the removal
of CHsO radicals by (8), must be accompanied by a completely compensating
increase in the rate of the H02 reaction (11) with CHs02. It seems more
likely to us that the former expectation is more nearly correct. In any
case the analysis points to the conclusion that our extimate of k^ is an
upper limit to the true k. value.
Estimate of the rate constant for the CH302-S02 reaction —
When S02 is present in the flash photolyzed azome thane -0
additional reaction of CH302 with S02 may occur:
CHs02 + S02 -+- Product
mixture the
(16)
Indeed in the S02 -containing mixtures we observed a suppression of the CH302
signal seen in otherwise identical runs but in the absence of S02. The study
of this system by optical methods is limited to a very small range of added
S02 concentrations since a light- scattering aerosol develops in experiments
at pressures of S02 above about 0.2 Torr. Utilizing the simple reaction
mechanism (l), (2), (4), (5), and (16) we expect the observed CHs02
absorbance (A.) at time t to be given by relation (17):
— "£
In
A,
el
(1 }
\b + c/
\
/
+ (b + c)t
(17)
Here £ = [S02]k,/-,and b = [Me2N2] k'. In a series of runs at constant
[Me2W2], the value of t> + c was obtained, as in deriving b in relation (6),
by an iterative Gaussian least squares treatment. These data are summarized
in Table 13. A plot of b + c versus P should in theory be a straight
— — bU2
line with slope = k, /• and intercept = k,. [Me2N2], Such a plot, given in
Figure 32, shows the expected linear dependence between the variables within
the experimental error. From this we derive for the apparent second order
rate constant for CH302 removal by S02, k,gj= (6.4 ± 1.4) x 10s I/mole-sec,
and by azomethane, k i= (1.8 ± 0.2) x 105 I/mole-sec. The latter value is in
reasonable accord with that derived from the azomethane-oxygen experiments
described in the previous section (^c-i~ (1*6 ± 0.6) x 105 I/mole-sec). If
the reaction (5) involves the addition of CH302 to the N=K bond, then it is
likely that a second CHs02 radical will add by analogy with the observed
tetramethyl hydrazine formation in the case of CHs addition to azomethane.17
Thus we would estimate the true value of k ^ ktr#/2 — (8 ± 3) x 104 I/mole-
sec. There is no other estimate of this rate constant, and the rather in-
direct method which we have employed to derive k here requires that it be
accepted with considerable reservation. The rate constant for the analogous
reaction of CHs addition to azomethane is somewhat smaller than k^ at room
temperature, 1.1 x 10s I/mole-sec.1T Our estimate of k is in line with the
rough extimate for the analogous CHs02 addition reaction with CsH6 for which
k ^ 3 x 104 I/mole-sec.1
72
-------
TABLE 13- RATE DATA FOR THE REACTION, CH302 +
S02 PRODUCTS (16) and CH302 + Me2N2 PRODUCTS
(5), DERIVED FROM THE FLASH PHOTOLYSIS 01 AZO-
METHANE-02-S02 MIXTURES"
a
-1
P^rt r torr k1/..lSO-] + k_,[Me_N_], sec
SOo i-O •& 5 22
0.122 80.6 ± 6.9 %
89.2 ± 4.2 I 86.9 ± 5.5
90.9 ± 7.1 '
0.10 68.5 ± 3.1 •
91.1 ± 7.6
67.2 ± 2.9
108.6 ± 8.4
65.9 ± 4.9 .
0.080 58.0 ± 4.2 %
73.6 t 4.3
73.8 ± 4.4
68.8 ± 5.6 ,
61.3 ± 1.5
54.7 ± 1.2
62.7 ± 2.7
71.7 ± 2.7
57.0 ± 1.9 '
> 80.3 ± 19.0
> 64.6 ± 7.5
0.060 55.4 ± 4.2 \
88.0 ± 2.7 > 74.8 ± 17.1
80.9 ± 8.8 J
0.040 68.2 ± 3.0 s
60.9 ± 2.6 \ 60.6 ± 7.8
52.6 ± 3.8 J
0.020 53.4 ± 5.1
V 48.9 ±6.4
44.3 ±4.4 /
5.0 torr;
= 50 torr; temperature, 23 ± 2°C;
absorbance followed at 265 run.
73
-------
Figure 32. Plot of the rate constant function, k'/-
[S02] 4- k ,[Me2N2] versus P ; data of Table II
for flash experiments using PM =5-0 Torr and
Pn = 50 Torr; temperature, 23 ± 2°C.
-------
The products of the reaction (16) have not been defined, but the
measured, apparent rate constant for this reaction must represent a maximum,
kl6 < (°*^ ± 1-Ij-) x 1C|6 I/mole-sec. The reaction channels of reaction (l6)
may involve either 0-atorn transfer from CH302 (l6a) or the addition of CHc,Oo
to S02 (l6b): -
CHs02 + S02 ->- CH30 + S03 (l6a)
-*- GH302S02 (l6b)
If (l6a) is the major pathway, then the addition of the CH30 product species
to S02 may compete successfully with the reaction (8) and the other normal
fates of this radical. If this occurs then there is a high probability that
a second CH302 radical will be removed by reaction with the CHsOS0202 species
which may form subsequently. If (l6b) is the major route then the initial
CHs02S02 radical, or the CH302S0202 species formed from it, may also react
with an additional CH302 species in a non-rate limiting step. Hence it is
likely that the true rate constant k.,/- is about one-half the measured ap-
parent rate constant k..,- . We suggest that the best present estimate of the
rate constant for the elementary step (16) is: (3.2 ± 0.7) x 10s < k,g <
(6.4 ± 1.4) x 10s I/mole-sec, and the lower estimate is probably most nearly
correct.
This is the first reported estimate of k../-. It is interesting to
compare it with that for the analogous reaction (17) studied by Payne,
et al.:is
H02 -i- S02 ->- HO + S03 (17)
They found k /k r/i = 11.8 ± 1.7 (I/mole-sec )'/2. Taking k.^ = 2.0 x 109
[13,14], k17 ^ 5.2 x 10s I/mole-sec. If we make the reasonable estimate that
the differences in the entropies of activation for these reactions, A S^ -
A S,7 ^ 0.8 eu, then the observed ratio of rate constants, k-^/k^ ^ 0.15,
corresponds to an activation energy difference, E-^ - E-^ ^ 1.4 kcal/mole.
This is a seemingly realistic result in view of the difference between the
enthalpy changes for the two reactions: A H17 - A H^ =7.6 kcal/mole. The
measurable reactivity of the CH302 radical with S02 observed here is not
typical of the other alkyperoxy radicals for which kinetic data are available.
In additional studies in our laboratories, Whitbeck, et al.,19 flash photo-
lyzed 2,2/-azoisobutane-oxygen mixtures with and without S02. In this case
there was no detectable effect of S02 addition on the very slow, second
order (CH3)3C02 radical decay. It was estimated that klg < 4.4 x 102 I/mole-
sec.
(CH3)3C02 + S02 ->- (CH3)3CO + S03
-^ (CH3)3C02S02
-------
Calvert, et al.,3 have concluded from a kinetic evaluation of the published
kinetic data that the acetylperoxy radical is similarly unreactive toward
30P with knr, < 7.8 x 102 I/mole-sec.
1^ - '
CH3COOp +• S02 -*~ CH3C02 + S03 (I9a)
-*-CH3C002S02 (I9b)
Cox and Derwent20 came to a similar conclusion in studies of peroxyacetyl
nitrate and S02 mixtures; they suggested k..q < k.2. x 102 I/mole-sec.' From
computer simulations of the S02 oxidation in the sunlight-irradiated, NOX-RH-
polluted troposphere, Calvert, et al.,3 concluded that the contribution to
S02 oxidation from the branched peroxyalkyl and the acylperoxy radicals was
negligible.
Ihe results of the present work suggests that the reaction (16) of the
CHs02 radical may result in a significant rate of S02 oxidation in the highly
polluted troposphere. Using the present estimate of k,x- Calvert, et al.,3
found that the contributions of S02 oxidation by HO, H02, and CHs02 were
about equal, and these accounted for a large fraction of the h%/hr maximum
rate of S02 oxidation observed in computer simulations. However the rates
of oxidation of S02 within a reasonable clean troposphere were found to
depend largely on the rate of HO-radical attack on S02.3
In view of the present results we suggest that the reaction (16) be in-
cluded in the future quantitative considerations of the chemistry of the
sunlight-irradiated, NGX-RH-polluted troposphere, with k ^ ^ 3.2 x 1.06 I/mole-
sec.
76
-------
REFERENCES
1. K.L. Demerjian, J.A. Kerr, and J.G. Calvert, Adv. Environ. Sci. Technol.,
4, 1 ----
2. J.G. Calvert and R.D. McQuigg, Int. J. Chem. Kinet., Symp. 1, 113 (1975).
3. J.G. Calvert, F. Su, J.W. Bottenheim, and O.P. Strausz, "Mechanism of
the Homogenous Oxidation of Sulfur Dioxide in the Troposphere",
Proceedings Internation Symposium on Sulfur in the Atmosphere, Sept.
17-14, 1977, Dubrovnik, Yugoslavia, Atm. Environ.., 12, 375 (1978).
4. R. Renaud and L.C. Leitch, Can. J. Chem., 32, 545 (1954).
5. (.a) D.A. Parkes, D.M. Paul, C.P. Qciinn, and E.G. Robson, Chem. Phys.
Lett., 23, 425 (1973), (b) D.A. Parkes, Int. J. Chem. Kinet., 9,
W(1977).
6. C.J. Hochanadel, J.A. Ghormley, J.W. Boyle, and P.J. Ogren, J. Phys.
Chem., 81, 3 (1977).
7. J. Weaver, R. Shortridge, J. Meagher, and J. Heicklen, J. Photochem.,
4, 109 (1975).
8. J.R. Barker, S.W. Benson, and D.M. Golden, Int. J. Chem.. Kinet., 9, 31
(1977).
9- L. Batt, R.D. McCulloch, and R.T. Milne, Int. J. Chem. Kinet., Symp.
1, 44l (1975); L. Batt and R.T. Milne, Int. J. Chem. Kinet., 6, 9*4-5
(197*0.
10. S.W. Benson, "Thermochemical Kinetics". 2nd Ed., Wiley, N.Y. 1976.
11. D.F. Dever and J.G. Calvert, J. Amer. Chem. Soc., 84, 1362 (1962).
12. M.T. Leu and W.B. Demore, Chem. Phys. Lett., 4l, 121 (1976).
13. T.T. Paukert and H.S. Johnston, J. Chem. Phys . , 56, 2824 (1972).
14. E.J. Hamilton, Jr., J. Chem. Phys . , 63, 3682 (1975).
15. P. Gray, R. Shaw, and J.C.J. Thynne, Prog. Reaction Kinet., 4, 63 (1967).
16. W.G. Alcock and B. Mile, Comb. Flame, 24 125 (1975).
17. J.A. Kerr and J.G. Calvert, J. Phys. Chem., 69, 1022 (1965).
18. W.A. Payne, L.J. Stief, and D.D. Davis, J. Amer. Chem. Soc., 95? 76l4
(1973).
77
-------
19- M.R. Whitbeck, J.W. Bottenheim, S.Z. Levine, and J.G. Calvert, "A
Kinetic Study of the CH302 and the (CH3)3C02 Radical Reactions by
Kinetic Flash Spectroscopy", Abstracts 12th Informal Conference on
Photochemistry, U.S. Bureau of Standards, Gaithersberg, Md., July,
1976, pp Kl-1 to Kl-5.
20. R.A. Cox and R.G. Derwent, "Laboratory Measurements of Free Radical
Reactions Involving 302"j paper presented at COST 6lA Technical
Synposium, Ispra, Nov. 15-17, 1976.
-------
MECHANISM OF THE HOMOGENEOUS OXIDATION OF SULFUR DIOXIDE IN THE TROPOSPHERE
Introduction
The day has arrived when it is possible to include more than an empirical-
ly assigned, first order rate constant for S02 conversion to "sulfate" in
our atmospheric transport and conversion models for S02. Atmospheric
scientists have shown a new interest in the quantitative evaluation of the
significance of the possible paths which convert S02 to its oxidation
products, S03, H2S04, NH4HS04, (m4)2S04, etc. It is now generally recognized
that the development of scientifically sound and lasting strategy for
"sulfate" aerosol control requires that we understand the nature of the
various atmospheric reactants which oxidize S02, the chemical nature of the
products formed in its reactions, the rates at which these reactions will
occur in a given atmosphere, as well as the physical processes which
transport these gases and aerosols about the atmosphere and those which
remove them from it. Our group is engaged in the study of the kinetics and
mechanisms of the elementary gas phase reactions which can lead to the
homogeneous chemical transformation of S02 within the troposphere. The
understanding of the homogeneous chemistry of S02 appears at first sight to
be one of the simplest of the several important tasks faced by atmospheric
scientists. The evaluation of the mechanisms and rates of the heterogeneous
paths of S02 oxidation within the troposphere, the significance of surface
removal processes, and the transport and diffusion processes are less
amenable to laboratory study. However as we shall see in our discussions,
there are as many unsolved problems as there are well defined areas of
knowledge related to the homogeneous atmospheric chemistry of S02.
In any case the time is right to pause and take note of those aspects of
the problem which appear to be well understood, as well as to reconsider some
of the apparently conflicting observations, the unexplained results, and the
alternative hypotheses which are so common in this area of research. The
state of this science has improved greatly in recent years, and some new and
definitive conclusions can be formulated from the wealth of existing in-
formation. In many cases where confusion remains, at least in our minds,
we will offer our speculation on the possible resolutions. We hope that
our prejudices will be readily distinguished from the experimental facts we
quote and that they will serve to stimulate the reader to an active
participation in the future solution to some of the many problems which remain
unresolved.
We begin this study with an evaluation of the possible atmospheric
reactions of the electronically excited S02 molecules. Next we consider the
various possible reactions of the ground state S02 molecule with the reactive
atmospheric components. From this we derive a set of our "preferred" values
for the rate constants of the various elementary reactions. In the final
section we attempt to evaluate the relative importance of the various homo-
geneous S02 conversion modes in the several major types of gaseous atmospheres
which we encounter within the troposphere: the "clean" troposphere; the NOX-
hydrocarbon-CO-aldehyde-polluted regions of the troposphere; and finally the
gaseous mixtures peculiar to the stack gas plumes.
79
-------
iropospheric Reaction of Electronically Excited Sulfur Dioxide
Ihe Nature of the Reactive States in S02 Photochemistry--
The photochemistry of sulfur dioxide excited within the lower atmosphere
provides in principle several reaction pathways which may lead to the ox-
idation or other transformation of S02. There has been general agreement
in recent years that several other paths of S02 oxidation are probably much
more rapid than those involving the excited states of S02 . However this
consensus was reached on the basis of a very limited body of information,
and a new quantitative look at the reactions of excited S02 is in order.
Some new insight into the rates and mechanisms of these reactions is pos-
sible from the results of recent studies.
Sulfur dioxide absorbs light within the ultraviolet region of the solar
radiation incident within the troposphere. Its two near-ultraviolet ab-
sorption bands can be seen in Figure 33- The dashed curve represents the
wavelength dependence of the relative quanta cm~2s~1 of the solar actinic
flux incident near sea level at the solar zenith angle of ^0°( Peter son, 1976).
The regions of overlap of these curves show that the absorption of solar
energy by S02 in the troposphere occurs within the long wavelength tail of
the relatively strong "allowed" band of S02 (2900-3300 A) as well as within
the much weaker "forbidden" band in the 3^00-^000 A region. Only when S02
is photoexcited at wavelengths less than 2180 A is photodissociation of S02
energetically possible, S02 + hv(A < 2180 A) •>- 0(3P) + SO(3S "). Thus solar
radiation absorbed by S02 within the troposphere leads to non-dissociative,
excited states of S02, and it is the nature of these states and their chemical
reactions which concerns us here.
It is clear that excitation within the "forbidden" band of S02 leads
to the population of the S02(3B1) molecule, reaction I (Brand et al., 1970;
Merer, 1963; Su et al., 1977a) :
S02(X ^3.) + hv(3^00 < A < 4000 A) -*» S02(SB1) (l)
Excitation of S02 within its first allowed absorption band leads to the
generation of two emitting singlet species (Brus and McDonald, 1974; Su et al.,
1977b), a very short-lived state which may be the S02(1A2) species, and a
long-lived state which is probably the S02(1B1) molecule:
S02( X XAX) + hv(2400 < 7\ < 3300 A) •*• SG^Ag) (ll)
-*- S02(1B1) (III)
Collisional perturbation of these singlet states results in an additional
efficient pathway for the formation of the S02(SB1) species:
S02(1A2,1B1) + M ->- S02(3BX) + M (l)
In addition to these optically detectable or emitting states of S02, indirect
chemical and physical evidence has led to the conclusion by some workers
that two other non-emitting triplet states, presumably S02(3A2) and S02(3B2),
are important in the photochemistry of S02. For example, see Cehelnik et al.
-------
400
2400
3200
o
Wavelength, A
4000
Figure 33- Comparison of the extinction coef-
ficients of S02 within the first allowed band
(left), the "forbidden" band (right), and a
typical wavelength distribution of the flux
of solar quanta (relative) at ground level
(dashed curve).
81
-------
(1971) and. Ke-J.ey e_t al. (1976/7?) and the references therein. However
recent observations of Rudolph and Strickler (1977), Su e_t al. (1977c), and
Su and Calvert (1977&) open to serious question this conclusion. It has
been observed that the phosphorescent lifetime of the S02(3B1) molecules at
atmospheric pressure devia.tes markedly from the simple Stern-Volmer behavior
characteristic of the law pressure regime. The lifetime of the S02(3B1)
molecule is about 2.h times longer in one atmosphere of air than extrapolation
of the low pressure quenching rate data would suggest. The phenomenon was
first observed and rationalized by Rudolph and Strickler (1977) in terms of
the saturation of the triplet-quenching at high pressures, an expectation of
the recent theory of Freed (1976). A consideration of the details is not in
order here, but a qualitative picture will suffice. The rate of triplet
molecule quenching to ground state becomes independent of added gas pressure
when the nearly degenerate rotational sublevels of the different singlet
vibronie levels of the ground electronic state have collision-induced
broadening which exceeds their spacings, so that the rovibronic manifold
becomes an effective quasieontinuum. In the case of S02 this saturation
occurs at pressures above about 100 Torr of added nitrogen gas. On the
other hand, the chemical quenching of S02(3B1) molecules follows second
order kinetics as expected from theory, and for these cases, the low pressure
bimolecular rate constants are applicable even at 1 atmosphere pressure.
Su and Calvert (1977&) have reevaluated the previous high pressure studies
of S02-C0 and S02-C2H2 mixture photolyses with S02 excitation within the
"forbidden" band. They have found that a simple mechanism involving only a
single excited state of S02(3B1) species, and the newly discovered saturation-
quenching effect explain well all of the experimental data. Thus the
postulate of KeHy et al. (1976/77) that the photolysis of S02-C2H2-added
gas mixtures within the "forbidden" band involves the reactions of three
different excited states of S02 appears to be incorrect.
When S02 in the atmosphere is excited within the first allowed band, it
is unlikely that the newly formed S02(1B1) and S02(1A2) species will live
to react with impurity molecules, since they are quenched by collisions with
the major atmospheric gases so very efficiently (Su et al. 1977b). The
possible chemical interaction of the excited singlet states with 02 may
occur in principle to generate an 804 excited triplet species or 80s and
0(3P) in the reactions: S02(ZA2,3'B1) + 02(3Z ") ->- S04 (or S03 + 0(3P)).
o
However the near equality of the rates of quenching of each of the singlet
states by the potentially reactive oxygen molecule and the "unreactive"
nitrogen molecule argues against this possibility (Su et al., 1977b). The
hypothetical energy transfer reaction, S02(1A2,1B1) + 02(3S ~) -*- S02(x ^A.^) +
, S
02(1A y1?- ), is spin forbidden and is probably unimportant also (Davidson
O O
et al., 1972/73). However the quenching does form the reactive S02(3B1)
molecule. Ground state molecules and conceivably some other unreactive or
reactive species are formed in reaction 2.
S02(1A2,1B1) + M •*- S02(X 1Ai} + M (or other products) (2)
Important for our considerations here is the intersystem crossing ratio,
k1/(k1 + k2); this is the fraction of the originally excited singlet species
82
-------
which form the reactive S02(3Bi) molecules. This fraction has been measured
for various M species in experiments at relatively low pressures and rather
short wavelengths of S02 excitation at 2650, 2662, and 2875 A (Horowitz and
Calvert, 1972a,b; Wampler et al., 1973a) . In view of the recent saturation
quenching effect observed with S02(3B1), it may be inaccurate to use these
low pressure data for our estimates for S02 in one atmosphere of air. The
density of the vibronic states of the electronically excited singlet and
triplet states of S02 are significantly less than that of the ground state
S02 molecule at energies in the vibronic manifold equivalent to the
excitation of the S02(3B1) or the S02 ( 1B1 , ^ ) states. As a result one would
expect that the rate constants for the reaction 1 would not pressure saturate
at as low a pressure as the rate constant for the reactions 2 and 3:
S02(3B1) + M •*- S02(x iAj.) + M (3)
This may lead to an increase in the fraction of the singlets which are
transformed into triplets when higher pressures of the quenching gases are
employed in the photochemical experiments. Although there is now no direct
experimental evidence of the saturation effect in singlet quenching, the
possibility exists, and one must exercise caution in the use of the low-
pressure values for ki/k^ + ka) to estimate S02(3B1) formation rates in air
at 1 atm. Furthermore recent observations of Su and Calvert (I977b) suggest
that there is an increased efficiency of intersystem crossing for S02
singlet excitation at the longer wavelengths. Thus the extimates of kx/
(kx + ka) from the experiments in the 2650 to 2875 A range are probably in-
appropriate for S02 excited singlets populated at the relatively long wave-
lengths present in the solar spectrum within the troposphere (7\ > 3000 A) .
Both the effects outlined would lead to the more efficient generation of
S02(3B1) on excitation of S02 at the long wavelength region of the first
excited singlet band than would be anticipated from the low pressure, short
wavelength data. In view of this discussion we should accept with great
reservations the postulate of the participation of other reactive states of
S02 to explain the "excess" triplet observed in singlet excited S02-
containing systems at high pressures and long wavelength excitation.
The evidence at hand today appears to support the original contension
of Okuda et al. (1969), that the S02(3B!) molecule is the major photochemically
active species formed in the photoexcitation of S02 even within the first
allowed singlet absorption band at high pressures. Thus if we assume an
average intersystem crossing ratio of 0.10 for M = S02 in Okuda' s experiments,
an effective first order rate constant for S02(3B1) quenching to ground state
of 1.5 x lO6^"1 for 730 Torr of S02 (Su et al., 1977 c) , bimolecular chemical
rate constant for reaction k, k, = 7-0 x lO1^ cm3 molec'^-s"1 (Chung _et al.,
1975) , $ = 0.026 from the singlet excited S02, we derive the observed
so2(3B1) + so2 -*- sos + so(3ir)
value of $nrt = 0.08 reported by Okuda et al. (1969).
SOs --
83
-------
We may update the earlier estimates of the rate of S02(3B1) generation
in the lower atmosphere (Sidebottom et al., 1972; Penzhorn et al., 197^a)
utilizing all of the current data and alternative "reasonable" estimates of
the effective intersystem crossing ratio. These data are shown in Table 14.
The calculations have been made for two cases: (l) ¥e have chosen the ex-
perimental values for the intersystem crossing ratio kl/(k1 + k2) determined
by Wampler et al. (1973) from experiments at low pressures of added
atmospheric gases; i.e., we assume that there are no pressure saturation or
wavelength effects on the intersystem crossing ratio; see section A of Table Ik.
(2) In section B of the Table we aave taken the intersystem crossing ratio
as 0.11 at 1 atm for all of the atmospheric gases; i.e., we have assumed that
pressure saturation effects and the longer solar wavelengths populating the
excited singlet S02 resulted in an increase in the ratio observed at low
pressures. It can be seen from the data of Table 14 that the maximum rate
of excitation of S02(3B1) for a solar zenith angle of 0° (in a clean lower
troposphere at 100% relative humidity) ranges from 4.7$ hr"1 in the first
case to 12.9$ hr"1 in the second case.
The Nature and Rate of the S02(3B1)-02 Quenching Reaction—
Obviously the rate of excitation of S02(3B1) does not alone establish
the maximum rate of S02 photoreaction since quenching by the unreactive gases
such as N2, Ar, C02, and H20 leads to no net chemical change. However there
are several types of reactive species in the atmosphere which are of special
interest as possible reactants with S02(3B1) molecules. The first of these
reactants is oxygen. Table 15 summarizes the estimated fractions of S02(3B1)
which will be quenched by the various atmospheric components; we have used
the newly determined, pressure saturation quenching data in deriving these
estimates. It should be noted that the quenching of S02(3B1) by 02 does not
pressure saturate, as does that by N2, Ar, C02, and other chemically un-
reactive gases. This points to some type of chemical action which ac-
companies this quenching action. The rate constant for the quenching of
S02(3B1) by 02, k = 1.6 x 10~13cm3 molec~1s~:L, is very similar to that ob-
served for N2, k = 1.4 x 10~13cm3 molec"3^-1 ( Sidebottom et al., 1972).
We have argued previously that this might indicate a physical, non-chemical,
quenching of S02( 3B1) by 02 as well as N2. However if this were the case
then, contrary to the fact, the saturation of S02(3B1) quenching by 02 would
be seen at high 02 pressures. From the data of Tablesl4 and!5 we can estimate
that the maximum rate of S02(3B1)-02 chemical interaction in air at 50$
relative humidity (25°C) and solar zenith angle of 40°, will range from 1.8$
hr'1 with mechanism A and 3-4$ hr"1 with mechanism B of Table ll)-. Obviously
these rates are very significant if chemical changes in S02 do result from
the interaction.
We have considered previously several possible chemical changes which
could occur in theory as S02(3BX) is quenched by 02 (Sidebottom et al., 1972).
Ozone formation from excited S02(3B1)-02 interactions is possible through
two reaction routes:
84
-------
TABLE Ik. ESTIMATED RATES OF SOg^Bj.) GENERATION BY SOLAR
RADIATION IN THE LOWER TROPOSPHERE*
Solar zenith Rate by I ' Total rate of SC>2{ B^) generation,
angle,8 Rate by 11,1 % hr"1
rH: 0% 50% 100% 0% 50% 100%
A. Intersystem crossing ratios from the low pressure experiments at 2662 °,
of Wampler et al.(1973); no singlet quenching saturation is assumed.§
0 0.093 0.085 0.078 4.03 4.36 4.72
20 0.099 0.090 0.083 3. 4 3.96 4.28
40 0.125 0.113 0.104 2.59 2.84 3.06
60 0.243 0.220 0.203 1.04 1.15 1.22
80 0.569 0.515 0.475 0.18 0.18 0.18
B. Intersystem crossing ratios for all gases assumed to be 0.11 at 1 atm pressure.
0 0.0273 12.9
20 0.0289 11.7
40 0.0365 8.2
60 0.0710 3.1
80 0.166 0.43
'Percentage of singlet S02 quenching by the various atmospheric components
at 25°C and 1 atm of air were estimated using kN2As02= 0-326-' k02/*S02 =
0.321; k^Ago = 0.275 (Su et al. , 1977b) ; XH Q/ kso2 = 1-2 (Mettee, 1969);
Rates of direct excitation of S02 (3B1) and sOj^Aj,1^) are from Sidebottom
et al. (1972).
+These rates refer to the reaction I and II followed by 1 of the text; the
ratio gives the relative population of S02 (3B]_) through absorption in the
forbidden band to that for the first allowed band.
§The values used are: ^/O^ + k ) = 0.033 (M = N2> ; 0.030 (M = 02> ; 0.025
(M - Ar); 0.09 (M= H2O) ; 0.095 (M = S02) .
-------
IA3LE15. P2P.CSNTAGE OF S02(3~S,-L)-QUENCHING BY VARIOUS ATMOSPHERIC GASES
II: AIR AT 25°C and 1 atm
Component
Percentage of SO2(B1)-quenching
rH: 0% 50% 100%
Nitrogen
Oxygen
Argon
Water
55.5
44.1
0.4
0
45.7
41.7
0.3
12.2
38.
39.
0.
21.
8
9
3
1
S02(3B1) + 02(3£ ") -*- SG4(cyclic); AH = -79 kcal mole-1
o
S04(cyclic) + 02 •*- S03 + 03; AH = 15 kcal mole"1
SOa(3B1) + 02(3Z ~) -*~ S03 + 0(3P); AH = -38 kcal mole"1
o
0 + 02 (+M) •*- 03 (+M)
(5)
(6)
(7)
(8)
Presumably the singlet S04(cyclic) species could be formed in 5 and- stabilized
by collisions. However even when vibrationally equilibrated, this species is
rather unstable with respect to decomposition into S02 and 02 (AH = 5 kcal
mole"1). A vibrationally rich S04 species initially formed in 5 could in
principle react in 6, but the thermally stabilized SO^ cannot react sig-
nificantly to generate 03 and S03 in reaction 6 at atmospheric temperatures,
since this reaction is about 15 kcal mole"1 endothermic. See the following
section for a more in depth consideration of possible roles of the S04 inter-
mediate in atmospheric chemistry. Reaction 7 is energetically possible, but we
would expect it to be slow since it involves an electron spin inversion.
Indeed there is a great deal of experimental evidence which suggests the
relative inefficiency of these reactions in pure air. Thus the photooxida-
tion of dilute mixtures of S02 in oxygen leads to very small quantum yields
of S02 oxidation to S03 or H2S04. In those previous S02 photooxidation
studies where a relatively large fraction of the gaseous mixture was S02
(Hall, 1953; Sethi, 1971; Allen et al., 1972a,b), it has been shown by
Chung et al. (1975) that the quantum yield of S03 can be accounted for in
large part through reactions k and 9:
so
SO + S03 -*- 2S0
so3 + so(3z-;
(M
(9)
Relative large quantum yields of S03 formation in S02-rich mixtures are ob-
served at short irradiation times (Skotnicki, 1975) or in flow systems in
86
-------
which S03 is removed quickly and the reverse of reaction 9 is unimportant
(Chung et al., 1975). In steady state photoHyses in static systems, the
rapid build up of the relatively low levels of S03 and SO is followed largely
by reformation of S02 in 9. Much of the SOS (or H2S04) which was formed in
previous studies of the solar ultraviolet-irradiated S02-02 or S02-clean
air mixtures probably was derived from the reaction 4 and bears no relation
to the occurrence of the theoretically possible photooxidation steps 5 and 6
or 7 and 8. Several studies show that there is a small, but not insignificant
rate of 80s (H2SC>4) formation in very dilute S02-air mixtures; thus for
typical solar intensities at ground level in clean air, the following S02
oxidation rates (at low [S02]) were estimated experimentally (% hr"1): 0.65
(Cox and Penkett, 1970); 0.02-O.OU (Cox, 1972, 1973); < 0.1 (Junge, 1972);
0.7 (Kasahara and Takahashi, 1976); 0.000 (Friend et al., 1973).
Altwlcker (1976) reported recently that the generation of small amounts
of Os occurs in dilute S02-containing purified air samples irradiated in
sunlight; for example, 1-3 pphm more ozone was developed in a pure air
mixture with S02(3 ppm) after hO min irradiation than was formed in the ir-
radiation of a similar S02-free sample of purified air (0.6-1.0 pphm). His
observations are interesting and in qualitative accord with the occurrence
of the reaction sequences 5 and 6 and/or 7 and 8 at the rate of about 0.7$
hr-1 or less. However further more detailed studies of the ozone formation
in S02-air mixtures must be carried out before meaningful conclusions con-
cerning the mechanism of 03 generation in Altwicker's experiment can be es-
tablished. However with the exception of Friend's observation, all of the
present studies show a small, but not negligible, rate of the direct photo-
oxidation of S02 occurs in pure air. Since it is so difficult to obtain
pure air samples which have no small impurities of NOX, Os, EH, etc., one
should not give great weight to the photooxidation experiments as proof of
reactions 5 and 6 and/or 7 and 8 occurring. We conclude that the present data
are consistent with an upper limit for the rate constants, k_ +• k = 3 x
10~15cm3molec"1s~1.
At most a few percent of the S02(3B!) quenching collisions with 02 lead
to S03. Thus some other "chemical" result must account for most of the
quenching. Some further clues about this interaction can be had in con-
sideration of the nature of the oxygen product formed in this process and
the seemingly similar organic triplet molecule quenching by oxygen. It ap-
pears to us that the major chemical result of S02(3B1)-02 quenching en-
counters is likely energy transfer to generate excited singlet-02 species:
S02(3BX) + 02( S -) ->- S02(X lAx) + 02(1Z +) (10a)
B °
+• S02(X ^x) + 02(1Ag) (10b)
The studies of the Abrahamson group (Davidson and Abrahamson, 1972; Davidson
et al., 1972/73) have shown convincingly that reaction lOa occurs. The
fraction of the reaction which leads to the alternative singlet states of
02 is not now known. However the experimental data are not inconsistent
with k + k s 1.6 x 10"13 cuAiolec-^s-1, near the total quenching- rate
lOa lOb
87
-------
constant for S02(3B1) reacting with 02. Support for this view is added from
a comparison of these data with those obtained from organic triplet molecule
quenching by 02; in this case the analogous energy transfer reactions occur
efficiently. In Figure 3^ is shown a plot of the logarithm of the rate
constants for the quenching of a series of organic triplet molecule by oxygen
versus the energy difference between the first excited triplet and ground
state of the various molecules (Em - Ec ) . Data shown as open squares are
J-i >=>o
from studies in benzene solvent; open circles are from hexane solutions
(Patterson et al., 1970; Gijzeman et al., 1973a) . Since these reactions at .
high Em - E0 values are not diffusion controlled, and non-polar solvents
•Li bo
were employed, a comparison with gas phase rate constants should be meaning-
ful. It is seen from the positions of the closed circles that our gas phase
rate constant data for 02-quenching of triplets of S02 and (CH3CO)2 follow-
well the trend observed with the other triplet molecules (Horowitz and
Calvert, 1972; Sidebottom et al., 1972). Thus it seems most reasonable that
the same triplet energy transfer mechanism invoked for organic triplet M(T;j.)
quenching by 02
M(TX) + 02(3Z-) ->- M(S0) + 02(lS+ , ^-A) (11)
is descriptive of that for S02(3B1) as well (Kawaoka et al., 1967; Kearns,
1971; Gijzeman and Kaufman, 1973; Gijzeman,
In summary the present data are consistent with singlet oxygen formation
as the major result of S02(3B1) quenching by 02 (k + k = 1.6 x 10~13
cm3 molec~1s~1) , and the oxidation of S02 to SOs results at most a small
fraction of the quenching collisions (k + k = 3 x 1015 cm3 molec~1s"1).
The Nature and Rates of the S02(3B1)-Chemical Quenching Reactions with the
Common Atmospheric Contaminants--
The extensive rate data related to S02(3BX) quenching reactions points
to some potentially significant chemical reactants with S02(3BX) among the
common air pollutants. Thus the very large rate constant for the quenching
of the triplets with WO (1.2 x 10~locm molec~1s~1; Sidebottom, 1972) suggests
come important chemistry at first sight. Although the possible reaction,
S02(3BX) + NO S0(%~) + N02 is exothermic by 15 kcal mole~x, there is no
evidence that this chemical change or any other of significance occurs. N
Nitric oxide has been found to quench efficiently the triplet excited states
of many molecules. The lowest excited electronic state of NO, a 4jt state,
lies 355000 cm'1 above the ground state, so energy transfer from S02(3B1)
25,766 cm"1 above S02(X ^A^}, is not a possible quenching mechanism here.
However theoretical studies show that intersystem crossing of an excited
triplet molecule (3M*) to its ground state (^-M) may be catalyzed by its
interaction with nitric oxide (Gijzeman et al., 1973):
N0(2 H) + 3M* -*- '2(M-NO) + NO(2H) + % (12)
Gijzeman et al. (I973b) have observed a correlation between the magnitude
of the rate constant for triplet molecule quenching by NO and the triplet-
88
-------
11
10
o
in
o
E
9
\
1
,°
1
10
E-r-l
15 20
, , cnrfj1 x10~3
25
Figure 3^-- Relation between the rate constants for
the triplet- quenching reaction with oxygen and the
energy separation between the first excited triplet
and the ground state of a series of molecules:
aromatic triplets measured in hexane solution
(squares), benzene solution (open circles) from
Patterson et al. (1970) and Gijzeman et al. (I973a);
(CH3CO)2 and S02 data from gas phase measurements
(Horowitz and Calvert, 1972; Sidebottom et al.,
1972) .
ground state energy difference of the molecules; see Figure 35- The open
circles are data from hexane solution studies of a variety of organic triplet
molecules. They have suggested a theory which can rationalize this cor-
relation in terms of reaction 12. Observe the magnitude of the quenching rate
constant increases dramatically with Em - E0 value for energy separations
greater than 1.5 x 104 cm
"1
This trend is just the opposite to that observed
in Figure 2 and the 02-quenching of triplets where the mechanism is electronic
energy transfer to 02. Our data from the gas phase studies of the quenching
of SQa and biacetyl triplet molecules by NO are plotted in Figure 35 as closed
circles. It is seen that these molecules follow the same trend observed
for the organic triplet molecules. Thus it is probable that the net effect of
of S02(3B1) quenching by NO can be represented simply by reaction 13; no
interesting net chemistry is expected here.
S02(3B1) -f
S02(x
NO(2H.)
(13)
-------
9.5
'« 9.0
tn
o
E
«=-<
T—
^8.5
dT
o
8.0
I
- oN^o0 °
: *> .
\ 0
O O M
D \ OO
D
\
\
: (cH3C0^(9)o* \
i \
\
scv>
-------
TABLE 16. THE SOa^Bj.) "chemical" QUENCHING BATE CONSTANTS FOR VARIOUS ATMO-
SPHERIC COMPONENTS AND IMPURITY SPECIES IN THE OVERALL REACTIONS SHOWN
VD
Reactant
Reaction
3 , -1-1
k, cm molec s
Isobutane (17) SO2 ( Bj^) + iso-C4H10 •* HOSO +
1.4 x 10
-12
Reference
•> Oo T
Oxygen
Sulfur
Nitric
Carbon
dioxide
oxide
(5,6)
(7)
(10)
(4)
(13)
monoxide (14)
Acetylene
cis~-2-Butene
(15)
(16)
SO
2
+ °2
+ S02
+ N0(
+ CO
+ C2H
->• so4
- S°3 +
-»• so2 +
-so3
2n) - 2
-* SOI3!
2 * (?)
+ cis-C4Hg -+
+ S°3 + °3 f
0(3P) J
°2( &g ' V
+• so(3s:")
(SO2-NO) •* SO2 + NO(2H)
~) + co2
-* CO + other products
3(S02-C4H8) * trans-C4H8
<3
•v-1.
7.
1.
4.
2.
h-
X
6
0
3
3
7
2
io-15
x 10~13
x IO"14
x ID'10
X 10~15
x 10~12
x 10'10
See the text.
Sidebottom et al.(197:
Chung et al. (1975)
Sidebottom et al. (197;
Jackson and Calvert (IS
Su and Calvert (1977a)
Kelley et al. (1976) , £
and Calvert (1977a)
Sidebottom et al . (1971
Demerjian and Calvert(1974)
Wampler et al.(1972),
Badcock et al.(1971),
Su and Calvert(1977b)
-------
its formation appears to correlate best with the extent of S02 molecule
quenching of S02(3B1); it is thought to be S03 (or ultimately H2S04) which
causes the scattering observed following reaction k in the system. So
although the alkenes quench S02(3B2_) with very high efficiency, the net
trivial chemistry which results from this process is apparently of no
importance to the removal or the chemical transformation of S02 in the atmo-
sphere.
The alkanes quench S02(3B1) reasonably effectively, and the magnitude of
the quenching rate constant increases with decreasing strength of the C-H
bonds and the increasing number of C-H bonds in the hydrocarbon (Badcock
et al., 1971; Wampler et al., 1973b). The quenching act has been related to
an H-atom abstraction by the S02(3B1) molecule (Badcock et al., 1971?
Penzhorn et al., 1975). Thus in the case of isobutane-S02 mixture photolysis,
the measured chemical rate of quenching of S02(3E1} to form sulfinic acids
and other complex products is very nearly equal to the rate of S02(3B1)-
isobutane quenching as measured by lifetime studies (Su and Calvert, 1977b).
SQa(3Bi) + iso-C4Hl0 •*• HOSO + C4H9 (17)
The enthalphy change which accompanies reaction 17 is about -22 kcal mole"1,
not very different from that for the analogous HO-radical reaction with
isobutane (-26.6 kcal mole"1). If the abstraction of an H-atom occurs by
S02(3B1)-alkane interaction in the atmosphere, the ultimate result of the
quenching likely will not be the formation of sulfinic acids or other products
characteristic of the laboratory studies of S02-RH mixtures. Thus the HOSO
and R radicals formed in the quenching reaction 18 will probably react with
02, the dominant reactive species in the air. The HOSO radical may combine
with oxygen and ultimately form S03 or H2S04 by a reaction sequence such as
the following:
S02(3BX) + RH -*r HOSO + R (18)
(a) f (b)
HOSO + 02 , HOSCO >- HO + S03
(c) ^ j (d)
^ 9 *
HO^-0
? (+RH)
H2S04
The overall enthalpy change for steps a and b_ above is about -25 kcal mole"1;
AH = -65? AH , = 40 kcal mole-1. Thus path d will not occur. It seems less
Q CL —
likely to us that the HOSO will react with 02 in the less exothermic step,
HOSO + 02 -*- S02 + H02; AH = -6 kcal mole-1. The radical R formed in 18
will give R02, RO, aldehydes, etc., in subsequent reactions in the atmosphere.
In view of these considerations, net photooxidation of S02 could result from
the quenching of S02(3B1) by alkanes.
The predicted rates of the various "chemical" quenching reactions of
S02(3B1) by impurity molecules in the atmosphere have been estimated for
typical solar intensities encountered near ground level for a solar zenith
92
-------
angle of about W>. See Table 17. The hypothetical atmosphere contains 1 ppm
of the different impurity molecules. Various amounts of S02 are chosen
which range from 500 ppm, common to fresh stack gas emissions, to 0.05 Ppm
which is more representative of the levels of S02 present in ambient polluted
air. The rates (ppm hr"1) of all of the reactions shown are very slow.
Even under the most favorable conditions which one might choose to enhance
the importance of these reactions, the rates are far below those which we
extimate for the reactions of HO, H02, etc., with these impurities and S02
in the lower troposphere.
From our considerations we may conclude that the major chemical effect
of S02 photooxidation by sunlight within the polluted atmosphere is the
generation of O^S ) and 02(1A ) through reactions lOa and lOb; see Table 17.
O &
Although the expected rates are relatively large, the rates of this reaction
are about I/20th of those expected from excited N02 reactions with 02 when
H02 and S02 are at comparable levels (Jones and Bayes, 1971; Frankiewicz
and Berry, 1972; Demerjian et al, 197^). The available data do support as
well the occurrence of a slow, but not insignificant rate of S02 photo-
oxidation, presumably through reactions 5 and 6 or 7 and 8. The maximum
rates for these reactions correspond to less than O.O^fo hr"1 (See Table 17).
The observed rates of S02 oxidation in air are much higher than this. Ob-
viously the oxidation of S02 by reactions other than those involving photo-
excited S02 molecules must be important within the troposphere. We shall
consider these in the remaining sections of this study.
Reactions of Ground State S02 with Reactive Molecules and Transient Species
in the Troposphere
Many reactive species present in the lower .atmosphere are potentially
important reactants for S02. A number of reviews have appeared in recent
years in which the significance of the various reaction pathways for S02
conversion in the atmosphere has been estimated (Calvert and McQuigg, 1975J
Davis and Klauber, 1975; Sander and Seinfeld, 1976; Bottenhem and Strausz,
1977; Levy et al., 1976). Some significant new rate data related to these
systems have become available recently, and it is timely for us to re-
evaluate the various potentially important S02 reaction pathways. The
increasing use of these chemical reaction mechanisms in atmospheric models
makes a careful, continuing reassessment of the kinetics and rate constant
data related to these reactions of special value. We have summarized the
thermochemistry of several reactions of interest in Table 18, Most of the
reactions shown are exothermic, and from these energy considerations alone,
they warrent our attention, since they could occur at significant rates at
the temperatures of the troposphere. Also listed in Table 18 are our
recommended values for the various rate constants. Each is expressed as
an apparent second order rate constant which should be applicable for the
pressures of the lower atmosphere near 25°C. In this section we will con-
sider in some detail the data upon which each of these estimates were based.
The Reactions of 02(1A0.) and 02(1Z+) with S02--
g &
93
-------
TABLE 17. THE THEORETICAL RATE OF REACTION (ppm hr"1) OF S02
(3Bi) REACTIONS WITH VARIOUS IMPURITY SPECIES AMD 02 IN A
HYPOTHETICAL SUNLIGHT-IRRADIATED LOWER TROPOSPHERE*
Re act ant
molecule
NO
CO
C2H2
cis-2-C4Hg
iso-C4H8
so2
o,5
Reaction
No.
13 A1"
Bt
14 A
B
15 A
B
16 A
B
17 A
B
4 A
B
10 A
Initial [SO,] , ppm
500 50
3.1
8.9
9.9
2.9
.6.1
1.8
5.0
1.5
3.4
9.7
8.1
2.4
<9.0
x
x
X
X
X
X
X
X
X
X
X
X
10-2
10-2
ID'7
10~6
io-4
10-3
ID'2
10-1
io"4
io"4
io-3
10-2
3.1
8.9
9.9
2.9
6.1
1.8
5.0
1.5
3.4
9.7
8.1
2.4
<9.0
x
x
X
X
X
X
X
X
X
X
X
X
X
10-3
10-3
lO"8
10-7
10-5
io-4
10-3
10-2
io"5
io"5
ID"5
io-4
10-1
5
3.1 x
8.9 x
9.9 x
2.9 x
6.1 x
1.8 x
5.0 x
1.5 x
3.4 x
9.7 x
8.1 x
2.4 x
<9.0 x
ID'4
io-4
10-9
10-8
10 "6
ID'5
io-4
10-3
lo"6
io"6
io-7
10-6
10-2
0.
3.1
8.9
9.9
2.9
6.1
1.8
5.0
1.5
3.4
9.7
8.1
2.4
<9.0
,5
x
x
X
X
X
X
X
X
X
X
X
X
X
IO-5
10-5
10-10
10-9
10-7
io-6
10-5
1C'4
io"7
io-7
io-9
lO-3
10-3
0.
3.1
8.9
9.9
2.9
6.1
1.8
5.0
1.5
3.4
9.7
8.1
2.4
<9.0
.05
x ID"6
x lO"6
x 10-11
x 10-1°
x 10"8
x 10" 7
x 10 ~6
x ID"5
x 10~8
x 10"8
x IO-11
x 10-1°
x ID"4
B <17.0 <1.7 <1.7 x 10"! <1.7 x 10"2 <1.7 x 10"3
5,6,7 A <1.0 x 10-1
-------
TABLE .16. ENTHALPY CHANGES AND RECOMMENDED RATE CONSTANTS FOR
POTENTIALLY IMPORTANT REACTIONS OF GROUND STATE S02 and S03
MOLECULES IN THE LOWER TROPOSPHERE
Reaction _aH°* kcal k,f c
(19)
(20)
(21)
(22)
(23)
(24)
(25)
(26)
(27)
(28)
(29)
(30)
(31)
°2
°2
°2
°2
0(
t*
<*g
(li9
I1!*
3p)9
) +
) +
) +
+!
S02 •* SO4 (biradical) ;S04 (cyclic)
S02 -»• S03 + 0(3P)
S02 *°2<3V> + S02
+ SO2 •* S04 (biradical) ; SO (cyclic)
+ so2 •* so3 •*• o(3p)
- *• *• T
+ S02 (+M) •*• S03 (+M)
O, -t- SO, •+• 0, -t- SO,
J 4 £ 3
NO
NO
2 +
+
ONOO +
N->
2
HO
oc +
5
, +
so2
so
so
so.
•* NO + S03
* NO + SO
2 - N02 + S03
, ->• N,0. + SO,
2 24 3
-* HO + SO.
-13
22
"!
1-40 ;M3\
15. l)
83
57
10
32
•^30
24
19
.3
.8
.0
.8
.6
•3^
(3
6
(5
<8
8
<7
<7
<4
>(8
.9
.6
.7
x
.8
x
X
X
.7
+
X
±
0.9) x
ID"16
0.5) x
10-2°
10-14
10-24
x
10
10
10-30
-21
-21
ID'23
+
1.3) x
ID'16
(33) CH.,0, + SO, + CH,0 -t- SO, ^27 J. (5_3 ± 2^} ^ 1Q-15
(32) H02
(33) CH3i
(34) CH3O2 -t- SO2 •* CH302S02
(35) (CH ) CO + SO •* (CH,)3CO + SO3
• <7.3
(36) (CH,),CO- + SO, -* (CH,),CO,SO, "-30
•3JZ £ J 3 £ 4
\
J
(38) CH3C002 + S02 - CH3C002S02
(39) HO + S02 (+M) •+ HOS02 (-t-M)
(40) CH O + S02 (+M) * CH3OS02 (+M)
(41) RCH CHR + SO •* 2RCHO + SO,
(42) RCH-CHR + SO2 ->• 2RCHO + 50^
(43) RCHDO- + S02 -f RCHO + S03
(43a)RCHOO- + ^0 •* RCOOH + H2O
0-
(44) RCHO.-t- SOj-* RCHO + S03
(45) SO + 0(3P) (+M) •* S04 (biradical) ;S04 (cyclic)
(46) S03 + 0(3P) (+M) * S02 -t- 02
(47) S03 + H20 * H2S04
"33\ <1.3 X ID'18
^37 )
0,37 (1.1 ± 0.3) x 10"12
•v-24 ^6 X 10"15
^98 \ K43a - 1Q-5
-147 J k43
35.9 )
24.6 (9.1 ± 2.0) x 10~13
*Enthalpy change estimates were derived from the data of Benson(1976,1977), O'Neal
-139-
95
-------
The quenching reactions of 02(XA ) and 02(XZ +), reactions 19 through
£ &
2k, have been studied quantitatively in recent years. Penzhorn et al. (197*0
estimated the total quenching rate constant for 02(1A ) by S02, k + ^20 +
---^~ = (3-9 ± 0.9) x 10-20cm3molec-1s~1.
Op^Aj + S02 S04(biradical; S04(cyclic) (19)
Op^A.) + S02 -*- S03 + 0(3P) (20)
o
0£(XA ) + S02 •*- 02(3Z -) + S0a (21)
O O
The endothermicity and spin inversion of the possible reaction 20 ex-
cludes it from further consideration. The possible formation of the transient
S04 species in reaction 19 may account for at least a portion of the total
quenching reaction observed. Although many researchers have invoked the
intermediate S04 in their considerations of S02 photooxidation since this
suggestion was first made by Blacet (1952), only recently has any rather
direct evidence for the existence of the S04 intermediate species been ob-
served. The infrared kinetic studies of the 0(3P)-S03 reaction in the gas
phase by Daubendiek and Calvert (1972, 1977) gave strong indirect evidence
that the S04 intermediate was involved in the formation of transient molecules
S309 and S30s observed in their system. Kugel and Taube
(1975) have prepared an S04 species in C0a(s) and Ar(s) matrices at 78° and
15°K} respectively, through the action of 0-atoms on S03. It is probably
significant that Kugel and Taube saw no S04 formation when SQp was irradiated
with 2537 A in an 02 matrix at 15°K. Presumably 02(1A , 1S ) would be formed
& &
by interaction of the adjacent partners, S02(3B1)-02 under these circumstances,
and reactions such as 19 or 22 could then occur. Neither of these reactions
appear to be important, and the mechanism of deactivation of 02(1A ) by S02
D
remains unclear. It probably involves no chemical change in the S02 molecule,
but only electronic relaxation occurs as in 21. In any case for the usual
steady state levels of 02(1A ), about 10s molec cm"3, which we might
o
anticipate in a typical polluted urban atmosphere (Demerjian et al., 197*0>
the maximum rate of quenching by S0a amounts to an insignificant, i.k x 10~6%
hr'1.
Kear and Abrahamson (197V75) determined the rate constant for the
quenching of 02(XS +) by S02 to be: k22 + k + k , = 6.6 x 10~lscm3 molec'1
s'1.
02(X2 +) + S02 -+• S04(biradical); S04(cyclic) (22)
S
02(1Z+) + S02 + S03 + 0(3P) (23)
&
) + S02 •+• S02 + 02(1Ao.) (2k)
96
-------
The rate of 02(1Zg+) attack on S02 in the lower atmosphere must be very low
also, about l.k x 10-r% hr'1, for typical values of [02(1Z +)] = 6 x 102
molec cm" (Demerjian et al., 197*0. I^us we can concludesthat the rate
of singlet oxygen reactions with S02 in the lower atmosphere is insignificant,
and we may confidently neglect them in our further considerations.
The Reaction of 0(3P) with S0a --
The earlier kinetic work on the reaction 25 has been reviewed by
Schofield (1973), Hampson and Garvin (1975), and Westenberg and deHaas
(I975a).
0(3P) + S02 (+M) ->- S03 (+M) (25)
From the data available in 1973 Schofield's choice of constants gave k
(1.3 ± 1.3) x 10~l3cm3 molec'^-s"1 for the apparent second order constant at
1 atm of H2 (25°C). Hampson and Garvin picked the value of Davis et al.
(197^a) as their preferred rate constant: k - 2.0 x 10-14cm3 molec-1s-1
(1 atm N2, 25°C). Westenberg and deHaas (I975a) have concluded, however,
that most of the earlier estimates of k?c- which were based upon 0(3P) disap-
pearance rate are too high by a factor of two. Their recent findings shew-
that the reaction, S03 + 0(3P) -*- S02 + 02, is so very fast that a second
0-atom loss is expected to occur almost every time the reaction 25 occurs
.for the conditions employed in many of the studies. From their extimates,
k25 = ^'9 ± °*2') X 10"14cm3 molec~1s-1 for M = N2 at 1 atm, 25°C. The
only studies of kpc. with 02 as the third body come from the work of Mulcahy
et al. (1967); correcting their estimate for the occurrence of reaction h6}
in accord with the Westenberg and deHaas suggestion, we derive k^,- = (9-2 ±
1.7) x 1014 for M = 02 (1 aim, 25°C). Combining the best estimates for M =
N2 and M = 02, we derive, k = (5.7 ± 0.5) x l(Tl4cm3 molec ^s"1, M = air
(1 atm, 25°C). The estimate which we have chosen is in reasonable accord
with that observed by Atkinson and Pitts (197*0 using W20 as the third body:
k = 7.6 x 10~l4cm3 molec~1s-1 at 1 atm, 25°C; the greater number of
internal degrees of freedom of N20 than for H2 &&& °2 would result in a
somewhat higher value of k for M = N20. Our present estimate is con-
siderably higher than that of Davis et al. (197*0 with M = N2J "the
peculiarly high efficiency of S02 as M which was found by Davis et al. may
have resulted from an overcorrection for the contribution of the reaction,
0(3P) + S02 + S02 -*~ S03 + S02, to the total rate measured in the N2-S02
mixtures .
Our estimate of k coupled with a fairly normal [0(3P)] for a sunlight
c ^
irradiated, FOX- polluted, lower troposphere, 2 x 10 molec cur , the
anticipated oxidation rate for S02 by reaction 25 is about 1.2 x 10" % hr~ .
Thus we expect reaction 25 to be relatively unimportant for the conditions
normally encountered in the lower troposphere. We should retain this
97
-------
reaction in our simulations, however, since it is important, at least in
theory, during the early stages of stack gas dilution.
The Os-SOp Reaction —
In view of the great exothennicity of reaction 26, it is surprising
slow; the upper limit on its rate constant has been set by Davis et al.
(19?4b) as k2g < 10~22 and by Daubendiek and Calvert (1975) as kgg < 8 x
10~24cru
03 + S0a ->- S03 + 02 (26)
"3
Using our upper limit estimate of k?/- and an [03] = 5 x 1012 molec cm"
typical of photochemical smog in a highly polluted atmosphere, it can be
shown that the S02 oxidation rate through reaction 26 will be less than
l.U x 10" 5 % hr"1, truly unimportant for our further consideration.
Die Reactions of the Oxides of Nitrogen with S02--
The extrapolation of the high temperature rate data for the N02-S02
reaction 27 from Boreskov and Illarionov (19^0) leads to the estimate, k™ -
8.8 x 10-30 cm3 molec^s"1 for 25°C.
N02 + S02 -»- NO + S03 (27)
Data of Daubendiek and Calvert (1975) and Davis and Klauber (1975) give
kpn < 7 x 10~21 and about 10~21, respectively; these workers also report
k~0 < k x 10~23 and about 10~23cm3 molec~1s~1, respectively.
N03 + S02 ->• N02 + S03 (28)
W205 + S0a ->- N204 + S03 (30)
For typical concentrations of N02(5 x 1012 molec cm"3; 0.20 ppm), the con-
centrations of W03 and N20s which are expected in heavy photochemical smog
are about 2.5 x 107 and 2.5 x 109 molec cm"3, respectively. Thus the rates
of attack on S02 by N02, N03, and N205 should be of the order of: 1.6 x
10"11, 6.3 x 10"8, and 3.6 x 10~8
-------
The possible reactant OWOO, the intermediate species formed by the
N0-02 bimolecular interaction, is presumably in equilibrium with 10 and 0P
in air:
NO +
From the data of Benson (1976) we may estimate that the [(MX)]/ [NO] ratio
in air at 1 aim and 25°C is about 0.029. If the rate constant for reaction
29 is near equal to that for the analogous reaction 28 of the symmetrical
N03 species which it parallels in enthalpy change, then for the [NO] = 5 x
1012, [OKOO] = 1.5 x 1011 molec cm~3, and the estimated S02 oxidation rate
by reaction 29 will be less than 3.8 x 10~4f0 hr"1.
ONOO + S02 •+- N02 + S03 (29)
In summary, current data rule out the significant contribution of all
of the oxides of nitrogen (N0a, N03, ONOO, N205) to the oxidation of S02
within the lower troposphere.
The H02-S02 Reactions --
The study of Payne et al. (1973) provides the only experimental estimate
of reaction 31 of which we are aware. It was based upon a photochemical
competitive, 1802- labeling technique in which rate measurements of reaction
31 versus reaction h<3 were made:
H02 + S02 -*- HO + S03 (31)
2H02 -*- H202 + 02
They derived the estimate k^/kj* = (k.8 ± 0.7) x 10"10
taking the then preferred value of k^ = 3-3 x 10~12 cm3 molec'^-s"1 (Baulch
et al., 1972), they extimated k = (8.7 ± 1.3) x lO"16 cm3 molec-^-s'1. The
error limits shown are those inherent in the measurement of the rate constant
ratio alone. The uncertainty in k> is an additional source of error. Thus
Hampson and Garvin chose k^ = 5.6 x lO"12 at 25°C from a review of the rate
data available in 1975, while Lloyd (197^) chose k^ = 3-2 x lO"12 cm3 molec-1
s"1 at 25° in his evaluation. A brief review of the estimates of k^_ is
important here in that this value determines our k^ estimate through the
measured ratio, k_.,/k. '/a.
The first extimate of k^q was made by Foner and Hudson (1962); they
found k, = 3 x 10~12 cm3molec~:Ls"1 in a fast-flow, mass spectrometric
method of H02 detection. The later measurements of Paukert and Johnston
(1972) and Hockanadel et al. (1972), based upon ultraviolet detection of
H02, gave k^ = 3.65 x 1Q-12 and 9.58 x 10-12 cm3molec-1s-1, respectively,
99
-------
at 25°C. Harailton (1975) noted that the major difference between the reactant
systems employed by Paukert and Johnston and Hochanadel et al. was the
presence of H20 vapor (21 Torr) in the latter work; no H20 was present in
the former study. He reinvestigated the reaction 49 using varied amounts of
HpO vapor in a 1.5 MeV electron pulsed gaseous mixture of H2 (2 attn) and
02 '5 Torr). He followed H02 by uv-absorption spectroscopy. Indeed he
found that the k, q was sensitive to the pressure of HgO vapor present;
using Paukert and Johnston' s absorption coefficient for H02 for experiments
at PTT = 0, he derived a value, k, = 3-15 x 10~12 cm3molec~-Ls~:L, in
112 0 ^"9
excellent agreement with that of Paukert and Johnston. In experiments with
added water the k, Q/e values derived were near equivalent to those of
Hochanadel et al. In view of these results Hamilton and Kaleway (1976) have
proposed that a complex between water and H02 radicals forms in the H20-
containing mixtures:
H02 + H20.5£H02'H20 (50)
The apparent increase in rate constant for reaction 49 with increasing H20
seemed to result from the occurrence of the more rapid reaction 51 in ad-
dition to 49:
H02 + H02'H20 +- RsOs + 02 + H20 (51)
They estimated the thermodynamic properties of the H02*H20 species theoreti-
cally, and noted that the ratio [H02'H20]/[H02] was equal to about 0.037 in
air at 100$ relative humidity (25°C). About 3.5$ and 1.8$ of the H02
radicals in the atmosphere at 100 and 50$ relative humidity, respectively,
are in the complex at 25°C. The observations of Hamilton and Naleway
introduce a new and probably significant complication into the already
horrendous problems of the atmospheric scientist who desires to simulate the
rates of chemical changes in the troposphere. If the interpretation of
Hamilton and Naleway is correct as outlined, then modelers should employ a
value of k. q which varies with the humidity as well as the tempera-tare. We
have used the estimates of Hamilton and Naleway to calculate the fraction
of the H02 which will be in the form of the complex for various humidities
and temperatures; see Figure 36. Although the amount of H20 in the air at
a given relative humidity rises exponentially as the temperature is in-
creased, the stability of the complex decreases somewhat with increasing
temperature. The net result seen in Figure 36 is that there is small in-
crease in the fraction with increasing temperature (constant relative
humidity) . Although about 1$ of the H02 is expected to be complexed with
H20 at 0°C and 50$ relative humidity, about 4.5$ is complexed at 40°C and
100$ relative humidity. The immediate significance of these observations
for our purposes is the effect on the choice of value for k, which should
be used with the data of Payne et al. in deriving k . Note" that 20 Torr
of ^0 vapor was employed in all of the experiments of Payne et al., and
it may be more appropriate to use the higher value for the rate constant
k^ (9.6 x 10-12 cm3 molec-^-s"1) in estimating k . Thus the rate constant
k may be as high as 1.5 x 1Q-15 cm3 molec""^"1. Although we have chosen
100
-------
o
X
x
•2- 3
Q. °
O
u
tn 2
rt
CM
O
o
o
rH=100°/o
= 50%
O
10
20
30
40
Temperature, °C
Figure 36. Hie theoretical percentage of H02 in the
gas phase which is complexed with H20 vapor as a
function of the temperature and relative humidity;
estimated using the thermodynamic data from Hamilton
and Waleway (1976).
the lower value in our summary of Table 18, we have shown it as a lower
limit which must be increased if further experimentation proves the
Hamilton and Naleway hypothesis correct.
A typical [H02] expected in a sunlight-irradiated region of the lower
atmosphere which is highly polluted is about 6 x 109 molec cm~3 (Demerjian
et al.j 1974). This will lead to a rate of S02 oxidation through reaction
31 of about 1.9% hr"1. The [H02] for a fairly clean atmosphere may be lower
by a factor of 10 or so, and the rates of conversion of S02 by 31 will be
reduced correspondingly. However it is seen that this reaction can be a
major source of S02 conversion in the atmosphere, and it must be included in
our further considerations.
Reaction 32 which results in H02 addition to S02 has not been observed
experimentally and only speculation on its possible significance can be
made at this time.
H0
S02
H02S02
(32)
101
-------
Calvert and McQuigg (1975) have estimated by analogy with similar reactions
that k_? = 10~16 cm3 molec~1s~1. If this magnitude is correct, then
reaction 32 will result in only a few tenths of a percent conversion hr"1
for the highly polluted atmosphere, conditions which favor it. Benson's
enthalpy estimates (1977) suggest that the intermediate adduct H02S02 formed
in 32 will not be stable at 25°, since the decomposition reaction, H02S02
HO + S03 will be energetically favorable (AH = -12 kcal mole-1). Thus the
net effect of the occurrence of reaction 32 in the experiments of Payne
et al. (1973) may be the formation of HO and S03, hence their estimate
probably refers to the rate constant sum, k,,.. + k~2.
The CH302-S02 Reaction--
The first preliminary estimates of reactions 33 and 3^- have been
reported recently by Whitbeck et al. (1976).
CHs02 + S02 -*• CHsO + S03 (33)
CHs02 + S02 -+• CH302S02
They observed the kinetics of the decay of CH302 spectroscopically following
its generation in flash photolyzed mixtures of azomethane and oxygen, and
azomethane, oxygen, and sulfur dioxide. The range of initial concentrations
of S02 which could be employed was very limited (Por, < 0.3 Torr), since
b02
the products of the reaction formed an aerosol which interferred with the
optical measurements for experiments at high S02 pressures. The rate constant
estimate, independent of the absolute extinction coefficient for CH302, gave
k + k , = (5.3 + 2.5) x 10-15 cm3 molec"1^"1. If k > k,,> then the
magnitude of this estimate is in reasonable accord with that for the analogous
reaction 31. Taking the difference in entropies of activation for reactions
31 and 3^- as 0.8 eu, then the ratio of the observed rate constants k_../
k__ = O.l6 corresponds to an activation energy difference, E~, - E^o = 1*3
kcal mole"1, not unreasonable in view of the differences in enthalpy between
these reactions: AH,,, - AH.-,,, = 7.6 kcal mole"1.
The CH302 radical is probably the most abundant of the many organic
peroxy radicals in the atmosphere. In a highly polluted atmosphere it is
expected to be present at about 109 molec cm"3 (Demerjian et al., 197^),
and the oxidation of S02 may occur through 33 and 3^- by a rate as large as 2%
hr"1. In the fairly clean troposphere, rates of 1 x 10-2$> hr"1 are expected
(Crutzen and Fishman, 1977). Obviously we do want to include these reactions
in our further considerations.
It is not possible to determine the extent to which each of the reactions
33 and 3^- occur from the existing data, but thermochemical arguments favor
the reaction 33; thus the CH302S02 intermediate formed in 3^ may decompose
readily into CH30 and S03 after a short delay.
102
-------
The Tert-butylperoxy Radical-S02 Reaction--
Whitbeck et al. (1976) have studied the tert-C4H902 decay kinetics by
flash photolysis of (CH3)3C¥=NC(CH3)3 mixtures with pure oxygen and in
mixtures with 02 and S02. In this case the reactivity of the radicals
toward S02 was sufficiently low so that aerosol formation was not a serious
problem at short times. At long times (one minute) aerosol appeared. This
was attributed to the subsequent addition of (CH3)3CO radicals to S02
following their generation in the major C3H902 loss reaction in this system:
2C3H9Oa -*- 2C3H90 + 02. No difference in the decay rate of the tert-CsH902
radicals could be detected in runs with added S02 (0.2 Torr), so in this
case only an upper limit can be set for the rate constants: k.~ + k ,
7.3 x 10~19cm3
(CH3)3C02 + S0a -+- (CH3)3CO + S03 (35)
(CHs)3C02 + S02 -*• (CH3)3C02S02 (36)
This value is smaller than the estimate of k + k , by a factor of 10" 4;
thus we conclude that reactions 35 and. 36, as well as those of the other
highly hindered tert-alkylperoxy radicals, will be unimportant in the atmo-
spheric conversion of S02. The very marked difference in reactivity of the
tert-C4H902 and the CH302 radicals noted here with S0a, is also seen in the
R02-R02 reactions of these species: 2t-C4H902 •*- 2t-C4HgO + Oa; 2CHsOa ->-
2CH30 (or CHs-O + CH3OH) + 02. The rate constant of the former is 10~* to
10~5~ times that of the latter (Parkes et al., 1973; Parkes, 197^', Whitbeck
et al., 1976).
The Acetylperoxy Radical-S02 Reactions —
The peroxyacyl radicals are significant participants in photochemical
smog formation. Some of them formed in the polluted atmosphere react with
N02 to form the peroxyacyl nitrates:
RC002 + N02 -*- RC002N02 (52)
The possible significance of these radicals in S02 oxidation has not been
studied directly, but there is some information related to the reactions 37
and 38 which we can extract from existing published kinetic data:
CH3C002 + S02 •*- CH3C02 + S03 (37)
CHsC002 + S02 ->- CHsC002S02 (38)
Heicklen (1976) has quoted unpublished data of Shortridge which presumably
gives evidence that the peroxyacetyl radical reacts readily with S02 in
reaction 37. He found SOs addition to photolyzed mixtures of CHsCHO and 02
did not affect the rate of acetaldehyde oxidation, although S02 was removed
and an aerosol formed. It is not clear how this evidence proves uniquely
the occurrence of reaction 37, since many other chain carriers such as HO,
H02, CH30, CH302, etc., will be present and may also oxidize S02 without
terminating the chains .
103
-------
perhaps somewhat more definitive results related to the reactions 37
and 3c were reported by Pate et al. (1976). They observed the rate of
reaction of peroxyacetyl nitrate (PAN) with many compounds including S02.
The apparent bimolecular rate constant at 296°K for the PAN-S02 reaction
was found to be less than 1.35 x 10~23 cm3 molec~1s~1. However it has been
rather well established that PAN is in equilibrium with CH3C002 and N02,
and we can estimate from the data of Pate et al. rate information related
to reactions 37 and 38. The data from the Pitts group and those from Hendry
and Kenley (1977) clearly demonstrate the dynamic character of the PAN
present in gaseous mixtures :
CH3C002W02 -*- CHsC002 + N02 (53)
CH3C002 + N02 •+- CHSC002N02
Hendry and Kenley have estimated the rate constant k,-~ as a function of
temperature. For 23°C, the temperature of the experiments of Pate et al,
k „ = 2.65 x 10~4s~1. Taking their theoretical estimate of k,_h = 1.04 x
10" 12 cm3 molec~1s~1, we may estimate the [CH3C002] at equilibrium:
[CH3C002N02](2.6 x 10s)
[CHsC002]eq = - - molec cm"3 (55)
It seems reasonable to attribute any reaction of PAN to that of the CHsC002
radicals in equilibrium with it. Thus the observed rate constant (k )
may be related to the rate constant sum, k~7 + k~o, by relation 56:
(k + k o)[CH3C002][S02] = k [PAN] [S02 ] (56)
3 1 O4-1 G.A.JJ
Estimating [CHsC002] from the equilibrium data and relation 55 gives: k__ +
k c, = k [N02]/2.55 x 10s cm3 molec~1s~1. It is not reported what levels
.30 exp
of N02 were present in the experiments of Pate et al. , but it is .likely that
no more than 1 ppm (2.5 x 1013 molec cm"3) was present on the average. With
this estimate the data of Pate et al. lead to a value of k_ + k Q < 1.3 x
10~18 cm3 mole~1s~1. For the theoretically estimated levels of the per-
oxyacetyl radicals expected to be present in heavy photochemical smog ( ~ 2 x
10s molec cm"3), the new rate constant estimate suggests a maximum rate of
oxidation of S02 through 37 and 38 of about 1 x 10~4$, hr"1. We conclude
that the oxidation of S02 by CHsC002 radicals, and peroxacyl radicals in
general, is negligible for the usual atmospheric conditions. The results
of Fox and Wright (1977) can also be interpreted to support this conclusion.
They carried out sunlight-irradiated smog chamber experiments using dilute
mixtures of CsH6 and NOX in air and matched mixtures of these compounds but
with S02 added (0.75 Ppa) . They observed that S02 addition did not inhibit
significantly the formation of PAN; the peak concentration was not lowered
appreciably. The rate of [PAN] increase with time, if altered at all by
104
-------
S02 addition, -was somewhat more rapid in the SC^-containing system. Thus
in our opinion all of the definitive evidence at hand point to the un-
importance of the reactions 37 and 38 in the tropospheric conversion of S02.
The H0-S02 Addition Reaction--
The kinetics of free radical addition to S02 have been observed for a
number of free radical species; thus H-atoms (Halstead and Jenkins, 1968),
CHs-radicals (Good and Thynne, 196?a; Calvert et al., 197!; James et al.,
1973}t C2H5-radicals (Good and Thynne, 196?b), fluoroethyl radicals, CHa18
FCHs (Milstein et al., 197*0, cyclohexyl and other radicals formed during
recoil tritium reactions with cyclohexene (Fee et al., 1972), all react with
reasonably large rate constants which approach those of the analogous reactions
with oxygen in some cases. Of course the H-atoms and alkyl radicals formed
in the troposphere do not live to encounter S02 present at low impurity
concentrations. They react with oxygen largely, to form alkylperoxy radicals
and ultimately alkoxy radicals, H02, and other free radicals and molecular
products. As we have seen in our earlier discussions, the additions of
some alkylperoxy and acylperoxy radicals (reactions 36 and 38) to S02 appear
to be rather slow reactions, while those of H02 and CHs02 additions remain
unevaluated. Much of the important homogeneous chemistry of the impurity
molecules within the troposphere is dominated by the reactions of the
ubiquitous radical pair, HO and H02. Indeed it appears that the homogeneous
S02 removal paths also depend largely on the reactions of these radicals.
In particular the HO-addition to S02, reaction 39, seems to be the most
important of the several homogeneous reaction paths of S02 in the troposphere
for many different atmospheric conditions.
HO + S02 (+M) -*- HOS02 (+M) (39)
There has been a significant and productive effort of several research
groups to determine the rate constant for this seemingly important reaction.
It was apparently first suggested by McAndrew and Wheeler (1962) as one of
the reactions necessary to rationalize the effect of S02 on radical chain
terminations in propane-air flames. Following the suggestions of Fair and
Thrush (1969), the McAndrew and Wheeler rate constant estimations can be
used to derive a third order rate constant, k~Q = 1.1 x 1031 cm6 molec~2s-1
at 2080°K. Many of the estimates of k,,g were obtained in competive rate
studies using the HO reaction with CO as a reference.
HO + CO •+- E + C02 (57)
With this technique Davis et al. (1973) reported a third order rate constant
k = 3 x 10-31 cm6 molec^s-1 with M = H20. In similar experiments Payne
et al. (1973) found k = 1.5 x 10"31 cm6 molec^s"1 using the Stuhl and
Wiki (1972) estimate of the rate constant for reaction 57. Cox (197^/75>
1975) also carried out competitive S02-C0 reaction studies by photolyzing
HONO in air at 1 atm pressure; they found k^ = (6.0 ± 0.8) x lO"1 cm
molec-1s-1 in air at 1 atm, using k™ = 1.5 x 10'13 cm3 molec^s-1 as
105
-------
recoarr.enc.ed by Baulch and Drysdale (1974). The Castleman group (Wood et al.,
1974) also nave made rather extensive competitive rate studies of 39 i*1
experiments which have extended over a number of years. They photolyzed
H20 in a mixture of S02, GO, and 1^0 in N2 carrier gas. Their first work
gave k~ = 3.8 x 10"13 cm3 mole~1s-1 for the high pressure limit. They
recognised early the probable importance of this reaction in S02 conversion
in the atmosphere (Castleman et al., 1974; Wood et al., 1975)* In a later
similar study (Castleman et al., 1975), they found a pseudo-second order^
rate constant at 1 atm of N2, k = 6.0 x 10~13, taking k™ = 1.4 x 1CT13
cm3 iriolee~1s~1 (Hampson and Garvin, 1975). In the most recent report of
this group, Castleman and Tang (1976/77) extended their competitive rate
study from 20 to 1000 Torr of added N2 gas. Again accepting k , they
calculated: k = 6.0 x 10~13 cm3 molec~as~:L. In experiments at low
pressures carried out over a range of temperatures from 2k to -20°C, they
estimated Eocr = -2.8 kcal mole"1.
35
Some more direct measurements of k,,Q have been made by flash photolysis
HO-resonance fluorescence techniques. Using this method Davis and Schiff
(197*0 reported the reaction 39 to be in the high pressure fail off region
above .--10 Torr of He. At 500 Torr of added helium, the effective bimolecular
rate constant was reported to be: 2.5 x 10~13(2.5 x 10~3 shown, must be in
error) cm3 molec~1s~1, and the low pressure value of the third order rate
constant, k~q = 2 x 10~32cm6 molec~2s~1 with He = M (work of Davis and
Schiff quoted in Payne et al., 1973). In further experiments Davis (1974b)
reported preliminary values for the effective bimolecular rate constants
for k~ : with He as M, values varied from 0.87 x 10~13 at 50 Torr to 2.7 x
lO"13 at 500 Torr; for M = Ar, 1.37 x 10"13 at 50 Torr to 3.7 x lO"13 at
500 Torr; for M = N2, 0.80 x 10~13 at 5 Torr to 2.4 x 10"13 cm3 molee"1^"1
at 20 Torr. Harris and Wayne (1975) studied reaction 39 using the HO-
resonance technique but in a discharge flow system at low pressures. They
derived the third order rate constants: k-Q = (4.5 ± 1«5) x 10"31 with M =
Ar, and k = (7.2 ± 2.6) x 10~31 cm6 molec~2s~3- with M = N2. Atkinson et
al. (1976) also utilized H20 flash photolysis with HO-resonance spectroscopy
to determine k~q over a wide range of argon pressures. The estimated second
order rate constant for 25°C and 1 atm (M = Ar) was: k.~Q = (6.7 ± 0.7) x
10~13, with the high pressure limit (extrapolated) of 8.3 x 10~13 cm3 molec'1
s"1. Gordon and Mulac (1975) employed pulsed radiolysis and absorption
spectroscopy to follow the HO radical at 3087 A; in H20(g) at 1 atm. and
435°K; they obtained k,,Q = 1.8 x 10"12 cm3 molec-xs"1.
Recently Sie et al. (1976) and Cox et ai. (1976) have presented strong
evidence that the rate constant for the reaction 57j used as a reference
value in many competitive studies of reaction 39s was pressure sensitive.
Both groups of workers concluded that the rate constant k™ was very much
larger at 1 atm total gas pressure than had been assumed in view of the
106
-------
earlier low pressure data. A unique test of the pressure dependence of k
was not possible in the work of Cox et al. (1976), since their experiments
were carried out at a fixed pressure of 1 atm (largely air). However they
found the ratio k^/k^g to be about twice that observed by other experiment-
alists who used much lower pressures of reactants and added gas; they sug-
gested the phyothesis of a pressure dependent k seemed most compatible
with their results. ?'
HO + H2 -*- H20 + H (58)
in the work of Sie et al. (1976) the pressure of added gas (H2, He, SF6) was
varied from 20 to 77^ Torr with an observed increase in k,_7/k Q with pressure
by over a factor of two in the case of added H2 or SF6. Chan et al. (1977)
confirmed the conclusions of these workers. They reported from competitive
HO rate studies with dilute mixtures of CO and iso-C4Hlo in air, that the
rate constant ratio k^7/k was about a factor of two higher in experiments
at 700 Torr of air than in runs with 100 Torr in air.
HO + iso-C4H10 -*» C4H9 + H20 (59)
There is no reason to believe the pressure dependence observed by these
workers resulted from the normal bimblecular H-atom abstraction reactions
58 and 59j it is reasonable to assume that these rate constants remain at
their low pressure values: kg = 7.0 x 10~15, (Cox et al., 1976); k =
2.3^- x 10"12 cm3 molec"3^"1 (Greiner, 1970). Accepting this seemingly
plausible premise, we have derived estimates from the various competitive
experiments for k and have plotted these together with the data from the
direct measurements at low pressures in Figure 37- Excluding the one very
divergent point of Sie et al (1976) at 570 Torr of added SF6, the general
increase of the rate constant k „ with increasing pressure of H2, H2, 02,
and SF6 seems clear. Obviously values of the H0-S02 reaction rate constants
from the data at the higher pressures based upon the low pressure value of
k.-- must be corrected to take into account the pressure sensitive character
57
of k . Thus the estimates of k_Q by Davis et al. (1973), Payne et al.
(1973), Cox (1975), Castleman et al. (1975), and Castleman and Tang (1976/77)
should be corrected. We have made such a correction to the extensive and
seemingly accurate data set of Castleman and Tang, using the values cor-
responding to those of the dashed curve given in Figure37 for k . The
apparent second order rate constant derived in this fashion are plotted
versus the reciprocal of the pressure of N2 in Figure38 (triangles). Also
plotted here are the values of koo"1 determined more directly by resonance
fluorescence of HO in experiments at various pressures of Ar; the open
circles are data from Atkinson et al (1976) and the closed circles are data
from Davis (197^b). The two sets of data from the direct HO-measurements
do not check very well at low pressures but appear to reach about the same
value at high pressures. The data of Atkinson et al. give k (M = Ar, 1
107
-------
CO
*b
t>
in
To
V
0
e
in
3 —
2 —
20O
4OO
Pressure, Torr
6OO
8OO
Figure 37. The variation of the apparent second
order rate constant for the reaction 57? HO +
CO (M) -*- H + C02 (+M), with pressure of added
gases; data for H2, Sie et al., 1976 (open
circles); air, Cox, 1975 (diamond), Chan et al.
1977 (closed squares); SF6, Sie et al, 1976
(open triangles), Overend et al., 197^- > and
Paraskevopoulos, 1976 (inverted triangles);
He, Paraskevopoulos, 1976 (hexagons); low
pressure points shown: Stuhl and Niki, 1972
(inverted closed triangle); Greiner, 1967
(closed diamond); Mulcahy and Smith, 1971
(on- end triangle); Davis et al., 197^-c (open
diamond); Westenberg and deHaas, 1973 (closed
triangle); Smith and Zellner, 1973 (closed
hexagon) .
atm) = (6.7 ± 0.7) x 10-13 and a high pressure limit of 8.3 x 10~13 cm3
molec~1s~1. Since Ar is somewhat less efficient as a third body than 5T2
and 02J the latter number is probably the best estimate from the Atkinson
et al. data which is applicable to air at 1 atm. The Davis (197^b) data
are less abundant, but the extrapolated upper limit to k-Q for M = Ar is
*~yy
compatible with this estimate as well. The Davis data for k_Q using M =
N2 consisted of only 3 points (not shown in Figure 37) at low pressures (P <
kO Torr), so the value obtained by extrapolation of these data to 1 atm N2
(k,,q = 9 x 10"13) cannot be very accurate. We will take as the best estimate
of the Davis group the number quoted by Davis and Klauber (1975) for the
bimolecular rate constant in the troposphere, k sz 8 x 10"13 cm3 molec"1
s"1. The Castleman and Tang (1976/77) corrected estimates (using our
108
-------
Figure 38. Plot of the reciprocal of the apparent
second order rate constant for the reaction 39,
HO + S02 (+M) -*- HOS02 (+M), versus the reciprocal
of the pressure of added gas (M); estimates for
M = N2 were calculated from the data of Castleman
and Tang (1976/77) using corrected values for k,
(triangles); for M = Ar data are from Atkinson
et al., 1976 (open circles) and Davis,
(closed circles).
'57
preferred choice for k,_ = 3.0 x 10~13 cm3 molec~1s~1, Chan et al., 1977)
gives k = l.U x 10-12 cm3 molec""^"1 for M = N2 (1 atm, 2U°C). The Cox
(1975) extimate corrected to our preferred choice for k „ gives k~~ = (1.2 ±
0.2) x 10~12 cm3 molec"1s~1. We feel the best current choice for k,,^ in
the lower troposphere (l atm air, 25°C) is derived from an average of the
results of the four most extensive studies at the highest pressures, i.e.,
the data of Davis (197^b), Atkinson et al. (1976), and the corrected data
of Cox (1975), and Castleman and Tang (1976/77). We suggest the value:
k = (1.1 ± 0.3) x 10~12 cm3 molec-^-s""1 which is given in Table 18.
109
-------
1'he typical [HO] which is anticipated theoretically in the highly
polluted (NO*, EH, S02), sunlight-irradiated, lower atmosphere, is about 7 x
10s molec cia"^ (Calvert and McQuigg, 1975). That expected for a relatively
clean atmosphere in the midday summer sun in the midlatitudes of the northern
hemisphere, is somewhat lower, typically 1 x 106 molec cm"3 (Crutzen and
Fishman, 1977). Using these estimates we predict that S02 removal by
reaction with HO in 39 may be as high as 2.7% hr"1 in the dirty atmosphere
and typically O.k% hr"1 in the clean troposphere. Our evaluation certainly
confirms the current view of atmospheric chemists that reaction 39 is a
very important factor in the homogeneous oxidation of S02 in the atmosphere.
The Fate of the HOSOa Product of Reaction 39--
The HOS02 species formed as reaction 39 occurs is not a stable product,
but it is a free radical which will react further to form final products.
Calvert and McQwigg (1975), Davis and Klauber (1975), and Benson (1977) have
speculated on the subsequent events expected following the formation of
HOS02. It is important to review these possible steps in light of recent
information concerning them.
AH , kcal mole"1
HO + S02 (+M) -*- HOS02 (+M) -37 (39)
HOS02 + 02 -*- HOS0200 -16 (60)
HOS0200 + NO -*- HOS020 + N02 -25 (6l)
HOS0200 + N02"±Z^HOS02OON02 ? (62)
HOS02OON02 -*- HOS020 + N03 ? (63)
HOS0200 + N02 -V HOS020 + NOS -2 (6k)
HOS0200 + H02 -»- HOS0202H + 02 -if-3 (65)
2HOS0200 ->- 2HOS020 + 02 -22 (66)
HOS020 + NO -*- HOS02OWO -26 (67)
HOS02OI[0 + he -*~ HOS020 + NO (68)
HOS020 + N02 ->- HOS02ON02 -22 (69)
HOS020 + H02 ->- HOS02OH + 02 -57 (70)
HOS020 + C3He -*- HOS02OH + iso-C^ -10 (71)
HOSO^O + CsHg •*• HOS02OCH2CHCH3 ? (72)
H2S04 + aerosol (H20,KH3,CH20,CnH2n..) -^ (growing aerosol) (73)
110
-------
HOS02ON02 + aerosolCHgO) -*- aerosol(H2S04,HON02..) (7k)
HOS02ONO + aerosol(H20) •>- aerosol(HaS04j HOWO..) (75)
The first step In the sequence of HOS02 reactions,, reaction 60, should be
the major fate of the HOS02 radical in the troposphere, it is exothermic by
16 kcal mole-1. The alternative disproportionation reaction, HOS02 + 02
H02 + S03, will be non-competitive with 60 since it is endothermic by about
8 kcal mole"1. As with the H02 and R02 radicals, it is likely that the
HOS0200 radical will oxidize NO, reaction 6l, or react with N02 in reaction
62, since NO and N02 impurity will be present with S02 in most of the
impurity-laden air mixtures encountered. Reaction ,6l is more exothermic
(-25 kcal mole'1) than the comparable, analogous, fast reaction, R02 + NO
RO + N02 (-17 kcal mole'1). The coupling reaction 62 should be somewhat
exothermic, and it forms in theory an inorganic analogue to peroxyacetyl
nitrate (PAN) as suggested by Calvert and McQoigg (1975). This compound
would be expected to be somewhat unstable toward decomposition, and reaction
62 is written as reversible. Benson (3-977) has not reported estimates for
AHf(HOS02OON02), but he noted that both the adducts of the HOS0200 radical
with NO and N02 will not be stable; the reaction 64 with N02 to form HOS020
and N03 is slightly exothermic. Disproportionation of the HOS0200 radical
may occur with the H02 radical in 65, or, in the absence of other radicals,
it may react with another HOS0200 radical in 66. It is unlikely that HOS0200
will oxidize S02 at any significant rate: HOS0200 + S02 -*- HOS020 + S03
(AH = -35 kcal mole'1), in view of the slowness of the reaction 37 of near
equal AH . If this is important, then rates of S02 oxidation estimated in
this work are minimum rates.
The HOS020 radical in this sequence is somewhat analogous to the HO
radical. It may form sulfuric acid by abstracting a hydrogen from a hydro-
carbon, aldehyde, or other H-containing species such as H02 in reactions
70 and 71. However the H-atom abstraction reaction of the HO-radical
analogous to the typical reaction 71 is much more exothermic (AH = -25 kcal
mole-1) than reaction 71, so one may expect the rate constant for H-
abstraction by HOS020 to be somewhat smaller than the analogous HO reaction.
The HOS020 radical may add to alkenes and generate organic sulfate containing
species in reaction 72. It may react with NO in reaction 67 and form the
well known reactive reagent for nitrosation and oxidation of organic com-
pounds and diazotization of amines, nitrosylsulfuric acid, ONOS02OH. It is
interesting to observe that this compound is available now commercially
(duPont) in a sulfuric acid solution. Nitrosylsulfuric acid may photolyze
in sunlight in 68, react in aerosol solutions to hydrolyze to H2S04 and
HONO, or react with organic matter present in the H2S04 -rich aerosol in
75. Alternatively the HOS020 species may combine with N02 in 69 and form
the nitrylsulfuric acid. In principle this compound could react on aerosols
to hydrolyze to H2S04 and HON02, or it may act to nitrate certain organic
compounds present in the aerosol solution, reaction 7^.
It seems probable to us that the reactions of the HOS0202 and HOS020
radicals which are shown do occur in the atmosphere, and the reactive
species formed in these reactions should be considered as prime candidates
for the active forms of the "sulfate" aerosol of our urban atmospheres.
Ill
-------
The assignment of rate constants for these reactions of the basis of our
present knowledge is very speculative. However recognize that once the HO
radical has added to S02, the reactions proceed to form sulfuric acid,
peroxysulfuric acid, and other compounds which will eventually lead to
sulfate and nitrate containing aerosols. If the aerosol is rich in un~
neutralized H2S04, then much of the nitrate will be lost to the atmosphere
as nitric acid.
Davis and Klauber (1975) and Davis et al. (1974d) have suggested an
alternative reaction route for the HOS020 radical which we should consider
farther here. This is the chain reaction sequence 76, 77 > ancL 6l which can
presumably pump NO to N02 as in the H02, HO cycle in smog.
HOS020 + 02 (+M) -*- HOS0203 (76)
HOS0203 + NO -»- HOS0202 + N02 (77)
HOS0202 + NO -*- HOS020 + N02 (6l)
The reaction 76 is analogous to the reaction 78 involving the HO radical:
HO + 02 -*- HOOO (78)
Using Benson's (1976) estimates for AHf(H03) = 60 eu (1 aim, 25°C), we
calculate that at equilibrium in the lower troposphere at 25° the ratio
[HOOO]/[HO] = 2.6 x 10~19. The greater endothermicity of the reaction 76
compared to 78 will lead to a still lower ratio of [HOS0203]/[HOS020] in
the lower troposphere. Thus it is improbable that the proposed reactions
76 and 77 occur in the atmosphere.
The observed 03 -bulge which Davis and colleagues saw late in the
transport of a stack gas plume does not require reaction 76 and 77 for
explanation. The conventional reactions involving NO to N02 conversion
chains of typical smog are the likely 03 developing mechanism in a plume
which is well diluted with the usual contaminants of polluted urban air.
The Methoxy Radical Addition to S02—
There are no kinetic data related to reaction kO of which we are aware.
CHsO + S02 -*- CHsOS02
Calvert and McQuigg (1975) have presented a very rough estimate: k. _ s 6 x
10~15 cm3 molec~1s~1, which was based upon comparisons with analogous
reactions of other radicals. An experimental estimate should be made. In
contrast to the CH302 and other alkyl peroxy radicals, the CH30 radical and
other alkoxy radicals are reactive toward molecular 02; e.g., CH30 •*• 02 -*-
CH20 + H02. Thus the steady state concentrations of CHaO and other alkoxy
radicals within a sunlight-irradiated, polluted, lower atmosphere are quite
low, about 5 x 10s molec cm3, Using this estimate the rate of S02 reaction
in ifO should be about 0.01$ hr"1. This reaction seems to be an unimportant
112
-------
loss mechanism for S02 in the atmosphere. However in view of the reactive
nature of the potential alkylating agents which should form eventually
following reaction 40, an experimental determination of the rate constants
for alkoxy radical reactions with S02 should be made to allow an accurate
assessment of the significance of this reaction.
The Oxidation of S02 by Products of the 03 -alkene Reaction--
As we have seen in the preceeding section, the rate of oxidation of S02 by 03
reaction 26, is negligibly slow at smbient air temperatures and at the lew-
levels which are characteristic of 03 and S02 impurities in the troposphere.
However when a third reactant, an alkene is present in the 03-S02-air mixture,
a fairly rapid oxidation of S02 occurs even in the dark. This very sig-
nificant observation was first reported in detail by Cox and Penkett (1971a,b).
For their conditions the rate of removal of S02 was 3% hr"1 with the alkene
cis-2-pentene and 0.4- RCHO + RCHOO* (80)
RCHOO- + S0a *- RCHO + S03 (43)
In a later more extensive study, Cox and Penkett (1972) also considered the
original ozonide (molozonide) product of 79 as an alternative reactant in
41:
0-0-0
RCH- -CHR + S02 -*• 2RCHO + S03 (4l)
The rate data from a series of different alkenes (cis-2-butene, cis-2-
pentene, trans-2-butene, 2-methyl-l-pentene, and 1-hexene) were rationalized
well by a simple mechanism in which an intermediate, presumably the original
molozonide or the Criegee intermediate, oxidized S02 through the reaction
series 79, 80, 43, and/or 4l, or underwent decomposition, reaction at the
wall, or some other fate unproductive to S02 oxidation. Although a strong
inhibition of the S02 oxidation occurred with increased [H20] in the reaction
mixture, the effect of H20 remained puzzling and unexplained by Cox and
Penkett. Wilson et al. (1974) accepted a similar mechanism in their computer
simulation of the Cox and Penkett Os-alkene-SOs-air results.
In the reaction schemes considered here for this system, two other
reactive entities related to the molozonide and its products should be noted
as well as potential reactants: the open form of the original molozonide
(reaction 42), and the rearranged Criegee intermediate (reaction 44):
113
-------
RCH-CHR + S02 -*• 2RCHO + S03 (42)
RCH-0' + S02 -*- RCHO + S03 (43)
Other possible reactants for S02 in the alkene-03~S02-air system have been
suggested. Demerjian et al (1974) speculated that the reactant diradical
of 44 might be formed from the original Criegee intermediate by reaction
with 02:
RCHOO- + 02 -*- -OOC(R)HOO- (8l)
p-o
•OOC(R)HOO- -*• 0 ^0 ->- 02 + -OCH(R)0« (82)
RCH
Alternatively O'Neal and Blumstein (1973) envisaged a rapid rearrangement of
the Criegee intermediate through reaction sequence 83:
RCHOO' •*- hc(R)H -*--OCH(R)0' (83)
0
Thus at least four different reactants should be considered as potential
S02 reactants in this system; note in Table 18 that all of the potential
reactions 4l-44 are considerably exothermic. However there is some
uncertainty whether one needs to invoke any of these reactions to explain
S02 oxidation. Demerjian et al. (1974) have argued that the 'OCH(R)0*
radical will react readily with oxygen to generate other reactive species
(RC02, H02, R02, etc.), and Calvert and McQuigg (1975) suggested that these
may be responsible for the S02 oxidation observed in the alkene-Os-S02
system.
The 03-alkene reactions are very complex, and the simple Criegee
mechanism cannot be the only route of fragmentation and rearrangement to
products in the gas phase system. There is abundant evidence today that
fragmentation of the ozonide formed with the simple alksnes in gas phase
reactions creates highly excited free radicals including HO and carbonyl
species (Kuramer et al., 1971J Pitts et al., 1972; Finlayson et al., 1972;
Atkinson et al., 1973). It is difficult to rationalize the formation of
these highly excited species through the Criegee mechanism alone. O'Neal
and Blumstein (1973) have suggested several alternative paths for de-
composition of the original ozonide which can account better for some of the
chemiluminescent products of the gas phase ozone-alkene reactions. Various
modified forms of the 0 Neal and Blumstein mechanism have been adopted by
various groups of modelers of smog chemistry; for example, see Whitten and
Hogo, 1976; Sander and Seinfeld, 1976. However the extent and the nature
of the fragmentation and rearrangement paths which are chosen must neces-
sarily be rather arbitrary at this stage of our knowledge. Use of the
-------
Benson (1976) thermochemical-kinetic considerations have been made by O'Neal
and Blumstein (1973) and others to derive "best estimates" of the importance
of the alternative paths of reaction of the ozonides. In most detailed
modeling schemes involving S02 removal which are in use today, reactions
such as IkL, h2, ^3, or kk are not considered to be important, but the S02
oxidation in the alkene-03-S02 system is made to occur exclusively through
partial fragmentation of the ozonide to various free radicals, followed by
reactions 31-^0 or Table 18 and their various analogues. For examples see
Sander and Seinfeld, 1976; Walter et al., 1977.
However it appears to us that the ideas related to the mechanism of the
gas phase 03-alkene reactions must be modified again. Recent FTS-IR spectro-
scopic observations of the Os-alkene-air systems have been made by Mki et
al. (1977) • They provide striking new evidence 'for the reasonable stability
of many of the gas phase ozonides. Thus the role of reactions ^1-^ should
not be discarded. Hull et al. (1972) had observed through low temperature
(-175 "to -80°C) infrared studies in the condensed phase, the primary ozonides
of C3Hs, iso-C^iHs, cis- and trans-2-C4H8, cyclopentene, cyclohexene, tri-
methylethylene, and tetramethylethylene ; these decomposed to various products
upon warming the frozen mixtures to room temperature. In 1959 Hanst et al.
concluded from long-path infrared experiments that they observed ozonide
formation from the gas phase reactions of 03 with 1-hexene and 3-heptene
but not from 1-pentene and the smaller alkenes. However the recent work of
Niki et al. (1975) shows clearly that ozonide formation can be observed in
the gas phase from olefins as small as propylene. In their work, Niki et al.
reacted 03 (5 ppm), cis-2-C4Hs (10 ppm), and CH20 (10 ppm) in 700 Torr of air.
The identified products (ppm) included: C02(1.9), CO, CH4(0.66), CH02H(0.25),
CHsOH(O.if), CH2CO, and, significantly propylene ozonide (0.88 ppm). .Thus
they presented unambiguous evidence that the Criegee intermediate CHsCHOO-
9
(or conceivably the rearranged radical, CH3CHO*) had a much longer lifetime
in air than has been suggested in recent years by O'Neal and Blumstein (1973)»
Demerjian et al. (197^), and many others; it lived to react with the added
CH20 and formed propylene ozonide:
/O-O
+ CH20 •*
For the purposes of this review another most important observation was
made by Niki et al. (1977); they found that addition of S02 at the 5 ppm
level to the 03, CsHs, CH20 system quenched ozonide formation completely,
and the S02 was consumed to an extent comparable to the ozonide yield ob-
served in its absence. Furthermore the rearrangement of the CHaCHOO* species
to acetic acid was not observed in these experiments, although it is ex-
pected in the O'Neal and Blumstein considerations:
CHsCHOO- -: CHsCH
0 -0-
CH3CH -*- CH3COOH (85)
0 0-
115
-------
liiki (1-977) has indicated to us that the ozonide yield is about 20$ for all
of the alkenes .studied except CgR^ where no ozonide formation was detected.
The "radical" product yield was less than
At this writing it is not clear what species we should invoke as the
reactants in the 03-alkene-S02 system, but Niki's work demonstrates that the
Criegee intermediate is a strong candidate. Some reaction from various free
radical fragmentation products may occur as well. With this apparent return
to the original mechanism suggested by Cox and Penkett (1971a,b)s it is
instructive to reconsider the detailed results of Cox and Penkett (1972).
A major problem in the quantitative evaluation of the rates of reactions
^1-M+, is the lack of information related to the absolute rate constants of
the intermediate, S02-oxidizing species. In the mixture of gases encountered
in the polluted troposphere, NO is commonly present with S02. If the
Criegee . intermediate is formed by ozone-alkene interactions, then S02 must
compete with N.O (reaction 86) and other reactants as well in order to be
oxidized by RCHOO-.
CH3CHOO' + NO -*~ CHsCHO + N02 (86a)
;•
or CHsCH-O' + NO -*- CH3CHO + N02 (86b)
The enthalpy changes for these reactions (AHo/r = -97; AHo/:, = -51 kcal
mole"1) may be compared to that for the oxidation of NO by 03 (AH o7 =~48
kcal mole"1), a reactive species which is very similar to the Criegee inter-
mediate .
03 + NO -»~ 02 + N02 (87)
It seems likely that kg,- > kg/- = kn7 = 1.6 x 10~14 cm3 molec'-'-s"1 (Clyne
et al., 1964), and a gooa competition with the possible S02 oxidation
reactions must be provided by NO in the atmosphere. In the absence of an
experimental basis for the evaluation of the inhibition of S02 oxidation by
NO addition to the 03-alkene-S02 system, it will be impossible to judge
accurately the significance of reactions such as ^3 and hk in the atmosphere.
It is interesting to note that AH< * is comparable to AHpx-, and reaction
26 is immeasurably slow at room temperature. This evidence seems to favor
the direct involvement of the Criegee intermediate and not the rearranged
0*
I
form, RCH-0' in S02 oxidation.
Let us adopt for the purposes of our considerations here an expanded
version of the Cox and Penkett mechanism for the cis-2-butene, Os, SOo,
H20, air system, and focus our attention on the Criegee intermediate as
the major source of S02 oxidation in these experiments. The competitive
reactions of the Criegee intermediate in the NO-free system may occur in
principle with S02, the alkenes, ozone, oxygen, water, as well as the
unimolecular decay to other products :
116
-------
•
03 + cis-2-C4H8 ->- (molozonide) ->- CH3CHOO« + GHsCHO
-»- (ECHO, RC02H5 etc.)
CHsCHOO- + S02 -*- CHsCHO + S03
*
CH3CHOO • + C4He -*- CH3CHO + CtfsO (and other products)
CHsCHOO- + 03 ->- CHsCHO + 202
0-
I
CH3CHOO- + 02 -*• CH3CH-0- (other products)
03 + CHsCHO
CHsCHOO- + H20
CHSCHOO- •
CHsCOOH + H20
1=
(CHsCOOH) •*- CH4 + C02(CHsOH, CO, etc.)
(88)
(89)
(90)
(91)
(92)
(93)
(94)
(95)
(96)
Each of these steps is energetically feasible, but the extent of the
participation of each is unclear. Since the rate of Os-alkene reaction
and the rate of S02 oxidation in alkene-03-S02 mixtures was relatively un-
affected by replacing the air with N2 ( 0.2$) in the experiments of Cox
and Penkett (1972), then reactions 93 and 9^- are probably unimportant.
Deviations of the 03-alkene stiochiometry from 1 to 1 in reactant mixtures
with excess of 03 or excess of alkene, suggest that 91 and 92 may occur to
some extent. The marked inhibition by H20 which they observed is consistent
with some reaction such as 95- One might speculate that H20 catalyzes the
rearrangement of CHsCHOO- to CHsC02H by way of some complex between the
Criegee intermediate and water:
CH3 — C—H-
CH3 — C
0
The Cox and Penkett data for cis-2-butene, S02, 03, 1^0, air mixtures
can be reconsidered in terms of these reactions. Following Cox and Penkett
we may assume that the reactive Criegee intermediate achieves a steady state
concentration. Let us assume further that S02 oxidation occurs only in
reaction 90 and that E « Egg + EQ . Then we expect relation 97 to hold:
-R§3 / % +
~\
k,
(97)
117
-------
The rather limited data cannot provide a test of the functional form for
each of the reactants in relation 97, but certain observations can be
made. In the Cox and Penkett experiments the [03]° and [C4H8]° were held
essentially constant at 0.5 and 1.0 ppm, respectively. For these conditions
the form of the relation 97 simplifies to 98, where A and B are constants:
1 (98)
[S02]
From relation 98 we expect a linear relation between -R° /Rqn an<^
for a series of runs made at constant [H20] . This was observed by Cox and
Penkett, and such a plot has been redrawn from the original data; see Figure
39. Observe that the slope to intercept ratios for the linear plots at each
[H20] should give the quantity, (A + [H20]k )/k , where A = k g + [C4Hs]
kq, + [03]k . For runs at 76, kO and 10$, relative humidity, the slope/
intercept ratios in Figure 39 are: 1.^3 ± 0.26, 0.83 ± 0.21, and 0.30 ±
0.02 ppm. These are plotted versus [H20] in Figure kO and the expected linear
dependence is seen. The intercept and slope in Figure hO give A/k = (6.1 ±
0.3) x 10~5, respectively. The kinetic treatment given here differs somewhat
from that of Cox and Penkett in that we have attempted to include explicitly
the reaction of H20 with the intermediate. The data do suggest that the
reaction with water is by far the dominant reaction of the intermediate; the
ratio of rate of the reaction with water to that for all other reactions of
the intermediate varies from about it- for experiments at ^0$, relative
humidity to about 8 at 76% relative humidity. This striking effect for
water noted by Cox and Penkett suggests indirectly that R02, H02, HO, and
other radical species which may be formed in the alkene-S02-0 -H20 air
system are not the important oxidizing species present. Their reactivity
toward H20 is thought to be very low, and their concentrations will be
altered insignificantly with H20 increase.
There is one other rather thorough study of the alkene-03-S02-air
system of which we are aware. McNelis (197^-) made a kinetic study of the
CsHs, Os, S02, air system, and it is interesting to compare these results
with those of Cox and Penkett. First it must be observed that McNelis
did not vary the relative humidity over a wide range in his study so a
quantitative test of the effect of H20 is not possible. However he did
observe in otherwise similar runs at 20$. and 36$ relative humidity, a
conversion of 0.071 and 0.066 ppm S02/ppm 03 consumed, respectively, the
direction of the trend seen by Cox and Penkett. In retreating the McNelis
data we have assumed that the reaction with H20 is present and have used
the same mechanism outlined for the butene studies. We recalculated the
apparent second order rate constants for the Oa-CsHg reaction from the
McNelis data, using the conventional second order rate law and the observed
data for [Os] and [CsH6] versus time. The initial rate method used by
McNelis seemed less accurate to us. The seemingly consistent set of data
from the McNelis runs 26, 27, 31, 87, 88, 89, and 90 were used. The data
gave k /k = (7-^ ± 1.5) x 10~5 for the propylene data, in reasonable
118
-------
o
Of
s
n
CSO23T
Figure 39. Plot of the ratio of rate of 03 loss
to rate of S03 formation versus [S02]-1 for the
data of Cox and Penkett (1972) from the dark
reaction in cis-2-butene~S02~03-H20-air mixtures
at various relative humidities (22°C).
[H2O], ppm, xlCf
Figure kO. Plot of the ratio of slope/intercept
for the plots of Figure 7 versus [H20] derived
from the data of Cox and Penkett (1972); the
slope gives the k /k estimate used in this
study.
-------
accord with the estimate from the C^Hs data. Reactions 95' &&& 90' refer to
the reactions of the species 'CH200* as well as CHsCHOO which are both formed
in the CsH6 system.
\-:^ can estimate the rate of S02 oxidation "by the reactive intermediates
formed in the 03-alkene reactions using the data derived here. Taking a
concentration of total alkenes = 0.10 ppm, [03] = 0.15 PP11^ a^d [S02] -
0.05 ppm, typical of a highly polluted, sunlight-irradiated urban atmosphere,
and using Mki's observation that about 20$ of the 03 molecules which react
with alkene form the Criegee intermediate, we extimate that the rate of S02
oxidation, presumably through ^3, will occur at about 0.23 and 0.12$ hr"1
at 50 and 100$ rH (25°C) when the reactivity of the alkene toward 03 is
typical of that for an alkene with a terminal double bond (k s 1 x 10~17 cm3
molec~1s~1) . In the unlikely event that 0.10 ppm of a highly reactive
alkene such as cis-2-butene were present together with 0.15 PF11 of ozone,
then much higher rates of S02 oxidation would be expected. However we have
neglected completely any loss reaction for the Criegee intermediate with
NO in these considerations, so the rates estimated represent theoretical
upper limits. Although these rates do not seem large in comparison with
those expected for HO, H02, and CHs02 reactions, they are not insignificant.
The Os-alkene reactions will continue to occur during the nighttime hours
when both reactants are present, and the S02 oxidation from the products of
this interaction will carry on as well; however, the rates are not expected
to be large.
The S03 Reactions in the Atmosphere- -
Reactions ^5 and h6 have generally been neglected in kinetic simulations
of the tropospheric chemistry of S02 and S03.
S03 + 0(3P) (+M) -»- S04
S03 + 0(3P) (-H4) -*- S02 + 0
Jacob and Winkler's 1972 estimate of the bimolecular rate constant for the
S03-0(3P) reaction, k^_ + k^ s 5 x lO"17 cm3 molec~1s-1, justified this
action. However more recently Daubendiek and Calvert (1973, 1977) and
Westenberg and deHaas (1975^) have found that these reactions are very much
faster than the earlier measurements suggested, and a new look at their
potential role in the troposphere should be taken. Daubendiek and Calvert
reported infrared kinetic studies from the 03 photolysis (A > 590 run) in
the presence of S03; they noted a rapid reaction (compared to 0 +• Os ->-
202) which destroyed S03, but surprisingly, there was a delay in the S02
appearance which is expected if reaction k-6 occurs. A metastable product
absorbing at 6.78 n was seen which decayed in the dark (-r,/2 = 280 s) to
form SOs and S02 . From the stoichiometry of the gases evolved the species
appeared to be a mixture of S30s and S30g. These seemed to form in the
reaction between the S04 initial product of ^5 and S02 and S03. The 0-S03
reaction appeared to obey second order kinetics for pressures of S03 and 03
above about 6 Torr. From the competitive rates of S03 and 03 reactions
with 0(3P),
120
-------
0(3P) + 03 -K 202
they derived k^. + k^ s 7 x 10-13 cm3 molec^-i, accepting k = 8.5 x
10"15 cm3 molec-is"1 at 2500 (Hampson and Garvin, 1975).
Westenberg and deHaas (I975b) reported rate studies of the 0(3P)-S03
reaction using a discharge flow method with ESR detection of the 0(3P).
They found no evidence for an S04 intermediate for their conditions,, but the
fast overall reaction 46 seemed to occur with third order kinetics up to
7 Torr of He. They derived k^ = 1.1* x ICT3^785/ cm6 molec~2s-1(298-507°K) .
The third order nature of the reaction which they observed at low pressures
suggests that an 304 intermediate is formed prior to rearrangement to their
observed products, S02 and 02. Indeed the results of both recent studies
give strong independent evidence for the rapid occurrence of reactions 45
and 46; where the data do overlap in pressures employed, good agreement is
found between the Daubendiek and Calvert and Westenberg and deHaas data.
Both sets are consistent with the estajuate: k, + k. /- = 7 x 10" l3 cm3 molec"1
s"1 in air at 1 atm. and 25°C.
Consider the possible implications of these findings on the atmospheric
chemistry of S02 and S03. Westenberg and deHaas (1975a) concluded from a
consideration of the two reactions 25 and 46 alone, that the main effect of
the presence of S02 on 0 in the atmosphere would be to catalyze its re-
combination with little or no net S03 formation.
0(3P) + S02 (+M) -*- S03 (+M) (25)
0(3P) + S03 (+M) •+- S02 + 02 (+M) (46)
Thus using our present estimates of the rate constants for air at 1 atm,
and assuming that S04 always forms S02 + 02 ultimately, then we find [S03]/
[S02] = 0.08 at the steady state. However there is no chance that a steady
state of SOs and S02 will be established involving reactions 25 and 46 in
the real atmosphere. The S03 molecule encounters with 0-atoms will be very
much less frequent than those with the abundant water molecule within the
lower troposphere. Recent evidence of Castleman et al. (1975) confirmed
the conclusion of Goodeve et al. (1934) that the overall reaction 47 is very
fast; they estimated k^ = (9.1 ± 2.0) x 10'13 cm3 molec"1s":L from flow
experience at low pressures. The rate constant estimate is surprisingly
S03 + E20 ->- H2S04
(47)
large for such a complex rearrangement of reactant molecules that must ac-
company this reaction; the observation of the mass 98 product in the ex-
periments of Castleman et al. may reflect the formation of an S03'HpO
adduct which is a precursor to the final rearranged, stable product molecule,
H2S04. In any case the ultimate removal of S03 by its net reaction 47 with
water is so very fast that the reaction of S03 with any other species in the
moist lower atmosphere should be unimportant. We conclude from the evidence
at hand that the usual assumption of modelers of the S02 conversion in the
121
-------
troposphere, namely, that H2C04 formation will always follow the generation
of SOa, appears to be sound.
I valuation of the Relative Importance of the Homogeneous SQ2 Reactions with
Various Reactive Species in the Troposphere
It is instructive to use the present evaluation of the S02 rate data
to estimate in greater detail the rates of S02 homogeneous oxidation which
we expect to occur in various clean and polluted regions of the troposphere.
First let us consider the rate of oxidation of S02 which is anticipated for
the relatively clean troposphere. The concentrations of the important
species, HO, H02, and CHs02 were estimated for various elevations within
the troposphere of the northern hemisphere', and for various latitudes and
seasons. If one considers the presence of S02 to be the ppb level or below
in these atmospheres, it is a fair approximation to assume very little
perturbation will occur in the estimated radical concentrations. Thus we
can use the Crutzen and Fishman (1977) estimates to calculate the theoretical
S02 conversion rates by reactions with HO, H02, and CHs02 in the troposphere.
Shown in Figure kl are the rates averaged over all northern latitudes and the
entire depth of the troposphere for each month of the year. Obviously the
rate for the HO-radical reaction with S02, reaction 39 > accounts for most
of the oxidation; a significant fraction is also contributed by the H02-
radical reaction 31 and a smaller amount from the CHs02 species. The total
average rate of S02 oxidation from these three most important homogeneous
reactions varies from a low of 0.09% in January to a maximum of 0.2% hr"1
in July. Of course these rates are directly related to the solar irradiance,
the temperature, and the atmospheric composition at the particular point
within the troposphere. One can observe in Figure 42 the theoretical rates
which are representative of the maximum values of the S02 oxidation by HO
and H02 which occur near the 1200 hr averaged over each day in July. These
are as high as 1.5$ hr"1 at ground level and about 32° N latitude. The rate
is somewhat lower at most latitudes as the elevation above sea level is
increased. At 8 km. a maximum rate of about 0.4$ hr"1 is seen near 28° H
latitude. At higher elevations (12 km) a small increase in rate can be
observed. It is clear from these considerations that the rate of S02
homogeneous oxidation in the relatively clean troposphere can be very sig-
nificant at certain time periods and positions within the troposphere.
Theoretical rates of S02 oxidation in parcels of highly polluted air
are of special interest to us as well. Previous computer simulations of the
reactions within a sunlight- irradiated, NOX, RH, RCHO-polluted air mass
have been made by Calvert and McQuigg (1975)? Sander and Seinfeld (1976),
and Graedel (1976). All workers agreed as to the importance of the HO- and
H02 -radical reaction rates with S02. For the simulated polluted atmosphere
chosen by Calvert and McQuigg and for an older less accurate set of rate
constants, a maximum S02 oxidation rate of 1.1$ hr"1 was predicted in
simulations carried out at a solar zenith angle of 4-0° . With a somewhat
different choice of elementary reactions, rate constants, and initial
reactants, Sander and Seinfeld predicted a maximum S02 homogeneous oxidation
rate in their simulated polluted atmosphere of 4.5$ hr"1, with HO and H02
radicals again accounting for much of the reaction rate. Graedel's
simulations employed the energetically unfavorable and unlikely chain
122
-------
H
.20
o
i I • * *~
$
o'10
.05
x
8
Q.
a
r i r
2 3 4 5 6 7 8 9 10 11 12
Month
4l. The theoretical monthly average of the
rate ( % hr""1) of S02 oxidation within the northern
atmosphere as a function of month of the year; rates
are shown for the HO reaction 39> the H02 reaction
31, the CH302 reaction 33, and the total of these
three rates; calculated using [HO], [H02], and
[CHs02] estimates of Crutzen and Fishman (1977).
c. 1.4
1.2
1.0
.8
.6
.4
.2
O km
fl oL_J I
15 25 35 45 55 65 75
Latitude (N), degrees
Figure ij-2. The theoretical rate (% hr"1) of
S02 oxidation by HO (reaction 39) and H02
(reaction 31) at various elevations within
the troposphere and at various latitudes in the
northern hemisphere; data are for the average
rates for the 1200 hour in the month of July;
calculated using the [HO] and [H02] estimates
of Crutzen and Fishman (1977).
-------
reaction sequence of Davis and Klauber (1975)> so comparisons with his
results are of questionable value. In Figure if-3 is given an updated version
of the Calvert and McQuigg simulations of S02 removal rates based upon the
new data and the new rate constant choices presented in this study. The
simulated atmosphere contained initially (ppm): [NO] = 0.15; [N02] = 0.05;
[cis-2-C4H8] = 0.10; [CO] = 10; [CEt] = 1.5; [CH20] = 0; [CH3CHO] = 0;
[S02] = 0.05; relative humidity, 50f0 (25°C); solar zenith angle, ^0°; a
stagnant air mass without dilution was considered to simulate the conditions
of highest smog-forming potential. Hie Os concentration in this simulation
rose to 0.15 ppm at the 120 min point of the irradiation at which time the
allcene had decreased to 0.013 ppm. The length of the ordinate within each
area of Figure 11 represents the theoretical % hr"1 of S02 oxidation by
the radicals shown. Note that the total oxidation rate from the several
species rises to OTer k% hr'1 at about 60 min into the irradiation. There
is one major difference between these results and those of the previous
simulations. The newly estimated rate constant for the CHs02-S02 reactions
33 and 3^, has been included, and this reaction has added an additional
increment to the S02 homogeneous oxidation rate expected in theory. It
appears that about equal rates of oxidation of S02 occur in a highly polluted
atmosphere through the reactions of the H02, CHs02, and the HO species
(reactions 31j 33, 39)- We have picked conditions, a high concentration of
very reactive alkene (trans-2-butene) and Iw- = 0, which should accentuate
the contribution of the Criegee intermediate. Even so the rates of S02
oxidation from reaction ^3 are relatively small compared to the rates of
the three major reacting species; the contribution from this reaction in-
creases to a maximum of 0.25% hr"1 at about 60 min into the irradiation
period. It is improbable that an actual polluted atmosphere would contain
this much reactive aUcene and that kn/- = 0. Hence we conclude that the
Criegee intermediate is probably never a major reactant for S02 oxidation
in the highly polluted atmosphere, although during nighttime hours this
reaction is likely the major homogeneous S02 oxidation mechanism which is
operative.
There is one other major class of S02 polluted atmosphere which is of
special interest to us. It is that contained in a stack plume from a power
plant or other industrial combustion operation. These gaseous mixtures
are of unique composition; the original effluent is oxygen-depleted and NO-
and S02- rich. Because of the very high NO levels and the very low hydro-
carbon and CO levels, the long oxidation chains which can lead to the
reasonably rapid conversions of S02 to SOs, HgSO^., etc., are suppressed
during the first stages of the plume transport. The initial [N02]/[NO]
ratio is very low so that the [Os] generated through the reaction sequence
100, 101, and 87,
N02 + hv(A < ^100 A)-*- WO + 0 (100)
0 + 02 (-tM) -*- 03 (+M) (101)
03 + NO -*- N02 + 02 (87)
124
-------
.c
.o
-RCHOO, 0(3P),
30 60 90
Irradiation Time, min
120
Figure 43 • The theoretical rate of attack of various
free radical species on S02 (% hr"1) for a simulated
sunlight-irradiated (solar zenith angle = ^0°),
polluted atmosphere; initial concentrations (ppm):
[S02] = 0.05; [NO] = 0.15; [N02] = 0.05; [CO] = 10;
[CH*] = 1.5; [CH20] = 0; [CHsCHOl = 0; relative
humidity, 50$(25°C).
125
-------
will be very much below the 0.04 ppm values observed in relatively clean
ambient air. For the conditions present in the plume after a short mixing
period, the 03 should reach a photostationary state for which the [03] will
be related to the [N02]/[NO] ratio, the apparent first order rate constant
for the rate of photolysis of N02 in 100 (k100), the rate limiting process
for Os formation in 101, and the rate constant kn7 for the dominant 03 loss
reaction 87.
[03] = [N02] k/[N01 k (102)
k,oo is a function of the solar irradiance in the 4100 to 2900 A region and
a typical value for k100/k87 is °*02 ppm for the 100° to lli00 time Period
(Calvert, 1976). Thus the ozone level will not reach the "clean air"
background values within the plume until the [N02]/[NO] ratio climbs to
near 2. Actual plume data show significant depletion of the ambient ozone
levels within the plume for very great distances. For examples see Davis
et al. (I974d) and Wilson et al. (1976). Only after the plume gases have
mixed with sufficient quantities of reactant hydrocarbons, aldehydes, CO,
etc., present in the ambient air can extensive NO to N02 conversion be
effected and 03 levels climb. However there are at least two mechanisms
which in theory can contribute to an initial burst of SOa oxidation involving
the elementary reactions of Table 18.
The first of these involves the photolysis of N02 (reaction 100) early
in the plume dilution when the [02] is relatively low, [S02] is high, and
the fraction of 0-atoms captured by S02 can be significant. N02 represents
a small fraction of the plume gases released to the atmosphere. Additional
N02 is formed as the ambient air mixes in with the NO-rich gases : 2NO + 02
-*- 2N02. For these conditions the S02 in the plume gases can compete
somewhat successfully with NO, N02, as well as 02 for the 0-atoms formed from
N02 photolysis in 100:
N02 + hv -*- 0 + NO (100)
0 + 02 (-W) -. 03 (+M) (101)
0 + N02 -*- 02 + NO (102)
0 + N02 (+M) -*- N03 (+M) (103)
0 + NO (4M) •*• N02 (+M) (iQlf)
0 + S02 (+M) -+- S03 (+M) (25)
The rates of S02 oxidation from reaction 25 alone in the sunlight-
irradiated stack gases can be significant during the early stages of
dilution. Thus if we assume stack gases to be at the concentrations: [NO] =
500, [S02] = 500, [N02] = 20; [02] = 1 x 10s ppm, for midday sunlight we
expect an instantaneous rate of S02 oxidation by 25 alone to be about l.V$
hr"1. This rate will drop fairly quickly as further dilution occurs with
126
-------
the transport of the plume. As the plume is further diluted by factors of
4, 8, 16, and 32, the instantaneous rates of S02 oxidation by reaction 25
are expected to fall to 0.1*3, 0.20, 0.10, and 0.05$ hr"1. A similar early
peak in the rate of the homogeneous oxidation of S02 by reaction 39 is
expected with the HO-radicals formed from the photolysis of HONO. The
HONO + hv(A < kOOO A) -*• HO + NO (105)
generation of HONO early in the dilution of stack plumes may occur in theory
through the homogeneous reaction 106 (Chan et al., 1976) or through hetero-
geneous paths involving water condensation or particulate surfaces:
NO + NQa + H20 -*- 2HONO (106)
Thus the development of as much as 10 ppm of HONO within a short dilution
period of certain stack effluents with very high NO, N02, and H20 con-
centrations may occur. When HONO is at the 10 ppm level with [S02] = 500,
[NO] = 500, [NO] = 20, [CO] =1 ppm, the instantaneous rate of S02 oxidation
by 39 is expected to be about 1.2$ hr"1 in noonday sunlight. In this case
02 is not a competitor for HO but the NO, N02, and CO are:
HO + S02 (H-M) -*- HOS02 (+M) (39)
HO + NO (+M) -»- HONO (+M) (107)
HO + N02 (+M) -*- HON02 (+M) (108)
HO + CO (-(M) -*• H + C02 (+M) (109)
After dilution of the reaction mixture by a factor of ten with clean air,
the rate of S02 oxidation by 106 again falls presipitously to 0.1$ hr'1.
Recently Miller (1977) has simulated the chemical rates of change
within a stack plume; the only S02 removal mode which he included was
reaction 39. It is interesting that he did observe a maximum rate of S02
oxidation (about 0.1$ hr"1) early in the dilution of the stack gases with
either unpolluted or polluted air.
As we have seen one anticipates in theory that S02 may be oxidized at
rates up to several percent per hour for a short period during the early
stages of stack gas dilution through the occurrence of the homogeneous
reactions 25 and 39 in a sunlight-irradiated plume. It is evident that the
apparent order of S02 conversion will appear to be higher than first during
the first stages of plume dilution. Such orders have been observed in
stack plumes . They have been rationalized well in terms of heterogeneous
reactions alone (Schwartz and Newman, 1977).
The homogeneous S02 conversion rates at long transport times are ex-
pected to increase as dilution of the stack gases with urban air containing
hydrocarbons, aldehydes, ozone, etc., occurs, and the rapid chain reaction
sequence familiar to smog chemistry take over. The S02 to sulfate conversion
rates estimated in plumes at long transport times in the study of Wilson
127
-------
et al. (1976) were observed to increase from about 1.5$ hr"1 during the 0.7
to 1.5 hr transport tirae period (10-21 km path) to about 5$ hr"1 after 2.3
to 3.2 hours of transport (32-^5 km path). Gillani et al. (1977) observed
in recent studies in the St. Louis area, maximum rates of S02 conversion
to particulate sulfur of the order of 3$ hr"1 in plumes with significant
transport time and dilution. They also presented some significant evidence
that homogeneous photochemical processes may be important in the S02 con-
versions observed.
From the considerations presented in the present study we conclude that
the homogeneous oxidation of S02 in the troposphere does occur at rates
which constitute a major fraction of the rates observed experimentally. It
is the hope of the authors that the discussion and the data presented in
this study will be useful to the atmospheric scientists in their continued
development of more quantitative models of the tropospheric S02 oxidation
and transport.
128
-------
REFERENCES
E.R. Allen and J.E. Bonelli (19?2b) The photooxidation of sulfur dioxide in
urban atmospheres. Abstracts, 6th Informal Photochemistry Conference
Oklahoma State University, p. 6.1-6.2. '
E.R. Allen, R.D. McQuigg, and R.D. Cadle (I9?2a) The photooxidation of
gaseous sulfur dioxide in air. Chemo sphere 1, 25-32.
E.R. Altwicker (1976) Ozone formation during sulfur dioxide irradiation.
J. Environ. Sci. Health All,
R. Atkinson, B.J. Finlayson and J.N. Pitts, Jr. (1973) Photoionization mass
spectrometer studies of gas phase ozone-olefin reactions. J. Amer.
Chem. -soc. 95, 7592-7599.
R. Atkinson, R.A. Perry and J.N. Pitts, Jr. (1976) Rate constants for the
reactions of the OH radical with N02 (M = Ar and N2) and S02 (M = Ar).
J. Chem. Phys. 65, 306-310.
R. Atkinson and J.N. Pitts, Jr. (197*0 Rate constants for the reaction of
0(3P) atoms with S02 (M = N20) over the temperature range 299-392 K.
Chem. Phys. Letters 29, 28-30.
C.C. Badcock, H.W. Sidebottom, J.G. Calvert, G.W. Reinhardt, and E.K. Damon
(1971) Mechanism of the photolysis of sulfur dioxide -paraffin hydrocarbon
mixtures. J. Amer. Chem. Soc. 93, 3115-3121.
D.L. Baulch and D.D. Drysdale (197*0 An evaluation of the rate data for the
reaction CO + OH-» C02 + H. Combust. Flame 23, 215-225.
D.L. Baulch, D.D. Drysdale, and A.C. Lloyd (1972) Evaluated kinetic data for
high temperature reactions, Vol. 1, Butterworths , London.
S.W. Benson (1976) Thermochemical kinetics, Second Edition, John Wiley &
Sons, Inc., New York, pp 1-320.
S.W. Benson (1977) Thermochemistry and kinetics of sulfur containing molecules
and radicals. Preprint of a paper to appear in Chem. Revs.
F.E. Blacet (1952) Photochemistry in the lower atmosphere. Ind. Eng. Chem.
*A, 1339-13^6.
G.K. Boreskov and V.V. Illarionov (19*1-0) J. Phys. Chem. Moscow ik, 1428;
quoted in Kondratiev V.N. (1970) Rate constants of gas-phase reactions,
Science Publishing House., Moscow, p. 233.
J.W. Bottenheim and O.P. Strausz (1977) Review of pollutant transformation
processes relevant to the Alberta oil sands area and proposals for
further studies. Teport to Atmospheric Environment Service, Dept.
Supply and Services Canada, Contract No. OSU 76-00169, from Hydrocarbon
Research Center, University Alberta, Edmonton, Alberta, Canada, T6G 2G2.
129
-------
J.C.D. Brand, C. diLauro, and V.T. Jones (1970) Structure of the 3BX state
of sulfur dioxide. J. Amer. Chem. Soc. 92, 6095-6096.
L.E. Brus, and J.R. McDonald (197*0 Time-resolved fluorescence kinetics and
1B1(1Ag) vibronic structure in tunable ultraviolet laser excited S02
vapor. J. Chem. Phys. 6l, 97-105.
•J.3. Calvert (1976) Test of the theory of ozone generation in Los Angeles
atmosphere. Environ. Sci. Techno!. 10, 248-256.
J.G. Calvert and R.D. McQuigg (1975) The computer simulation of the rates
and mechanisms of photochemical Smog formation. Int. J. Chem. Kinet.
Symp. 1, 113-15l|.
J.G. Calvert, D.H. Slater, and J.W. Gall (1971) The methyl radical-sulfur
dioxide reaction. In Chemical Reactions in Urban Atmospheres, Tuesday
C.S., editor, American Elsevier Publishing Co., pp. 133-158.
A.W. Castleman, Jr., H.R. Munkelwitz, and B. Manowitz (197*0 Isotopic studies
of the sulfur component of the stratospheric aerosol layer. Tellus 26,
222-234.
A.W. Castleman, Jr. and I.N. Tang (1976/77) Kinetics of the association
reaction of S02 with the hydroxyl radical. J. Photochem. 6, 349-354.
E. Cehelnik, J. Heicklen, S. Braslavsky, L. Stockburger III, and E. Mathias
(1973/7*0 Photolysis of S02 in the presence of foreign gases. IV.
Wavelength and temperature effects with CO. J. Photochem. 2, 31-48.
E. Cehelnik, C.W. Spicer, and J. Heicklen (1971) Photolysis of sulfur
dioxide in the presence of foreign gases. I. Carbon monoxide and
perfluoroethylene. J. Amer. Chem. Soc. 93, 5371-5380.
W.H. Chan, R.J. Nordstrom, J.G. Calvert, and J.H. Shaw (1976) Kinetic study
of HONO formation and decay reactions in gaseous mixtures of HONO, NO,
N02, H20, and N2. Environ. Sci. Techno!. 10, 674-682.
W.H. Chan, W.M. Uselman, J.G. Calvert, and J.H. Shaw (1977) The pressure
dependence of the rate constant for the reaction: HO + CO-*-H + C02.
Chem. Phys. Letters 45, 240-244.
K. Chung, J.G. Calvert and J.W. Bottenheim (1975) The photochemistry of
sulfur dioxide excited within its first allowed band at 3130A and the
"forbidden" band (3700-4000A). Int. J. Chem. Kinet. 7, l6l-l82.
M.A.A. dyne, B.A. Thrush and R.P. Wayne (1964) Kinetics of the chemilum-
inescent reaction between nitric oxide and ozone. Trans. Faraday Soc.
60, 359-370.
R.A. Cox (1972) Quantum yields for the photooxidation of sulfur dioxide in
the first allowed absorption region. J. Phys. Chem. 76, 814-820.
130
-------
R.A. Cox '1973) Some experimental observations of aerosol formation in the
photooxidation of sulfur dioxide. Aerosol. Sci. 4, 1^73-14.83.
R.A. Cox U97V75) The photolysis of nitrous acid in the presence of carbon
monoxide and sulfur dioxide. J. Photochem. 3, 291-3C4-
R.A. Cox (1975) The photolysis of gaseous nitrous acid -- A technique for
obtaining kinetic data in atmospheric photooxidation reactions. Int.
J. Chem. Kinet. Symp. 1, 379-398.
R.A. Cox and S.A. Penkett (1970) The photooxidation of sulfur dioxide in
sunlight. Atm. Environ, k, 425-1433,,
R.A. Cox and S.A. Penkett (I971a) Oxidation of atmospheric S02 by products
of the ozone-olefin reaction. Nature 230, 321-322.
R.A. Cox and S.A. Penkett (I971b) Photooxidation of atmospheric SOP. Nature
229, 1+86-1+88.
R.A. Cox and S.A. Penkett (1972) Aerosol formation from sulfur dioxide in
the presence of ozone and olefinic hydrocarbons. J. Chem. Soc. Faraday
Trans. I 68, 1735-1753-
R. Criegee (1957) The course of ozonization of unsaturated compounds. Record
Chem. Progr. 18, 111-120.
P.J. Crutzen and J. Fishman (1977) Average concentrations of OH in the
northern hemisphere troposphere, and the budgets of CH4, CO, and H2.
Geophys. Res. Letters, submitted for publication May, 1977j in a private
communication to one of the authors, computer readouts of [OH], [H02 J,
and [CHs02] were provided.
R.L. Daubendiek and J.G. Calvert (1973) A photochemical study involving S02,
02, S03 and 03. Abstracts 165th National Meeting of the Amer Chem. Soc.,
Dallas, Texas, Physical Section, paper #19.
R.L. Daubendiek and J.G. Calvert (1975) A study of the N205-S02-03 reaction
system. Environ. Letters 8, 103-116.
R.L. Daubendiek and J.G. Calvert (1977) An infrared study of the 0 + S03
reaction. Paper in preparation.
J.A. Davidson and E.W. Abrahamson (1972) The S02 photosensitized production
of 02(xZg+). Photochem. photobiol. 15, ^03-^05-
J.A. Davidson, K.E. Kear and E.W. Abrahamson (1972/73) The photosensitized
production and quenching of 02(12g+). J. Photochem. 1, 307-316.
D.D. Davis (I97^a) A kinetics review of atmospheric reactions involving Hx
O compounds. Can. J. Chem. 52,
131
-------
D.j. _,avis 'I974b) Absolute rate constants for elementary reactions of
atmospheric importance: results from the University of Maryland s gas
kinetics 1974; quoted in Hampson and Garvin (1975) 5 P« 55-
D.D. Davis, S. Fischer, and R. Schiff (I974c) Flash photolysis reasonance
fluorescence kinetics study: temperature dependence of the reactions OH +
CO-»-C02 + E and OH + CH4-*~H30 + CH3. J. Fhys. Chem. 6l, 2213-2219.
D.D. Davis and G. Klauber (1975) Atmospheric gas phase oxidation mechanisms
for the molecule S02. Int. J. Chem Kinet. Symp. 1, 5^3-556.
D.D. Davis, J. Prusazcyk, M. Dwyer and P. Kim (I974b) A stop-flow time-of-
flight mass spectrometer kinetics study. Reaction of ozone with nitrogen
dioxide and sulfur dioxide. J. Phys. Chem. 78, 1775-1779-
D.D. Davis and R. Schiff (1974); unpublished work quoted in D.D. Davis (I974a),
D.D. Davis, R. Schiff and S. Fischer (1974a); unpublished work quoted in
D.D. Davis (I974a).
D.D. Davis, G. Smith, and G. Klauber (I974d) Trace gas analysis of power
plant plumes via aircraft measurement: 03, NOX, S02 chemistry. Science
186, 733-736.
K.L. Demerjian, J.G. Calvert and D.L. Thorsell (1974) A kinetic study of the
chemistry of the S02(3B1) reactions with cis- and trans-2-butene. Int.
J. Chem. Kinet. 6, 829-848.
K.L. Demerjian, J.A. Kerr and J.G. Calvert (1974) The mechanism of photo-
chemical smog formation. Adv. Environ. Sci. Technol. 4, 1-262.
E.S. Domalski (1971) Thermochemical properties of peroxyacetyl (PAN) and
peroxybenzoyl nitrate (PEN) Environ. Sci. Technol. 5, 443-444.
R.¥. Fair and B.A. Thrush (1969) Reaction between hydrogen atoms and sulfur
dioxide. Trans Faraday Soc. 65, 1550-1570.
D.S. Fee, S.S. Markowitz and J.K. Garland (1972) Sulfur dioxide as a radical
scavenger in alkene systems. Radiochim. Acta 17, 135-138.
B.J. Finlayson, J.N. Pitts Jr. and R. Atkinson (1974) Low pressure gas
phase ozone-olefin reactions. Chemiluminescence, kinetics and mechanisms.
J. Amer. Chem. Soe. 96, 5356-5367.
S.N. Foner and R.L. Hudson (1962) Mass spectrometry of inorganic free radicals
Advan. Chem. Ser. 36, 34-49.
D.L. Fox and R.3. Wright (1977) Photochemical smog mechanisms-HC-NOx-S02
systems. Paper presented at the 4th International Clean Air Congress,
May 16-20, Tokyo, Japan; the authors are grateful for a preprint of this
paper.
132
-------
T. Frankiewicz and R.S. Berry (1972) Singlet molecular oxygen production
from photoexcited N02. Environ. Sci. Technol. 6, 365-366.
K. F. Freed (1976) Theory of collision induced intersystem crossing. J.
Chem. Phys. 64, I6o4-l6ll.
J.P. Friend, R. Leifer and M. Trichon (1973) On the formation of strato-
spheric aerosols. J. Atm. Sciences 30, 465-479.
O.L.J. Gijzeman (1974) Interaction between oxygen and aromatic molecules.
J. Chem. Soc. Faraday Trans. II 70, 1143-1152.
O.L.J. Gijzeman and F. Kaufman (1973) Oxygen quenching of aromatic triplet
states in solution. Part 2. J. Chem. Soc., Faraday Trans. I-I 69, 721-726,
O.L.J. Gijzeman, F. Kaufman and G. Porter (I973a) Oxygen quenching of
aromatic triplet states in solution. Part 1. J. Chem. Soc., Faraday
Trans. II 69, 708-720.
O.L.J. Gijzeman, Kaufman F., and Porter G. (I973b) Quenching of aromatic
triplet states in solution by nitric oxide and other free radicals. J.
Chem. Soc. Faraday Trans. II 69, 727-737.
N.V. Gillani, R.B. Jusar, J.D. Husar and D.E. Patterson (1977) Project
MISTT: Kinetics of particulate sulfur formation in a power plant plume
out to 300 km. Paper in International Symposium on "Sulfur in the
Atmosphere", Dubrovnik, Yugoslavia, Sept. 7-l4, 1977; the authors are
grateful for a preprint of this.
A. Good and J.C.J. Thynne (I967a) Reaction of free radicals with sulfur
dioxide. Part 1-Methyl radicals. Trans. Faraday Soc. 63, 2708-2719.
A. Good and J.C.J. Thynne (I967b) Reaction of free radicals with sulfur
dioxide. Part 2-Ethyl radicals. Trans. Faraday Soc. 63, 2720-2727.
S. Gordon and W.A. Mulac (1975) Reaction of the OH(X2jj_ ) radical produced
by the pulse radiolysis of water vapor. Int. J. Chem. Kinet. Symp. 1,
289-299.
T.E. Graedel (1976) Sulfur dioxide, sulfate aerosol, and urban ozone.
Geophys. Res. Letters 3> 181-184.
R. Greiner (1967) Hydroxyl-radical kinetics by kinetic spectroscopy. I.
Reactions with H2, CO, and CH4 at 300°K. J. Chem. Phys. 46, 2795-2799-
T.C. Hall (1953) Photochemical studies of nitrogen dioxide and sulfur
dioxide. Ph.D. Thesis, University of California, Los Angeles.
C.J. Halstead and D.R. Jenkins (1969) Sulfur dioxide-catalyzed recombination
of radicals in premixed fuel-rich hydrogen + oxygen + nitrogen flames.
Trans. Faraday Soc. 65, 3013-3022.
133
-------
E.J. Hamilton Jr. ''1975) Water vapor dependence of the kinetics of the
self-reaction of H02 in the gas phase. J. Chem. Phys. 63, 3682-3683.
E.J. Hamilton Jr. and C.A. Naleway (1976) Theoretical calculation of strong
complex formation by the H02 radical: H02 • H20 and H02 • NHs- J. Phys.
Chem. 80, 2037-2040.
E.F. Hampson and D. Garvin (1975) Chemical kinetic and photochemical data
for modeling atmospheric chemistry. U.S. National Bureau Standards
Technical Note 866, U.S. Government Printing Office, 1-113.
P.L. Hanst, E.R. Stephens, W.E. Scott and R.C. Doerr (1959) Atmospheric
ozone-olefin reactions. Paper presented before Division of Petroleum
Chemistry, 136th National Meeting American Chemical Society, Atlantic
City, Sept. 13, paper #28, p. 7R.
G.W. Harris and R.P. Wayne (1975) Reaction of hydroxyl radicals with NO,
N02 and S02. J. Chem. Soc. Faraday I 71, 610-617.
J. Heicklen (1976) Atmospheric Chemistry, Academic Press, New York, pp.
343-344.
D.G. Hendry and R.A. Kenley (1977) Generation of peroxy radicals from
peroxy nitrates (R02N02). Decomposition of peroxyacyl nitrates. J.
Amer. Chem. Soc. 99, 3198-3199-
C.J. Hochanadel, J.A. Ghormley and P.J. Ogren (1972) Absorption spectrum
and reaction kinetics of the H02 radical in the gas phase. J. Chem.
Phys. 56, 4426-4432.
A. Horowitz A. and J.G. Calvert (I972a) The S02-sensitized phosphorescence of
biacetyl in photolyses at 2650 and 2875A; the intersystem crossing ratio
in sulfur dioxide. Int. J. Chem. Kinet. 4, 175-189.
A. Horowitz and J.G. Calvert (1972b) A study of the intersystem crossing
reaction induced in gaseous sulfur dioxide molecules by collisions with
nitrogen and cyclohexane at 27°. Int. J. Chem. Kinet. 4, 191-205.
A. Horowitz and J.G. Calvert (1972c) Emission studies of the mechanism of
gaseous biacetyl photolysis at 3450, 3650, 3880 and 4380A and 28°C.
Int. J. Chem. Kinet. 4, 207-227.
0. Hov, I.S.A. Isaken and E. Hesstvedt (1977) Diurnal variations of ozone
and other pollutants in an urban area. Report No. 24, Institutt for
Geofysikk, Universitetet i Oslo, Jan., 1977.
L.A. Hull, I.C. Hisatsune and J. Heicklen (1972) Low-temperature infrared
studies of simple alkene-ozone reactions. J. Amer. Chem. Soc. 94,
4856-4864.
134
-------
G.E. Jackson and J.G. Calvert (1971) The triplet sulfur dioxide-carbon
monoxide reaction exited within the S02(1A1) -*- S02(3B1) "forbidden"
band. J. Amer. Chem. Soc. 93, 2593-2599.
A. Jacob and C.A. Winkler (1972) Kinetics of the reaction of oxygen atoms
and nitrogen atoms with sulfur trioxide. J. Chem. Soc. Faraday Trans. I
68, 2077-2082. —
F.C. James, J.A. Kerr and J.P. Simons (1973) Direct measurement of the rate
of reaction of the methyl radical with sulfur dioxide. J. Chem. Soc.
Faraday Trans. I 69, 212^-2129.
I.T.I. Jones and K.D. Bayes (1971) Energy transfer from electronically
excited W02 . Ghem Phys. Letters 11, 163 -166.
C.E. Junge (1972) The cycle of atmospheric gases-natural and man made.
Quart. J. Roy. Meteor. Soc. 98, 711-729.
M. Kasahara and K. Takahashi (1976) Experimental studies on aerosol
particle formation by sulfur dioxide. Atm. Environ. 10,
K. Kawaoka, A.U. Khan and D.R. Kearns (1967) Role of singlet excited
states of molecular oxygen in the quenching of organic triplet states.
J. Chem. Phys. k6, 181*2-1853.
K. Kear and E.W. Abrahamson (197V 75) Electronic energy transfer in the gas
phase: the quenching of 02(3-Zg+). J. Photochem. 3, ^09-^16.
D.R. Kearns (1971) Physical and chemical properties of singlet molecular
oxygen. Chem. Rev. 71, 395-^27.
N. Kelly, J.F. Meagher and J. Heicklen (1976/77) The photolysis of sulfur
dioxide in the presence of foreign gases. VIII. Excitation of S02 at
3600-^lOOA in the presence of acetylene. J. Photochem. 6, 157-172.
J.A. Kerr and M.J. Parsonage (1972) Evaluated kinetic data on gas phase
reactions . Butter-worths, University of Birmingham, England.
R. Kigel and H. Taube (1975) Infrared spectrum and structure of matrix-
isolated sulfur tetroxide. J. Phys. Chem. 79, 2130-2135-
W.A. Rummer, J.W. Pitts Jr., and R.P. STeer (1971) The cherai luminescent
reaction of ozone with olefins and organic sulfides. Envir. Sci.
Technol. 5,
A. Levy, D.R. Drewes and J.M. Hales (1976) S02 oxidation in plumes: a
review and assessment of relevant mechanistic and rate studies. Report
EPA-Jf50/3-76-022 prepared for the U.S. Environmental Protection Agency,
Office of Air Quality Planning and Standards, Contract No. 68-02-1982
by Battelle Pacific Northwest Laboratories, Richland, Washington 19352.
135
-------
A.C. Lloyd 1974) Evaluated and estimated kinetic data for gas phase
reactions of the hydroperoxyl radical. Int. J. Chem. Kinet. 6, 169-228.
R. Louw, J. vanHam and H. Nieboer (1973) Nitrogen trioxide: key intermediate
in the chemistry of polluted air? J. Mr Pollut. Control Ass. 23, 716.
R. McAndrew and R. Wheeler (1962) The recombination of atomic hydrogen in
propane flame gases. J. Phys. Chem. 66, 229-232.
D.H. McNelis (197^) Aerosol formation from gas-phase reactions of ozone and
olefin in the presence of sulfur dioxide. Final report 650/4-7^-034
to the Environmental Protection Agency, Research Triangle Park, Worth
Carolina, for program element 21 MB, ROAP 38, August, 1974.
A.J. Merer (1963) Rotational analysis of bands of the 3800A system of S05.
Discussions Faraday Soc. 35, 127-136.
H.D. Mettee (1969) Foreign gas quenching of sulfur dioxide vapor emission.
J. Chem. Phys. 73, 1071-1076.
D.F. Miller (1977) Simulations of gas-phase reactions in power plant plumes.
Preprint of a paper presented at 173rd National Meeting Amer. Chem. Soc.,
New Orleans, March, 1977-
R. Milstein, R.L. Williams and F.S. Rowland (1974) Relative reaction rates
involving thermal fluorine-18 atoms and thermal fluoroethyl radicals
with oxygen, nitric oxide, sulfur dioxide, nitrogen., carbon monoxide
and hydrogen iodide. J. Phys. Chem. 78, 857-863.
M.F.R. Mulcahy, J.R. Steven and J.C. Ward (1967) The kinetics of reaction
between oxygen atoms and sulfur dioxide; an investigation by electron
spin resonance spectrometry. J. Phys. Chem. 71» 2124-2131.
M.F.R. Mulcahy and R.H. Smith (1971) Reactions of OH radicals in the H-N02
and H-N02-CO systems. J. Chem. Phys. 54, 5215-5221.
H. Niki (1977) Private communication to one of the authors.
H. Niki, P.D. Maker, C.M. Savage and L.P. Breitenbach (1975) Fourier trans-
form spectroscopic studies of organic species participating in photo-
chemical smog formation. Paper at International Conference on Environ-
mental Sensing and Assessment, Las Vegas, Nevada, Sept. l4, 1975, Vol.
2, 24-4.
H. Niki, P.D. Maker, C.M. Savage and L.P. Breitenbach (1977) Fourier
transform IR spectroscopic observations of propylene ozonide in the gas
phase reaction of ozone-cis-butene-formaldehyde. Chem. Phys. Letters
46, 327-330.
S. Okuda, T.N. Rao, D.H. Slater and J.G. Calvert (1969) Identification of
the photochemically active species in sulfur dioxide photolysis within
the first allowed absorption band. J. Phys. Chem. 73, 44l2-44l5.
136
-------
H.E. O'Neal and C. Blumstein (1973) A new mechanism for gas phase ozone-
olefin reactions. Int. J. Chem Kinet. 5, 397-413.
R. Overend, G. Paraskevopoulos and R.J. Cvetanovic (1974) Hydroxyl radical
rate measurement for simple species by flash photolysis kinetic spectro-
scopy. Paper 6-4, Abstracts llth'Informal Conference of Photochemistry,
Vanderbilt University, Nashville, Tennessee, 248-252.
G. Paraskevopoulos (1976) Private communication to one of the authors.
D.A. Parkes (1974) The roles of alkylperoxy and alkoxy radicals in alkyl
radical oxidation at room temperature. Paper presented at the 15th
International Symposium on Combustion, Japan, 1974.
D.A. Parkes, D.M. Paul, C.P. Quinn and R.C. Robson (1973) The ultraviolet
absorption by alkylperoxy radicals and their mutual reactions. Chem.
Phys. Letters 23, 425-429.
C.T. Pate, R. Atkinson and J.N. Pitts, Jr. (1976) Rate constants for the
gas phase reaction of peroxyacetyl nitrate with selected atmospheric
constituents. J. Environ. Sci. Health--Environ. Sci. Eng. All, 19-31-
L.K. Patterson, G. Porter and M.R. Topp (1970) Oxygen quenching of singlet
and triplet states. Chem. Phys. Letters 7, 712-6l4.
T.T. Paukert and H.S. Johnston (1972) Spectra and kinetics of the hydro-
peroxyl free radical in the gas phase. J. Chem. Phys. 56, 2824-2838.
W.A. Payne, L.J. Stief and D.D. Davis (1973) A kinetics study of the
reaction of H02 with S02 and NO. J. Amer. Chem. Soc. 95, 76l4-76l9.
R.D. Penzhorn, V.G. Filby, H. Gusten (I974a) Die photochemische abbaurate
des schwefeldioxids in der unteren atmosphare mitteleuropas. Z.
Naturforsch. 29a, 1449-1453-
R.D. Penzhorn, W.G. Filby, K. Gunther and L. Stieglitz (1975) The photo-
reaction of sulfur dioxide with hydrocarbons. II. Chemical and physical
aspects of the formation of aerosols with butane. Int. J. Chem. Kinet.
Symp. 1, 611-627.
R.D. Penzhorn, H. Gusten, U. Schurath and K.H. Becker (1974b) Quenching of
singlet molecular oxygen by some atmospheric pollutants. Current Res.
10, 907-909.
J.T. Peterson (1976) Calculated actinic fluxes (290-700 nm) for air
pollution photochemical applications. U.S. Environmental Protection
Agency Report 600/4-76-025, p. 21.
J.N. Pitts, Jr., B.J. Finlayson, H. Akimoto, W.A. Kummer and R.P. Steer
(1972) The chemiluminescent reactions of ozone with olefins and organic
sulfides. Advan. Chem. Ser. 113, 246-254.
137
-------
--..I:. r.udoliDh and S.J. Strickler (1977) Direct measurement of the lifetimes
of the BL state of S02 in air at atmospheric pressure. J. Amer. Chem.
Soc. 99, 3871-3872.
S.P. Sander and J.H. Seinfeld (1976) Chemical kinetics of homogeneous atmo-
spheric oxidation of sulfur dioxide. Environ. Sci. Technol. 10, 111^-1123.
K. Schofield (1973) Evaluated chemical rate constants for various gas phase
reactions. J. Phys. Chem. Ref. Data 2, 25-77.
S.E. Schwartz and L. Newman (1977) Processes limiting the oxidation of
sulfur dioxide in stack plumes. Manuscript of a paper to be submitted
for publication.
H.W. Sidebottom, C.C. Badcock, J.G. Calvert, B.R. Rabe and E.K. Damon (1971)
Mechanism of the photolysis of mixtures of sulfur dioxide with olefin
and aromatic hydrocarbons. J. Amer. Chem. Soc. 93S 3121-3128.
H.W. Sidebottom, C.C. Badcock, J.G. Calvert, B.R. Rabe and E.K. Damon
(1972a) Lifetime studies of the biacetyl excited singlet and triplet
states in the gas phase at 25°. J. Amer. Chem. Soc. 9^-j 13-19-
H.W. Sidebottom, C.C. Badcock, G.E. Jackson, J.G. Calvert, G.W. Reinhardt,
and E.K. Damon (1972b) Photooxidation of sulfur dioxide. Environ. Sci.
Technol. 6, 72-79.
B.K.T. Sie, R. Simonaitis and J. Heicklen (1976) The reaction of OH with CO.
Int. J. Chem. Kinet. 8, 85-98.
P.A. Skotnicki, A.G. Hopkins and C.W. Brown (1975) Time dependence of the
quantum yields for the photooxidation of sulfur dioxide. J. Phys. Chem.
79, 2450-2452.
I.W.M. Smith and R. Zellner (1973) Rate measurements of reactions of OH by
resonance absorption. Part 2. Reactions of OH with CO, C2H4, and C2H2
J. Chem. Soc. Faraday Trans. II 69, 1617-1627.
E.R. Stephens and M.A. Price (1972) Comparison of synthetic and smog
aerosols. J. Colloid Interface Sci. 39, 272-286.
F. Stuhl and H. Niki (1972) Pulsed vacuum-uv photochemical study of OH with
H2, 02, and CO using a resonance-fluorescence detection method. J. Chem.
Phys. 57, 3671-3677.
F. Su and J.G. Calvert (1977&) The mechanism of the photochemical reactions
of S02 with C2E2 and CO excited within the S02(3B1) ^-S02(xxA1)
"forbidden" band. Chem. Phys. Letters, in press.
F. Su and J.G. Calvert (I977b), work in preparation for publication.
138
-------
F. Su, J.W. Bottenheim, H.¥. Sidebottom, J.G. Calvert and E.K. Damon (1977b)
Kinetics of fluorescence decay of S02 excited in the 2662-3273A region.
Int. J. Chem. Kinet., in press.
F. Su, J.W. Bottenheim, D.L. Thorsell, J.G. Calvert and E.K. Damon (I977a)
The efficiency of the phosphorescence decay of the isolated S02(3B1)
molecule. Chem. Phys. Letters, in press.
F. Su, R.B. Wampler, J.W. Bottenheim, D.L. Thorseil, J.G. Calvert and E.K.
Damon (I97?c) On the pressure saturation effect of the quenching of
S02(3BX) molecules. Chem. Phys. Letters, in press.
T.A. Walter, J.J. Bufalini and B.W. Gay, Jr. (1977) Mechanism for olefin-
ozone reaction. Environ. Sci. Technol. 11, 382-386.
F.B. Wampler, J.G. Calvert and E.K. Damon (1973a) A study of the bimolecular
intersystem crossing reaction induced in the first excited singlet of
S02 by collsions with 02 and other atmospheric gases. Int. J. Chem.
Kinet. 5, 107-117.
F.B. Wampler, A. Horowitz and J.G. Calvert (1972) The mechanism of carbon
dioxide formation in 3130A-irradiated mixtures of sulfur dioxide and
carbon monoxide. J. Amer. Chem. Soc. 9^, 5523-5532.
F.B. Wampler, K. Otsuka, J.G. Calvert and E.K. Damon (1973b) The temp-
erature dependence and the mechanism of S02(3BjJ quenching reactions.
Int. J. Chem. Kinet. 5, 669-687.
A.A. Westenberg and N. deHaas (1973) Rates of CO + OH and H2 + OH over an
extended temperature range. J. Chem. Phys. 58, 4o6l-4o65.
A.A. Westenberg and N. deHaas (I975a) Rate of the reaction 0 + S02 + M
S03 + M. J. Chem. Phys. 63, 5^11-5^15•
A.A. Westenberg and N. deHaas (I975b) Rate of the 0 + S03 reaction. J. Chem.
Phys. 62, 725-730.
M.R. Whitbeck, J.W. Bottenheim, S.Z. Levine and J.G. Calvert (1976) A
kinetic study of the CH302 and (CH3)3C02 radical reactions by kinetic
flash spectroscopy. Abstracts of 12th Informal Conference on Photo-
chemistry, U.S. National Bureau of Standard, Gaithersberg, Maryland,
July, 1976, pp Kl-1 to Kl-5-
G.Z. Whitten and H.H. Hogo (1976) Mathematical modeling of simulated photo-
chemical smog. Final Report EF 76-126 by Systems Application Inc., to
the U.S. Environmental Protection Agency, August, 1976.
W.E. Wilson, R.J. Charlson, R.B. Husar, K.T. Whitby and D. Blumenthal (1976)
Sulfates in the atmosphere. Paper presented at 69th meeting of the Air
Pollut. Control Ass., Portland, Oregon, June, 1976.
139
-------
W.E. Wilson, M.C. Dodge, D.H. McNeils and J. Overton (197^) S02 oxidation
mechanism in olefin-NOx-3Gp smog. Paper presented before Division of
nvironmental Chemistry, American Chemical Society, Los Angeles, April,
W.E. Wilson, A. Levy and D.E. Wimmer (1972) A study of sulfur dioxide in
photochemical smog. II. Effect of sulfur dioxide on oxidant formation
in photochemical smog. J. Air Pollut. Control Ass. 22, 27-32.
W.P. Wood, A.W. Castleman, Jr., and I.N. Tang (197^) Mechanisms of aerosol
formation from S02. Paper presented at 67th annual meeting Air Pollut.
Control Ass., Denver, Colorado, June 9-13, 197^-
W.P. Wood, A.W. Castleman, Jr., and I.N. Tang (1975) Mechanisms of aerosol
formation from S02. J. Aerosol Sci. 6, 367-375.
-------
SECTION 3
STUDIES RELATED TO FOEMALDEHYDE REMOVAL MECHANISM IK THE ATMOSPHERE
QUANTUM EFFICIENCY OF THE PRIMARY PROCESSES IN CHP0 PHOTOLYSIS AT 3130A
AND 25°C
Introduction
Efforts to elucidate the mechanism of formaldehyde photolysis have
increased greatly in recent years. The renewed interest in this system
has been stimulated largely by two factors: (l) the high potential of
formaldehyde for successful laser-photolysis, isotope separation; and (2)
the need for the quantitative evaluation of the significance of CI^O reactions
in smog formation.1 In this study our attention was focussed on the second
area of interest. The extensive efforts of many research groups have defined
many features of the formaldehyde photolysis mechanism,2"34 but several
qualitative and quantitative aspects are not fully understood today. It is
•well established that primary reactions 1 and 2 occur when formaldehyde is
excited to the first excited singlet state.
+ hv -*- H + HCO (1)
CEfeO + hv •+- E2 + CO (2)
Uncertainty remains about the detailed reaction paths through which these
processes occur. McQuigg and Calvert15 rationalized the wavelength dependence
of b-i/^o 3-n "terms of a simple kinetic model which implied the occurrence of
internal conversion of the original excited state to a vibrationally excited
ground state reactant common to both decomposition modes „ Some recent more
quantitative experimental and theoretical studies also favor this interpreta-
tion;20?22 ? 3°,32~34 others show the possible involvement of the lowest
triplet excited state in the process 1.2 0,22, 32 -34 pjj. recent studies in-
dicate that the dominant final products of formaldehyde photolysis are H2
and CO, and their quantum yields suggest that ^ + 4>2 may be near unity
at most wavelengths within the first absorption band (1Ap •«- x^-A, ). The
effects of radical scavengers such as iodine,3 alkenes,1^ oxygen,37 and
nitric oxide24'28 on these quantum yields, the relative chemiluminescence
of excited HNO formed by H-atom capture by NO following process 1 in CH20-
NO mixture photolyses,29 and isotopic scrambling experiments with CH20,
CD20, CHDO,8s15>^J31 have all been used in the estimation of the ^/^
ratio as a function of wavelength. It is generally accepted today that the
importance of the free radical mode of decomposition, reaction 1, increases
with decreasing wavelength used to excite the CH20. However, the absolute
values reported for the primary quantum yields are widely divergent. Thus
at 3130 A previous estimates of *2 vary from 0.013 to O.5.15'27 The apparent
disagreement between the quantum yields measured in the different studies
may have resulted at least in part from different conditions employed, i.e.,
-------
temperature, light intensity, etc., as well as unrecognized complexities
in the reaction pathways in many of the CHaOradical scavenger studies and
the subsequent misinterpretation of the results.
We initiated our present work as part of an effort to estimate the
importance of CH20 photolysis as a source of H-atoms in the atmosphere. We
have carried out formaldehyde photolyses in the presence of several different,
seemingly simple, H-atom and HCO-radical scavengers in order to define the
best system for the accurate determination of 4> and . The varied results
from the several systems which we have studied in this work provide a new
and reliable set o£ self-consistent estimates of , and
-------
the highest pressure employed (12 Torr) the pressure drop as a result of
polymerization was negligible during the short irradiation tijnes made
possible by the use of the high intensity source. When other gases were
added to formaldehyde prior to irradiation, the mixture was allowed to
stand for 20 min to insure complete mixing. After photolysis, the gases
were pulled through a glass bead-containing trap immersed in N2(l) and
separated from CH20 using a Toepler pump. The non-condensible gases H2 and
CO were transferred to a calibrated volume and the total pressure of the
mixture measured. ES and CO were determined using a gas chromatograph
(Varian 2?00) equipped with a TC detector and l/8th in x 12 ft column filled
with NO-treated Molecular Sieve 5A held at 100°C. Nitrogen was used as the
carrier gas when the concentrations were high, while at low CO concentrations,
helium was used. H2 and CO were the onl^r non-condensible products (except
for experiments with added NO), and therefore one of them, could be directly
determined and other by difference from the known total pressure. The
results obtained with both direct product measurements and those by pressure
change checked well with one another.
Results and Discussion
The Photolysis of Pure Formaldehyde --
The photolysis of formaldehyde at 3130A was studied in a series of ex-
periments at room temperature (25 ± 2°C) and over a range of pressures (1-12
Torr) and light intensities. The data, summarized in Table 19, confirm the
previous observation of near equality of the quantum yields of carbon monoxide
and hydrogen products; these results give $ /$ '= 1.01 ± 0.09(2cr). Further-
00 n2
more the ratio of twice the pressure change (Ap) during the course of the p
photolysis to the pressure (A) of the recovered noncondensable gases (H2 and
CO) is unity within the experimental error; the data of Table 19 give 2Ap/A =
1.00 ± 0.08(2). Thus we conclude that H2 and CO are the only significant
products of formaldehyde photolysis and the observed rate of pressure in-
crease during an experiment is equal to the rate of CH20 decomposition. Then
for formaldehyde photolyses at low pressures for which the limiting form of
Beer's law applies well, we anticipate that the total pressure p_ in the system
with the initial formaldehyde pressure p0 should be described by the expres-
sion I:
la [(2£o - p)/p_0] = -kt (I)
Indeed the results of an experiment at 2.49 Torr of CH20, plotted in Figure
44, show that this simple first order rate law describes the data well, at
least over the first 30% of the reaction which was followed here.
In one series of experiments at 2.50 Torr of CH20, the rate of pressure
increase was monitored using different incident light intensities. These
data, shown in Figure 45, indicate that the rate of CH20 decomposition is
directly proportional to the iaflicent light intensity, so that for our con-
ditions the quantum yield of formaldehyde decomposition is intensity in-
dependent. This result is consistent with that observed by Sperling and
Toby19 in experiments at low intensities but differs from what they observed
at the higher temperatures (80-120°C), higher pressures (up to 40 Torr),
and at the highest intensities used in their experiments.
143
-------
TABLE
Ik. EXPERIMENTAL DATA FROM THE
PCH20' T°rr
Initial Average
1.044
1.033
1.037
1.036
1.036
1.035
1.006
1.007
1.007
3.240
3.236
3.249
3.244
3.346
3.238
3.158
3.157
3.156
3.173
3.187
3.153
8.268
8.400
8.303
8.303
8.030
8.438
8.072
8.008
8.062
8.062
8.060
12.10
12.12
1.015
1.011
1.008
1.007
1.002
1.004
0.975
0.976
0.975
3.135
3.133
3.158
3.156
3.256
3.161
3.079
3.082
3.082
3.090
3.094
3.060
8.169
8.302
8.221
8.193
7.924
8.311
7.960
7.895
7.954
7.949
7.952
11.95
11.97
Irradia-
tion time
sec
900
900
900
900
900
900
1000
1000
1000
900
900
900
900
900
900
900
900
900
1000
1000
1000
450
450
450
600
600
700
600
600
600
600
600
600
600
PHOTOLYSIS OF FORMALDEHYDE AT 3130A AND 25 °C
Quanta abs. 2Ap, Torr A, Torr 2Ap/A
, /cell x KT18 x 103b x 103c
(0.
(0.
(0.
(0.
(1.
(1.
(0.
(1.
1.
3.
3.
2.
2.
2.
2.
2 .
2.
2.
2.
2.
2.
2.
3.
2.
3.
3.
4.
3.
3.
3.
3.
3.
4.
4.
92)
91)
91)
91)
13)
13)
99)
01)
00
21
21
73
73
81
44
38
39
39
63
58
86
96
00
53
23
34
07
59
42
39
49
23
71
72
118
127
118
118
137
125
125
123
130
390
404
353
361
373
313
330
298
296
328
376
375
396
404
344
460
403
499
445
449
435
456
417
572
568
117
120
116
121
137
141
125
124
129
419
413
364
361
350
310
316
302
298
330
372
371
396
391
328
442
423
510
448
449
434
451
420
601
596
1.01
1.06
1.02
0.98
1.00
0.89
1.00
0.99
1.01
0.93
0.98
0.97
1.00
1.07
1.01
1.04
0.99
0.99
0.99
1.01
1.01
1.00
1.03
1.05
1.04
0.95
0.98
0.99
1.00
1.00
1.01
0.99
0.95
0.95
x 103d
59.8
59.7
60.1(56.
64.8(55.
71.3
70.9
62.3
59.8
64.9
200
(211)
176 (179)
176 (195)
167 (182)
152 (159)
150 (166)
146
144
(164)
167
(183)
199
195
161 (158)
215 (208)
207
247
227
226
215
(225)
(200)
307
305
5) 59
9) 65
202
185
167
169
152
150
167
189
170
234
226
221
x 103d
(56.8)
(60.3)
.7(56.1)
.1(56.2)
(66.0)
(70.2)
(62.5)
(64.4)
(64.0)
(219)
(187)
(185)
(183)
(159)
(165)
(156)
(154)
(185)
(197)
(196)
(167)
(226)
(197)
(262)
(222)
(223)
(219)
(294)
(291)
s
a
a
a
a
a
a
a
a
1.10
1.06
1.12
1.10
1.15
1.05
i.oa
1.13
1.04
1.02
1.05
1.23
1.08
1.14
1.10
1.07
1.11
1.05
1.03
1.07
1.12
1.08
1.09
1.05
1.10
1.10
*CO
a
a
a
a
a
a
a
a
1.09
1.16
1.07
1.15
1.09
1.06
1.08
1.12
1.11
1.10
1.08
1.22
1.12
1.13
1.11
1.13
1.21
1.00
1.09
1.05
1.11
1.10
1.10
1.16
1.05
1.05
*CO/*H2
0.95
1.01
0.99
1.01
0.93
0.99
1.00
1.08
0.99
1.10
0.96
1.05
0.95
1.01
1.00
1.00
1.07
1.07
1.02
0.99
1.03
0.99
1.01
1.05
1.09
0.95
1.06
0.98
0.99
1.02
1.00
1.11
0.96
0.95
aln these experiments e could not be determined accurately, and quantum yields are not reliable. Ap is the
pressure change in the cell. CA is the pressure of noncondensables. Yield by pressure difference (parentheses).
-------
Time, sec,
Figure kk. First-order plot of a function of the
pressure in formaldehyde photodecomposition at
3130 i, p0 = 2.U88 Torr, temperature 25°C.
,
0.5
1.0
Relative absorbed light intensity
Figure 45- Effect of incident light intensity on
the rate of formaldehyde photodecomposition at
3130 A. PnTT n = 2.50 Torr, temperature 25°C.
-------
Consider the present kinetic results in terms of the following simple
reaction mechanism:
CH20 + hv -*- H + HCO (1)
CH20 + hv -*- He + CO (2)
H + CH20 -*• H2 + HCO (3)
HCO + HCO •*- CH20 + CO (4)
HCO + HCO -*- H2 + 2CO (5)
HCO + HCO -*- (HCO)2 (6)
HCO + M -*- H + CO + M (7)
H + HCO ->- H2 + CO (8)
Since the present results show $CO/$H is unity within the experimental error,
some restriction on the relative importance of these reactions is imposed.
The occurrence of reaction 6 would lead to / > 1, and a different
M2 LO
pressure change during photolysis would be observed here. Thus it is likely
that kx- < (k + k. ), and reaction 6 can be neglected here. This conclusion
is in accord with the findings of Khan, Mbrrish, and Porter6 who also found
no evidence of glyoxal formation following the flash photolysis of CHsCHO.
Kinetic considerations show that reactions 7 and 8 are also unimportant
here. Thus the present best estimates related to the thermochemistry of
reaction 7 give AH ° ^ 17.5 kcal mole'^Ref. 28,29,31,37-40] and AS ° ~
20.9 eu (25°C, 1 atm);41 from these data we may estimate k = k /k s 2.3 x
10~10 mole I"1 at 25°C. Taking the measured value of k = (k.O ± 0.6) x
107 t2 mole"2 sec"1 for M = H2 at 25°C,4a we estimate k & 9 x 10~3^ mole"1
sec-1 at 25°C. From these data, the measured value of k. + k = 2.2 x 101Q
$. mole'1 sec"1 [Ref. ^3] > the measured absorbed light intensity in the ex-
periments at 12 Torr of CH20, and the steady state assumption for [H] and
[HCO], we calculate that the rate of HCO radical loss by dissociation in 7
to that in reactions h and 5 will be negligible: Ry/2(R, + R ) s 2 x 10~7.
Furthermore taking k = 2.8 x 10Y jfmole"1 sec"1, an average of the recent
determinations,44?46 and a reasonable estimate of kn = 2 x 1010 jfmole"1
sec"1, the ratio of the rate of HCO loss in 8 compared to that in k and 5 is
also very small: R^/2(R, + R ) s 3 x 10~4. Thus for our conditions, 1-5
appear to be the only reactions of H and HCO which will occur at significant
rates; reactions 6,7, and 8 may be neglected in our further considerations
here.
-------
The results of Table 1 give the average values: $ = 1.09 ± 0.09(2cr)
and $H = 1.10 ± 0.09. Clark has concluded, from the results of his recent
study2 that $co = $H = 1.0 at wavelengths near 3130 A employed here, but
the experimental error in his determinations is somewhat higher than that
of the present work. Although our estimates of $ and 0 include unity
H2 CO
within the 2cr error limits, the results of the CH20-Me3SiH experiments sug-
gest (See the next section.) that the true values are somewhat larger than
unity as our average estimates do indicate. Formation of the photolysis
products hydrogen and carbon monoxide in equal yields is consistent with
the occurrence of both reactions k and 5. However in view of the unim-
portance of reaction 7 and chain reactions for our conditions, values of
$ and $co greater than unity suggest that reaction 5 as well as k must
occur to some extent. If ^ + $2 = 1.0, as seems likely from other results
in this study, then from the measured values of $ = $ = 1.10 and the
HP CO
reaction sequence 1-5, we anticipate in theory that 0, = 0.10(k, + k )/k .
Glicker and Stief16 assumed that reaction 5 was the only fate of HCO
radicals in their short wavelength CH20 photolysis studies; i.e., k^/(k, +
kj.) = 1.0. If this were the case here, then our results would require
$.. = 0.10 and = 0.90. This conclusion is inconsistent with the results
of all other studies of CH20 photolysis at the 3130 A region where estimated
values of $2 = 0.01,3 O.I?,8 0.48,14 0.5,15 0.2,19 O.k^,27 0.58 (at 3172 A)28
and 0.37 (at 31^0 A),29 and it is not considered realistic. Only if our
actinometry were seriously in error and the actual 0 and <3? values were
112 ^U
much lower than we estimated could the hypothesis of Glicker and Stief lead
to consistent $p values. It appears more likely to us that reaction 5 is
not the sole fate of HCO radicals as they suggested, and in view of Clark's
results and conclusions,28 there must be some question as to whether reaction
5 occurs at all. In order to assess better the relative importance of
reactions k and 5 and to define well the magnitude of the primary quantum
yield sum, $ + 3> , we have carried out CI^O photolyses in the presence of
trimethylsilane.
Photolysis of Formaldehyde-Trimethylsilane Mixtures --
We reasoned that trimethylsilane would be an excellent trap for H-atoms
and HCO radicals through the reactions 9 and 10, and conceivably the occur-
rence of the usual reactions 3, k, and 5 of pure formaldehyde photolysis
could be suppressed completely.
H + Me3SiH ->- Ha + Me3Si (9)
HCO + Me3SiH -*• CH20 + Me3Si
Me3Si + CH2C -*• Me3SiH + HCO
-------
2Me3Si -*- Me3SiSiMe3 (12)
However abstraction of H-atoms from CH20 by Me3Si radicals in reaction 11
may occur to some extent since D(Me3Si-H) = 89 ± k [Eef . 4?] while D(HCO-H)
= 86.3 ± 1.5 kcal mole-1 [Refs. 28, 29 ,31, 37-^0] . In Table 20 the results
of CH20-Me3SiH photolyses are summarized. Note that with added trimethylsilane
the quantum yield of H2 is lowered from the value of 1.10 characteristic of
pure formaldehyde to unity within the experimental error [$„ = 1.00 ± 0.08
j±2
(2cr)], even at the lowest concentration of the silane added here. This is
consistent with the occurrence of primary processes 1 and 2 with $, + $2 =
1.0, followed by reactions 9, 10, and 12 alone. Presumably the extent of
lowering of $TT reflects the contribution of hydrogen formation from reaction
H2
5 which cannot occur under these conditions. These results support the
"high" values of <1> and 0 found here in the pure CHgO runs, and place
OU rl2
the quantum yield of H2 formed by way of reaction 5 at about 0.10 for pure
formaldehyde photolyses under our conditions.
However not all of the results of the CH20-Me3SiH mixture photolyses
are explicable in terms of the simple mechanism outlined above. The <3>
CO
data and the pressure changes observed in the data of Table 20 cannot be ex-
plained using reactions 1-5 and 9-12 alone. If only these reactions occur-
red in this system then one would expect = $0 and $ - $ = $ ; the
L,(J c- ii2 (_/U -L
change in pressure (Ap ) observed in the system should be given by Ap = Pnr.j
ou
since no pressure change should result from the occurrence of reaction 1
followed by 95 10, and 12. It can be seen that the measured Ap in column
5 of Table 20 is not equal to P in column 8; the observed pressure change
\j\j
is always smaller than that anticipated from the occurrence of only reactions
1, 2, and 9-12. Furthermore the imbalance of Ap and P increases with
'• L/U
increasing trimethylsilane concentration. In rationalizing these results
additional reactions must be considered. The reaction sequence 1, 2, 9-12
together with the additional reactions 13-19 can qualitatively explain these
results :
Me3Si + CHeO ->- Me3SiOCH2 (13)
Me3SiOCH2 + CH20 -*- Me3SiOCH3 + HCO (l^)
Me3SiOCH2 + Me3SiH Me3SiOCHs + Me3Si (15)
Me3Si + HCO ->- Me3SiH + CO (l6)
Me3SiOCH2 + HCO ->- Me3SiOCH3 + CO (17)
2Me3SiOCH2 -*- Me3SiOCH2CH2OSiMe3 (18)
Me3Si + Me3SiOCH2 -*- Me3SiCH2OSiMe3 (19)
-------
TABLE 20. EXPERIMENTAL DATA FROM THE PHOTOLYSIS OF FORMALDEHYDE IN THE PRESENCE
OF TRIMETHYLSIIANEa
VQ
PCH20
Initial
1.033
1.033
1.037
3.252
3.205
3.228
3.234
3.033
3.187
3.187
8.263
8.292
, Torr
Average
1.010
1.006
1.012
3.175
3.137
3.148
3.165
2.958
3.006
3.058
8.239
8.241
pMe3SiH.
Torr
0.504
1.257
5.225
0.487
1.143
1.249
4.730
7.266
8.668
28.5
4.275
4.422
Quanta
/cell
0.
0.
0.
2.
2.
2.
2.
2.
2.
2.
3.
3.
abs.
x 10-18
91
91
91
71
43
69
33
42
57
50
59
59
Ap, Torr
x 103c
29
16
-22
85
21
57
0
50
—
_-
33
-70
A , Torr
x 103d
98.5
104.4
94.5
282
247
283
245
263
267
260
356
331
X SB
55.1
54.7
54.4
154
135
159
139
151
166
151
199
204
; "co-Io^
39.8
39.9
40.1
139
95.2
122
94.1
113
105
109
173
142
H2
1.02b
1.02b
1.01b
0.96
0.95
1.01
1.01
1.06
1.07
1.03
0.94
0.96
*CO
0.74b
0.75b
0.74b
0.87
0.67
0.77
0.69
0.79
0.68
0.74
0.82
0.67
*CO/*H2
0.72
0.73
0.73
0.91
0.70
0.76
0.69
0.75
0.63
0.72
0.87
0.70
aExcitation at 3130 A, temperature 25°C. In experiments at 1 Torr pressure of Q^O, tp, and Q-, were estimated
relative to pure CH^O photolysis with *Q-,
#
1.10. °Ap is the pressure change in the cell. A is the
pressure of noncondensables.
ations.
Except for results marked with an asterisk, all values are from direct determin-
-------
The occurrence of reaction 13 is in accord with the smaller pressure in-
creases, and the pressure decreases, which are observed in some of the ex-
periments at high silane concentrations. Reaction 13, the addition of
Me3Si radicals to CH2Oj is not unprecedented, although we had not anticipated
this happening. The addition reactions of silyl radicals to carbonyl com-
pounds have been observed previously;43 they seem to occur readily,
presumably as a result of the high oxygen affinity of the silyl radicals.
Consider the effect of trimethylsilane addition to CI^O on the CO
formation in terms of this more complete mechanism. When Me3SiH is added to
CH20, reaction 9 takes over from reaction 3 as the main sink for H-atoms,
and 10 presumably may compete with the bimolecular HCO loss reactions k and
5. However since HCO radicals can still be formed in reactions 11 and Ik,
this effect need not necessarily lead to a decrease in $.-,.-• The results of
0\J
Table 20 show that the addition of even a small amount of Me3SiH causes a
marked decrease in $,,,0 and the ratio <1> / approaches a constant value
LO L,(J Jig
of about 0.7. Since is practically constant at unity for these conditions,
%
this result taken alone appears to favor $^ s 0.7. However an alternative
interpretation is possible, and indeed it is likely. In the presence of
MesSiH, CO is probably formed by reactions 2, 16, and 17 largely, since the
radical concentrations [Me3Si] and [Me3SiOCH2] are probably much larger than
[HCO]. Hence reactions 4 and 5 a^e probably much less important. Thus for
these conditions the seemingly constant value of observed at high
CO
[Me3SiH] may be a consequence of reasonably constant rates of 1, 16, and 17
which are achieved for these conditions. In this case the data require that
$P < $ =0.7. Obviously the mechanism of CO formation in the CH20-Me3SiH
system is very much more complex than we had anticipated, and we must resort
to other less complicated kinetic systems to estimate §p accurately. The
isobutene-CH20 system seems to fulfill our needs.
Photolysis of Formaldehyde- 1 sobutene Mixtures —
The complete reaction sequence, in addition to reactions 1-5, which
we would expect to describe the product formation in the CH20-i sobutene
system is as follows:
H + i-SO-C-iHs -»- tert-C^Hg (20a)
H + iso-C4H8 ->- iso-C4H9 (20b)
2C4H9 -*- C4H10 + C4H8 (21a)
2C4H9 + C8H18 (21b)
HCO + C4H9 -*- C4H10 + CO (22a)
HCO + C4H9 -*• C4H9CHO (22b)
150
-------
C4H9 + CH20 - C4H10 + HCO
HCO + C4Hs -*- C4H8CHO (25)
HCO + C4H8 •*• CO + C4H9
The rate constant for H-atom addition to isobutene, k = (1.5 ± 0.5)
x 109 £ mole-^-sec'1 [Ref. 1*9], is about 50- times larger than that for H-atom
abstraction from CH20 by H-atoms, reaction 3 [Bef. l+k-k6]s so hydrogen
formed in reaction 3 should be suppressed essentially completely by the
addition of relatively small amounts of isobutene to CH20. This expectation
is borne out by our experimental findings; see the data of Table 21. In
experiments with added isobutene at each pressure of formaldehyde employed,
the $ is decreased to the same limiting value within the experimental
error, 0.32 ± 0.03; we believe that these data provide our best estimate for
. In view of the results from the CH20-Me3SiH experiments, $ + $ = 1.0,
and hence = 0.68 ± 0.03. Using the estimate for $, ajad the relation
which one anticipates in theory for pure CH20 photolysis with reactions 1-5,
k,/k = ($ - 0.10)/0.10, we find k, /k = 5.8. Although the accuracy of
this estimate is low, our results do suggest that reaction 5 does occur to
a small extent in addition to the reaction k.
A kinetic test of the mechanism of isobutene inhibition of Hg formation
in formaldehyde photolyses was made in experiments reported in Table 22.
The P_TT _ was kept as high as possible without inducing significant poly-
Un.2U
merization, and the irradiation times were as short as possible to insure
that a wide range of $TT values could be covered by C4Ha additions without
H2
the significant depletion of the isobutene concentration during the ir-
radiation time. These conditions were not met in the experiments of Table
21 at the lowest added isobutene pressures. Considering reactions 1 and 3
as the major sources of H2 in the CH20-C4Hs photoJyses, we expect $H to be
given by the approximate relation II: 2
H2
_ x _ ^ (II)
"- ~ ks[CH20]
In Figure k6 is shown a plot of the quantum yield function of relation II
versus the ratio [C4H8]/[CH20], taking ^ = 0.683 and $2 = 0.317, the
average of the nine values derived from the experiments at the highest
[C4H8]/[CH20] ratios in Table 21 (marked with asterisks). A reasonably
good linear relationship between the variables can be seen. The least
squares treatment of the data gives k2Q/k3 = ^ ± ^^ and ^ intercept
151
-------
TABLE 21. EXPERIMENTAL DATA FROM THE PHOTOLYSIS OF FORMALDEHYDE IN THE PRESENCE
OF ISOBUTENEa
ro
PCH20' Torr
Initial Average
0.959
0.998
3.193
3.198
3.191
3.194
3.203
3.207
3.176
3.194
3.186
3.185
3.217
8.265
8.268
8.242
8.272
8.248
8.228
12.12
12.10
12.09
11.83
0.943
0.981
3.156
3.139
3.137
3.139
3.152
3.150
3.127
3.144
3.151
3.137
3.171
8.200
8.210
8.185
8.219
8.202
8.176
12.04
12.02
11.85
11.76
PC4H8. Irradia-
Torr tion time ,
sec
5.
5.
0.
0.
0.
0.
0.
0.
1.
1.
3.
5.
5.
0.
0.
1.
3.
10.
10.
1.
2.
10.
30.
185
185
067
067
214
207
517
506
004
004
537
188
211
259
519
141
095
37
37
145
408
10
25
900
900
600
600
900
900
900
900
900
900
900
900
900
500
500
500
500
450
500
600
600
600
600
Quanta abs./ 2Ap, A, Torr PH ,
cell x 10~18 Torr . x 103C 2x
x 103b
1
1
2
2
3
3
3
3
3
3
2
3
3
3
3
3
3
2
3
4
4
4
4
.01
.05
.16
.15
.22
.22
.23
.23
.21
.22
.43
.22
.20
.30
.30
.29
.30
.97
.29
.44
.68
.36
.49
10
10
108
95
78
102
56
70
64
--
41
48
44
172
30
72
—
—
52
--
—
—
— -*
63.3
68.0
171
157
218
200
203
205
198
199
141
192
186
272
234
228
212
186
207
312
308
219
291
19.
20.
80.
73.
79.
79.
66.
64.
64.
62.
40.
62.
56.
108
80.
74.
67.
54.
62.
107
101
85.
86.
Torr
103
5
3
0
0
6
6
3
8
8
4
9
0
0
4
9
1
8
5
4
3
Pco, Torr *H a
x 10J
46.2
47.7
90.5
83.8
138
143
140
140
133
137
100
130
130
163
153
153
145
131
144
205
207
206
208
0.327*
0.327*
0.629
0.577
0.420
0.419
0.348
0.340
0.343
0.328
0.286*
0.327*
0.297*
0.558
0.413
0.386
0.344
0.313*
0.322*
0.415
0.366
0.325*
0.326*
*CO
0.775
0.768
0.711
0.662
0.729
0.751
0.733
0.735
0.702
0.718
0.695
0.685
0.690
0.840
0.788
0.787
0.744
0.748
0.744
0.796
0.756
0.785
0.786
W\
2.37*
2.35*
1.13
1.15
1.74
1.79
2.11
2.16
2.05
2.19
2.43*
2.10*
2.32*
1.51
1.31
2.04
2.16
2.39*
2.31*
1.92
2.05
2.41*
2.41*
aExcitation at 3130 A, temperature 25°C. Ap is the pressure change in the cell. °A is the pressure of non-
condensables. Experiments marked with asterisks are the runs at the highest lC4Hg]/[CH2O] ratios which were
used to derive the best estimates of the primary quantum yield 4>2 and the limiting ^cO/^Ho ratio-
-------
TA3LL 22. THE MIES OF H2 FORMATION IN FORMALDEHYDE PHOTO-
LYSES WITH SMALL AMOUNTS OF ADDED ISOBUTENEa
PCH20
13
12
12
12
12
12
Torr
t
.1
.0
.0
.0
.0
.0
PC4H8',Torr
x KT
101
204
299
402
604
994
Irradiation
time , sec
34.8
75
75
75
75
75
PH ,
16
28
24
22
20
17
Torr
103
.6
.8
.9
.6
.0
.7
0
0
0
0
0
0
*H2
.866
.767
.663
.602
.553
.471
o
[C4HgMH2CO], x103
Figured. Plot of [$H - ®2] - 1 versus
= 0.683 and $ = 0.317-
153
-------
near zero (-0.12 ± 0.14), as anticipated from relation II. This kpQ/k
estimate is in reasonable accord with that calculated from the ratio of the
individual rate constants, k /k = 5k ± 15(lcr) ,44~46>49 and the result adds
some credence to the mechanism choice, and the selection of the limiting
$ at high butene pressures as equal to $2 appears justified.
H2
Again in these C^JIs-CHsO mixture photolyses, as in the MesSiH-CHsO
mixture studies, the mechanism of the HCO destruction reactions and CO
formation appear to be ill-defined. A consideration of the reactions other
than H-atom removal will illustrate the unresolved problems. Die structurally
undesignated C^-Hg radicals shown in the reaction series should be largely
of the tert-butyl structure, since kon /kOA s 20 [Ref. ^9] . In view of the
- -•- - £_U£l £_OD
relative magnitude of D(tert-C4H9-H) = 91 [Ref. 50] and D(HCO-H) = 86 kcal
mole"1, the occurrence of 23 cannot be ruled out as an additional source of
HCO radicals. Reactions 2k and 25 should be followed by H-atom transfer from
formaldehyde, or addition to isobutene, thus leading to possible chain
reactions; some of these may also generate HCO radicals. The pressure change
in this system suggests that there may be short chains involving C4Hs poly-
merization reactions such as 2h. There is obviously no chain involving CO
formation. The reaction 26 has not been observed or considered previously
to our knowledge, yet its occurrence here seems to be another reasonable
source of CO. It is the analogue to the somewhat more exothermic, well
studied reactions of HCO with NO and 02: HCO + NO -*- MO + CO, AH = -28;
AH26 ^ ~21 kcal mole~:L'
Note from the results in Table 21 that $„,-/$„ climbs from a value of
1.0 for pure CE20 to a seemingly constant value of about 2.3 at high iso-
butene concentrations, independent of the [CH20]. Since § is much larger
00
than the $_ we have estimated from the $ values at high [C^], reaction
t-. il2
25 cannot be very important. It can be seen from the quantum yield data
that a decrease in $ n occurs significantly even at very low pressures of
Uw
added isobutene. We believe that this rapid onset of the attenuation of CO
reflects not only the decrease in reaction 3 as a source of HCO but the
suppression of the bimolecular reactions h and 5 which dominate CO formation
in the uninhibited CH20 photolysis; the incomplete suppression points to
new sources of CO. If the tert-butyl radical concentration is considerably
higher than that of the HCO radical, presumably through the significant
occurrence of reaction 26, the reaction 22a as well as 26 may become major
HCO loss steps. From the rate constants k = 9.3 x 109 [Ref. 51] and k, +
k = 2.2 x 1010 [Ref. 43], the approximate value of k22 s 2[k (k, + k )]'/2 =
2.9 x 1010 It, mole~1sec~1, can be estimated. From this it can be shown that
the [HCO]/[C4Hg] ratio must be lowered below 0.066, presumably through the
HCO radical removal reactions 25 and 26, before 90
-------
in reactions k and 5 even in the presence of isobutene, we expect f = 2(1 -
*Ha)/{l + [kj/Ck^ + k5)]} and ^ = 2(1 - ^/{[^/(k^ -f k^]}; using
measured values of $CQ - O.?l+ and *H = 0 for the isobutene-CH20 experiments
we would estimate ^ = 0.84; $2 = 0.16, and k^/k = 1.6. These values are
untenable in terms of the other results reported previously and determined
by other methods in this work. We conclude that reactions 4 and 5 are
probably unimportant loss reactions for HCO radicals and CO formation in
the CH20-C4Hs experiments, but the relative contribution of the reactions
22a and 26 to CO formation remains unclear.
Further confirmation of the conclusions concerning 0 and $ formulated
here can be had from the results of the CHgO-NO experiments described in the
next section.
Photolysis of Formaldehyde-Nitric Oxide Mixtures
As a further test of the $ and $ estimates derived from the CH20-C4H8
system experiments, CH20 was photolyzed in the presence of nitric oxide;
these data are summarized in Table 23. Note that the addition of NO to CH20
results in an increase in 0 and a decrease in $ . Formation of N2 was
L/O M2
also observed with yields which increased with NO concentration. No attempt
at the quantitative determination of this product was made. These ob-
servations are in line with the results of other photolytic studies of
Strausz and Gunning,12 Tadasa et al.s24 Harger and Lee,52 and Clark;28
however in the latter study an increase in 4> was observed only for ex-
citation at 2991 A but not for the longer wavelengths.
In order to rationalize the major features of these results consider
the simple reaction scheme 26-28:
H + NO (+NO) -»- HNO (+NO) (26a)
H + NO (+CHpO) -*- HNO (+CH20) (26b)
HCO + NO -*- HNO -f- CO (27)
2HNO •*- H20 + N20 (28)
Obviously this sequence is incomplete since it does not involve N02
formation which occurs in experiments at high NO.12'24 The results of Table
23 show that with the addition of the smallest amount of NO, $co reaches its
maximum value while $H2 is only slightly reduced. This suggests that at
low NO concentrations, the following reaction rate inequalities must exist:
R > R, + R and R > R^. Indeed, taking the measured values of k^ + k^
[Ref. 43] and k = 5.2 x 109 £ mole^sec-1 [Ref. 53] it can be shown that
I *-| £- n *l
at the absorbed light intensities of the present work (1.1 x 10- quanta !r
sec'1 at 8 Torr of CH20), practically all of the formyl radicals will react
155
-------
TABLE 23. EXPERIMENTAL DATA FROM THE PHOTOLYSIS OF FORMALDEHYDE IN THE PRESENCE
OF NITRIC OXIDEa
Initial
1.007
3.131
3.043
3.230
1 1 AQ
3.139
3.061
8.075
8.021
8.019
8.010
8.079
8.023
8.018
8.048
8.001
8.396
8.060
, Torr
Average
0.987
3.038
2.950
3.137
1 AtC
3.046
2.968
7.969
7.911
7.910
7.899
7.969
7.913
7.908
7.938
7.891
8.286
7.948
PNO, Torr
20.1
5.89
8.12
12.1
20.6
38.1
0.405
1.025
1.865
1.919
3.034
4.610
6.418
15.2
27.6
70.0
103
Irradiation
time, sec
1000
1000
1000
1000
1 AAA
1000
1000
600
600
600
600
600
600
600
600
600
600
600
Quanta abs./
cell x 10~18
1.04
2.79
2.67
3.02
2.87
2.80
3.37
3.34
3.38
3.37
3.32
3.29
3.32
3.27
3.25
3.14
3.30
x 103
24.4
137
107
79.5
84.5
53.1
177
178
173
153
151
144
107
90.9
61.3
70.3
PCO- T°rr *H2
x 10J
0.398
0.825
0.663
0.464
•>•>*»
0.508
0.327
0.881
0.902
313
0.863
0.780
0.778
0.736
0.550
0.474
0.331
0.361
*CO
1.57
— —
aExcitation at 3130 A, temperature 25"C.
-------
with NO in experiments with P = 8 and PNQ = 0.5 Torr. The extent of the
H-atom removal reactions 26a and' 26b is more difficult to assess since
values for k^ with M = ClfeO and NO are not known. However k , = 2.2 x
1010 ^mole^sec-1 [Refs. 53,5^1 with 1 atra of H2:
H + NO (+Ha) -*- MO (+H2) (26C)
Assuming k^ £ 10^^, it can be shawn that $H Wili be reduced *>y less
than lU% for experiments at CH20 and NO pressures as before. Thus the NO-
CH20 experiments provide a unique system in that NO quenches the HCO radicals
efficiently but affects the H-atoms very little in experiments at low NO and
total gas pressures.
According to the reaction scheme outlined above, at high NO concen-
trations all of the H atoms will be scavenged and the $TT will be reduced
H2
to $2. As expected the lowest values of $H of 0.32? and 0.331 at formalde-
hyde pressures of 3 and 8 Torr, respectively, are very close to $ as
determined in the C^O-C^Ks system, $2 = 0.32 + 0.03. At a formaldehyde
pressure of 1 Torr, $ is reduced by the addition of 20 Torr of NO, but
only to a value of 0.398; this probably reflects the difference in the rate
constants, k26b > k2ga, and the less efficient quenching of H-atoms for the
low pressure conditions. The average of the two determinations at 8 and 3
Torr of CH20 and at the highest added NO pressures gives <£> = 0.3^. This
H2
compares well with $„ = 0.31 reported by Clark from experiments at 3172 A
H2
using P = 10 and P = 50 Torr.a8
In terms of the mechanism 1-3 and 26-28 outlined above, the relation III
should describe $, in experiments at low NO concentrations:
This expectation appears to be qualitatively correct; using it we estimate
$, = 0.724 (3 Torr CHaO) and 0.717 (8 Torr CH20); this should be compared
with $ = 0.68 from the isobutene experiments. However, it can be seen frcm
an inspection of the data at both 3 and 8 Torr of CH20 that $co - $H in-
creases as NO increases. A similar observation was made by Clark.28 There
are several possible explanations of these results. First, ^ could change
as the result of a perturbation of the distribution between the precursors
of molecules and radical products. Houston and Moore22 have observed that
addition of NO to formaldehyde photolyzed at 3055 A increased the rate of
formation of vibrationally excited CO (v = 1). They rationalized this
effect assuming that collisions with NO enhanced the molecular mode of
formaldehyde decomposition. Following this line of argument we would expect
- $ to decrease with increasing W concentration; however the reverse
CO H2
157
-------
effect is observed. Careful examination of the assumptions made in the
derivation of expression III points to an alternative explanation for this
effect. The equality between - $ and $, is based upon the simple
CU rig J-
reaction scheme which requires that for each H-atom lost through the reaction
26, a formyl radical and therefore a molecule of CO will not form. It is
possible that this assumption does not hold at high NO concentrations.
Under these conditions both N2 and N02 are formed, and the simple reaction
scheme used in the derivation of expression III cannot account for this.
Conceivably one of the intermediates involved in the reaction sequence
that leads to the N2 and N02 might be a free radical that abstracts hydrogen
from formaldehyde. In this case a decrease in $ would not be accompanied
H2
by a concurrent decrease in 3> •
158
-------
REFERENCES
1. a) J.G. Calvert, J.A. Kerr, K.L. Demergian, and R.D. McQuigg, Science,
175* 751 (1972); b) K.L. Demerjian, J.A. Kerrs and J.G. CalvertT Advan.
Environ. Sci. Technpl., 4, 1 (19714-). -
2. R.G.W. Norrish and F.W. Kirkbride, J. Chem. Soc.. 1932, 1518.
3. E. Gorin, J. Chem. Phys.3 7, 256 (1939).
4. E.I. Akeroyd and R.G.W. Norrish, J. Chem. Soc., 1936, 890.
5. F.E. Blacet and W.J. Blaedel, J. Amer. Chem. Soc., 62, 3374 (1940).
6. M.A. Khan, R.G.W. Norrish and G. Porter, Proc Roy. Soc., (London).
A219, 312 (1953).
7. J.G. Calvert and E.W.R. Steacie, J. Ghem. Phys., 19, 1176 (1951).
8, R. Klein and L.J. Schoen, J. Chem. Phys., 24, 1094 (1956).
9. E.G. A. Horner, and D.W.G. Style, Trans. Faraday Soc., 50, 1197 (1954)
10. J.F. McKellar and R.G.W. Norrish, Proc. Roy. Soc., (London),
(I960).
11. A.G. Harrison and F.P. Lossing, Can. J. Chem., 38, 544 (1960).
12. O.P. Strausz and H.E. Gunning, Irans. Faraday Soc., 60, 347 (1964)
13. M. Venugopalan and K.O. Kutschke, Can. J. Chem., 42, 2451 (1964).
14. B.A. DeGraff and J.G. Calvert, J. Amer. Chem. Soc., 89, 2247 (1967).
15. R.D. McQoigg and J.G. Calvert, J. Amer. Chem Soc., 91, 1590 (1969).
16. S. Glicker and L.J. Stief, J. Chem. Phys., 54, 2852 (1971).
17. E.S. Yemg and C.B. Moore, App. Phys. Lett., 21, 109 (1972).
18. E.S. Yeung and C.B. Moore, J. Chem. Phys., 58, 3988 (1973).
19. H.P. Sperling and S. Toby, Can. J. Chem., 51, 471 (1973).
20. R.G. Miller and E.K.C. Lee, Chem. Phys. Lett., 27, 475 (1974).
21. E.S. Yeung and C.B. Moore, J. Chem. Phys., 60, 2139 (1974).
22. p.L. Houston and C.B. Moore, J. Chem. Phys., 65, 757 (-1976).
23. R.S. Lewis and E.K.C. Lee, unpublished work which is referenced in
Lewis, et al., ref. 29.
159
-------
2k. K. Tadasa, N. Imai and T. Inaba, Boll. Chem. Soc. Japan, ^9, 1758 (1976),
25. T.L. Osif and J. Heicklen, J. Phys. Chem., 80, 1526 (1976).
26. R.G. Miller and E.K.C. Lee, Chem. Phys. Lett., 33, 104 (1975).
27. T.L. Osif, Ph.D. Thesis, "The Reactions of 0(XD) and OH with CH3OH, the
Oxidation of HCO Radicals, and the Photochemical Oxidation of Formalde-
hyde", Pennsylvania State University, August, 1976.
28. J.H. Clark, Ph.D. Thesis, "Laser Photochemistry and Isotope Separation
in Formaldehyde", University of California, Berkeley, 1976.
29. R.S. Lewis, K.Y. Tang, and E.K.C. Lee, J. Chem. Phys., 65, 2910 (1976).
30. K.Y. Tang, P.W. Fairchild, and E.K.C. Lee, J. Chem. Phys., 66, 3303
(1977).
31. J. Marling, J. Chem. Phys., 66, ^200 (1977).
32. W.H. Fink, J. Amer. Ghem. Soc., 9^, 1073, 1078 (1972).
33. D.M. Hayes and K. Morokuma, Chem. Phys. Lett., 12, 539 (1972).
3^. R.L. Jaffe, D.M. Hayes, and K. Morokuma, J. Chem. Phys., 60, 5108 (1971!-).
35. R. Spence and ¥. Wild, J. Chem. Soc., 1935, 338.
36. J.G. Calvert and J.N. Pitts, Jr., "Photochemistry", John Wiley and Sons,
New York, 1966, p. 732.
37. P. Warneck, Z. Naturforsch., 26A, 204? (1971).
38. E. Murad and M.G. Inghram, J. Chem. Phys., In, h-Ok (1964).
39. R. Walsh and S.W. Benson, J. Amer. Ghem. Soc., 88, ^570 (1966).
40. A. Horowitz and J.G. Calvert, work in preparation for publication.
4l. JAMAF Thermodynamic Tables, 2nd Edition, U.S. Department of Commerce,
N.B.S., 1971.
if-2. T- Hikida, J.A. Eyre, and L.M. Dorfman, J. Chem. Phys., 5*f, 3^22 (1971).
43. a) M.J. YeeQuee and J.C.Y. Thynne, Trans. Faraday Soc., 63, 1656 (1967);
b) Ibid, Ber. Bunsenges. Phys. Chem., 72, 211 (1968).
¥l-. W.R. Brennen, I.D. Gay, G.P. Glass, and H. Niki, J. Chem. Fhys., ^3,
2569 (1965).
45. A.A. Westenberg and N. deHaas, J. Phys. Chem., 76, 2213 (1972).
160
-------
h6. B.A. Ridley, J.A. Davenport, L.J. Stief, and K.H. Welge, J. Chem. Phys.,
57, 520 (1972).
l47. a) I.M.T. Davidson and A.B. Howard, J. Chem. Soc. Chem. Comm., 323 (1973);
b) P. Potzinger, A. Ritter, and J. Krause, Z. Naturforsch., 30A, 3^7
1975); c) R. Walsh and J.M. Wells, J. Ghem. Soc. Faraday I. 72. 100
(1976).
1$. J. Cooper, A. Hudson, and R.A. Jackson, J. Chem. Soc., Perkin Trans. II3
1973, 1933.
1+9. J.A. Kerr and M.J. Parsonage, "Evaluated Kinetic Data on Gas Phase
Addition Reactions", Butterworth, London, 1972, p. 32; we have averaged
the six independent estimates shown here which were derived from Hie
experiments at the low pressures comparable to those employed in this
work.
50. "Handbook of Chemistry and Physics" 50th Edition, R.C. ¥est editor.
The Chemical Rubber Company, Cleveland, 1969, p. F-165-
51. D.A. Parkes and C.P. Qainn, J. Ghem. SPG., Faraday Trans. I, 72, 1952
(1976).
52. R.A. Harger and E.K.C. Lee, Abstracts, 173rd American Chemical Society
Meeting, New Orleans, La., March 20-25, 1977.
53. I. Tanaka, private communication of the results of Dr. Shibuya, Ph.D.
Thesis, Tokyo Institute of Technology, Tokyo, Japan, 1976.
5^. J.J. Ahumada, J.V. Michael and D.T. Osborne, J. Chem. Phys., 57, 3736
(1972).
55. R. Atkinson and R.J. Cvetanovic, Can. J. Ghem., 51, 370 (1973).
161
-------
THE WAVELENGTH DEPENDENCE OF THE QUANTUM EFFICIENCIES OF THE PRIMARY PROCESSES
IN CH20 PHOTOLYSIS AT 25°C
Introduction
The absorption of solar radiation by formaldehyde vapor present in the
polluted atmosphere results in its photodissociation by the primary processes
(1) and (2):
CHpO + hv -*• H + HCO (1)
CH20 + hv •*- H2 + CO (2)
The H and HCO radicals formed in (l) react with oxygen in the air to generate
H02 and HC002 radicals:
H + 02 + M ->- H02 + M (3)
HCO + 02 + M ->- HC002 + M
HCO + 02 ->- H02 + CO (5)
Thus CH20 photolysis in the atmosphere may be a significant source of the
hydroperoxy and peroxyformyl radicals; both radicals may be of significance
in photochemical smog formation and in other atmospheric reactions.1'7
Previous estimates of the rates of processes (1) and (2) in the lower atmo-
sphere4"6 were based upon the primary quantum yield versus wavelength data
of McQuigg and Calvert7 which were determined by a very indirect and necessarily
imprecise method. The wavelength dependence of $, and $ has been estimated
in several siudies.which have extended over nearly a forty year period.7"15
In many of these studies CH20 photolyses were made with added radical
scavengers (alkenes, NO, I2) and the limiting values of and $„.. at high
ri2 CO
scavenger concentrations were used to derive estimates of $ and cb . However
the published estimates of these quantities are so widely divergent that
trends in $, with A are not well defined; see Figure lj-7. However if one
examines the trend of $, with ?\ within each set of the several recent results
(Calvert and McQuigg,7 triangles; Lewis, et al.,12 closed circles; Clark,14
closed squares), it is clear that the value of 0, increases with decreasing
wavelength of the absorbed light. However the large divergence between $,
estimates of the different groups points to the critical need for a further
accurate definition of the $ values as a function of A.
The wavelength dependence of $, can in principle be used to es ;ablish
the wavelength of the onset of the process (1). If the dissociation
originates from the vibrationally excited ground state of the CH20 molecule,
as seems likely, then the threshold wavelength may be related to the bond
162
-------
1.0 r-
0.8
0.6
0.4
0.2
• A
» A
I I I I
280O 30OO 3200 3400 360O
o
Wavelengh, A
Figure 4?. The previously published estimates of the primary quantum yield
of the free radical-forming process (l) in CH20 photolysis: CH20 + hv
H + HCO (1); data are from the following studies: Gorin,8 open circles;
DeGraff and Calvert,10 open diamonds; McQuigg and Calvert,7 open triangles;
Sperling and Toby,11 closed diamonds; Lewis, Tang, and Lee,12 closed circles;
Clark,14 closed squares.
163
-------
dissociation energy, D(H-CHO). Recent studies12*14'15 suggest strongly that
the threshold for photodissociation of CH20 is near 3^-00 A; the energy at
this wavelength (Qk kcal mole"1) is in reasonable agreement with the D(H-CHO)
estimates based upon kinetic16 and photoionization experiments.17 On the
other hand results of several earlier studies7'9'11*18 indicated that the
free radical process (l) in CH20 photolysis can occur at wavelengths con-
siderably longer than 3^-00 A. A new unambiguous determination of the
threshold wavelength for (1) is required to settle this issue.
In a recent study of the 3130 A photolysis of CH20 at 25°C, we have
found that $, and 9 can be determined relatively simply using isobutene as
an H-atom scavenger.19 ¥e have applied this technique in the present study
to provide seemingly reliable new estimates of $, and $~ at selected
wavelengths from 2890 to 3380 A. It has been possible to derive a new
estimate of D(H-CHO) from these data. Also we have used them to estimate
the rate of H and HCO generation from CH20 photolysis in the lower atmosphere.
Experimental
The experimental setup and the method of gas chromatographic analysis
were the same as that used in the previous work.19 In order to isolate, the
different wavelengths from the high pressure mercury arc (Osram HBO 500W),
we used a high intensity grating monochromator (Schoeffel CM 250). This
had a grating with 2365 grooves mm"1 that gave reciprocal linear dispersion of
of 16.5 A per 1 mm slitwidth. The monochromator was. calibrated with the
3132 and 33^-2 A lines of the mercury arc with the monochromator set at 0.25
mm (hbw, U.I A). The amount of scattered light at wavelengths longer than
3380 A did not exceed 0.2%, and in most cases, it was considerably smaller
as determined by introducing a WG-385 filter (cut off, ~ 3&50 A) between
the monochromator and the photodiode (RCA 935) placed at the back of the
cell.
For the non-monochromatic source employed here, the "effective"
excitation band is determined by the wavelength distribution of the absorbed
light j this is controlled by the emission spectrum of the source on which
a near Gaussian intensity distribution is superimposed by the slit of the
monochromator, and by the absorption spectrum of CH20. Taking these factors
into account we have calculated the effective bands for the different
monochromator settings used in this work. In general for the 33 A hbw
experiments, about 90fo of the absorbed light fell within the A ± 33 A
range; we define the "effective" wavelength here as that wavelength at which
the absorbed light intensity versus wavelength plot could be divided into
two parts of equal area. As a result of the presence of strong mercury
lines at 2968, 3023, 3132, and 33^2 A and the fine structure of the CH20
absorption spectrum, some of the band profiles obtained for the 50 and 66 A
hbw experiments showed long, and at times, intense tails that extended well
beyond the AQ ± hbw region. However even in these cases A -- was probably
not displaced greatly from AQ as indicated by the agreement between the
experimental results at the same A but with different hbw values.
-------
Azome thane was used for actinometry taking $ = 1.00. In addition, in
order to ensure an accurate determination of the quantum yields, each run
with added isobutene was preceded by a run with pure CH20 at the pressure
employed in^the GH20- isobutene photolyces. In all experiments the temperature
was near 25 C, and the PC,H Q was 10 Torr. Polymerization ol the CH20 was
negligible for these conditions. The conversion of CH20 into CO and H2 did
not exceed 2% in any experiment.
Results
The photolysis of CH20 mixtures with isobutene was studied over the
wavelength range from 2890 to 3380 A? and the results of these experiments
are summarized in Table 2k. At 3380 A the addition of either 1 Torr of 02
or 25 Torr of NO to 10 Torr of CH20, did not have any significant effect of
$TI . In addition the effect of C02 addition on hydrogen formation was
H2
examined both in the absence and presence of isobutene. The results of these
experiments are given in Table 25 .
Discussion
The Wavelength Dependence of the Quantum Yields of the Primary Photodis-
soeiation Processes in Formaldehyde --
Previous studies have shown that the photolysis of pure formaldehyde at
25°C results in the formation of hydrogen and carbon monoxide via reactions
(1), (2), (6)-(8):
CH20 + hv H +-HGO (1)
+ hv -»~ H2 + CO (2)
H + CH20 -*~ H2 + HCO (6)
2HCO -*- H2CO 4- CO - (7)
2HCO -*• H2 + 2CO (8)
In our previous CH20 study at 3130 A and 25°c,:L9 we found that isobutene
was an effective scavenger for H-atoms, and when sufficient isobutene was
added. $ was reduced to a value which corresponded to 3> through the
n-2
occurrence of reaction (9):
H + iso-C4H8 ->- C4H9 (9)
In view of the simplicity of this system, we have adopted the same approach
to $ and $p determination in the present study. The absolute value of
/-\
3> was determined in experiments with pure CH20 at 25°C and at several
H2
wavelengths over the range important for the lower atmosphere photochemistry
of CH20: 2930, 3130, 3150, 3335, and 3380 A. With the exception of the
runs at 3130 A and 3380 A, the data confirm the view that the quantum yield
of the product H2 is unity within the experimental error.14 The results
165
-------
TABLE 2k. THE EFFECT OF ISOBUTENE ON HYDROGEN FORMATION
IN THE PHOTOLYSIS OF FORMALDEHYDE AT 25°C AND WITH EX-
CITATION AT 2890 TO 3380A
Wavelength (^eff)b. A
2890
2894 (2894)
2930
2970 (2975)
3030 (3032)
3130 (3137)
3150
3175
3195 (3180)
3230
3250
3270 (3271)
3300 (3297)
3300
3335
3340 (3338)
3360
3380 (3385)
3380
o o c
hbw, A *n2
66
33
66 1.02
41
33
33 1.12
33 1.06
50
33
41
50 1.01
33
33
50
33 1.0
17
66
33
66 0.75
-------
TABLE 25. THE EFFECT OF C02 ON HYDROGEN FORMATION IN THE PHOTOLYSIS OF
FORMALDEHYDE AT 25°Ca
0
Wavelength, A
2890
2890
2930
2930
3130
3300
3300
3300
3300
3360
3380
3380
3380
3380
(hbw)
(66)
(66)
(66)
(66)
(33)
(50)
(50)
(50)
(50)
(66)
(66)
(66)
(66)
(66)
Piso-C4H8' Torr
10
10
10
10
0
0
14
14
14
0
0
0
0
0
PC02' Torr
0
250
0
350
400
305
0
315
320
295
154
200
276
330
* /* ° b
*H2/*H2
0.313
0.319
0.275
0.280
0.996
0.905
0.660
0.582
li
0.546
0.761
0.910
0.810
0.757
0.606
aPCH 0 = 10 Torr in each run. bThe quantum yield ratio is $H2 from the
run at the added CO2 pressure shown to that for pure CH2O photolysis
at this wavelength and PCH20 = 10 Torr.
167
-------
from the experiment at 3130 A and those which we reported previously for this
wavelength region,19 give estimates of $u2 somewhat higher than unity. This
was interpreted in our previous work as suggesting the occurrence of reaction
(Q) with k /ko s 5.8. In view of the present results at the four other
short wavelengths studied, we now feel that it is likely that $„ = 1.0
2
over the entire range from 3335 to 2890 A. Thus as Clark has suggested,14
reactions (6) and (7) are probably the dominant fates of the H and HCO
radicals in pure CH20 photolysis at 25°C. We have estimated previously that
kp./kx- ^ 43, so that in our present experiments employing PnT, A = 10 Torr,
9 D \ji\Q U
and P T = 10-14 Torr, about 98% of the hydrogen atoms formed in (l) will
C4H10
be scavenged by isobutene. Thus the primary quantum yield ($-,) of the free
radical process (l) can be estimated reliably from our results using the
relation (10):
1 -
He
1 +
43P
(10)
Here 3> and $1 are the quantum yields of H2 in pure formaldehyde photolyses
and in CH20-isobutene mixture photolyses, respectively. The $, values
estimated in this fashion are given in Table 26.
Inspection of the results of Table 26 shows that the onset of
formaldehyde photodissociation in primary process (1) is at 3370 ± 10 A.
This corresponds to an H-CHO bond dissociation energy of 84.8 ± 0.3 kcal
mole"1, in reasonable accord with the other recent estimates of the photo-
dissociation limit: Ty, > 3300 A, D(H-CHO)<87 kcal mole"1 [Ref. 12], and
AJ-J ^3392 A, D(H-CHO) ^ 84 kcal mole-1 [Ref. 14]. The limit which was
estimated for CHDO dissociation is Ap = 3250 ± 50 A, D(H-CDO) = 88 ± 1.4
kcal mole"1 [Ref. 15]. Our present estimate of D(H-CHO) is somewhat lower
than, but in accord with the kinetic estimate of D(H-CHO) = 87 ± 1 kcal
mole"1 [Ref. 16] and the photoionization value of D(H-CHO) = 88.0 ± 1.6
kcal mole"1 [Ref. 17].
Our data show that for CH20 photolyses from 3380 to 3130 A, $, in-
creases rapidly until it levels off at about 3130 A. The general features
of the A dependence of $, are in agreement with the observations made in
recent studies of Lewis, et al.,12 Marling,15 and Clark14 using very dif-
ferent methods. The results of these studies are also given in Table III
for comparison. In the case of the study by Marling, CHDO photolyses were
made, and the relative yields of H2, D2, and HD were measured in photolyses
using either monochromatic laser radiation or a high pressure mercury arc-
monochromator combination with a hbw of 30 A. In this system H2 and D2
could be formed only in free radical reactions following (1), while HD~
could be generated from both the molecular and free radical decomposition
168
-------
TABLE 26. THE WAVELENGTH DEPENDENCE OF THE PRIMARY QUANTUM
EFFICIENCY OF THE PHOTODECOMPOSITION OF FORMALDEHYDE INTO H
AND HCO (0>)
X, A
2840
2882
2890
2894
2930
2934
2970
2982
2991
3030
3035
3040
3050
3088
3130
3140
3166
3172
3175
3180
3195
3210
3230
3250
3260
3264
3270
3296
3298
3300
3310
3324
3335
3340
3360
3378
3380
3392
This work Marling
0.701
0.711
0.740
0.737
0.760
0.760
0.735
0.760
0.692
0.636
0.554
0.519
0.460 0.456
0.442
0.438
0.330
0.097
0.097
0.212
0.113
0.048
0.111
0.00
of primary proces
[15]a Lewis, et
Original
0.27
0.34
0.26
0.33
0.36
0.26
0.32
0.37
0.33
0.27
0.22
0.15
al.[12]a Clark [14]b
Normalized Low NO High NO
0.57
0.72
0.55
0.70
0.48 0.70
0.76
0.55
0.68
0.78
0.70
0.42 0.69
0.57
0.46
0.42 0.48
0.32
0.31 0.38
-0.03 0.01
formalized at 3040 A to give *! = 0.76. Experiments with low NO, assuming *co
- *H =()>]_ as in the original work; experiments with high NO, assuming *H2 = 4>2
and taking *! + *2 = 1 for X * 3300 A and *x + *2 = 0.75 at 3392 A.
169
-------
modes of CHDO. According^ the lower limit of $ should be given by the
fraction (H2 + D2)/(H2 + D2 + HD), provided that "^ + 3>2 = 1. Furthermore
as Marling noted, this ratio should be proportional to $, values. We may
use our data to determine the proportionality factor. At short wavelengths
the difference between $, in CH20 and CHDO should be small, so we have
assigned our experimental value of 3> =0.76 at 3030 A to his ratio of 0.55
measured at 30^0 A. The $ versus A results of Marling normalized in this
fashion check reasonably well with those which we have estimated in this
work. Compare the $ values so derived from the Marling data (squares) with
ours (closed circles) in Figure48. In view of the isotopic difference
between the CHaO and CHDO molecules, we would not expect a more exact cor-
respondence between the two sets of data.
Our results may be compared as well with those of Clark14 who employed
a tunable laser to photolyze CH20 with NO added as a HCO-radical and H-atom
scavenger. He estimated and $_- in experiments with both large and
rig CO
small amounts of added NO. In the runs at low added NO, Clark assumed that
3> = 5> - $ . However in his experiments with 3392 A excitation and with
-L L/U lig
large amounts of added NO, $„ decreased to a value significantly below the
"2
value of <3>p expected from the relation <& = 1 - and his - H + HCO in experiments with excitation wavelengths as long as 3392 A,
so collisional perturbations should not enhance the dissociation. Further-
more, we find no significant reduction of $ with NO addition and with
2
CH20 excitation at 3380 A; the quantum yield of H2 (0.75) in the absence of
NO is lowered only slightly ( = 0.73) by the addition of 25 Torr of NO.
h.2
Therefore it seems more likely to us that Clark's $u values at high [NO]
i!2
are more appropriate estimates of than those derived from the low [NO]
data. Indeed, the $, values estimated from Clark s results with this assump-
tion are in excellent agreement with those reported in this work; see Table
26 and the plot in FigurekQ (open circles).
Lewis, Tang, and Lee12 photolyzed CH20-NO mixtures at low pressures
using a tunable laser ( ~1 A hbw). In a very unique experiment they
measured the intensity of the chemiluminescence from the excited HNO molecules
formed in the reaction, H + NO -*- HNO , following process (1). Their method
gave relative values of $_ which were then converted to absolute values
taking $, = 0.36 at 3035 A. The reference value was obtained in another set
of experiments in which the low pressure photolysis of CH20-l-butene
mixtures was studied. The vibrationally-rich C^g radicals, which resulted
170
-------
2800 3OOO 3200 3400
Wavelengh, A
Figure 48. The wavelength dependence of the primary
quantum yield $, of the free radical-forming process
(1) in CH20 photolysis: CES0 + hv -*~ E + ECO (l);
estimates shown are from experiments at room temp-
erature: this work, closed circles; Lewis, Tang,
and Lee,12 triangles (
-------
from H-atom capture by 1-butene following process (1), were monitored through
their fragmentation to C3H6 ( and CH3) [Kef. 20]. It can be seen in Table
III that the trend in our results parallels that observed by Lewis, et al.,12
but our absolute values are significantly higher than those of Lewis, et al. ,
at each wavelength. The origin of the difference appears to be related to
the choice of $, value by Lewis et al., for the standard at 3035 A. Since
the details of the CH20-l-butene study have not yet appeared, it is dif-
ficult to judge the wisdom of their choice. If we accept our estimate of
$ at 3030 A as the standard to adjust the relative $, data of Lewis, et al.,
at 3035 A, then our data and those of Lewis, et al., agree well over a wide
range of wavelengths; in Figure kQ compare our data (closed circles) with
those of Lewis, et al. (triangles). Thus the seemingly logical adjustments
of the data outlined here lead to a set of self consistent estimates of
$, and its variation with A. The data from our study and all of the recent
studies carried out at room temperature, 12>14»15 define the same 0 versus A
function within the experimental error of the determinations; see Figure 48
All the data support the conclusion that there is a much more rapid initial
rise in $, with decreasing A for excitations below 3^-00 A than had been
suggested from previous studies using less definitive methods of $, de-
termination.7"11 Furthermore the plateau in $, values which occurs well
below unity, appears to be well documented from all of the recent results.
For excitation at wavelengths less than 3100 A, the decomposition of the
excited formaldehyde by processes (l) and (2) appears to occur in a nearly
constant ratio of about 3/1 , independent of the excitation energy.
The absence of a C02 effect on $ in the photolysis of CH20-isobutene
2
mixtures at 2890 and 2930 A (see Table 25.), indicates that the theoretically
possible formation of "hot" hydrogen atoms that cannot be quenched by iso-
butene is unimportant in this system. If this were the case then the ad-
dition of the large amount of C02 to the Cl^O-C^HQ mixture photolysis should
have resulted in a decrease in 3>. for these conditions; no such effect is
J12
seen experimentally.
A further test of the mechanism of CH20 photolysis at 3380 A was made
in this work. ^We observed in another recent study of CH20-02 mixture photo-
lyses at 3130 A that H2 is formed in an unexpected, long-chain process which
appears to be initiated by H02 radical interactions with CH20. Quantum
yields as high as $ =6 were observed in these experiments at 25°C.21
M2
In the present study we added 1 Torr of 02 to CH20 (10 Torr) in photolyses
at 3380 A and observed no increase in $„ which characterized the H-atom
H2
chain in the experiments at 3130 A. This confirms our conclusion based upon
the CH20-C4Hs and CH20-NO photoJyses that there is very little, if any,
dissociation of CH20 into free radicals when it is excited at 3380 A.
The Effect of C02 Addition on the Photodissociation of Formaldehyde —
If one is to extrapolate our findings regarding $x and $2 variation with
172
-------
7v to estimate the rates of processes (1) and (2) in the lower atmosphere,
then the possibility of quenching of the excited CH20 molecules by air at
1 aim must be evaluated. All present evidence suggests that decomposition
in processes (1) and (2) proceeds via a highly vibrationally excited ground
state molecule;22 hence relaxation of the excited molecule before decomposi-
tion is a real possibility at pressures of the lower atmosphere. The ad-
dition of a representative atmospheric gas such as nitrogen in these ex-
periments was precluded by the analytical methods which we employed, so C02
gas was added to determine the possible significance of the relaxation
processes. G02 is somewhat more efficient than the diatomic gases N2 and 02
in energy relaxation processes; thus it is 2.4-times more efficient than
02 and Na in the relaxation of excited singlet S02 [Ref. 23], Thus about
300 Torr of C02 should provide roughly the equivalent relaxation medium for
CH20 as 1 aim of air. Our results from experiments with CH20 and CH20-
isobutene mixture photolyses with added C02 are summarized in Table 25 . It
can be seen that in experiments at 2890, 2930, and 3130 A, there is no sig-
nificant lowering of the $ with C02 additions of up to 400 Torr. Thus
we can conclude that the quantum efficiencies of primary processes (l) and
(2) are not sensitive to added C02 gas pressures equivalent to 1 aim of air
for CH20 excitations at 3130 A or shorter^wavelengths. However the data
from experiments with excitation at 3300 A show a small quenching of hydrogen
with C02 addition. The $ value estimated in CH20-isobutene mixtures in the
absence of C02 is 0.660; this is lowered somewhat (4> = 0.582, P = 315j
0.546, Ppn = 320 Torr) in runs with added C02. It is important to establish
whether any concurrent reduction in 5> occurred for these experiments. We
can attempt this comparing the data from the pure CH20 and CH20-C02 runs at
3300 A. The $„ values from these two runs show that *, + $„ is lowered
jn^ -*- *—
from 1.0 to 0.905 on adding 305 Torr of C02. From the experiments with
added isobutene and C02-isobutene we would expect a lowering by a fraction
of (0.564 ± 0.0l8)/0.660 = 0.85 ± 0.03 if both ^ and $2 were quenched
equally by C02 addition; the lowering of « to 0.34 + (0.564 ± 0.018) =
"S
0.904 ± 0.018 would occur if only $2 were affected significantly by C02
addition. Obviously the experimental value observed here, 0.905, suggests
that the latter alternative is correct. Although the accuracy of our data
does not allow a firm conclusion, the present data suggest that the effect
of added C02 on $ in CH20 photolyses at 3300 A is largely on the quenching
of primary process (2); $, appears to be affected very little, if any, by
the addition of 320 Torr of added C02, at least for the vibronic bands of
CH20 populated by CH20 excitation at 3300 A.
A much more pronounced quenching of $H by added C02 was observed in
experiments at 3380 A. Here the quantum yield of hydrogen formation is well
below unity even in the absence of added C02 (9 = 0.75). The ratio of
this $ to that observed with added C02 decreased regularly from 0.91 to
0.6l as" 154 to 330 Torr of C02 was added. Our observations here are in accord
173
-------
with the effects of added C02 on CH20 quantum yields observed by Sperling
and Toby11 in their experiments at 3130 and 33^0 A and l60°C. Our quenching
data do not fit well a simple one excited state quenching mechanism of the
Stern-Volmer type. It seems likely to us that a multistep mechanism of
deactivation of the vibrationally excited CH20 must be operative with dis-
sociation rates which increase markedly with increasing vibrational
excitation of the CH20. Only the molecular primary process occurs at 3380 A,
and it is apparent from these data that the efficiency of fragmentation by
this mode can be lowered significantly at the pressuresoof the lower atmosphere
for excitation of CI^O at wavelengths longer than 3380 A.
Estimation of the Rate of Formaldehyde Photolysis in the Lower Atmosphere—
We may use the primary quantum yield data derived in this study together
with the CH20 extinction data of McQuigg and Calvert"7 and the recent actinic
irradiance data of Peterson,24 to derive new estimates of k-^, the apparent
first order rate constant for the sunlight-initiated, free radical-forming
primary process (l) for CH20 in the lower atmosphere. The standard
procedures outlined by Leighton25 were applied in our calculationsj
the e and $.. values were averaged over wavelength intervals of 50 A. No
correction was necessary for the quenching of the 1 atm of air on the primary
quantum yields of process (1), since quenching of $ became important in
the experiments with added C02 only at the wavelengths were $, was unimportant.
The k., values so derived are summarized in Table 27 for typical conditions
near ground level and at various solar zenith angles. These estimates were
made using Peterson's data calculated assuming no reflection from the earth's
surface. They should be increased by the factor 1 + a where a is the
fraction of the solar ultraviolet light (2900-3700 A) which is reflected
from the surface of the earth at the point of interest.
Because of the strong dependence of <1> on the presence of added gas
for CI^O excited at the longer wavelengths (A > 3360 A), and the undetermined
efficiency of N2 and 02 on this process, no attempt was made to estimate
rate constants for the photodecomposition of CH20 in sunlight which occurs
by process (2). We did estimate the rate constant for light absorption by
CH20 in lower atmosphere (k ), and these data are also presented in Table 27.
This constant can be used as an upper limit to the apparent first order rate
constant for the total rate of CH20 photodecomposition by processes (1) and
(2) in sunlight. Because of the stated uncertainties in the extent of air
quenching of process (2) in C^pO excited at the long wavelengths, k0 < k -
<-. a
kn. Hie present estimates of k and k are shown graphically in Figure ^9
_L J_ cL
for various solar zenith angles. These estimates are, on the average, about
15% lower than those derived by our group previously from less accurate
l
and solar irradiance data.3*25
It is clear that H-atoms formed in process (l) will react almost ex-
clusive^ in the lower atmosphere to form H02 radicals in reaction (3). The
fraction of the HCO radicals which will react by the two paths (k) and (5)
-------
TABLE 27. ESTIMATED APPARENT FIRST ORDER RATE CONSTANTS FOR SUNLIGHT AB-
SORPTION BY CE20 (ka) AND FOR PHOTODECOMPOSITION BY REACTION (l), CH20 + hv
H + HCO (1), IN SUNLIGHT IN THE LOWER ATMOSPHERE (k-,)a
Solar Zenith 0 10 20 30
Angle, deg
ka, min"1 x 103b 7.74 7.62 7.38 6.90
kj, min"1 x 103 2.31 2.22 2.17 1.98
40 50 60 70 78 86
6.18 5.14 3.80 2.19 0.96 0.21
1.71 1.35 0.92 0.46 0.17 0.028
actinic irradiance was taken from the estimates of Peterson [24] for conditions of zero ground
reflectivity. These values represent an upper limit to the specific rates of the total decomposition
of formaldehyde by both primary processes (1) and (2); uncertainties in the extent of the collisional
quenching of excited CH2O at the wavelengths greater than 3350 A where only reaction (2) is important,
require k2 - ka - kj.
175
-------
CO
O
c
E
0 20 40 60 80
Solar zenith angle, degree
Figure ^9. Variation with solar zenith angle of the
apparent first order rate constants for sunlight
absorption by CH20 (k ) and the decomposition of
CH20 by primary process (1): CH20 + hv ->- H + HCO
(l); values shown are estimated for the lower
atmosphere employing Peterson's actinic irradiance
estimates,24 assuming no reflection from the earth's
surface.
176
-------
in the lower atmosphere is 111101660: at present. Hie recent observations of
Osif and Heicklen26 suggest that %/R,- s 5 in air at 1 atm pressure, while
Niki, et al.j suggest that R/Rc > 9 for these conditions. However our
recent studies of the CBeO-02 system have disclosed some hitherto unforseen
complications in the mechanism of the formic acid and CO formation in the
irradiated CH20-02 system, and it appears that this rate ratio may be much
lower than the other recent studies suggest. Further work is now underway
in our laboratory to derive this important rate constant ratio for atmo-
spheric conditions .
177
-------
1. 3. Heieklen, K. Westberg, and N. Cohen, Pennsylvania State Center Air
Environment Studies, Publication No. 115-69 (1969).
2. K. Westberg, N. Cohen, and K.W. Wilson, Science, 171, 1013 (1971).
3. J.G. Calvert, J.A. Kerr, K.L. Demerjian, and R.D. McQuigg, Science,
175, 751 (1972).
h. K.L. Demerjian, J.A. Kerr, and J.G. Calvert, Advan. Environ. Sci.
Technol., 4, 1 (197M-
5. J.J. Bufalini and K.L. Brubaker in "Chemical Reactions in the Urban
Atmospheres", C.S. Tuesday, Ed., Elsevier, Amsterdam, 1971, P« 225.
6. J.T. Peterson, K.L. Demerjian, and K.L. Schere, "Actinic Solar Flux and
Photolytic Rate Constants in the Troposphere", in Proceedings, Vol. 2,
International Conference on Photochemical Qxidant Pollution and Its
Control, EPA Report 600/3-77-OOlb, January, 1977, P 763-
7. R.D. McQuigg and J.G. Calvert, J. Amer. Chem. Soc., 91, 1590 (1969).
8. E. Gorin, J. Chem Fhys., 7, 256 (1939).
9. R. Klein and L.J. Schoen, J. Chem. Fhys., 24, 109^ (1956).
10. B.A. DeGraff and J.G. Calvert, J. Amer. Chem. Soc., 89, 22*4-7 (1967).
11. H.P. Sperling and S. Toby, Can. J. Ghem., 51, lj-71 (1973).
12. R.S. Lewis, K.Y. Tang, and E.K.C. Lee, J. Chem. Phys., 65, 2910 (1976).
13. K. Tadasa, N. Imai, and T. Inaba, Bull. Chem. Soc. Japan, ^9, 1758 (1976)
14. J.H. Clark, Ph.D. Thesis, "Laser Photochemistry and Isotope Separation
in Formaldehyde", University of California, Berkeley, 1976.
15. J. Marling, J. Chem. Phys., 66, ^200 (1977).
16. R. Walsh and S.W. Benson, J. Amer. Chem. Soc., 88, ^570 (1966).
17. P. Warneck, Z. Naturforsch., 26A, 20^7 (1971).
18. M. Venugopalan and K.O. Kutschke, Can. J. Chem., k-2, 2^51 (196^).
19. A. Horowitz and J.G. Calvert, Int. J. Chem. Kinet., submitted for
publication.
20. R.S. Le-wis and E.K.C. Lee, referenced by Lewis, et al. [20].
178
-------
21. A. Horowitz, J.G. Calvert, and F. Su, Int. J. Chem. Kinet., to be
submitted.
22. a) E.S. Young and C.B. Moore, J. Chem. Phys., 58, 3988 (1973); b) Ibid.,
60, 2139 (197^).
23. F. Su, J.W. Bottenheim, H.W. Sidebottom, J.G. Calrert, and E.K. Damon,
Int. J. Chem. Kinet., in press.
2k. J.T. Peterson, "Calculated Actinic Fluxes (290-700 ran) for Air Pollution
Photochemistry Applications", EPA-600/1|~76-025, 1976.
25. P.A. Leighton, "Photochemistry of Air Pollution", Academic Press, New
York, 1961.
26. T.L. Osif and J. Heicklen, J. Phys. Chem., 80, 1526 (1976).
27. H. Niki, P. Maker, C. Savage, and L. Breitenbach, Abstracts, 173rd
American Chemical Society Meeting, New Orleans, La., March 20-25,
1977.
179
-------
Al; 'JrlUSUAL H2-FORMING REACTION IN THE 3130A PHOTOLYSIS OF CH20-02 MIXTURES
AT 25°C.
Introduc tion
The chemistry and photochemistry of the formaldehyde-oxygen system have
been studied extensively for a number of years.1"16 The importance of CHgO
oxidation in hydrocarbon combustion and the potentially important role of
CH20 photooxidation in the chemistry of the atmosphere have helped to renew
the interest in this system in recent years. However several important
aspects of the photochemistry of the CH20-02 system remain unexplained: the
detailed mechanism of formic acid formation,14?16 hydrogen formation in
quantum yields that apparently exceed unity;9 and the detrimental effect of
oxygen addition to HDCO on the isotope enrichment of the products of the
laser photolysis of HDC0.18s19
It is well established that pure CH20 vapor photoexcited within its
first electronic absorption band (2^00-3800 A) decomposes by way of the
primary processes (l) and (2), and CO and H2 products are formed ultimately
in about equal amounts through the reactions (3), (^-), and (5) in experi-
ments near 25°C:
CH20 + hv -*- H + HCO (1)
CH20 + hv -* H2 + CO (2)
H + CH20 ->- H2 + HCO (3)
2HCO -*• H2 + 2CO (k)
2HCO +- CO + CH20 (5)
02 addition to CH20 leads to several additional products since 02 will
scavenge H-atoms and HCO radicals20"22 rather effectively. Indeed, in ad-
dition to H2 and CO products of CH20-02 photolyses, the formation of formic
acid, C02, and H20 have been observed, and the possible products, peroxyformic
acid14?15 and H202 [Ref. 2] have been inferred from indirect methods of
analysis. Recent results of Niki, et al.16 suggest strongly that peroxyformic
acid is not formed in the photooxidation of CH20. In view of the fact that
C02 is an unimportant product of the reactions in this system, the mechanism
of formic acid formation through the HC02 radical reactions is unclear; one
might expect the highly exothermic reaction, HC02 + 02 —^ H02 + C02, to
compete successfully with formic acid formation: HC02 + CH20 -*- HC02H + HCO.
The early work of Horner, et al.9 showed that the quantum yields of
formation of H2 and CO as well as formic acid were above unity in CH20-02
mixture photolyses at temperatures > 100°C. This observation seems to have
attracted little attention in the years since its publication. In view of
the relative stability of the HCO radical and the efficient scavenging of
the HCO radical by oxygen, a chain sequence for H2 formation in this system
is unexpected; certainly confirmation of this unusual observation is re-
quired.
180
-------
Another interesting effect of 02 addition on CH20 photolysis has been
observed. Oxygen may change the efficiency of the primary steps (i) and (2)
as Houston and Moore have suggested.23 They found that the quantum yield
of vibrationaUy excited C0(v = 1) significantly increased in the laser
photolysis of CH20 at 3055 A in the presence of small amounts of 02. This
result was attributed to the 02-enhancement of the molecular reaction mode
(2) of CI^O photodecomposition. It is of both theoretical interest and of
great practical value to establish the mechanism and importance of this
process in order to evaluate adequately the rate of H-atom and HCO radical
generation in the earth's 02-rich atmosphere.
Our interest in the CH20-02 system was stimulated by the several un-
answered problems which we have reviewed, and it was heightened further
by a most unexpected result which we obtained in another recent study of
formaldehyde photolysis.24 We redetermined the primary quantum yields of
free radical (l) and molecular (2) processes in photolyses at 3130 A.
Various free radical scavengers were employed as potential reactants for H
and HCO radicals in an attempt to determine $0 from $ values in experiments
e- rig
at high scavenger concentrations. Isobutene, nitric oxide, and trimethyl-
silane were found to be effective reactants, and CH20 photolyses in mixtures
with each of these compounds gave useful data related to $, and . However
we were very surprised to find that the addition of small amounts of 02 to
CH20 did not suppress the hydrogen quantum yields in experiments at 25°C.
On the contrary $ significantly increased to values as high as $ = 6,
n2 ii2
yet C02 was only a minor product in these experiments. Thus the unusual
observations of Horner et al.9 made in experiments at temperatures of 100°C
and above, are confirmed for CH20-02 photolyses even at 25°C. It was ap-
parent that new kinetic information was necessary to rationalize these
unusual findings. The present work was initiated to attempt to answer these
many puzzling questions related to this interesting system. The results
presented here provide some of the answers.
Experimental
Materials--
Formaldehyde was prepared from paraformaldehyde (Eastman) according to
the procedure of Spence and Wild25 and stored in a trap cooled to dry-ice
temperature. Isobutene (Phillips) and C02 (Matheson) were purified by trap-
to-trap distillation. As a further precaution against moisture introduction
to the cell, each time that C02 and 02 were introduced, their storage bulbs
were cooled first to dry-ice temperature and then the gases were expanded
into the cell.
Procedures and Product Analysis--
The experimental apparatus was the same as that described in our previous
study.24 In the present work actinometry was based upon the photolysis of pure
CH20 at 25°C, accepting §j^=1.10 as determined for these conditions in our
earlier work. Any polymer and acid which deposited on the walls of the cell
as the CH20-02 photolysis progressed in each experiment, was removed between
runs by warming the cell to about 110°C while evacuating it with the high vacuum
181
-------
system. The technique of gas transfer from, the cell and chromatographic
procedures were somewhat modified from those of our previous work. The
products E2) CO, and C02 and the amount of 02 which reacted were directly
determined in this study. The Molecular Sieve 5A column held at 30°C and
with !J2 as carrier gas was employed here to determine H2 in the presence of
relatively large amounts of oxygen. The same column held at 100°C and with
helium as carrier gas was used in 02 and CO determinations. In these
analyses at least four determinations of the sample were made to increase
the accuracy of the 02 disappearance measurement. In a few experiments C02
was estimated by Toepler- pumping the product gases through a glass bead -
containing trap immersed in pentane-dry-ice slush. The untrapped gases
were then analyzed for C02 using a Poropak Q. column held at 100°C and He
carrier gas. The C02 quantum yields were found to be very small; in ex-
periments with 8 Torr of CH20, $> = 0.21, 0.11, and O.l6 (± hQPjo uncertainty)
with P_ = 1, 2, and k Torr, respectively. Thus this analysis procedure
02
was not continued in the later experiments. In most runs only the gases
which did not condense at liquid nitrogen temperature (Hg, CO, and 02) were
determined. Duplicate runs were carried out for each set of experimental
conditions. The gases from one of the runs were analyzed for H2 while CO
and 02 were determined in the other run. The total pressure of these gases
was also measured in each run as an independent check on the overall re-
producibility of the system.
The key factor which limited the accuracy of the quantum yield measure-
ments in this work was the estimation of the final and the average form-
aldehyde concentrations during the run and the number of quanta of light
absorbed during the reaction. In theory the final P_TT _ could be estimated
from the measured absorbance at the end of the run; for our conditions
this would amount to a 15-20% change from the initial value. However the
addition of 02 in the CH^O photolyses leads to a much lowered stability of
the monomer CH20; apparently acid and H20 products of the CH20-02 mixture
photolysis catalyzes the polymerization of the monomer, even during very
short light exposures. As a result the transmitted light intensity was not
a true measure of the gaseous formaldehyde absorption as the run progressed.
Thus the formaldehyde remaining after a run was estimated by two other less
direct methods: (1) the material balance of the products; and (2) the
thermal jump in the pressure which was observed at the start and the finish
of the irradiation. In the former method we have assumed that HC02H is
the only acid product as Mki, et al.16 have shown recently. We also as-
sumed that H20 (and not H202) is a final product and have estimated the
chemical loss of formaldehyde during the run (other than polymerization)
from the mass balance relation: $ = $ + 2$ . . The amount of CH?0
— Lxl2U Jrl2 ~"02
that polymerized was obtained from the recorded pressure-time profiles of
the reaction. Typical pressure records are shown in Figure50 for the
CH20-02 system (curve A) and the CH20-02-C02 system (curve B) . The sudden
jump in pressure (point a) as the photochemical reaction began resulted
largely from the very exothermic nature of the overall oxidation reactions
which occurred as well as the ultimate conversion of the absorbed light
energy to heat. The former source is dominant here. After a short in-
duction period following the initial exposure, the pressure dropped at a
182
-------
0
Time(A), sec
50 100 150
o
309.2H
-------
rate which was essentially constant during the entire irradiation period.
As the light was extinguished a sudden pressure drop was observed (point
b In Figure 50) • This was followed by a further decrease in pressure which
proceeded with a rate equal to or lower than the rate observed when the
ceil was irradiated. The amount of products formed was found to be in-
dependent of the time elapsed between the end of the irradiation and the
tine that the gases were transferred to the analytical system. Evidently
the acid- and water- catalyzed polymerization of CH20 was the sole origin
of the pressure drop after the irradiation. Since the pressure in the cell
decreased nearly linearly throughout the photochemical run, the rate of
polymerization had to be essentially constant and equal to the rate observed
immediately after the thermal jump at the end of the run. This rate was
taken from the pressure- time profile and multiplied by the irradiation time
to determine the amount of CH20 which polymerized during the run; this
together with the estimated amount of CHgO which reacted chemically, gave
the final CHeO pressure using the mass balance method.
The thermal jump in pressure provided an independent method of estimat-
ing the final CHgO pressure at the end of the run in our experiments with
added C02. The large excess of C02 insured that the heat capacity of the
reactant mixture was essentially unchanged during the run. In a typical
experiment at initial pressures: P. = 1, P^TT ^ = 8, and P = 155 Torr,
(Jp Ln2U UU2
the heat released per mole of quanta absorbed was about l46l kcal.26>27
Only about 91 kcal mole"1 of quanta would be released from the conversion
of the light energy to heat and about 151 kcal mole"1 from polymerization;28
the latter energy probably would be released at the wall of the cell. The
equilibrium between heat generating and heat dissipating processes at the
wall is reached rapidly at a temperature which is slightly higher than the
initial temperature. Our studies of the thermal jump observed in irradiated
S02-isobutane systems29 have shown that when the heat capacity of the system
is relatively unchanged during the run, the pressure jump is directly pro-
portional to the heat generated in the cell as a result of the light ab-
sorption and subsequent chemistry. This condition is met in our system at
high C02 pressures. The results of Tables 28-30 show that within a narrow
pressure range the rate of the chemical reaction was directly proportional
to the absorbed light intensity and independent of the PQ . Thus in the
CHpO-02-COp system, the initial and final pressure jumps should be pro-
portional to the rate that light was absorbed by CH20 at these points in the
reaction. Since the limiting form of Beer s law applies to the CH20 ab-
sorption for our conditions, the ratio of the initial and final pressure
jumps could be used to estimate the final PQJJ Q. The average value of
PCHPO calculated by both the product balance method (column 1, Table
ancTthe pressure jump method (in parentheses, column L, Table
re;u'.on.'i.bly well, ;ui
-------
TABLE 28. EXPERIMENTAL DATA FROM THE PHOTOLYSIS OF CH20-02 MIXTURES AT 3130A AND 25
o,ia
Pressure of react ant, Torr Irrad.
CH20 02 02 timo' sec
(aver) (init) (aver)
7 Oft
7.92
7 67
7.92
7.84
7.88
7.76
7.75
7.49
7.49
7.49
7.21
7.94C
7 ^1
7.O2
7-05
7. SI
7.02
7.23
7.76
7.00
0.050
O 12O
O.15O
0.150
0.300
O.50O
0.500
1.00
l.OO
l.OO
1.00
l.OO
2.0O
2.00
4.00
4.0O
4.OO
8.00
8.00
0.035
0.131
0.119
0.277
0.453
0.44?
0.887
0.885
0.882
0.774
O.Q7Q
1.77
1.78
3.68
3.79
7.9
7.71
2O
1 qn
30
60
30
75
75
150
150
150
3OO
l CQ
300
300
7C
3OO
3OO
1OO
4oo
Quanta ahs./ S^
cell-sec, x
10-15
7.02
6 80
7.02
6.95
7.01
ft n-»
6.88
6.87
6.64
6.64
6.64
6.50
7.O4
6 66
6.22
6.26
6 67
6.22
6.41
6.88
6.21
i aM
4.50(4.
™ f\Q(
(4.
4.42(4.
4.37(4.
L /.of
4.26(4.
(4.
3.75(3.
(3.
(3.
(3.
o 33(0
(2.
(2.
2.11(2.
(2.
1.44(1,
1.41(1.
*CO
)b , ^
79)
\
66)
63)
45)
\
20)
2O)
67)
68)
68)
66)
•10)
\
91)
91)
\
14)
16)
48)
40)
7.01(6.54)
/ \
7.07(6.41)
6.77(6.38)
6.66(6.21)
t \
6.15(5-95)
6.35(5.95)
6.00(5.40)
5.61(5.41)
5.51(5.41)
5.44(5.35)
i 15(1 oo)
5.03(4.60)
4.82(4.61)
/ \
4.02(3.87)
3.80(3.92)
3.19(3.37)
2.85(3.16)
«-02
' _ ( ~)
3.60(4.19)
/ \
2.98(4.19)
2.53(4.15)
3.75(4.17)
/ \
3.08(4.11)
3.50(4.11)
3.83(3.98)
3.90(3.98)
4.00(3.98)
3.92(3.84)
O 69(0 68)
4.18(3.75)
3.89(3.77)
/ \
5.75(3.75)
3.71(3.86)
(4.11)
3.90(3.74)
*C02
(0.25)
/ \
(0.25)
(0.24)
(0.24)
i \
(0.22)
(O.22)
0.21(0.20)
(0.20)
(O.2O)
(0.20)
/ \
. .
0.11(0.10)
(0.16)
/ \
0.16(0.12)
(0.12)
(0.09)
(0.09)
R,
>
0.99
0.96
0.96
0.92
0.87
0.88
0.78
0.78
O.80
0.8O
0.63
0.63
0.45
0.44
____
0.26
«H2- ^2
R^ + Rfi
4.2
-__
4.3
4.4
4.5
___
4.4
^
4.5
4.0
«--
- —
k.2
alnitial PCH,,Q = 8.OO Torr. The values in the parentheses were calculated by computer solution of the rate
equations based upon the reaction mechanism and rate constants given in Table IV. Isobutene at 2.O Torr was
added in this experiment.
-------
TABLE 29. THE EFFECT OF ABSORBED LIGHT INTENSITY ON THE PRODUCT QUANTUM
YIELDS IN THE PHOTOLYSIS OF CH20-02 MlXTURESa
Pressure (average) Irrad. time, Quanta abs./cell- $jj
Torr sec sec, x 1O
*CO
'->
A) PgH 0 = 8.0O; PQ =1.00 Torr
7.96
7.67
7.63
7.63
7.64
7.40
7.14
0.980
0.971
O.94o
0.938
0.938
0.863
0.827
B) PCH20 = 8-°°5 P(
7.4o
7.47
6.60
7.26
O.818
0.877
0.594
0.754
10OO
1000
770
770
770
450
4oo
>2 = LOO;
1500
6OO
770
300
2.70
4.51
11.3
11.3
11.3
26.6
38.3
PC02 = "7 Torr
4.39
11.1
24.2
38.8
2.1O
2.22
2.58
____
„__
2.84
3.19
4.49
5.02
5.79
5.19
3.55
4.1O
4.O5
3.89
3.89
4.75
5.45
7.22
8.11
8.69
8.26
2.50
2.15
2.35
2.43
2.4l
3.90
3.84
6.81
6.27
7-41
7.18
Excitation was at 3130 A; temperature, 25 C.
186
-------
TABLE 30. EXPERIMENTAL DATA FROM THE PHOTOLYSIS OF CH2&-02-C02 MIXTURES AT 3130A
AND 25 Ca
Pressure reactants, Tbrr Quanta abs./cell- 9., IQQ *-Oa
(average) sec, x 1O~*5 2
CH2Ob 02 C02
7.24(7.24)
7.21(7.35)
7.24(7.36)
7.22(7.41)
7.28(7.35)
0.775
0.769
0.781
0.744
0.754
"initial pressures
55
117
117
155
300
of reactants:
6.68
6.6?
6.68
6.67
6.62
P =8
6.05
5.53
5.42
4. 02
.00; P
(5.6l)c
(5.21)
(5-20)
(5.10)
(4.01)
=1.00 Torr;
8.62
8.79
8.74
8.60
6.69
(8.46)°
(8.32)
(8.32)
(8.58)
(7.38)
irradiation time,
7
7
7
8
8
15O sec
.62
.86
.44
.68
.34
(7.79)c
(8.72)
(8.76)
(9.28)
(9.79)
in all experiments.
(0.39)°
(0.46)
(0.46)
(0.52)
(0.56)
"the
two values shown here were estimated by the mass balance method and the pressure jump metliod (in parentheses) as
described in the text. Calculated by computer from the reaction mechanism and rate constant data given in Table IV;
k7 was assumed to be 2.0 x 10 i. mole" sec~ for M = CO .
-------
the r-ost puzzling results is the formation of H2 in a chain mechanism; note
in Tables 28-30 that the quantum yields of hydrogen are as high as 6 in some
of these experiments at 25°C. One anticipates that formic acid, or possibly
peroxyformic acid, carbon monoxide, and carbon dioxide may be formed in
chain reactions following primary process (1) in this system through the
reactions (6)-(12):
CH20 + hv -*- H + HCO (1)
H + CH20 •+- E2 + HCO (3)
H + 02 + M •*- H02 + M (6)
HCO + 02 + M •*- HC002 + M (?)
HCO + 02 ->• H02 + CO (8)
HC002 + CH20 HC002H + HCO (9)
2HC002 •>- 2HC02 + 02 (10)
HC02 + CH20 •+- HC02H + HCO (11)
HC02 + 02 -»- H02 + C02 (12)
Apparently peroxyformic acid is not formed in significant amounts in this
system,16 so we have neglected reaction (9) in. our further considerations.
However conventional reaction schemes cannot rationalize H2 formation in a
chain for experiments at 25°C. The formyl radicals formed in the CH20-02
system will react almost exclusively with 02, even in experiments at P_ -
U2
0.05 Torr, and H2 formation through the reaction (4) which may occur in 02-
free CH20 photolyses at 25°C, is unimportant here.21'22'30'31 However H2
formation through reaction (3) is expected to compete with the H-atom
reaction (6) with 02 for the relatively low [02]/[CH20] ratios which we
employed. Only a small fraction of the H-atoms formed in this system will
be captured by 02 in experiments at low added 02 pressures; most of them
will react with CH20 in reaction (3).* Note in Table 1 the rate ratio
R0/(Ro + R/;); in experiments with P° = 8 Torr, the ratio decreases from
33 o On2u
0.99 with P° = 0.050 Torr to 0.26 with P° = 8 Torr. However without some
Op O2
*In making these rate estimates we have used the following rate constants:
k = 2.8 x 10T £. mole-1sec-1 [Ref. 20], k^ = 1.64 x 1011 (M = CH20; assumed
equal to that for M = CE^ [Ref. 20e]), and 2.1 x 1010 ^.2mole~2sec~1 (M =
02 [Ref. 20a,20e]).
188
-------
chain reaction which regenerates H-atoms or H2, the 3> resulting from the
HP
sequence (1)-(12) must equal 5> + $ R /(R + RA), 1.00 at P° = 0.05 Torr
^ -^ o j o 02
and 0.50 at P° =8 Torr. Obviously some undefined reaction steps must
regenerate H-atoms to explain the much higher $ values actually observed
H£>
for these conditions: 4.5 at P° =0.05 and 1.4 at P° =8 Torr. Observe
2 °2
in Table 28 that the ratio of the quantum yield of H2 .exclusive of H2 produced
in the primary molecular process (2), $H - $2, to the fraction of H-atoms
formed which lead to H2,R~/(R + Rg), is essentially a constant for all of
the conditions employed in Table 28. This suggests that the unknown H-atom
generating species provides a large and fairly constant source of H-atoms
which has about the same rate in runs at both low and high pressures of added
02; in experiments at the higher 02 pressures, H2 formation in (3) is
ultimately attenuated through the increasing importance of reaction (6).
The reactive species which generates H-atoms cannot be the HCO or the
HC02 radicals. The conventional reactions (l), (3)5 and (13) cannot
constitute a significant chain, since reaction (13) is very slow compared
to (7) and (8) even at very low pressures of 02:
HCO + M -*- H + CO + M (13)
HCO + 02 + M •*- HC002 + M (?)
HCO + 02 -*- H02 + CO (8)
Taking k [M] + kg = 3.4 x 109 &. mole"1sec-1 [Ref. 21] for an experiment
with P_ = 0.05 and P__ _ = 8 Torr, R,Q/(^ + Ba) 5 x 10'10 [Ref. 24]. It
U2 UngU Jo ( °
is equally unlikely that the H-atom generating species is the HC02 radical,
since its reaction with 02 in (12) or its decomposition in (1^) would
generate C02 in amounts equal to those of H-atoms; experimentally it is ob-
served that « *„ .
xl2
HC02 + 02 •*- H02 + C02
EC02 (+M) -*- H + C02
It is clear that the obvious conventional reaction steps which one might
invoke are inadequate to account for the present results.
One reaction sequence which seems consistent with our findings, in-
volving an H02-H0 radical chain (15) -(19) and radical addition to the C-atom
of the CH20:
H02 + CH20 -*- (H02CH20) -^ HO + HC02H (15)
HO + CH20 +- HaO + HCO* , (l6)
189
-------
HO + CH20 -*- H20 + HCO (17)
HO + CHaO -*• (HOCH20) -*- HC02H + H (18)
- H + CO (19)
The reaction (15)? exothermic by 60 kcal/mole overall, is somewhat analogous
to the overall reaction proposed by McKeller and Worrish10 in the flash
photolysis of CH20-02 mixtures at temperatures li|-OoC: H02* + CH20 -*- H20 +
CO + HO; here H02* is a vibrationally-rich H02 radical. Reaction (15) ap-
pears to require a much less complicated rearrangement of the reactants,
and hence it appears to us to be a more realistic choice. In reaction (15),
presumably the simultaneous shift of an H-atom on the carbon atom to an
0-atom must accompany the rupture of the HO- bond as this reaction occurs.
Reaction (18) is exothermic by 22 kcal/mole and requires rupture of single
C-H bond following HO addition to CH20 to form a formic acid molecule and
an H-atom. One might speculate that H-atoms may result as well from a
vibrationally excited HCO product of the reaction (l6). Hie reaction (17)
is exothermic by 33 kcal/mole, and it is conceivable that a fair fraction
of the reaction will proceed by (16) in which the initial HCO product will
contain more than the ~l8 kcal/mole necessary for the decomposition of the
radical:
HO + CH20 •*• KC+ H20 (16)
HCO*-^ H + CO (19)
HCO* + M -*• HCO + M (20)
Indeed without reactions (16) and (19), all of the CO molecules formed in
CH20-02 mixture photolysis at 25°C must be generated in reactions (2) and (8),
and for these conditions one expects $ - $ < fl> = 0.68; this is
CO Hg J_
contrary to the experimental results. By analogy with the proposed H02
radical addition to CH20, reaction (21), the addition of the peroxyformyl
radical to CH20, and the subsequent rearrangement by H-atom transfer should
be considered as well:
HC002 + CH20 •*- (HC002CH20) -*- HC02 + HC02H (21)
The relative insensitivity of 5> to variations in the absorbed light
HQ
intensity (Table 29) points to the first order termination of the radical
chains at the wall of the reaction cell for our conditions. A reaction
such as (22) is suggested: •
H02 + Wall -*• Products (H20,02) + Wall (22)
Presumably the 3> values increase with added C02 (Compare 0 values in
•"•2 ns.
sections A and B in Table 29.) as a result of the slowing of the diffusion
of H02 (and/ or other chain carried radicals) to the wall and the increased
length of the chains which would result for these conditions.
190
-------
Although the nature of the chain source of H-atoms must remain unclear
at this time, it appears to us that the reactions (18), (l6) and (19) are
attractive candidates. Whatever the nature of the chain carrier, it is
clear that it and/or its precursor radicals react rapidly with added butene.
Thus in the experiment involving an initial mixture of 8 Torr of CH20,
1 Torr of 02, and 2 Torr of isobutene shown in Table I (footnoted with a
"c"), the residual $„ found (0.33) is equal within the experimental error
to that estimated earlier in binary mixtures of CH20 and isobutene at 3130 A;
this limiting value of $ was attributed to the quantum yield of EQ from
the molecular process (2). The transients H, HO, and H02 which are involved
in our sequence would react with isobutene as these experiments require.33
The proposed reaction sequence including reactions (15)-(19) and (21)
also provides a realistic alternative to the HC02H formation in the CH20-02
photolysis system. It avoids the apparent contradiction between the required
rapidity of the HC02 reaction (11) with CH20 and the required slowness of the
highly exothermic reaction (12) of the HC02 radical with 02(AH,g = -53 kcal
mole"1) which seems necessary to avoid significant C02 formation in the
system.
Using the steady state assumption for the transient radicals HO, H02,
etc., in our system and the reaction mechanism (l)-(3), (6)-(8), (11), (12),
(15)-(19), (21) and (22), the relation (A) can be derived:
•s.
1,2'*16
k
(A)
For our experimental conditions [CH20] is approximately constant (average
PCH 0 = 7.5 i 0.5 Torr), and hence in terms of the mechanism outlined, the
function ($ - *2)/[Ro/(Eo + Rg)lj should be approximately constant as is
observed experimentally; see Table 28.
A more quantitative test of the compatibility of the proposed mechanism
with the present data will be made after a consideration of the mechanism
of the HCO reactions with 02 in the next section.
The mechanism of the formyl radical reactions with 02. There has been
considerable interest in recent years concerning the relative importance
of the two competing reactions (7) and (8) of the HCO radical with 02:
HCO + 02 + M -*- HC002 + M (7)
HCO + 02 -*- H02 + CO (8)
Direct measurements at low pressures suggest that the HCO reaction with 02
is in the second order region and that reaction (8) is dominant.21'22'30
However in the recent study of the Cl-atom sensitized decomposition CH20-02
191
-------
mixtures, Osif and Heicklen14'15 estimated the ratio of Ry/Rg = 5 ± lj
independent of the pressure in the range 60-700 Torr. Niki, et al.,16 came
to the conclusion that R?/Rg > 9 in similar C12-CH20-02 studies at atmo-
spheric pressures. However the conclusions of Osif and Heicklen and those of
I'liki, et al., must be reexamined in view of the present findings. Both
research groups have made the reasonable assumption on the basis of the data
then available that the rates of HC02H and CO formation can be taken as
measures of the rates of reactions (7) and (8), respectively. They have
assumed that all HC002 radicals formed in (7) will ultimately form HC02H,
presumably through the reactions (10) and (11). However in view of the
present findings, other sources of HC02H in addition to HC002 radicals formed
in (7) are likely.
The absence of peroxyformic acid among the products of the photooxidation
of CH2016 and the very exothermic nature of the HC02 reaction with 02, sug-
gest to us that C02 formation may represent a large fraction of the ultimate
products formed following the HCO radical addition to 02, reaction (7). The
possible direct reaction of HCO with 02 to give C02, HCO + 02 -*- HO + C02,
is considered here to be an unimportant source of C02. Assuming the mechanism
outlined and the inequalities R, ? » R-,7j we may estimate a lower limit to
the ratio, R7/(R7 + Rn) through the approximate relation (B):
R $
•p ~ rfi i rf> rh \&)
R + R^ $ + $ - 5>
7 8 H2 1 2
From our data obtained in experiments at 8 Torr of CH20 and 1, 2, and 4 Torr
of 02, we estimate R?/(R? + Rg) ^ 0.051, 0.031, and 0.065, respectively.
The CH20 molecule should be a much more effective M in reaction (7) than 02,
so that for our conditions R?/(R7 + Rg) ~ k7[CH20]/(k7[CH20] + kg) = 0.049 ±
0.017 for [CH201 = 4.0 x KT4 M. Taking kg = 3.4 x 109 1. mole-1sec-1 [Ref.
21], we find the lower limit for the rate constant k > (4.4 ± 1.6) x 1011
•0.2mole~2sec~1. This estimate for k is in reasonable accord with those for
other third order reactions of similar complexity of reactants and similar
exothermicities (AH? = -39 kcal mole"1): HO + S02 + N2 -*- HOS02 + N2, AH =
-37, k = (2.6 ± 1.0) x 1011 [Ref. 34] j NO + HO + N2 -*- HOM) + N29 AH = -36
kcal mole'1, k = (2.1 ± 0.4) x 1011 ^.2mole~2sec~1. Obviously the quantum
yield of C02 which would give a clearer picture of the pressure dependence
of R7/Rg, could not be determined from the present experiments with added
C02. The only measurements of $ in CH20 photooxidation at high pressures
are those of Osif and Heicklen15 who found $ values ranged from 0.08 to
C02
0.31 in the presence of 81-458 Torr of N2, 21-122 Torr of 02 and 1-10 Torr
of CH20. This is the magnitude of the $ values which we predict for the
higher pressures from the present mechanism (Table30). Since k for M=
192
-------
N2 would be smaller than that for C02, the $ values found by Osif and
OUg
Heicklen should be smaller than those presented here.
The relative dominance of reaction (8) compared to (?) for the P°
o . CH2 0
« iorr experiments at low added 02 pressures is also consistent with the
results of our experiment in which 2 Iorr of isobutene was added to 8 Torr
of CH20 and 2 Torr of 02 (the run footnoted with a "c" in Table I). In this
run about 91% of the H-atoms formed should add to isobutene,24 so that $
should be very close to $g = 0.32 [Ref. 2k 1. Indeed we found 4 = 0.33.2
¥e also found in this experiment that 3> = 1.15 and * _ = 0.69? for these
_ v(J ~02
conditions and Rg » R7, we expect $CQ s i.o = $ + and f = $ = 0.68.
If R^ were equal in magnitude or greater than Eg, then we would expect $ <
*-]_ + *2- Obviously the present data appear to require that Ro » R7, at
least in the 8-12 Torr pressure range employed in these experiments.
Although the present data do not establish an accurate estimate of k?,
they do lead to a theoretically reasonable value and point to some un-
recognized complications in other recent studies of this system. If we
accept the true k7 value as the lower limit of this constant which we have
estimated here (k_ = k.h x 1011 ^.2mole"asec~i), and we assume that the
relative efficiency of N2 and 02 as M in (7) is about one-quarter of that
for CH20 and that the third order dependence of reaction (7) extends to a
pressure of 1 atm. of air, then we predict that about equal fractions of HGO
radicals formed in the atmosphere will react by (7) and (8). Obvisouly
further more definitive experiments are required to redetermine accurately
the pressure dependence of the important rate ratio, R7/(R7 + Rg). These
are now underway in our laboratory.
Tests of the compatibility of the present mechanism with the ex-
perimental data obtained in this work were made using reasonable choices
for the rate constants and a computer solution of the differential equations.
The rate constant choices are summarized in Table 31 together with the
references and reasoning used in arriving at the rate constants used. The
quantum yields of the various products which are calculated using these
constants are shown (in parentheses) with the experimental 0 data in Tables
1 and 3. Note that there is reasonable accord between the experimental and
calculated quantum yields. Obviously the mechanism suggested is compatible
with the present data, and it appears to be deserving of further detailed
testing.
The Effect of 02 Addition on the Laser Isotope Enrichment in C1IDO--
Clark18 and Marling3-9 have studied the laser photolysis of CJJDO at 3050A
and 3llK)A, respectively, in the absence and presence of small amounts of
added 02. Both workers have found that the addition of 02 to CHDO sharply
increased the H2/HD and HD/D2 ratios in the products. Thus Marling found
the addition of 02 (2 Torr) to CHDO (k Torr) increased the H2/D2 ratio to
193
-------
TABLE 31. SUMMARY OF THE REACTION MECHANISM AND RATE
CONSTANTS EMPLOYED IN THE SIMULATION OF THE CH2Oj 02)
AND C02 MIXTURE PHOTOLYSES
Reaction
Rate constant
Reference
(1) CH2O + hi> -» H + HCO
(2) CH20 + hi> -* H2 + CO
(3) H + CH20 -» H2 + HCO
(4) 2HCO -* H2 + 2CO
(5) 2HCO -» CO + CH2O
(6a) H + 02 + CH20 -* HO + CH2O
(6b) H + 02 + 02 -* H02 + 02
(6c) H + 02 + C02 -»• H02 + C02
(7) HCO + O2 + M -»• HCOO2 + M
(8) HCO + 02 -* H02 + CO
(21) HCOj + CH 0 •*• HCOgH + HC02
(10) 2HCO- •*- 2HC02 + 02
(11) HCO + CH20 -*• HC02H + HCO
(12) HCO 1 + 02 -* H02 + C02
(15) HO + CH20 -» HO + HCOoH
(16) HO + CH2O -*- H2O + HCO*
(17) HO + CH20 -*• H20 + HCO
(13) HO + CH^ -*• HC02H + H
(19) HCO* -*• H + CO
(22) H02 + wall •»• Products(H90,Op) + wall
3.68 x 1O-5
1.73 x 10-5
2.8O x 10'
0.32 x lO1
1.88 x 1O
1.64 x 10n
2.10 x 1010
2.50 x 1O10
4.2 x 10X1(M
3.40 x 10^
5.35 x 10
1.50 x 1O
8.6O x 1O5
3.0O x 10^°
7.30 x ur
2.92 x K>9
3.44 x 107
5.64 x 10^
7.0O x 10
11 - 4.1
|>6lb
P6]b
C20J
(24J
f24]
L20eJC
d
= CHJO) This work
L21J
e
e
e
e
L17bJ
[37] f
[37] f
p7]f
g
h
Reactions (l), (2), (19), and (22) are first order (sec"1); reactions (6a), (6b),
(6c), and (7) are third order (J.2mole"2sec-1), and all others are second order
(t. mole" sec"1). ^Values for kj and k2 were calculated from the measured light
intensity and ^ and $2 values of ref. (JJ6]. cSee footnote 1 of the text.
value was chosen to optimize the fit of the data in Table III. eThe value of
k21 was picked to be consistent with that taken for the analogous reaction
(15); kit; was assumed to be equal to the total HO2-CH_0 reaction rate constant
(presumably for H-abstraction) estimated in reference [l7b] ; k1Q was assigned
by analogy with the observed rate constant for sec-alkyl peroxy radical reaction,
2R2CHO2 -* 2R2CHO + 02, in ref. [38] » the relative rate constants,
were picked to match best the 5>£o data; the 4jj? and 9(-Q va]
relatively insensitive to the choices of kg, k^Q' kll' and k!2'
kj£ + k _ + kjo was taken as the total measured HO-CHgO rate constant of ref.
[37J, trie relative values were assigned to give the best computer fit to the
data. 9The rate constant kj_ was estimated to give a fast unimolecular dissoci-
ation rate which would be relatively unperturbed by the deactivation step
(20) for the pressures of added gases we have employed here. k22 was assigned
so that a match of the experimental chain length was achieved; values of k22 =
11, 5.2, 4.6, 4.3, and 4.1 sec"1 were used for runs at 0, 55, 117, 155, and 300
Torr, respectively, of added C02.
calculated were
fThe value of
194
-------
70 from its value of 5.^ in the absence of 02. This observation is in ac-
cord with the isotopic discrimination which one might expect with the H-atom
chain reactions observed in this work. A small degree of isotopic discrimina-
tion is expected in the primary decomposition of CHDO in (la) and (Ib) such
that $la > »lbs
CHDO + hv -*- H + DCO
CHDO + hv -»- D -t- HCO
Also the subsequent reactions of the H and D atoms will favor H2 formation in
CHDO mixtures with added 02, since E~ < E0, = E0 < E •
ja 30 3c 3d
H + CHDO -*• H2 + DCO (3a)
H + CHDO •>- HD + -HCO (3b)
D + CHDO -*- HD + DCO (3c)
D + CHDO ->- DS + HCO (3d)
H + 02 + M •*• H02 + M (6a)
D + 02 + M -*- D02 + M (6b)
If the reactions (1) and (3) were the exclusive sources of H2 and D2 in the
02-CHDO mixture photolysis, and sufficient 02 is present to capture all HCO
and DCO radicals (P > 0.05 Torr), then the ratio of E2/~D2 in the products
02
should be given by relation ( C ) :
k3d + k6d M M
Da \k0, / V$1TJU~Q + kQK + k,o [02] M
When the reactions (6a) and (6b) remove a major fraction of the H and D atoms
([02]/[CHDO] > 0.5), then one expects the residual H2/D2 formed to be given
approximately by: (H2/D2) = (k3a/k3d) ($la/$lb) (^b/^a^ ' ¥e may make the
reasonable estimates that Iw = kga, $la/$llD ^ 2 » and k3a/k3d~^' thus ln
the unscavenged H2 and D2 product from CHDO photolysis with several Torr of
added 02, we expect H2/D2<10. The much higher discrimination favoring H2
formation in the studies of Marling and Clark can be rationalized in terms
of the additional hydrogen forming chain reactions considered here. In this
case the intermediate which forms H or D atoms in the chain sequence,
reactions (l6a)-(l6d), (19a), (19^), and (l8a)-(l8d), will also reflect iso-
topic preferences for H-atom formation.
HO + CHDO -*• EDO + HCO* (l6a)
HO + CHDO •+- BsO + DCO* (l6b)
195
-------
DO + CHDO -*• D20 + HCO* (l6c)
DO + CHDO •*- EDO + DCO* (l6d)
HCO*->- H + CO (I9a)
DCO*-*-D + CO (I9t>)
HO + CHDO -*- (HOCHDO) -*- DC02H + H (l8a)
HO + CHDO •* (HOCHDO) -»- HCOpH + D (l8b)
DO + CHDO -*- (DOCHDO) -*- DC02D + H (l8c)
DO + CHDO -*• (DOCHDO) -*- HC02D + D (l8d)
The reactions (l6a)-(l6d) all have very large rate constants and are expected
to show only a small isotopic discrimination favoring DCO over HCO formation.
However we expect the inequalities: kiq > k , k-Q > k-.g,, and k-,0 >
k.o,. Thus it appears that the combined isotopic preferences shown in the
chain sequence for ES, HD, and D2 formation can rationalize the experimentally
observed detrimental effect of 02 on the laser isotope separation in CHDO-02
mixtures. Without such a chain, the effect is difficult to explain.
Marling19 also observed that the addition of 1 Torr of 02 to k Torr of
CH20 with natural isotopic abundance of 13CH20 (1.1$), decreased the isotope
enrichment in the CO product induced by 22EFe laser photolysis at 3323«7 A.
In this case however the decrease was by a factor of 2-3, while in the CHDO
experiments the H2/D2 ratio changed by a factor of 13. The present mechanism
is not inconsistent with these results since CO formation occurs in primary
process (2) rather efficiently at this wavelength ($p =0.8 [Ref. 36]), and
in reactions (8a)-(8d) only a small difference in the rates of the reactions
of the 13C and 12C species is expected:
H13CO + 02 •+- H02 + 13CO (8a)
H12CO + 02 -*• H02 + 12CO (8b)
D13CO + 02 -*- D02 + 13CO (8c)
D12CO + 02 •*- D02 + 12CO (8d)
The Effect of 6~2 Addition on the Primary Photodissociation Steps in Formalde-
hyde--
In their study of CH20 photolysis at 3050A, Houston and Moore23 found
that the addition of WO or 02 to formaldehyde increased markedly the yield
of vibrationally excited C0(v = 1) and the vibrationally relaxed C0(v = 0).
The relative yield of C0(v = 1) was higher for NO addition than for 02 ad-
dition, while the reverse trend was observed for C0(v = 0). The effect was
196
-------
attributed to interactions between NO or 02 and. the vibronicaliy excited
ground state [So(v)] formaldehyde molecules. The decompocition of CHoO by
primary processes (1) and (2) was assumed to occur from So(v), and col-
lisions with NO and 02 presumably "catalyzed" the latter molecular mode of
decomposition at the expense of the free radical decomposition. The authors
did not, however, exclude the possibility that the increased rates of CO
formation were caused by chemical reaction between the HCO radicals and
added NO and 02. The present results allow a seemingly unambiguous choice
between these alternatives.
In our previous study of CH20 photolysis at 3130 A,24 the addition of
a sufficient amount of isobutene and NO lowered the hydrogen quantum yield
to the same limiting value, presumably equal to §p. The agreement between
the two values seems to indicate that the molecular mode of decomposition
is not enhanced by NO. The present results with added isobutene which we
have discussed previously show that this is also true for oxygen. In the
experiment in which 2 Torr of isobutene was added to 8 Torr of CHgO and 2
Torr of 02 (see Table 28.), we find $TT = °*33S* Houston and Moore23
H2 d.
observed a two-fold increase in the relative C0(v = 1) quantum yield when
0.5 Torr of 02 was added to 2 Torr of ClfeO; for the same CH20/02 ratio we
find no effect of 02 on . Thus we feel that reactions (8) and (23) are
responsible for the enhanced formation of the vibronically excited CO.
HCO + 02 -*- H02 + C0(v = 1) (8e)
HCO + NO •*- HNO + C0(v = 1) (23)
The reactions (8) and (23) are sufficiently exothermic (30 and 3^ kcal mole"1,
respectively27) that they can in theory provide readily the necessary 6.2
kcal mole-1 to create vibrationally excited C0(v = 1) molecules. The stronger
effect of NO on the relative yield of C0(v = 1) could possibly be related to
the somewhat higher exothermicity of reaction (23). Finally the observed
smaller effect of NO on C0(v = 1) formation at the longer wavelengths can
also be explained by the occurrence of reaction (23), since the amount of
HCO formed in reaction (l) is significantly decreased at the longer wave-
lengths .
197
-------
1. R. Fort and C.N. Hinshelwood, Proc. Roy. Soc., A, 129, 284 (1930).
2. W.A. Bone and J.B. Gardner, Proc. Roy. Soc., A, 154, 297 (1936).
3. J. Spence, J. Chem. Soc., 1936, 649-
4. J.E. Carruthers and R.G.W. Worrish, J. Chem Soc., 1936, 1036.
5- F.F. Snowdon and D.W.E. Style, Trans. Faraday Soc., 35, 426 (1939).
6. D.W.E. Axford and R.G.W. Norrish, Proc. Roy. Soc., A, 192, 518 (1948).
7. D.W.G. Style and D. Summers, Trans. Faraday Soc., 42, 388 (1946).
8. E.G.A. Horner and D.W.G. Style, Trans. Faraday Soc., 50, 1197 (1954).
9. E.G.A. Horner, D.W.G. Style and D. Summers, Trans. Faraday Soc., 50,
1201 (1954).
10. J.F. McKellar and R.G.W. Norrish, Proc. Roy. Soc., A, 254, 14? (I960).
11. J.J. Bufalini and K.L. Brubaker in "Chemical Reactions in Urban Atmo-
spheres", C.S. Tuesday Ed., Elsevier, Amsterdam, 1971, p. 225.
12. D.E. Hoare and G.S. Milne, Trans. Faraday Soc., 63, 101 (1967).
13. a) R.R. Baldwin, A.R. Fuller, D. Longthorn, and R.W. Walker, J. Chan.
Soc., Faraday I, 68, 1362 (1972); b) Ibid., 70, 1257 (1974).
14. T.L. Osif, Ph.D. Thesis, "The Reactions of 0(1D) and OH with CH3OH, the
Oxidation of HCO Radicals, and the Photochemical Oxidation of Formaldehyde
Formaldehyde", Pennsylvania State University, 1976.
15. T.L. Osif and J. Heicklen, J. Fhys. Ghem., 80, 1526 (1976).
16. H. Niki, P. Maker, C. Savage, and L. Breitenbach, Abstracts 173rd
American Chemical Society Meeting, New Orleans, La., March 20-25, 1977.
17. a) J.G. Calvert, J.A. Kerr, K.L. Demerjian, and R.D. McQuigg, Sci., 175,
751 (1972); b) K.L. Demerjian, J.A. Kerr, and J.G. Calvert, Adv. Environ.
Sci. Technol., 4, 1 (1974).
18. J.H. Clark, Ph.D. Thesis, "Laser Photochemistry and Isotope Separation
in Formaldehyde", University of California, Berkeley, 1976.
19. J. Marling, J. Chem. Phys., 66, 4200 (1977).
198
-------
20. a) D.L. Baulch, D.D. Drysdale, D.G. Home, and A.C. Lloyd, Evaluated
Kinetic Data for High Temperature Reactions, Vol. 1, Butterworth,
London, 19?2; b) W.P. Bishop and L.M. Dorfman, J. Chem. Phys., 52,
3210 (1970); c) T. Hikida, J.A. Eyre, and L.M. Dorfman, J. Ghem. Phys.,
54, 3422 (1971); d) M.J. Kurylo, J. Phys. Chem., 76, 3518 (1972); e)
W. Wong and D.D. Davis, Int. J. Chem. Kinet., 6, 401 (1974).
21. K. Washida, R.I. Martinez, and K.D. Bayes, Z. Maturforsch., 29a, 251
(1974).
22. a) I. Tanaka, private communication of the results of Dr. Shibuya, Ph.D.
Thesis, Tokyo Institute of Technology, Tokyo, Japan, 1976; b) I. Tanaka,
private communication of the results of Dr. Ebata, Ph.D. Thesis, Tokyo
Institute of Technology, Tokyo, Japan, 1976.
23. P.L. Houston and C.B. Moore, J. Chem. Phys., 65, 757 (1976).
»
24. A. Horowitz and J.G. Calvert, Int. J. Chem. Kinet., in press.
25. R. Spence and W. Wild, J. Ghem. Soc., 1935, 338.
26. R.A. Fletcher and G. Pilcher, Trans. Faraday Soc., 66, 79^ (1970)
27. S.W. Benson, Thermochemical Kinetics, Sec. Ed., John Wiley, Blew York,
1976.
28. G.S. Parks and H.P. Mosher, J. Polymer Sci., A, 1, 1979 (1963).
29. F. Su and J.G. Calvert, Int. J. Chem Kinet., submitted for publication.
30. J.H. Clark, C.B. Moore, and J.P. Reilly, paper submitted for
publication; the authors are grateful to Dr. Clark for a preprint of
this work.
31. a) M.J. YeeQuee and J.C.Y. Thynne, Trans Faraday Spc., 63, 1656 (1967);
b) Ibid., Eer. Bunsenges. Phys. Chem., 72, 211 (196b).
32. F. Su, J.W. Bottenheim, H.W. Sidebottom, J.G. Calvert, and E.K. Damon,
Int. J. Chem. Kinet., in press.
33. A.C. Lloyd, Evaluated and Estimated Kinetic Data for the Gas Phase
Reactions of the Hydroperosy Radical, N.B.S. Report 10447, 1970.
34. G.W. Harris and R.P. Wayne, J. Chem. Soc., Faraday I, 71, 610 (1975).
35. J.G. Anderson, J.J. Margitan, and F. Kaufman, J. Chem. Phys., 60, 3310
(1974).
36. A. Horowitz and J.G. Calvert, Int. J. Chem. Kinet., in press.
37. E.D. Morris and H. Niki, J. Chem. Phys., 55? 199 (1971).
199
-------
-------
SECTION 4
STUDIES RELATED TO THE REMOVAL MECHANISMS OF NITROGENOUS COMPOUNDS IN THE
ATMOSPHERE
THE NEAR UV ABSORPTION SPECTRUM OF GASEOUS HONO AND H203
Introduction
In recent years there has been a renewed interest in the study of the
gas phase reactions of HONO. This has been stimulated largely by two
factors: (1) The need for a quantitative evaluation of the reaction path-
ways which form and destroy HONO in the atmosphere; and (2) the Increasing
use of HONO as a convenient photochemical source of HO radicals in the
laboratory.
The atmospheric chemistry of nitrous acid remains poorly defined today.
Nash1 seemingly identified HONO in the atmosphere of southern England (O.k-
11 ppb), but the analytical method employed could, at best, provide only
a very indirect identification of this compound.2 In theory nitrous acid
is expected to be present in the NOx-RH-polluted, sunlight-irradiated atmo-
sphere at levels in the range estimated by Nash. It may be formed in the
atmosphere through several reaction pathways; for example, the reactions
(l)-(3) have been considered in the computer simulation of the chemistry of
the polluted atmosphere:3'4
HO + NO (-tM) •*- HONO (+M) (l)
H02 + N02 -> HONO + 02 (2)
NO + N02 + H20 -*- 2HONO (3)
The ultimate concentration to which HONO will build is limited by the rates
of its various removal reactions such as
2HONO -*- NO + N02 + H20
HONO + HO -*- H20 + N02 (5)
HONO + hv(A < 400nm) ->- HO + NO (6)
Existing rate data suggest that reaction (6) is probably the dominant loss
reaction for HONO in the atmosphere,4'5 but the large difference between
recent estimates of the absorption cross sections of HONO, " make its rate
uncertain by as much as a factor of six. The methods used in the previous
studies appear to be uncomplicated. In the HONO spectral studies of Johnston
and Graham,7 nitrous acid was prepared by reactions (3) and (4), presumably
at its equilibrium concentration, from starting gaseous mixtures of NO, N02,
and H20 in He. Corrections were made for absorptions due to N02, N204, and
201
-------
N203 in the spectrum. In the studies of HONO by Cox and Derwent,8 relatively
pure, dilute mixtures of HONO were prepared in a nitrogen carrier gas which
was allowed to flow over an acidified nitrite salt solution. With this
method corrections for N02 absorption were relatively small, and those due
to N204 and N203 "were negligible. In principle both of these recent quanti-
tative studies of HONO absorption should provide reliable absorption cross
section data, and it is not clear from the published information as to what
problems led to the very different results. Obviously a further study of
this system is necessary to establish the correct HONO absorption cross
sections and to provide a more quantitative evaluation of the potential
reactions of HONO in the atmosphere.
Cox has shown recently that nitrous acid vapors are an excellent photo-
chemical source of HO radicals through reaction (6).9"11 Following his lead,
we have studied some of the thermal and photochemical reactions of HONO
vapors employing an FTS IE, long-path spectroscopy system to follow HONO and
various reactants and products.5'12*13 In the course of these studies, we
observed that the rates of HONO photolysis were very much faster in dilute
gaseous mixtures than we had anticipated from the measured intensities in
our experiments at simulated solar intensities (290-4lOnm region) and the
HONO absorption cross section data then available.7 The present spectral
study of HONO was initiated at that time in an attempt to reevaluate the
near uv absorption cross sections of HONO and to allow an unambiguous and
independent check on our measured rates of reaction 6 in the reaction chamber.
We utilized the equilibrium method employed by Johnston and Graham. As our
work neared completion, the recent study of Cox and Derwent8 appeared. The
new results which we report here are in qualitative accord with those reported
by Cox and Derwent, although some significant differences appear.
Experimental
Spectral measurements of the gaseous mixtures of HONO, N02, N20s, N204,
NO and H20 were made using a Gary 17 uv-visible spectrophotometer in a 10 cm
path gas cell equipped with a Teflon vacuum stopcock and Supracil windows.
The optical resolution varied with wavelength but was kept less than 1 nm
in the 325-^50 nm range and less than 5 nm in the 300-325 nm range. Since
King and Moule6 found no trace of fine structure in the HONO spectrum when
they employed a 20 ft grating spectrograph with a resolving power of 150,000,
our equipment was sufficient to resolve the structure of the HONO. The
reactant gases NO and N02 (Matheson), and H20 were purified by fractional
distillation and degassing in two conventional high vacuum systems which
were equipped with Teflon high vacuum stopcocks, calibrated volumes, pressure
gauges, and a quartz spiral manometer.
In the study of the N203 spectra, small quantities (l-llj- Torr) of an
NC2-N204 gaseous mixture were added to the cell. Then a large measured
quantity of NO (112-753 Torr) was introduced into the cell. The cell
contents were allowed to equilibrate for about 30 min, and then the absorption
spectrum of the mixture (300-^50 nm) was determined. The N02 pressure in
the cell was estimated from the visible absorption at ^30, Ifto, and Mf8 nm
using the absorption cross section data for these wavelengths reported by
Hall and Blacet;14 the three independent estimates of the P checked well
N02
202
-------
with one another, usually within a few percent. From the measured P in
the equilibrium mixture and the P°Q, P^^ and P were calculated using
the known equilibrium constants K and KR for the given temperature of the
particular experiment. '
(7)
NO + N02-±5.N203 (8)
The values of K~ and Kg which we employed were derived from the JANAF
enthalpy, entropy, and heat capacity data15 and should be reliable for
temperatures used in this work (2U-30°C):
KT = exp[6989T-1 + 0.791 la( 1/298) - 21.52] (9)
Kg = exptl^T-1 - O.l¥j- Jji(T/298) - 16.85^] (10)
These expressions yield estimates which are in reasonable accord with the
measured values of K and Kg at 298°K [Ref. 16,18] and 300°K [Ref. 17], The
absorption data of Bass, et al.19 were used to determine absorbances due to
B02 and N204 in the 300-400 nm range, and these were subtracted from the
measured absorbance to determine that due to H20s. Absorbances determined
in this fashion at several wavelengths followed well the linear Beer's law
dependence on [N20s] over the entire pressure range which could be studied
here (0.5*4-6.32 Torr); data for the 300, 310, and 320 nm measurements are
shown in Figure 51 • Summarized in Table 32 are the absorption cross sections,
cr, = In (I _/I, )/[H203]^ CB^molec"1 for N203 at each nm in the 300-^00 nm
range .
In the HOWO studies, HONO was formed from N02, WO, and H20 vapors
through the equilibrium 11:
W + N02 + H20±^.2H01\TO (11)
Mixtures were prepared by adding small amounts of N02, H204 as before. Then
H20 vapor at its equilibrium pressure at 0°C was introduced to a calibrated
volume in the manifold, and a measured fraction of this was distilled into
the cell. This procedure provided the same 1.57 Torr (298°K) initial
pressure of H20 in each run which was made. A measured quantity of nitric
oxide gas was then added to the cell as before. The mixture was allowed to
warm to room temperature and to equilibrate. Spectra of the mixtures were
determined every day for three successive days; they did not change de-
tectably after the first 2k hour period, so it seems clear that equilibrium
was established in the mixtures used for the HONO absorption cross section
measurements in this work. The equilibrium P was measured as before.
Using the initial pressures of H20 and HO, ancl $he measured amount of N02
in the equilibrium mixture, the pressures of the other components in the
equilibrium mixture were calculated using relations (9)5 (10)? an-d (12):
203
-------
-------
TABLE 32. DINITROGEN TRIOXIDE ABSORPTION CROSS SECTIONS*
Wavelength,
EUQ
300
301
302
303
304
305
306
307
303
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
c, cm moleo"^-
x 1020
150
146
141
136
132
128
123
118
113
109
106
101
97.4
94.0
90.9
88.6
85.6
83.3
81.7
79.8
78.3
77.5
76.8
76.4
76.0
75.6
75.2
74.9
74.5
74.1
73.7
71.8
70.7
69.9
Wavelength,
run
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
o, cm2 molec"1
x 1020
68.4
66.3
65.7
63.8
62.3
60.3
58.1
57.3
56.1
55.4
54.2
53.5
53.1
52.7
52.3
51.9
51.6
50.4
49.7
48.9
48.1
47.7
47.0
45.3
44.7
43.9
43.2
41.2
39.7
38.2
36
35
33
31
Wavelength,
run
368
369
370
371
372
373
.374
375
376
377
378
379
380
381
332
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
a, cm2 molec"!
xlO20
29
27
26
26
26
26
26
26
26
27
27
27
27
25
24
23
21
20
13
17
15
13
12
10
8.8
7.3
5.7
4.6
3.8
3
2
1
0
In
-------
Kxl = exp[-15.56 + 4.73 x lO3!-1] (12)
Relation (12) was derived by Chan, et al.5 from, the combined experimental
data of Wayne and Jost,2° Waldorf and Babb,21 and Ashmore and Tyler.22
The JMFAF data for the HONO species must be in error since constants derived
from these data do not reproduce the experimental equilibrium data.5 Using
these composition data, the known absorption cross sections for N02 and
N204 [Ref. 19] together with our absorption data for N203 (Table32), ab-
sorbances due to N02, N204, and N20s were calculated at 1 nm wavelength
intervals from 300 to 400 nm. These were subtracted from the total measured
absorbance of the equilibrium mixture to obtain the absorbance due to HONO.
The fraction of the measured absorbance at 332 and 368 nm which was at-
tributable to HONO was in the range 15 to 67$ of the total absorbance
measured. The absorbances for HONO followed reasonably well the linear
dependence on concentration expected from Beer's law for the very limited
range of concentrations which could be determined for this system (1.47-3-73
Torr, 25°C). The absorption cross sections which we have estimated from
this study are summarized in Table 33. The accumulated error in these
estimates which could result from the absorbance measurements and equilibrium
data employed should be less than 10$ at the maxima in the HONO absorption
spectrum.
The near uv absorption spectrum of N50s vapor
The absorption cross sections estimated for NaOs(g) in the near ultra-
violet region are presented in Table32 and plotted in Figure 52. The con-
tinuous nature of the absorption reported previously is confirmed.23 However
there is a weak, but readily discernible structure which can be seen. We
attribute this tentatively to the N20s(g). It is not likely that such an
effect would result from impurity bands of HONO which may have formed in-
advertently from the reaction of the oxides of nitrogen with H20 impurity
released from the cell wall. The amount of such water vapor must be below
a few hundred ppm at most, and the equilibrium level of HONO in such a
mixture could not create the relatively intense absorption of the broad
structure observed here. Furthermore the positions of the strongest HONO
bands at 368 and 3^0 nm are near minima in the broad structure exhibited in
Figure 52. Conceivably the structure could arise due to the faulty correction
for N02 present in the N20s mixture, but experiments in which N02 absorption
was matched in the reference beam of the spectrometer also showed this
broad structure. We are not aware of any tabular data with which we can
compare these cross section estimates. Johnston and Graham7 did not publish
their measured data for N20s. That of Shaw and Vesper23 were presented in
graphical form only, and the scale of the plot chosen makes impossible an
accurate reading at the wavelength of most intense absorption in our work
(300 nm); however it appears to be qualitatively consistent with our
estimates.
The near uv spectrum of HONO vapor
The absorption cross section data for HONO determined in this study are
summarized in Table33 and plotted in Figure 53(the solid curve). These data
206
-------
TABLE 33. NITROUS ACID ABSORPTION CROSS SECTIONS*
Wavelength,
ran
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
a, cm2 molec"^
x 1020
0.0
0.0
0.2
0.42
0.46
0.42
0.3
0.46
3.6
6.10
2.1
4.27
4.01
3.93
4.01
4.04
3.13
4.12
7.55
6.64
7.29
8.70
13.8
5.91
5.91
6.45
5.91
4.58
19.1
Wavelength,
nm
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
a, cm molec"!
x 1020
16.3
10.5
8.70
33.5
20.1
10.2
8.54
8.32
8.20
»7.49
7.13
6.33
17.4
11.4
37.1
49.6
24.6
11.9
9.35
7.78
7.29
6.83
6.90
7.32
9.00
12.1
13.3
21.3
35.2
Wavelength,
nm
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
o, cm molec"*
x 1020
45.0
29.3
11.9
9.46
8. 35
7.44
4.77
2.7
1.9
1.5
1.9
5.8
7.78
11.4
14.0
17.2
19.9
19.0
11.9
5.65
3.2
1.9
1.2
0.5
0.0
0.0
0.0
0.0
0.0
* a - In (I,/! )/[HONO]£, cm2
20?
-------
300
350
Wavelength, nm
400
Figure 52. The absorption cross section for N20s(g)
as a function of wavelength.
208
-------
T •»
'o
30
-------
can be compared with those estimated in the studies of Johnston and Graham7
(the lower dashed curve) and Cox and Derwent8 (the upper dashed curve). It
is apparent that the magnitude of the cross sections for the peaks in ab-
sorption estimated in our work are more nearly equal to those of Cox and
Derwent, although the vibrational structure which we observe correlates
better with the data of Johnston and Graham.7 The peaks seen in the Cox
and Derwent spectrum near 325, 330, 3^5, 3^8, 357, 361, 373, 376, 380, and
391 nm do not appear in our HONO absorption spectrum or in those reported
by Johnston and Graham7 and King and Moule.6 We noted that maxima occur
very near these wavelengths in the spectrum of N02 gas.19 The magnitude of
these peaks is within the uncertainty in the HONO cross sections reported
by Cox and Derwent. However if these peaks are real, then it is likely that
the Cox and Derwent spectral data for HONO include some uncorrected absorptions
from N02 impurity in their mixtures. In making the corrections for N02 ab-
sorption, they employed the Johnston and Graham absorption data for N02
which lists only values at wavelength intervals of 5 nm [Ref. 7]• It seems
inevitable that some peak structure of N02 must be retained in the corrected
spectrum of HONO using this procedure. If one assumes that this is the
origin of the difference between our absorption data and that of Cox and
Derwent and that they have correctly estimated the concentration of HONO in
their mixtures, then we expect the uncorrected NQ2 to be given by: [N02]
= CHOW0^^HONO - 0HONO)/0"N02); °HONO iS the ^^e^ absorption cross
section observed by Cox and Derwent, and crHmTn and
-------
A test of the reliability of the HONO absorption cross section data
can be made using rate data derived in our large photochemical reactor
system. The ultraviolet light •which bathes the cell contents rather uniformly,
mimics quite well the ground level solar flux both in intensity and wave-
length distribution in the 2900 to 4100 A region. FTS IR spectrometric
analysis of reactants and products can give a fairly continuous record of
the chemistry which occurs in the reactor. The apparent first order rate
constant for reaction (7) was determined in a series of experiments employing
low concentrations of HONO in 700 Torr of air and in mixtures with small
quantities of added CO, NO, and N02 in 700 Torr of air.24 Rate data on
reaction 13 were obtained in another series of experiments at low con-
centrations of N0£ carried out both in the absence and the presence of 700
Torr of air.
HONO + hv -*- HO + NO (7)
N02 + hv -*- 0 + NO (13)
The kinetic treatment of these data gave k_, = 0.11 and k, „ = 0.60 min'1
[Ref. 13]. The relative light intensity versus wavelength distribution.
function F(A) for the uv light within the cell was determined using an
absolute spectrofluorimeter as detector. These data were coupled with our
present cr (A) estimates, cr (A) data of Bass, et al.,3-9 primary quantum
yield data for (Ref. 13) , <1> (A), of Jones and Bayes,24 and the assumption
that 3> = 1 at all HONO absorbing wavelengths in the 290-^10 nm range,8 to
derive the theoretically expected relative rates of reactions (7) and (13)
in our photochemical reactor:
J cr(A) F(A) *(A) dA
Using relation (14) and integrating over the 290 to 4lO nm range of non-zero
values, we obtained R7/R-,o = O'1^ ± 0.02. This checks well with the ratio
of the independently estimated experimental rates: Ry/Rjo sO.ll/0.06s;
0.18 ± 0.0k-. Thus some support is provided for the accuracy of the present
HONO cross section data and the conclusion of Cox and Derwent that =1.0
within the first absorption band of HONO. If the o~HONO values of Johnston
and Graham are used in relation (1^), we obtained R^/R-^ = 0.065; use of the
Cox and Derwent estimates of cr™™ yields R^/R13 S °*32-
We feel that our present estimates of the absorption cross sections
of HONO are the most reliable of those available today, and we recommend
their use in the modeling of the tropospheric chemistry. We may combine
our cr values with Peterson's recent estimates of the flux of solar
HONO
radiation at ground level25 to derive new theoretical estimates of k^
211
-------
applicable for the lower troposphere at various solar zenith angles. These
estimates are shown in Figure 5^ for the choice of zero surface albedo. They
may be increased by the appropriate reflection factor which describes best
the properties of the earth's surface at the point of interest.
The unambiguous detection of HOHO in ambient air has not yet been ac-
conplished to our knowledge. Simulations using these estimates suggest
that for a typical NOx-RH-polluted, sunlight-irradiated atmosphere, the
levels of HOKO, controlled largely by reactions (l)-(6), which are expected
to be present near ground level should be in the ppb range. Obviously a
sensitive detector system will be necessary to successfully identify HONO
in the ambient air.
212
-------
CM
O
r—
X
'c
I
CO
2O 40 60 80
Solar zenith angle, cleg.
Figure 5 4. Estimated first order rate constants for
HONO(g) photolysis in sunlight, HONO + sunlight -*-
HO + MO; data are calculated for sea level and
surface albedo equal zero for various solar zenith
angels; actinic flux data from Peterson.25
2L3
-------
1. I. Hash, Tellus, 26 (197*0 175-179-
2. I. I'lash, Am. Occup. Hyg., 11 (1968) 235-239.
3. K.L. Demerjian, J.A. Kerr, and J.G. Calvert, Adv. Environ. Sci. Technol.,
4 (1974) 1-262.
4. J.G. Calvert and R.D. McQuigg, Int. J. Chem. Kinet., Symp. 1 (1975)
113-154.
5. W.H. Chan, R.J. Nordstrom, J.G. Calvert, and J.H. Shaw, Environ. Sci.
Technol., 10 (1976) 67^-682.
6. G.W. King and D. Motile, Can. J. Chem., 40 (1962) 2057-2065.
7. H.S. Johnston and R. Graham, Can. J. Chem., 52 (1974) l4l5-l423.
8. R.A. Cox and R.G. Derwent, J. Photochem., 6 (1976/77) 23-34-
9. R.A. Cox, J. Photochem, 3 (1974) 175-188.
10. R.A. Cox, J. Photochem., 3 (1974) 291-304.
11. R.A. Cox, Int. J. Chem. Kinet. Symp. 1 (1975) 379-398.
12. W.H. Chan, W.M. Uselman, J.G. Calvert, and J.H. Shaw, Chem. Phys.
Letters, 45 (1977) 240-244.
13- S.Z. Levine, W.M. Uselman, W.H. Chan, J.G. Calvert, and J.H. Shaw,
Chem. Phys. Letters, 48 (1977) 528-535.
14. T.C. Hall, Jr., and F.E. Blacet, J. Chem. Phys., 20 (1952) 1745-1749.
15. JANAF Thermodynamic Tables, 2nd Edition, Nat. Stand. Ref. Data Ser.,
Nat. Bur. Stand. 37 (1971).
16. F.H. Verhoek and F. Daniels, J. Amer. Chem. Soc., 53 (1931) 1250-1263.
17. E.W. Kaiser and C.H. Wu, J. Phys. Chem., 8l (1977) 187-190.
18. E.R. Beatte and S.W. Bell, J. Chem. Soc. (1957), l68l-l686.
19. A.M. Bass, A.E. Ledford, and A.H. Laufer, J. Res. Nat. Bur. Stand. Sect.
A, 80A (1976), 143-166.
20. L.G. Wayne and D.M. Yost, J. Chem. Phys., 19 (1951) 4l-47.
21. D.M. Waldorf and A.L. Babb, J. Chem. Phys., 39 (1963), 432-435; ibid.,
4o (1964) 1165.
214
-------
22. P.G. Ashmore and B.J. Tyler, J. Chem. Soc. (196!) 1017-1021.
23. A.W. Shaw and A.J. Vosper, J. Chem. Soc., Dalton (1972) 961-96^.
2U. I.T.N. Jones and K.D. Bayes, Chem. Phys. Letters, 11 (1971) 163-166.
25. J.T. Peterson, "Calculated actinic fluxes (290-700nm) for air pollution
photochemical applications", Report EPA~600/i|-76-025, June, 1976,
Environmental Protection Agency, Research Triangle Park, N.C.
215
-------
THE KINETICS OF DIMETHYLAMINO RADICAL REACTION IN SIMULATED ATMOSPHERES:
THE FORMATION OF DIMETHYLNITROSAMINE AND DIMETHYLNITRAMINE
Introduction
There has been increasing concern over the possible generation of
nitrosamines within the atmosphere because of their unusually high
carcinogenic activity.
In 1956, dimethylnitrosamine (DMN) was shown to cause cancer in rats,1
and concern has continued to grow about this and other N-nitrosamines and
their relation to human health., Although there have been no studies directly
linking DMN with cancer in humans, DMN has been shown to cause tumors in
mice, hamsters, mink, rats, guinea pigs, rabbits and rainbow trout.2 To date,
no animal species studied has shown any resistance to the carcinogenic
action of DMN. It would seem rather presumptuous to assume that man is the
uniquely immune species.
DMN has been shown to be an effective carcinogen regardless of the
manner of application. Ingestion, inhalation, injection and topical ap-
plication of DMN have all been shown to cause tumors, generally far from
the site of administration. For example, long-term, low-dosage ingestion
of DMN was found to lead to liver cancer, while higher concentrations in the
foodstuff produced kidney cancer in a shorter amount of time — in several
cases after just one dose.3 A single dose of 5 PP11 caused tumors in 70$> of
the animals in another study.4 Inhalation of DMN was shown to be an ef-
fective means of exposure at levels of 200 M-g/m3 (67 ppb); furthermore, the
study showed that a small daily dose led to a smaller cumulative dose before
the appearance of cancer.5
It is from the overall picture rather than from any individual study
that concern arises. It is these considerations that prompt researchers
to classify DMN as one of the most potent carcinogens.3>G>7
The first reported finding of nitrosamines in the atmosphere, the
detection of DMN and diethylnitrosamine (DEN) near a factory producing
secondary amines in Germany,8 generated considerable interest in the atmo-
spheric occurence and behavior of these species. Since that time,
nitrosamine concentrations in several U.S. cities have been measured.2»9»10
The highest concentrations of DMN were found near an FMC plant synthesizing
unsymmetrical dimethylhydrazine and utilizing DMN as an intermediate. These
concentrations peaked at 36 p-g/m3 (12 ppb), and levels in the nearby community
of Baltimore ranged from 0.03 to 8 ng/m3 (0.01 - 2.6 ppb). In Belle, W. Va.,
where both DuPont and Union Carbide manufacture or use dimethylamine (DMA),
levels of from 0.2 to 1.0 M-g/m3 were found.10 This is comparable to levels
observed in New York City.11 These levels are much higher than those found
for other airborne carcinogens, e.g., benzo(a)pyrene at 2 ng/m3 [Ref. 12],
and thus these levels are cause for concern.
'The origins of DMN in ambient air are not clear. Presumably, the
nitrosamine will either be emitted as a primary pollutant, or conceivably
it could be formed from precursors in the atmosphere.
216
-------
The only known major emitter of Dm, the IMC plant near Baltimore, was
shut down in April, 1976.1S Small amounts of DMN have also been found to be
by-products of the combustion of rocket fuel14 and tabacco;15 DMN is also
produced in the preparation of fish products,16'17 which are rich in amines.
The main precursors of nitrosamines in the atmosphere are felt to be
secondary and tertiary amines, and oxides of nitrogen. Both are widely
distributed. Amines are emitted as a result of a wide variety of industrial
activity, including amine manufacture, food processing, coking, the refining
of petroleum, and incineration. They are also given off in appreciable
concentrations from sewage treatment plants, animal feed lots, and swamps.11
Nitrogen oxide sources include fossil fuel combustors, refuse incineration,
home heating, automobile exhausts, and industrial processes related to the
manufacture and use of nitric acid.12
Bretschneider and Matz8 have shown that mixtures of dimethylamine and
nitrogen dioxide at 50 to 100 ppm combine within seconds to form DMN.
Neurath et al.18 have shown that the stoichiometry of this reaction is 1:1,
but the mechanism of the reaction remains unclear. Recent work by Hanst19
at much lower concentrations (2 ppm NO, 2 ppm N02, 1 ppm DMA) indicate that
this reaction is too slow to be an important source of nitrosamine under
atmospheric conditions.
Hanst19 has also studied the reaction of HONO with DMA. The reaction
of 0.5 ppm HONO (in equilibrium with 2 ppm NO, 2 ppm N02, and 13,000 ppm
H20) with 1 ppm DMA. gave DMN from DMA in about 10 to 30fo yield. Glasson20
repeated this work, finding the same rate of loss of the amine, but obtain-
ing only about 1% yield of the nitrosamine. He suggests that this is
evidence that the reaction does not take place homogeneously, but is con-
trolled by a reaction occurring at the cell wall.
Further work by Hanst19 includes a study of the photolysis of DMN in
sunlight. Hie sample was prepared by mixing 20 Torr DMA with 20 Torr NO,
adding room air, and pumping out the volatile components. A value for
t / of 30 minutes for photolysis in bright sunlight at noon was obtained,
corresponding to an apparent first order photolysis rate constant, ^ =
0.02 min"1.
(CH3)2N-NO + hv •*- (CHs)2N + NO (l)
Products found include nitric oxide, carbon monoxide, formaldehyde, and an
unidentified compound with absorbances at 1560 cm~1(s), lM30 cm-1(m-s),
1310 crn'Ms), 1000 cm-x(m) and 775 cnrMw).
In light of the work of Hanst, Pitts et al.21 have initiated a study
of amine-NOx photolyses. One of the amines employed was DMA. DMA was
combined with NO and N02 at a relative humidity of about kctfo and allowed
to react for two hours in the dark before being exposed to sunlight. The
qualitative results agree with Hanst; that is, DM is formed in the dark,
and disappears during photolysis with t../2 s 1 - 2 hr. The unknown product
of Hanst is identified as dimethylnitramine, (CH3)2N-N02.22 This is the
217
-------
same product obtained by Althorpe et al.23 by the sunlight photolysis of
dilute solutions of DMN in hexane.
The detailed mechanism of nitrosamine formation is still obscure, al-
though some speculation about it has been made. Attack of the amine by a
hydroxyl group has been suggested as an initiating step for reaction of the
amine.21?24
(C%)2NH + OH -*• (CHs)2N + HOH (2a)
+ HOH (2b)
Atkinson et al.25 have measured the rate constants for the reaction of
OH with dimethylamine and trimethylamine to be 9«6? x 104 ppm"1^!!""1 and
9.00 x 104 ppnT^ffiin"1. If we make the naive assumption that the rate con-
stant can be separated into a value for each bond (analogous to Greiner's
work with alkanes),26 then the results of Atkinsori et al. lead to k =
0— il
1.00 x 104 ppm~1min~1J and k^ = 3«6? x 104 ppm"1^!!"1. From this we can
estimate the rate constants of k0 = 3.67 x 104 ppm~1min~1 and kc, = 6.00 x
£_a c-D
104 pptif^-min"1, or a relative rate of k?,/k = 1.63.
Once the radical (CHs)2N has been formed, it is expected to react in the
atmosphere according to the following scheme:
(CHs)2N + 02 -*- (CHs)2N-02 and other products (3)
(CH3)2N + HO (CHs)2N-NO (k)
(CHs)2N + N02 - (CH3)2W-H02 and other products (5)
(CHs)2N-NO + hv -. (CHs)2lT + NO (l)
At first glance, it might seem that the only reaction of any importance
in the 02-rich atmosphere would be reaction 3; oxygen is present at con-
centrations approximately 106 times greater than NO or NOs in urban air,
and most organic radicals react very quickly with oxygen. However,
Lesclaux27*28 has shown that the radical HH2 is very much less reactive
towards 02 than it is toward NO.
NHa + 02 ->- products (6)
NH2 + NO -»~ products (7)
His flash photolysis studies show that kg < 5 x 10~3 ppm"1!^"1 [Ref. 27],
while k = 3 x 104 ppnr-'-min"1 [Ref. 28]. If the dimethylamine radical is
analogous to the NH2 species, it is conceivable that its reactions with
pollutant molecules could dominate the system, and may contribute heavily
to the nitrosamine concentrations observed in urban atmospheres.
218
-------
This hypothesis was tested in this work through the quantitative study
of the reactions 1, 3, 4, and 5 using FT IE spectroscopy. Relative rates
for the reaction of the amine radical with NO, N02, and 02 were determined.
The data presented here may be used to derive the first quantitative estimates
of the importance of the nitrosamines and nitramines formed through these
homogeneous reactions in the atmosphere.
Experimental Study of the Dimethylamino Radical
In this study, reactions of amines of probable importance in the atmos-
phere were investigated. To understand these reactions, it is advisable to
study gaseous systems which contain reactants at concentrations approximating
those found in ambient air and which can be irradiated at light levels
similar to sunlight.
Although many techniques have been proposed for monitoring low levels
of gaseous pollutants, infrared absorption remains among the most specific,
sensitive and accurate methods for the detection of many molecules. The
ability to detect a given molecular species increases with path length and
with spectral resolution.
The time required to collect an infrared spectrum typically increases
as the spectral resolution or the spectral interval surveyed is increased.
The slow response typical of grating and prism infrared systems has been
largely overcome by Fourier transform spectroscopy (IRFTS), which can
collect spectral information over wide spectral regions in relatively short
times.
The instrumentation available for the present study includes a glass-
walled multiple pass cell with a maximum path length of more than 500 m.
'The cell is surrounded by fluorescent bulbs which result in an irradiance
inside the cell similar in spectral distribution and magnitude to ground
level solar radiation between 300 and kOO nm.
The infrared spectra were taken using a Michelson interferometer
(Digilab Model FTS 20). The time required to collect a single scan inter-
ferogram from 0 to 3950 cm"1 at 1 cm"1 resolution and to store the in-
formation on magnetic tape is 28 seconds. This rapid collection speed
allows chemical reactions to be monitored rapidly so that rate constants
can be determined.
The FTS Photolysis System
Photolysis Cell —
The main body of the cell consists of four 1.5 m lengths of Corning
low-expansion borosilicate glass separated by stainless steel spacers with
Teflon gaskets. Stainless steel endplates are also used. One endplate
contains KBr windows to allow infrared radiation to enter and leave the cell;
the other has an outlet to a vacuum pump and mechanical feed throughs to
adjust the mirrors. Each endplate and spacer has been provided with a gas
219
-------
inlet port to help ensure a uniform distribution of gases throughout the
cell.
The entire length of the cell can be irradiated by four banks of
fluorescent lights surrounding the cell. The lamps installed in the system
are GE FT2T12/BL/HO and F^-OBL black lights. The emission spectrum of one
of these lamps after passing through the cell wall was measured with a
Turner spectrophotometer operated in the luminescence and energy mode. The
spectrum obtained is shown in Figure 55$together with the estimated relative
spectral distribution of solar radiation for a solar angle of ^5° [Ref. 29].
320 36O
Wavelength, nm
400
Figure 55- Comparison of irradiance inside photolysis
cell with solar irradiance.
(a) spectral distribution of fluorescent lamp
output after passing through glass walls
of cell (unpublished data of Eugene Beer);
(b) relative spectral distribution of solar
energy at ground level.
220
-------
The fluorescent lamp assembly is surrounded by an aluminum foil reflector
to increase the intensity and improve the uniformity of the light within
the cell.
Gas Introduction System __
Figure56 shows the gas handling system used to inject known amounts of
gases into the photolysis cell. This grease-free, mercury-free system was
constructed entirely of glass with Teflon stopcocks. Connections are made
with stainless steel Cajon unions using Viton 0-rings. Pressures are read
on one of two Wallace and Tiernan absolute pressure guages (0-20 and 0-800
Torr) which are separated from the system by a quartz spiral manometer used
as a null meter.
This gas handling system was designed to mininize dead volumes and to
permit positive injections of samples with a carrier gas; positive injection
capability is essential for forming multicomponent mixtures in the photolysis
cell.
The seven volumes built into the vacuum system (2.7 to 92.k cm3) allow
convenient injection of samples from less than 0.01 to over 200 ppm.
Larger concentrations can be injected by attaching larger bulbs to the
system at the direct inlet port. Weighed amounts of solids or liquids with
reasonable vapor pressure can also be injected through this port.
A manifold of stainless steel tubing (0.25 i-n« o.d.) connects the gas
handling system to the photolysis cell. Five inlet ports (one in each
spacer and endplate) help to ensure a uniform distribution of gases
throughout the cell.
Optical Arrangement Within the Photolysis Cell —
A standard White cell is shown in Figure 57, and consists of three
spherical mirrors of equal focal length. The light from the source is
focused into a real image at the entrance aperture of the cell, and above
the horizontal centerllne of the field mirror Mf (see Figure 58). It then
diverges and is collected on mirror DI . Since D, is located two focal
lengths from the image, the inverted image from D, is focused on Mf. D^^ is
adjusted so the image (marked 2) falls below the horizontal centerline of
Mf. The reflected diverging beam falls entirely on D2, and Dg is adjusted
so the real image formed (marked 4) is beside the aperture image. If this
image is placed in the exit aperture, the total number of traversals is
the minimum of four. If it is placed symmetrically opposite the first image
(marked 2), there will be at least four more passes through the system.
The optical system used in this study was a modified White cell, and
has been described'by Hanst.30 As shown in Figure 59? the single field
mirror M in Figure 3 has been replaced by four rectangular spherical mirrors
M
a
, M, , M , and M,, and the pair of mirrors D^ and D2 have been replaced by
221
-------
ro
IX)
RH
Ports
-a
To Cell
-V.T— - \^civ-iyv-« _ >
Direct Inlet Port
^
— s
\
-\
^•~"\j\__j*j
To Pump
Trap
Figure 56. Gas handling system for introducing known amounts of chemicals into the photolysis cell.
-------
ro
D
1
Figure 57- Optical paths in a standard White cell (kn pass).
-------
0
4n
4n-2
Figure 58, Placement of images on ]VL in a standard
White cell. The numbers refer to consecutive
reflections by the mirrors.
22k
-------
ro
IX)
VJl
Ma
Mb
Mc
Md
4 12
0
8 16 (8)
14 6
1O 2
16.5cm
19.7cm
Figure 59- Mirror system for the modified White cell. The numbers refer to consecutive ref-
lections by the mirrors. The number of traversals is altered by turning D^.
-------
four quadrant shaped spherical mirrors D-^, D2, D~, and D^. In an aligned
system, an image of the source is formed in the plane of the mirrors near
M, (marked 0). Rows of images are formed on M , M,, M , and M, after
reflections by the mirrors D , Dp, D~, and D> . The beam leaves on the
opposite side of M, .
'The number of traversals used throughout this work was 32. The
distance between the mirrors is 5-31 m, giving a total path length of 170 m.
Transfer Optics --
A schematic of the optical arrangements used in this work is given in
Figure 60. The entire optical system outside the cell is mounted on an
aluminum slab and enclosed in a transparent plastic box sealed to be es-
sentially airtight. The box is purged with dry nitrogen gas during
operation, and residual water and carbon dioxide in the gas are removed by
placing traps of sodium hydroxide and activated alumina on the floor of the
box.
Radiation from a Hernst glower (N) is reflected by the spherical mirror
M, (f, =15 cm) to an off-axis paraboloid mirror IVL (f? = 27 cm) which
results in parallel radiation to the interferometer. Within the inter-
ferometer, the collimated beam is amplitude-divided at a beam splitter (BS)
consisting of germanium on a KBr substrate. Half of the radiation is re-
flected to the movable mirror MLi and half is transmitted to a stationary
mirror MS (Figure 6l). These two beams recombine at the beam splitter and
exit the interferometer. The resultant beam is reflected by two plane mirrors
(Mg and MB in Figure60) and two spherical mirrors (My, f_ = 60 cm; Mg, fq =
lj-5 cm) before entering the absorption cell. Optics are adjusted so an image
of the Nernst is formed at NS in the plane of the field mirrors and an image
of the beam splitter is formed on the top left quadrant at the far end of
the cell. Intermediate images of the Nernst are also found at positions
N, and Ng.
Results and Discussion
DimethyInitrosamine (DMN) has been found to be one of the most
carcinogenic chemicals in experimental animals.3s6»7 In view of this, the
detection of DMN in industrial atmospheres2s8?11 has generated considerable
interest in the atmospheric sources and sinks of this and. other similar
compounds.
Formation of Dimethylnitrosamine by the Photooxidation of Dimethylamine in
Simulated Atmospheres —
Through the formation of nitrosamines from amines in the condensed
phase has been studied extensively,31 few investigations of their formation
from amines in the atmosphere have been made. Our original intention in
226
-------
M
DO
tV)
Cu:Ge Detector
Interferometer
F±gare 60. Optical transfer system from interferometer to photochemical cell.
-------
CO
Input
Output
BS
A
V
] M
r"i
M5
>
0.5cm
Figure 6l. Optical diagram of interferometer
-------
this work was .to characterize the products and kinetics of the photo-
oxidation of dimethylamine (DMA) in simulated atmospheres, with particular
attention being paid to the question of nitrosamine formation.
Results from experiments designed to investigate the reaction of DMA
with nitrous acid, concentration- tine profiles were derived from the data;
a typical one is shown in Figure 62.
Formation of Dimethylnitrosamine in the Dark --
Dimethylnitrosamine was produced by the reaction of DMA with HONO in
the dark. A yield of DMN from DMA of about 5% was found in each ease
before photolysis was initiated, though a considerably lower yield (< 2%}
was found in the dark period directly following the photolysis. These values
are greater than those reported by some workers20'21 and less than that
obtained by others,19 supporting the idea that the reaction is heterogeneous.20
Initially, DMN was produced quite rapidly, with a concurrent rapid loss
of HONO and DMA. The rate of DMN formation decreased to a fairly steady value
within five minutes. This is similar to behavior noted by Pitts et al.21
However, in our system considerable nitrous acid and amine were left, in-
dicating that the reaction was limited by wall reactions rather than by the
availability of nitrous acid.
Nitrosamine Formation During Photolysis —
The rate of nitrosamine formation in both experiments increases
dramatically when photolysis is initiated, and slows as photolysis is con-
tinued. In the first experiment, an initial rate of DMN formation of 0.26
ppm/min decreases to Q.Qk ppm/min after 10 minutes. In the experiment
with 10 ppm NO added, the initial rate of 0.18 ppm/min decreases to about
0.08 ppm/min after 10 minutes.
The sequence of reactions leading to the formation of DMN presumably
begins with the photolysis of nitrous acid to produce OH radicals, which
then attack the amine.
HONO + hv -*- OH + NO (22)
OH + (CHs)2NH •*- (CH3)2N + H20 (2a)
The radical (CHs)2N formed in reaction 2a can, among other possibilities,
react with NO to form DMN.
(CH3)2W + NO -*- (CH3)2N-NO
In the experiment where 10 ppm NO was added, the reaction of the di-
methylamino radical with NO will be enhanced. In this experiment, a
maximum yield of DMN from DMA of 31% was found during the first few minutes
of photolysis. Ihis indicates that at least 31$ of the amine reacts with OH
229
-------
0 20 40
Time, min
6O
Figure 62. Concentration/time profile for the
reaction of dimethylamine with nitrous acid.
Open symbols indicate data collected during
the photolysis. Circles, dimethylamine;
triangles, nitrous acid; squares, dimethyl-
nitrosamine.
230
-------
to form the dime thy lamino radical, or ^2a.^2b = °'^' This is in good
agreement with our estimation of 0.6l from the data of Atkinson et al.,25
and supports the contention of Pitts et al.21 that H-abstraction from the
N-H bond is competitive with H-abstraction from the C-H bonds in secondary
alfcyl amines.
The reduction in the rate of formation of nitrosamine can be explained
by a combination of factors. These include lower concentrations of the
reactants HONO and DMA., and removal of DMN from the system, either by
photolysis or reaction with other species in the system. The observed
reduction in the yield of DMN from DMA. can best be explained by the removal
of DMN.
At this point in the study we became aware that the group directed by
Dr. J.N. Pitts in Riverside, California, had initiated an IRFTS study of
the irradiation of amine-NOx systems essentially identical to the one we
had begun.22 Rather than duplicate their efforts, the emphasis of our study-
was changed from a general consideration of amine reactions in NOX -polluted
atmospheres to concentration on the reactions of the dimethylamino radical
in these atmospheres.
Determination of the Photolysis Rate of Dimethylnitrosamine--
The ultraviolet absorption spectrum of dime thy Initros amine and the
measured spectral irradiance inside of the photolysis cell (Figure 55)0verlap
in the region from 300 to ^20 nm. The dissociation energy of the N-N bond
in DMN has been given by various authors as 32 kcal/mole,32 ^-3 kcal/mole,33
and 55.2 kcal/mol.34 Taking the largest (and most recent) of these values,
we can expect possible dissociation of the nitrosamine at wavelengths below
518 nm. Thus our photolysis cell is suited to this study.
Photolysis of Dimethylnitrosamine in Nitrogen --
Dimethylnitrosamine is expected to photolyze according to reaction 1:
(CHs)2N-NO + hv -*- (CHs)2N + NO (1)
Data from the photolysis of DMW in nitrogen were determined in a series of
runs and shown as circles in Figure 63. As can be seen from a plot of the
logarithm of the nitrosamine concentration as a function of time, (circles
in Figure 64, a first-order loss of nitrosamine is not observed. Presumably
this is because of the recombination of the amine radical with nitric oxide.
(CH )2N + NO •*- (CH )2N-NO (*0
Photolysis of Dimethylnitrosamine with Nitric Oxide in Nitrogen--
To test this contention, a sample of nitrosamine was photolyzed in the
presence of 5 ppm nitric oxide. This should enhance reaction k, sup-
pressing the radical concentration and lowering the rate of nitrosamine
decay. The expected behavior is observed, and the rates of nitrosamine
photolysis with and without nitric oxide are compared in Figure 63.
231
-------
1.0
'CDMN30
^s.
n
Q0.9
-A **** A
_° *** A*
o
- 6°<£°o 0 *** A A
^>°6 A A ^ AA
— o o A
— OO ^ A ^
o A
O
O Q
— o v o
0 o°
0° 0
o °0oo
I I I I I I °!
0 20 40 60
Time, min
Figure 63. Plot of the disappearance of dimethyl-
nitrosamine as a function of photolysis time.
Circles, [DMT] = 3-96 ppm, [NO] = 0; triangles,
[DMN] = 3.16 ppn, [WO] =5.0 ppm.
232
-------
o o
oo
o0o
00
OB
0 30
Time, min
60
Figure 64. Plot of the logarithm of dime thy Initrosamine
concentration as a function of time. Closed symbols
indicate data taken before photolysis was begun. Circles,
nitrosamine in nitrogen; triangles, nitrosamine with 50
pptn isobutane, slope = -0.0046 min"1; squares, nitro-
samine with 6.4 ppm trimethylsilane, slope = -0.0061
man
-l
233
-------
In addition, a plot of 1/[DMN] as a function of time gives a straight
line (Figure 65). This is consistent with a rapid equilibrium followed by
bimolecular decay of the radical.
hv
(CH3)2H-NO •*-; (CH3)2N + WO (fast)
2(CH3)2N - *- products (slow) (23)
In this event -the radical concentration at any time is given by the equilibrium
expression.
k,[DMN]
In the case where 5 PP31 °f nitric oxide is present initially, its concen-
tration will remain fairly constant during the photolysis. The radical
concentration will then be proportional only to the concentration of the
nitrosamine, and the behavior of the radical will be reflected by the
behavior of the nitrosamine.
Photolysis of Dimethylnitrosamine with Trimethylsilane and Isobutane in
Nitrogen —
One way to measure the photolysis rate of DMN is to add another species
that will react with the dimethylamino radical much faster than it reacts
with NO. Two candidates for this added species are trimethylsilane (TMS)
and isobutane (IBU), either of which may react to donate a hydrogen atom.
(CH3)2N + (CH3)3CH >- (CH3)2WH + (CE3}3C (25)
(CHs)2N + (CH3)3SiH *- (CHs)2m + (CHs)3Si (26)
Data from these two experiments were used to estimate the concentration-
time profiles, and the logarithm of nitrosamine concentration as a function
of time is plotted in Figure 6k. The slopes of the linear portions of the
two graphs give rate constants of 0.006l min"1 and 0.00^6 min""1 for the
experiments with TMS and IBU, respectively. These low values, together
with the fact that the concentrations of nitrosamine at the onset of photo-
lysis predicted by these lines are considerably lower than the actual con-
centration measured before photolysis, indicate that an appreciable fraction
of the radicals are still reacting with NO to reform the nitrosamine. The
use of larger concentrations of either TMS or IBU was not practical because
of their strong infrared absorbances.
Estimation of Photolysis Rate Constant from Initial Slope Data __
An estimate of the photolysis rate constant k, can be obtained by con-
sidering only the initial data and extrapolating back to the onset of photo-
lysis when the concentrations.) of radical and/ or nitric oxide are zero.
This technique is illustrated in Figure 66 for the photolysis of DMN in
nitrogen, giving a value of k, = O.Qk-2 min"1 in this case.
23^
-------
035
£
Q.
a
034
033
O32
031
40
Time, min
80
Figure 65* Plot of the reciprocal concentration
of dimethylnitrosamlne as a function of time for
the photolysis of DMN with ^.0 ppm nitric oxide
in nitrogen.
235
-------
137
a
a
Q
1.35
1.33
— \
"— \
o
Time, min
Figure 66. Initial data from the photolysis of
DMN in nitrogen illustrating determination of
initial slope (shown as dashed line).
236
-------
Values obtained using this technique for the experiments indicated are
listed in Table 3k- The deviation from first-order behavior is considerable
even at the shortest times measured here, and so the values given are subject
to the prejudices of the person making the measurements. However, as the
concentration of oxygen increases, the absolute value of the initial slope
increases, indicating that the oxygen acts to remove the radical. Thus the
larger values (i.e., 0.11 - 0.12 min-1) should be closest to the true photo-
lysis rate constant.
Theoretical Estimation of the Photolysis Rate from Ultraviolet Absorption
and Quantum Yield Data --
The photochemical rate constant for a species X in the cell is pro-
portional to the integral
F = / "2 €(A)
•"• / \~i
where e(A) is the extinction coefficient at wavelength A, $(A) is the quantum
yield for the decomposition, and I°(A) is the relative light intensity. The
ratio of two rate constants is therefore equal to the ratio of their integrals,
or
Thus the rate constant for photolysis of DMN can be estimated from its
ultraviolet absorption spectrum and the relative light intensity by com-
parison to a system whose photochemical decomposition is more easily measured.
One such system is the photolysis of W2 at low pressure. Absorption
coefficients35 and quantum yield data36 are available in the literature,
and the primary photolysis rate constant, k , is readily measured since,
at low pressures of nitrogen, only reactions 29 and 30 are important.31
N02 + hv •*- HO + 0 (29)
N02 + 0 -*- NO + 02 (30)
In this case, two molecules of N02 are destroyed for each molecule destroyed
photochemically, and the primary photolysis rate constant k^ is given by
the relationship k = (l/2t)ln(Ao/A), where AQ and A are the values of the
absorbance due to M)2 at time zero and time t, respectively.
Data from the photolyses of M)2 at low pressures were determined and
the Figure 6? consists of a plot of the logarithm of N02 absorbance as a
function of time for each run. The slopes of these plots give values of
k of 0.280 ± 0.00lf(2
-------
TABLE 3^- VALUES OF THE RATE CONSTANT FOR THE PHOTO-
LYSIS OF DIiyiETHYLNITROSAMINE OBTAINED FROM THE INITIAL
SLOPE OF THE LOGARITHM OF THE DIMETHYLNITROSAMINE CON-
CENTRATION AS A FUNCTION OF TIME
0
0
0
0
0
0
0
0
0
fc,
min
.042
.081
.084
.112
.122
.008
.014
.057
.056
4.0
4.0
2.9
1.8
11.
8.2
5.9
5.9
2.9
ppm
ppm
ppm
ppm
Re act ants
DMN
DMN, 50 ppm isobutane
DMN, 50 Torr 0
DM51, 140 Torr 0
4 ppm DMN, 730 Torr O
ppm
ppm
ppni
ppm
DMN, 5.5 ppm NO, 69 Torr O
DMN, 5.2 ppm NO, 138 Torr 0
DMN, 3.7 ppm NO, 1.2 ppm NO., 700 Torr 0
DMN, 20 ppm HO, 4.8 ppm NO , 145 Torr 0.
Expt #
770917-1
770929
770919
770917-2
770930-2
770930-1
771003-1
771003-2
770920-1
238
-------
•r -1
E
o
in
N
in
C\J
O
-
-3
0
1 2
Time, min
Figure 67. Plot of In (absorbance) for three low-
pressure photolyses of nitrogen dioxide. Circles,
slope = -0.559 ± 0.008(2cr) min"1; triangles, slope =
-0.505 ± 0.010(2cr) min-1; squares, slope = -0.519 ±
0.010(2cr) min"1.
239
-------
Table 35. shows the light intensity, extinction coefficient and quantum
yield data used in the estimation of the photolysis rate of DMN in our
photolysis cell. A quantum yield of unity is assumed for the nitrosamine,
and the integral of Equation 29 is approximated by the sum at 1 nm increments.
The following values were obtained:
F..T = 51693 (arbitrary units)
F^,.,- = 15183 (arbitrary units)
From equation 28, ^ = (15l83/5l693)(0.264 min'1)
= 0.078 min""1
This value is lower than that predicted by the method of initial rates.
The most probable explanation is that the light intensity within the cell
has decreased during the course of this work. The measurement of kpq =
0.264 min"1 was made about six months after the bulk of the nitrosamine
photolyses were performed. Measurements of the same rate constant made in
the same way some two years earlier gave values of k? = 0.515 min"1. An
intermediate value of k?Q = 0.40 min"1 gives k.. = 0.12 min-1, in line with
the values predicted from the initial rate data.
Photolysis of Dimethylnitrosamine in the Presence of Oxygen --
The relative rates of the reactions of dialkylamino radicals with
oxygen and oxides of nitrogen at the concentrations typical of polluted
atmospheres are not now known, although some workers have assumed that the
reaction with oxygen is small enough that it can be ignored.21 If this is
the case, then the photolysis of DMN in nitrogen/oxygen mixtures should be
largely independent of the amount of oxygen present. By comparison of the
work in the previous section (DMN photolyzed in nitrogen) with the results
of Hanst et al.19 in air, it can be seen that this is not the case. Thus
if we are to understand the chemistry of dimethylnitrosamine in the atmo-
sphere, it is imperative that a direct experimental estimate of the relative
importance of the reactions of the radical (CHs)2N with oxygen and the
oxides of nitrogen be made.
Formation of Products: The Mechanism of Photooxidation—
The two main products formed during the initial stage of the photolysis
of DMN in the presence of oxygen are dimethy Initramine21'22 (DMNA) and a
species having a sharp absorbance at 1025 cm"1. This absorbance has been
tentatively assigned to tetramethylhydrazine ((CHs)2N-N(CHs)2) by Tuazon
et al.,22 but our work shows it to be due to monomethylmethyleneamine (MMA.).
This assignment was verified by comparison with the spectrum of an authentic
sample.
240
-------
TABLE 35- EXTINCTION COEFFICIENTS, QUANTUM YIELDS AND
VALUES OF THE LIGHT INTENSITY WITHIN THE CELL USED IN
THE DETERMINATION OF THE PHOTOLYSIS RATE OF DIMETHYL-
NITROSAMINE
X,
nun
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
Nitrogen dioxide Dimethylnitrosamine
I°(X),a e(X),b *{X),C F(X),d e(X),e F(X),d
i/mol-cm 1/mol-cm
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.01
0.02
0.04
0.05
0.08
0.10
0.11
0.13
0.15
0.19
0.23
0.27
0.30
0.30
0.32
0.38
0.42
0.50
0.55
0.60
0.72
0.85
0.96
1.10
1.20
1.30
1.42
1.52
1.65
1.75
1.87
2.02
2.19
2.30
2.55
33.5
35.1
39.2
45.5
45.6
47.3
45.2
46.6
46.2
52.5
50.2
53.7
56.0
58.1
55.5
64.3
60.9
66.5
70.9
66.1
72.4
75.7
75.6
79.0
76.4
79.5
82.2
83.1
87.8
85.6
85.3
87.1
85.8
106.4
85.1
98.6
100.2
98.9
99.3
113.9
110.8
118.9
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.990
0.985
0.985
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.46
0.91
2.08
2.48
4.25
5.54
6.33
7.14
9.55
11.46
15.14
18.95
19.63
21.50
23.98
28.44
32.85
37.82
43.29
48.83
59.23
73.88
81.35
92.89
103.47
110.42
149.58
128.06
161.06
173.60
183.09
198.58
246.95
251.02
295.65
2.40
2.28
2.28
2.03
2.28
1.90
2.28
2.40
2.28
2.53
2.66
3.29
3.42
3.29
3.42
3.93
4.94
5.45
6.08
6.46
7.09
8.49
8.87
10.26
11.27
11.91
13.30
13.93
15.20
16.85
17.73
18.62
19.63
20.65
21.91
23.31
24.95
26.60
28.50
30.53
32.30
33.44
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.03
0.04
0.10
0.13
0.27
0.34
0.36
0.44
0.59
0.94
1.25
1.64
1.94
2.13
2.72
3.37
4.31
5.64
6.55
7.98
10.03
12.92
16.17
19.51
22.34
25.52
29.32
33.31
38.45
43.67
49.74
57.57
66.85
74.29
85.27
(Continued)
-------
TABLE 35. (Continued)
Nitrogen dioxide
e(X),b *(X),C F(X).d
i/mol-cm
DimethyInitrosamine
e(X),e F(X),d
i/mol-cm
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
2.68
2.83
2.92
3.13
3.25
3.45
3.60
3.70
3.88
4.15
4.30
4.49
4.62
4.73
4.88
5.05
5.18
5.35
5.55
5.75
6.20
6.25
6.40
6.49
6.60
6.65
6.69
6.70
6.72
6.65
6.52
6.46
6.42
6.36
6.30
6.27
6.22
6.17
6.13
6.12
6.09
6.08
6.06
109.4
101.2
115.9
116.2
122.6
122.1
137.6
131.7
117.0
129.0
126.7
113.9
143.9
146.5
131.5
159.3
143.8
130.0
128.8
153.8
143.9
146.2
139.1
165.0
154.1
148.0
152.5
147.9
154.7
148.8
170.8
157.1
153.0
152.8
178.0
161.9
147.7
156.1
170.9
161.6
160.9
153.4
170.3
0.985
0.985
0.985
0.984
0.984
0.984
0.984
0.984
0.983
0.983
0.983
0.983
0.983
0.982
0.982
0.982
0.981
0.981
0.980
0.979
0.978
0.977
0.976
0.975
0.974
0.973
0.972
0.971
0.970
0.969
0.968
0.967
0.966
0.965
0.964
0.963
0.962
0.961
0.960
0.958
0.956
0.954
0.952
288.74
282.10
333.35
357.89
392.07
414.51
487.43
485.97
446.24
526.25
535.55
502.72
653.52
680.47
630.17
789.98
730.73
682.29
700.54
865.78
872.55
892.73
868.87
1044.08
990.62
957.63
991.66
962.19
1008.40
958.84
1077.98
981.38
948.86
937.79
1081.03
977.55
883.78
925.57
1005.71
947.45
936.77
889.77
982.48
34.83
35.97
36.73
38.13
39.39
41.80
44.97
48.13
51.42
54.21
55.48
55.73
54.72
54.21
54.08
53.70
55.10
57.12
60.54
64.98
70.04
72.70
73.34
71.31
68.40
64.22
60.42
57.50
54.84
53.20
53.07
54.08
57.25
60.42
62.44
61.96
59.28
55.22
51.55
46.61
43.57
39.39
35.72
93.35
101.80
107.26
119.34
128.02
144.21
161.87
180.49
199.53
224.98
238.56
250.23
252.80
256.42
263.93
271.21
285.41
305.62
336.02
373.62
434.27
454.40
469.36
462.81
451.42
427.05
404.20
385.28
368.56
353.77
346.03
349.39
367.56
384.26
393.40
388.35
368.71
340.73
316.01
285.26
265.35
239.51
216.45
(Continued)
-------
TABLE 35- (Continued)
X,
mn
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
Nitrogen dioxide
I°(X),a e(X),b *U),C F(X),d
i/nol-cm
6.02
6.00
5.99
5.95
5.90
5.85
5.78
5.70
5.62
5.55
5.48
5.42
5.35
5.27
5.22
5.21
5.22
5.45
5.68
5.75
5.86
5.87
5.73
5.63
5.40
5.05
5.05
4.75
4.60
4.48
4.45
4.40
4.35
169.6
157.2
159.9
170.7
171.9
171.3
166.4
172.7
155.7
158.2
168.1
175.4
161.7
182.9
160.9
193.0
186.3
163.0
145.7
173.2
180.4
153.9
135.0
170.2
168.4
164.8
164.0
164.0
164.0
164.0
164.0
163.0
163.0
0.950
0.948
0.946
0.944
0.942
0.940
0.935
0.933
0.930
0.925
0.920
0.905
0.885
0.860
0.800
0.680
0.600
0.500
0.440
0.360
0.330
0.280
0.230
0.180
0.160
0.150
0.120
0.080
0.060
0.050
0.030
0.020
0.000
969.94
894.15
906.08
958.79
955.37
941.98
899.28
918.44
813.78
812 . 16
847.49
860.25
765.61
828.94
671.92
683.76
593.49
444.18
364.13
358.52
348 . B6
252.95
177.92
172.48
145.50
124.84
99.38
62.32
45.26
36.24
21.89
14.34
0.00
Dime thy Initrosandne
e(X), F(X).d
4/mol-cm
32.43
29.51
26.35
23.69
21.53
19.63
17.73
16.21
14.19
13.17
11.40
10.89
9.37
8.61
7.60
6.46
6.08
5.32
4.43
4.31
3.67
3.17
2.91
2.53
2.03
1.77
1.90
1.52
1.27
1.14
0.89
0.89
0.89
195.20
177.07
157.81
140.93
127.04
114.85
102.50
90.25
79.73
73.11
62.47
59.04
50.15
45.39
39.67
33.65
31.74
26.99
25.18
24.76
21.53
18.59
16.69
14.26
10.94
8.96
9.59
7.22
5.83
5.10
3.95
3.90
3.86
Unpublished work of Eugene Beer; arbitrary units.
bA. M. Bass, A. E. Ledford, Jr., A. H.-taufer, J. Res. Nat. Bur.
Stand., 80A, 143(1973).
Values from J. G. Calvert's curve from the data of I. T. N. Jones and
K. D. Bayes, J. Chem. Phys., 59_, 4836(1973).
Arbitrary units.
-------
Concentration/time profiles of several experiments are shown in
Figures 68 through 75. Figures 68 through 70 depict photolyses of pure DMN
samples in synthetic nitrogen/oxygen mixtures of varied composition. Figures
70 through 74 illustrate a similar series of experiments where 5 PP11 NO was
added initially. General observations from these data include the following.
1. The loss of DMN is enhanced by both 02 and N02, but in-
hibited by HO.
2. The formation of DMNA is enhanced by N02, but inhibited by
NO (cf. Figures 69, 73 and 75).
3. The formation of MMA is enhanced by 02, but inhibited by NO.
k. The formation of HONO is enhanced by both oxygen and NOX.
The simplest mechanism which can account for the products found ini-
tially and the general trends noted above includes the following reactions
of DMN and the dimethylamino radical.
(CH3)2N-NO + hv -»~ (CH3)2N + NO (l)
(CH3)2N + 02 -*• CH2=N-CH3 + H02 (3)
(CH3)2N + NO *- (CH3)2N-NO (4)
(CH3)2N + N02 -*- (CHs)2N-N02 (5a)
If these are the only important reactions of DMN and the radical
(CH3)21T, then the rates of change for species involved can be expressed as
follows.
= k^DMN]
/k [02] + k [N02]\
= -k [DMN] I -^ —2. 1 (3la)
J
(32)
at - •"•-,!_— «J \ TT / (33)
where
-------
a
a
•V
c
o
CTJ
U
C
o
u
100
Time, min
Figure 68. Concentration-time profile from the
photolysis of dimethylnitrosamine in a mixture of
49 Torr oxygen and 651 Torr of nitrogen. Open
symbols indicate data taken during photolysis;
circles, dimethylnitrosaminej upright triangles,
dimethyInitraminej inverted triangles, mono-
methylmethyleneaminej squares, nitrous acid;
diamonds, nitric oxide; hexagons, nitrogen dioxide,
-------
Or—
E
Q_
Q.
*
C
O
C
-------
E
a
a
cf
o
05
-------
O
30 60
Time, min
Figure 71. Concentration-time profile from the
photolysis of dimethylnitrosamine with $.k ppm
nitric oxide in a mixture of 23 Torr oxygen and
67? Torr nitrogen; symbols are same as Figure 68.
-------
o-o
D—D—D-q—0—0—O—O
o
50 100
Time, min
150
Figure 72. Concentration-time profile from the
photolysis of dimethylnitrosamine with 5«5 PP1
nitric oxide in a mixture of 69 Torr oxygen and
621 Torr nitrogen; symbols same as Figure 68.
-------
E
CL
Q.
Cti
L.
•*->
U
c
o
U
0
60 120
Time, min
Figure 73. Concentration-time profile from the
photolysis of dimethylnitrosamine with 5-5 PP31
nitric oxide in a mixture of 138 Torr oxygen and
562 Torr nitrogen; symbols as in Figure 68 .
250
-------
E
Q.
Q_
05
-*-»
C
-------
30 60
Time, min
Figure 75. Concentration-time profile from the
photolysis of dimethylnitrosamine with 25 ppm
nitric oxide in a mixture of 1*4-5 Torr oxygen
and 555 Torr nitrogen; symbols same as in
Figure 68.
252
-------
Determination of kjYk- —
From th- rate equations, the following expressions can be derived.
From equations 31a, 31b, and 32,
_ -d[DMM1/dt - _
+ d[DMN]/dt = B (35)
and from equations 31b and 33,
dfMMAI/dt
d[DMN]/dt
(36)
The quantities B and B' can be calculated from values of the concentrations
obtained experimentally and a knowledge of the photolysis rate constant k, .
Only data from the first few minutes of the photolysis were used in
the calculation of B and B' . At later times, secondary reactions of the
products DMNA and MMA become important, as one might expect. This is
especially evident in the case of MMA., which is seen to decrease at later
times in most of the photolyses.
The results of these calculations for several experiments using k, =
0.12 min"1 are listed in Table 36. A plot of B and B/ as a function of [02]/
[NO] should have a slope of k_/k. and an intercept of zero. This is il-
lustrated in Figure 76. The reciprocal of the slope of the line shown gives
k^/k
= (6.76 ± 0.33(2
-------
TABLE 36. CALCULATED VALUES OF B AM) B'' USED IN THE
ESTIMATION OF k^/k . CONCENTRATIONS LISTED ARE TAKEN
FROM GRAPHS OF THE DATA FROM THE EXPERIMENTS IN-
DICATED
time,
min
[DMN],
ppm
[DMNA] ,
ppm
Experiment
43.00
44.00
45.00
46.00
47.00
48.00
49.00
50.00
51.00
52.00
53.00
54.00
55.00
2.570
2.438
2.330
2.236
2.154
2.087
2.026
1.972
1.925
1.882
1.843
1.807
1.772
0.074
0.121
0.161
0.195
0.222
0.246
0.268
0.284
0.296
0.306
0.314
0.320
0.325
Experiment
33.00
34.00
35.00
36.00
37.00
38.00
1.342
1.234
1.152
1.077
1.012
0.952
0.118
0.153
0.182
0.207
0.229
0.247
Experiment
42.00
43.00
44.00
45.00
46.00
47.00
48.00
8.73
8.13
7.61
7.16
6.75
6.36
6.02
0.740
0.947
1.136
1.300
1.456
1.590
1.712
Experiment
32.00
33.00
34.00
35.00
36.00
37.00
38.00
40.00
42.00
8.244
8.187
8.128
8.071
8.020
7.976
7.933
7.842
7.752
0.007
0.043
0.072
0.095
0.116
0.132
0.148
0.176
0.200
[MMA]
ppm
number
0.142
0.211
0.265
0.310
0.349
0.382
0.412
[NO] ,
ppm
770919.
0.18
0.26
0.33
0.38
0.42
0.46
0.49
0.53
0.56
0.59
0.62
0.64
number 770917-2.
0.299
0.364
0.18
0.23
0.28
0.30
0.33
0.36
number 770930-2.
1.910
2.320
2.645
2.900
3.100
1.40
1.67
1.87
2.03
2.15
2.27
2.38
nvimber 770930-1.
5.50
5.52
5.54
5.56
5.58
5.59
5.61
5.65
5.69
[NO]
49 Torr 0,.
2.93
2.19
1.82
1.61
1.47
1.36
1 .27
1. 19
1.13
1.07
1 . 02
0.98
140 Torr C
9.01
7.31
6.40
5 . 84
5 . 38
Ba
.0.51
0.38
0.33
0.30
0.23
0.21
0 • 20
0 . 19
0.18
0.17
0 . 16
0.17
P2'
1.57
0.87
0.85
0.71
0 . 73
B-b
0.41
0.30
0.25
0.22
0.18
0.16
1.40
730 Torr 0 .
6.26
5.43
4.93
4» 60
4.35
4.13
69 Torr O.
^
0.165
0.164
0.164
0.163
0.163
0.162
0.161
0.160
0.96
0.78
0.66
0 * 60
0.645
0.54
,.
0.023
0.033
0.037
0.033
0.030
0.030
0.035
0.037
1.00
0.77
0.58
( Continued)
254
-------
TABLE 36. (Continued)
time,
min
1DMN],
ppm
[DMNA]
ppm
Experiment
34.00
35.00
36.00
37.00
38.00
39.00
40.00
5.561
5.494
5.436
5.378
5.322
5.271
5.219
0.097
0.129
0.157
0.182
0.204
0.225
0.244
[MMA] , [NO] ,
ppm ppm
number 771003-1.
5.22
5.22
5.22
5.22
5.23
5.23
"[NO]*10
138 Torr 0
0.353
0.353
0.353
0.352
0.352
0 352
B* B.»
2'
0.059
0.050
0.056
0.061
0.051
0.057
B is defined in equation 35.
B' is defined in equation 36.
255
-------
CO
c
CO
m
(CO23/CNO])x10'
Figure 76. Plot of B and Bx as a function of [02]/
[NO] used to estimate k./k~; open symbols, B;
closed symbols, B'.
-------
TABLE 37. CALCULATED VALUES OF C
ESTIMATION OF k- /k
pa-
AND k- /k».
AND C ' USED IN THE
CONCENTRATIONS
LISTED ARE TAKEN FROM GRAPHS
EXPERIMENTS INDICATED
OF THE DATA FROM THE
time.
Bin
tDMN],
ppm
(DMNA]
ppm
Experiment
42.00
43.00
44.00
45.00
46.00
8.73
8.13
7.61
7.16
6.75
0.740
0.947
1.136
1.300
1.456
Experiment
34.00
35.00
36.00
37.00
38.00
39.00
40.00
41.00
5.755
5.430
5.155
4.920
4.730
4.550
4.395
4.245
0.066
0.291
0.462
0.600
0.715
0.802
0.888
0.965
Experiment
38.00
39.00
40.00
41.00
42.00
43.00
44.00
45.00
2.812
2.644
2.501
2.380
2.271
2.169
2.081
1.995
•0.052
0.178
0.298
0.408
0.508
0.597
0.663
0.728
[MMA],
ppm
[N02], _[C
ppm [I
number 770930-2. 730
1.910
2.320
2.645
2.900
3.100
0.182
0.229
0.264
0.295
0.318
number 771003-2. 700
0.040
0.139
0.230
0.316
0.395
0.471
0.542
0.606
1.11
0.97
0.86
0.765
0.685
0.620
0.560
0.510
number 770920-1. 145
4.880
4.742
4.618
4.507
4.409
4.318
4.228
4.148
^io'€
i^]*iu
Torr 0, .
4.67
3.90
3.44
3.13
Torr 0.
0.89
1.01
1.13
1.27
1.41
1.56
1.72
Torr O
0.040
0.041
0.042
0.043
0.044
0.045
0.046
; ca
1.90
1.75
1.74
1.63
0.59
0.61
0.70
0.65
1.07
0.80
0.95
0.33
0.19
0.10
0.09
0.15
0.33
0.32
c-b
1.98
1. 72
1.55
1.28
0.48
0.53
0.62
0.69
0.87
0.83
0.83
3C is defined in equation 37.
C' is defined in equation 38.
257
-------
The quantities C and C/ are plotted as a function of [02]/[I\T02] in
Figure 23. The linear relation between the variables is in qualitative
agreement with expectations. A weighted least squares gives the line
shown, and the inverse of the slope of this line gives k<- /kQ = (2.^-0 ± 0.17
,?a O
(2cr)) y- 106. A non-weighted least squares give k /k« = (2.56 ± O.l8) x 10G.
pa 3
Ref inement of the Mechanism —
Hydrogen Atom Abstraction by Nitrogen Dioxide: The Determination of k /k —
There are two deficiencies in the simple mechanism considered in the
previous section. First, it does not explain the non-zero intercept found
in the plot of C and c' as a function of [02]/[N02]. Secondly, this
mechanism does not predict the quantities of nitrous acid that were ob-
served experimentally.
Both of these shortcomings may be overcome by the addition of a
reaction in which N02 abstracts a hydrogen atom from the dimethylamino
radical.
(CH3)2N + F02 •+- CH2=N-CH3 + HOWO (p~b)
If this reaction is added to the scheme, the following more complete ex-
pression for the relation of C and C/ to the concentrations of reactants
may be derived.
From the intercept of the graph in Figure 77, kc,/kc = O.l^k ± 0.063.
po pa
Dividing by the slope, k /k = (3.7 ± l.l(2cr)) x 105.
The effect of this change in mechanism on B and Bx is also of interest.
Using the new mechanism,
Since the experiments used to evaluate k^/k^ were chosen to be those in
which little N02 was present, this should have little effect on the value
of kj,A obtained.
Evidence for the Direct Oxidation of Dime thy Initrosamine--
Some indication is given that DMN may react directly (without dis-
sociation) with oxidizing species to produce DMNA. Specifically, in the
258
-------
u
T3
C
CO
U
1
o
c
8
D
I
2
I I
4
,-6
Figure 77. Plot of C and C ' as a function of [02]/
[W02] for the estimation of the rate constant ratio
and 1
'k_; open symbols, C; closed symbols,
259
-------
experiment where 11.4 ppm DM was photolyzed in 730 Torr 02, an initial
enhancement of DMNA formation seemed to be present. This can be thought of
as occurring from the subsequent reactions of the H02 radical formed in
reaction 3.
(CHs)2N-NO + hv -*- (CH3)2N + NO (l)
(CH3)2N + 02 -*- CH2=N-CH3 + H02 (3)
When NO is present, either as an added gas or as a photolysis product at
later times, ROS will react with it rapidly to convert it to N02.
H02 + WO ->- OH + N02 (4l)
However, in the beginning of an experiment such as the one under consideration,
no oxides of nitrogen are present, and the H02 may react to oxidize DMN
directly.
H02 + (CHs)2N-NO -»- OH + (CH3)2N-N02 (42)
From the present data, no crucial test of this hypothesis is possible,
although they are consistent with it. A series of photochemical ex-
periments using C12/H2/02/DMN mixtures might be used to test this point
further.
Summary and Conclusions
A Fourier transform infrared spectrometer system was employed to
study the kinetics of the reactions of the (CHs)2N radical in simulated
atmospheres. The radical was prepared by the photolysis of both di-
methylnitrosamine and tetramethyl-2-tetrazene, and by the reaction of OH
with dimethylamine. The following reactions appear to be important in the
determination of nitrosamine levels in the atmosphere.
(CH3)2N-NO + hv -*~ (CH3)2N + NO (1)
(CH3)2NH + OH ->- (CH3)2N + H20 (2a)
^CH3NHCH2 + H20 (2b)
(CH3)2N + 02 ->- CH2=N-CH3 + H02 (3)
(CHs)2N + NO •>- (CH3)2N-NO (4)
(CH3)2N + N02 -*- (CH3)2N-N02 (5a)
-*- CH2=N-CH3 + HONO (5b)
A kinetic analysis of the data gave the following rate constant ratio
estimates:
260
-------
= (6.76 ± 0.33) x 105
k /k = (2.UO ± 0.17) x 10s
pa j
= (3.7 ± 1.1) x 105
These data may be used to make a preliminary estimate of the con
centration of dimethylnitrosamine in a polluted atmosphere.
If the above scheme is valid,
[DMN]
Assuming that a steady state concentration of dimethylnitrosamine is reached,
ku[ito]
+ MO + + hro +
k[°H][NO]
In sunlight irradiated polluted atmospheres, [OH] is generally estimated to
be about 10~7 ppm. Using this value, the value k = 3.67 x 104 ppm-^-minr1
derived from the data of Atkinson et al.,25 and values of the rate constant
ratios and k. derived in this work,
Assuming W and N02 concentrations in the range 0.1 to 0.5 ppm normally
encountered in highly polluted air, then nitrosamine levels of about 1$ of
the dimethylamine in the air will result at the steady state.
Dimethylnitrosamine is not an end product in the photooxidation of
dimethylamine in the atmosphere, since it photolyzes in sunlight to reform
the dime thy lamino radical. The two other products formed in our system
from the reactions of the dimethylamino radical were monomethylmethyleneamine
and dimethylnitramine. Of these two, monomethylmethyleneamine went on to
further oxidation, while dimethyInitramine was reasonably stable.
If this stability of dimethylnitramine carries over to the atmosphere,
as seems likely, the potential for a considerable accumulation of this
compound exists. Using the rate constant ratios derived in this work,
together with the estimation of kg& derived from the work of Atkinson
261
-------
et i.1.,25 the rate of formation of dimethylnitramlne from dime thy lamine in
a polluted atmosphere can be estimated.
r k[N°2]
= (3.67
Even at the moderate N02 concentration of 0.1 ppm, this gives a rate of
conversion of dime thy lamine to the nitramine of about 0.2^/minute. Haus oiir
data predicts that appreciable levels of dime thy Initramine may be built up
where amine levels are high. Since dimethy Initramine has demonstrated
carcinogenic activity in laboratory animals J38s39 these high levels are of
concern. These results suggest that analysis for dimethylnitramine in the
real atmosphere be initiated.
262
-------
REFERENCES
1. P.N. Magee and J.M. Barnes, Brit. J. Cancer, 10, Ilk (1956).
2. D.H. Fine, D.P. Rounbehler, N.M. Belcher, and S.S. Epstein, Science,
192, 1328 (1976).
3. P.N. Magee and J.M. Barnes, Adv. Cancer Res., 10, 163 (1967).
4. B. Terracini, P.N. Magee, and J.M. Barnes, Brit. J. Cancer, 21, 559
(1967). ~~
5. G.E. Moiseyev and V.V. Benemanskiy, Vopr. Onkol., 21, 106 (1975).
6. H. Druckrey, R. Preussman, S. Ivankovie, and D. Schmahl, Z. Krebsforsch.,
69, 103 (1967).
7. P. Walker, J. Gordon, L. Thomas, and R. Ouellette, MITRE Technical
Report $4TR-7152, February 1976.
8. K. Bretschneider and J. Matz, Arch. Geschwulstforsch., 42, 36 (1974).
9. D.H. Fine, D.P. Rounbehler, A. Rounbehler, A. Silvergleid, E. Sawicki,
K. Krost, and G.A. DeMarrais, Environ. Sci. Technol., 11, 58! (1977).
10. D.H. Fine, D.P. Rounbehler, E.D. Pellizzari, J.E. Bunch, R.¥. Berkeley,
J. McCrae, J.T. Bursey, E. Sawicki, and G.A. DeMarrais, Bull. Environ.
Contain. Toxicol., 15, 739 (1976).
11. Atmospheric Nitrosamine Assessment Report. Environmental Protection
Agency Draft, March 12, 1976.
12. R.B. Faoro, J. Air Pollut. Control Assoc., 25, 638 (1975).
13. D. Shapley, Science, 191, 268 (1976).
14. B.R. Simoneit and A.L. Burlingame, Nature (London), 234, 210 (1971).
15. J.W. Rhoades and D.E. Johnson, J. Nat. Cancer Inst., 48, l84l (1972).
16. T. Fazio, J.N. Damico, J.W. Howard, R.H. White, and J.O. Watts, J. Agr.
Food Chem., 19, 250 (1971).
17. D.F. Gadbois, E.M. Ravesi, R.C. Lundstrom, and R.S. Maney, J. Agr.
Food Chem., 23, 665 (1975)-
18. G. Neurath, B. Pirmann, H. Luttich, and H. Wichern, Beitr. Tabakforsch.,
3, 251 (1965).
19. P.L. Hanst, J.W. Spence, and M. Miller, Environ. Sci. Technol., 11,
403 (1977).
263
-------
20. '-,7.A. Ola-son, General Motors Research Laboratory, Warren, Michigan,
private communication with J.G. Calvert.
21. J.i;. Pitts, Jr., D. Grocjean, K.A. Van Cauwenberghe, J.P. Schmid, and
D.R. Fitz, Environ. 3ci. Technol., 12, 9^6 (1978).
22. E.G. Tuazon, A.M. Winer, R.A. Graham, J.P. Schmid, and J.N. Pitts, Jr.,
Environ. Sci. Technol., 12, 954 (1978).
23. J. Althorpe, D.A. Goddard, D.J. Sissons, and G.M. Telling, J. Chromatogr.,
53, 371 (1970).
2k. J.G. Calvert, Report of Ad Hoc Committee on Assessment of Nitrosamines
in the Atmosphere, August 1976, preliminary draft.
25. R. Atkinson, R.A. Perry, and J.N. Pitts, Jr., J. Chem. Phys., 68, 1850
(1978).
26. N.R. Greiner, J. Chem. Phys., 53, 1070 (1970).
27. R. Lesclaux and M. Demissey, to appear in Nouv. J. Chem.
28. R. Lesclaux, P. Vankhe, J.C. Soulignac, and J. Joussot-Dubien, Abstracts,
VII International Conference on Photochemistry, Edmonton, Canada,
August 1975.
29. K. Ya. Kondratzev, "Radiation In the Atmosphere", Academic Press, New
York, N.Y., 1969, p. 225.
30. P.L. Hanst in "Advances in Environmental Science and Technology," Vol.
II, J.N. Pitts, Jr. and R.L. Metcalf, eds., Wiley Interscience, New
York, N.Y., 1971, pp. 154-164.
31. C.H. Wu and H. Niki, Environ. Sci. Technol., 9, k6 (1975).
32. C.H. Bamford, J. Chem. Soc., 12 (1939).
33. E.G. Gowenlock, F. Jones, and J. Mayer, Trans. Faraday Soc., 57? 23
(1961). ~~
3^. B.L. Korsunskii, V.N. Pepkin, Yu. A. Lebevdev, and A. Ya. Apin, Izv.
Akad. Nauk SSSR, Ser. Khim., 525 (1967).
35. A.M. Bass, A.E. Ledford, Jr., and A.H. Laufer, J. Res. Nat. Bur. Stand.,
80A, 143 (1976).
36. I.T.N. Jones and K.D. Bayes, J. Chem. Phys., 59, ^836 (1973).
37. E.G. Gowenlock and K.E. Thomas, J. Chem. Soc., B, ^09 (1966).
264
-------
38. H. Druckrey, R. Preussmann, D. Ochmahl, and M. Mailer, Waturwissensehaften,
Ij8, 13^ (1961).
39. C.M. GoodaH and T.H. Kennedy, Cancer Lett., 1, 295 (1976).
265
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
\- =E?ORT NO. |2.
EPA-600/2-80-024 |
- TIlLE ANDSUBTITLE
KINETIC STUDIES OF SIMULATED POLLUTED ATMOSPHERES
7. AUTHORISi
Jack G. Calvert
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Department of Chemistry
The Ohio State University
Columbus, Ohio 43212
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Sciences Research Laboratory-RTF, NC
Office of Research and Development
U.S. Environmental Protection Agency
ii Research Triangle Park, North Carolina 27711
3. RECIPIENT'S ACCESSION- NO.
5. REPORT DATE
January 1980
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT NO.
1AA603A AC-24 (FY-78)
11. CONTRACT/GRANT NO.
R804348-01
13. TYPE OF REPORT AND PERIOD COVERED
Final 1/76 - 4/79
14. SPONSORING AGENCY CODE
EPA/600/09
|15, SUPPLEMENTARY NOTES
\
16. ABSTRACT
The kinetics and reaction mechanisms of several important atmosheric con-
taminants - SO-, formaldehyde, nitrous acid, and the nitrosamines - were assessed
to help quantify some key aspects of the chemistry of polluted atmospheres. The
reactions and lifetimes of excited sulfur dioxide with various atmospheric com-
ponents including hydroxyl, hydroperoxy, and methylperoxy radicals were studied.
These data and other published rate data were reviewed and evaluated. The photolysis
of formaldehyde was investigated as a major source of hydroperoxyl radicals, and a
quantitative evaluation made of its apparent first order rate constants at various
solar zenith angles. The absolute extinction coefficients for nitrous acid were
determined, and estimates made of the rates of hydroxyl radical generation in the
troposphere by photolysis of nitrous acid. Long path Fourier transform infrared
spectroscopy was used to help evaluate the potential for nitrosamine formation in
the polluted atmosphere.
•I
>;
6
»17. KEY WORDS AND DOCUMENT ANALYSIS
;a. DESCRIPTORS
"A"Air pollution
i *0zone
*Nitrogen oxides
*Sulfur inorganic compounds
^Photochemical reactions
13. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
b. IDENTIFIERS/OPEN ENDED TERMS
19. SECURITY CLASS (This Report/
UNCLASSIFIED
20. SECURITY CLASS (This page)
UNCLASSIFIED
c. COS ATI Field/Group
13B
07B
07C i
07E
21. NO. OF PAGES
280
22. PRICE
EPA Form 2220-1 (9-73)
266
------- |