Ecological Research Series
MECHANISM OF PHOTOCHEMICALLY
INITIATED OXIDATIONS
Environmental Sciences Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into five series. These five broad
categories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ECOLOGICAL RESEARCH series. This series
describes research on the effects of pollution on humans, plant and animal
species, and materials. Problems are assessed for their long- and short-term
influences. Investigations include formation, transport, and pathway studies to
determine the fate of pollutants and their effects. This work provides the technical
basis for setting standards to minimize undesirable changes in living organisms
in the aquatic, terrestrial, and atmospheric environments.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
BISCEAIMER
This report has been reviewed by the Environmental Sciences Research •
Laboratory, 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 connercial products consti-
tute endorsement or reconmendation for use.
ii
-------
EPA-600/3-76-070
June 1976
MECHANISM OF PHOTOCHEMICALLY
INITIATED OXIDATIONS
Jack G. Calvert
Chemistry Department
The Ohio State University
Colxirbus, Ohio 43210
Grant No. R800398
Project Officer
Joseph J. Bufalini
Atmospheric Chemistry and Physics Division
Environmental Sciences Research Laboratory
Research Triangle Park, N.C. 27711
U. S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF RESEARCH AND DEVELOPMENT
ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
RESEARCH TRIANGLE PARK, N.C. 27711
-------
MSCIAIMER
This report has been reviewed by the Environmental Sciences Research
Laboratory, 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 ccmnercial products consti-
tute endorsement or recommendation for use.
ii
-------
ABSTRACT
Several significant new observations have been made relative to chemical
reactions that occur in sunlight-irradiated Npx/hydrocarbon/aldehyde/CX)/S02
polluted atmospheres. Many of the primary reactions that are needed to quanti-
tatively evaluate the photooxidation mechanisms of S02 in the atmosphere were
measured. Rate constants for the reactions of the excited S02(%i) state of
SO2 with various atmospheric gases, alkanes, aUcenes, NO, 00, etc., were
determined. In view of these results, the rate of SO2 photooxidation in the
atmosphere is estimated, and the possible role of excited-SO2/alkene inter-
actions that generate aerosols is evaluated. Rate constants for the homo-
geneous reaction of SO2 with 03, N03, and N20s were also estimated. All of
these reactions are relatively slow for conditions that usually exist in
polluted atmospheres. The unusual reaction of SQ3 with N02 was observed,
although its importance in the atmosphere cannot be evaluated accurately
from the existing data. An evaluation was made of the photochemical smog
mechanisms using a computer to simulate the rates of change in various
polluted atmospheres. Several important features of special interest in
developing control strategies were observed.
iii
-------
Table of Contents
Introduction 1
Part I. The Atmospheric Removal Paths of S02 2
I-A. The Efficiency of S02 Triplet Formation Induced in the Excited
Singlet S02 by Collisions with 02 and Other Atmospheric Gases. 3
1. Experimental Methods and Results 3
2. Discussion of Results o 6
The mechanism of 3S02 formation in the 2^00-3200 A
irradiation of SOg 6
Estimation of the inter system crossing ratio in -"-SOg-M
collisions with various partners 7
Effect of the nature of the collision partner on the
intersystem crossing rate in excited singlet sulfur
dioxide 9
Efficiency of 3S02 formation in N2- and CO-containing
mixtures 9
"Excess" triplet in the 3130 A-irradiated S02-C0
mixtures at high pressures 11
The theoretical maximum rates of S02 photo-oxidation
in sunlight-irradiated, S02-containing atmospheres 11
I-B. A Study of the Nature of the SQa Excited States Formed at
High Added Gas Pressures 11
1. Experimental Methods and Results 12
2. Discussion lU
The "excess" biacetyl phosphorescence from 2875 A-
irradiated S02-biacetyl mixtures at high added-gas
pressures Ik
Treatment of the data in terms of the Wampler,
Horowitz, Calvert mechanism IB
Treatment of the present data in terms of the Cehelnik,
Spicer, Heicklen mechanism of S02 photochemistry 19
The nature of the intermediate species involved in S02
photochemistry at high added-gas pressures 21
The theoretical maximum rate of 50s photo-oxidation in the
lower atmosphere 24
I-C. The Temperature Dependence and the Mechanism of the S02(3Bi)
Quenching Reaction 25
1. Experimental Methods 26
2. Results and Discussion 27
The mechanism of S02(3Bi) quenching 27
Mechanism of S02(3Bi) quenching by the atmospheric gases 30
The SQ2(3Bi) quenching by the paraffinic hydrocarbons 39
The S02(3Bi) quenching by NO and by the olefinic and
aromatic hydrocarbons kO
-------
Table of Contents (continued)-
I-D. A Kinetic Study of the Photoexcited S02(3B1)-Alk.ene
Reactions ^2
1. Experimental ' *t-3
2. Discussion of Results kl
Mechanism of S02(3Bi) Induced cis-trans isomerization of
2-butenes o 47
Mechanism of aerosol formation in 3500-4100 A irradiated
SO^-2-butene mixtures 5^
I-E. The Mechanism and Kinetics of the Alkene-SOp Reactions Excited
Within the First Allowed Band of S02 (3130 A) 56
1. Experimental Procedures and Results 57
2. Discussion 57
The simple two-state reaction mechanism of cis-2-butene
isomerization sensitized by electronically excited S02 57
The isomerization mechanism in experiments at low
[S02]/[C4Ife] ratios 62
The isomerization mechanism in experiments with high
pressures of added C&2 67
I-F. The Mechanism and Kinetics of S03 Formation in the Photolysis
of SQa Mixtures 69
1. Experimental 70
2. Discussion of the Results 76
Dependence of quantum yields of S03 formation on reactant
flow rates 76
Mechanism of SOs formation in 3600-14-000 A S02(3Bi)
Excitation experiments 8l
Mechanism of S03 formation in S02 excited at 3130 A 83
Conclusions 86
I-G. The Nature of the Excited Singlet States in S02 Photolysis
Within the First Allowed Absorption Band 87
Part II. Some Thermal Reactions of Possible Significance in the Sunlight-
Irradiated, Polluted Troposphere 92
II-A. A Kinetic Study of the SQ2-03, S02-NOs, and S02-N2C>5
Reactions 92
1. Experimental 93
2. Discussion of the Results 95
Estimates of the rate constants for the N0s-S02 and
N20s-S02 reactions 95
The predicted influence of N2C>5 photodecomposition on
the levels of N205 and N03 in the sunlight-irradiated
polluted urban atmosphere 98
The reaction between gaseous S03 and N02 98
-------
Table of Contents (continued)
II-B. The Reaction of 03 with Perfluorinated Polyolefins . . . . . .99
1. Experimental ' 99
2. Results and Discussion 101
Part III: The Computer Simulation of the Rates of Chemical Reactions
in Sunlight-Irradiated Polluted Atmospheres .......... Ill
III-A. The Relative Importance of Various Active Intermediates in the
Attack on Alkenes in the Polluted Atmosphere ......... 112
1. Discussion 112
III-B. The Effect of CO on the Chemistry of Photochemical Smog
Systems ....................... 1 ... .123
1. Results and Discussion
III-C. Computer Simulation of the Rates and Mechanisms of Photo-
chemical Smog Formation .................... 133
1. Results and Discussion 136
Computer analysis of the chemistry of smog formation in a
simulated NOx-hydrocarbon polluted atmosphere 136
The theoretical concentrations of HO and W^. i° the
simulated smoggy atmospheres and related data from
real atmospheres Ih-J
Theoretical mass balance of the nitrogen- containing
compounds formed in the simulated polluted atmosphere 15U
Effects of the variation of the concentrations of the
impurities on product formation in simulated sunlight-
irradiated polluted atmosphere 155
The saturated "reactive" hydrocarbons 155
Effects of variation of NOx concentrations 156
Effects of variation in initial CO concentrations 159
Effects of S02 addition on the reactions in the
simulated polluted atmospheres 162
Summary 167
References ................................ 169
Conclusions ....
vii
-------
List of Figures
Figure No.
1 The extinction coefficients of sulfur dioxide gas 2
2 Plot of the equation B for M = 02s Ar, and CO 8
3 Plot of equation B for M = C02 and -Ha 8
h Absorption spectra of SOa and excitation spectra of biacetyl
phosphorescence emission measured at 5120 A as excited in
S02-Ac2-C02 mixtures 15
5 Plot of l/$sens vs* V^sl for data from S02-Ac2-C02 mixture
photolyses 20
6 Plot of l$>§ens - * sens (2)) vs- V^Cfe] 22
7. Plot of l/(faenfl- * sens (2)) vs. 1/[N2] or 1/[CO] 25
8 Temperature. dependence of the S02(3Bi) lifetimes in pure S02 3^
9 Arrhenius plots of the rate constants for the S02(3Bi) quenching
reactions for several atmospheric gases: 02, Ar Jh
10 Arrhenius plots of the rate constant data for the S02(3Bi)
quenching reactions for several atmospheric gases: NO, C02j
N2, CO 35
11 Arrhenius plots of the rate constant data for the S02(3Bi)
quenching reactions for several paraffinic hydrocarbons:
CH4j C2H6.. C3H8, n-C4Hlo 35
12 Arrhenius plots of the rate constants for the S02(3Bi)
quenching reactions for some paraffinic hydrocarbons:
iso-C4Hio, neopentane, cyclohexane 36
13 Arrhenius plots of the rate constants for the S02(3Bi)
quenching reactions for the olefinic and aromatic hydro-
carbons 36
ik- Arrhenius plots of the S02(3Bi) quenching constants per
hydrogen atom for primary, secondary, and tertiary C-H bonds UO
15 Time dependence of [t-2-C4H8]/[cis-2-butene] ratio in 3500-
1*100 A irradiated mixtures of 2-butenes and S02 ^7
16 Plot of reciprocal of initial quantum yield of isomerization
of 2 -butanes versus [S02]/[2-butene] ratio ^9
17 Stern-Volmer plot of S02(3Bi) lifetime data in S02-2-butene
mixtures U9
viii
-------
List of Figures (Continued)
Figure No. Page
18 Plot of the reciprocal of the initial quantum yield of
t-2-butene versus [S02]/[C4H8] ratio 63
o
19 Plot of function C versus [S02]/[C4H8] in -3130 A photolysis
of S02-cis-2-butene mixtures 63
20 Plot of function C versus [S02]/[C4H8] from Cox data 6k
21 Plot of function D versus [SQal/Cc.^] ratio in 3130 A
photolysis of S02-cis-2-butene mixtures 66
22 Plot of function D versus [SOalt^Hs] ratio from Cox data 66
23 Plot of function G versus l/[ CCg] from 3130 A photolysis of
S02-2-butene-C02 mixtures 69
2k Molar extinction coefficients for the first allowed band
and the forbidden band of S02 at 25% 71
25 Gas handling and flow reaction system used in this study 72
26 Effect of residence timf of S02 on the quantum yield of S03
formation in 3600-lj-OOO A excited SQ2 78
27 Effect of residence time of 862 on the quantum yield of S03
formation in 3130 A irradiated S02 78
28 Plot of the ratio of the quantum yield of S03 formation in
pure SOa to that in SOs-added gas mixtures: NO-S02, COa-SC^,
02-SC-2 82
29 Plot of function G versus [CQ2]/[S02] in 3130 A irradiated
mixtures of S02-C02 in the flow system 85
30 Diagram of the apparatus used to prepare ^Os 95
31 The molar extinction coefficient for N205 vapor 96
32 Infrared spectral changes used to monitor the reaction of
ozone and Teflon 102
33 Infrared spectrum of products resulting from the reaction
of ozone with Teflon 103
3^ Rate of formation of CF20 plus C02 products of the 03-Teflon
reaction versus the 03 concentration 107
35a The major reaction paths for degradation of t-2-butene in a
sunlight- irradiated, NO -polluted atmosphere: 03, 0(3P), 02
(V
o
ix
-------
List of Figures (Continued)
Figure No. Page
35b The HO radical reaction paths with trans -2 -but ene llU
36 The .HOg, CH30, and N03 reaction paths with trans-2-butene 115
37 Theoretical rates of product formation in a sunlight-
irradiated simulated auto-exhaust polluted atmospheres 116
38 Comparison of experimental and computer simulated chemical
changes in N0-N02-C0 mixtures irradiated in moist air
(100 ppm CO) 126
39. Comparison of experimental and computer simulated chemical
changes in NO-NOa-CO mixtures irradiated in moist air
(1*00 ppm CO) 12?
1*0 The expected effects of water vapor and carbon monoxide
addition on the products of the NO photooxidation in air 128
hi The NO-NOa- CO-polluted atmosphere; simulation of the effects
of added CO 129
1*2 The photooxidation of NO in CO-containing mixtures; comparison
of experimental and computer simulated chemical changes in
O mixtures irradiated in relatively dry and moist air 132
1*3 The photooxidation of NO in CO-containing mixtures; comparison
of the experimental and computer simulated chemical changes in
CO mixtures irradiated in moist air 133
1*1* Theoretical rates of product formation in a computer simulated
sunlight-irradiated N0x-hydrocarbon-aldehyde-polluted atmos-
phere 137
1*5 Theoretical rates of reaction of various free radical species
with alkene in the simulated sunlight-irradiated, polluted
atmosphere of Figure 1*1* 1^3
1*6 Theoretical HO and H02 concentrations as a function of ir-
radiation time in simulated sunlight-irradiated, polluted
atmospheres of Figure 1*1* 150
1*7 The theoretical effects of varied alkene concentration on
the HO concentration in the simulated sunlight-irradiated,
polluted atmospheres 151
1*8 The theoretical effect of varied [N0]° on the HO concen-
tration in the simulated sunlight-irradiated polluted
atmospheres . 152
1*9 The theoretical effects of varied [CO]0 on the HO and H02
concentration in simulated sunlight- irradiated, polluted
atmospheres 153
-------
List of Figures (Continued)
Figure No. Page
50 The time dependence of the theoretical composition of the
nitrogen containing compounds formed in the simulated,
sunlight-irradiated polluted atmosphere. 155
51 Relation of eye irritation to hydrocarbon and oxides of
nitrogen levels in smog chamber experiments from Faith,
Renzetti and. Rogers 157
52 Relationship between the maximum daily 1-h average oxidant
levels and the 6-9 A.M. average concentration of nonmethane
hydrocarbons (Schuck, Altshuller, Earth, and Morgan) 157
53 The theoretical effect of reactive hydrocarbon concentration
on the integral of the [03] and [PAN] versus time data de-
rived from the simulated polluted atmosphere 158
5*1 Theoretical effects of variation of the concentrations of
the nitrogen, oxides on the 8-h integrals of the [03] and
[PAN] versus time curves obtained from the simulated polluted
atmosphere 158
55 Theoretical effect of increased [N0]° on the [o3]-time profile
in simulated polluted atmospheres 160
56 The theoretical effect of dilution of a highly polluted
atmosphere on the 8-h integrals of the [03] and [PAN] versus
time curves obtained in simulated polluted atmospheres 160
57 The theoretical effect of variation of the initial concen-
tration of the carbon monoxide on the 03 and PAN formation
in a simulated polluted atmosphere l6l
58 The theoretical rate of attack of various free radical
species on SOg for a simulated sunlight-irradiated, pol-
luted atmosphere. 168
-------
List of Tables
Number
II
III
IV
VI
VII
X
XI
XII
XIII
XIV
XV
XVI
XVII
Relative intensities of phosphorescence determined at t = 0 -'after
laser pulse and at 25 C- ..
Relative 1S02 quenching efficiencies of various atmospheric gases
Summary of the relative 3S02. formation rate constants and inter-
system crossing ratios for various atmospheric gases at 25 C
The 2875 A-excited SOs-sensitized phosphorescence of biacetyl in
Ac2-S02-C02 mixtures at room temperature
2&*75 A-excited S02 -sensitized phosphorescence of biacetyl in
Ac2-S02-C02 mixtures at room temperature
2875 A-excited SQa-sensitized phosphorescence of biacetyl in
Ac2-S02-C0 and Aca-SQs-Na mixtures
The effect of excitation wavelength on the quantum yield of
sensitized phosphorescence excited in S02-Ac2-C02 mictures
VIII Summary of parameters derived from the plots of l/$sens versus
Kinetic parameters related to the reactions of ^he 1SC>2 and X
species formed in the 2875 A-irradiation of S02-Ac2-mixtures
with added CQa, CO, or N2
The S02(3Bi) lifetime at various S02 concentrations and tem-
peratures
Rate constants for S02(3Bi) quenching reactions with various
atmospheric components at several temperature
Rate constants for the S02(3Bi) -quenching reactions with
various saturated hydrocarbon gases at several temperatures
Rate constants for the S02(3Bi)-quenching reactions with
various olefinic and aromatic hydrocarbon gases at several
temperatures
Summary of Arrhenius parameters for SOa(3Bi ) -quenching rate
constants with various collision partners
Arrhenius parameters Ea and logi0A for the SOs(3Bi)-quenching
rate constants per H atom for C-H bonds of different type
Quantum yields of SOs(3Bi) -sensitized isomerization of the
2-butenes
Aerosol formation in the S02(3Bi) reaction in S02-2-butene
mixtures
9
10
13
13
lU
17
19
23
28
29
31
33
37
xii
-------
List of Tables (Continued)
Number Page
XVIII Lifetimes of S02(3Bi) molecules excited in S02-2-butene mixtures 1*7
XIX Initial quantum yields of the S02-photosensitized isomerization
of cis-2-butene at 3130 A -' 58
XX Effect of flow rate on the quantum yields of S03 formation in
pure S02 irradiated within the first forbidden band 73
XXI Effect of foreign.gases on S03 quantum yields in S02 mixtures
irradiated within the first forbidden band in the flow system 7 4
XXII Quantum yields of S03 formation in pure S02 irradiated at 3130 A
in the static system 75
XXIII Effect of flow rate on the quantum yield of S03 formation in
pure S02 irradiated at 3130 A in the flow system 75
XXIV Effect of foreign gases on the quantum yield of S03 formation
in 3130 A irradiated S02 mixtures in the flow system 76
XXV Comparison of the rate constant ratios derived from the present
product S03 quantum yield data and S02(3Bi) lifetime data 83
XXVI Summary of new rate constant estimates derived in this work 86.
XXVTI Comparison of quenching rate constants for excited singlet S02
(long-lived component) 88
XXVIII Estimated ratio of intensities of the fluorescence from the
short-lived S02 singlet species to that for the long-lived species 89
XXIX Estimation of 0f from lifetime data of Brus and McDonald and
data of Calvert et al., and Mettee and Co-workers 90
XXX Rate data for the reaction of gaseous ozone with Teflon gasket
material at room temperature 105
XXXI Rate data for the reaction of gaseous ozone with different
samples of perfluorinated polyolefins at room temperature 106
XXXII The rate of attack of various reactive intermediate species on
trans-2-butene in a sunlight-irradiated, simulated atmosphere 117
XXXIII Comparison of the theoretical rates of the HO-radical forming
reactions in a simulated sunlight-irradiated auto-exhaust
polluted atmosphere 119
XXXIV Comparison of the theoretical rates of the H02-forming reactions
in a sunlight-irradiated, auto-exhaust polluted atmosphere IP]
XXXV The effect of carbon monoxide level on the concentrations of
products formed in the sunlight-irradiated moist atmospheres 130
xiii
-------
List of Tables (Continued)
Number
XXXVI
XXXVII
XXXVIII
XXXIX
XL
XLI
XLII
XLIU
XLIV
Rates of the major NOs formation and decay reactions in a simu-
lated smoggy atmosphere •-
The theoretical primary rates of the major radical generation
reactions in the simulated polluted atmosphere
Theoretical rates of the major HO formation and loss reactions
in the simulated polluted atmosphere " '
Theoretical rates of the major HOg formation and loss reactions
in the simulated polluted atmosphere
Theoretical rates of ROg formation and NO oxidation to
the major ROg radicals formed, in the simulated polluted atmos-
phere
The theoretical chain lengths for the HO, HOa, and
reactions in the simulated polluted atmosphere
radical
Page
Ik6
The theoretical rates of the major chain termination reactions
in the simulated polluted atmosphere
The fractional NO -» NOa conversion rate due to the various re-
actants in the simulated polluted atmosphere at 10 min into the
reaction
Estimated enthalpy changes, rate constants of possible homogeneous
elementary reaction paths for S02 in a polluted urban atmosphere 163
XIV
-------
INTRODUCTION
Interest in the sulfur dioxide removal mechanisms in the atmosphere remains
high among atmospheric scientists. There is substantial evidence that health
hazards are associated with urban atmospheres containing moderately low levels
of sulfur dioxide (above 0.2 ppm), and that these harmful effects are apparently
enhanced when there are significant levels of suspended, participates. The
special problem of aerosols containing high sulfate levels has focused new con-
cern on atmospheric SOg levels since the recent CHESS report. However, there
is now little evidence as to the chemical nature of the active species responsible
for these effects. The detailed mechanism of sulfur dioxide oxidation remains
unclear. For examples, see the reviews of Alshuller and Bufalini, Cadle and
Allen,3 and Urone and Schroeder.5 The majority of existing evidence suggests
that a major fraction of the sulfur dioxide is ultimately converted to sulfuric
acid and sulfate salts, but the intermediate species involved and the reaction
paths which lead to these products are open to question. The relatively slow
rates of photooxidation of sulfur dioxide in air exposed to sunlight""" and. the .
demonstrated catalytic influence of certain solids and moisture on the rate of
S0% oxidation9~15 have led to the common belief that heterogeneous paths for
SOg oxidation probably far outweigh the homogeneous modes. Although this con-
clusion may prove to be true for certain atmospheric conditions, it is by no
means established today. There is a real question as to the availability of
sufficient reactive metallic oxide, catalyst particles, and acid-neutralizing
compounds in many atmospheres to promote S02 removal at the observed rates.
The most compelling argument which has favored the importance of the hetero-
geneous removal processes has been the apparent lack of alternative homogeneous
reactions of sufficient rate which might be invoked. It now appears to us that
there are several such possible processes which are open for sunlight-irradiated
atmospheres, but these remain quantitatively unevaluated. Our approach to this
problem is based on the hypothesis that it is more realistic to evaluate the
contributions from these homogeneous processes and arrive indirectly at the
contribution of the heterogeneous modes than to establish unambiguously and
directly the importance of heterogeneous paths in the real atmospheres.
A major portion of the work completed under this grant bears directly on
the evaluation of the possible removal mechanisms of SQg in the troposphere.
These kinetic studies are described, in Section I of this final report. In
Section II we summarize the results of our kinetic studies of some chemical
systems of major interest in the understanding of photochemical smog mechanisms
and the development of scientifically sound control strategies. These relate
to several important intermediates and their reactions. In Section III we
discuss our computer simulation studies which allow an extrapolation of our
current knowledge to conditions which simulate those of the polluted atmosphere.
We derive in this section theoretical rates of development of ozone, peroxy-
acetyl nitrate, and other compounds of special interest in simulated polluted
atmospheres of varied composition. Predictions are made of the effects of
alteration of the concentrations of the reactive hydrocarbon, NO, NOa, CO, al-
dehydes, and SOg on the levels and dosages of Q3, PAN, and other compounds of
.special interest in air pollution control. Finally conclusions can be formulated
from these results which provide a reasonable guidance in the development of more
refined control strategies. These simulations provide as well recognition of
areas which are still ill-defined, and in which further definitive work is needed
to provide unambiguous conclusions concerning SOa removal paths and the mechanism
of photochemical smog formation.
-------
PART I
THE ATMOSPHERIC REMOVAL PATHS OF S02
.The absorption of solar energy by sulfur dioxide in the troposphere occurs
within the relatively strong, "allowed" band which extendsefr,om 2*100-5300 A and
the "forbidden", weaker band which extends from jUOO-UOOO A; see Figure 1.
,400
300-
e
200
100-
2200 2400 2AOO 2600 3OOO 3200
WAVELENGTH,A
(e)
0.10
0.08
O.O6
€
0.04
0.02
0
3400
36OO 3800
WAVELENGTH,A
4000
(b)
Figure 1. The extinction coefficients of sulfur dioxide gas
in the first allowed band (a) and the "forbidden" band (b); €
= [logio(l0A)Vcl (liter
Excitation of S02 within its first absorption band (2^00-3300 A) leads to the
generation of three emitting species: (l) a very short-lived singlet state
[perhaps SOj^Aj.)]; (2) a long-lived singlet state [probably S02(1B1)]; (3)
a triplet state [S02(3Bi)]. In addition to these optically detectable or
emitting states, indirect chemical and physical evidence which we will review
later suggests that at least one other non-emitting triplet [presumably S02(3A2)]
and possibly a third triplet [conceivably S02(3B2)] may be important in the photo-
chemistry of SCfe. An understanding of the kinetics and mechanisms which control
the population of these SCfe states and the rate constants for the various decay
reactions of these states are essential ingredients to our evaluation of S02
removal mechanisms in the atmosphere. Significant progress toward this end was
accomplished in this work and is described in the following sections.
-------
I-A. The Efficiency of S02 Triplet Formation Induced _in the Excited
Singlet SOg by_ Collisions with Og_ and Other Atmospheric Gase'sT10
Molecules of sulfur dioxide in the first excited triplet state (3Bi),
designated here as 3S02, have been shown to be important reactants in several
irradiated S02-containing systems1?~22. Excited singlet sulfur dioxide molecules
(-"•SOs), formed by absorption within the 2UOO-3200 A band, generate the triplet
species largely in a second-order, collisionally induced reaction [reaction (2)].
The XS02 deactivation also occurs in a competitive second-order process [reaction
(l)] of ill-defined mechanism involving the ground-state, nonoptical excited
states or isomeric S02 states:
(1) 1S02 + M -» (S02 - M) -* S02 + M (or products)
(2) 1S02 + M -» 3S02 + M
The efficiency with which 3S02 molecules are formed in the spin-inverting re-
action (2) is of primary importance in elucidating the kinetics of product
formation in S02-reactant mixtures irradiated within the 2UOO-3200 A absorption
band of S02.
18 2^ 22 2.h 25
Rao et al., ' J Wampler et al., and Horowitz and Calvert ' 'ohave de-
termined the quantum efficiency of the production of 3S02 (2UOO-3200 A irradiation)
by reaction (2) and. the bimolecular intersystem crossing ratio, k2/(ki + k2), for
M = S02, N2, CO, and cyclohexane. They employed an indirect method based on the
3S02-sensitized phosphorescence of biacetyl. This method suffers from the dis-
advantage that it cannot be used to determine the intersystem crossing ratio in-
duced in 1S02 by collisions with oxygen or other compounds which are good quenchers
of the triplet state of biacetyl. However, a knowledge of k2/(k1 + k2) for M = 02
is of special importance in evaluating those processes which occur in S02-polluted
atmospheres.
In this work we used a new method to determine intersystem crossing ratios
for 1S02; the relative intensity of the phosphorescence from 3S02 molecules was
determined as a function of time following the 2662 A laser excitation of 1S02
molecules. Extrapolation of the phosphorescent intensity to zero time following
the laser pulse allowed the estimation of intersystem crossing efficiency using
02 as a collision partner for 1S02. Also, we have applied this method to esti-
mate this important ratio with the other major atmospheric gases acting as col-
lision partners.
I-A—1. Experimental Methods and Results
Laser excitation source
The first excitedosinglet state of S02 was populated by absorption of
radiation from a 2662 A laser pulse. This wavelength was produced by the appli-
cation of nonlinear optical techniques to a Q-switched neodymium laser. The
10,650 A neodymium fundamental was frequency doubled by the use of a precisely
oriented, 2-in.-long potassium dihydrogen phosphate crystal. The 100-megawatt
neodymium fundamental generated a 2-megawatt 532U A laser pulse which was
further doubled by a precisely oriented ammonium dihydrogen phosphate crystal
to give a 2662 ± 1 A laser pulse of 30 kilowatts having a half-intensity duration
-------
of 20 nanoseconds. Because of the peculiar limitation of our equipment it was
most convenient to have the laser and associated optical apparatus in one hori-
zontal plane and the reaction cell and associated equipment in a different hori-
zontal position. The laser beam could be readily aligned with the reaction cell
using two adjustable aluminium-coated mirrors. Because of the reflectivity
characteristics of the aluminium mirrors, the laser intensity entering the cell
was reduced to about 15 kilowatts, but this was quite adequate for the experi-
mental requirements.
Radiation detectors and treatment of the data for the decay of phosphorescence
Because the laser power does vary somewhat from pulse to pulse, it was neces-
sary to monitor the power of each laser pulse so that the initial intensities of
phosphorescence could be corrected for variations in laser pulse intensities.
Power measurements of the 2662-A radiation exiting the cell were made by con-
necting only the first stage of an RCA 1P28 photomultiplier tube. Phosphores-
cence was monitored at right angles to the cell using a RCA 7265 photomultiplier
tube. A Jena WG-2 filter having a sharp wavelength cut-off at 3500 A was placed
in front of the photomultiplier unit to help discriminate against the fluorescent
emission envelope of SOg and to protect the photomultiplier tube from the laser
radiation. The fluorescence was reduced largely by the WG-2 filter. For the
reactant pressures employed and the sweep speed of 20 p-sec/cm, the intensity of
fluorescence became negligible within a very short period, and the phosphorescent
decay was clearly defined. The oscilloscopic traces of the phosphorescence and
the 2662-A laser pulse signals were photographed with high-speed hlD Polaroid
film; the phosphorescent decay photographs were digitized.and treated as described
by Sidebottom et al. °
The sweep of the oscilloscope to record the response from the photomultiplier
tube was triggered externally at the time of the laser pulse. The relative in-
tensities of the phosphorescent traces at t = 0 following the laser pulse were
determined from the antilogarithm of the intercept of the least-squares fit of
the In Ip-versus-time data. Data for the ratio of the phosphorescent intensity -
at t = 0 for S02, [M]7* 0 (I M) to that for S02. [M] = 0 (lp) are given in
Table I.
Table I. Relative intensities of phosphorescence determined at t = 0 after
the laser pulse and at 25° C.
1 +
Pu, torr [M]/[SOJ (V//,) «-o \ (*,. + *i.)[SO
,)(¥),
1.04
2.04
2.54
3.07
3.64
6.81
9.95
13.3
0.534
2.04
2.54
3.02
3.46
3.96
4.77
1.95
3.82
4.76
5.75
3.50
6.55
9.57
12.8
1.00
3.82
4.75
5.65
6.48
7.42
8.93
M = O2
0.729
0.663
0.659
0.536
0.666
0.561
0.480
0.500
M = N2
0.882
0.751
0.627
0.631
0.697
. 0.531
0.601
1.16
1.43
1.60
1.46
1.37
1.67
1.86
2.42
1.14
1.58
1.49
1.67
2.01
1.67
2.16
-------
Table I. (Cont'd')
Pti, torr
[M]/[S02
(/PM //
1 +
0.994
1.49
1.94
2.92
3.79
5.14
10.0
0.534
1.04
2.04
2.04
2.55
3.02
3.42
3.53
4.00
0.169
0.338
0.626
1.25
1.88
2.50
3.82
7.64
11.5
3.64
5.47
3.64
5.47
3.64
9.62
9.62
1. 00
1.95
3.82
4.13
4.78
5.66
6.40
7.15
7.49
0.379
0.758
1.40
2.80
4.22
5.61
3.67
7.34
11.1
M = Ar
0.623
0.607
0.647
0.587
0.675
0.522
0.469
M = CO
0.880
0.774
0.792
0.833
0.713
0.624
0.747
0.633
0.691
M = CO2
0.937
0.839
0.699
0.655
0.649
0.630
0.737
0.718
0.660
.19
.44
.24
.39
.29
.78
.60
1.21
1.35
1.94
2.14
2.01
1.97
2.56
2.35
2.66
.18
1.28
1.38
1.92
2.54
3.07
2.60
4.35
5.72
1^ represents the relative phosphorescent intensity in a given experiment
divided by the laser pulse power in that same run; Ip^ is the same quantity
determined in a run at equal [SOa! but with added quencher gas M.
This quantity is the ordinate employed in Figures 1 and 2 to obtain the rate
constant ratio kgb/kga; see relation [B] in text. _
Reaction cell and associated apparatus
The reaction cell body was constructed of Pyrex and was 88 cm in length
and 2.5 cm in diameter. Suprasil windows were cemented to the cell at the
Brewster angle, using low-vapor-pressure epoxy cement. Reactants were mixed
using a thermal gradient pump. The cell was connected to a grease-free vacuum
line, and pressure measurements were made using a quartz spiral manometer.
Materials
Burdett high-purity nitrogen was used in experiments with added nitrogen.
All other gases used in this work were obtained from Matheson and were of the
quality indicated; argon (high purity), 02 (research grade), CO (C.P. grade),
and COa (Coleman instrument grade); SQs (anhydrous grade) was purified further
-------
by bulb-to-bulb -distillation in a high vacuum system, and. the middle third was
taken for the experiments.
I-A—2. Discussion of Results
The mechanism of 3S02 formation in the 2kOO-32.QQ A irradiation of SOg
In recent studies the JIJO-A of irradiation of the SOa-CO system, Cehelnik
et al. ^ have postulated that two nonradiative excited states of SOa as well as
the optical states of S02 are involved as reactants leading to CQa formation.
In a recent similar study of Watnpler et al.^2, we concluded that 3S02 molecules
were the major reactant in this system even at high gas pressures; we suggested.
that a bimolecular route of forming 3SOa in addition to reaction (2) became im-
portant at high gas pressures. We noted "that our kinetics were consistent with
the generation of 3SQ2 molecules by .some unknown precursor, X, which was produced
by 1SQz + M •* X + M, and it formed 3S02 at_a significant rate only when reactant
pressures were high enough for the reaction X + ~M -» 3S02 + M to compete with the
unimolecular disappearance of X.. Investigations are presently under way to gain
more definitive information regarding this proposed pathway to 3SQ£ and the
alternative hypothesis ofeCehelnik et al.^1. Preliminary results from experi-
ments utilizing the 2662-A laser excitation of SOa indicate that in the presence
of added gases, the observed phosphorescent lifetimes in experiments below about
30 torr total pressure are nearly identical to those observed for S02(3Bi) mole-
cules produced by direct excitation utilizing a J829-A laser pulse. However, in
experiments above 30 torr, the measured decay of the phosphorescence is slower
than the low pressure rate data would have suggested. Possibly the apparent
enhancement in the 3S02 lifetime may be due to the bimolecular participation of
X to produce 3S02 during the monitored decay of SS02. If this hypothesis is
correct, a nonexponential decay of the phosphorescence is expected. Because of
experimental limitations, our phosphorescent measurements cannot be extended now
over a sufficiently large time scale to unequivocally verify or disprove the
presence of a nonexponential decay. Further work is necessary before we can
properly assess these findings in terms of alternative hypotheses. ^>^. However,
our observations are clear on one result: below about 30 torr total pressure.
the major bimolecular route of producing 3S02 is from the ^SO^ precursor. Ac-
cordingly, for the experimental conditions employed (total pressure less than
30 torr), the participation of X in the previously proposed reaction scheme may
be neglected, and we now introduce only those reactions which are pertinent for
our present conditions:
(l) S02 + hv (2662 A) -» 1S02.
(la) -"-SOa + S02 -» (2S02 or products)
(2a) XS02 + S02 -» 3SOa + S02
(Ib) 1S02 + M -» (S02, M or products)
(2b) 1S02 + M -» 3S02 + M
(6) 3S02 -» S02 + hv
(7) 3S02 -» S02
-------
(8a) 3S02 + S02 -» S03 + SO
(8b) 3S02 + S02 -» (2S02 or products not S03)
(9) 3S02 + M -» (S02, M, or products)
Estimation of the intersystem crossing ratio in 1SQ2-M collisions with various
partners
Immediately following the laser excitation of S02 to 1S02, the relative
intensity of phosphorescence, (lp)t=0? i-s directly proportional to the concen-
tration of triplets formed by intersystem CJtiossing. (The unimolecular decay of
1S02 to 3SQ2 is unimportant for our reactant pressures, and the fast bimolecular
reactions steps (2a) and (2b) are responsible for the population of molecules in
the S02(3Bi) state.) From a consideration of the above mechanism at t = 0, we
have the following relationship for the ratio of initial phosphorescent intensity
with S02 alone at a given pressure to that in which S02 is at the same pressure
but an added gas M is present:
IpM x 1 + (kabAaa)
I J
p t=0 1 + [klb + ksb)/(kia + k23)] [M]/[S02]
Equation [A] may be arranged to give the useful relationship [B]
M
/Ip)t=0 d + C(klb + kab)/(kla + kea)] [M]/[S02]}
= 1 +
Relationship [B] is tested graphically in Figure 2. using the data from Table I
for M = 02, Ar, and CO: a similar test for M = COa, and N2 is shown in Figure J.
According to equation [B], the intercept value should be unity. The experimental
intercept values from a least-squares treatment of the data for M = 02, N2, Ar,
C02, and CO are: 0.98 ± 0.15, LOU ± 0.30, 0.98 ± 0.13, 0.86 ± 0.20, and l.OU ±
0.28, respectively.* Within the error limits the graphical representations are
*A11 error limits reported represent the 95^" confidence limit (twice the standard
deviation).
in agreement with the functional form required by equation [B]. The slopes of
the plots in Figures 2 and 3 are equal to k2t>/k2a for the various added quenching
gases. These results are summarized in Table III. Weomay take the value of
k2a/(kia + ^23) = °-°82 ± 0.003 as determined at 2650 A2^ and the average value
of the ratio (ki-b+ k2-)3)/(kla + k2a) from the published data summarized in Table II.
With these and our present data we can estimate the intersystem crossing ratio;
these results are shown in Table III also*
•In calculating the error limits for k2-b/(k1'b + k2b)3 the error limits for
k2b)/(kia + ^2a) were estimated by the difference between the average and the
extreme values.
-------
oJ ^a.
M=CO
0
M=Ar
I I I I
10
Figure 2. Plot of equation [B] for M = Os, Ar, and CO. The
slope of each plot is equal to kgb/ka for the particular
quenching gas. Data used are from Table I.
all
/
M=C02
0 5 10
M
[so2]
Figure 3- Plot of equation [B] for M = COs and Ng. The slope
of each plot is equal to kab/^aa for the particular quenching
gas. Data used are from Table I.
8
-------
Table II. Relative ^SCfe quenching efficiencies of various atmospheric
gases: (k^ + k2t)/(kla + ksa).
Relative 'SOz quenching efficiency
Compound Sidebottom et al.° Mettee6 Rao et al. c Horowitz et al.d
so»
Ar
N3
0,
CO
CO2
1.0
0.28
0.30
0.31
0.43
0.71
1.0
0.21
0.28
0.27
0.35
0.63
1.0
—
—
0.31
—
0.74
1.0
—
0.29
—
0.35
Q
Reference 27
Reference 28
°Reference 29
Reference 2.k and 22
The Effect of the nature of the collis ion partner on the intersystem crossing
rate in excited s inglet sulfur dioxide
Compare the estimates for k^/k^ and kab/(kib + k2b) for the various col-
lision partners in Table III. These data support the earlier conclusions of
Horowitz and Calvert which were based on much more limited data. ^ The indi-
vidual rate constants for the collis ion- induced spin inversion of ^-SOg increase
significantly with the increase in the internal degrees of freedom of the col-
lision partner; the values for k2b/k2a with M = Ar, 02, N2, CO, CQ^, S02} and
cyclohexane are: 0.075, 0.11, 0.12, 0.21, 0.1*U, 1.0, and 1.1, respectively.
However, the intersystem crossing ratio, k2b/(klb + keb)> is relatively less
sensitive to the nature of the collision partner M; for M = Ar, 02, W2, CO, C02,
S02, and cyclohexane, kab/(kib + keb) = °-025, O.OJO, 0.03**, O.OU5, 0.052, 0.082,
and 0.073, respectively. The extent of the partitioning of the excited 1S02
between the alternative energy degradation paths of reactions (1) and (2) is
probably controlled largely by the properties of the S02 molecule, that is, the
relative density of states in the triplet and the singlet state manifolds near
the vibronic level of the excited singlet and the extent of the mixing of the
singlet and triplet states.
The efficiency of 3SQ2 formation _in M2- and CO- contain ing mixtures
The present estimates of k2b/k2a and kgb/(kib + k2b) f°r M = N2, agree well
with those of Horowitz and Calvert2' who employed the triplet-sensitized biacetyl
phosphorescent technique. However, the k2t,/k2a ratio for M = CO reported here
(excitation at 2662 A) is significantly higher than that found previously using
the biacetyl technique and excitation of S02 at 2875 A. The reason for this dif-
ference is not clear. In view of the nitrogen data from the two studies and the
apparent insensitivity of the k2a/(kia + k2a ) to the wavelength of the exciting
light for M = S02 (see Table III), it seems unlikely that the apparent discrep-
ancy arises from the difference in wavelength employed in the two studies. The
biacetyl method, employed previously is very sensitive to the presence of oxygen
-------
Table III. Summary of the relative 3S02 formation rate constants, kab Asa > and the
H
o
intersystem crossing ratios, k2b/(kib +
gases at 25° C.
kgb ) , for various
ktb/kta kit/ (kib •+•
Gas This work (2662 A) Previous work This work (2662 A)
Oi 0.11 ± 0.02 — 0.
N2 0.12 ±0.05 0.10 ± 0.03 (2875 A) - 0.
Ar 0.075 ± 0.021 — 0.
CO2 0.44 ± 0.04 — 0.
CO 0.21 ± 0.06 0.063 ± 0.027 (12875 A)6 0.
030 ±
034 ±
025 ±
052 ±
045 ±
SO2 1.0 1.0 0.082 ±
(assumed)
c-CtHn — 1.11 ± 0.37°
—
0.
0.
0.
0
,013
029
,005
.014
0.028
0.
,003
0
0
0
0
0
0
0
0
.033
.017
.082
.095
.080
.09
.10
.073
atmospheric
*«)
Previous work
±
±
0
0
± 0
± 0
± 0
(2963
(3020
±
0
—
.008
—
—
.010
.003
.005
.014
A)-
A)-
.024
(2875 A)°
(2875 A)»
(2650 A) '
(2875 A) '
(2875 A)-*
a
a 25
From Horowitz and Calvert .
From Wampler, Horowitz, and Calvert
From Horowitz and Calvert
j i ft
From Rao, Collier, and Calvert
e 23
From Rao and Calvert .
-------
impurity in the quenching gas employed. This is not a problem with quenching
gases which can be condensed and degassed well. The removal of traces of Qg
from CO gas is not a simple matter, and as little as $ of oxygen in the CO
could cause the difference seen here. Although the stated analysis of the CO
reagent used in the previous work did not suggest this extent of contamination,
it may have been the source of the difference. In any case the present value of
keb/C^ib + ksb) is relatively insensitive to oxygen impurity in the reactants,
and for this reason it is considered to be more reliable data for the case of
CO as the quenching gas .
The "excess" triplet _in the 3130 A- irradiated SQg-CO mixtures ert high pressures
In view of the present more reliable estimates of k.2b/(kib + ^sb ) for M =
CO, a reevaluation should be made of the relative importance of the previosuly
reported "excess" triplet SOs reactant leading to COa in the 3130 A-irradiated
SOg-CO mixtures at high pressures2-1- >22. We estimate that the quantum yield of
COg from the 3SO^ species derived only from the ^^SOg precursor may be as high
as 1^.3 x 10~3 in experiments at high [COl/CSOa] ratios; our previous data sug-
gested a maximum of 2.0 x 10"3. Our previous work showed that the excited
singlet S02 will contribute about l.U x 10~3 to the total § cCb v81*16 at niSh
[COl/CSQal ratios. The total measured quantum yields of COa from the 313° A-
irradiated SO^-CO mixtures exceed the sum of these maximum triplet and singlet
yields (5.7 x 10~3) in experiments at high [CO]/[SOa] ratios and at pressures of
CO above about 100 torr. Thus it remains necessary to invoke an additional
source of triplet molecules in runs at pressures above 100 torr, but the im-
portance of this source is somewhat less than we and Cehelnik et al. had
suggested in the earlier studies.
The theoretical maximum rates of SOg photo-oxidat ion in sunlight- irradiated
S02- contain ing atmospheres
The present data allow an improved estimate of the rate of 3S02 generation
in the sunlight- irradiated atmosphere of the earth^ . For a relatively dry
atmosphere, the effective intersystem crossing ratio will be about 0.033- Ih
making our previous estimate we used the only intersystem crossing data then
available, which was for M = SQ2 in reaction (2b), k2b/(kit + kgb) = 0-09- Use
of the more appropriate present rate constant data will result in values for '
the theoretically maximum rates of 3S02 formation from (2b) which are roughly
one third lower than those estimated previously from the less complete infor-
mation. However, the additional "high pressure" mechanism of triplet sulfur
dioxide formation must be present at the pressure of the atmosphere and must
now be included, in any revised quantitative treatment of the problem. Our
recent work suggests that the rate of triplet generation from this source is
roughly two to three times that formed in reaction (2b) in air. A more sophis-
ticated estimate of the theoretical maximum SOg photo-oxidation rates is pos-
dible utilizing our results outlined in our further work described in the
following section.
I-B. A Study of_ the Nature of the S0g Excited States
Formed at High Added Gas Pressures. -^-L
Recently, some very interesting and unexpected observations concerning the
nature of the excited states of S02 have been made using irradiated SOa-CO
11
-------
7O
mixtures. Jackson and Calvert generated 3S02 by direct absorption within the
forbidden 3Bi *- 1A1 band (3^00-3970 A) of S02; the 3S02 molecules were found to
be rather efficient reactants with CO to form C02 for these- conditions. However.
PI • '
Cehelnik, Spicer, and. Heicklen observed that the kinetics of 'the formation ofo
C02 in S02-C0 mixtures irradiated within the first allowed band of S02 at 3^30 A
required the participation of reactive states of SOa in addition to the 3S02
species. At high CO pressures, there was an "excess" of COs. product formation
which exceeded, that anticipated from the measured rate constants for the •'•SOa
and 3S02 states. Cehelnik et al. invoked as reactants both the emitting
singlet and triplet states as well as two new nonemitting excited states of SOa-
Wampler et al^2 reinvestigated this system and confirmed most of the obser-
vations of Cehelnik et al*; however, some key differences in the results led
these workers to interpret their results in terms of the involvement of the
and 3SQa states plus one additional nonemitting state of SOa; this species was
not a reactant to form COa, but was an additional source of 3SOa generated at
high added-gas pressures.
In this phase of our work, we have attempted to gain further insight into
the nature of the intermediate states involved in SOa photochemistry which might
allow a more definitive test of the mechanistic alternatives in the photochemistry
of SOg-CO mixtures. ¥e have employed biacetyl as a triplet energy acceptor in
2&T5 A-irradiated mixtures of SOa, biacetyl, and added gases (COa, CO, or Na).
In other experiments, the wavelength dependence of the intensity of the sensi-
tized emission has been determined at varied added-gas pressures. The kinetic
functions which define the quantum yields of the sensitized phosphorescence
emission in biacetyl.provide interesting new details of the mechanism of SOa
photochemistry.
I-B—1. Experimental Methods and Results
Equipment
The absolute spectrofluorometer of Turner (Model 210) was used in all of
these studies. The SOa was excited at 2875 A using a band width of 150 A; the
triplet biacetyl emission was scanned using a 250 A band width. The same band.
widths were employed in the determination of the excitation spectra. Biacetyl
excited at U350 A was used as a standard for the quantum yield determinations as
described previously".
The quantum yields of the 2875 A-excited S02-sensitized phosphorescence of
biacetyl ($sens) were measured at several S02 an(3 Ac2 reactant concentrations
and added C02 pressures (25-750 torr). The results from several series of ex-
periments carried, out at constant S02 an(3 COa concentrations and room tempera-
ture (~ 25°C) are reported in Table IV. Several other series of experiments
were made with S0a-Ac2-C0a mixtures in which the concentration ratio of the
three components was kept constant in a given series; these are given in Table V.
The results from similar series of experiments using S0a-Ac2-C0 and S02-Ac2-Na
mixtures at constant concentration ratios are summarized in Table VI.
For our experimental conditions, no biacetyl emission could be detected in
the direct 2875-A excitation of Ac2 in a mixture containing [C02] = 3 x 10~2
and [Aca! = 1 x 10~4 M; it is clear that the second excited singlet state of
biacetyl was not stabilized measurably against dissociation, even at the highest
pressures of added gas employed here.
12
-------
'Table IV. The 2875 ^-excited S02-sensitized phosphorescence of biacetyl in
Ac2-SC>2-C02 mixtures at room temperature; constant pressure ex-
periments • •
*»M X 10s
[Ac2], M X 106 4 X 10-W 10 X IQ-'M° 25 X IQ-'A/" 40 X
2
3
4
7
10
50
100
2.25
2.87
3.20
4.07
5.04
7.75
9.08
2.23
2.65
3.30
4.64
5.32
8.54
9.19
1.84
2.52
3.23
4.74
5.94
10.35
11.54
2.05
2.48
3.41
6.35
6.75
(6.54)*
13.84
14.1
r\
C02 concentration, [C02].
[S02] = 1.0 x 10~4 M in all cases except for this run, in which [SOa] =
1.0 x ICr5 M.
rv^ °
Table V. 2875 A-excited S02-sensitized phosphorescence of biacetyl in Ac2-
SQa-COa mixtures at room temperature; constant reactant ratio ex-
periments
[C02
1, M X 10s
1.36
1.63
2.14
2.56
3.34
4.00
0.75 X
10-»°
1.88
1.98
2.12
2.29
2.60
2.87
1.00 X
10-"*
1.99
2.13
2.36
2.60
2.89
3.20
*.„•»
1.75 X
X 10s
2.50 X
12.
5 X
25.0 X
io-»° io-" io-«« io-«-
3
3
3
3
3
4
.02
.22
.21
.40
.77
.07
'3
3
3
4
4
5
.28
.47
.86
.14
.62
.04
5
5
6
6
7
7
.43
.83
.30
.63
.21
.75
6
7
7
7
8
9
.68
.03
.48
.87
.43
.08
5.55
6.65
8.72
10.4
13.6
16.3
21.4
25.6
33.4
40.0
0.50 X
10-"
—
.
1.16
1.39
1.70
2.05
1.00 X
10-"
— ^_
—
—
—
1.59
2.04
2.37
2.94
3.41
1.75 X
10-"
2.59
2.99
3.67
4.29
5.44
6.35
2.50 X
10-"
2: 11
2.51
2.99
3.47
4.15
4.93
5.95
6.75
12.5 X
10-"
5.63
6.29
7.32
8.19
9.59
10.5
12.4
13.8
25.0 X
10-"
5.28
5.74
6.66
8.24
8.35 •
9.05
10.4
11.4
12.8
14.1
£J
rAca]/[C02] ratio; [COa] in the low concentration region; rS02]/LC02~' =
2.52 x 10~2 throughout the low concentration series.
[Ac2]/[co2] ratio; [ CC^l in the high concentration region: [S02]/[C02] =
2.50 x 10~3 throughout the high concentration series.
13
-------
Table VI. The 2875 A-excited SOs-sensitized phosphorescence of biacetyl in
Ac2-S02-C0 and Ac2-S02-N2 mixtures in experiments at constant
concentration ratioa
[CO], M X 10"
*„„. X 103
[N,], M X 10>
X 10"
3.13
3.76
4.91
5.87
7.70
9.20
12.0
14.8
18.8
22.5
4.71
4.96
5.60
5.98
6.36
7.11
7.57
8.18
9.00
9.76
3.56
4.21
5.49
6.58
8.59
10.3
13.5
16.1
21 ;0
25.2
4.56
4.92
5.44
6.02
6.65
7.14
8.02
8.63
9.74
10.4
The reactant concentration ratio was held constant in a given series of
runs: [CO]: [SOfe]:[Acs], 225:1:1; [N2]:[S02]:[Ac2], 252:1:1.
The effect of excitation wavelength on $Sens was determined in various
SQs-Ac2-COs mixtures using a 150-A band width. These data are summarized in
Figure h together with the absorption spectrum of S02 determined at the same
excitation band width (150 A).
I-B—2. Discussion
The "excess" biacetyl phosphorescence from 2&75 A-irradiated S02-biacetyl
mixtures _at high added-gas pressures; deficiencies in the "low pressure"
photolys is mechanism
The previously proposed mechanism of S02 photochemistry-'-" has been found
to be quantitatively consistent with the results of the S02 emission studies
reported to date3^~^; however, because of the very large second-order quenching
rate constants for the 1S02 species (~ 1010 - 1011 4./mole-sec), the previous
studies were necessarily carried out at relatively low total gas pressure (P <
10 torr) in order to allow detection of the weak singlet emission. This "low
pressure" mechanism and the measured rate constants require that the only signif-
icant triplet-SOa formation for experiments above about 1 torr arise in the
second order reaction (2):
S02 + hv (2*1.00-3200 A) -» XS02
+ M -»' (S02-M)
+ M -» 3S02 + M
(I)
(1)
(2)
When the concentration of added gas M is relatively high and biacetyl (Ac2) is
added to the S02-M mixture, as in this work, then the first-order decay reactions
of the ^SOz and 3S02 are unimportant and only the following reactions of the
and Ac2 are expected to occur following reaction (2):
3S02
-------
(6)
(7.)
(8)
(9)
3S02 + M -» S02 + M (or products)
3S02 + Ac2 -» 3Ac2 + S02
3Ac2 -» Ac2 + hv
3Ac2 -» Ac2
zoo
ISO
2600
2800 3000
Womlinglli, t
t
~G
a
i
3
Figure U. Absorption spectra of S02 (curve l) and excitation
spectra of Macetyl phosphorescence emission measured at 5120 A
as excited in irradiated SQ2-Ac2-C02 mixtures; curve 2: [S02] =
1.0 x 10-4 M, [Ac2] = 2.0 x 10-6 M, [OOfe] = 2.0 x 1Q-2 M;
curve J>: [S021 = 1.0 x 10'4 M, [Ac2] = 2.0 x 10'6 M, [C02] =
0.0 M; curve k: [S02] = 1.0 x 10~4 M, [Ac2] = 5.0 x KT5 M,
[ C02] = 2.0 x 10~2 M. For all these measurements, the exci-
tation and emission band widths were set at 150 and 250 A,
respectively; the ordinate for emission intensity is in arbi-
trary units which are different for each of the curves 2, J>,
and h.
In terms of this mechanism of the irradiated S02-Ac2-M system and for our
present conditions of essentially complete quenching of the 1S02 by added gas
M, relation (A) should describe the quantum yield of the sensitized biacetyl
emission ($):
sens
(A)
[M]k<
Thus, for experiments at constant [M]/[Ac2] ratio, we expect $sens
dependent of the added gas pressure.
concentrations employed here (up to
to be in-
Note in Table V that at the high gas
x 10"2 M) , the' expectations of the above
15
-------
"low pressure" mechanism are not borne out. Furthermore, for experiments at
fixed [S02] and [Ac2], one expects from relation- -(A) that the $sens would de-
crease as the pressure of the added gas in increased. 'The data of Table IV do
not support this prediction either; the $sens continues to rise as the total •
pressure of the gas mixture is increased. Although the simple mechanism outlined
in (l), (l), (2), and (6)-(9) is followed well at low gas pressures'^ j^O^ obvi-
ously it is not a sufficiently complete mechanism to account for the results
obtained here at high pressures (up to 1 atm).
There are two possible trivial sources of these unusual results which should.
be evaluated before considering other more sophisticated alternative mechanisms:
1. Conceivably, the increase in biacetyl emission with increasing pressure
could result from the decreasing importance of diffusion and deactivation of
biacetyl triplets at the wall of the cell. Although such an effect has been
observed in experiments at much lower pr.es sure^^j 39 5^°5 we have found that it is
unimportant for the conditions employed here. Thus, when biacetyl was excited
by direct absorption at ^350 A, there was no significant increase in quantum
yield of phosphorescence on increasing the pressure of added gas from 70 to 700
torr.
2. The "excess" biacetyl emission at high pressures might arise from a
reaction between ^SQg ant^ biacetyl. In theory, this could occur by either of
two paths. Fjrst, an excited singlet of S02 may lead to excited singlet of
biacetyl by singlet energy transfer:
(10) ^-SOa + Ac2 -* 1Ac2* + S02
(ll) 1Ac2 -» decomposition products
(12) rAc2* + H -» XAC2 + M
(13) 1Ac2 -» 3Ac2
o
There is good overlap between the 2^00-3200 A absorption band of S02 and the
absorption band corresponding to the second excited singlet of Ac2. If re-
laxation of the -""SOs is unimportant, we would expect energy transfer of this
type to form biacetyl in the second excited singlet state (1Ac2*), reaction (10).
Presumably the excess emission seen at high added-gas pressures could have arisen
as a result of vibrational and electronic energy dissipation through collisions
which would ultimately populate the lowest triplet of biacetyl, reactions (12)
and (13)- This hypothesis cannot be correct either, since we have found that
the direct excitation of biacetyl within its second allowed absorption region
(2200-3200 A) produces no detectable emission even when J60 torr of added gas
is present; obviously, kn » k12 [M] even at high [M].
Alternatively, one may assume that at the very high inert gas pressures
employed here, the 1S02 which survive the electronic quenching collisions with
M and finally encounter biacetyl molecules are vibrationallyo relaxed to near
the lowest vibrational level of the excited singlet (~ 3370 A). Thus, singlet
energy transfer for these circumstances will form singlet biacetyl equivalent
to that generated by light absorption within the first excited singlet band of
Ac2; in this case, vibrational quenching of the initially dissociative state is
significant, although for 33^0 A-excited Ac2, stablization of xAc2 is only
at 1 atm of quenching gas. Taking the relative rate constants for 1S02
16
-------
quenching by S02. C02, and Macetyl (1.00, 0.7358? ancj 1.75^3^ respectively)
and ke/(k8 + kg) = 0.15 , we estimate that the maximum contribution of singlet
energy transfer to the measured §sens must be less "than-7 x 10~4 when [C02] =
k x 10-2 and [Ac2] = 1 x 10~4 M;the observed $sens is l.U' x 10 ~2 for these
conditions. Thus we must conclude that although as much as 5$> of the observed
§sens may result from singlet energy transfer for these conditions, it cannot
be the dominant source of "extra" biacetyl phosphorescence in these experiments.
Further support for this conclusion is had from the observed wavelength
independence of the $sens in S0a-Ac2 and S0a-Ac2-C02 mixtures (see the excita-
tion spectra shown in Fig. 4). Note the wavelength dependence of the 3Ac2
phosphorescence intensity is the same for a mixture of [S02] = 1.0 x 10~43
[Ac2] = 2.0 x 10~6 M as that for the same mixture with [d02] = 2 x 10"2 M'added;
furthermore, each spectrum mirrors the dependence of S02 absorption on wave-
length. This effect can be appreciated in a more quantitative fashion by
inspection of the data of Table VII; although the absolute quantum yields of
sensitized biacetyl phosphorescence increase significantly with increase in
pressure at a given wavelength (see Table IV), there is no significant wave-
length variation of the $sens values for experiments at either a fixed high
pressure or a fixed low pressure (see Table VII). If singlet energy transfer
were important here, one would expect to see a significant increase in the
ratio of $sens at 2650 A in'the high-pressure runs compared to the same ratio
from the low-pressure experiments. Obviously this is not the case.
A second possible alternative source of 3Ac2 in these experiments is the
intersystem crossing reaction (2) with Ac2 acting as M. It is difficult to
determine this ratio accurately since singlet energy transfer can be important
if the [Ac2] represents an appreciable fraction of the quenching medium present,
and the two reactions have identical kinetics with respect to [1S02] and [Ac2].
It appears that at most 17$ of the singlet S02 molecules quenched by collisions
Table VII. The effect of excitation wavelength on the quantum yield of
S02-sensitized phosphorescence excited in S02-Ac2-C02 mixtures
*MM, relative"
Excitation
wavelength, A
2650
2700
2800
2900
3000
3100
3150
[SO,]:
[Ac,]:
[CO,]:
1
1
1
1
1
1
1
1.0 X 10-4 M
2.0 X 10-" M
0.0
.07
.03
.00
.00
.06
.07
.01
1.0 X 10-" M
2.0 X 10-6 M'
2.0 X 10-' M
0.97
0.95
1.00
1.00
1.03
1.14
1.15
1.0 X 10-" M
5.0 X 10-5M
2.0 X 10-2M
1.09
1.00
0.99
1.00
1.04
1.06
1.04
Q
The $sens values are normalized for each series of runs at a given concentration
of reactants by taking the value at 2900 A as unity; all the values were de-
termined by dividing the measured value of the sensitized biacetyl phosphores-
cence intensity by the apparent e for these conditions (see Fig. k).
17
-------
with biacetyl are converted to 3S02 in reaction (2)^3 5^5. Even with the most
favorable conditions which we employed for S02 quenching by Ac2, [C02] = h x
10~3 M and [Ac2] = 1 x 10~4 M, the maximum quantum yield of sensitized emission
from biacetyl would be about i.k x 10~3. For the higher concentration conditions
employed, [C02] = h x 10~2 M and [Ac2] = 1 x 10~4 M, the maximum contribution
to the measured $sens from this effect will be 1-5 x 10"4, about 100 times lower
than the observed quantum yield for these conditions.
In summary, it appears to us that the observed "excess" biacetyl triplet
emission, seen here in experiments at high added-gas pressures, cannot be
rationalized by invoking singlet energy transfer, enhanced intersystem crossing
in (2) with Ac2 as M, or, for that matter, any other possible mechanistic feature
which involves only the emitting XS02 and 3S02 species. Recourse to the somewhat
unusual mechanisms such as those invoked by Cehelnik et al. and Wampler et al.
seems necessary.
Treatment of the data _in terms of_ the Wampler, Horowitz, Calvert mechanism of
"excess" triplet formation _in S02 ajt high added-gas pressures
It is instructive to consider the present results in terms of the two
alternative mechanisms for "excess" triplet SQ& formation in these SOg-added
gas systems at high pressure. Consider first that suggested by Wampler, Horowitz,
and Calvert^:
(I) SO2 + h (2400-3200 A) -» 'SO*
(1) >SO2 + M -» (S02—M)
(2) 'SO;. + M -» 3SO2 + M
(3) 'SOt + M -> X + M
(4) X + M -» 3SOS + M
(5) X —-> SO2 (or nonemitting product)
(6) 3SOZ + M -* (SO2—M) (or products)
(7) 3SO2 + Ac2 -> SO2 + 3Ac2
(8) 3Ac2 —» Ac2 + hvf
(9) 3Ac2 -» Ac2
X in the sequence represents some unidentified species derived from iSQ2. Pre-
sumably it could be an excited, nonemitting singlet state of S02 or one of its
unstable geometrical isomers (S-0-0, S |). It cannot be a nonemitting triplet
stats of S02 which could transfer energy to Ac2. For the conditions employed
here, practically all of the 1S02 molecules are quenched by the added gas M
(COs.. CO. or N2). The steady-state treatment of the above sequence for these
:cr-di~.ions leads to relation (B):
«' »'--
*6..n5
18
-------
Thus, at constant [M], l/$sens sh°uld be linear in l/[Aca]. A test of this de-
pendence is shown in Figure 5 for the data of liable IV at [C02] = 2.5 x 10~2 M,
[S02] = 1-0 x 10~4 M. The data appear to follow well the expected relation.
In fact, each series of runs at other constant" CQs. concentrations also fit the
the expected linear relation demanded by relation?(B). The parameters which
describe the least-squares lines of these plots are summarized in Table VIII.
According_to the functional dependence of (B), the ratio of slope to intercept
Table VIII. Summary of parameters derived from the plots of l/$sens versus
l/[Ac2] from experiments at various fixed pressures of
[C02],
M X 10s
4.0
10.0
25.0
40.0
Slope, M
X 104
6.7 ±0.9
7.2 ±0.8
9.4 ±0.1
9.1 ± 1.2
Intercept
X 10-1
13 ±2
11 ±2
7.7 ±0.4
5.7 ± 3.0
Slope/Intercept,
M X 106
5.2 ± 1.1
6.5 ± 1.4
12.2 ±0.7
16 ±9
(*«A7) X 10*
12 ± 3
6.2 ± 1.3
4.8 ±0.3
4.0 ± 2.2
In each series of experiments, [S02] = 1-0 x 10 ~4 M; data were estimated
by least-squares methods from Figure h and similar plots using the data
of Tableiv. Fjrror limits shown represent the 95$ confidence limits
(twice the standard deviation), assuming only random errors are present
in the data.
from plots of the data of Table IV should be given by relation (c):
(C) [" sl<>pe "I MM]..
|_ intercept Jpig. 5 = A7
Values of (k6/k7) x 104 estimated from relation (c) and the slope/intercept
ratios for each series of data for a given M are: 12 ± 3, 6.2 ± l.J, U.8 ± 0-3,
and h.O ± 2.2; the average of these estimates gives ks/k7 = (6.8 ± 5) x 10~4,
where M = CQ2. This estimate checks reasonably well with the more accurate
estimate of ks/kY which can be calculated from the individual rate constants
determined using the measured lifetime data for 3SCb directly excited by 3829-!
laser pulses: for M = C02, k6 = (l.lU ± 0.07) x lO8^1; k7 = (l.U ± O.l) x 1011
j^./mole-sec42; thus, ks/k7 = (8.1 ± 0.8) x 10~4. It is not clear whether the
apparent trend in the ks/k7 values with increasing pressure is real or a con-
quence of the necessarily high uncertainty which is inherent in the phosphores-
cence data and the slope/intercept method of treating these data to derive
ke/k7. It seems to us that the latter explanation is most realistic. The data
appear to add credence to the kinetic interpretation of "excess" triplet proposed
by Wampler et al . However, one should consider the present results in terms
of the alternative mechanism of Cehelnik et al. before any reasonable conclusions
between alternatives is made.
Treatment of the present data rn terms of the Cehelnik, Spicer. Heic.^len mechanism
of S02 photochemistry a_t high added-gas pressures
22
The major differences between the S02. photolysis mechanisr:. -f warr.pler et al
and that proposed by Cehelnik et al. lies in the involvement of ntv: excited
19
-------
CM
b
- 3
[Ac2], jf/mole x IO"5
Figure ^. Plot of l/$sens vf' V^s^ for data from S02-Ac2-C02
mixture photolyses at 2&T5 A; [S02] = 1-0 x 10~4 M, [C02] = 2.5 x
HT2 M.
SO2 + hv (2400-3200 A)
'SOj + M -> (SOr-M) (or products)
'SOz + M -» 3SO2 + M
singlet (SOa*) and triplet (S02**) species which are presumed to be some non-
emitting excited states, possibly the S02(1A2) and S02(3A2) unidentified states
of S02. Their mechanism for our experimental conditions of high [M] may be
represented as follows:
(I)
(1)
(2)
(10)
(6)
(11)
(12)
(13)
(14)
(15)
(7)
(16)
(8)
(9)
'SO2 + M -> SO2* + M
3SO2 + M -» SO2 + M (or products)
SO2* -» SO2**
SO2* + M -^ SO2 + M (or products)
SO2* + M
SO2**
M
S02**
SO2
** + M
3S02
SO2**
SO2
» SO2 + M (or products)
Ac2 -» S02 + 3Ac2
Ac2^SO2 + 3Ac2
3Ac2 — > Ac2 + Ay,
3Ac2 — *• Ac2
20
-------
There is not a complete correspondence between the S02* of this mechanism and
the X species of the Wampler mechanism although it might appear so at first
sight. X does not generate triplet species unimolecularly as in (11), and it
is not a reactant with CO as postulated by Cehelnik et al for the S02* species.
Presumably, the two triplet species, 3S02 and S02**, can transfer triplet energy
to Ac2. and the expected relation of $sens for this case is given by relation (F):
tn * = ( k* \ |7 *' ^ ( I**!*' \
1 } 9en8 Us + kj |_Ui + *, + kj MAcs]*7 + [M]*,/
+ ( ho \ / *u + *»[M] \ /
U + h + W Un + (*« + *u)[M]/ \[Aci]iu + [M]*u + Jfc
The experimentally observed linear form of the l/$sens-versus-l/[Ac2] plot can
be expected only if both the following equalities hold:
= {An + (*i, +
Although this fortuitous match of rate functions may occur at a given [M], a
perfect match cannot occur at more than one [M] value unless certain peculiar
inequalities exist: ki;L » (k12 + k13) [M] and [M] k15 » k14. However, these
conditions are not tenable. If reaction (11) were the dominant fate of SOa* at
all [M] values, then one would not see the increase in SOs**, and the experi-
mentally required increase in $sens as [M] is increased. The experimental fact
is that the l/$sens-versus-l[Ac2J plot is a straight line within the experimental
error for values of [M] varied over a factor of 10 (U x 10~3 to UO x 10~3 M).
Thus, it appears to us that the Cehelnik mechanism cannot explain the present
results.
The nature of the intermediate X species involved _in S02 photochemistry at_ high
added-gas pressures
In terms of the Wampler et al. mechanism outlined above, we can treat the
present data to derive kinetic constants for the reactions of the species X.
From relation (D),
(D) *.».<» = (^ + ^ + kj (kt + kj V(
we may estimate that part of the quantum yield of biacetyl phosphorescence which
is derived from 3S02 formed in reaction (2). We would expect from relation (D)
that the value of $ sens' ' will be a constant for a given series of experiments
at constant [M]/[Ac2], Combining relations (B) and (D). equation (E) results:
( '
/ *s \ Ai + *i + kt\ (k, + kt\ f *,[M] \
\ * k<(M]j \ k, / \ A, A . *7[Ac2]/
.
Thus5/,the Wampler et al, mechanism predicts a linear relation between l/(^sens "
$sens ) and 1AM^ for runs at constant [M]/[Ac2] ratio. The data of Table V
may be used to estimate some of the kinetic parameters in (E). The most re-
liable data to test this function are those from the highest biacetyl concen-
tration. For these conditions, any error in the value k6/k7 used has a minimum
effect on the $ sense's see relation (D). Results of this treatment are shown
in Figure 6 for C02 as M in Figure 7 for M = CO or N2. If $sens^' ^s assume(J
to be zero, the curved function (defined by the squares in Figs. 6 and 7) is
obtained. For a certain choice of $ sens 5 ^ne ^es"t least-squares straight line
21
-------
(defined by the circles in Figs. 6 and 7) was found. The values of
obtained in this fashion are summarized in Table IX for M = C02, CO,
These estimates may be coupled with the values of ke/(kQ) +
''
(2)
sens
nd NP
-Sin.
to derive from
the most reliable estimate of kg/k7 = (8.1 ± 0.8) x 10
relation (D) new estimates of the intersystem crossing ratio, k2/(ki + k2 + k3).
These data and values of the intercepts and slopes of the linear plots in Figs. 6
and 7 provide the additional rate constant ratios ks/(ki + k2 + k3) and
From these estimates, we can calculate as well the ratio ^/(k^. + k2 + k;
All
of these rate constant data are summarized in Table IX for the three different
M species employed in this work.
The present estimates of k2/(k! + kg + k3) may be compared to those derived
previously from the low-pressure experiments of very different kinetic treatment.
In the earlier work j»^ , a significant correction for 3Ac2 destruction at the
wall was necessary^"' . We find here for M = CC^, CO, and. N2, respectively,
i.o
0.8
|O
I
O
Si.
§ 0.4
0.2
[C02]"', //mole x KT2
Figure 6. Plot of l/($Sens ~ ^sens > vs- l/ICOs]- Data are
from the S02-Ac2-C02 mixture photolyses at 2875 A; fC02]/[Ac2]
400; [S02] =
squares, $sens
1°
circles, $sens(2) = 2.J x 10"3;
22
-------
Table IX. Kinetic parameters related to the reactions of the 1S02 and X
species formed in the 2875 A-irradiation of S02-Ac2 mixtures
with added C02, CO. or N2 gases at high pressures.8
M = CO,
M = CO
M = N8
Slope, M (Figs. 6 or 7)
Intercept (Figs. 6 or 7)
•W" X 10s
k*/(ki + k, + *,)
kt/(k\ + Aj + ij)
*i/(*i + *! + *j)
A 6/^4, mole /I.
1.3 ±0.2
65 ± 10
2.3 ±0.3
0.020 ±0.010
0.14 ±0.02
0.84 ±0.03
0.020 ±0.004
1.4 ±0.2
89 ± 12
2.8 ±0.4
0.021 ±0.010
0.085 ±0.012
0.89 ±0.02
0.016 ±0.003
1.5 ±0.3
72 ± 10
2.5 ±0.3
0.019 ±0.010
0.11 ±0.02
0.88 ±0.03
0.021 ±0.005
Error limits in most cases represent twice the standard deviation as
determined by the method of least squares, assuming only random errors
in the data. For the kg/C^i + kg + k3) estimates, the errors are based
on reasonable estimates of the sensitivity of the linear fit to the
choice in $sens' ' and the other uncertainties in the method employed
here.
0.5
0.4
03
02
o
t
O
„ O.I
r*
=>
"*«
ji 0.5
I
i 0.4
«
_-•
0.3
O.Z
O.I
(a)
(b)
[CO]"', //mole x 10'2
Figure 1 . Plot of l/(§sens - *sens(2))
vs
(a),
or vs.
sens
l/[CO] (b). Date are from the S02-Ac2-N2 end S02-Ac2-C0 mixture
photolyses at 2875 A. In (a), circles, $sens'2^ = 2-5 x 10~3:
in (b), circles. $sens = 2.8 x 10~3; squares in each figure
represent $„ (2) = 0.
-------
W(ki + k2 + ^3) =0.020 ±-0.010, 0.021± 0.010, and 0.019 ±0.010. Within . •'
the error limits of the data for CO and R2j these new estimates check well with
those derived previously at low pressures; with M = CO, k&/(k± + k2 + k3) =
0.01? ± O.OKP°; for M = N2, this ratio is 0.033 ± 0.008^°. There is no previous
estimate of the ratio for M as
Note that neither the fraction of 1S02 quenching collisions which form 3SOa
in (2) nor the fraction of these collisions which form the undefined species X
is very sensitive to the nature of M. For M = COa, CO-, and N2, respectively,
W(ki + ka + ks) = O.lli ± 0.02, 0.085 ± 0.012, and 0.11 ± 0.02. The largest
fraction of the •'•SOa-quenching collisions (8^ - 89$) occurs by reaction (l),
presumably proceeding by energy cascade through the ground-state SOa vibrationaj.
manifold.
The present data give few clues as to the identity of the intermediate
species X. The efficiency of the bimolecular quenching of this species to form
3S02 on collision with M, reaction (4.), is rather insensitive to the nature of
M. Our data give the ratio of the first-order rate constant for the decay of X
to S02 to that for the bimolecular reaction (U); for M = COa, CO, and N2, re-
spectively, k5/k4 = 0.020 ± O.OOU, 0.016 ± 0.005, and 0.021 ± 0.005 mole/4.
Since only a ratio of rate constants involving X can be d.erived here, the life-
time of X or other useful data which might allow characterization of X cannot
be estimated. However the results reported here offer further support for the
existence of the undefined species X in irradiated S02 systems at high pressure.
All of our data support the hypothesis that the only triplet state of S02
involved in both the chemical conversion of CO to COa observed earlier^ s^l,
and. in triplet energy transfer to biacetyl in this work is the optical triplet
state of S02(3Bi). If a second triplet species of S02 is involved, then it must
have properties almost identical to those of the optical triplet. The present
evidence precludes the intermediate state X being a triplet species [e.g.,
SQa(3Aa)] which can transfer energy directly to biacetyl. Since we postulate
that X forms 3SOa rather efficiently on collision with added gases in reaction
(4), we must conclude that it has a combined electronic and internal energy near
equal to or greater than that of the 3SOa state, 73 -T kcal/mole above the
ground-state S02. Conceivably, X could be the S02(1A2) state of SOa, but there
is no experimental evidence which warrents this suggestion now. It may equally
well be a high-energy isomer of sulfur dioxide. The identity of X must await
further definitive experimentation which is aimed specifically at its characteri-
zation.
The theoretical maximum rate of SOa phot o - ox id a t io n jln the lower atmosphere
In a previous study, we have estimated from the rather limited rate constant
and mechansims information then available on SOa photochemistry^ , the rate of
3SOa formation in the sunlight- irradiated lower atmosphere. The existence of
the high-pressure. mechanism of 3SOa formation was not known at that time. Further-
more, the assumption was made that the rate constant ratio k2/(k! + k2 + ks) was
the same for M = N2 and Oa as that found for M = S02. Recently, we have been
able to determine this rate constant ratio for M = Oa as well as N2 and S02; the
ratio for Oa is equal within the experimental error to that for N2 as M> k2/
(kj + ke + k3) = 0.03U ± 0.029 for M = 02 and 0.030 ± 0.013 for M = Na16. How-
ever, these values are well below that for S02 as M, 0.095 ± 0.00539. We do not
now have information on the rate ratios k3/(k! + k2 + ks) and k5/k4 for M = 02;
since 3Ac2 molecules are quenched effectively by 02, the present method cannot
2k
-------
be used to determine these constants. However, in view of the apparent insensi
tivity of these rate ratios to the nature of M for the compounds CQ2, ]&, and
CO, which is reported here, and the observed near' equality of the ratio kg/^
ka + k3) for M = N2 and Oa , it is probably reasonable to assume that these
other rate ratios for M = Og are near equal to those for % as well. If we
assume this, then the present data predict that the rate of 3S02 formation re-
sulting fron sunlight absorption by SOg in the lower atmosphere should be given
by relation (G):
(0
= /. \h + *; + kt +(ki + *; + J (
where Ifl is the rate of solar energy absorption by SOs in the allowed band at
21*00-2200 A, and IB' is the rate of absorption in the "forbidden" band at 3^00-
l^OOO A. Ia and Ia values have been estimated by us previously for various
solar zenith angles^-1-. Substituting in (G) the rate constant ratios derived
here for N2, and taking [M] = O.OUl M, we find
= 0.092 ia + v
Fortuitously, this is nearly identical to the final expression employed by us
previously. Thus , the values derived previously for the maximum theoretical
rates of photo-oxidation of SOa in the lower atmosphere (1.9^" hr at z = 20° , etc.)
are reliable as present rate information will
I-C. The Temperature Dependence and the Mechanism
of the SO_2(fBjJ Quenching Reaction*'?
The detailed mechanism of the singlet quenching reactions (la) and (Ib)
is incompletely understood at present. However, the data suggest that for experi-
ments at pressures below about JO Torr, the singlet quenching reaction (la) leads
to nonemitting products, including ground state SQ^, and little chemical change.
(I) SO, + hv (3200-2400 A) -> 1SO,
(II) SO, + hv (4000-3400 A) -» SO2('Bi)
(la) 'SO, 4- SO* -» (2SOS)
(2a) 'SO, + SO, -» SO,('Bi) + SO,
(Ib) 'SO, + M -» (M-SO,)
(2b) 'SO, + M -» SO,(«B,) + M
(3) 'SOj -» SO, + hvf
(4) 'SO, -* (SO,)
(5) >SO, -» SO,(»Bi)
(6) SO,('Bi) -* S02 + hvp
(7) SOzCfiO -» (SO,)
(8a) S0,('fii) + SO, -» S03
(8b) SO,('B,) + S02 -» (2SO,)
(9) SO,(8B,) + M -» (SOr-M)
-------
where M represents some quencher molecule other thaniOa..-, The product species
designated in the above sequence by^the general fornfulas, (2802), (M - SOg), and
(SOa), are either ground-state SO^' and M molecules., till defined nonemitting
excited states, or other -products '-formed in the qufn^ing step.
The mechanism of S02(3Bj.) quenching in reaction (9) also remains open'ii^
question for most collision partners ,M.u It has been suggested that a large
of the quenching is directly associated with a chemical reaction; .when M is
' , CaHe, CaHe, and the higher paraffinic hydro carbons 19 ? or CaE^, C3H6,
the higher olefinio hydro carbons^5 . However, at room temperature, less than
of the SOa (3Bi) -quenching collisions with CO result in a chemical change (COa
format ion p2 , 22 ^ while it appears that a much smaller fraction of the SO
Qg quenching encounters lead to final chemical change (SOs formation
With Qa as M in (9), an energy transfer reaction, SOa(3Bi) + QsC^g") -» SOs.
* QaC^g^), has been observed to occur at an undetermined efficiency^.
at 25 C the quenching rate . constant for S02(3Bi) with N2 is near equal
to those for" CO and Oa, it is apparent that no chemical quenching is possible
for ]fe as the reactant, since formation of the possible products NaO and SO is
highly endothermic (AH = h'J .6 kcal mole'1). Furthermore, such chemically un-
reactive species as COa, Ar, and He quench S02(3Bi) with fair efficiencies.
Obviously, some unidentified form of physical quenching is operative for many
of these systems.
Our work, described, in this. .section was initiated to help .elucidate the SQa
) -quenching mechanisms through a determination of the Arrhenius parameters
for the quenching rate constants for a variety of chemically reactive and un-
reactive molecules. The rate constants for SOa(3Bi) decay were calculated from
phosphorescence lifetime measurements. The triplet species were generated both
by direct absorption of a 3829-A laser pulse within the "forbidden" band and by
intersystem crossing reactions (2a) and (2b), following excitation of singlet
SOa using a broadband Xe-flash which overlapped the first allowed absorption
band (21*00 -J200 A).
The temperature dependences of the S02(3Bi) -quenching rate constants re-
ported here provide significant new insight into the mechanism of the quenching
reactions involving a wide variety of quencher molecules.
I-C— 1. Experimental Methods
The laser excitation equipment
The source of J829-A light pulse used in one phase of this work, a Raman-
shifted, frequency-doubled ruby laser, has been described previous lyv^^o, jt
generated a 50-^W pulse of 3829-A radiation of 20-nanosec duration. In these
experiments, the laser beam passed through the center of an all-Suprasil cylin-
drical reaction cell, 88 cm in length and 2.5 cm in diameter. The front and rear
windows were fused to the cell body at the Brewster angle. The cell was sur-
rounded by a specially constructed oven which maintained a selected temperature
to within ± 1°C. Front and rear windows on the oven allowed passage of the
excitation beam; phosphorescence emission was detected by an RCA 7265 photo-
multiplier tube which was mounted at right angles to the cell axis and received
light through an additional side window in the oven. A Kodak Wratten gelatin
2B filter placed in front of the detector to remove scattered J829-A radiation.
The phsophorescence intensity decay curves were photographed from the screen
of an oscilloscope, and the data were reduced as described previously^ .
26
-------
The Xe-flashlamp excitation equipment
The majority of the Xe-flashlamp equipment has been described previously^
and it need not be outlined here in detail. However, there were some significant
improvements which were made in the older system. A. fast extinguishing flash
lamp (Suntron 6, of the Xenon Corporation) was employed with a 0.5-M.F, low-
impedance capacitor (Xenon, Model No. C-10-0.5). The flash energy of 10-20 J
created a burst of light (2^00-3200 A) of less than 2 |j,sec half-peak time. The
cell, oven, and detection system were the same as those described previously^l.
Experimental procedures and treatment o£ the data
Gaseous mixtures were prepared, in mercury-free systems using a calibrated
transducer-digital voltmeter combination in the case of the flash excitation ex-
periments. A quartz spiral manometer was employed in the laser excitation
experiments. Homogeneity of the SQa-added gas mixtures was effected using a
thermal gradient pump which was in series with leads near the front and rear of
the photolysis cell. The absorption properties of SOsj "the timing of the excit-
ation light pulse, and the sensitivity of the detector system limited the range
of SQa concentrations which could be employed from about k x 10~6 to U8 x 10~6 M
in the flash experiments and from about 9 x 10~5 to 30 x 10~5 M in the laser
experiments. In the flash excitation experiments, the concentration of the
added quencher gas was varied only over the range of values which ensured a
reasonable accuracy in the lifetime observations; that is, the concentration of
added gas M was controlled so that the lifetime of S02(3Bi) in the mixture with
added gas was no less than one half that observed in SOg alone at the concen-
tration employed in the mixture. In all of the Xe-flash experiments and some
of the laser experiments, about six to seven different [M] values were used at
each temperature with a fixed [SOa]. The rate constant kg was determined from
the least-squares slope of the 1/T-versus-[M] plots. These data for the pure
SQg system are summarized in Table X. Because of space limitations all of the
measured lifetimes, data obtained from SOg-M mixtures are not presented here,
but the rate constants derived from these estimates at several temperatures are
•tabulated in Tables XI, XII, and XIII, where M is an atmospheric gas, a satu-
rated hydrocarbon, and an unsaturated hydrocarbon gas, respectively. In the
third column of these tables, the symbols F and L refer to the experiments in
which a Xe-flash and the laser, respectively, were used for excitation. The
rate constants are shown with 95$ confidence limits for those cases where the
number of data points allowed a meaningful statistical treatment of the random
error; this is the case for all of the Xe-flash experiments. In most of the
laser excitation experiments, the kg value was determined from repeated measure-
ments of T in experiments at fixed [M] and [SOal values. In these cases, the
reproducibility of the rate constant measurements was within about ± 1C$> of the
mean value; this approximate error limit is shown in parentheses following the
rate constant for these cases.
I~C— 2. Results and Discussion
•The mechanism of_ S02(3Bi) quenching
According to the mechanism of S02(3Bi) quenching outlined in the reaction
sequence (l)-(9), the lifetime of the triplet molecules, following their gener-
ation in a pulse of short duration, should be given by relation (A):
(A) 1/T = (*8a + *Sb)[S02] + kt + kl + *9[M]
27
-------
Table X. The S02(3Bi) lifetime at various SOa concentrations"and temperatures.
Temp., — —. tg, (1. mole"1
°K 0.286*0.637 0.875 1.57 2.24, 2.81 3.71 4.26 sec-1) X 10'8
298
309
318
326
333
344
361
369
378
388
399
1
1
1
2
2
2
2
2
3
3
3
.48
.54
.74
.07
.11
.53
.96
.70
.12
.12
.87
2.90
2.95
3.34
4.21
3.98
4.34
5.52
5.52
5.64
6.11
7.48
3.94
4.52
5.02
5.31
5.78
7.05
7.00
7.62
8-43
8.88
10.6
7.37
8.13
9.80
9.70
11.8
11.0
11.5
12.6
16.2
1B.B
19J4
9.65
10.8
12.3
14:0
14.2
15.8
17.1
18.2
21.0
22.3
25.0
11.1
10.8
14.3
13.8
16.0
17.7
19.8
20.9
21 .2
25.2
.28.1
14.3
15.4
16.0
17.8
19.5
21.1
24.7
30. f
27.9
-29.4
30.1
17.7
20.0
21.9
22.8
24.4
27.0
31.1
33.4
36.5
40.0
38,- 5
3.90
4.28
4.68
4.86
5.31
5.83
6.79
7.74
7.77
8.61
8.38
± 0.28=
±0.56
±0.68
±0.62
±0.64
± 0.58
±0.44
±0.36
± 1.08
± 1.02
± 1.16
• Excitation of SOsC^O by 3829-A laser pulse.
»[SO,], M X 10«.
«Error limits shown here and elsewhere in this paper refer to the 95% confidence
limits (2
-------
Table XI. Rate constants for S02(3B;i.) quenching reactions, S02(3Bi) + M -»
(S02-M)(9)5 with various atmospheric components at several tempera-
tures.3
! '
X
IT
10'
*.,
sec"1
(1. mole-'
) X 10~«
(a) 0, =
• 3
3.
3.
36
36
24
0
3-20
3.
3.
3.
3.
14
12
07
00
3.00
2.
2.
2.
2.
2.
2.
91
77
71
64
58
51
2
2
3
2
3
4
.92
.27
•45.
-.02,
.43
.21
.23
.41
.97
.10
.38
.31
.52
.89
.15
±0
(±0
,M
.13
.13)
(±0.15
(±0
±0
(±0
±0
(±0
(±0
(±0
(±0
(±0
(±0
(±0
(b) N2 =
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
2.
2.
2.
2.
2.
2.
2
2
2
42
36
36
35
29
25
24
24
19
14
08
07
00
00
91
79
77
71
64
64
58
51
44
0
0
0
0
0
0
0
1
0
2
2
1
2
2
2
3
.669
.916
.880
.848
.816
.701
.697
.09
.760
.17
.06
.19
.16
.17
.42
.49
.08
.66
.74
.41
.30
.30
.09
±0
±0
(±0
±0
±0
±0
±0
(±0
±0
(±0
±0
(±0
±b
(±0
(±0
±0
(±0
(±0
±0
(±0
(±0
(±0
(±0
.14)
.35
.12)
.24
.20)
•21)
-24)
.33)
.25)
.39)
.42)
M
.072
.136
.09)
.072
.066
.081
.105
-11)
.144
.12)
.17
-12)
.20
.12)
.14)
.10
.21)
.27)
.25
•24)
.23)
.23)
.31)
Excita-
tion
method *>
F
L
L
f
'\
F
L
l/T
X 10'
3.
3.
kt, (1 mole-1
sec-*) X H)-«
(c) Ar -
37 0.475 ±0.
M
0%
36 0,558 (±O.O56)
3-24 0.721 (±0.
3.
3.
3.
3.
072).,.
Excita-
tion
method*
F
L
;-.'L -
Ifi 0.486 ±0.068 . , F . •
14 0.673 (±OJ67)
07 0.617 (±0.062)
06 0.683 ±0.
A-F 3.&F 0.848(=tO.
I,
L
L
L
L
L
L
F
F
L
F
F
F
F
L
F
L
F
L
F
L
L
F
L
L
F
L
L
L
L
2.
2.
2.
2.
2.
2.
2.
2.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
2.
2.
2.
2.
2.
2.
Sft 0.709 ±0.
91
77
71
64
58
51
44
38
36
35 (
30 (
29
26 (
24
19
14
09
07
00
00
00
91
77
.15 (±0.
.45 (±0.
.71 (±0.
.40 (±0.
.42 (±0.
.73 (±0.
.47 (±0.
(d) CO =
.01 ±0.
.22 (±0.
).900 ±0.
).900 ±0.
.13 ±0.
).984 ±0.
.22 (±0.
.13 ±0.
.26 (±0.
.26 ±0.
.55 (±0.
.38 ±0.
.37 ±0.
.45 (±0.
.68 (±0.
.64 (±0.
71 2.30 (±0.
64 2.55 (±0.
58 3.01 (±0.
51 3.36 (±0.
081
085)
094
12)
15)
17)
14)
14)
17)
15)
M
11
12)
12
11
17
19
12)
16
13)
20
16)
23
16
15)
17)
16)
23)
26)
30)
34)
•L '
1
F
A
F
L
L
L
L
L
L
L
F
L
F
F
F
F
L
F
L
F
L
F
F
L
L
L
L
L
L
L
(continued)
29
-------
Table XI, (continued)
1/T-.
X 10'
3.36
3.24
3.14
3.07
3.00
2.91
2.77
2.71
5.64
2.58
2.51
A», (1- mole'1
sec'1) XI O'8
(e) COj = M
1.73 (±0.17)
2.10 (±0.21)
2.06 (±0:21)
2.01 (±0.20)
2.63 (±0.26)
2.93 (±0.29)
-4.39 (±0.44)
3.81 (±0.38)
.£.06 (iO.SJX
4.80 (±0.«S)
5.08 (±0.51)
Excita-
tion
method *
L
L
L
L
L
L
L
L
.£'•
L
L
i/r
X 103
3.36 (
3.24 (
3.14 (
3.07
3.00
2.91
2.77
2. -71
2.64
2.58
2.51
k,, (1. mole-'
sec-') X 10-11
(f) NO = M
3.94 (±0.09)
3.91 (±0.09)
3.86 (±0.09)
.10 (±0.11)
.13 (±0.11)
).99 (±0.10)
.51 (±0.15)
..62 (±0.16)
.58 (±0.16)
.$^(±0.17)
.15 (±0.12)
Excita-
tion
method 6
L
L
L
L
L
L
L
L
L
L
L
I « In Xe-flash experinjents, [SO«) S 4.3 X IQ-'Af, and [M] was varied from 0 to cpn-
i centration shown
-------
Table XII. Rate constants for the SC>2(3Bi)-quenching reactions, SOa^Bi) + M
(S02-M)(9)3 with various saturated hydrocarbon gases at several
temperatures.3
\/T
X 10'
3.36
3.35
3.35
3.35
$Jl
3.24
3.23
3.J4
3.14
3.08
3.07
3.04
3.00
2.98
2.91
2.81
' 2.77
2.71
2.67
! 2.58
2.54
2.51
2.42
2.23
3
3
3
3
3
3
3
2
2
2
2
2
2
2
.36
.35
.33
.18
.14
.07
.00
.98
.91
.77
.74
,71
.64
.58
k», (1. mole"1
sec-1) X 10-"
(a) CH4 = M.
1,15 (±0.12)
.26 ~±0.*5(1)»
.34 ±O.U(2)>
.28 ±p.l.6jf3)b
.30 (±0.13),
.48 ±0.20
.57 ±0,25
.42 (±0.14)
.67 ±0.21
.40 (±0.14)
.69 ±0.15
.59 (±0.16)
2.07 ±0.34
1.95 (±0.20)
2.23 ±0.13
2.69 (±0.27)
2.74 (±0.27)
2.69 ±0.12
3.82 (±0.38)
3.16 ±0.33
3.31 (±0.33)
3.87 ±0.45
5.49 ±1.10
(b)C
1.69
1.43
1.62
1.71
2.27
2.38
2.67
2.46
2.88
3.46
3.44
4.35
5.34
5.18
Excita-
tion
method 6
L
F
F
F
F
•F
L
F
F
L
F
L
F
L
F
L
F
L
L
F
L
F
L
F
F
iH6 = M
(±0
±0
±0
±0
(±0
(±0
(±0
±0
(±0
(±0
±0
(±0
(±0
(±0
.17)
.16
.09
.25
.23)
.24)
.27)
.13
.29)
.35)
.28
.44)
.53)
.52)
L
F
F
F
L
L
L
F
L
L
F
L
L
L
i/r
X 103
2.53 ->
2.51
2.32
2.15
3.40
3.36
3.35
3.35
3.35
3.35
3.25
3.24
3.14
3.09
3.07
3.00
2.99
2.98
2.91
2.77
2.74
2.71
2.64
2.58
2.54
2
2
2
3
3
3
3
3
3
3
3
3
.51
.30
• 21
.38
.36
.35
.35
.35
.35
.35
.34
.32
sec"1)
Excita-
mole-1 tibn
X 10-" method*
4.57 ±0^40
6.86 (±0.69)
7.13 ±0.64
9.59 ±2.3
(c) C,H8 = M
3.45 ±0.^4"
3.23 (±0.32)
4.22 ±0.37(1)'
3.98 ±0.41(2)'
3.68 ±0.26(1)-'
4.14 ±0.56(2)'
3.86 ±0.47
3.87 (±0.39)
3.66 (±0.37)
4.51 ±0.43
4.64 (±0.46)
3.93 (±0.39)
6.11 ±0.98
5.45 ±0.18
5.12 (±0.51)
6.46 (±0.65)
6.91 ±0.35
7.98 (±0.80)
7.60 (±0.76)
9.00 (±0.90)
10.0 ±0.53
10.0
14.7
14.9
(d) n-C
8.48
9.14
10.0
8.57
8.65
8.49
10.13
8.57
8.66
(±1
±1
±1
4" 10
±0
(±o
±1
±0
±0
±0
±0
±0
.0)
.8
.8
= M
.50
.91)
.2-
.48(iy
•5i(2y
.87(3)'
.83(4)'
.48
±0.89
F
L
F
F
F
L
L
L
F
F
F
L
L
F
L
L
F
F
L
L
F
L
L
L
F
L
F
F
F
L
L
F
F
F
F
F
F
(continued)
31
-------
Table XII. (continued)
\/T
X 10"
3
3
3
3
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
3
2
2
2
2
2
2
2
.24
.14
.07
.00
.91
.77
.71
.64
.58
.51
.38
.35
.35
.35
.35
.35
.24
.20
.14
.08
.00
.99
.91
.79
.77
.64
.60
.58
k>, (1. mole-1
sec-1) X 10-"
9
11
13
14
'17
16
21
49
25
23
7
11
7
8
10
11
10
8
9
9
11
10
12
12
16
18
14
22
.87
.5
.0
.2
.1
.3
.1
.8
.1
.5
(e)
.88
.2
.85
.04
.3
.0
.4
.75
.15
.51
.9
.3
.8
.3
.8
.4
.7
.2
(±0
(±1
(±1
(±1
(±1
(±1
(±2
(±2
(±2
(±2
.99)
-2)
-3)
•4)
•7)
•6) ,
.1)
• 0)
.5)
.4)
Excita-
tion
method*
L
L
L
L
L
L
L
L
L
L
iso-CUHio
±0
dbl
±0
dbO
±0
±1
(±1
±0
(±o
±1
(±1
±1
(±1
±0
(±1
(±1
±1
(±2
.52
.5'
.33(1)0
.53(2)'
• 8(3>
• 3(4>
.0)
.50
.92)
.07
-2)
.1
.3)
.6
.7)
.8)
.0
•2)
F
L
F
F
F
F
L
F
L
F
L
F
L
F
L
L
F
L
\/T
xio8
2
2
2
2
2
2
.51
.42
.34
.30
.26
.21
*», (1. mote-1
sec-') X 10-«
24.4
17.2
17.9
20.6
24.4
21.8
(±2.4)
±1.0
±2.0
±3.3
±3.5
±1.0
Excita-
tion
method b
L
F
F
F
F
F
' 'v ' (f) Neopentane
3
3
3
3
2
2
2
2
2
.31
.28
.12
.00
.91
.58
.44
.36
.21
3.63
3.60
4.56
5.66
6.20
9.64
11.1
12.9
15.7
±0'.17
±0.21
±0.45
±0.48
±0.42
±0.82
±0.6
±0.6
±1.3
F
F
F
F
F
F
F
F
F
(g) cyclo-C«Hu
3
3
3
3
3
2
2
2
2
2
.36
.24
.14
.07
.00
.91
.71
.64
.58
.51
24.2
27.0
32.3
35.4
30.1
36.3
42.9
45.7
45.7
48.8
(±2.4)
(±2-7)
(±3.2)
(±3.5)
(±3.0)
(±3.6)
(±4.3)
(±4.6)
(±4-6)
(±4.9)
L
L
L
L
L
L
L
L
L
L
0 In most of the Xe-flash experiments, labeled F under excitation method, the [SOj]
= 4.3 X 10-" M and [M] was varied from 0 to the concentration shown (M X 10'):
CH4, 2.6; CjH,, 2.9; C,Hs, 1.0; n-CJijo, 0.50; iso-C^jo, 0.45;neo-CsHls, 1.1. In most
of the 3829-A laser experiments, labeled L under excitation method, [SOj] = 9.17 X 10~*
M, and [RH] was at the following concentrations (M X 104): CH4, 1.75; C2He, 1.91;
s, 0.789; n-C
-------
Table XIII. Rate constants for the S02(3Bi)-quenching reactions, S02(3Bi) + M -»
(S02-M)(9), with various olefinic and aromatic hydrocarbon, gases at
several temperatures.8
i/r
X 10»
(a)
3.36
3.24
3.14
3.07
3.00
2.91
2.77
to, (1. niole-1
sec->) X 10-10
CjH, = M
3.59(±0.36)
3.83(±0.38)
4.47(±0.45)
3. 23{ ±0.32)
4.70(±0.47)
3.71(±0.37)
4.85(±0.49)
2.71 6.29(±0.63)
2.64 4.85(±0.49)
2.58 5.08(±0.51)
2.51 5.83(.±0.58)
(b) CF.CHF = M
3.36 1.05(±0.11)
3.24 0.95(±0.10)
3.14
3.07
3.00
2.91
2.71
2.64
2.58
2.51
/ X
'
3.36
3.24
3.14
0.98(±0.10)
.21(±0.12)
.19(±0.12)
.28(±0.13)
.71(±0.17)
.79(±0.18)
.77(±0.18)
.88(±0.19)
C>H« = M
10.0 (±1.0)
8.70(±0.87)
10.0 (±1.0)
\IT *9, (1. mote-*.
X 10" sec-1) X 10-10
3.07
3.00
2.91
2.77
2.71
2.64
2.58
(d)
3.36
3.24
3.14
3.07
3.00
2.91
2.77
2.71
2.64
2.58
2.51
10.3 (±1.0)
13.2 (±1.3)
10.3 (±1.0)
11.5 (±1.2)
11.6 (±1.2)
10.6 (±1.1)
: 15.9 (-dsti6)
14.6 (±1,5)
V
cis-2-C«H8 •= M
13.9 (±1.4)
13.5 (±1.4)
13.9 (±1.4)
14.9 (±1.5)
14.9 (±1.5)
16.3 (±1.6)
16.0 (±1.6)
20.5 (±2.1)
18.4 (±1.8)
16.0 (±1.6)
17.4 (±1.7)
(e) Cyclopentene = M
3.36
3.24
3.14
3.07
3.00
2.91
2.77
15.7 (±1.6)
14.6 (±1.5)
16.7 (±1.7)
16.2 (±1.6)
13.5 (±1.4)
16.9 (±1.7)
20.0 (±2.0)
i/r
X 10»
2.71
2.64
2.58
2.51
(0
3.36
-3.24
31 A
.-I*r
3.07
3.00
2.91
2.77
2.71
2.64
2.58
2.51
(K)
\o/
3.36
3.24
3.14
3.07
3.00
2.91
2.77
2.71
2.64
2.58
2.51
*», 0- mole-1
sec-1) X 10-10
18.6 (±1.9)
14.2 (±1.4)
16.3 (±1.6)
19.0 (±1.9)
C,H« = M
11.2 (±1.1)
9.6 (±1.0)
IO'.B (±i!i)
11.6 (±1.2)
9.5 (±1.0)
10.6 (±1.1)
15.6 (±1.6)
15.6 (±1.6)
12.3 (±1.2)
10.6 (±1.1)
C«F« = M
0.706(±0.071)
0.936( ±0.094)
0.887( ±0.089)
0.795( ±0.080)
0.659( ±0.066)
0.875( ±0.088)
0.760( ±0.076)
0.780( ±0.078)
0.883( ±0.088)
1.23 (±0.12)
1.09 (±0.11)
,) excitation was by 3829-A laser; [SO»] = 9.17 X 10"4 M; (RH) was at the
following concentrations (M X 10"): CjH* 1.32; CFjCHF, 3.68; C.H., 0.301; w-2-C4Hfc
0.288; cyclopentene, 0.295; C«H6> 0.301; C«Fe, 4.87.
or 3Bi (0,0,1) levels at 905.7
cm
-i
(2.6 kcal mole"1) and near 96&-1071 cm'1
(2.S-3-1 kcal mole"1),* involving one quantum of vibration in the symmetric and
The range of values has been estimated approximately assuming that the ratios
of the frequencies for the antisymmetric and symmetric stretching modes of
in the S02(3Bi) state are the same as those in the SOg^Ai), 1.18 and S02(1B1)
s-ates, 1.0655.
sn-isymmetric stretching modes, respectively. Note that Brand and coworkers5^
have observed that the line structure in the 3730-3750 A bands within the SOa
(3Bi) 4- S02(1Ai) absorption region shows the transitions in this region to be
perturbed through the mixing of the S02(3Bi) levels with those of an electronic
state other than the ground state. They concluded that in the 3Bi(l,l,0) «-
33
-------
FigureiS. Temperature dependence of the SOi('Bi) lifetimes in pure SO,.
Stern-Volmer plots of the reciprocal lifetime! of the SO,(«Bi) molecules vs. [SO,]
for experiments at several temperatures; excitation of the triplets was made using a
3829-A laser plus.
9.0-
28.0
0>
7.6
2.5
3.0
3.5
I/T
10s
Figure 8. Arrhenius plots of the rate constant for the SOi('Bi) quenching reaction
(9), SOj('Bi) + M -» (SO» - M), for several atmospheric gases: M = SOj
(circles), Oj (squares), and AT (triangles); open symbols are data from experi-
ments with the Xe-flash excitation, and closed symbols are data from runs using the
3829-A-laser excitation of the triplets.
-------
3.5
I/T . "K
FjpurefcnArrhenius plots of the rate constant data for iHc SOjt*Bi) quenching
rcacfion (9), SO,(»Bi) + M — (SO,-M), for several atmospheric gases: M = NO
(diamonds), CO, (triangles), N, (circles), and CO (squares); open symbols are
data from experiments using the Xe-flash excitation, and closed symbols are data
from runs using the 3829-A laser excitation of the triplets; arrows shown with each
compound formula indicate the ordinate scales which apply.
-C4H
10
3.4
plots of the rate constant data for the SOjCB,) quenching
reaction (9), SO2('Bi) + M — (SOj-M), for several paraffinic hydrocarbons:
M = CH4 (circles), C2H, (squares), C8H8 (triangles), and n-C4H10 (diamonds);
open symbols are data from experiments with the Xe-flash excitation, and the
closed symbols are data from runs using the 3829-A laser excitation of the triplets.
35
-------
I 9.4
2.2
it*
H/T , '"K
10s
3-4
Figure laArrfaennis plots of the rate constants for the SOj('Bi) quenching reaction
(9kStWJ*i) 4- M •«••, (SOj-M) > for sotne paraffin^ bydrocarb«ffl»;: :M,, = iso-
C.iHi, (circles), neopentane (triangles), and cyclohexane (squares); open symbols
are data from experiments with the Xe-flash excitation, and the closed symbols
are data from runs using the 3629-A laser excitation of the triplets.
10.0-
I/T ,
FiguretaArrhenius plots of the rate constants for the SOi('Bi) quenching reaction
(9), SO2('Bi) + M -* (SOj-M), for the olefinic and aromatic hydrocarbons which
are indicated on the figure. All these data are derived from experiments em-
ploying the 3829- A laser excitation of the triplets.
36
-------
Table XIV. Summary of Arrhenius parameters for S02(3Bi)-quenching rate constants
with various
M
SO,
i ' "**
02
Ar
CO,
N,
CO
CH4
CjH,
C,H,
n-C4Hi0
iso-C4HJO
Neopentane
cyclo-CgHu
NO
C,H4
CFjCHF
C,H6
cis-2-C4H8
Cyclopentene
C.H.
C.F,
collision partners
: S02( Bl) +
logjo 1/4(1. mole-1 sec-1)] Ea, kcal mole-1
10.04*0,11
10. 60 ±0.47
10.05 ±0.15
10.16 ±0.71
9.69 ±0.60
10.35 ±0.35
9.61 ±0.64
9.30 ± 1.00
9.78 ±0.37
10.26 ±0.44'
9. 90 ±0.50-
9.57 ±0.48
9.98 ± 0.28
9.80 ±0.35
9.50 ±0.54
9.98 ±0.26
10.29 ±0.57
9.85 ±0.11
9.93 ±0.14
10.53 ±0.43
10.39 ± 0.15
10.43 ±0.18
10.21 ±0.34
10.37 ±0.16
10.31 ±0.19
10.64 ±0.20
10.73 ±0.15
10.51 ±0.41
10.10 ±0.18
10.18 ±0.20
10.48 ± 0.06
10.56 ±0.18
11.96 ± 1.02
11.36 ±0.83
11.22 ±0.42
11.65 ±0.85
11.66 ±0.50
11.44 ±0.79
11.38 ± 1.03
10.37 ±0.94
£.0±0.2
2.8 ±0.7
2.0 ±0.2
2.9 ±0.6
2.4 ±0.8
3.2 ±0.5
2.6±0.7
2.2 ±1.5
2. 8 ±0.6
2.8 ± 0,4 ;
-2.7 ±0.6
2.3 ±0.7
2.9 ±0.4
2.4 ±0.6
2.1 ±0.7
2.7 ±0.4
3.1 ±0.6
2.4 ±0.2
2.5 ±0.6
3.2 ±0.4
3.1 ±0.2
3.1 ±0.3
2.3 ±0.5
2.6 ±0.2
2.4 ±0.3
2.3 ±0.3
2.4 ±0.2
2.1 ±0.6
1.6 ±0.3
1.7 ±0.3
2.7 ±0.1
1.6 ±0.3
1.4 ±0.7
1.1 ±0.6
1.7 ±0.4
0.9 ±0.6
0.7 ±0.4
0.4 ±0.5
0.5 ±0.7
0.7 ±0.7
M-» (S02-M)(9)
Source
laser data
flash data"
all data6
laser data
flash data
all data'
laser data
flash data
all data1
• laser data
laser data
flash data
all data'
laser data
flash data
all data'
laser data
flash data
all data
laser data
flash data
all data
laser data
flash data
all data
laser data
all data
laser data
flash data
all data
flash data'
laser data
laser data
laser data
laser data
laser data
laser data
laser data
laser data
laser data
aData of Otsuka and Calvert^1.
All data from present laser and flash experiments of Otsuka and Calvert ex-
cluding the high-temperature point at 373'2°K.
CA11 data from the present study plus data of Sidebottom, et al.. for 298.2°K5°.
dAll data from present study plus data of Jackson and Calvert for 298.
-------
1Ai(0,0,0) band there is a vibronic perturbation by b2 vibrational levels of a
neighboring 3A2 state; also in this band and in the 3Bi(l,0,0) +• ^(.OjOjO)
band, rotational-type perturbations occur which may possibly arise from a neigh-
boring 3Bs state. It is probably not fortuitous "that the energy region above
the 3B!(0,0,0) level at which marked perturbations are observed, in the S02
spectrum, 900-1000 cm"1 or 2.6-2.9 kcal mole"1, is exactly coincident with the
region to which the S02(3Bi) molecule must be excited by collision in order to
effect physical quenching as measured, in this work: Ea = 2.8 ± 0.3 kcal mole'1.
One might hypothesize that in this energy region, the potential energy surface
for the SO^'(3Bi) state intersects that for another lower-lying state, con-
ceivably the S02(SA2) or the SOsC3^) states. SCFMO calculations suggest that
the S02(3A2) state lies somewhat above the S02(3Bi) and that the S02(3Bs>) may
lie somewhat below the SOg^Bi) Ievel55j5^. Such approximate calculations are
unreliable in predicting small energy differences between states, but it seems
likely that the 3B2, 3Bi, and 3Ag states may all have very similar energies.
The theoretically predicted large singlet-triplet energy separation between the
B2 states (> l£,800 cm"1) is in contrast with the much smaller separation
(3856 cm"1) observed experimentally between the BI states. However, this dif-
ference is consistent with the ji •* jt* character of the S02(1'3B2) «- SOa^Ai)
transitions and the n -» jt* character of the S02(lj3Bi) «- S02(1Ai) transitions.
Then conceivably, the physical quenching act may occur as the SQ2(sBj.) mole-
cules are promoted to the energy region (~ 900-1000 cm"1) where intersystem
crossing to another, potential energy surface may occur; this may be the S02(3A2)
or the SOa(3B2) state. If the potential surface crossing is to lead to an ef-
ficient energy sink for the SC^^Bi) molecules and result in quenching of the
emission, then we require that the new state must have either one of the two
following special properties: (l) the new state is very short-lived because it
couples much better with the ground state and allows internal conversion of the
electronic energy through relaxation within the vibrational manifold of the
ground state; (2) the new state is very long-lived, and when vibrational re-
laxation within the new lower state has occurred so that the efficient return to
the S02(3Bi) state is impossible, it cannot emit readily; its radiative lifetime
must be very much longer than that of the S02(3Bi) state. In the latter case,
the failure to detect such a low-lying state experimentally may be rationalized
in terms of the very low intensity of the emission and the probable heterogeneous
destruction of a large fraction of the molecules at the walls of the cell in the
previous emission studies carried out at low pressures. A proper choice between
alternative states and mechanisms of quenching of S02(3Bi) molecules cannot be
made from the information at hand. It should be noted here that some of these
considerations among others have been offered earlier by Sidebottom et al.^6 in
their attempt to rationalize the low quantum yields of emission from the isolated
SC^^Bi) molecules. Heicklen and • coworkers have postulated that the S02(1A2) and
S02(3A2) states, as well as the optical states. S02(1B2) and S02(3Bi), are in-
volved in the chemistry of S02-(XP°3 S02-C2F45D.] and S02-thiophene57 mixtures
irradiated within the first allowed absorption band, of S02. However, their
interpretation and other alternatives suggested in subsequent work by Calvert
and coworkers remain open to question. Calvert's group found that the optical
states alone could account for the chemical reactions in their studies of the
S02-C0 system^,31. their kinetic results required that some ill-defined, non-
reactive, nonemissive state formed from the excited singlet constituted a source
of S02(3B!) reactant in addition to (2a) and (2b) at high pressures. Obviously,
further spectrescopic studies are needed to characterize better the nature of
the unidentified state(s) involved in the physical quenching of the S02(3Bi)
molecule.
38
-------
The S02(3Bi ) quenching by_ the paraf-finic hydrocarbons
The quantum yields of sulfinic acid formation in irradiated SOg-hydrocarbon
mixtures have been studied by Dainton and.,.rvin'° using- the full mercury arc and
Timmons59 employing JIJO-A irradiation. Badcock et al ° have rationalized these
quantum yields well in terms of their SC>2(3Bi) quenching data. It appeared that
the quenching for methane was largely physical in nature. This conclusion is
supported here as well; the kg, A, and Ea values for CH4 as M are near- identical
to those for 02) Ng, Ar, CO, and C02, where this mechanism most certainly pre-
vails. However, in the higher paraffins the rates of product formation paralleled
closely the anticipated rate of S02(3Bi) quenching by the paraffinic hydrocarbon.
The present information on the temperature dependence of these rate constants is
also consistent with this interpretation. The observed activation energies (kcal
mole"1) for SQ2(3Bi) quenching decrease as M changes in the series: ethane, 3-1 ±
0.3; neopentane, 2.7 ± 0.1; propane, 2.U ± 0.2; n-butane, 2.^ ± 0.2; cyclohexane,
1.6 ± 0.3; and isobutane, 1.7 ± O.J. If H atom abstraction is the primary chemi-
cal event in the quenching reaction in these cases, then the observed trend is
consistent with the lowered C-H bond strengths of the hydrocarbons in the order
shown above. In testing this hypothesis it is instructive to attempt to sepa-
rate the quenching constant into components which reflect the different re-
activities for primary, secondary, and tertiary H atoms.
¥e may estimate the reactivity in reaction (9) per methyl group from the
relations (B) and (c):
._. ,
- (B) ACH, =
L.
- • , ineopentane
(C) KCH, = - -
Relations (B) and (c), respectively, applied to the present combined laser and
flash data, give the following rather consistent k.Qj values (£. mole"1 sec"1)
for the temperatures indicated: 298° K: 7.l£ x 10Y, .7790 x 107; 350° K: 1.56 x 108,
1.56 x 108; 1+00°K: 2.72 x 108, 2.53 x 108. Using these estimates, we can treat
the data for C3H8, n-C4HiO3 and cyclohexane to attempt to separate the reactivity
of a CHa group toward S02(3Bi) quenching. According to this simple picture, the
following relations may apply:
— 2*CH,)/2
Relations (E) and (F) yield near-equal estimates of ^QJI while those derived
from (D) are somewhat lower. Values calculated from our present data for
(SL. mole"1 sec"1) from (D), (E), and (F), respectively, are as follows: 29
2.0k x 10s, 3.91 x 10s, k.06 -x 10s; 350° K: 3.36 x 108, 6.96 x 10s, 6.06 x 10s;
1*00° K: ^-72 x 108, 10.5 x 10s, 8.08 x 10s. If we couple the kcng estimates
with the present data for isobutane, the reactivity for the tertiary C-H group
can be estimated from (G):
(G) *CH = *uo-«cHio — 3£CH,
Values of kcjj (I . mole^sec"1) are: 298° K: 6.3! x 108; 350°K: 8.1+5 x 108, UOO°K:
9.9^ x 108.
Values for the S02(3Bi) quenching per hydrogen atom type estimated from
39
-------
these data are shown in the Arrhenius plot of Figure ik, where kp, kg, and tj, .are
the rate constants per primary, secondary, and tertiary H-.-atom, respectively.
The activation energies and preexponential factors which result.from these plots
are summarized in Table XV. The preexponential factor for the S02(3Bi)-quenching
rate constants per H atom is the same for the three types of C-H bonds within the
experimental error. However, it can be seen that the activation energies (kcal
mole."1) decrease regularly from about 2.9 for primary C-H bonds, to 2.0 for.
secondary OH bonds, and to 1.1 for tertiary C-H bonds. . The Evans-Polanyi-type
relation, Ea (kcal mole"1) = 0.26 [DR_n(kcal mole"1) - 86.7] fits this variation
in E reasonably well.
Figurel'4.Arrhenius plots of the SOj('Bi) quenching constants per hydrogen atom
for primary (i,), •secondary (*,), and tertiary (kt) C-H bonds in the paraffinic
hydrocarbons. Data were derived from quenching data of the compounds shown
in parentheses.
Thus it appears that the present data for the temperature dependence of
the S02(3Bi).-quenching rate constants are in accord with the dominance of
chemical quenching in the case of ethane and higher paraffinic hydrocarbons.
The H atom abstraction mechanism favored by Badcock et al-*-° would be consistent
with the data reported here; the insertion reaction suggested by some previous
workers5°,59 cannot be excluded, but it seems less probable for reasons cited
previously-^.
The SOg(3Bi) quenching by_ NO and by the olefinic and aromatic hydro carbons
The very large rate constants for S02(3Bi) quenching by NO and the olefinic
and the aromatic hydrocarbons observed at room temperature by Sidebottom et al^O
are confirmed in this work. In Table XIV we note that most of these compounds
exhibit very large preexponential factors; Iog10 [A (4. mole"1 sec'1)]: NO,
•11.96; C2H4, 11.36; CF2CFH, 11.22; C3H6, 11.65; £is-2-C4H8, 11.66; cyclopentene,
11.Ml; C6H6, 11.38; the value of A for C6F6 is about a factor of 10 lower than
that for C6H6. The activation energies for these reaction are very low: in some
cases Ea is equal to zero within the experimental error. Ea values (kcal mole"1)
are: NO, 1.4 ± 0.?; C2H4, 1.1 ± 0.6; CF2CHF, 1.7 ± 0.4; C3HS, 0.9 ± 0.6; cis-2-
C4H8, 0.7 ± O.U; cyclopentene, O.U ± 0.5; C6H6, 0.5 ± 0.7; and C6F6, 0.7 ± 0.7.
-------
Table XV. Arrhenius parameters Ea and loglo A for the SOa^B].)-quenching rate'
constants (kg) per H atom for C-H bonds of different type.
C-H Bond type
Primary
Secondary
Relation used
• in estimate
(C)
(0)
(10
Ea, kcal mote"1
3.1
2.7
Average 2.9
2.0
2.3
1.6
Average 2.0
mo.-^c%]
9.65
9.40
9.53
,9.44
9.97
9.48
9.63
; Tertiary (G) 1.1 9.58
Recent chemical studies of the SOa(3B1)-cis-2-butene SOa(3B1)-trans-2-
butene60j S02(3Bi)-£is-CHDCHD, and S0g(3Bi)-trans-CDHCDH61 show that the
quenching act in these cases results primarily in an isomerization reaction
which occurs with near-perfect quantum efficiency. The triplet levels of these
hydrocarbons are probably too high to allow triplet energy transfer to the
olefin with the observed very high efficiency. Triplet energy transfer has
been suggested by Cundall and Palmer0^ as the mechanism of isomerization in
SQg-cis-2-butene mixtures irradiated within the SOa^Bi) *- S02(1A1) band. The
primary addition of triplet S02 to the olefinic double bond appears to rationa-
lize all of the results best in our view. In accord with the previous sug-
gestion of Sidebottom et al. , the quenching reaction may proceed through a
highly polarized, charge transfer-like intermediate formed between S02(3Bi)
and the jt-system of the olefin. The collapse of this complex into an addition
product which allows bond rotation probably occurs as observed in the o(3P)-
olefin reactions^. However, in the case of S02(3Bi) addition to a double
bond in the gas phase, the addition complex seems to be unstable toward de-
composition and regenerates isomerized olefin and S02 largely^'°1. The present
data are in accord with this view of the primary quenching act. The preex-
ponential factors for reaction (9) are about the same for all of the olefins
studied and are presumably characteristic of the SOa^BxJ-jt-bond interaction
which is common to all these compounds. There is an apparent decrease in Ea
(kcal mole'1) as one progresses in the series of olefins in the order CF2CFH,
1.7; C2H4, 1.1; C3H6, 0.9; £is_-2-C4H8, 0-7; cyclopentene, Q.h. This is just
the trend expected if the activation energy barrier in these reactions is
inversely related to the polarizability of the n-bond in the formation of the
charge transfer-like complex between S02(3Bi) and the n-system of the olefins.
The accuracy of the present data is not sufficient to establish the origin of
the comparative slowness of the S02(3Bi) quenching rate for C6Fs compared to
C6H6. It appears that in this case a lower preexponential factor as well as
a higher activation energy may exist.
The rate constant for the S02(3Bi)-NO reaction has a low Ea (l.U ± 0.7 kcal
mole"1) and a very high preexponential factor which appears to be peculiar of a
chemical quenching mechanism. Conceivably the 0 atom transfer reaction, NO +
la
-------
+ SO(^Cg~), may occur in this case, as it is somewhat exothermic:
AH = -16 kcal mole'1. However, no spectroscopic or chemical information .has been
obtained from this system to test this or other alternative hypotheses;
I-D. A Kinetic Study of the Photoexcited SOg(3Bi)-Alkene Reactions"^
The rate constants for the SQ2(SB1) quenching reactions have been deter-
mined for a great variety of reactant molecules using phosphorescence lifetime
studies in SOg quencher molecule mixtures30>19520. r^g ^g^ rapi(j reactions
observed in the .previous work were those involving the olefinic hydrocarbons as
the quencher molecule; the rate constants in this case were very hear the col-
lision number*^ j ^7, This result focuses new interest on the SOa(3Bi)-alkene
reaction, since it may occur at a significant rate even in very dilute mixtures
of alkene in the polluted atmosphere^ the rate constant for S02(3Bi) quenching
with Qg and Ife is only about 1/1^00 of that for the butenes3°.
.Little unambiguous information is now available concerning the chemistry
of the SQ2(3Bi)-alkene interactions, although some significant insight into the
physical parameters of the reaction have been derived, in recent years^j^'.
Sidebottom and coworkers^ have.suggested that the reaction may proceed through
a highly polarized, charge-transfer-like intermediate formed between S(>2(3Bi)
and the it-system of the olef in. The collapse of this complex into-an inter-
mediate addition product may occur as observed in the analogous 0(3P)-olefin
reactions"^.
There are no published chemical studies which can be attributed unambig-
uously to the S02(3Bi)-alkene reactions. There is considerable information
related to the reactions in SOa-alkene mixtures irradiated in the SOa(1Bi) «-
SQ2(A, """Ai) band, but the fraction of the observed products which is derived
from the S02(3Bi) state is unclear. Thus the previous studies of SOg-alkene
mixture photolysis of Bristow and Dainton°5, Cundall and Palmer^2, Penzhorn
and Gusten66, and Cox°7 suggest that S02 initially excited to the xBi state
may lead to the isomerization of the olefin. It is probable that the 3Bi
state is one of the reactants which leads to this overall change. Penzhorn and
Gusten°6 have interpreted their results in terms of S02(3Bi) as the sole re-
actant with the olefin. Coxy7 gave this interpretation as well but concluded
that some participation of other excited states may lead to olefin isomerization
as well. The initial driving force for the change was thought to be the result
of a. triplet energy transfer from SQs to olefin"2, although this interpretation
has been questioned by Sidebottom and coworkers^O. Dainton and Ivin5° photolyzed
SOa in gaseous mixtures with 1-butene and found a product which had character-
istics of an unsaturated sulfinic acid. Polysulfones are formed in irradiated
SC>2-olefin solutions"!?. Cehelnik and coworkers^l found that an 0-atom transfer
occurred from photoexcited SOa i° gaseous CF2CF2 mixtures, and CF20 was an
observed product. Jones and Adelman°" identified cyclic sulfites as products
of irradiated SC>2-alkene-02 solutions. The very great variety of products
which have been observed in the 2500-3200-H irradiated SOa-alkene mixtures have
been rationalized by a .number of different reaction schemes.
In this portion of our work we have attempted to characterize quantitatively
the chemical pathways in which one of the excited states of sulfur dioxide, the
phosphorescent S02(3Bi) state, reacts with the olefins cis-2-butene and trans-2-
butene. We have excited this state directly by irradiating S02 within its
"forbidden" S02(3Bi) «- S02(X, 1Ai) band at 3500-UlOO' A. Quantum yields of
olefin loss and isomerization were determined in a variety of experiments. In
-------
addition aerosol growth was .monitored using light scattering measurements. In
other experiments we have excited S02(3Bi) molecules using a 3630-A laser beam
and have determined directly the quenching rate constants for the 2cbutenes..
These results give a substantially new insight into the nature of the chemical
mechanism of the S02(3Bi)-alkene interactions.
I-D r-1. - Experimental
Equipment for Photochemical Studies
Photolyses were carried out in a cylindrical Pyrex reaction cell (diameter
4.5 cm; length 50 cm). The cell was connected in series with a thermal gradient
pump for reactant mixing. The grease-free vacuum line and gas-handling system
employed only Teflon .stopcocks.. Pressures of 20 torr or less were measured
directly using a calibrated transducer-digital voltmeter combination. Pressures
greater than 20 torr were measured on a mercury manometer using the transducer
as a null device. The light source was an Osram XBO ^50 high-pressure xenon arc
held in an air-cooled housing equipped with a quartz condensing lens assembly.
The light was filtered by passing it through several solutions and glass filters
in combination6^: (a) cuS04 (5.0 g/100 ml HsO), 10-cm path; (b) 2,7-dimethyl-
3,6-diazocyclohepta-l,6-diene perchlorate (0.01 g/100 ml of HsO), 1-cm path;
(c) Jena WG-1, 1-mm glass plate; (d) Corning 7-51 (59TO), 5-mm glass plate;
(e) the Pyrex front window of the photolysis cell. The light incident on the
reactant mixture extended in a band from 350° to ^10° A and matched well the
long'wavelength region of the 3Bi «- X, xAi transition in SOgf®. The intensity
versus wavelength spectrum of the arc and filter combination was determined
using a Turner spectrofluorometer (model 210) in the luminescence-energy mode.
A nearly uniform parallel beam of light was formed using light stops and con-
densing lens. The intensity of the beam was determined at regular intervals
employing a 3.3-cm path of a 0.012 M potassium ferrioxalate solution and the
procedures described by Hatchard and Parker?!. The relative intensity of the
incident light was also monitored continuously during photolyses by reflecting
a small fraction of the beam with a quartz plate, angled with respect to the
optical axis of the cell, onto a 935 phototube. These reflectance measurements
and actinometry performed both after lamp warmup of one hour and at the finish
of the eight-hour runs, showed less than ±3$ variation in incident intensity
during the run. However, because aerosol is formed in this system of alkene-
SO& as the run progresses, the uncertainty in the absorbed intensities is some-
what greater, especially in runs at high pressures of SOs- The average ex-
tinction coefficient of SOa for the band of incident light employed was esti-
mated by two procedures. The small fraction of light absorbed by SOa was
measured directly using a thermopile potentiometer-galvanometer system. This
method gave the average decadic extinction coefficient e = 0.0266 ± 0.006U A./
mole*cm. We also calculated e from the experimentally measured relative in-
cident intensity versus wavelength data obtained, from our lamp-filter system
and experimental e versus wavelength data for SOa in the forbidden 3Bi «-,
X xAi band' determined on a spectrophotometer;
i = f ** t lad(\)/\ X* /orf(X) = 0.0206 ± 0.002 l./mole-cm, A, = 3400 A and A2 = 4300 A.
J xi J xi
In one series of experiments the intensity of light scattering from aerosol
was determined throughout the photolyses using a system designed by Dr. Charles
Badcock. A horizontal He-Ne laser beam transversed the photolysis cell at
right angles to the cell axis. An 1P28 phototube detector system was mounted
-------
in a light shield which sucroun'ded the cell and. received scattered radiation at
90° from the laser.'beam and--at Bright angles to the cell axis.
Equipment for SOg(3Bi) Phosphorescence, quenching studies
A tunable dye laser, using 3,3/-dimethyl-2,2/-6xatricarbocyanine iodide
(Candella Co.) in .acetone solution.as the active medium and a 75-MW ruby laser
as the pumping source, was employed to excite SQg directly from the ground
state to the 3Bj. state. The 20-nanosecond 5-10 kW dye laser pulse had a half-
width of about 2 A and was centered at 363° A. The cell used for excitation
was a 21-liter Pyrex flask to the sides of which were attached cylindrical
sections with Supracil windows sealed at the Brewster angle. The relative
intensity of the emission from the SOfe(3Bi) as a function of the time was
observed at right angles to the incident laser beam using a phototube receiver
and photographing the oscilloscopic trace of the signal.
A Perkin-Elmer (model Fll) flame ionization gas chromatograph was used in
the analyses, of the cis- and trans-2-butene. A J-ft by 1/8-in stainless-steel
column packed with !($> by weight Nad on 9$ by weight activated alumina (90-
100 mesh) gave excellent separations of the butene isomers. The column was
prepared as described by Brookman and SawyerT2. it was found that optimum
isomer separation of the butenes was attained at a column temperature of 130 C
rather than 200 C for our conditions. We found that large concentrations of
SOg introduced error in the analyses, first by increasing the noise level in the
FID signal and second (in cases of very large SQg concentrations, i.e., PgQg >
300 torr) by sensitizing the column to catalytically convert one butene isomer
to the other. This problem was solved by use of a 1-ft by 1/4-in stainless
steel precolumn packed with NaOH-coated glass beads, which stripped away the SOa-
Calibration curves from the butenes were made with and without the stripping
precolumn, and the results were identical within the experimental error. The
major limitation to the use of the alumina-salt column was that of sample size.
With butene samples of 0.07 torr in 10 cm3 (3.8 x 10~8 moles) or greater, tailing
was observed; therefore, samples of this size were reduced by expansion before
analysis. The column gave linear response to the butenes from 0.07 torr in
10 cm3 down to the limit of the FID response (0.0001 torr). Peak areas were
automatically calculated by an Infotronics (model CRS-11 HDB) electronic digital
integrator with paper tape print out.
Quantum Yield Measurements of_ SOg(3Bi)-sensitized. Isomerization of the 2-Butenes
The photolysis of S02-2-butene (cis- or trans-) mixtures within the triplet
absorption band of SOa results largely in isomerization. Aerosol formation was
observed in experiments with high concentrations of S02> but 2-butene loss was
very small. Photolysis studies in which the photostationary concentrations of
isomers were determined, showed. < 6.0 ± l.C$> butene loss in 1000 min of ir-
radiation time. From the known rate of light absorption in these experiments
we estimate the quantum yield of butene loss to be less than or equal to 0.0l£ ±
0.003. A summary of the quantum yields of isomerization from photolysis experi-
ments of SQ2-cis-2-butene and S02-trans-2-butene mixtures is given in Table XVI.
Quantum yields of isomerization were calculated using initial rates determined
from the least squares fit of the percent isomerization versus time curves. In
most cases data with conversions ^ 15% and irradiation times less than 100 min
were used in the treatment of the data. This reduced the possible error intro-
duced by the occurrence of the back reaction, butene loss, and. uncertainties in
the absorbed intensity of light which resulted at long exposure times from light
-------
Table XVI. Quantum yields, of SOa^Bi)-sensitized isomerization of the 2-butenes.a
Run
No.
1
2
3
It
5
6
7
8
9
10
11
12
13
Hi
15
16
17-
18
19*"
20?
22!;
23c
25 c
26
27
26
29
Incident Light Initial Quantum
Par. . Tbrr Pn u , Torr Ron_/P/< u Intensity, quanta Yield of Isomeri-
802 «Vfe • '802' c4Ho cell-iBec-i x 10-ie zatjon of j.Butene
17.U
97:-8
60.8
97-9
80.1
123
98.7
173
89:9
59.9
93-2
28T
96.1
375
. 96-3
518
«£=T._ •-
325
109
137
192
301
297
302
299
23.6
270
385
737
0.803
2.0k
0.683
1.00
0.513
0.6U3
10.508'
0.617
0.283
0.187
0.273
0.655 •••
o,W9-
OI6S1-
0.131
0.601..
'0.883- -
0,201
0.981
0.71k
0.612
0.523
0.523
0.523
0.523
0.383
0.697
0.1*83
0.885
21.7
k7-9
89.0
97-9
156
191
19*.
280
318
320 ;
3*1
5U8'
598
750 ..
658
930
l£l7
111
13
3.81
2*95
3.21.
.}.*
3.2U
S-.60
2.76
2.68
2.75
2.93
3.32
3-33
3-36
0.1.8
0.1(1
0.50
0.1(8
0.37
0.1(0
0.53
OM
o.to
0.35
0.1)6
0.29
0.30
0.30
0.25
O.23
o.ei
0-.13
0.35
0.33
0-33
Q.Sh
0.28
0.23
0.22
0.31
0.25
0.1)l
0.15
"Excitation of SOj('fij) was effected by irradiation using 3500-4100-A light; in runs
; 1-25 the initial olefin was the cw-2-butene isomer, and in runs 26-29 the initial olefin was
; the ut varied
butene concentration. These data are given in Table XVIII. Each lifetime point
was derived from the average of several (usually five) determinations. The error
limits shown for each data point represent the rms deviation of the individual
determinations of I/T from the average.
-------
Table XVII. Aerosol formation in the SOa(3Bi) reaction in S02-2-butene mixtures.8
-P-
a\
Run
No.
i
2
J>
k
5
6
7
8
9
1C
11
12
13
li*
it;
J*s
16
IT
18
19
20
21
22
S3
21*
25
26
Reactant Pressure, Rate of Increase Tine of First Rate of Decrease Relative kite
Torr b of 6J28 k Scattered Detection of of 1*100-3500 A of SOaC^i)
SOa 2-C4He, Light Intensity, 6328 i Li,jht Transmitted Light Quenching:
Relative Units Scatter, icin. Intensity, $ nln"1 80s 2-C4He
650
635
626
617
609
519
509
5C2
371*
368
3»
559
287
283
27Q
<- 1 y
215
212
pno
c.\jy
T37
727
716
707
38U
379
373
368
0.000
0.683
0.671*
o.66i»
0.655
o.eoit
0.596
0.587
0.627
0.618
0.610
0.601
0.635
0.625
0.617
0.612
0.603
O ^Q^
w 'Jjj
0.886
0.871*
0.861
0.850
0.1*83
0.1*76
0.1*70
0.1*63
3-1
5-3
5-5
k-3
3-9
3-7
5.U
U.6
3-9
2.6
2.5
1-3
0.0
1-3
f) O
w * v
0.0
0.71
0*r(\
• [ Q
3-0
3-7
U.I*
3.*
0.0
1.9U
0.97
0.63
1
6
3
1
3
12
8
6
13
16
16
/ 18
1*2
28
if.
yj
55
39
Tin
yi
2
5
9
5
30
19
2U
22
0.81
1.08
1.25
1.15
1.22
0.76
0.86
0.83
0.58
0.62
0.67
0.50
c
c
0.22
c
£
1.30
1.17
0.83
0.96
0.3
0.6
0.5
0.6
1.00
o.Tl
0.69
0.69
0.68
0.56
0.55
0.5U
O.J6
0-35
0.35
O.JU
0.25
0.25
Opll
• CH
0.17
0.16
016
• JJO
0.80
0.78
O.T7
0.76
0.1*1
O.UO
0.1*0
0.39
0.00
0.27
0.27
0.27
0.26
0.21*
0.2}
0.23
0.22
0.82
0.21
0.81
0.20
0.19
Ola
••Ly
0.17
0.16
n ifi
V *4JD
0.3U
0.31*
0-33
0.33 •
0.18
0.18
0.16 .
0.17
(6)
W
»
u.o
1..6
i*.3
b.7
3-2
3-7
3-6
2.6
2.8
3-2
2.l«
-_~
---
1.3
3-8
3A
2.5
• 2.9'
1.7
3-3
2.8
3-5
(7)
Rl
0.8
1-5
1.8
1.7
1.8
1.1*
1.6
1.5
1.6
1.8
1.9
1.5
--.
1-3
i;6
1.5
1.1
1-3
0.7
1.5
1-3
1.5
(8)
«
a>
20
20
16
15
15
23
20
18
12
12
6
0
7
0
l».l*
ti 8
*f »O
• 8.8
11
13
00.
0
11
5.U
3-7
(9)
$
3-1
7-5
8.0
6.2
5-7
6.6
9.8
8.5
10.8
7. 1«
7-1
3.8
0.0
5-2
Of)
• w
0.0
u.u
L n
** . o
3-8
l*.7
5-7
!*.!«
•o.o
U.9
2.1*
1.6
" All runs were carried out at room temperature. No detectable aerosol was present at the start of each photolysis period. Ex-
citation was effected using 3500-4100-A light.
* The initial reactant, 2-butene, was the m-isomer in runs 1-18 and the trans-isomtr in runs 19-26.
e Data not obtained in this experiment.
-------
Table XVIII. Lifetimes of S02(3Bi) molecules excited in S02-2-butene mixtures.8
Pressure, cis-£-C4He
x lO^Tbrr
0.0
1.81*
Ml
6.50
8.71*
/aft.59-
.1^.72
15.05
17.25
l/T X 10"4,
sec"1
' i*.66 ± 0.32;
5.79 ± 0.58
6.77 ± 0.50
9.17 ± 0.1*3
11.1*7 4 0.77
13.85 ± 1.1*2
13.60 ± 0.58
ll*.95 ± 0.68
15.97 ± 0.96
Pressure, trans -2-C^Ha
x 10s Torr
0.0
2.21*
3.06
l*.l*8
•J..52
6.51*
6.98
9.23
11.1*0
16.1*6
19-33
l/T X 10-*
sec"1
5:32 ± 1.13
6.12 ± 1.15
6.25 ± 0.90
7.1*9 ± 0.70
7.86 ± 0.1*3
9.69 ± 0.89
ID. 59 ± l-^S
12.12* 1.59
1«*.23 ± 2.02
16.32 A 5.09 '
l£.87 ± 0.68
« The SOt('Bi) molecules were excited at 21°C using a 3630-A laser pulse in mixtures
containing 1.55 torr of SO j and pressures of the 2-butene isomer indicated.
* Error limits shown represent the rms deviation of the measured values from the
reported average 1 /r value.
I-D —2 . Discussion of Results
Mechanism of 'S02(3Bi) Induced c is -trans Isomerization of 2-Butenes
The major chemical result of the excitation of SQs^Bi) molecules in cis-2
butene or trans -2-butene containing mixtures is an isomer ization of the alkene.
Note in Figure 15 that the composition of a mixture containing initially either
pure trans- or pure cis-isomer with SOa reaches the same photos tat io nary state
after extended irradiation with light absorbed within the S02(3Bi) *• SOa(X, x
500 1000
Irrodiotion Time, min.
1500
5.Time dependence of [rranr-2-butene]/[n>-2-butene] ratio in 3500-4100-A.
irradiated mixtures of 2-butenes and SOs. Upper curve—initial pressures (torr):
SOS 737,
-------
band; [trans-2-butane]/[cisr-2-butene] = 1.9 ± 0.1. Some interesting mechanistic
details concerning this chemical change can be had from a consideration of the
initial quantum yields of SC>2(^Bi)-sensitized isomerization of the 2-butenes
reported in Table .XVI. First consider the simplest possible reaction mechanism
(l)-(5) which can rationalize the data. Other alternatives and possible compli-
cations will be considered subsequently.
Here we have assumed that the S02(3Bi)-alkene interaction leads to transient
triplet addition complexes which, within the time scale of events measured in
this work, are structurally indistinguishable and decay rapidly to ground state
cis- or trans-2-butene" and SOg- If these reactions occur, then the quantum yield
results should be described by relations (A) and (B):
(A) i~«=
(B) 'J--
+
ib\/*». + *«>\ [so,]
Xk* + k
*3b
+
The data of Table XVI have been plotted in Figure 16 to test the theoretically
expected linear relationship between the reciprocal of the initial quantum yield
of isomerization and the [S02]/[2-C4Hel ratio. Within the experimental error of
the data a reasonable fit to relations (A) and (B) is seen. The slopes and inter-
cepts of the least squares lines which best fit these data provide the following
rate constant estimates in terms of the mechanism outlined; error limits represent
the 95$ confidence limits (2a).
W(*ta + *5b) = 0.60 ±0.10
W(*i. + *5b) = °-39 ±0.15
[(*«. + *2b)A3a][*6a + *6b)/Aib] = (3.45 ±0.48) X 10-*
[(**. +-**b)/A«,][*6. + *6b)/*ta] = (5.18 ± 1.60) X 10-»
Within the large experimental error the sum of the ratios kgg^ksa +
k5^/(k5a + KSI-,) is unity as one would expect if a common precursor to the cis-
and trans -2-butenes is involved. Combining the slope to intercept ratios
obtained here with the experimental values for k.2a + ^2b = (3-9 * O.l) x 108
.£ ./m°3_e'£ec at 25° C^O 3 26 ,4? } the data give the following rough estimates for
-------
ksa and ksb:
A3a = (1.9 ± 0.4) X 10" l./mole-sec
.*3b'= (1.9 it 0.9) X 1011 l./mole-sec
According to the suggested mechanism the ratio of [trans-2-butene]/[ cis-2-
butene] at the photostationary state should be equal to (k3a/k3b) (ksbAsa) 5
using the above rate constant estimates the expected trans/cis ratio is 1.5 ± 1.0.
Within the very large error limits this is in accord with the directly measured
ratio of 1-9 ± 0.1.
The quantum yield data presented suggest the. near equality of the
quenching rate constants for cis-2-butene and trans-2-butene as quenchers. To
test this point more directly we determined these constants independently in
SQ2(3Bi) lifetime studies in this work using a 5630-! laser pulse excitation.
A Stern-Volmer plot of these data (Table XVTIl) is shown in Figure 17 . The
slopes of these lines give direct estimates of ksa and ksb at the temperature
of the experiments, 21 C:
*8a = (1.29 ± 0.16) X 10" l./mole-sec
3b = (1.22 ± 0.15) X 10"
300
600 900 1200
[S02]/[2-C4HB]
1500 1800
Figure 1.6.Plot of reciprocal of initial quantum yield of isomerization of 2-butenes
versus [SOj]/[2-butene] ratio from 3500-4100-A irradiated mixtures. Upper
curve—/ranj-2-butene and SOi; lower curve—ro-2-butene and SOj; temperature
25°C.
201-
15
70IO
— trons-2-butene—,
-t-
20
O
15 *•
10
5 :
0 5 10 15
Pressure of 2-Butene x io5, Torr
20
Figure! 7. Stern-Volmer plot of SOzCBi) lifetime data. Upper curve—in lrans-1-
butene-SOj mixtures; lower curve—in m-2-butene-SO2 mixtures; excitation of SOj
by 3630-A laser pulse; pressure of SO. constant at 1.55 torr; temperature 21°C.
-------
These estimates ..agree well within the experimental error with the less accurate
indirect estimates derived from the slope/intercept ratios obtained from the l/§
versus [S02]/[2-C4H8] plots in Figure 16. The lifetime data confirm the con-
clusion that the SC>2(3Bi) quenching rate constants for cis- and trans-2-butenes
are the same within the experimental error. These estimates are in fair agree-
ment with those derived indirectly by Cox°7 from photochemical studies of SOa-
2-butene mixtures excited within the first allowed band of SOa which he attributed
to these reactions: k3a = (1.62 ± 0.08) x 1011 and ksb = (1.U2 ± 0.09) x 1011 &./
mole'sec. However, this agreement may be somewhat fortuitious, since we have
observed some previously unrecognized complicating factors related to -the Cox
interpretation of the reactions of singlet excited SOs in 2-butene mixtures; see
the section I-E following.
The present data for k3a show that the magnitude of this rate constant is
insensitive to the degree of the initial vibrational excitation of the S02(3Bi)
molecule formed, at least over the range of the triplet energies excited in the
broadband quantum yield experiments reported here. In our previous work S02(SB1)
excitation was effected near the (0,1.0) level at 5829 ± 1 A (7^.6 kcal/mole),
and. we found k3a = (l.jlf ± 0.10) x lO11 SL . /mole-sec at 25° (T2 and (1.38 ± 0.15)
x 1011 jfc./mole'sec at 21° C, calculated from the measured Arrhenius parameters
derived from studies at several temperatures^'. Thus our present neglect of the
S02(3Bi) vibrational relaxation steps and alternate reactions (2) and (3) from
various vibronic levels is justified.
The best estimate of the rate constant ratio kgb/k^ is obtained using our
directly measured values of k3a and k3b reported here at the observed photo-
stationary state ratio of trans /cis = 1.9 ± 0.1. These give ksb/ksa = 1.80 ±
0.35. If the triplet SC^-olefin addition complex is truly structurally in-
distinguishable for the cis- and trans -2-butene reactions (3a) and (3b), and
hence kga/^a + ^sb) + k5-b/(k5a + kgb) = 1 as the quantum yield data suggest
within the large error limits, we can derive our most reliable estimate of the
rate constant ratios: k5a/(k5a + kgb) = 0.35 ± 0.05 and ksb/tksa + ksb) = 0.65 ±
0.05. Note that the distribution of the trans- and cis-2-butenes is very dif-
ferent from the thermally equilibrated mixture at 25 C for which trans/cis = 3 -85*,
it-
Obtained from the least squares extrapolation to room temperature of the equili-
brium data of ref. 73-
and the observed ratio of 0.92-0.9^ obtained when benzene triplet was the sensi-
tizer of the butene isomerization'^~77. Obviously real mechanism differences
exist between these different systems.
Consider now the experimental results which bear on several mechanistic
alternatives to the suggested reaction scheme.
Conceivably the triplet energy transfer from S02(3Bi) may occur directly to
excite the triplet alkene which then may relax to form the cis- and trans-isomers
in their ground states. The major problem which one must face in the evaluation
of this alternative is the estimation of the energies of the different configu-
rations of the triplet 2-butene molecules. Solution phase studies of Fukano and
Sato?" suggest that many organic sensitizers with triplet energies greater than
TO kcal/mole (pyrazine, ET = 8^.8, to acetophenone, E^ = 73-6, 76.3) efficiently
isomerize cis- or trans-2-butene. Even some sensitizers with triplet energies
less than 70 kcal/mole cause isomerization (benzophenone, Ef = 68.5, 69-33 to
50
-------
naphthalene, ET = 60.9), although the reaction is very slow in these cases. Also
Shekk and Alfimov79 found that cis- and trans-2-"butane' quenched the phosphorescence
of various organic triplet excited species in polycrystalline glasses of the alkenes
at liquid nitrogen temperature, when the triplet donors had. energies which lay
above 73-6-71-8 kcal/mole. Penzhorn and Gusten°° estimated from the oxygen per-
turbed singlet-triplet spectra of 2-butene solutions that for the trans-2-butene
ET > 76 kcal/mole and for cis-2-butene ET = 70 ± 2 kcal/mole. Indeed triplet
energy transfer was suggested by Cundall and Palmer0^ as the mechanism of the
cis-2-butene ** trans-2-butene isomerization observed in gaseous SOg-butene mix-
tures irradiated within the first allowed band, of SO^', this alternative has also
been considered among others in more recent studies of this systemP°5°7. However,
the possibility that the cis-trans isomerization and the phosphorescence quenching
observed in the previous solution phase studies resulted from the addition of the
triplet donor to the alkene double bond cannot be discounted from the evidence at
hand. Furthermore, as Wagner and Hammond have pointed out previously^, the fact
that the triplet energy transfer may occur in solution even at the diffusion-con-
trolled rate merely means that its inefficiency is not so great that reaction
cannot occur during the characteristic prolonged encounter time between reactants
in solution.
The results of gas phase studies appear to give more definitive data on the
nature and efficiency of the triplet energy transfer to the alkenes. The transfer
of triplet energy from excited acetone (ET = 77 ± 2 kcal/mole)* to the olefinic
An upper limit of 79 kcal/mole can be set for acetone triplet energy by the
position of the 0-0 band of the SQ -» Si transition at 3600 A°2. A lower limit
has been suggested by Sidebottom and coworkers20} who found that quenching
(3Bi) by acetone is rather inefficient (kg = 1.2 x 109 4./mole*sec), yet S
quenches acetone triplet at a rate near the collision number"?. Thus it is
highly probable that the acetone triplet lies above 73-6 kcal/mole. If the
rate of quenching of S02(3Bi) by acetone is from the energy transfer reaction
exclusively and the relative slowness is a consequence of the endothermicity
of the reaction alone, then ET for acetone - 77 kcal/mole.
hydrocarbons is relatively inefficient in the gas phase; in fact the probability
that energy transfer will occur on collision between triplet acetone and 2-pentene
is only 2.1 x 10~5. Rebbert and Ausloos"1 attribute this inefficiency to the
endothermicity of the energy transfer reaction. As Sidebottom and coworkers
have argued^O, if in the acetone triplet-olefin reaction, energy transfer is
already inefficient because of its endothermicity, the energy transfer from
S02(3Bi), ET = 73-6 kcal/mole, to olefins must be much less efficient.
Ion energy-loss spectres copy with the olefins has been carried out by
Moore°^ who estimated that the energy for the maximum transition intensity (re-
lated to the ground, state to first excited triplet transition) both in cis- and
t.rans-2-butenes occurs at h.2 eV (97 ± 2 kcal/mole); recent results from sililar
studies of Flicker, Mosher, and Kupperman°5 gave preliminary values for the
vertical transition energy for cis-2-butene of U.22 ± 0.07 eV (97 ± 2 kcal) and
for trans-2-butene of U.25 ± 0.06 eV (98 ± 1 kcal). Moore's data suggest that
the low-intensity onset of the transition with cis-2-butene appears to be near
the S02(3Bi) energy of 73-6 kcal/mole.
5.1
-------
On the "basis of all available information it appears to us that the
energy may be insufficient to excite the alkene triplets with a. rate constant
which is near the collision number'as was observed experimentally, since most
S02(3Bi)-alkene encounters must involve molecules of the olefin in the near-planar
configurations expected at 25° C in the gas phase; direct triplet energy transfer
is probably considerably endothermic for most of these encounters. We also note
from the data of Sidebottom and coworkers that the measured SOa(3Bi) quenching
rate constant for trichlorethylene, (2.3 ± 0.2) x 1010 ^./mole*sec, is less than
that for ethylene, (H.2 ± 0.5) x 1010 H ./mole*sec, although the triplet energy
for the chloroethylenes lies below that of ethylene°^>°°>°f. This again does not
support a hypothesis of quenching by a direct triplet energy transfer mechanism.
The rate constants (& ./mole'sec) for the 3Bi quenching by the simple olefins
correlate well with the polarizability of the olefinic hydrocarbon double bonds^0:
C2H4, h.2. x 1010; CsH6, 8.5 x 101Q; cis-Z-Cfa, 13 x 101O; CF2CFH, 1-1 x 1010;
CClaCClH, 2.3 x 101"; cyclopen|ene, 11 x 10*°. This correlation is similar to
that observed for the 0(SP) reactions with the alkenes°3 where a charge transfer
intermediate has been suggested and the ultimate addition of the 0 atom to the
double bond of the alkene does occur. The preexponential factors for the quench-
ing rate constants for the simple olefins are all about the same and very near
the collision number; logioA (&./mole*sec): QalU, 11-36; CF2CPH, 11.22: Cgle>
11.65; cis-2-butene, 11.66; cyclopentene, ll.W*i . Presumably this reflects a
relatively unhindered SOa(3Bi)-jt-bond interaction which is common to all of these
alkenes. The rate constant differences observed with structural changes of the
olefins originate largely in differences in the activation energy (kcal/mole) for
the quenching rate constants: CgK^, 1:1; CFaCHF, 1.7; C3H6, 0.9; £is_-2-C4H8, 0-7;
cyclopentene, 0.14-. This is .just the trend, expected if the activation energy
barrier in these reactions is inversely related to the polarizability of the it
bond. The observed differences in activation energies do not seem to be in line
with those expected for the occurrence of direct triplet energy transfer and
reasonable triplet energy differences between the structurally different species.
?0
In our opinion the chemical mechanism outlined by Sidebottom and coworkers
seems most consistent with all the results observed here and previously for this
system. That is, the quenching may proceed through a highly polarized charge
transfer-like intermediate formed between S02(3Bi) and the n system of the alkene.
Presumably this complex may collapse into a transitory addition product which
allows rotation of the groups about the central C - C bond with the ultimate dis-
sociation of the complex into the alkene isomers and S02. This is very similar
to the mechanism favored by Caldwell and James°° for the benzophenone triplet
sensitized isomerization of cis- and trans-2-butene in benzene solution where
the observed deuterium isotope effect on the quenching rate constants is very
small, and addition to the butene double bond is evident from the formation of
some oxetane product. It is also analogous to the mechanism of Kochevar and
Wagner who studied, the olefin quenching of the Norrish type II photoelimination
in butyrophenone triplet in benzene solution"?. Both groups of workers conclude
that triplet energy transfer is not the major quenching reaction in the simple
olefin-RCOCsHs-triplet systems, but suggest that some type of charge transfer
complex is involved in the rate-determining step of the quenching process.
The effect of added gases on the quantum yield of isomerization gives some
insight into the timing of the isomerization reactions. Large quantities of He
gas were added in runs 19-21 and small quantities of a commonly employed triplet
quencher gas 02 were added in runs 22-25 of Table XVI. Presumably added gases
could, affect the quantum yield in at least three critical points in the reaction
52
-------
mechanism: (a) The S02(3Bi) can be quenched by the added gas in reaction (6):
(6) SO»(»Bi) + He (or O2) -> (SOj) + He. (or O2)
(b) The internal energy of the S02-2-butene triplet complex may be altered by
the collisional relaxation, and some product other than the olefin isomers might
be stabilized, (c) Oxygen may trap the S02-2-butene triplet complex and form
some stable oxygenated. SOg-olefin containing species such as
The quantum yield data show that the main influence of' the added, gases is in
reaction (6). Assuming that the only effect on the rate of isomerization will
be that due to the deactivation of the triplets in reaction (6) and taking the
experimental estimates of k6 = 6.8 x 107 (He) and 9.6 "x. 107 (o^) ^./mole'sec1",
the average of our best estimates of k3a = l.jk x 1011 A./mole*sec and k5b/(ksa
ksb) = 0.65, we may calculate the theoretically expected values of §^ from the
relation (c) where M is He or 0%:
,\/kt 4- ka\ k^. -4- £0,
, _^- - • ^w •«-•- *J \" 6& I rw*M/ I L""~J"Wlf **5fl I " OO I I ™ ' **OD
*^, = \ [m-2-C4H8]A3a A *6b / *6b
The experimental quantum yields for runs 19-21 with added He (0.35, 0.33, and
0.33, respectively) check reasonably well with the theoretical values (0.4l, 0.34,
and 0.27, respectively). Again in runs 22-25 with added 02 the experimental
values (0.24, 0.28, 0.23, and 0.22 respectively) are in accord with the theoreti-
cal values (0.24. 0.24, 0.24, and 0.24, respectively). The lifetime of the pre-
cursor S02-2-butene triplet adduct must be very short lived in that there is no
evidence of its interaction with these added gases.
It is possible that the reactions (3a) and (3b) are rate determining, yet
(5a) and (5b) proceed through the intermediacy of the stable twisted 2-butene
triplet state. Conceivably as rotation about the central C—C bond occurs in
the addition complex, and the original planar configuration of the carbon skeleton
approaches more nearly that of the stable twisted triplet of 2-butene, the po-
tential energy surface of the complex will cross that leading to the triplet
olefin and ground state S02 as well as the surface leading to two ground state
molecules in reaction (5a) and (5b). Such a reaction path avoids the endothermic
route of vertical triplet energy transfer. One might designate this potential
reaction as a "chemically induced triplet energy transfer". The ultimate in-
volvement of the triplet alkene here seems unlikely in that it would appear to
favor a near equal distribution of the isomers as is seen in the benzene-triplet
sensitized cis-trans isomerization of the 2-butenes, both in solution? --75 anc|
in the gaseous state7">^1. jn these experiments with the benzene triplet,
probably a true triplet energy transfer to the olefin does occur.
The indefinite product designation (S02) and (2S02) in reactions (ib) and
(2b) specifies some undetermined state or states of S02 such as S02 (X, 1A1),
53
-------
S02(3A2), and S02(3B2)• The latter two excited states could be involved in .the
isomerization of the 2-butenes as well as the initially excited species S02(3Bi).
This possible complication must be considered, seriously in view of the mechanistie'f
suggestions of Heicklen and coworkers^5^' ° . In-addition to these possible
reactive products of reaction (Ib) and (2b), S0(3£~) and 863 formed in reaction
(2a) could conceivably .react with olefin to induce isomerization. Our recent
results show that 10$ of the quenching collisions of S02(3Bi) with S02 lead to
the chemical path (2a) at 25° C?1. Perturbations in the S02(3Bi) «- S02 (X^Ai)
spectrum seen by Brand and coworkers° suggest that the potential energy sur-
faces for the 3Ba and 3A2 states cross that of the 3Bi state somewhere near the
(1,1,0) level of the 3Bi state. SCFMO and INDO-CI calculations on S02 fils'o are
consistent with this view^3j5\ one cannot rule out entirely some contribution
to the isomerization from the reaction of these intermediates and products with
the olefin, but it can be seen from the present results that such reactions can
provide no more than a minor contribution; the increase of the values of l/$
with the increase in the [SOg]/[2-butene] ratio seen in Figure 2 is only con-
sistent with the formation of those products of reactions (2a) and (2b) which
are relatively inactive chemically in producing 2-butene isomerization. A
further piece of experimental evidence supporting this view comes from our ex-
periments with added gases. He and 02 are thought to quench a large fraction
of the S02(3Bi) molecules to the 3A2 or 3Bs states because of the exact match of
the experimental activation energy of reaction (6) with the difference between
the (0,0,0) energy of the S02(3Bi) molecule and that of the (1,1,0) region of
this molecule where a potential curve crossing appears to occur*'. However, in
this work we have seen that the effect of added gas is to quench the isomerization
of the 2-butene by a fraction which is, within the error, limits, just that fraction
of the 3Bi molecules which have been quenched in (6). If the 3B2 or 3A2 molecules
are formed in reaction (6) as seems likely to us^i, then these species are not as
chemically reactive toward the olefinic hydrocarbons as is the S02(3Bi) species,
and they play little or no part in the isomerization reaction.
The evidence considered above also excludes SO(3E~) and 80s as important
reactants in the isomerization of the 2-butenes. However, it is possible that
some of the scatter in the quantum yield, data comes from a small irrepreducible
heterogeneous alkene isomerization reaction that is catalyzed by 1^804 formed
from the S03 product of reaction (2a). Aerosol formation is observed in this
system, and its nature and possible mechanistic significance -are considered in
the following section.
Mechanism of Aerosol Formation _in 3500-U100 A Irradiated S02-2-Butene Mixtures
From a study of the data presented in Table XVII some further clues can be
had as to the chemistry, reaction mechanisms, and origin of the light scattering
observed in the S02-2-butene mixtures irradiated within the forbidden band.
There are three independent qualitative measures of the initial rate of aerosol
formation in these experiments: (a) the rate of increase of 6328-A scattered
light observed early in the experiment, column 1 of Table XVII; (b) the time of
first detection of 6328-A light scatter, column 2; and (c) the rate of decrease
of the transmitted UlOO-3500-A light as monitored at the rear of the cell early
in the experiment. In general it can be seen that the two direct measures of
light scatter (a) and (c) increase with increase in S02 pressure, and the time
at which aerosol light scattering is first detectable is shortest in runs with
the highest S02 pressures. Furthermore, note that pure S02 with no added olefin.
run 1 of Table XVII, shows significant light scatter, although runs with both
olefin and SOa show a somewhat enhanced scattering. In columns U and 5 of
-------
Table XVII are listed the calculated relative rates of -quenching of SC>2(3Bi) '
molecules by SOa and 2-butene, respectively. We" have made the reasonable as-
sumption in this calculation that the incident light intensity was essentially
constant over the 26 runs listed. The ratios of the rates in column 1 or
column 3 to the quenching rates in columns h and 5 are shown in columns 6
through 9 of Table XVII. The degrees of constancy of the ratios with varied
[SOa] and [2-butene] can be compared. The best correlation with the observed
rate of aerosol formation is with the rate of SOa(3Bi) quenching by SOa- This
quenching reaction results in S03 formation by reaction (2a) about !($> of the
time91. Presumably S03 will lead to HaS04 aerosol by reaction with traces of
water in the cell. Once HaS04 aerosol is present in the cell, olefin 'molecules
may in theory participate-in the growth mechanism by polymerizing in the aerosol
droplet. The fraction of the total olefin removed by this path is very small
however, since the maximum quantum yield of olefin removal is equal to or less
than 0.016 ± O.OOJ, as estimated in the runs with the longest exposure time (see
Fig. l). This is in reasonable accord, with the results of Cox"' which were
obtained in SOa-2-butene mixtures irradiated within the first allowed, band of
SOa (near 3H5 A); he reported a quantum yield of 2-butene loss = O.OOJ, some-
what lower than the reported yields of "sulfinic acid" of 0.033 observed by
Dainton and Ivin5°\ cox observed a small quantum yield, of 1-butene formation
as well (O.OOOli)-) in his SOa^Bi) studies. No detectable isomerization to
1-butene was observed in this work, although such a product would not be un-
expected in runs at long times since the heterogeneous, H2S04 catalyzed iso-
merization of the butene-2 could become important for these conditions.
Since the formation of S03 is insured in all of these irradiated SOa-olefin
mixtures, the ultimate formation of sulfuric acid aerosol is expected; the com-
plete elimination of water from the glass wall of the photolysis cell is impos-
sible with the techniques which we and the previous investigators have employed.
Thus one suspects that the aerosol formation in irradiated SOa-olefin mixtures
may be initiated largely from excited. SOa reactions with SQ2. A reasonable
extrapolation of these results shows that aerosol formation observed in dilute
NOjjSOa-olefinic hydrocarbon mixtures in irradiated air probably does not origi-
nate either from S02(3Bi)-S02 or SOa(3Bi)-olefin interactions. It is likely
that in this case aerosol forms following very different reactions which involve
free radical attack on the SOa molecule; reactions such as the following have
been suggested as the primary steps in SOa removal" •» 93. These and other
alternative reactions are now under extensive study in many laboratories in-
cluding our own:
HO2 + SO2 -» SO3 + HO
HO2 + SO2 + M -» HOjSOs + M
HO + SOs + M -» HOSO2 + M
RO2 + SO2 -» RO + SO,
RO2 + SO2 -* RO2SO2
Aerosol formation in the irradiated. SOa-olefin systems at high SOa pressures
is a serious source of error, particularly in runs at long exposure time. In the
present work the scatter of the incident ^00-hlDO-L exciting light lowered the
apparent transmitted light at the end of the cell by as much as 20$ after 60 min
of irradiation in runs with 600 torr of SOa- Some of the appreciable scatter in
the quantum yield data must reflect this source of error which could not be
entirely eliminated, even by extrapolation of the rate data to zero time. To
our knowledge other researchers apparently have either not recognized or not
55
-------
reported this problem. Those who study irradiated SOa systems at high S02
pressures should be warned of this unexpected source of error in any optical ab-
sorption or emission measurements which are made with these systems.
I-E. The Mechanism and Kinetics of the
Alkene-SOg Reactions Excited Within ,
the First Allowed Band of SQg (3130 A)94
The mechanism and rate constants for the interactions of photoexcited S02
with the alkenes are of special interest because of the potential of these
reactions to form condensation nuclei and sulfate containing aerosols. Previous
It is a reasonable extrapolation of existing data to conclude that the SQs(3Bi)
species is one of the reactants which leads to this change. Directly excited
SQaC^i) molecules are quenched at a rate which is equal to that of formation
of an addition compound which leads rapidly to isomerized alkene and SO^P^.
Furthermore S02(3B!) molecules are formed from excited singlet SOg through col-
lisional perturbation in a significant fraction of the excited singlet quenching
collisions with a variety of quenching partners^-33.
Penzhorn and Gusten"" have interpreted the results of their study of singlet
S02 excited in 2-butene-S02 mixtures in terms of S02(3Bi) as the sole reactant
leading to isomerization of the alkene. Cox?? aiso employed this rationale in
his treatment of results from a similar study. He concluded, however, that pos-
sibly some other excited states may be reactants as well. Indirect evidence has
been presented that the S02(1B1) and several other electronic states of S02,
possibly &BZ, 1A2, and 3A2, which may be generated, from the S02(IBi) species,
may be involved in the chemistry ofothe SCb molecules excited within the first
allowed absorption band, 3200-2500 A22,31,21,57 ,88-901 certainly the possible
participation of these additional species in the alkene isomerization reaction
must be considered among the other alternatives. The data at hand do not allow
an unambiguous designation of the state of the S02 reactant in this system.
Existing product studies offer somewhat divergent views as to the nature
and extent of the chemistry in irradiated S02-alkene mixtures. Dainton and Ivin5^
suggested that an unsaturated sulfinic acid formed with a quantum yield of about
O.Qhh in a mercury-arc irradiated mixture of S0a and 1-butene. Cox°T found the
quantum yield of olefin loss, attributed to sulfinic acid formation, to be 0.007
for SQg-cis-2-butene mixtures excited with a band of 3000-J200-A light. He also
reported a small quantum yield of 1-butene formation in this mixture; $i-c4H8 =
l.U x 10~4. Cehelnik andocoworkers5° found evidence of an 0-atom transfer from
S02 photoexcited at 3130 A in CF2CF2 mixtures where CF20 was an observed product;
$CF20 = °*°5 at"high pressures of C2F4. Irradiated solutions of S02 and alkenes'
gave polysulfones65, while cyclic sulfites apparently form in irradiated S02-
alkene-02-containing solutions0".
In this study the quantum yields of the products of the 3130-1 excited S02-
£is_-2-butene gaseous mixtures were determined in an attempt to further define
the nature of the reactive states of S02 involved in this system. At the time
that this work^was completed^ the reports of the somewhat similar studies of
Penzhorn and Gusten°° and Cox"''' appeared. In general the work reported here
supports the conclusions of the previous workers, but some interesting differences
-------
both in the results and the interpretation of these results exist. The kinetic
treatment of the data presented here and a reevaluation of the data of the
previous studies provide new and more definitive information about the detailed.
mechanism of the excited SOa-alkene reaction system.
I-E—1. Experimental Procedures and. Results
Reactions were carried out at 25 ± 3 C in a cylindrical quartz reaction
cell (diameter 3.5 cm, length 39.3 cm). The cell was attached in series with a
thermal gradient pump which was used to ensure uniform mixing of the reactants.
The vacuum system and pressure measuring-components were the same as those de-•
scribed previously^. Light from a medium pressure (PEK 200 W) mercury arc was
collimated by a lens and light stops and filtered through a series of solutions
and glass filters to isolate the 3130-A band°9. The incident light intensity
was measured at regular intervals using potassium ferrioxalate actinometry' .
The relative intensity was monitored continuously d.uring photolyses by reflectance
of a small fraction of the incident beam into a 935 phototube by means of an
angled quartz plate placed in the light beam. The lamp was allowed to warm up
for one hour before starting irradiations. The variation in incident intensity
during a run was found to be less than ±2$, as indicated by both actinometry and
reflectance monitoring. The fraction of the incident light which was absorbed
was measured with a 935 phototube mounted at the rear of the cell. Within the
range of SOa pressures of 2-11 torr employed in this study, the absorption fol-
lowed Beer's law well with the decadic extinction coefficient for SOa, 6 =29-7
H ./mole* cm.
Reactant purification and. chromatographic product analyses were made as
described in the earlier SOa(3Bi)-alkene study°^. Initial quantum yields of
the SOa-sensitized isomerization of cis-2-butene were determined by extrapolation
to rates at zero time using the least squares fit of the cis- and trans-2-butene
composition versus time curves. These data from experiments at 3130 A and 25 ±
3° C with varied, initial SOa and cis-2-butene pressures, are summarized in Table
XIX. COa gas was added to the SOa-alkene mixtures in experiments 16-18.
No detectable aerosol was observed by light scattering in these studies at
the relatively low pressures of SOa employed here. The quantum yield of 2-butene
disappearance was below the detection limits in these runs of short duration, and
the only chemical change which occurred measurably was the eis -» trans isomeri-
zation of the 2-butene.
I-E —2. Discussion
The Simple Two-State Reaction Mechanism of_ cis-2-butene Isomerization Sensitized
by_ Electronically Excited
The present kinetic data provide some new insight into the detailed mechanism
of the electronically excited SOa-sensitized isomerization of cis-2-butene. In
the consideration of our findings we will draw also on the results of Penzhorn
and Gusten"0 and Cox°7 which appeared soon after this work was completed"1-1. Both
previously published studies support the view that S02(3Bi) is the major reactive
SOa state in this system. Penzhorn and Gusten derived rate constant estimates
for the SOa(3Bi)-quenching reactions with NO and Oa through kinetic studies of
the inhibition of the rate of the excited SOa-sensitized cis- to trans-2-butene
isomerization by added NO and. Oa- Quenching rate constants for SOa(3Bi) with
57
-------
Table XIX. Initial'quantum yields of the SOa-photosensitized isomerization of
cisr2-butene at 51JO A.
"Run
.-- *>•
\
i
2
3
14
5
6
-7
8
-• j^j" "
ID
"11
12
13
Jfc
3J5-
16C
17 c
18=
Rreaaure :
2.58
2.61
2.62
2.6}
5.H.
5.01
5.06
, 5-09
- .5.09 :
5-16
9-97
9-81
9-86
9.85
9>99
2.67
5-17
10.02
reartaot. tbrr
0.17}
0.085
o.o»a
0.026
1.081.
0.335
O.lfl.
0.082
0.-051
0,033
2.100
0.317
0.159
.0.098
0.065 '
0-259
0.503
0.97}
tao,j/[c»n,]
H..9
30.9
62.0
100.6
I..7
Ik. 9
30.9
62.0
KB. 6
151".}
. • 1* .7 '
30.9
62.0
100.6
15"..}
10.3
1C.}
10.3
lot quanta/aec-
cell flee, x KT"
7.15
7.15
T.W-
• 7.JO
7.03
7.5*
7.15
7.<>,8
7«B-
7.JO
70>3
7.15
7.1*8
7.30
7-30
7.01.
7.01.
7.0»
Initial quantua yield of tno£-2-butene
foration
Ezperlnnital ' - Calculated
Method- 1" Method 2b
0.22
0.18
O.ll.
0.11
0.21.
0.2}
.0.19
O.ld
6.U
0.096
0.31
0.20
0.35
0.11
0.102
O.ll.
0.20
0.26
0.062
0.060
0.055
0.050
0.06k
0.062
0.060
0.055
0.050
O.OMt
0.06k
0.060
0.055
0.050
O.OW.
0.026
0.022
0.022
0.17
0.15
0.1}
0.11
0.25
0.17
0.15
0.13
0.11
0.096
0.25
0.15
0.13
0.11
0.096
o.u
0.20
0.27
Calculated from relation (A) assuming tu/(*W + *») = **• + *») = 0.10, Ar4» -f
= 0.052 as derived from* tiacetyl-SOj studies at shorter wavelengths together with
other rate constant estimates specified in the text.
'Calculated from relation (A) for runs 1-15 and relation (G) for runs 16-18, using
present estimates of *,./(*i» + *«,) = 0.85, *4»/(*4. + *«>) = °-14' *»»/(*u + *«>) =
0.116, *n/(/tio» + fciob) = 0.026 mole//, together with the other rate constants specified
in text.
«COt added in runs 16, 17, and 18 was 105.4, 308.0, and 609.7 torr, respectively.
NO and Q2 relative to that for SOa were derived. Coupled with our previous direct
rate constant estimates of S02(3Bi) quenching with SOa, they estimate the NO and
02 quenching constants as (7.8 ± 1.9) x 1010 and (2.U ± 0.5) x 10s ^./mole'sec,
respectively. These values are in reasonable accord with the direct measured rate
constant estimates from our SOa^Bi) lifetime studies^0, (J.k ± 0.3) x 1010 and
(0.96 ± 0.05) x 108 £./mole*sec for NO and 02) respectively as quencher*.
In the case of Oa as the S02(3Bi) quencher, the agreement would be improved sig-
nificantly if proper account were taken of the difference in triplet SOa forming
facility on collision of singlet-excited S02 with SOa and Q2. It was assumed
by Penzhorn and Glisten that the intersystem crossing ratio for excited singlet
quenching collisions with Oa was equal to that for collisions with S02. Our
recent work suggests that the fraction of the excited SOa singlet quenching
collisions with 02, which result in S02(3Bi) formation, is only about one third
of that with SOa as quenchei"
Coxb' assumed also that S02(3Bi) is the major reactant leading to alkene
isomerization in 3000-3200-A irradiated S02-cis-2-butene mixtures, and he de-
rived estimates of the S02(3Bi ) -quenching rate constants with the 2-butenes
relative to that for SOa- Using our published rate constant for S02(3Bi) quench-
ing with S02, he obtained (1.62 ± 0.08) x 1011 and (1.1*2 ± 0.09) x 1011 ^./mole-
sec for the S02(3Bi) quenching constant with cis-2-butene and trans-2-butene,
respectively. Again these indirect estimates check well with those found from
our direct SOa^B) lifetime studies within the experimental error of the de-
terminations. (1.29 ~ 0.18) x 1011 and (1.22 ± 0.15) x 1011 ^./mole'sec for the
58
-------
cis- and trans-2-butene, respectively"^. Thus the results of these previous
studies and our recent direct study of the SQ2t3Bi)-sensitized isomerization of
the cis- and trans-2-butehes"^ suggest that the phosphorescent S02'(3Bx) state of
SOp. may be an important, or perhaps the exclusive, reactive SOfe state in the
3130-A excited SOg-sensitized 2-butene isomerization. Then it is reasonable to
attempt first a fit of the present data to this simplest possible mechanism
choice, j"
'Recent spectroscopic analyses of the first allowed absorption band in SOp.95 and
S02 fluorescence lifetime determinations-?0 suggest-'that structural features in
the 33^0-3^00-1 region belong to the -""A^ «- X, xAi transition which derives in-
tensity from the allowed ^-BI *- X, xAi transition by way of b2(>vibronic coupling.
Presumably the -"-fii *- X, 1Ai transdsfcian origin lies near 3200 A. We have assumed
that the long-lived component of the SOa fluorescence for which we have de-
termined, rate constants previously, is the xBi species. A more definitive,
and detailed description of these processes is not possible at this time.
(I) SO, +/»<3130 A-* SO»('Bi)
0) . jsOiOft) -> so, + b>,
(2a) 80,050 + SO, -» SO,(3B,) + SO,
(2b) S02(»fi,) + SO, -^ (2SO,)
(3a) SO2(1fi,) + w-2-C4H8 -» SO^'fl,) + c«5-2-C4H8
(3b) SOjCBi) + w-2-C4H8 -> (SO,) + m-2-C4H8
(4a) S02(1B!) + C02 -» SOsC'BO + CO,
: (4b) SOzCB.) + CO, -» (SO,) + CO,
(5a) SOi('Bi) -» SO, + hvp
(5b) S02(3Bi) -* (SO,)
(6a) SO,(3BO -f SO2 -» SO, + SO(8Zr)
(6b) SO,('B,) 4- SO, -» (2SO,)
(7) SO2(3B,) + m-2-C4H8 -> (SO2—C4H8)8
(8a) (SO2—C4H8)3 -» m-2-C«H8 + SO,
(8b) (SO2—C4H8)3 -> trans-2-CtHt + SO,
(9) SO,(»B,) + C02 -> (S0») + C02
The indefinite product designation (SOa) and (2S02) shown in the above sequence
signifies a ground state or some excited nonemissive state of SOg- (SOp-C^s)3
represents the transient triplet adduct suggested in our previous work^. The
concentration conditions employed in the present work ensure that the fraction
of the excited S02 molecules which react in the first-order decay paths is
negligibly small; 5.5 •/. 10~4 in reaction (l) and 2.Q y. W~2 ir. reactions (5a)
59
-------
and (5b). Thus if this simple mechanism is operative then relation (A) should
describe the present results from the cis-2-butene-SOg-CQa mixture phdtolyses.
L \ t
""ob V I --*i»|,— — aj .
_l_ L f J L ' *- \r^M«»i 1~
/.\ w. v*oa i~ **8b/ V*8
(A) *c-t = 7
(A
u
.. (. , (^6a + Agb)tSD^] *9[COS
x{1+ *7[c«H8] +;
In Table XIX it. can be seen that the experimental quantum yield data from
S02-cis-2-butene binary mixtures are in accord with one of the expectations of
relations (A). For a given [S023/[c4Hel ratio the quantum yield should be inde-
pendent of the total S02 pressure within the experimental error. This is seen
to be the case; for [SOal/tc^] = 15^, *) = 0.10 as have the previous workers°°>67. With
this assumption and the other rate constant estimates given, we may calculate
from relation (A) the theoretically expected initial values of §cr»t f°r various
reactant concentrations employed, in the experiments summarized in Table XIX.
60
-------
These estimates are given in the column headed "Calculated, Method 1." If you
compare 'these values with those found experimentally, it can be seen that the
calculated values are much less than the experimental values. In experiments
with binary mixtures of SQg and butene they are roughly a factor of 2 or 3 less.
The disagreement is even greater in runs with added CO& where the calculated
quantum yields are lower than the measured values by a factor of about ten
(runs 17, 18, of Table XIX. If the simple mechanism were operative and kaa/
(k2a + kat) = ksa/(ksa +.k3b)> tnen in runs with mixtures containing only S02
and butene relation (A) simplifies to
Thus for these assumed conditions we expect l/$c-»t to ^e a linear function of
the LSOal/CC.^] ratio. The dashed line labeled "simple theory" in Figure 18
shows the theoretical dependence expected from Relation (B). Compare this in
Figure 18 with the results obtained by Cox°f and those observed in this work.
The Cox data for ISQal/I^Hs] ratios above 50 fit reasonably well with the ex-
pectations of simple theory and relation (B), but the l/$c-*t values from this
work are consistently lower than those of Cox and simple theory by a factor of
about 2 within the range of [S02]/[C4Hs] values which is common to the two ex-
perimental studies. The reasons for the difference is not clear, but obviously
the absolute values for $c-»t from both sets of data cannot be correct. It
appears to us that the absolute values of $ c-,^ which we report in this work may
be more accurate than those of Cox. His estimation of the fractions of the
broad wavelength band (half intensity width of 3000-3200 A) of the incident light
which were absorbed by the S02 and the actinometer solution was seemingly much
more involved and presumably subject to greater error than the same estimates in
our experiments. We employed a nearly monochromatic beam of light, and a Beer's
law dependence of absorption on- pressure of SOa was observed. However, Cox kept
the S02 pressure essentially constant at about 30 torr in all of his experiments.
So if there were an error in measurement of the intensity of the absorbed light.
his results would give experimental $ c->t values which would only differ from
the true $c-»t values by a constant multiplying factor. Furthermore the values
of the slope-to- intercept ratio of the linear portion of the l/$c-»t versus [S02]/
[ C4Hel ratios should be independent of any constant factor of error, and should
provide reliable values for (kea + k6t>)/k7. indeed the Cox data give (2.UO ± 0.09)
x 10~3 which does agree reasonably well with that measured directly from our life-
time studies, (2.91 ± 0.23) x 10~3 26,51,42,64. ^e application of the potential
absolute error hypothesis to our data suggests that the slope-to- intercept ratio
from our data should also give a reliable estimate of (ksa + kgtJ/ky. We may use
the data for the highest [S02]/rc4H8] ratio to test this hypothesis, since for
ratios equal to or greater than 100, the quenching of singlet SOg by S02 is
ensured, and relation (B) should apply reasonably well regardless of the magni-
tude of the unknown rate ratio k3a/(k3a + k3t>). Our present data give (kea +
k-sb)/k7 = (2-7 ± 1-2) x 10~3, a reasonable check with the indirect value from the
Cox data and our previously determined direct estimate of this ratio.
The magnitude of the intercept observed from the extrapolation of the I/
§£_»-(; versus LS02]/[C4H8] plot in Figure 18 for the linear high ratio range gives
in theory an estimate of the rate ratio, [(k2a + k2b)/k2a]C (ksa + k8b)/ksbJ.
Using our estimate of k8b/(ksa + kst>) = 0.65 ± 0.05 and the intercept of the Cox
data, we find k2a/(k2a + k2ti) = 0.10 ± 0.01, in almost exact agreement with our
61
-------
estimates of this quantity from the S02-triplet sensitized biacetyl phosphores-
cence data for 2875 ^ (0-095 ± 0.005). The intercept of our present data, how-
ever, gives k2a/(k2a + ^b) = °-21 ± °-04. The factor of 2 difference in the :,,£.%. •
absolute quantum yields of* our work and that of Cox is reflected in this observed
difference. There has been no direct measurement of k2a/(kea + k2b) by the
biacetyl method for experiments at JIJO A. Previous estimates of kga/Cksa + k2b)
are 0.082 ± 0.003 at 2650 P8; O.p95 ± 0.005 at 2875 A24; ~ 0.10 at 3020 A38.
It is not clear whether the high value for this constant estimated here represents
a real wavelength effect or whether it is an artifact related to the uncertainties
in the present data and the detailed mechanisms involved in the sensitized ciSr-2-
butene isomerization and the sensitized biacetyl phosphorescence methods of
estimation of triplet quantum yields. Certainly a new S02-biacetyl mixture study
at 3130 A is highly desirable. From the present work we conclude tentatively
that the intersystem crossing ratio kea/(k2a + ksb) = 0.21 ±0.04 for S02 excited
at 3130 1. All of the published data from the high CS02]/[C4HQ] ratios, which
guarantee singlet quenching by SOg, appear to be consistent with the simple
mechanism and relation (B).
The Isomerization Mechanism ^n Experiments a_t Low [SOjg/Cc^Ha] Ratios
The curvature in the l/S^t versus [SC^l/Cc^s] ratios seen in Figure 18
for both the Cox data and .those determined here, points to an increasing ef-
ficiency in the cis -» trans isomerization as the extent of S02-singlet quenching
by the alkene increases. This could arise from an enhanced singlet-to-triplet
intersystem crossing ratio for excited singlet S02-butene collisions, that is,
kaa/C^sa + k3b) may be greater than k2a/(k2a + k2t>). Indeed Cox has shown tha,t
his data for the low [S02]/[cis-2-C4H8] ratios may be interpreted in this 'fashion.
We may treat our present data using this more realistic mechanism alternative
which allows k3a/(k3a + ksb) to be an adjustable variable which we may evaluate.
Relation (A) may be rearranged, substitution of the known rate constant ratios
made, and the consideration restricted to the COa-free systems. Relation (C)
is then generated:
/ [SOj] \ / [SOj] \
(C) *c_,(l+2.91.X 1^87hri)(°-55|FHl+1)
\ IV4H8J/ \ lljmgj /
a ~t~ fab [C^Hs] Aja 4~ *3b/
.*s. + k
The left-hand side of relation (C) has been calculated from our present data and
this function is plotted versus [SC^j/Cc^s] in Figure 19- A reasonably good
linear relation is observed between the variables. Accepting the modified simple
mechanism, the slope and intercept of this plot may be used to derive rate con-
stant estimates. The most accurate estimate of the slope, derived from all the
data, gives [k8b/(k8a + ksb)l[k2a/(k3a + k3b)] = 0.0728 ± 0.0042. The best
estimate of the intercept, derived from the data from low [S02]/[C4H8] ratios
(< 30.9), gives [k8b/(kaa + k8b)][k3a/(k3a + ksb)] - 0.55 ± 0.24. The slope-to-
intercept ratio gives k2a/k3a = 0.13 ± 0.06. Taking k2a = 8.3 x 109 A./mole*
sec9T,99,28,38,24j we estimate kga = (6.3 ± 2.9) x 101Q Z./wle-sec. From the
intercept and our previous estimate of kab/(kaa + ksb) = 0.65 ± 0.05 , we find
^sa/(^3a + ksb) = 0.85 ± 0.38. Combining this with our present estimate of k3a,
we get ksa + ksb = (7.4 ± 4-7) x 1010 ^./mole*sec. This should beocompared with
our directly measured value from singlet lifetime studies at 2662 A, k3a + k3b =
(6.9 ± 0.4) x 1010 £./mole*sec97.
62
-------
40
30
ft5
i 20
10
- This Work
200. 400
605 ' §00
. Figure! Bjlot of the reciprocal of the initial quantum yield of irmu-2-butene vs.
, [SOi]/[C4H»] ratio. Photolysis of SOr-cu-2-butene mixtures from data of Cox67
i with broadband excitation of SOi (3200-3000 A) and from this work at
3130 A; upper dashed curve labeled "simple theory" is calculated from relation
(B) using our published rate data from very different experiments.
21"
: 10-
6
50
100
ISO
Figure 19Plot of function (C) vs. [SO2]/[C4H8]. Data from 3130-A photolysis
of SOi-cij-2-butene mixtures derived in this work.
The Cox data"? from low [S02]/[C4He] ratios may be retreated to obtain an
independent estimate of these rate constants. In his original treatment Cox did
not have access to our experimental values for (k2a + k2b)/(ksa + k3b) = 0.55,
and he assumed that this ratio was unity. If we employ only his data for [S02]/
[C4H8] ratios in the range 0.96 to lU.5, the fraction of triplet S02 quenched by
the alkene is near unity (0-997 to 0.958), and the first term multiplying $£-»£
in relation (c) is essentially unity for these conditions. The retreated Cox
data are shown in Figure 20. In this case the slope is O.OUf ± 0.005 and the
intercept is 0.2U ± O.OJ. In terms of the mechanism given, the slope-to-intercept
ratio gives k2a/k3£ = 0.18 ± 0.03- Within the large error inherent with such
slope-to-intercept methods, this value is in fair accord with the present esti-
mate of this ratio derived from our data (0.13 ± 0.06). The intercept and our
previous estimate of k8b/(kea + k8b) give k8a/(ksa + ks^) « 0.37 ± 0.05. The
difference between this value and that estimated from our data reflects again the
63
-------
1.0
O 04
Figure ad.Ptot of function (C) vs. [SOi]/[C«H8]. Data from 3000-3200-A
photolysis of SO i-m-2-butene mixtures by "
same factor of 2 difference between o.ur absolute quantum yields and those of Cox.
However, in every case, the rate constant ratios derived from slope-to-intercept
ratios, which should be independent of any constant percentage error in the data,
check well within the experimental error of the measurements, and the present
results are in most details consistent with those of Cox. Both these sets of
date suggest strongly that there is a high efficiency of triplet formation, or
at least isomerization, which results from singlet-S02 quenching collision with
cis-2-butene.
The results of many previous studies for 2660-2&T5-A excitation of SOs have
shown that the fraction of singlet quenching collisions, which form the
species, varies somewhat with quenching partner, S02, 0.082-0.095 , Na-f
O.OJU1", 0.01931; 02, 0.030^; Ar, 0.025"5; CO, 0.01T22, 0.01951, O.O^; C02,
0.0521° O0.02o3!; cyclohexane, 0.07325. NO studies have been made with excitation
at 3130 A. However, from all the data at hand none of the singlet quenchers
studied has an efficiency of S02 triplet formation which approaches that sug-
gested here for cis-2-butene. The reason for the unusual behavior of the alkene
may lie in the unique nature of the singlet -» triplet conversion process in this
case. The spin inversion which is perturbed by collision of excited singlets
with a chemically unreactive collision partner presumably results in a certain
fraction of the quenching collisions and is followed by triplet separation from
the perturbing molecule. Interaction of this triplet then occurs with one of a
variety of competing molecules in a subsequent collision. Our previous work
has shown that the interaction between SOa(sBi) and alkene molecules is strong
with a rate constant for triplet quenching and subsequent alkene isomerization
which is near the collision number°^. If a perturbation occurred creating a
triplet SOa during an S02-excited singlet-alkene collision, then it seems to us
that it is unlikely that the two molecules could separate before reaction of the
triplet and alkene would occur, and efficient isomerization of the olefin would
result. The abnormally high efficiency of isomerization then might result since
there will be no competing reactions which can remove the triplets in this case.
If this mechanism is operative, then we expect a somewhat modified form of
the rate law to hold for mixtures at low [S02]/[C4H8] ratios. Excited singlet
B02 quenching by cis-2-butene which involves entrapment of the triplet generated
on collision requires that reaction (3a) be reformulated as follows:
-------
(3a') SOtCfii) + m-2-C4H8 -» [SO8(3£i)-C4H8] -
(3b) -» (SO,) + cu-2-C4Hi
This mechanism choice leads to the theoretical rate law (D) for the initial quan-
tum yield of trans-2-butene formation:
1 /*8» + *8b\ Aaa- + *3b\ / (A:,. + *«.) [SO,] \
* ' .»- r1 I t IX ^ I \ (L _L i. \ t/~" u 1 I
.*c-.l — ii \ *8b / \ *3a' / \ V.*3a' T *3bJ Ivj4ri8j/
where
[SO,]
W IC4H8]
/(*«. + *»b) [SO,] \ / (*«»,+jUb) JSO*1\
\(*,.' + *,b) [C*H8] A *7 '" ICaUl/
From our present data one may calculate (l/S^^ - E)*-and test its dependence on
We have arbitrarily chosen our previous estimate of k8b/(k8a + kg^) = 0.65 de-
rived from S02(8B!)-cis-2-butene and S02(3Ex)-trans-2-butene rate data; the value
applicable to the species (S02-C4Hs)3 formed in reaction (3a ) may be nearer to
0.50 as observed in the triplet benzene interaction with cis-2-butene7^~7f, since
the total energy of the excited singlet S02 (91 kcal/mole) is available to the
S02-C4H8 triplet complex formed in reaction (3a7) rather than 73-6 kcal/mole
which is characteristic of the S02(3Bi) vibrationally relaxed molecule.
[SOg]/[C4Hs]» these data are plotted in Figure 21. The functional form of re-
lation (D) is followed quite well over the range of data for which the $c-»t - E
difference is of meaningful accuracy. However, the intercept determination is
highly inaccurate since few data at low [ SOa"!/[ C4He] ratios were determined here.
The ratio of slope to intercept, equal in theory to (k^a + k2b)/(k3a' + k3b) gives
O.l8 ± 0.25. We may treat our present data in an alternative fashion to test this
singlet quenching mechanism if we accept the previously measured value of (kea +
kab)/(k:3a + ^sb) = 0.55 together with the other rate constant ratios derived in
the present work. Then we may calculate the values of k3a'/(k3a/ + k3b) from re-
lation (D). This procedure gives k3a'/(^3a' + k3b) values in the range of 0.8l
to 2.U. Although our present results are not inconsistent with the mechanism
(3a') of singlet quenching by cis-2-butene, they provide no real definitive test
of it.
The data from Cox cover a wider range of [S02]/[C4He] ratios and are better
suited to test this relation (D). We may evaluate E using for consistency the
estimate k2a/(ksa' + k3b) = 0.055 which we derived from his data, and our pre-
vious estimate of k8b/(ksa + keb) = 0.65, (kaa + k2b)/(k3a' + k3b) = 0.55, and
(ksa + ksb)/k7 = 2.91 x 10"3. A plot of l/($c-+t - E) versus [S02]/[C4H8] is
shown in Figure 22, determined with data up to 6.^5 = [S02]/[C4H8], where
reasonable accuracy in the difference $^^ - E is maintained. The ratio of slope
(2.3D± 0.38) to intercept (3.7 ± l.U) gives 0-57 ± 0.1^; this estimate should be
independent of any constant fractional error in quantum yield values. The value
is in good accord with the ratio of the direct estimates of (k2a + k2b)/(k3a' +
k3b) made from S02 singlet lifetime studies, 0.55 ± 0.0^97-99. Thus the limited
65
-------
40r
Figures!'. Plot of function (D) vs. fSdj]/[C4H(] ratio. ' Data "from! 3130-A ,
photolysis of SOj-c"-2-butene mixtures dl-rived in t nis work. '
2 46
[S02]/[ci«-2-C4H,]
Figure22.Plot of function (D) vs. [SO,]/[C4H8] ratio. Data from 3000-3200-A
photolysis of SOi-«>-2-butene mixtures by Cox
data now in hand do favor the modified interpretation of the singlet SQs-alkene
interaction given by reaction (3a' ) , although no definitive choice of alternative
reaction mechanisms (3a) and (ja') can be made at present. In any case we can
conclude that excited singlet gQ2-cis-2-butene collisions do result in the gener-
ation of S02(3Bi) and/or (S02-C4.H8)a complex a very large fraction of the time.
According to the mechanism outlined, this fraction varies from 0.85 ± 0.38 to
0.37 ± 0.05. With the alternative mechanism (ja') replacing (3a)5 Cox's data
suggest ksa'/(k3a' + ksb) = °-^2 * °-l6. It should be noted, however, that the
latter estimate may be low by a factor of 2.
In preliminary reports from Horowitz and Calvert-'1 and Fatta and coworkers" }
a high efficiency of triplet formation has been observed also with excited singlet
SOa-biacetyl collisions. About 17$ of the quenching collisions for SOg excited
at 2875 A create triplet SOg- In this case the very raPi£ triplet energy transfer
from triplet S02 to biacetyl (k = l.k x 1011 -£./mole*sec could occur following
triplet formation in perturbing collisions between excited singlet SQ2 an& biacetyl.
Another interesting mechanistic variation is possible to rationalize the high
efficiency of triplet generation from S02(1Bi) quenching reactions with both cis-2-
butene and biacetyl. Conceivably S02(3A2) and/or S02(SB2) molecules, as well as
S02(3Bi) molecules, may be created whenever quenching of the SOs^Bi) occurs by
66
-------
collision with any quencher molecule. If the lifetime of the SOa(3Aa) and/or SOg
(3Ba) states were very short as a result of efficient collisional quenching or
internal conversion processes, then in the usual dilute mixture of the triplet
quencher and reactant gases only the 3Bi species would be counted by the techni-
ques which we have employed in the previous SOa-biacetyl mixture studies. However,
when the quenching collisions and subsequent intersystem crossing of the singlet
to triplet is induced by the triplet trapping agent itself, then the total of the
S02(3Bi), S02(3A2), and SOa^Ba) maybe counted. Since nothing is now known about
the lifetimes of the S02(3A2) and S02(3B2) species, it is impossible to test this
alternative mechanism at this time.
The Isomerizatipn Mechanism in Experiments with High Pressures of -Added CO?
It appears to us that all of the available photochemical data for the binary
SOg-cis-2-butene system are rationalized well in terms of the well-known phos-
phorescent SOa(3Bi)" and fluorescent SQs(1Bi) states, and there is no compelling
reason to invoke the participation of the 3B2, 3A2, or XA2 states of SC>2 in the
kinetic scheme. However, in our study of the ternary mixtures of S02~cis-2-
butene-COs reported here, this simple mechanism is not adequate. It predicts
isomerization quantum yields which are about a factor of ten too low for experi-
ments at very high pressures of add.ed C02 gas. Compare the experimental values
for runs 16, 17, and 18 of Table XIX with those calculated from relation (A).
This result is entirely analogous. to that observed by Cehelnik and co workers •*•
and Wampler, Horowitz, and Calvert22 in photolyses of CO - SC-2 mixtures at high
CO pressures. We must invoke some new state or states of 50% to describe these
results quantitatively.
The excess quantum yield of isomerization observed at high COs pressures
may be considered in terms of several alternative mechanisms. Heicklen and co-
workers5°j57 ,88-90 have speculated on the role of SC-2 singlet and triplets other
than the optical states in the photochemistry of SQa mixtures with high pressures
of CO, C2F42°, and thiophene^T. Wampler, Horowitz, and Calvert22 '51 have pro-
posed an alternative reaction scheme to rationalize their study of CQg product
quantum yields in 3130-A photolyses of SOa - CO mixtures at high CO mixtures at
high CO pressures. They suggested that some undefined intermediate state (X)
was formed in the quenching SOa^Bi) in reactions (2b), (Ub), and (Ub). This
state was unreactive toward CO to form CO& directly, but it could generated SQ2
(3Bi) on collision with other species in experiments at the higher pressures.
According to the mechanism of Wampler and coworkers, the following additional
reactions should be important in our present experiments with high pressures of
added C02 gas:
(4b') SO«('B,) + CO, -» X + CO,
(10a) X + CO, -» S02('flj) + CO,
(10b) X + CO2 -» SO, + CO,
(11) X->S02
•The excited singlet SOa will be quenched largely by C02 in runs 16-18 of Table XIX
so the above reactions coupled with reactions (4a), (^b), (6a ) , (6b), (7), (3a),
(8b), and (9) should describe the major happendings related to alkene isomerization
in these experiments. The steady-state treatment of these reactions leads to
relation (F) for the initial quantum yield of trans-2-butene formation:
67
-------
. tab' :
taa 4- tab "I" tab'
(Ic v, \P k
ff8b \ «4a
taa + tab/ Ltaa + tab + tai
X I :—
,_, , \[COj](*jo» -h *iob) + kn
(F) 3>c^t = . ,,^.
.(*„.+ *•>) [SO,]
We may rearrange the terms of relation (F) to obtain the more useful form of
reaction (G):
(G) ; '- _ = r«'^^i ^jp^- ^ +
'. P-
where
H = (''Tb ;*•*' (1 + A*[c«H8]+ ***t Icj^j)
This function is plotted versus 1/[C02] in Figure 27). From the ratio of the
slope'(0.0159 ± 0.0017) to the intercept (1.52 ± 0.19) of this plot we obtain
kn/(k10a + kiob) = 0.012 ± 0.002 mole/£. This may be compared with the esti-
mate of the same rate constant ratio derived from the very different measurements
of the quantum yields of SOa-sensitized biacetyl phosphorescence with excitation
of SQz at 2&T5 A in SOa-biacetyl-C02 mixtures, 0.020 ± O.OOU51. The intercept
of the plot given in Figure 2J gives k4b'/(k4a + k4b + k-tb') = O.j6 ± 0.11, a
number considerably larger than the previous estimate of this ratio from experi-
ments at 2875 A, O.lU ± 0.0251. The data from the experiments with added COs
can be treated in an alternative fashion to find a fit to the mechanism outlined.
We may allow k4a/(k4a + k^b + k4t/) to be a variable as well as the slope and
intercept values of relation (G) and solve directly for these three unknowns.
This treatment gives k4a/(k4a + k4b + k4b') = 0.14 and k!i/(kloa + kiob) =
0.026 mole/4. The latter value is in reasonable accord with the previous esti-
mate, 0.020 ± O.OOU mole/A.51. The ratio k4a/(k4a + k4bo+ k4b') is somewhat
higher than the previous estimate of this ratio ato28?5 A (0.052lb). There have
been no other measurements of this value for JIJO-A excitation of S02 of which
we are aware. Considering the magnitude of the complications involved in the
two very different experiments used to determine the kn/(kloa + klob)3 the
degree of agreement observed is considered to be satisfactory. It can be seen
by extrapolation of the high-pressure rate constant data to the conditions
employed in runs 1-15 of Table XIX that the excess triplet mechanism will be of
no significance at the pressures of reactants employed in these runs.
There are, of course, other mechanism alternatives which one may choose to
treat the present data. Thus we may designate the species X as SOa^a) and
invoke S02(3A2), or conceivably SOfe(3Bfe), as the product of reaction (lOa) in a
fashion analogous to that of Cehelnik and. coworkers. However, if this is done
there are several significant changes which must be made in the original mechanism
68
-------
p
6
50
100
I / [COj] .
BO
Figures 3,PLot of function (G) vs. 1/[CO.] from 3130-A. photolytii of
SOj-m-2-butene-COi mixtures derived in this work.
choice of Cehelnik and coworkers. Thus we must require that the SOaAa) state
be unreactive toward isomerization of the alkene and the 802(^2)5 and/or the
S02(3B2) states must be near identical in their reactivity to that observed for
SOg(3Bi) in its quenching reactions with SQa, cis-2-butene, t ra ns-2-butene, NO,
Oa, and COa- At this stage in the development of our knowledge of these systems,
no completely unambiguous choice of mechanism can be made. However, all of the
available data from the butene-SC>2 mixture photolysesoin both the first allowed
band (2UOO-3200 A) and the forbidden band (3UOO-1000 A) seem to be rationalized
well in terms of the participation of the known phosphorescent triplet state
(3Bi), the long-lived excited.singlet state (1Bi), and at least one other chemi-
cally unreactive state of SC>2 which either regenerates S02(3Bi) in part by a
collisionally induced reaction at high added gas pressures, or which gives other
triplets (3B2, 3A2) of aljnost identical reactivity to that observed for SC>2(3Bi).
However, the introduction of the SC>2(3B2) and SC>2(3A2) species as potential re-
actants which promote isomerization of the alkene in irradiated SQ2-cis-2-butene
systems seems to us to be both unnecessary and speculative at this stage of our
knowledge of these systems.
I-F. The Mechanism and Kinetics of SQ-:
in the Photolysis of SO? Mixtures10
Tins
1
In principle the photochemistry of the pure SOa gas offers the simplest chemi-
cal system from which we may gain information on the chemical reactivity of the
different excited states of S02- However, the two most recent studies of S03
formation in pure S02 irradiated within the first absorption band gave widely
divergent results. Thus Cox^02 estimated that the quantum yield of S03 for-
mation in irradiated SO^ was (3.8 ± 1.0) x 10~3, only about one-twentieth of
the value observed by Okuda and coworkers-^: $gQ = (8 ± 2) x 10~2. In both
studies a broad band, of light which overlapped tSe first allowed singlet band
was employed. In rationalizing the difference between these results, Cox has
suggested that the very nonuniform light absorption which must have occurred
in the studies of Okuda and coworkers, carried out at 1 atm pressure of SQ2 gas,
may have favored a possible heterogeneous wall reaction forming SQ3. However
Cox's explanation of the difference between the results of the two studies is
difficult to accept, since in the study of Okuda and coworkers a fast flow system
69
-------
was employed with a very short residence time for 'the products in the reaction
cell; while Cox used rather extended irradiations period (2-17 hr) in a static
system over a range of pressures ,(25 - 730 torr). It appeared to be more likely
that Some difference between the two results might enter as.a result of the con-
trasting flow and static experimental conditions employed in the two studies.
In the previous study of Okuda and coworkers the extent of inhibition of
S03 formation which resulted from small additions of biacetyl to the SOa gas
was rationalized well in terms of the known S02(3Bi) quenching rate constants,
and it was concluded that S02(3Bi) is the dominant chemically active species
which leads to S03 in SOa photolyses. Presumably the important chemical re-
action leading to S03 in the system is the exothermic, spin conserving reaction
(la):
' (la) SO2(3£i) -I- SO, -» SO, + S0(»2-)
The experimental data from both recent studies are very limited, and no rig
orous test of mechanism or unambiguous proof of reactant states now exists. It
is obvious that further work is necessary to resolve the observed differences in
the quantum yields of SOs from the photolysis of pure SOa and to establish the
mechanism of product formation.
In this work we have determined the quantum yield of SOa formation in both
static and flow systems using pure SOa and SOa-NO, SQa-COa and S0a-0a gaseous
mixtures. In one series of experiments SOa was excited directly^to the SOa(3Bi)
state by light absorption within the "forbidden" SOa(3Bi) «- SOa(X, 1Ai) band at
3700-4000 A. In another series BQ2 was excited to the first excited singlet
state (^-Bi) by light absorption within the first allowed band at 313° A. The
results of these varied experiments provide new information on the detailed
mechanism of SOQ formation in irradiated SOa and allow a meaningful rationali-
zation of the seemingly divergent results of Okuda and coworkersl7 , Allen and
coworkers^"' 103 and -^2
I-F — 1. Experimental
Photolysis Systems
Two different photochemical systems were employed in this work. Excitation
of SOa within the singlet band was effected using a narrow band of light near
3126-3132 A. This was generated from a medium-pressure mercury arc (PEK 200 W),
filter solutions, and glasses°9. Dashed curve A in Figure 24 shows the exci-
tation was made with emission from a high-pressure xenon arc (Osram XBO 450 W)
which was filtered through a series of solution and glass plates to provide a
band of light in the range of 3600-4100 A0"^. Dashed curve B in Figure 24 defines
the excitation region employed here and the overlap of this with the triplet
bands. In each set of experiments a nearly uniform parallel beam of light was
obtained using lenses and light stops. in triplet and singlet excitation experi-
ments a cylindrical Pyrex cell (diameter 4-5 cm, length 50 cm) and a quartz cell
(diameter J cm, length 15 cm), respectively, were used. The absolute incident
light intensity was measured, at frequent intervals between photolysis runs using
potassium ferrioxalate actinometryfl. The relative intensity was monitored during
photolyses by reflecting a small fraction of the incident light into a 935 photo-
tube by means of an angled quartz plate placed in the incident beam. The lamps
were allowed to warm up for a 1-hour period before photolysis was started. The
light intensity was constant to within ±5$ during a given photolysis period. In
70
-------
400
2200
. Motar extinction coefficients (« " Ppfo»(V/]/c£ L/mole/cin) for the
•first aljowed band and the "forbidden" band of SO,(g) at 25°C. Curves /I
and B represent the relative intensity versus wavelength distribution of the inci-
dent light used in Ac singlet and triplet exeitefion iexperhnentt.
the singlet and triplet experiments the incident light intensity was varied in
the ranges of 3.0 x 1014 to 1J.O x 1014 and 5.0 x 1017 to 8.5 x 1017 quanta (cell
face'sec)"1, respectively. The fraction of the incident light absorbed by S02
in the singlet experiments was measured using a phototube at the rear of the cell.
Over the pressure range employed in these experiments (3-2? torr) the absorption
followed Beer's law with e = 28.4 i ./mole*cm. The very small fractions of the
incident light absorbed by the SC^ in the triplet excitation experiments were
estimated using the average extinction coefficient (e = 0.0276 ± 0.0051 j&./mole*
cm) measured with a thermopile (Epply Laboratories) and potentiometer-galvanometer
system. This estimate is consistent with that calculated from the incident light
distribution and the S02 extinction data from the triplet band (e = 0.0206), and
it is considered more reliable than the latter value.
Experimental Proced.ures
A grease-free high-vacuum line (pressure in the region of 10~5 torr) was
used in all experiments. The gas-handling system with high vacuum is shown in
Figure 25. Pressures of 20 torr or more were measured with a calibrated pressure
gauge (Wallace and Tiernan) using a transducer (Dynascience, model FfD) with a
digital voltmeter combination as a null device. Pressures less than 20 torr were
measured directly with a calibrated transduced T. Flow rates for gases in the
system were changed as desired by choosing suitable size capillaries Ci and C2
which were jointed to the outlet side of the photolysis cell R. Fine control of
the flow rate was obtained using a needle valve F (Nupro SS-2SGD) connected to
the inlet side of the reaction cell. Flow rates were varied from 0.13 x 10~6 to
13.6 x 10~6 mole/sec for the singlet band experiments and 0.8 x 10"6 to 73 x 10
mole/sec for the triplet band experiments.
-6
The major fraction of the S03 produced remained gaseous; this together with
the large excess of reactant flowed into a cold trap C kept at liquid nitrogen
temperature. Following each run nitrogen gas (0.3 cm3/sec) was used to flush
the system for 30 min. When NO, C02, or 02 gases in addition to S02 were used,
71
-------
FROM GAS CYLINDER
VIA DISTILLATING LOOPS
TO PRESSURE- CAGE
TO BUBBLER
FJgure2S.Gas handling and flow reaction «ysteni used in this study.
these were separated by distillation before analysis. Then the product S03 and
excess SO^ collected in the trap were allowed to warm slowly and flow into two
bubblers and an ice-temperature trap placed in series with trap C. Each bubbler
contained 10 cm3 of 80$ isopropyl alcohol. The purpose of the last trap in the
ice bath was to test for possible loss of product by vaporization. Excess SOa
collected in the bubblers was flushed out by flowing nitrogen gas through the
bubblers for 3 to U hours.
The cell and associated glass tubing leading to the traps were disassembled
after each run and the walls washed with a small portion is isopropyl alcohol
(8C#>). The S03 recovered from the bubblers in flow experiments amounted to at
least 9<$ of the total S03 collected from all the system including the cell walls,
connecting tubing, and traps. The quantum yields of S03 reported here include
S03 from all sources. Since no measurable amount of S03 was found in the cold
trap, no loss of S03 in the transferring process is assured.
The S03 collected was titrated in the form of sulfate with barium chloro-
anilate using a spectrophotometric method10^5105. Measurements for the standard
curve and photochemical runsowere obtained with a 5~cm path length cell with
absorption measured at 5300 A using a spectrophotometer (Beckman model B). The
analytical method was very reproducible and gave accurate results for standard
samples of sulfate.
Following the analysis of a given run and before each new run, the cell and
connecting tubing were washed thoroughly with double distilled water and iso-
propyl alcohol (100%), dried in the oven for several hours, reassembled, and
evacuated until the pressure in the system reached 10~5 torr. Dark runs were
made which were identical in other respects with the photochemical runs. The
small amount of sulfate derived from these dark experiments was almost exactly
the same for all reaction conditions, about 0.07 M-mole. This suggests that the
72
-------
blank sulfate was formed in the analytical procedure itself, rather than in a
chemical reaction occurring in the flow system.
The reactant gases employed were the highest purity products of the Matheson
Chemical Company. All reagents which were condensable -at liquid nitrogen tempera-
ture were purified further and degassed by bulb-to-bulb distillation in the vacuum
lines to ensure the dryness and purity of all the reagents. The gases were pre-
mixed in the large storage bulb (M of Figure 25), diameter 15 cm, length 100 cm, .
and then allowed to flow into the reaction cell through the capillaries.
Other possible products collected in the trap besides SOg and S03 are SO
and possibly its subsequent reaction products. In theory we expect equal amounts
of SO and 803 to be formed from each bimolecular reaction of SQ& and the excited
SOs molecules. Therefore if there is no secondary reaction involving SO mole-
cules, the amount of SO trapped must be equivalent to the amount of S03 collected.
SO probably reacts with water to form HsS03 and HgS in the aqueous solution.
Previous results suggested that HaSOa (HgO + SO) and S04= are not final products
from SO in water10°^^-0°; the overall reaction may be,- JSO + 3^0 -» 2HsS03 + HsS.
Thus the possible loss or increase of S04= due to the aqueous reaction products
of SO is not expected. It seems unlikely to us that SO reacted with S03 during
the transfer of products or the analysis. Our evidence suggests that these mole-
cules do react extensively in the gas phase in the slow flow and static systems.
However, the high solubility and reactivity of S03 to form H2S04 in water solu-
tions suggest the unimportance of S03-S0 reactions in the fast-flow system. The
kinetic behavior of the S03 quantum yields at high flow rates bears out this
conclusion.
Table XX. Effect of flow rate on the quantum yields of S03 formation in
pure S02 irradiated within the first forbidden absorption band.8
Run Flow rate, Residence Light Photolysis Sulfate ^503
No. mole sec"', time, sec absorbed, period. Bin formed, Experiment Theory
x 10* quanta sec"1 umolc . _
x 10-"-
1
2
3
4
5
6
7
0.
2.
5.
7.
9.
12.
73.
813
73
16
35
63
95
38
1400
435
232
159
123
91
16
3
3
3
3
2
3
3
.16
.18
.39
.12
.88
.20
.16
120
120
120
120
120
120
60
0.
0.
0.
0.
0.
0.
13
.22
31
34
.29
.36
0.20
0
0
0
0
0
0
0
.036
.057
.077
.093
.086
.094
.108
0.034
0.058
0.075
0.086
0.092
0.098
0.108
• Pressure of SO8, 25 ± 1 torr; Radiation in band at 3700-4100 A.
6 Calculated from relations (A)-(D) using a = 0.108 and rate constants kt = 6.0 X 106
and k* = 5.0 X 1051./mole-sec and integrating by computer.
73
-------
Experimental Results
Several series of experiments were made in an attempt to find the origin
of the conflicting results of the previous published studies of the irradiated
pure SOa system-^ '^^ and to establish the mechanism and the nature of the ex-
cited states involved, in S03 formation. In the first series of runs direct
triplet excitation was employed using radiation in a band near 3700-^000 A (see
Fig. 2U. The data in Table XX summarize the effect of the reactant flow rate on
the S03 product quantum yields. In Table XXI the effects of added NO, C02, and
Oa on the quantum yield of S03 formation in SOg mixtures are shown. In the
second series of experiments excited singlet sulfurodioxide was generated by
irradjation with a narrow band of light near 3130 A (see Fig. 2U). Quantum
yields of S03 formation were determined for experiments using both a static
system (Table XXIl) and a dynamic system with varied reactant flow rates
(Table XXIIl). In Table XXIV are given the quantum yields of S03 from experiments
using singlet excited SOg mixtures with added nitric oxide, carbon dioxide, and
oxygen gases and with varied flow rates.
Table XXI. Effect of foreign gases on S03 quantum yields in S02 mixtures
irradiated within the first forbidden band in the flow system.8
*"" V Torr -M"S02 "SO, -S030/-S03*
WO. * "* J
a) M • NO
8
9
10
11
12
13
4.75
2.50
0.20
0.15
0.10
0.050
0.000
0.190
0.100
0.0080
0.0060
0.0040
0.0020
0.0000
0.000 i 0.005
0.000 ± 0.003
0.046
0.053
0.064
0.081
0.108b
2.35
2.04
1.69
1.33
1.00
b) M - CO2
14 90.0 3.60 0.052 2.08
15 61.3 2.45 0.060 1.80
16 40.8 1.63 0.071 1.52
17 25.0 1.00 0.089 1.21
18 20.2 0.81 0.094 1.15
0.0 0.00> 0.108b 1.00
c) M • O2
19
20
21
22
11.54
7.04
1.91
0.87
0.00
0.46
0.2B
0.076
0.035
0.000
0.077
0.085
0.092
0.083
0.108b
1.40
1.27
" 1.17
1.3O
1.00
• Pressure of SO5, 25 ± 1 torr; radiation in band at 3700-4100 A; flow rate 9.63 X 10-<
mole/sec.
6 Limiting quantum yield derived from computer fit of *sot versus flow rate data at
high flow rates for the pure SOj system.
-------
Table XXII. Quantum yields of S03 formation in pure SC>2 irradiated at JIJO A
in the static system.8
Run Light absorbed. Photolysis Sulfate Quantum yield 503 formation
No. quanta secr'cell"1 period, hr formed. Experiment Calculated
x 10"'^ pinole -Method Ib Method
23
24
25
26
37 i ..
28
13.84
18.85
8.30
15.80
12:90
'24.88-
.t .
3.00
6.25
8.33
16.30
21.7S
- . '*** ! * "
0.22
0.23
0.18
0.38
0.21
**'-*
0.0090
0.0032
0.0043
0.0020
P.0012 ,
0.00 TO
0.0019
0.0010
0.0011
0.00058
O.OS9R1
f O.OWSS1
O.OO4O
0.0033
0.0043
0.0033
O.O035
'"OMMtM
. .* . , , >-*,», - .c ,.
* Calculated from relations (AMD) using a = 0.090 and rate constant* tj = 1.2 X
10«, it = 5.0 ^X lO'L/molc-sec, *. = 0.0, and integrated by computer.
' Calculated from relation (A)-(D) with one alteration: the heterogeneous wall re-
moval of SOi gas wai assumed to occur with a rate constant kw = 4 X 10~4 sec~l. The
other rate constants were as employed in method I.
Table XXIII. Effect of flow rate on the quantum yield of 80s formation in
pure S02 irradiated at 3130 A in the flow system.8
Run
No.
29
30
31
32
33
3*
35
36
37
38
39
40
41
42
43
44
45
46
47
48
50
51
52
Plow rate.
nole sec* ,
x 106
Residence
time, sec
Light absorbed
quanta sec"'
cell"1 x 10" ^
Photolysis Sulfate Quantum
period, nin formed. Yield of
unole SO3
0.13
0.81
0.81
0.81
1.59
1.59
3.50
3.50
5.16
5.16
7.35
7.35
9.63
9.63
12.95
12.95
0.81
1.59
5.16
12.95
13.60
4.22
7.35
15.05
1000
200
2OO
200
130
130
45
45
31
31
21
21
16
16
12
12
200
125
31
12
10
37
21
10
3.73
3.64
3.56
3.73
3.64
3.64
3.59
3.30
3.S6
3.73
3.30
3.70
4.12
4.12
2.73
3.64
5.64
5.64
5.64
5.64
B.50
12.92
12.54
13.35
240
60
60
240
60
60
60
240
240
240
240
205
60
30
205
6O
120
120
120
120
60
240
240
60
0.17
0.04
0.07
0.36
0.09
0.10
0.15
0.45
0.62
0.63
0.58
0.75
0.23
0.13
0.59
0.20
0.03
0.14
0.56
0.60
0.45
0.95
1.13
0.66
0.010
0.015
0.031
0.041
0.041
0.046
0.072
0.057
0.073
0.071
0.074
0.084
0.093
0.111
0.086
0.09S
0.005
0.020
0.084
0.090
0.090
0.031
0.038
0.083
0 Pressure of SO2) 25 ± 1 torr.
75
-------
Table XXIV. Effect of foreign gases on the quantum yield of S03 formation^in
3130-A irradiated SOs mixtures in the flow system.3
Run
No.
Residence
time, see
Quantum yield of
SO] formation
a) H '» HO"
S3
34
55
56
?7
58
59
60 1
61 '
b) H - 002
62
63
64
65
66
c) H - 02
2.10
2.10
2.10
2.10
2.10
0.21
O;M
0.10
o.iw
o.oo
90.0
67.0
40.0
20.0
30.0
0.0
125
31.
31
21
12
21
21
21
21
0
21
21
21
21
.21
0
0.084
0.084
0.084
0.084
0.084
0.0082
^0.0055
0.0045
0.0032
0.00
3.60
2.70
1.60
0.80
0.75"
0.00
0.008
0.023
O.027.
0.029
0.024
•6S.OS3
0.058
0.060
0.067
0.090*"
O'.04»"
0.057
0.066
0.070
0.072
67
68
69
2.76
1.50
0.44
0.00
21
21
21
0
0.11
0.060
0.018
0.00
0.080
0.083
0.079
0.090*>
• Pressure of SOj, 25 ± 1 torr in all but run 66 in which 40 torr was employed; 70 =
5 X 1014 quanta/cell-sec.
* Limiting quantum yield derived from computer fit of *sot versus flow rate data at
high flow rates for the pure SOi system.
I-F —2. Discussion of the Results
Dependence of Quantum Yields of S03 Formation on Reactant Flow Rates
The absorption of light quanta of wavelengths greater than 2160 A provides
insufficient energy to dissociate a sulfur dioxide molecule [SOa -» 0(3P) +
S0(32~)]. Thus it is reasonable to speculate that S03 formation from the photol-
ysis of SOa within the first allowed or. the "forbidden" band is derived from an
electronically excited SOg molecule reaction with ground state SOa molecules.
Reaction (la)
(la)
S02 -» SO, + SO(3S-)
involving the SC>2(3Bi) molecule has been suggested as. an important source of
S03 in S02 gas irradiated within the 2^00-5200-A band1''' '51,26 and the "forbidden"
triplet band of S02 (UOOO-3600 A). SO(3S~) was observed in the flash photolysis
of SOs by Norrish and Oldershaw1(^9j but in these experiments the authors attributed
its formation to the direct photodissociation of SOg at wavelengths shorter than
2200 A present in the flash. Only recently both products of reaction (la) have
been identified spectroscopically in irradil^jOa at the longer wavelengths.-
76'
-------
Thus Basco and Morse110 and James, Kerr, and Simons111 have observed. S0(3£~) in
the flash excitation of SOa with wavelengths greater than 2500 A. Until recently
S03 formation in irradiated SOg has been inferred either- from the observed S02 -4..v
^'
'1 5 or the identified sulfate ion product formed by the reaction of
the primary products with water" j11?,11^ AT,!02. Through the study of SOa
photolysis at 31JO A in very dry photolysis cells, Daubendiek and Calvert have
been able to characterize and follow the kinetics of gaseous S03 formation and
decay directly through infrared absorption spectres copy11' . In view of the
present data it seems highly probable that S03 and SO are primary products of
the electronically excited SQ2 interaction with SQ2 molecules.
Reaction (ib)
: (Ib) SOsCfli) + SO, -» SO, + SO(1A)
L. • - • - -' --- - • - - •
involving the excited singlet state should also be considered as ap possible
source of SOs in 3130-A irradiated SQ2. Both reactions (la) and (ib) conserve
spin and 'are significantly exothermic: AHia = -27-7 (for S02(3Bi) in the 0,0,0
level), and AHib = -2^.3 kcal/mole (for internal excitation of SOg^Bi) as
formed on light absorption at 3130 A). Conceivably other reactions involving
the other nonradiative excited states of SOg rcay also occur when S02 is irradi-
ated within the first allowed band21 ,57,88-90.
However, whatever, the nature of the state of SOa forming S03, it is apparent that
the molecule SO is formed in amounts equivalent to those of the S03 generated.
We and others have tended to ignore the fate of the SO species in the photolysis
of SOa and its mixtures. In particular it has been assumed that it does not inter-
act with the primary product S03. From our present results it seems clear that
this assumption is not valid. Note that the quantum yield of S03 formation is
a function of the flow rate or the residence time of the irradiated SO^; this
is the case whether the irradiation is effected within the triplet band (Fig. 26)
or the singlet band (Fig. 27). In each case the observed quantum yield increased
dramatically as the residence time of the irradiated SOa was decreased from 1000
sec to 10 sec. In both systems the quantum yield approached limiting values at
high flow rates. Also note in Table XXII that the quantum yields of S03 observed
in the static system are only about one one-hundreth of those observed at high
flow rates. This observation confirms the apparently conflicting results of
Okuda and coworkers1? and those of Coxl02 an^ Allen and coworkers^", 114,103.
There is an obvious mechanism choice which appears to explain the effect of flow
rate on §gQ . At low flow rates the S03 and SO primary products of the excited
SOa reactions (la), (ib), and/or (ic) react significantly before these primary
products can be trapped at low temperature and subsequently allowed to react
with water. Reactions (2) and (3') should be considered in this regard:
(Ic) SO2(M2, M2, 3B2) + SO2 -» SO, + SO('A or *Z-)
Basco and Morse110 have suggested that these reactions might explain qualita-
tively the ultimate decrease in SO concentration which they observed following
its formation in reaction (l) in the flash photolysis of SOg. To the best of
our knowledge the present data provide the first experimental evidence to sub-
stantiate this mechanism. Schofield11" haj^reyiewed the rate data pertaining
to reaction (3)11''r"120, and he sets an upp|p¥Sjfe-kfor k3 = 1.8 x 106 jfc./mole-
sec. Halstead and Thrush11? quote an uppeiVlimit' for both k2 and k3 of 5 x 106
77
-------
0.10
0.08
0.06
0.04 —
OQ2
/
I I
,1 .1
0.4
O8
'/V. Wc
1.0
6.25
. Effect of residence time r.taf SOjpn the quantum yield of SQitema-
toon in SGOtMOQ^Aipn^^p,. #», -. 25j?a;J.- (&6,=fe OJ5) X .10-".
einstcin/L/sec; the solid curve wai derived by computer integration of relationi
' (B)-(D) using kt - 6.0 X 10« and k, - 5.0 X 10< l./mok/sec; the dashed curve
' is the theoretical dependence estimated in the same fashion but using Jti =• 1.2 X
' 10* and i| = 5.0 X -Id1 l./mole/aec as derived from the excited singlet studies
shown in Figure^? •
aio
0.08
0.06
0.04
0.02
4
»c" x(0z
Figure27.Effcct of residence time T of SOs on the quantum yield of SO> forma-
tion in 3130-A irradiated SO2. PBo, = 25 torr; /„ = (5.6 ± 0.8) X 10-'
einstein/l./sec; the solid curve was derived by computer integration of relations
(B)-(D) using *j = 1.2 X 10" and k, = 5.0 X 10s l./mole/sec.
-------
.'./mole* sec. It is significant to note that if the kg and k3 values are this
small, then the time required for" SO to reach a steady-state concentration in
an irradiated 50% system such as we employed falls in the range of hundreds of
seconds, and only an insignificant fraction of the primary SO and S03 products'
formed should react at the shortest residence times for the fast flow system
employed here (10 sec). Thus in theory the primary products can be carried to
the trap and frozen out in our experiments at the fast flow rates before any
significant reaction by (2) and (J>) can occur.
Indeed a reasonably good quantitative fit of the $QQ versus flow rate data
is possible using the simple reaction scheme outlined. In theory the dependence
of the experimental overall quantum yield of S03 formation on residence time
should be given by relations (A)-(D).
, (A) fcso, = [S08]r//0T
/•(SO.), ft •_ -
(B) / rf(SO,] = / (/„« - ISO,HSQ)A,) dr
J [S0.]o J 0
/•ISOlr fr
(C) / ./mole -sec"1 , while that of Figure 27 with singlet S02 excitation was calculated
using k2 = 1.2 x 10s and k3 = 5.0 x 105 jfc./mole*sec. These estimates, the first
of which we are aware, are in accord with the upper limits of k2 and k3 which
were derived from the previous studies-'--'-^ >^2® . Some measure of the sensitivity
of the theoretical fit to the choice of the rate constant kg can be seen in
Figure 26. where the "best" fit k2 derived from the singlet data (a factor of
2 higher than that from the triplet data) is used together with k3 = 5.0 x 105
4. /mole-sec. The fit of the data is much less sensitive to the magnitude of k3
chosen, and the present estimate of this constant is subject to considerable
uncertainty; we estimate that a value of k3 = (0.5 ± O.U) x 10s £./mole'sec is
reasonably consistent with our results.
It is interesting that the rate constant k2 estimated from the singlet exci-
tation of S02 is a factor of two higher than that for the triplet system. It is
possible that this is due to the occurrence of a spin allowed reaction (2') in
79
-------
the singlet system where SO^A) may be the product of reactions (Ib) or (Ic):
r
! (2') SO(»A).~fSO,-»2SOt'
However there are several key points, which would seem to discount this hy-
pothesis, l) As we shall see in later discussion most of the SOs product formed
in the singlet experiments is derived from a quenchable triplet state which would
not form S0(1^) in its reaction with SOg. 2) The rate of the wall quench-ing of
the SOf1^) species to SO(3S"), observed by Breckenridge and Miller at low pres-
sures1^0, is large when extrapolated to our conditions, and this reaction is
probably a major fate of SO^A ) in our system. 3) The radiative decay rate of
SO(1A) is probably also fast on the time scale of the SO-SOs reaction (2)120
and it probably accounts for the loss of another major fraction of the SO(1^J
generated in our system. H) The actual rate constant for (2X) would have to be
several orders of magnitude larger than that for reaction (2) in order for this
reaction to be the origin of this result. It seems more likely to us that the
minor mismatch of the ka values derived from the two systems has its origin in
the necessarily large uncertainty in the measurement of the very small extinction
coefficient of the SQ2 in the triplet band system, the values of Ia, and hence
the $gQ estimates. Certainly within the experimental error the rate constant
data from both systems are compatible.
It is clear that the occurrence of reactions (2) and (3) can account quanti-
tatively for the observed variation of the quantum yield of S03 formation with
flow rate in both the singlet and the triplet excitation experiments. It is in-
structive to use this mechanism to attempt to rationalize the quantum yield data
obtained from the static photolysis experiments performed here and by other
researchers. We must, recognize that this procedure should underestimate SOs
quantum yields in the very long static runs, since the transport of SOs from the
gas phase to the wall will occur during these extended experiments, and its
reaction there with adsorbed water may form the relatively nonvolatile H2S04,
as evident from the much larger fraction of the S04= derived from- the wall of
the cell in these experiments. Water will diffuse to the surface of-the "dry"
cell from the interior structure of the cell wall during the long experiments,
and it will be a much more significant reactant than in the fast flow experi-
ments. When SOs ^s removed from the gas phase, reaction (2) becomes less in-
hibiting, and the $sOo> as derived from the analysis of the combined H2S04
condensed on the wall and the gaseous S03 in all of the studies, will rise.
Table XXII summarizes the static experiments carried out in this work. Compare
the experimental quantum yields with those calculated from relations (A)-(D).
We have used the rate constant estimates derived from the singlet excitation
of SOa in the flow system together with the measured values of la and integrated
the differential equations (B')-(D) for the very long photolysis periods shown
in Table XXII for runs 23-28. Although the irradiation times range from 3 to
3^ hours in these cases, and the rate constants were derived from experiments of
10-2000-sec irradiation time, clearly the approximate magnitude of the quantum
yield is predicted reasonably well through the application of the simple reaction
scheme (l)-(3). In the best case (run 27) the theoretical value differs from
the experimental one by only a factor of 2.3, while in the poorest case (run 23)
the experimental and theoretical values differ by a factor of k.J. In every case
the predicted S03 is somewhat lower than that observed experimentally; we would
expect this since SOS removal at the wall increases in importance in these long
runs.
One can match more closely the present $SQ data from the extended exposures
in the static system by adding an apparent first-order wall removal step for SOs:
80
-------
SO, 4- HjO(adsorbed on wall) -^ H2SO4(adsorbed on wall)
Of course the apparent rate constant for any such heterogeneous reaction must be
a complex function of the pressure of the reactant gas, the size of the cell,
and the nature of the cell wall and its conditioning prior to the run, and it is
expected to be of limited value in treating other systems. However, in our runs
made at constant pressure of 30^(25 torr) and with a set pattern of cell treatment,
perhaps there is some justification for an attempt to include this wall reaction
in a somewhat quantitative fashion. Assuming the values of k2 and k3 derived pre-
viously and taking ky = h x 10~4 sec"1, one can match quite well most of the §so
data shown in Table XXII; see the last column, labeled "Calculated, Method II".
It can be shown that the loss of S03 at the wall is only a minor reaction in the
fast flow system; the choice of the homogeneous rate constants k2 and k3 derived
from the flow data is not altered significantly by including this heterogeneous
reaction.
It is interesting to estimate the theoretical quantum yield of S03 formation
expected from the runs of Cox-^2. Using his IJ-hr exposure time in experiments
with PgQ^ = 29 torr and Ia = 7.63 x 10~8 einstein/£.'sec, we estimate from the
present nomogeneous mechanism (reactions (l), (2), and (3), only) that $sOa =
O.U8 x 10~3; Cox reported an experimental value of (3-^ ± 1.0) x 10~3. When
one considers the potential significance of S03 removal from the gas phase by
heterogeneous reaction, for which we have not attempted a correction here, the
agreement between theory and the Cox experiment is considered satisfactory.
The earlier study of Allen and Bonelli^? which employed a static system with
long exposure times using 3130-A radiation, gave quantum yields of S03 formation
similar to those reported here and to those of Cox. These results can be ration-
alized in terms of reactions (2) and (3) with about the same precision noted for
the other results described above. Inclusion of the wall removal reaction using
the value kw. = U x 10~4 picked for our system gives a surprisingly close fit for
most of the Cox data and the Allen and Bonelli results. Obviously the potential
importance of the unpredictable heterogeneous transport and capture of SOs at
the wall makes such calculation of limited confirmatory value. However, it
appears to us that all of the quantum yield data from the photolyses of pure
seem consistent with the simple mechanism involving reaction (l)-(3)5 and de-
viations from the values predicted from this mechanism for runs of extended time
periods are qualitatively consistent with the expected wall removal of S03 as
H2S04. The "best" estimates from the flow experiments give the rate constants
k2 = (1.2 ± 0.7) x 10s and k3 = (0.5 ± O.U) x 106 Jt./mole-sec.
Mechanism of S03 Formation in 3600-UOOO A S02(3Bi) Excitation Experiments
The rate data summarized in Table XXI offer a significant test of the mechan-
ism of S03 formation. These were determined in mixtures of SOs. with added HO,
CO^ , and 02 gases. In each case the pressure of S02 was constant at 25 torr, and
a fast flow rate of reactants was employed. Thus the experimental quantum yields
of S03 represent initial quantum yields since the S03 destruction in reaction (2)
is unimportant for these conditions. The data can be rationalized well in terms
of the following simple mechanism:
(I) SO»(Jt, Mi) + M3600-4000 A) — SO»('fli)
(la) SOt(»BO + S02 -» S0a + SOC2-)
(4) SOiCBO + SO* -» (2SO2)
(5) SOj('Bi) + M -» (SOs—M)
81
-------
M is an NO, C02, or 02 molecule and (S02-M) represents ground state
molecules or any nonradiative products of the quenching reactions.
and M
The limiting'quantum yield of S03 formation from the triplet excitation ex-
periments gives the first reliable experimental estimate of the rate constant
ratio kla/(kla + k4) = 0.108 ± O-0^ Using the measured value of kla + k4 =
(3-9 ± 0.1) x 108 jfc./mole'sec'1^1!26, we estimate kj. = (k.2 ± O.k) x ID7 H. /
mole*sec.
Applying the steady-state assumption to the S02(3Bi) species in reactions
(l), (la), (M<
TDUre o le»
by relation
)
(5), we anticipate the ratio of the quantum yield of SOs in
•t-.hn-ri n r, cr>_ m-5v-i-.m»oe with added gas M ($so« ) should be given
that in
.(E)
$80,°
-'1. +
(^SOa1
[M]
[SO,]
In Figure^S the quantum yield data for runs with M = NO, CQa, and. Og are plotted
in the functional form of relation (E). The solid lines are determined by the
least squares fit of the results without forcing the data through the intercept
of unity expected theoretically from relation (E). Within the error limits both
the NO and COa data do conform to this expectation; the least squares intercepts
for the NO and C02 data are 1.00 ± 0.0k and 0.91 ± 0.12, respectively. The slopes
of these plots give estimates of k5/(kla + k4) which may be compared in Table XXV
with the same ratios derived from directly measured rate constants for reactions
(U) and (5) in SQ2(3Bi) lifetime studies^0. The rate data from the very dif-
ferent experiments for NO and COa as M agree well within the experimental error,
and they add credence to the mechanistic interpretation given here. The data
for 02 are not in accord with this simple mechanism outlined; an increase in SOs
formation occurs with small 02 additions, and then a decrease is observed as
high pressures of 0% are added. This is analogous to the findings of Coxl02 who
proposed that the S03 increase at low 02 concentrations resulted from the re-
action (6):
(6)
SO + O« -» (SO—Os) -» SO,
0.010 tNO]/[SOz]
0.5 [Oz]/[SOi]
4-0
Figure 2 B .Plot of the ratio of the quantum yield of SOj formation in pure
SOj(*80j°) to that in SOradded gas mixtures (*soiM); most °f tne data are
from experiments with SOi excitation within the forbidden SOi('Si) *~
SOi(.?, lAi) band (3600-4000 A) using fast flow rates of gases. The open
circles are from singlet excitation experiments at 3130 A calculated from relation
(F) which corrects for SO> formed from "singlet excited SOj.
82
-------
Table XXV. Comparison of the rate constant ratios ^/(k^. + k4) derived from
the present product S03 quantum yield data and S02(3Bj.) lifetime
data*0:
(la) SO2('Bi) + SOS -» SO, + SO('S-)
(4) SOzCBO + SOj -* (2S08)
* (5) SOjCJBQ + M -* (SO. - M)
*5/<*l. + Id,)
H Present data Lifetime studies 136]
HO
002
O
2
170 ± 8
0.34 i 0.06
0.34 ± 0.46
190 ± 9
0.29 t
0.25 ±
0.02
0.01
In this mixture S03 may arise from reaction (7) as
O, -» (S04) -» SO, 4-
Little is now known about the absolute magnitude of ke and k7, although the
available evidence suggests that both rate constants are rather small-^'-1-02*^.
The limited data for the SQ2-02 system obtained here are consistent with this
view but add little new to quantify this thesis.
It is important to recognize that all the present triplet excitation data
are consistent with the S02(3Bi), phosphorescent state of SQ2> being the sole
reactant to form S03. Not only are the ratios of the quenching rate constants
obtained here nearly identical to those observed from lifetime measurements for
S02(3Bi), but essentially complete quenching of S03 formation occurs at high
added. NO pressures; see runs 8 and 9 of Table XXI. It has been reasoned that
the collisional quenching of SC>2(3Bi) by chemically "inert" gases such as CO&
promotes crossover to other near lying nonradiative triplet states (3A2, 3B2).
This proposal seems attractive since the activation energy for this quenching
reaction, CCfe + S02(3Bi) -» C02 + (S02), 2.7 ± 0.3 kcal/mole, corresponds closely
to the energy for the promotion of SC>2(3Bi) molecules to the lowest energy region
of the triplet manifold in which distinctive perturbations give spectroscopic
evidence of neighboring triplet states92. if this mechanism for quenching is
correct, then one must conclude that S02(3A2) and/or SC>2(3B2) molecules do not
form S03 in reactions analogous to (l). This conclusion is inconsistent with
the suggestion of James, Kerr, and Simons who feel that the 802(^2) state would
be ideally constituted to form S03 by reaction with S02111' Perhaps nonradiative
coupling of the 3A2 with the upper vibrational levels of the S02 ground state
is sufficiently strong to lower its lifetime and lessen its chance to react with
S02-
Mechanism of S0_3 Formation iri S0g Excited a^t 3130 A
The data from the fast flow experiments with pure S02 and S02-N0. SC^-COa,
and S02-02 mixtures (Tables XXIII and XXIV) may be used to illucidate'the nature
of the reactive states in singlet excited S02 mixtures. First it should be
noted that unlike the triplet excitation experiments outlined previously, the
addition of a relatively large quantity of NO gas to the S02 does not suppress
S03 formation completely. See the fast flow experiments 5U-57 of Table XXIII.
83
-------
Thus we conclude that some excited singlet state reaction does occur in this
system; possible contenders are reactions (Ib) and (Ic):
Qb) SOsCfl,) + SOj -» SO3 + SO('A or 32~)
(8) SOiCB,) +. SO, -» (2SO2)
(9) SO »(»£,) + SO, ^VSO»T:+ SO»
(Ic) SO,(M,) + SOS -+ SO, + SO('A or »2~)
(10) SO,(MO + 'SOi -» (2SOO ' '
(11) SO«(lyi,) + SO« -f SO2r -F SO*
It is impossible to infer from these data alone which of these reactions is -
important in S03 formation here. .However, if the SQ2(1Bi) molecule is this
source, then we can estimate from the quantum yields. for the NO inhibited runs,
the rate constant ratio kib/(^ib + k8 + ^e) = 0-026 ± 0.003.o From the measured
rate constant sum kxh + ke + kg for S02 excited at X ^ 3000 A (8.6 ± 1.8) x 101D
We estimate k^ = (2.2 ± 0.5) x 109 £. /mole-sec.
Using the data of runs 58-61 on Table XXIV we may investigate the reactivity
of the nitric oxide-quenchable species which leads to 80s formation in the singlet
experiments. For the very small amounts of NO added in these runs, it is clear
that the excited singlet is quenched largely by SQ& in these runs. If only the
^•Bi and one other state of unknown designation (SOaT) is involved in S03 for-
mation in the excited singlet experiments, then the rate function F should apply:
: ~ *so,°' - 0.026 , , *„ [NO]
(t)
*so,NO - 0.026 (kit + kn) [SO,]
Here ki2, k13, and k14 refer to the quenching reactions for the undesignated re-
active state 502^ which is quenchable with added NO.
(12) SO»r + S0» -» SO*, + SO(«Z-)
(13) SO»T + SO,->(2SO2)
(14) SO,T + NO -* (SOy-NO)
Function (F) is plotted in Figure 28 (open circles) along with the data from the
S02(3Bi) quenching experiments from the direct triplet excitation with SOg-NO
mixtures (filled circles). Obviously the nitric oxide-quenchable, ill-defined
state formed in the singlet excited experiments has the same rate constant ratio
for quenching with NO and S02 as that for SOs(3Bi) within the experimental error.
We conclude that either S02(3Bi) is the most important reactive state which leads
to S03 in singlet excited SO^ mixtures or another NO-quenchable state such as
3A2 or 3Ba is generated here and it reacts to form SOs or is quenched by NO with
the same relative rates as found for S02(3Bi). If the first alternative is cor-
rect, as was originally suggested by Okuda and coworkers1''', then we must speculate
that the intersystem crossing ratio for JIJO-A excited SOa to form S02(3Bi) is
much higher (0.59) than that estimated by us for the photolysis at the shorter
wavelengths (~ 0.10). The high pressure mechanism of S02(3Bj.) formation sug- •
gested by Wampler, Horowitz, and Calvert '*l cannot be an important source of
3Bi molecules at the relative low pressures of S02 employed in this work.
-------
The singlet excited S02 experiments with added C02 (runs 62-66, Table XXIV)
should be considered also in light of the mechanism suggested here. In this case
the quenching effect of CQa is small, and relative large quantities of C02 must .
be added to observe one at all. This leads to considerable complication in at-
tempting to sort out the states involved, since singlet quenching-by C02 as well
as S02 must occur in these cases. ¥e may simplify our considerations somewhat
by assuming that only two reactive states of S02 are present in this system,
S02(1Bi) and the undesignated reactive state S02T.- Reactions (15)-(17) must now
be considered in addition to reactions (ib), (8), (9), (12), and (lj):
05)
(16)
-"•• - IT '.-'ii, *
+ 00,
+ eo,
SOsT + CO,
•SOi.+ GO*
We may attempt to treat the data utilizing the measured rate constant ratios
(k15 + kls)/(k1b + k8 + £9) = O'.TI12?, and assuming ki7/(ki2 + kls) =0.29 as
measured for the SQa^Bi) state^0. Then the quantum yield of S03 formation in
singlet excited S02-C02 mixtures should be given by relation (G).
0.026
(G)
1 + 0.71
0.108
0.29
[CO,]
[SO,],
[CO,]AW
k, [SO,](*lb + k»
Figure 29.Plot of function (G) versus [COj]/[SO2]; data are from the 3130-A
irradiated mixtures of SOj and COj.
-------
The excess triplet mechanism will "be relatively unimportant for these conditions *-'
A test of relation (G) is shown in Figure 29. The data are not accurate enough
to define a reliable intercept, but they suggest (O.U6 ± 0.09). In theory the
intercept is equal to the rate constant ratio ks/(k:1^ + ks + ks) and should be
compared with the ratio 0.59 derived elsewhere in this work. The slope in theory
equals k15/(k1t + kQ + k9) = 0.6j ± 0.0^. The only other estimate of this rate
constant ratio of which we are aware, was derived by Demerjian and Calvert9^ from
the 3130-A irradiated S02 - C0g - c is -2 -but e ne mixtures at high COg pressures. This
gave the approximate value of 0.14. Within the large error inherent in this
method of data treatment which involves differences in quantum yields which are
near equal, the COg-SOs results are qualitatively consistent with the generalized
two-state reaction scheme outlined above, but they do not provide any meaningful
test of it.
The Og-SOs singlet data (runs 67-69 of Table XKIV) are, as observed in the
triplet study, complicated by the reaction of SO with 02) and no quantitative
rate information can be gleamed from them at this time.
Conclusions
The $S03 data obtained in this work using both flow and steady-state static
systems confirm the apparently divergent results of both Okuda and coworkerslT,
and Cox102. The clue to the quantitative rationalization of all of the results.
is the relative importance of reaction (2), S03 + SO -» 2S02, in static and slow
flow-rate experiments. The S03 product rate data from irradiated SOa mixtures
at high flow rates within the forbidden SOa(3Bi) «- SQ2(x., 1Ai) band point to the
involvement of S02(3Bi) as the sole reactant with S02 in forming the 80s product.
Similar rate data from JIJO-A irradiated SOa mixtures show that both singlet
(probably -""Bi) and some triplet state are reactants forming S03 in these systems.
The ratio 'of the quenching rate constants of the reactive triplet state toward
NO and SQs molecules is near identical to that observed for the S02(3Bi) molecule
in lifetime studies. However, if S02(3Bi) is the reactant here, then the data
require that the JIJO-A excited singlet S02 molecule form S02(3B1) molecules in
a much larger fraction of the quenching collisions with SOg than previous kinetic
data allow for the excitation of S02 at the shorter wavelengths. If another
triplet state such as S02(3A2) or S02(3B2) is the reactant here, then the results
demand that the relative reactivity of this state with SQ2 and NO be near identical
to that for S02(3Bi). It is not clear now which of these alternatives is correct.
We are engaged in the determination of the efficiency of S02(3Bi) generation in
SOs excited at 3130 A and the longer wavelengths within the singlet band in an
attempt to resolve this uncertainty in the mechanism. We have summarized the new
rate constants derived in this work in Table XXVI.
Table XXVT. Summary of new rate constant estimates derived in this work.
Reaction Bate Constant (1. mole"'sec'1)
(la) S02(3B,) + S02 - S03 + S0(3!!-) (4.2 ± 0.4) x 107
lib) SOzlX) t S02 -> S03 + SOl'A or 3E') (2.2 t 0.5) x 109
(2) SO + sn3 - 2sr>: ().? * (J.TI x l.O6
(3) 2SO -> S02 * S [or (S0)2) (5 ± 4) x 10s
86
-------
I-G. The Nature of the Excited Singlet States
in SOg Ptiotolysis Within the First Al-
lowed Absorption
Until the recent detailed study of Brus and. McDonald122 the kinetics of
fluorescence emission from SC^ excited within the first allowed absorption band
(2500-3200 A) had been rationalized in terms of a simple mechanism involving
only one emitting excited singlet state
(I)
(1)
3S02 + S02 (2)
SQj + to,. (3)
(S02) (4)
3S02 (5)
3S02 represents a triplet state, presumably S02,(3'B1) , and (250%} and (S02)
designate ill-defined, non-emitting (or very, long-lived) excited states, products
other than SOa, or ground state S02. The best current information from our group
suggests kg «ka + 1^ for all wavelengths of excitation of SOa1^ '. However, a
divergence of opinion has remained on the possible importance of the first order
non-radiative process (k). Mettee and co-workers 3^ ,35 have concluded from limit-
ing quantum yield measurements at zero pressure that k3/(k3 + k4 + 1%) « 1 at
all wavelengths. Calvert and co-workers18523J98 have suggested that k3 « k^ at
2650 A and k^ > k3 at the long wavelengths. The recently published note of Brus
and McDonald122 presented clear evidence for the existence of two emitting singlet
states of 80s which are presumably both populated by direct absorption within the
first allowed band of SQa- As these authors suggest, their findings bear directly
on the interpretation of previous steady-state studies of SQ2 emission. They
conclude that the unexpectedly long lifetime and double exponential decay which
they observe invalidates the conclusion of previous workers that SOa undergoes
a unimolecular internal conversion. We now concur with this conclusion as also
formulated by Mettee3^ several years ago, and. we offer here support for it.
First, one should question why Sidebottom et al98 ^^ not see two emitting
singlet states in their recent lifetime studies using SOa excited by a 2662 A
laser pulse. From the recent rate parameters obtained for the two emitting species
by Brus and McDonald.122, one can estimate that the mean lifetime of the two states
at pSOa = 10 M- are 1.6 and 37 fisec. respectively, and. at Pgo2 = 100 n , 0.1? and
6.3 (J-sec, respectively. Since most of our experiments designed to detect a second
component were carried out in this pressure range, it is somewhat surprising that
the sensitivity of our method was not suitable to find the short-lived component.
A clue to the reason for this failure may be had. from the observation of Brus and
McDonald who report that the ratio of the fluorescence intensities from the short-
lived to that of the long-lived species was so small for the 26lT A wavelength
excitation that it was not possible to form an accurate Stern-Volmer plot for the
lifetime data of the short-lived species for these conditions.
Note that the quenching rate constants observed by us previously1^ 3 98 are
near equal to those reported by Brus and McDonald for the long-lived species;
see the data of Table XXVII. It is clear that our data refer to the dominant
87
-------
Table XXVII. Comparison of quenching rate constants (kj. + k2) for excited
singlet SOs (long-lived component)
Wavelength (A)
(E mole"1 sec'1 a))
Reference and method of measurement
2617.4
2662
2662
2715.2
2860
2856.7
-2980.9
2980
3000
3004
(2.7 ± 0.2) X 1010
(3.8 ± 0.1) X 1010
(3.2±0.2)X 1010
(3.9 ± 0.2) X 10 10
(5.1 ± 0.8) X 1010
(4.9 ± 0.4) X 1010
(7.0 ± 0.2) X 1010
(8.6 ± 1.8) X 10 10
(7.8±0.7)X 1010
(8.3 ± 0.8) X 10 10
ref. isdaser excitation
ref.23 laser excitation
extrapolated from near linear plot of
Jtj + ki versus X. data of ref. 122
filtered Xe-flash data of ref. 23
filtered Xe-flash data of ref. 23
ref.i22laser excitation
ref.l22iflser excitation
filtered Xe-flash data of ref. 23
excitation at 2662 but observation at
long times^frorn ref 23 .-•
ref i£>g>1aser excitation
a) Error limits shown here and elsewhere in this paper are ± 2o.
long-lived state delineated by these workers. Our lifetime estimate from the
2662 A laser experiments, TO = 36 ± k p.sec, is in fair accord with that of Green-
ough and Duncan^ obtained using broadband flash lamp excitation, TO = 42onsec,
although it is somewhat less than that found by Brus and McDonald at 26l? A,
To = 79 ± 30 (isec. The only significant difference between the previously esti-
mated properties of this emitting state and those for the long-lived species
delineated by Brus and McDonald, lies in the lifetimes at the long wavelengths.
Sidebottom et al9° did not measure this quantity directly since a short lifetime
excitation beam in the range of 31°0 A was not available to them at that time.
Instead they attempted to observe the intensity of the fluorescence after a long
delay time when presumably the 2662 A excited species would, be near vibrationally
equilibrated and. its reactivity similar to that for a long wavelength excited SQ2
species. Since the pressure of SOa had to be such that vibrational equilibration
was ensured, the method is inherently inaccurate in establishing a zero pressure
lifetime for the excited state. However, the quenching rate constant so deter-
mined was in excellent accord with that which we found by direct long wavelength
excitation using a filtered Xe-flash lamp system, and we were thus led to accept
the lifetime go determined as a preliminary estimate of some merit. It is inter-
esting to note that the quenching rate constant so obtained is also in excellent
agreement with the recent estimate of Brus and McDonald for near vibrationally
equilibrated species; see Table XXVII. The unexpectedly long lifetime observed
directly by Brus and McDonald, about 500 usec, makes the Sidebottom et al lifetime
estimate meaningless in retrospect. One can be sure that an extensive loss of
excited singlet species through diffusion controlled processes must have occurred
at the lower pressures. In fact, it is this unusual lifetime observation which
requires a re-evaluation of the role of diffusional loss of singlets in the
steady-state experiments from which much of the evidence for the non-radiative
decay was obtained.
First, one must consider to what extent the presence of the two emitting
singlets is expected to complicate the Stern-Volmer plots of l/$f versus [802]
found in stead-state experiments. From the rate parameters estimated by Brus
and McDonald and their estimate that the concentration ratio of the short-lived
to long-lived singlet is « 0.28 immediately following the laser excitation, we
-------
may estimate the relative intensities of the fluorescence from the two states
shown in Table XXVIII. Obviously no more than a few percent of the fluorescence
Table XXVIII. Estimated ratio of intensity of the fluorescence from the
short-lived S02 singlet species (lf) to that for the long-
lived species (l^) in steady-state experiments6)
Wavelength (A) />so (jj) (/f )/(/fO
2617
'
2856.7
3043.9
150
50
5
1
0
150
50
5
1
0
150
50
5
1
0
0.011
0.013
0.029
0.081
0.28
0.036
0.037
0.051
0.098
0.28
0.17
0.17
0.17
0.20
0.28
n OO
a) Rate data of Brus and McDonald used for these estimates .
light which we observed in our steady-state experiments down to pressures of 1 M,
would be that from the short-lived species for experiments in the 2650-2900 range.
Clearly the steady-state experiments should apply well to the long-lived state
for these conditions. Arguments which we have enunciated elsewhere-^5 show that
diffusional loss of the singlet of lifetime kO p,sec would not be very significant
for our small cell in experiments down to about 10 ^ pressure. However, at life-
times of 500 n$ec, there can be no question that diffusional loss of singlets
must be important in a small cell at the lower pressures which we employed. Thus,
the limiting quantum yield of emission which we measured may be greatly under-
estimated in experiments at 3050 A, while they should be reasonably good for
experiments at 2650 A. For 2650 A excitation the zero pressure limit gives 0£ =
0.55 ± 0.^3 from the data of Rao et al^3 extrapolated from data at pressures of
50 (j, and above, and O.Ul ± 0.2U from Sidebottom et a!123j extrapolated to zero
pressure from data in the range 4l-l4l p>. If the Brus and McDonald estimate of
the initial population of the two states is correct and there is little generation
of one state from the other as they suggest, then we would expect a directly
measured limiting quantum yield of 0.78 from these runs in which no significant
number of quanta of the short-lived state would have been detected. In view of
the very large extrapolation necessary to derive 0^ from the relatively high
pressure data, the observed estimates agree reasonably well with this expectation.
In other experiments which we will describe in detail elsewhere^?, we have
measured the zero pressure quantum yield of the long-lived emission from SQ2 ex-
cited in a laser pulse at 2662 A using benzene as a fluorescence actinometer.
We estimate for conditions for which diffusional loss of singlets, would be unim-
portant that 0S = 0.46 ± O.lU. Certainly there is no evidence of a large unex-
plained inefficiency in light emission from any of the studies in the 2650 A
region.
89
-------
122
'Table XXIX. Estimation of 0£ from lifetime data" of Brus and McDonald
(ki + k2)/k3 data of Calvert et al12?'126'2? and Mettee et
and
Wavelength
(A)
2650
2750
2875
2963
3020
(8 mole"1 sec"1) (S mole"1)
X 10"10 X 10-*
3.1 ±0.2 1.9 ±0.4
1.9 ±0.8
5;3 ± 0.1
2.9 ± 0.3
4.1 ±0.2 3.4 ±0.3
3.4 ± 0.3
6.0 ±1.1
5.6 ± 0.4 5.6 ± 0.6
3.9 ± 0.9
6.6 ± 0.3
10.9 ± 3.2
6.6 ± 0.1 10.2 ± 1.0
8.1 ± 1.6
13.8 ±0.4
20.1 ± 1.0
7.2 ±0.2 10.7 ± 0.8 W
12.0±3.3b>
15.9 ± 0.9 b)
23.4 ± 3.0 b>
Msec"1) J£-K + S
v in-* ^ '
x 10 x lo-4
1.63 ±0.36 1.15*0.44
1.63 ±0.69
0.58 ± 0.04
1.07 ± 0.13
1.21 ±0.12 0.95 ±0.06
1.21 ±0.12
0.68 ±0.13
1.00 ±0.1 3 0.71 ±0.06
1.44 ±0.35
0.85 ± 0.07
0.51 ±0.15
0.65 ± 0.06 0.53 ± 0.02
0.81 ± 0.16
0.48 ± 0.02
0.33 ± 0.02
0.57 ± 0.04 0.41 ± 0.07
0.51 ±0.15
0.38 ± 0.03
0.26 ± 0.04
• fe,**,...,
1.41 ±0.60
1.41 ±0.90
0.50 ±0.19
0.93 ± 0.37
1.27 ±0.14
1.27 ±0.14
0.72 ±0.14
1.41 ± 0.22
.2.03 ±0.52
1.20 ±0.14
0.72 ± 0.22
1.23 ±0.12
1.53 ±0.31
0.91 ± 0.05
0.62 ± 0.04
1.39 ± 0.26
1.24 ±0.41
0.93 ± 0.17
0.65 ±0.14
Reference
source of
34
35
23
123
34
35
123
34
'35
18
123
-34
•05
23
123
34
35
23
123
a) *i -(• fcj and k3 +• £4 + fcs values were obtained by interpolation of.the rate constant-wavejength data of Brus and McDonald
b) These experimentally measured slopes were corrected to the "true" (*] + k2)/k3 values for the long-lived singlet species by mul-
tiplying by 1.17, since the measured quantum yields for this wavelength are expected to be 1.17-times the fluorescence quan-
tum yield for the long-lived singlet species at the pressures employed here; see table XXVUI.
122
On the oth'er hand, steady-state quantum yield studies carried out at 3020 A
gave 0£ = 0.061 ± 0.12 and 0.069 ± 0.08l, in the studies of Rao et a!23, and
Sidebottom et al12^, respectively. We are now of the opinion that most or all
of the apparent lowering of the 0^ with wavelength increase may be an artifact
of the diffusional loss of the excited singlet as a direct result of the then
unrecognized very long lifetime of the state formed at long wavelengths. We
may use our own and Mettee's steady-state data- to test this hypothesis in a
rather quantitative fashion which is independdent of diffusional problems.
The Stern-Volmer plots derived from the high pressure, stead-state data
should have no appreciable influence from diffusional loss of singlets. However,
in view of the well established, result that a vibrational cascade mechanism
operates for the quenching of the long-lived singlet formed at the short wave-
lengths122J-'-23)98j it is somewhat surprising at first consideration that no
significant curvature has been seen in Stern-Volmer plots of l/0f versus [SOs]
in any of the previous studies. This probably results from two main factors.
In steady-state experiments most of the light quanta which are detected are
emitted during the first few lifetimes of the emitting species, and during this
time there is a very little vibrational relaxation of the excited 80s species;
this is evident in the near linear character if theol/T versus [SOa] plot during
the first lifetime periods for S02 excited at 2662 A9o. Secondly, the failure of
a plot of [S02]/(lf)^ versus [SOg] to maintain linearity (where (lf)&\ is the
intensity of a small band of wavelengths near the exciting light) is much more
90
-------
readily detected experimentally^" than curvature in a Stern-Volmer plot of 1/0 f
versus [SOg], since vibrational relaxation does not preclude the ultimate emission
of a quantum at the longer wavelength. In any case, all previous studies give
good Stern-Volmer plots whose slopes from the high pressure data should have no
appreciable influence from diffusional loss of singlets. These slopes should give
reasonable estimates of (kj. + ke)/ks for the long-lived species. These are shown
in Table XXIX for several wavelength ranges along with estimates for kx + k2 and
k3 + k4 + ks = I/TO for the long-lived S02 state which we have interpolated from
the results of Brus and McDonald-1^, ye can estimate ks from our (k-L + k2)/k3
data and directly measured values of k'i + k2. Now the ratio of this value to the
reciprocal of the limiting zero pressure lifetime of the long-lived SQ2 state
(l/To = ks + k^. + ks) should give the limiting quantum yield of fluorescence of
the long-lived state at zero pressure: ks/(k3 + k4 + ks) = 0£. Note in Table XXIX
that the data from both the Mettee group and our own group show no.-'significant
trend of the data with X, and a value near unity is certainly not inconsistent
with these results.
All of the present data support the contention made by Mettee several years
ago and reinforced recently by Brus and McDonald, that radiationless decay in
excited S02 singlets is unimportant. It now appears that any small inefficiency
in the emission of light quanta for S02 excited at short wavelengths is largely
due to the population of a second very short-lived state which for all of our
conditions is quenched effectively even at the lowest pressures we employed.
The very low efficiency of quanta production which we observed at long wavelengths
appears to have its origin not only in the second easily quenched state, but more
importantly, in the diffusional loss of the singlet which has a 20-fold greater
lifetime than was expected previously for these conditions.
91
-------
PART II.
SOME THERMAL REACTIONS OF POSSIBLE SIC2TIFICANCE ON
THE SUNLIGHT-IRRADIATED, POLLUTED TROPOSPHERE.
The interaction of the free radicals HO, HOa. RO (alkoxy), ROa (alkylperoxy),
N03, and other reactive species such as 0(3P), O^D), NaOs, 03} etc., with im-
portant pollutant molecules such as SOa, NO, NOa, BH, etc., may be of significant
rates 'in the sunlight-irradiated, polluted troposphere. The recent computer
simulation studies of the chemistry of the polluted atmosphere carried out by
our group1^?, have pointed toward the possible importance of several thermal
reaction systems. In this section we discuss these results; some of these studies
have been completed and published during the period, while studies on others re-
main incomplete at this report writing. We will consider these systems in this
section.
II-A. A,Kinetic Study of the SOp-Or,. SOg-NO^, and the SOg-NgO^
Reactions1^".
In addition to the heterogeneous pathways which are often invoked to rational-
ize SOa conversion to HaS04 and.sulfate salts in the atmosphere, there are a
number of homogeneous elementary reaction paths which may control SOa removal in
the sunlight irradiated NOx-S02-hydrocarbon-polluted atmospheres93. The present
study provides the first direct kinetic data related to two of these potentially
important reactions 1 and 2:
S02 + N03 * S03 + N02 (1)
SO2 -f S2Os •» SO3 + N2Oi, (2)
-------
In this work NOa symbolizes the symmetrical nitrogen trioxide species. A re-
action analogous to 1 has been considered to be an important step in the NQjj-
catalyzed thermal oxidation of SOa at high teraperatures^°», 5^; presumably the .
unsymmetrical, peroxy-bonded nitrogen trioxide species, 0-N-O-O, was considered.
the reactant in these systems, since this is the primary species which is ex-
pected to form from the bimolecular interaction of molecular oxygen and nitric
oxide for the conditions employed in the previous work.
The enthalpy changes associated with reactions 1 and 2 (AHi° = -33;
-2U kcal mole"1 at 25° c) favor the consideration of these reactions as potentially
important among the atmospheric SOa removal processes. Indeed reaction 1 has been
suggested as an important source of SOg oxidation in irradiated NQx-SOa-hydrocarbon
mixtures in smog chamber studies^-3^-. We report in this work the first direct ex-
perimental evidence concerning the magnitude of the rate constants for reactions
1 and 2.
We have studied mixtures of SO& and NgOs with and without added Q3, and have
employed direct infrared spectroscopic detection of SOa1?2. The mechanisms and
kinetics of the N205 decomposition (reactions 3, k, and 6) and of the Na(%-
catalyzed decomposition of QS (reactions 3-6) are well understood today, and
reasonably accurate estimates of the rate constants for these reactions are
available^-33.
N2°5 * "O3 + N02 (3)
N02 + N03 -»• N205 (4)
N02 + 03 + N03 + 02 (5)
N03 + N03 •» 2N02 + 02 (6)
Our recent results confirm the conclusions of previous workers that the rate of
the homogeneous gas phase oxidation of SC>2 by Op, 03 + S02 -» 0& + S03) is very
slow at room temperature (k < 5 x 10~3 &. mole" sec"1) ^-52, Therefore it is pos-
sible, at least in theory, to estimate the rate constants for the reactions 1 and
2 from a kinetic study of the N205-S02 and NaOs-SOa-Os mixtures. The success of
our method depends upon the feasibility of the direct, quantitative, infrared
spectroscopic measurement of gaseous S03 concentrations. We have been able to
follow SOa quantitatively in a CaFa-windowed cell and a Pyrex vacuum system, free
of mercury, stopcock lubricants, and other materials reactive to SOs, and which
were made essentially water-free through prolonged bakeout at 100° C and evacu-
ation^2.
II-A —1 . Experimental
Synthesis o
was prepared by the dehydration of HN03 in an 03- contain ing atmosphere
by a procedure modified from that of Grunhut, et al-^ to insure the complete
removal of water. The method appears to be one which would be of general use to
atmospheric scientists .so it will be described here in some detail. The method
of dehydration of HN03 for NsOs preparation was preferred to that of 03 oxidation
of N204 since it seemed' to offer less danger of explosion^-35 . ^he apparatus is
shown in Figure ^>0. Fisher-Porter Teflon valves (U or 6 mm) were used throughout.
Silicone grease was used on standard tapers in the J-neck flasks FI and f2. Other
demountable joints were Fisher-Porter Solve-Seal connectors. Prior to use the
93
-------
apparatus was dried. .The trap Ti which contained P2Q5 on glass wool, and the
Dewar trap T2, were both outgassed for a 2k hr period at 100° C in a vacuum
(p < 10~5 Torr); the vacuum was maintained until the system,, was purged with an
03-02 mixture following HN03 preparation. The U-tubes, Oi and U2, filled with
6-16 mesh silica gel, were dried in a vacuum at" 150-180° for 2 hrs . , then stored.
in a drying oven with the remainder of the apparatus until assemblage just
before the start of the synthesis. During HW03 preparation, traps Ui -and U2
were kept at -7&° C to remove any traces of EsQ in the 02 and N2 gases.
HN03 was formed at 70-100° C from H^S04 (25 cc, DuPont Reagent) placed in
flask FI and RaN03 (7-9 g. Mallinkrodt Analytical Reagent) added by way of
the bulb Bj.. The HW03 vapors were transferred from FI to F2 with a stream of
N2 gas (Burdette extra dry) while flask F2 was cooled to .-78° C. During this
step, exposure of the other parts of the apparatus to HN03 was avoided, by
closing the appropriate valves. Abo.ut 10 cc of pure HN03 liquid was formed.
-The system was then purged with an 03-02 mixture which was generated from 02
(Burdette extra dry) by passing it through a high voltage discharge D. In
this and the two succeeding steps the valve Vj. to vessel FI and the nitrogen
vent V2 were closed, and the U-tube U3 was kept at -78° C to avoid back diffusion
of water vapor into the system. The Og-Os mixture for our conditions contained
2.$ 03 as shown by chemical analysis1?".
With HN03 at -78? C in vessel F2 and the OQ-QZ flow continued, the dehy-
dration of the HN03 was accomplished using P^OS (21 g, Mallinkrodt Analytical
Reagent). This was" added by upending the 125 ml bulb ¥& attached to F2 and
loosening the pinch clamp on the Tygon tubing connector. The reaction mixture
was warmed gradually to 100° C over 'a period of 1 hr; as the viscosity of the
mixture decreased sufficiently, the mixture was stirred magnetically. Dehy-
dration was further assured by transporting the gaseous W205 product by means
of the 02-03 stream through the trap TI filled with P205 powder suspended on
glass wool. The condensables were trapped on a U8 mm coldfinger maintained at
-78? C in the Dewar trap T2. PaOs was also placed in the bottom of this trap.
The condensed product N205 at -78° C was a white, fibrous solid. The 02-03 gas
flow was continued for an additional k hours after F2 had cooled to room tempera-
ture. Then the Dewar trap T2 was connected to a bakeable vacuum line1? , an$
the line and connecting tubing were outgassed at 100° C for several hours before
degassing the sample. When the line had returned to room temperature, trap T2
was evacuated while maintaining the coldfinger at -78° C, and then it was closed
off. Next, the bottom of the Dewar was cooled with liquid nitrogen and the Dry
ice-acetone mixture was removed from the central coldfinger chamber of T2. The
product condensed onto the P20s at the bottom of the Dewar, it was again out-
gassed, then recondensed onto the coldfinger. The sample of NgOs was then
allowed to warm to -55° C and any volatile fraction was pumped away. The fraction
of the product which vaporized in the range -J>0 to -20° C was transferred over a
1 hour period to a carefully baked and. dried storage Dewar, the coldfinger of
which was kept at -78° C.. This Dewar was similar to T2 shown in Figure J>0 except
it had only one side arm (demountable) near the top, in order to allow cooling
of a much larger portion of the exterior with a Dry ice-acetone bath. P20s was
also placed at the bottom of this trap. The product which remained in trap T2
at -20° C was discarded. The N20s product was stored at -78° C. The preparer
of N205 should be cautioned to store it either on a coldfinger or in a very thin
film of low mechanical strength on the inner walls of a bulb, since N20s solid
has a larger coefficient of thermal expansion than glass and fracture of glass
vessels may occur on warming^?.
-------
Fiq.aO. Diagram of the apparatus used to prepare N205i see the text for
I the explanation of its use.- '• • •&-,-. •••
The Absorption Spectra of
The gas phase infrared spectrum of the NgOs product matched closely that
given by Pierson, et al^-57. The visible-ultraviolet spectrum of IfeOs vapor was
published twenty-seven years ago by Jones and Wulf^o. Their sample of "^gOs"
intentionally contained either 03 or N02, presumably to suppress the decomposition
of the NgOs molecules. It is not clear how correction for these impurities was
accomplished. We have redetermined the absorption spectrum of IfeOs and this is
shown in Figure jl- During the time required for introduction and spectral
measurements of the N20s, some decomposition into NQa and 02 occurred. The low
resolution spectrum of NgOs given in Figure 31 was calculated from the measured
scans of N20s(g) at 27 Torr in a 10 cm cell by subtracting absorption due to the
small N02 itnpurity product. The molar extinction coefficient e versus wave-
length is shown where e = [loglo (lo/l)]/[N2Os].£, 4. mole"1 cm'1.
Procedures in Kinetic Experiments with
S02, and
The reactions were carried out in a CaF2-windowed cell which, prior to each
experiment, was preconditioned with pure samples of 80s to insure the stability
and the detectability of small quantities of the potential S03 product-^2. LOW
pressures (p ^ 0.05 Torr) of 80s could be measured easily through infrared
analysis, so the appearance or the nonappearance of 80s absorption was a fairly
sensitive test for the occurrence of reactions 1 or 2. Pressures of the in-
dividual components were measured in calibrated volumes on a vacuum line and
mixed using an all glass, electromagnetically operated stirring pump. There
was an interval of about 20-30 min between the start of the mixing and the
start of the ir scan. The total cell pressure was measured again prior to
removing the cell from the vacuum line and placing it into the spectrophotometer.
II-A—2. Discussion _of the Results
Estimates £f the Rate Constants for the N0a-S0g and N205-S02 Reactions.
Several mixtures were prepared and the ir spectra followed as a function of
time; in experiments at room temperature (~ J/0°C) the initial reactant pressures
(Torr) were as follows: l) S02, 1.96; N205,. 11.9; 2) S02, 8-5; N2C>5, 2.5; 03,
10.6; 02, ^50; 3) S02, l6.7; N205, 7.62; 03, 35-6. No absorption by S03 and no
95
-------
60
u
o
350
WAVELENGTH ,
250
Fig.31. The molar extinction coefficient e for N205 vapor; c = Ilog10(Io/I)]/
[H205]4, 1. mole"1 oar1.
decrease in SOs absolution could be detected in run 1 after a period of several
hours. Also no S03 was detected in run 2. After 20 min of contact of the re-
actant's in run J>, the cell pressure had increased by 10.0 Torr (NgOs catalysed
03 decomposition), and no S03 formation could be detected. Obviously the rates
of the reactions 1 and 2 are very slow. We may use these data to estimate the
upper limits for the kx and k2 values.
From the detection limits of the methods employed we know that the maximum
amount of S03 present in any of the experiments was less than 0.2 Torr (a gen-
erous upper limit). We may make the reasonable assumption that reactions 1 and
2 are second order elementary steps. Furthermore since S03 formation was un-
detectably low, it may be assumed that S02 does not significantly perturb the
steady state' concentration of N03, [N03]ss. W205 is a catalyst for the 03 de-
composition in the N205-03 system, so that [N205] is a constant as long as
there is a significant concentration of 03 in the system. The kinetic equations
which may be derived readily from the reaction sequence 1-6 outlined for these
conditions and which apply to the kx evaluation in the N205-03-S02 systems are
the following:
d[S03]/dt = ki[S02HN03]
[S02jt = [S02]0 - tS03]t
= K8stN205jV3{03jl/3
1/3 _l
(7)
(8)
(9)
(10)
I/ 2/ 1/
where K^ = (k3k5/2k4ks)^ and k« = (l/2)(k3k5/k4)/3 (2ke)/3. Relations 8-10
may be substituted in 7 to obtain the differential equation which relates [S03]
and the reactant concentrations as a function of time:
d[S03]
[S02)0- [SO 3]
(kiKss[N205]1/3t03]01
(11)
-------
Integration of function 11 gives the desired function from which kj. can be
estimated:
ln{[S02]0/([S02]0 - [S03]t)}
(12)
Kss(tN205]1/3[03]01/3t - i k'[N205Jt2)
For the temperature of 30° at which these experiments were performed, k1 = 7.3 x
10~3 (£./mole)V3 sec"1 and Kgs = 3 x 10~4 (mole/A.)^. For the conditions of
experiment J outlined above the appropriate concentrations (mole/J&.) are:
[N20s] = 4-03 x 10-4; [03]0 = 1.88 x 1CT3; [O3]t = 1200 sec = 8-2°" x 1°~4;
[SOalo = 8.84 x 10-4; [S03]t _ -^00 sec ~ lt06 x 10 * Substituting these
values in relation 12 we derive a lower limit for kj.:
In a similar fashion we may estimate a maximum limit for ka:
d[S03]/dt = k2[S02][N205] (13)
Substituting relation 8 in 13 and integrating between the limits of t = 0 and
t relation lU is obtained:
n>([so2]{/(tso2]0 - [so3lt)} (14)
*2 = . [N205)t
Using the values applicable to run 3 in relation Ih we find:
k2 s 2.5 x 10-2 1. mole-'sec-1
It should be noted that the pressure change observed in the experiment 3
is entirely consistent with that expected simply on the basis of the NgOs-
catalyzed decomposition of Os, reactions 3-6 5 and the known rate constants.
The integrated rate law for [03] as a function of time in this system is had
by cubing both sides of relation 10:
l03Jt = ([03]0l/3 - \ k'lHzOslZ^t^ (15)
Substituting the values applicable for run 3 for t = 1200 sec, we expect [03] =
7.67 x 10~4 M or PQ = 1^.5 Torr. Within the uncertainty in the value of the
rate constant functfon k' and the reactant concentrations, this is in reasonable
accord with that estimated for the observed pressure change of 10.0 Torr: PQ =
35.6 - 2(10.0) = 15.6 Torr at 1200 sec. It can be seen that this result provides
further evidence of the unimportance of reactions 1 and 2 for these conditions.
If 1 and/or 2 occurred to a measurable extent then the observed increase in pres-
sure would be significantly less than that calculated from the ^0^ catalysis of
the 03 decomposition in run 3- This is the expected result as well if these
reactions occurred measurably and a non-volatile solid adduct formed by reaction
between SQ3 product and NOa (See the last section of the discussion.) or the
nonvolatile product H2S04 resulted from the interaction of the initial product
S03 with impurity HN03 or HaO. There was no evidence of the formation of a
solid in these experiments although an amount less than that formed from 0.5
Torr of SOs would have been observed. Also there were none of the changes in
the ir transmission of the CaFa windows of the cell which appear characteristically
if HaS04 formation occurs; changes from HeS04 action corresponding to as little
as 0.2 Torr of S03 would have been seen. Thus all of the evidence at hand points
to the unimportance of the reactions 1 and 2 at 30 C.
97
-------
If we employ £he maximum values for the rate constants kx and k2 derived
here, it can be shown in computer simulation of the complex chemical changes
which occur in the atmospher.e that reactions 1 and 2 cannot make a significant
contribution to the rate of Sp2 removal which occurs in sunlight -irradiated
NOx-hydrocarbon-S02 polluted atmospheres93. other alternatives "must be con-
sidered to rationalize the smog chamber data in which these reactions have been
invoked previous
The Predicted Influence of NgOs Photode compos it ion cm the Levels £f N2C% and
N03 ^in the Sunlight-Irradiated Polluted Urban Atmosphere .
The extinction data for the near ultraviolet absorption by gaseous %(%,
presented in Figure 51. and the estimated solar irradiance data of Leighton^39
for typical atmospheric conditions, may be combined to estimate the rate of
sunlight absorption by N205 . These data for the solar zenith angle of kO° , give
ka = U.3 x 10~3 min"1, where the rate of Sunlight absorption = ka[N2Os]. This
rate is the maximum possible rate of the photodecomposition of NaOg in the
sunlight- irradiated lower atmosphere (z = hO° ) which presumably would occur
by reactions 16 and/or 17:
N205 + hv •* NC>3 + N02 (16)
N205 + hv * N20i, + O (17)
Compare the magnitude of the photochemical rate with that for the first order
thermal decomposition of N20s at 298° C in reaction 3; d[N205]/dt = k3[N2Os];
k3 = 1.5 x 101 min"1. It is clear that the maximum rate of photodecomposition
of N20s in the solar irradiated polluted atmosphere is only 1/3500 of the rate
of its thermal dissociation. Thus in the detailed theoretical modeling of the
rates of chemical changes in urban atmospheres,-^ the occurrence of reactions
16 and 17 may be neglected; they have a negligible effect on the levels of N03
and N20s which will develop in these atmospheres .
The Reaction Between Gaseous S0a and NOg.
Preliminary results which we have obtained show that a very dry mixture of
80s and N02 gases reacts rapidly to form a relatively nonvolatile white solid.
Thus when 7-8 Torr of S03 and 9-15 Torr of NOg ^ were allowed to come together
in the gas phase, a white solid coated the walls of the vessel immediately (in
less than about 2 sec), and a residual pressure of 1.82 Torr was observed. If
it is assumed that S03 reacts with NQs i-n a 1:1 ratio to form a nonvolatile
adduct and that the N204 2 2N02 equilibrium is maintained, then we calculate
that the residual pressure in the experiment described would be 1.80 Torr. equal
within the experimental error to that measured experimentally. The material so
formed was removed by raising the temperature to 100 C and evacuating the vessel.
No detectable solid remained after 1 hour. The stoichiometry . vapor pressure,
and other properties of this interesting compound are now under detailed study
in this laboratory. However the preliminary data reported here suggest that
the white solid is a 1:1 adduct with a relatively low vapor pressure at 30 C.
We could find no reference in the literature to the gas phase S03-N02
reaction observed here. However the white solid may be the same as that
observed by Urone, Schroeder, and Miller1^! and Paul, Arora, and Malhotral^2
in more complicated systems. Urone, et al. , irradiated dilute gaseous mixtures
of S02 and N02 in dry air. Chemical analysis of the white solid which formed
98
-------
under their reasonably anhydrous conditions had the simplest formula ,
the composition corresponding to a 1:1 adduct of S03 and NOa suggested for the
compound formed in the present work. It is reasonable to expect that this com-
pound could arise In the Urone system through the same S03-N02 reaction we
observed in -this work, following the generation of S03 in the sequence, NOa +
hv -» 0 + NO; 0 + S02 + M -* S03 + M. Paul, et al. ,lij-2 isolated a white solid, of
composition and molecular weight in nitrobenzene corresponding to rfeC^'SSOs,
from the reaction between NgC^ an^ S03 in liquid- SOg- Some suggestions con-
cerning the chemical structure of the white solid can be had from that proposed
for solids of this and similar compositions which have been reported. Thus
nitrosyl pyrosulfate1^?, IfeSaOs, and nitryl pyrosulfate1^, N2S2On, have been
prepared by a number of workers using a variety of reactants. From the limited
data at hand it is a reasonable hypothesis to suggest that the white solid
formed here may be nitrosyl nitryl pyrosulfate,
Since there is such a paucity of data concerning S03 reaction rates, in-
cluding the rate of the often invoked homogeneous reaction,. S03 •*• HgO -* HgS04,
it is not possible to say at this time how important the S03-N02 reactions
observed here might be relative to other 80s reactions in the polluted atmos-
phere. The significance of the S03-NC>2 adduct as a "participant in urban aerosol
formation is an intriguing possibility which cannot be evaluated without further
quantitative work.
II-B* The Reaction of Qa with Perfluorinated Polyolef ins
The use of Teflons in chemical reaction systems is widespread today. For
example, the Teflon valves and stopcocks have become increasingly popular for
use in the control of reactive gases. Environmental scientists commonly employ
Teflon tubing in the transport of ambient air samples to the detection equipment
for specific pollutants. The Teflons are often used as protective coatings for
metals in chemical reactors.-. .and. environmental reaction chambers. In fact large
bags constructed of sheets of the perfluorinated polymers are sometimes employed
directly as environmental chambers for simulated smog studies. The Teflon,
fluorocarbon polymers offer the important advantage of unreactivity toward. a
great variety of reagents-"-^ . it was therefore somewhat surprising to find in
the experiments described here that exposure of Teflon substrates to ozone led
to gaseous, infrared absorbing species. The results reported in this paper
serve to identify the resulting products and to provide some insight into the
reaction mechanism involved in their formation. Workers who utilize Teflon
fixtures in ozone containing atmospheres should be cognizant of the occurrence
of this reaction and the possible contamination which will result from this
practice.
II-B-1. Experimental
Apparatus
The Pyrex vacuum line was designed to permit thorough elimination of water
to facilitate the study of gaseous mixtures containing S03, S02, 03, and 02^52.
Briefly, it was grease and mercury free and was enclosed in an oven to permit
extensive baking (100° c), while pumping, of all components up to the N2(^) trap-
preceding the diffusion pump. Rotaflo Teflon stopcocks, whose design eliminates
0-rings of any kind, were used throughout. Pressures during line outgassing
99
-------
were monitored with a CVC ionization gauge. Typically a minimum pressure of less
than 5 x 10~6 Torr was attainable. Beactant and'product pressures were measured'
with a spiral manometer. A flow discharge system was used to generate 03 from
QS.. Traps filled with silica gel were located immediately prior to and im-
mediately after the discharge. The first one was used to dry incoming 02- the-
second trap removed 03 from the 03-02 mixturel^o. A U-tube completed the
synthesis loop; as a cold zone, its function was to prevent atmospheric water
from diffusing back through the outgoing 0^ stream. During 03 synthesis both
traps and the final U-tube were kept at -70 C except while desorbing 03.
Reactant and product concentrations were monitored with infrared spectros-
copy. The ir cell (designed for crossbeam photolysis) was constructed from
U8 nun Pyrex tubing. Infrared windows were of CaFa (^9-5 EM diameter, 6 mm
thickness); windows for photolysis light were Pyrex and were fused to the cell
body. The path length of the ir beam through the cell was 9.5 cm. Teflon rings
(gaskets), machined from a 2 in. diameter cylindrical bar of polytetrafluoro-
ethylene (TFE), were placed between the Pyrex body and CaF2 windows. The CaFa
windows were tightly held to the cell by means of brass end rings connected
together with threaded rods. Several coats of Glyptal were applied over ends
of the Pyrex .cell. Teflon gaskets, and edges of CaFg windows. The cell was
baked for eight hours at 100° C between successive coats. This cell was fixed
in the analysis beam of a Perkin-Elmer Model 21 Infrared Spectrophotometer
equipped with an ordinate scale expander.
In many experiments the ir cell was also the reaction vessel. For experi-
ments with powdered Teflon and FEP film, the reaction cell was a round-bottom
cylinder (volume 33-3 cm3) formed from 20 mm Pyrex tubing.
High resolution mass spectra were taken with an Associated Electric In-
dustries, Ltd., MS-9 instrument.
Materials
Three different Teflon samples were used. The first and most extensively
used was chosen unintentionally as an ozone reactant in that it was the gasket
material in an ir cell used in following gas phase reactions in 03-containing
mixtures. Assuming a smooth, nonporous surface for the gasket, we estimate
that its area was l.J x 10~3 m2. The second sample was powdered Johns-Manville
Chromosorb T, hO/6d mesh, screened from Teflon 6. As specified by the supplier,
the surface area was 7-8 tr^/g; the quantity used gave an area of 30-3 m2-
Finally a sample of FEP film of 0.005 in thickness was employed as reactant;
this material was supplied by the Plastics Department of the E. I. duPont de
Nemours Co. Assuming the sheet to be perfectly smooth, we estimate the area of
this sample to be O.l8 m2. To facilitate packing into the reaction cell, this
film was cut in strips 6 cm x 21.5 cm- All Teflon samples were thoroughly out-
gassed at pressures less than 5 x 10~6 Torr at 100° C before experimental use.
Oxygen was Burdette extra dry. Ozone was synthesized from it using the
apparatus described, previously. 03 collected in the silica gel trap was
separated from 0% impurity by gradually warming the trap to room temperature
while transferring 03 to a storage trap held at -196° C (while pumping). A
one-liter expansion bulb joined the storage trap, and care was taken to avoid
buildup of 03 which would give a pressure above 100 Torr on complete vapori-
zation in order to lower the danger of explosion '.
100
-------
Procedures.
The observation of an infrared absorbing product from 03 was made first in
routine blank experiments carried out in connection with a study of 03 photolysis
in the presence and absence of Oa- Such blanks were usually performed as follows:
after the photolysis of 03 or 03-02 mixtures (X = 590 run), reaction components
were allowed to stand in the -ir cell (which was also the photolysis cell) for an
extended dark period. In other'experiments also involving the Teflon gaskets,
the cell was simply filled with 03 at a particular pressure and allowed to stand.
in the dark. The ir absorbances were of course recorded at appropriate times.
When powdered Teflon or PEP film was used, Os was taken at a desired pres-
sure in a known volume and then transferred to a coldfinger, cooled with MaC^),
on the reaction bulb. Cell pressures obtained on vaporization were calculated
from known volume relationships assuming ideal gas behavior. After a suitable
reaction time, the cell contents were expanded successively into a series of
cold zones (-196°C), first from the cell into a single trap, then from the cell
and trap into a second trap, and finally from the cell and two traps into a
multiple U-tube with three cold zones, each separated by a zone.kept at room
temperature. Total volume of this trap train was l£0 cm3. After a holding time
of one minute, the noncondensable gas (presumably 0^) was slowly removed by
pumping through these cold zones. Finally, the outgassed 03 and condensable
products were transferred to the ir cell, the total pressure was measured, and
appropriate absorbances were recorded. From an absorbance-pressure curve for 03,
the pressure of Oa was determined; subtraction from the total cell pressure gave
an estimate of the partial pressure of the products.
Products of the reaction were separated from 03 following condensation of
the entire 03-product mixture in a trap at -196°C. The liquid nitrogen bath was
replaced, with an QZ(&} bath, and 03 was pumped away leaving a white solid. In
order to accumulate a sufficient amount of products for mass spectral analysis,
several such separations were carried out and the products were collected and
stored, at room temperature, in a two liter bulb.
Ozone concentrations were monitored using the characteristic weak absorption
at 4.69 and U.75 M-; CaFa window absorption precluded use of the more intense
9.5 |i band. Instrument response as a function of 03 pressure was calibrated over
the range 0.3 to 120 Torr. For small concentrations the ordinate scale expansion
was used. Since neither the optical nor electronic system was adjusted on ex-
pansion, absorbances calculated from responses on individual scales were identical,
within the experimental error. Absorbances were often small (0.001-0.030 was the
usual range) so that multiple readings were made (xl, x5, xlO, up to x20) and the
results were averaged with each measurable signal receiving equal weight.
II-B—-2. Results and Discussion
When 03 was introduced into the infrared cell in the dark at room tempera-
ture its absorption at h.6<2 and H.75 M* decreased slowly with time as absorption
at several new bands appeared; see Figure 32. Bands at 5-10 and 5-17 M- were
used to monitor continuously the product concentrations in every run, while the
other bands were checked occasionally to see alterations which had occurred.
This absorption was not due to an impurity released from the cell components
since no absorption was detected over the entire accessible wavelength range
when the evacuated cell was closed off from the rest of the vacuum line for a
101
-------
1OO
ui
u
4-75 5-25 475 5-25 475 525
WAVELENGTH (microns)
Fig. 32. Infrared spectral changes used to monitor the reaction of ozone and
Teflon; (a) zero time, PQ = 96 Torr; (b) same mixture after 25,280 sec;
(c) same mixture after 63,200 sec; Teflon sample employed in this, case was
the gaskets of the infrared cell; note the growth of the product bands at
5-10 and 5-lTli as the 03-Teflon contact time lengthens.
time comparable to that used in runs. 03 attack on the CaFg windows was further
excluded as a source of the products; when 03 was added to a cell containing
CaF2 powder an acceleration in the 03 decomposition reaction resulted, but no
absorption bands other than those due to 03 were detected. A more complete
spectrum of the product mixture that was retained after Q3 removal is presented
in Figure 33- In addition to the bands at 5-1° and 5-17 |Ji} absorption appeared
at ^.25, 8.0 and 9-75 M" This mixture was the accumulated product from eight
experiments in the ir cell. After each run the absorbances at 5-10 and 5-17 M-
were determined. When account was taken of the small deviations of the pressure
dependence of the absorbances from Beer's law, it was found that none of.the
product which absorbs at 5-10 and 5-17 M< was lost during a storage time of 2J days.
When 03 was allowed to act on powdered Teflon or FEP film, ir spectra of
products were obtained which were very similar to those given in Figure 33. These
results show that a reaction between Teflon and 03 gives rise to the gaseous
products having the observed absorptions. Teflon gasket material used in the
cell construction was the origin of the gaseous products in the first experiments
102
-------
100
80
60
40
20
3456789TO
W89ELEN6TH (microns)
Fig. S3.) Infrared spectrum of products resulting from the reaction of ozone
with Teflon; total pressure of the gaseous mixture, 7;T5 Twrr.
described above. The ratios of the total pressure (P^-) in Torr of products to
the absorbance (A) at 5-10 and 5-17 M- were the same, within the experimental
error, whether Teflon gaskets for the ir cell or powdered Teflon were used as
the reactant: P-t/As.icM- = 32-6 and Pt/As.iT^ = 28.0 for the Teflon gaskets, com-
pared to Pt/As-iop. =51.1 and Pt/A5.i7p, =28.6 for powdered Teflon.
The ratio of absorbances A4.25p,/A5.17|o, = 0.115 was obtained from the ac-
cumulated product mixture shown in Figure 33; this is typical of experiments
employing Teflon gaskets with reaction times averaging about 27 hours. At
longer times this ratio increased; thus after 60 hours, A4. 25^ /As. 17)0, values of
0.271 and 0.269 were found.
The ratio AQ.OH/AS.IT^ was equal to 1.23 ± 0.05 for products derived from
either the Teflon gaskets or powdered Teflon. For FEP film, the ratio was some-
what lower (1.05) and a very weak additional absorbance at about 8.85 ± was noted.
The analytical data suggest strongly that the product distributions are nearly
identical for the reaction of 03 with Teflon gaskets or with powdered Teflon
even though the Teflon samples had very different surface areas, reaction times,
and reaction vessels in the two cases.
Mass spectral analysis of the sample whose ir spectrum is given in Figure 33 }
gave peaks at m/e = 104, 85, 66, hf , k6 , ^5 , and Ml. The high resolution masses
of the 85 and 66 peaks were found to be: 8^.9727 and 65.9915, respectively; these
data show conclusively the peaks to be SiF3+ (84.9721) and CF20+ (65.9918), re-
spectively. Infrared peaks in the 2.60-2.80, 5.10-5-17, and 7-95-8-75 g regions
are attributed to CFaO;1^'1^ absorption at k. 20-4. 35 (J, is due to COa1^, and
that at 9.70-9.80 ^i is associated with SiF4-'-51. The high resolution mass spectra
confirm the SiF4, CFaO, and CQ2 assignments. Both the infrared and mass spectral
data indicate strongly that the product mixture is a three component system con-
sisting of CF20, 5^.6$; SiF4, 5-7^; and C02, 39-7%- The extinction coefficient
(base 10) for CFsO at 5.10^ may be estimated from our work to be 0.00^9 Torr"1
cm'1; this is somewhat different than that reported previously, 0.015 Torr"1 cm"1
by Saunders and Heicklen ' .
103
-------
The amounts of.CFaO and COs formed in the Os-Teflon reactions studied for
several different experimental conditions are summarized in Tables XXX and XXXI
in which the Teflon sample was the cell gasket material, Teflon powder, or FEP
film. Note .-that in the case .of the products of the FEP film, there is a somewhat
lower ratio of absorbances, AQ. en /As . ITJJ, , than obtained with the polytetrafluor-
ethylene samples, and the small new absorbance at. 8.85 |JL observed in.these experi-
ments, may be indicative of the formation of an additional fluorinated carbonyl
product, perhaps CF3CFO, in this case.
The present study demonstrates clearly that there is a reaction between
gaseous ozone and various solid Teflon materials. The major products of this
reaction are CF20 and COa- The rate of formation of the products (CFgO + C02)
is related to the ozone pressure and the surface area of the Teflon sample. See
Figure 5^- Ozone and some reactive site on or near the Teflon surface seem to
be the most probably reactants here. Carbonyl products are the normal result of
the ozonolysis of the olefinic hydrocarbons, so the formation of CF20 very likely
results from such a reaction involving terminal olefinic sites in the Teflon,
reactions 1, 2, and "$. The symbol ~ used in reactions 1, 2, 3, and elsewhere in
this work indicates attachment to the Teflon surface.
i i
03 -» ~CT • •' CFg (1)
I I
~CF '• CFa -»~CFCfe + CFS0 (2)
~CPO + CF2Ofe (3)
The source of the COa product is less evident. It has been pointed out that
COa is sometimes formed from CFaO by a slow, overall reaction k involving a glass
wall surface1^3.
2CF20 + Si(fe(glass wall) -» 2CCfe + SiF4 (*0
However reaction 4 does not seem to be important here for several reasons, (l)
The stoichiometry expected for reaction h, [C02]/[SiF4] ratio of 2, is not ob-
served here; at the shortest reaction times [COaJ/tSUT.^ = 7. (2) Since no
detectable loss of CFaO was observed over a period of 27 days, it is clear that
reaction h did not occur to any appreciable extent in the glass storage bulb
which was used. The exceptional stability of CFaO observed in this case is
undoubtedly the result of thorough elimination of water. (3) The fraction of
CF20 in the gaseous products was the same whether the reaction was carried out
using either a small Teflon surface area (gaskets only) and a relatively large
Byrex surface area and necessarily long reaction times, or a large Teflon sur-
face area and a relatively small Pyrex surface area with comparatively short
reaction time. It is unrealistic to assume that the glass walls of the two
vessels would provide equal fractions of destruction of CF^O and formation of COa
through reaction k under these very different conditions. It is more likely that
COa is formed by some other reactions in our present experiments with the very
dry cells. The appearance of SiF4 does imply, however, that a glass wall re-
action of some kind is involved. It seems reasonable to associate SiF4 and COa
formation with some other fragment than the CF20 which results from'the OQ attack
-------
TABLE XXX
o
vn
Rate Data for the Reaction of Gaseous Ozone with Teflon Gasket Material at Room Temperature; Rates Given
in the Last Column Have Been Normalized to 1.00 x 1020 Active Sites for Each Run.
Run Reactant Pressure, Run Time, 03 loss, CFgO + CCfe -Rg / Reactive R + R ^ ,
No. Torr hr molec x formed, (R ? R ) Sites at molic /sec, x
10-la molec. x U'2U "* Start Run, IQ-IS
1
2
3
1*
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
03
5-31
7.1*6
7.60
6.82
8.11
8.26
2.66
2.50
2.40
2.51*
10.31
9.W
10.12
11.17
11.07
1*5.6
19-75
80.6
127-5
9U. 6
111.1
115-3
93-0
Os
202
60.8
201
0
150
250
150
0
100
125
0
50
11*9
2l+9
399
0
0
0
0
0
0
0
0
1*8.1*
12.25
15-5
16.8
15.0
1*6.7
36.1*
15.0
16.7
1*5-5
lit. 3
ll*. 1
19-1
ll* .7
22.5
25-5
1*0.0
21.0
6.7
26.8
1*3-3
5-9
15.8
7.0
16.3
11*. 5
22.0
51-5
62.0
ll*. 5
2.03
0.53
0.1*5
0.60
0.68
l*-35
1.13
(0.08)
(0.08)
1.13
2.63
0.98
3-75
1.81
2.70
28.9
57-8
(0.1)
27-3
66.0
77-3
126.9
177-0
25.6
1*2.6
58.1+
52.6
101.6
116.8
20.6
KT18
1.78
0.80
0.92
0.85
0.89
2.29
0.55
0.32
0.31*
0.73
0.99
0.95
1.1*2
l.ll
1.69
5-68
7-99
2.38
1.82
6.38
8.33
1.81
l*-75
1.36
3-13
l*.26
5.50
8.1*5
8.65
2.05
l.ll*
0.66
0.1*9
0.71
0.76
1.90
2.05
(0.25)
(0.23)
1.51*
2.06
1.03
2.61*
1.63
1.60
5.09
7.23
(0.03)
15.0
10.3
9.3
70.1
37.3
18.8
13.6
13.7
9-6
12.0
13.5
10.0
x 10-is
11.6
11.6
11.6
11.6
11.6
11.5
1J..5
11.5
11.5
11.5
11.5
11.1*
11.1*
11.1+
11.1*
11.2
10.3
10.1
8.93
8.1*6
8.1U
7.72
6.20
0.89
1.56
1.1*2
1.22
1.1*2
1.18
0.37
0.51
0.5!*
0.1*3
1.66
1.61*
l.8o
2.00
1.83
5-53 .
3.05
7.1*5
10.5
6.27
10.2
9.03
6.1+3
-------
TABLE XXXI
Rate Data for the Reaction of Gaseous Ozone with Different Samples of Perfluor-
inated Polyolefins at Room Temperature
Run Pressure
Ho
a)
21*
25
26
27
28
29
b)
30
31
32
33
Ot . Torr
V J J J.U.L.L
Run Time,
sec.
03 Loss,
molec. x
10-18
CF20 + COa
formed
molec. x
ID'18
-RQ
*^
RCFaO. + RCOa
RCFaO + ^^COg'
molec. /sec. ,
Powdered Teflon Reactant:
53-2
11*. 7
69.5
99-2
35.0
J.Jk
1080
2000
1000
1000
l&OO
2000
7.13
2.11
7.68
6.33
3.07
(0.31,)
2.95
1.71
2.81
1*.21
3-16
1.23
2.1*2
1.23
2.73
1.50
0.97
(0.28)
27.3
8.56
28.1
1*2.1
17-5
6.15
FEP Film Reactant:
96.2
1*5.2
10.0
15.1
6300
1011*0
11700
7080
(50.5 )a
9-78
(0-38)
2.02
7-78
8.12
2.52
2.50
(6.1*9)
1.20
(0.15)
o.8l
12-3
8.0
2.16
3.51*
The high Q3 consumption in Run 50 probably resulted from sample contamination
during handling and cutting of the film into strips. After this first expo-
sure, Oa loss was much decreased; see Run Jl.
on Teflon. In line with the accepted mechanism of ozonolysis in solution-*-!^ the
initial ozonide product of reaction 1 will decompose by the two routes, 2 and 3>
which generate the so-called zwitterionic species, ~CFC>2 and CF202, in addition
to a carbonyl compound. In theory there are a number of possible exothermic
reactions in which the zwitterionic species may decay. For example they could
decompose ozone by reactions 5 and 6.
CF202 + 03 -» CF20 +202 (5 )
-CFQa + 03 -» ~CFO + 2C& (6)
However, other information to. be considered later suggests that these reactions
may not be important here. The alternative reactions with 02 to form ozone are
energetically possible in our experiments with added 02.
Os -» CF20 + 03
-CFOs. + Oa + ~CFO -i- 03 W
However, reactions 7 and 8 are probably unimportant here since the rates of CF20
formation and ozone loss are insensitive to the concentration of 02 present in
the reaction mixture; compare runs h and 5, 8 and 9, and 11 and 12 of Table XXX.
Some evidence concerning possible COa-forming reactions and the fate of the
zwitterions can, be had by a consideration' of the reaction •stoichiometry. Con-
sider the data of the Tables XXX and XXXI. Teflon is the only source of carbon
in the gaseous products. For ozone pressures varying from 2.4 to' 11.2 Torr
(runs 1-15; Table XXX) in experiments with Teflon gaskets, the average ratio of
106 -
-------
C B ' A
45
o
CM
O 8
O
(T
+
O
a.
30
15
i.a
0.8
0.4
16 24 32
[oJ.molec. cm'3,x IO'17
40
Fig-S4. Rate of Formation of CF20 plus COa products of the 03-Teflon reaction
versus the 03 concentration: curve A, Teflon gaskets of the ir cell was the
reactant in this case; curve B, Teflon powder reactant; curve C, FEP film
reactant.
molecules of ozone lost to those of COg plus CFgO formed = 1.U5 ± 0.70. For ex-
periments at higher ozone pressures (runs 16-23, Table XXX) this ratio is much
higher, in the range 5-70, and parallel paths for ozone decomposition and re-
action to form gaseous products are obviously important here. In these experi-
ments with the Teflon gaskets the ozone decomposition is most evident because of
the relative slowness of the reaction which is a direct consequence of the very
small Teflon surface area. For the experiments with powdered Teflon the average
ratio of the molecules of ozone lost to those of CF20 plus CQ2 formed = 1.77 ±
0.77. The average ratio from the two most reliable runs involving the 03-FEP
film is 1.0 ± 0.2. The observed variation between determinations of this ratio
is believed to be a consequence of the inherent inaccuracy in measurement of 'the
very small absorbance changes in the reactant ozone during the reaction. Although
the scatter in the estimates of the ratio is large, there appears to be a some-
what larger number of ozone molecules destroyed than molecules of CFaO and COg
produced. Of course the.difference between the observed ratio and unity may
simply be a measure of a parallel heterogeneous 03 decomposition reaction. The
observed stoichiometric ratios are not inconsistent with the consumption of a
fraction of the surface and gaseous zwitterions by the reactions with ozone,
reaction 5 and 6. If these occurred exclusively, then two ozone molecules would
be decomposed for each CF^O molecule formed as a product. However, other evidence
which will be discussed subsequently favors the relative unimportance of reactions
5 and 6.
Since the mechanism forming the products CO^ and SiF4 must couple with the
CF20 formation in order to maintain the same product distribution observed ex-
perimentally for a variety of reaction conditions, it seems logical to utilize
the gaseous zwitterion as a reactant for these products. One such possible over-
all reaction is <*):
107
-------
(fluorineted ..
(glass surface) glasa .surface)
.0~ - F^ ^
„ Si
oX N~
0! represents the fraction of those Si atoms which react at the glass surface and
form SiF4; 1 - a is the fraction which react to form a partially fluorinated glass
surface. With this interpretation the expected ratio of C02/SiP4 in the products
= (1 + Jy)/2a. For the observed ratio of C02/SiF4 = 7, o/ = 0.09. Obviously
reaction 9 is expected to be a composite of several elementary steps which re-
main undefined.
One other reaction is necessary to describe some of the rate data. In two
cases successive runs of increasing duration were made to test the time dependence
of the product formation; for example, see runs 18 and 22 of Table XXX. In these
cases [CFaO] in the mixture approached a constant value at long times. We esti-
mate by the technique described, later that the number of reactive sites present
in the Teflon was depleted by only about 1.6% during run 18 and about 20$ in
runn 22, so that the constant [CFaO] reached cannot be the result of olefinic
site removal at these times. It appears more appropriate to assume that some
reactive surface species, possibly the zwitterions, consumes CF20 to form a
secondary ozonide, reaction 10, in a process completely analogous to that
observed in solution, ozonolysis ' . It is impossible to say definitely at this
/0-ox
~CFQa + CF20 -» ~CF -,CF2 (10)
point whether or not the surface zwitterion is sufficiently stable (half-life of
days at room temperature) to function in the fashion'we require here. However,
our results from another st.udy^-32 tend to support this possibility. Thus treat-
ment of the infrared-photolysis cell with high ozone pressures gives rise to a
reactive species (as evidenced by its complex but rapid reaction with added S03)
which can not be pumped away at 25°C. It is removed by washing the cell with
distilled water and subsequently outgassing at 100° C. It is not unreasonable to
attribute this behavior to a surface zwitterion or surface ozonide.
It is instructive to estimate the rate constant for the ozone reaction with
the CF2=CF"" reactive site in the Teflon from the present data. Note that the
number of molecules of CFgO and COa product formed per unit time in Run 2J of
Table XXX is 26.6$ less than that of Run 20 although the initial pressure of 03
is nearly the same. It is a reasonable hypothesis that this lowering in rate
comes from the partial elimination of olefinic sites, and that a direct propor-
tionality exists between the rate the number of sites. From the amount of
products formed we can estimate that about 2.J x 1019 sites were consumed in
Runs 20, 21, and 22. Thus about 8.5 x 1019 sites would have been available at
the start of Run 20. Similar less precise estimates can "be made from other
combinations of runs assuming'the linear dependence of rate on the [03J. Thus
we estimate 5-5 x 1019 (Runs 18 and 20), 2.b x 1019 (Runs 21 and 22), 16 x 1019
(Runs 19 and 22) active sites. It appears that the estimate of 8.5 x 1019 sites
at the start of Run 20 is accurate to at least an order of magnitude. This
108
-------
number and observed initial rates have been used to estimate the number of
reactive sites at the start of each experiment shown in Table XXX. The rates
shown in the last column of this Table have been normalized to 1.0 x 1020 sites.
These are plotted as a function of [03] in Figure *>k} curve A. The rate of the
Os-Teflon gasket reaction is seen to be roughly proportional to the ^Oaj. Of
course this is also the observed result from the much higher surface area samples
of Teflon powder (curve B) and FEP film (curve C) where depletion of surface
series during the series of runs was negligible. Departure of the Teflon gasket '
data from linearity at high [03] may be an artifact which results in part from
errors inherent in the estimation of the number of olefinic sites with a sub-
sequent inaccurate normalization procedure. It is also probable that the
reaction 10 may have occurred to a significant extent in the runs at the highest
[03] and caused some of the nonlinear behavior seen at high [03] 's. Deviations
from the assumed linear Beer's law dependence of the absorptivity of CFgO could
also contribute to this curvature. In any case the initial rate data can be
used to derive a reasonable estimate for the rate constant kx; we find ki = 5 x
10~25 cm3sec~1site"1. The specific reactivity of the Teflon with ozone is con-
siderably lower than that observed for the gaseous perfluoro-olefins studied by
Heicklen-^j he reported rate constants for 03 plus C2F4, C^e, and 2-C4F8 mole-
cules as equal to: 1-3 x 10~19, 2.2 x 10~2°, and 1.8 x 10~2* cc/mole-sec, respec-
tively. The 72-fold decrease in rate constant for the reaction of ozone with
olefin in the sequence C2F4 to 2-C4F8 observed by Heicklen, is opposite to the
trend in reactivity observed for the hydrocarbon analogues. Thus Wei and
Cvetanovicl56 found that the rate constant for the ozone reaction with CyLt is
over 10-times e-lower than that for cis-2-€4He and over IJ-times slower than that
for the trans-2-C4Ha- However, the mechanism of Os-perfluoro-olefin gas phase
reaction is very complex and apparently in the case of 2-C4Fe involves a reaction
of ozonide with a second molecule of the olefin. Of course this feature of the
reaction is physically impossible for the relatively rigid Teflon structures
studied here, and this difference may in part account for the apparent slowness
of the reaction 1 in our system.
If we use our estimate of k^. derived here together with the observed rates
of the reaction for the different Teflon samples, we estimate that the number of
olefinic groups available on the powdered Teflon and FEP samples are 2.2 x 1021
and 9 x 1020, respectively. ' By comparison with other samples, the number of re-
active sites per unit of apparent physical surface area is high for Teflon gaskets,
Possible reasons for the apparent wide differences in the population of active
surface sites in different samples include the following: (l) The actual surface
area available for the gaskets is much greater than that calculated from its
physical size due to roughness caused by machining; (2) The internal structure,
which may contain more unsaturation than the surface, is uncovered by machining;
(3) Machining may lead to unsaturation in the polymer.
The surface area for Teflon powder, determined by the supplier, presumably
by the BET method, would appear most appropriate in determining the area oc-
cupied by each reactive surface site. The result is l.U x 10~ls crr^/site. This
would seem to indicate complete surface coverage by olefinic groups. Alternative
explanations are probably more realistic: (l) Surface area determinations using
the BET method for Teflons are possibly unreliable: this might be the case if,
for example, N2 adsorption occurs only at olefinic sites: (2; 03 attack may be
at saturated carbon rather than, or in addition to, that at olefinic sites; (3)
Perhaps new olefinic sites are generated in large part in the ozonolysis process,
so that the present method of estimating the number of sites is not appropriate;
(k] 03 may diffuse into the Teflon matrix and react with terminal olefinic groups
109
-------
situated somewhat below the surface.
Information on the magnitude of the rate constant for the reaction 10 can
be gained from the data of Table XXX and the related data from these runs for
the P(rp Q attained as an asymptotic value at long times. If one-equates the
rate of CF20 loss to that for its formation for this condition relation 11 should
hold:
= k.
So
T? x D r np- nT s v •'-U
KCF20
Here So and Sz are the number of olefinic surface sites and number of zwitterionic
surface sites, respectively, in the sample. All quantities required in the calcu-
lation have been estimated with the exception of Sz- A maximum for Sz can be
derived assuming that the surface zwitterion is destroyed only in reaction 10
with CFaO, and that the number of surface zwitterions generated is equal to the
number of CFaO molecules formed. From these assumptions and using the data for
Runs 18 and 22. we calculate the following approximate upper limits for k10:
2-3 x 10"22 and 2.2 x 10~22, respectively.
The present study has provided some new insight into the reactions present
in the ozone-Teflon system; althougn many fundamental questions remain open, the
experimental approaches which can be used to press the study further are evident.
It is clear that Teflon is not inert to ozone. Those scientists who study the
photolysis or the thermal reactions of ozone at moderate to high pressures must
be aware of the previously unexpected contamination which will result in the
ozone mixture from the use of Teflon stopcocks, gaskets, etc., in these systems.
Particular attention should be given to the complications possibly resulting
therefrom in smog chamber studies employing Teflon bags or FEP coated chamber
walls. Although reaction with Os is relatively slow for the low [03] 's commonly
encountered in environmental systems, the carbonyl fluoride product is of major
concern since it should be photochemically active to form highly reactive F-atoms
and FCO radicals, and it is known to hydrolyze rapidly in the presence of water
to form HF and COa so the result would not appear benign-*-5T. on a more positive
note, similar studies of ozone with related polymers have great potential for
use in the determination of the properties of intermediates and products of
ozonolysis.
110
-------
PART III.
THE COMPUTER SIMULATION OF THE RATES OF CHEMICAL
REACTIONS IN SUNLIGHT-IRRADIATED POLLUTED ATMOSPHERES.
A significant part of the study supported under this grant involved the
computer simulation of the chemical changes expected to occur in various pol-
luted atmospheres. Some new and seemingly significant findings were made which
bear on the further development of control strategy. These studies are described
in this section.
Ill
-------
Ill-A. The Relative Importance of Various Active Intermediates
in the Attack on Alkenes in the Polluted.. Atmosphere1^?.
We -have recently developed a detailed kinetic model for the simulation of
smog chamber results from several laboratories1^0 and applied it, (i) to determine
the predicted effect of carbon monoxide on the ozone levels "in photochemical
smog systems1"-1- and (ii) to determine the relative importance of the various
intermediate species in the olefin removal reactions ia-simulated photochemical
smog chamber studies1^2. The success of'the model in matching the product-time
curves from the smog chamber experiments gives confidence in the general cor-
rectness of the kinetic mechanism and prompts us to apply this model to the
computer simulation of the chemistry of a simple analogue to the sunlight-ir-
radiated auto-exhaust polluted atmosphere.
III-A—1. Discussion
There is considerable sophistication of kinetic detail which we can employ
with the mechanism in attempting to simulate the chemical events which occur in
an auto-exhaust polluted atmosphere. Our present knowledge of photochemical
smog systems does not justify any such elaborate attempt. Instead we have chosen
a relatively simple mixture of the important classes of reactants to attempt to
picture the chemistry of these systems. The mixture consists of 0.075 ppm NO,
0.025 ppm NQg; 1-5 ppm CH4, 10 ppm CO, 0.10 ppm trans-2-butene (typical level of
total olefin), 0.10 ppm OfeO and Oj06ppm CH3CHO (representative of an the higher
aldehydes) in air with a relative humidity of 5$ (25° c). These concentrations
are typical for the major classes of primary and secondary pollutants in photo-
chemical smog.
Figures 35A, B,36 show the skeleton reaction scheme for the photooxidation
of trans-2-butene in a mixture with NO-NOa-air. Full mechanistic details with
the selected rate constants are given in reference1^. The theoretical time-
dependences of the products expected when the above synthetic mixture is ir-
radiated in sunlight (z = hO° ) were calculated by computer simulation. These
data are shown in Figs. 3?a- and37fa. which have been calculated asing two different
assumptions concerning the rates of the reactions 1, 2, and 3.
N205 + H2O •* 2HONO2 (1)
NO + N02 + H20 •* 2HONO (2) '
MONO -»• NO + N02 + H20 (3)
There is at present an uncertainty as to the degree of involvement of these
reactions in the real atmosphere. Certainly the magnitude of the homogeneous
component of these rates remains unclear today. Thus we have carried out the
simulations using the two possible extreme views concerning the rates of reactions
1, 2. and J. In both of the simulations shown in Fig. 3?a and 37b we have assumed
that there is no prior establishment of the equilibrium level of HONO before the
sunlight-irradiation begins. In calculating the data of Fig. 37a we have taken
"high" values for k2 and k3 estimated by Wayne and Yost1"? using conditions of
high surface-to-volume ratio, and the value kx = 1.0 x 10~4 ppm"1 min"1. which
we derived from "best fits" in .the simulation of several smog chamber studies;
this value is about 25 times lower than the only published estimate1^ Of y^.
On the ether hand, in calculating the product concentrations in Fig. 3?b, we.
have assumed that reactions 1, 2, and 3 do not occur at all; that is, kx. k2 and
ks are all assumed to be zero.
112
-------
CHjCH-CHCH.
CH3CH=»CHCH3..
CH3CH(6)CH(62)CH3
'(3P)
CH2=CHCH(CyH)CH3
CH3CHO2 + CH3CHO
K .
c«3ct-KQ,)o2
CHjCH-CHCHj
OCH(CH3)C
OOOChKCH3)6
CH3CH(6)6 * O2
CH3CC
|
CH3*
I14.
}NO
CHjO
h
jLHj.0
1 R
CHO
jo.
CO] H
CO2
^^^a.
* ^^"2 ^^~^^-»
» RH
HO2
"CH3CHO* » N
1 R
CKjCO * R
1°' .
JNO,
ICH-jCOOgNOg
CHjONOg
|CrljVjlNC'
io2
H
NO , C||C
j as bcfo
= PAN COg
CHgCf
CO
Fig 35, The major reaction paths for degradation of tran«-2-butene
in a sunlight-irradiated, NOx-pol.luted atmosphere; Fig.SB a : O3, 0(3P),
and O2<1A_) reactions with trans-2-butene. Products marked with an
asterisk undergo significant photodecon^osition in sunlight.
113
-------
CI-LCH-CHCH-
-
rr?
CHjCHCHCht,
CH3CHCH(O2)CH2
NO
-?
HCH(O)CH2 * NO2
CH3CH(6)CHO
CHLCHOH * CH3CHO*
CH3CH(C2)OH
INO
CH3CH(6)OH * NO2
1
HCOOg
\NO
HCOO2NO2 HCO2»NO2
H
Pig. 3E> ]> Bie HO radical reaction paths with trans-2-butene.
llU
-------
CH3CH = CHCH3
CH3CH(O2H)CHCH3
JO,
CH3CH(02H)CH(62)CH
NO
CH,OCH(CH_)CHCH,
lo, 3 3
C HjOC HCCHpC H(62)C HL
]NO
CH3CH(02H)CH(6)CH3 » N02 CH3OCH(CH3X:H(6)CH3 » NO2
CH3CHdi| * HO
I
as before
C02
CO
PAN
CH3OCHCH3 * CH3CHO
NO •
CH£ CH3OCH(6)CH3 » NOg
CH,6 *
(NO,)
(HO,)
CH3CH=CHCH2» CH3OH
las before
-------
4O 80
IRRADIATION TIME. MINI
12O
Fig. 373 « Theoretical rates of product formation in a sunlight-irradiated
(z = 40°), simulated auto-exhaust polluted atmosphere; initial concentra-
tions
-------
TABLE XXXII.
The Rate of Attack (ppm min"1 x 101*} of Various Reactive Intermediate Species
on Trans-2-Butene in a Sunlight-Irradiated (z = 40°), Simulated Auto-Exhaust
Polluted Atmosphere; Initial Concentrations (ppm): [NO]' = 0.075; [N02l° =
0.025; [trans-2-CijH8J * 0.10; [CO] ° = 10; [CHaO] ° = 0.10; [CHsCHO]0 = 0.060;
[CHtJ0 = 1.5; relative humidity = 50% (25°C)a
Time, min 0(3P) 03
2
10
30
60
90
120
0.13
(0.13)
0.20
(0.20)
0.17
(0.18)
0.08
(0.11)
0.03
(0.05)
0.01
(0.02)
0.26
(0.26)
0.83
(0.79)
1.58
(1.60)
1.43
(1.59)
1.00
(1.15)
0.65
(0.74)
- - - iteacrtive species - -
H02 HO CH30
1.69
(1.62)
1.84
(1.69)
1.49
(1.48)
0.88
(0.93)
0.49
(0.52)
0.28
(0.23)
18.1
(17.2)
12.5
(11.2)
5.9
(5.5)
2.6
(2.7)
1.2
(1.5)
0.7
(0.6)
0.026
(0.025)
0.018
(0.016)
0.008
(0.009)
0.004
(0.004)
0.002
(0.002)
0.001
(0.001)
NO 3
0.00005
(0.00005)
0.00051
(0.00050)
0.0022
(0.0028)
0.0026
(0.0045)
0.0090
(0.0041)
0.0018
(0.0028)
o2('V
0.000029
(0.000029)
0.000027
(0.000027)
0.000020
(0.000021)
0.000012
(0.000013)
0.000007
(0.000008)
0.000004
(0.000004)
Kesults from two simulations are shown; one was made assuming finite litera-
ture values for the ki, ka and ka rate constants; the other (results shown
in parentheses) was made assuming ki, k2 and k3 equal zero.
much higher olefin concentration. The theoretical fraction of olefin attack
which results from N03 and Oa^O is seen to be negligible for these conditions.
Note in Table XXXII that the numbers in parentheses were calculated assuming
kij ^2? and ks equal to zero as in the simulation shown in Fig. 37^- The other
numbers were calculated using the finite values for these rate constants as
designated previously for the simulation shown in Fig. 37A. Obviously there is
no significant change in the rates of the several species obtained using these
alternative assumptions.
Much concern has been focussed on the question of what reactions actually
trigger the process of NO to NOa conversion in the sunlight-irradiated polluted
atmospheres159. From the present simulation we can see that the HO radical is
the most important transient species from the standpoint of initiating the
olefin oxidation. However, any reactant which creates either HO or R02 will
effectively generate HO radicals for us, since the reaction, HOa + WO ~> HO + NOg
is the major fate of HOa for these conditions. Then to understand this system
it is important to investigate the relative importance of the several possible
sources of both these radicals for the present system.
Again we should concern ourselves with the cases which represent the two
possible extremes for the real atmosphere. First let us examine the happenings
for the case of no initial HONO formation; that is, ki and k2 equal zero. For
our assumed impurity levels we will have only the following initial sources of
117
-------
stimulus to chemical change:'
N02 + hv -*• o(3p) + NO (4)j Initial rate = 120 x KT1* ppm mirT1
H2CO + hv -»• HCO + H (5); Initial rate = 2.2 x 10"1* ppm min"1
CH3CHO •¥ hv ->- CH3 + HCO (6); Initial rate = 1.5 x 10~4 ppm min'1
Although the rate of 0-atom formation is by far the largest of the intermediates
formed here, very little impetus to the rate of photooxidation of the olefin or
the conversion of NO to M>2 is given by reaction k. The occurrence of this re-
action establishes an appreciable initial rate of 03 formation, but this is never
seen experimentally since 03 destruction and the very fast regeneration of NOg
occur through the reaction 7-
:'o3 -i- NO •* oz + no2 <7'
The extent of chemical change in NOg and. 03 is limited very quickly. If only
NO and NOg were present as impurities in this system, we would expect that the
[NOg] would fall somewhat as k occurs, and 03 would rise to a relatively small
steady state value of about 5-1 x 10~3 ppm in about 2 min time. However, the
reaction h .provides only a relatively minute rate of 0-atom attack on the
olefin, 0.13 x 10~4 ppm min"1 at 2 min, unimportant in explaining the rates of
chemical change observed in Fig. 37 A.
Thus the major part of the initial push to oxidize NO and olefin in this
system must come from the aldehyde photolyses in this case. Since most HCO
radicals react to form HOa. for these conditions, the total initial rate of HOa
generation from the aldehyde photolyses is expected to be about 5.5 x 10"4 ppm
min"1. For the initial reactant concentrations employed, about 9k.8f> of the
radicals oxidize NO and about 5$ react with olefin:
' H02 + NO -»• N02 + HO (8)
• HO2 4- CH3CH=CHCH3 ->• CH3CH(H02)CHCH3 (9)
An additional reactant is formed in 6, the methyl radical, which will form CH3Oa
in large part. Thus RQJ Q = 1.5 x 10*4 ppm min -1. For these conditions nearly
all of the CH302 radical! will react to oxidize NO:
CH3O2 + NO + CH30 + N02 (10)
Thus the initial rate of NOg formation by way of 8 and 10 will be about 7 x 10~4
ppm min"1. Of course this rate of formation is very much smaller than the initial
rate of destruction of N02 by photolysis (120 x 10~4 ppm min"1). The major key
to the rise in NOa in the smog system is the participation of the HO-olefin re-
actions which generate the chain processes that allow indirectly each HO radical
to oxidize several NO molecules. It is instructive to look into the details of
these processes using the data for our present simulations.
Let us first note the rates of generation of HO and HOa free radicals from
the several major sources in our system for several times throughout the photo-
oxidation as pictured in Fig- 37- The major share of the HO radicals formed can
be accounted, for throughout the entire run by the following reactions:
118
-------
(A) HONO + hv> + HO + NO
(B) 0(1D) + H20 * 2HO
(C) H02 + NO •+• HO + NO2
(D) H202 + hv '-»• 2HO
(E) CH3CH02H •* CH3CHO + HO
(11)
(12)
(13)
(14)
(15)
We have calculated the rates of HO-generation from each of these steps as
function of the irradiation time, as shown in Table XXXIII.
TABLE XXXIII
Comparison of the Theoretical Rates of the HO-Radical Forming Reactions in
a Simulated, Sunlight-Irradiated, Auto-Exhaust Polluted Atmosphere3
Rate of HP-Formation (ppm min"1 x 10**) from the Reaction Indicated
A B C D E
Time, min (reaction 11) (reaction 12) (reaction 13) (reaction 14) (reaction 15)
0.05
0.50
2.0
10.0
30.0
60.0
90.0
120.0'
0.0088
0.0907
0.32
0.64
0.29
0.12
0.07
0.05
0.0015
0.0108
0.023
0.082
0.23
0.38
0.48
0.55
28.3
29.3
27.8
18.5
10.0
5.9
4.0
1.2
0.0003
0.0029
0.013
0.080
0.33
0.79
1.2
1.6
1.5
1.6
1.6
1.7
1.5
0.9
0.5
0.2
The simulated auto-exhaust polluted atmosphere has the same composition as in
Table XXXII. The rate constants for the reactions 1, 2 and J are assumed to
be equal to zero here as in Fig. J?B for this same system.
Obviously the photolysis of HONO is not important at very short times since
we have excluded its presence for this consideration by setting k2 = 0. However.
even in this case it is formed in other reactions as the run progresses, and its
photolysis contributes a maximum of about ^ to the total rate of HO generation
at 10 min. Ozone of course is absent at the start of the run, and hence 0(1D),
its photolysis product, is also absent at this time, and reaction 12 is not
important. However, late in the photooxidation, as 03 builds up. the 0(1D)
generation of HO does become a significant contributor to the HO formation rate.
The photolysis of E^Os, also unimportant initially, provides a steadily in-
creasing source of HO as HaOs builds up with time; in fact, after most of the
butene has been oxidized at 120 min, reaction Ik is the major source of HO-
radicals in the mixture. The reaction 15 provides a steady but rather minor
source of BO radicals as well. The radical CH3CH02H is one of the products of
119
-------
the reaction sequence which follows HQg addition to the butane.
Note in Table XXXIII that by far the largest rate of HO radical generation
is through the HOg-NO reaction 1J. In fact the magnitude of" this rate (e.g.,
29O x 10~4 ppm min""1 at 0.5 min) is much greater than the rate at which we
make H02 radicals in the sunlight initiated steps described previously; this
rate totals to 7.0 x 10~4 ppm min'1 , 5/5 x 10"4 from H and HCO formed by
aldehyde photolyses and about 1.5 x 10~4 ppm min"1' by way of CH3 formation from
(02) (NO) (02)
acetaldehyde photolysis: (CH3 -» CH302 -* CH30 -» H02 + CHgO). Obviously
the HQg-radicals must be regenerated in a chain reaction to provide the observed
rate of reaction 13- The details of this chain process can be understood by
comparing the rates of H02 formation from the various major sources of this
radical:
(F) CH20 + hv -»• HCO + H; H + 02 + M -»• H02 + M (16)
HCO + 02 •+ H02 + CO
(NO) (02)
-f HCO02 -* N02 + HCO2 -»• H02 + CO2
(N02)
H- (HCO02N02) •* HCWO2 + CO2
(G) CH3CHO + hv -" CH3 + HCO (17)
HCO + 02 [as in (A)] * oH02 + products
(H) HO + CO •*• CO2 + H; H + O2 + M -»• H02 + M (18)
(I) CH3CH(6)CHO •* CH3CHO + HCO
HCO + 02 [as in (A)] •* oH02 + products
(J) CH30 + 02 -»• H02 + CH20 (20)
(K) CH3CH(6)6 + 02 * CH3C02 + H02 (21)
The rates of KQ£ formation which arise from each of these sources are compared
in Table XXXIV. Obviously the photolyses of both formaldehyde and acetaldehyde
continue as significant sources of HOs through H and HCO formation; see
column F and G in Table XXXIV. Acetaldehyde photolysis also forms HOg with
(Qa) (NO)
rather high efficiency through the CH3 radical reactions: CH3 -» CH302 -» CH30,
as we have. seen. The CH30 radical generates HOg in reaction 20. However, only a
small part, about 1.5 x 10~4 ppm min"1, of the total rate shown in Column J at
'short times comes from this source; see the following discussion.
A third source of HOa is the HO-CO reaction l8. This reaction has been sug-
gested to be the major HOa regeneration step in the chain oxidation of NO to '
120
-------
TABLE XXXIV.
Comparison of the Theoretical Rates of the H02-Forming Reactions in a
Simulated, Sunlight-Irradiated, Auto-Exhaust Polluted Atmosphere3
Rate of HO2-Fonnation (ppm ain"1 x 10**) froa the Reaction .Indicated
F (reac- G (reac- H (reac- I (reac- J (reac- K (reac-
Time, min tion 16) tion 17) tion 18) tion 19) tion 20) . tion 21)
0.05
0.5
2.0
10.0
30.0
60.0
90.0
120.0
4.0
4.0
4.1
4.3
4.3
4.2
3.8
3.4
1.5
1.5
1.6
1.8
2.2
2.5
2.5
2.4
4.2
4.4
4.3
3.2
2.2
1.8
1.7
1.3
5.1
5.2
4.9
3.0
1.5
. 0.7
0.4
0.2
15.8
16.5
15.9
11.8
8.8
6.9
5.4
3.9
0.05
0.1
0.2
0.6
1.1
1.1
0.8
0.5
aThe simulated auto-exhaust polluted atmosphere has the same composition as in
Table XXXII. Rate constants for the reactions 1, 2 and 3 are assumed to be
equal to zero here as in Fig.33bfor this same system.
Although it is important in this regard here, it is by no means the main source
of the HO to HOa conversion in this simulated polluted atmosphere.
The sequence I provides another source of the H02 radical from the de-
composition of the intermediate radical, CH3CH(6)CHO. Theoretically this
radical is formed in the reaction sequence which follows the H-atom abstraction
from the butene by the HO-radical (or other radicals, to the much smaller extent
that H-abstraction by them occurs).
The rates in column j of Table XXXIV show that the largest source of H02
in this system is the methoxy radical reaction 20. Shortly after the photo-
oxidation of the butene containing mixture is initiated, the CH30 is formed
from CH3 radicals derived from several reaction paths. We have considered the
relatively minor source from acetaldehyde photolysis previously. In addition
it is formed in the reaction sequence which follows the HO-radical addition to
the butene. Specifically we picture its formation in the following step:
CH3CH(OH)6 * CH3 + HC02H (22)
Also HO-attack on acetaldehyde can lead to CH3 radical formation through the
(HO) (Cfe) (NO)
sequence: CH3CHO -* CH3CO -» CH3C002 -» CH3C02 •*• CH3 + C02- When the
[NO]/[N02] ratio is relatively high, the CH3C002 radicals oxidize NO in large
part to lead to the CH3C02 radical. PAN formation is the favored product only
at low [NO]/[NOa] ratios (at long run times).
In theory the reaction sequence which follows ozone attack on the olefin
121
-------
leads to the radical CHsCH(0)0. Its reaction with Og is the fifth source of
reaction 21; this contribution to the rate, shown in- column K of Table XXXIV, is
negligible at short times, but as 03 builds up at long exposure times'^ it has a
finite contribution.
From these theoretical considerations we see that there is a complex inter-
play between the -various reactants in this simulated smog forming atmosphere
which stimulates the generation of the important HO and HOg radicals. For the
reactant conditions which we have chosen here, the olefin acts as the major
source of conversion of HO to HOg radicals. Both the abstraction of H-atoms
from the olefin by HO and the addition of HO to olefin generates an EQ2 radical
eventually with fair efficiency. In the sequence of reactions in which this
occurs, at least one molecule of NO is oxidized to NOg by various RQ2 radicals
as well. Thus for short exposure times in Fig. 37A, when each HO radical
abstracts H-atoms from butene, about one molecule of NO-is converted to NOa and
one HOg radical is formed. For each HO radical which adds to butene about three
molecules of NO form NOa and again one HOa molecule is formed as well. The HOg
radicals are largely reconverted to HO at short times through the reaction, KQ2
NO -» HO + NOs, and hence the chain cycle of attack of HO on C4He continues to
degrade the olefin until the chain is terminated. The rate of HOg formation in
the primary photolytic processes (7-0 x 10~4 ppm min"1 at 2 min) and the total
rate of HO and RQ2 attack on the olefin (19-8 x 10~4 at 2 min), suggests that
there is a short chain reaction, about 2.8 cycles in length at 2 min, which in-
volves the HO and HQa.radical oxidation of olefin. Of course the chains are
stopped whenever an RQg or an HO radical is removed by reactions with another
radical or an odd electron molecule such as NO or NOa- The dominant chain
ending steps for our chosen concentration conditions in this system are the
following:
2H02 •* H202 •*• 02 (23)
HO + NO + M •* HONO + M (24)
HO + N02 + M -»• HON02 + M (25)
CH302 + H02 •* CH302H + 02 (26)
The rates of radical removal in these reactions for our simulation in Fig.
at 2 min, are as follows: ^3 = h.6 x 10~4; Ee4 = 1.2 x 10~4; R^ = 0.8 x 10"4;
Bg6 = 0.07 x 10~4 ppm min"1. The total of these rates accounts for a rate of HOa
and HO radical removal = 6.6 x 10~4, ppm min"1. This would match our rate of
primary radical production (7.0 x 10~4 ppm min"1) if all radical sources and
termination reactions were included in our considerations.
Now' let us return to consider the alternate hypotheses of HONO pre-
equilibration in our simulated polluted atmosphere. We will allow HONO to have
been formed at its equilibrium value, [HONO] = 6.1 x 10"3 ppm, before sunlight
irradiation of the mixture. All other compounds are at the concentrations em-
ployed as before, and the HONO generating and destruction reactions, 2 and 35
are assumed to have their finite literature values. The initial rate of radical
generation in this system will include the same rates of H02 and CH302 formation
from CHaO and CH3CHO photolyses'as before, but there will be an additional source
of HO radical.formation from the HONO photolysis:
HONO + hv •+ HO + NO (11)
122
-------
This rate will be about 7 ...3 x 10~4 ppm min"1 for z = 1*0° ." Compare this with
the rate of HOa formation from all „of the aldehyde photolysis processes; RJJQ =
7.0 x 10~4. Obviously the HONO presence in the atmosphere can give a significant
boost to the initial rate of olefia photooxidation and NO to NOa conversion; it
would approximately double these initial rates for our simulations if we allowed
HONO preequilibration before irradiation. Assuming preequilibration of HONO and
allowing the finite values for the rate constants kls kg, and ks as before, we
predict from simulations not shown here that the maximum in the [NOa] would .
occur at 22 min. This compares with 2k min for the time of [N02]max for the
same mixture but with HONO absent originally; see Fig. 37A. A period of 31 min
is required to reach [NOgJmax when there is no HONO initially and we assume ka =
ks = 0 as in Fig. 37B.
We have clearly seen that HONO is not a necessary component of the polluted
atmosphere to account for the general features of smog formation,, but its
presence can enhance the initial rate of the. smog forming reactions. Of course,
its presence must be invoked to rationalize the photooxidation of NO in moist
atmospheres containing carbon monoxide as the only other oxidizable component.
It remains to be determined experimentally what levels of HONO are present in
the real atmospheres, and which, if either of the two extremes considered above,
best represents the real situation. It is our educated guess that the real
auto-exhaust polluted atmospheres have levels of HONO which lie between the two
extremes considered here.
We have also considered the expected effects of the presence of one ad-
ditional constituent not originally included in our simulated atmosphere, namely
n-butane, as representative of the saturated hydrocarbons. If n-butane is
present in our simulated atmosphere at the same level as the olefin, 0.10 ppm,
the rate of NOa formation is only increased by 2$>. Thus with paraffin hydro-
carbon additions at the levels near those of the olefins, the situation which
exists in the real auto-exhaust polluted atmosphere in the early morning hours1"?,
we can expect only a relatively small perturbation of the above general reaction
scheme.
III-B. The Effect of CO on the Chemistry of Photo-
chemical Smog Systems.
A detailed kinetic model for the simulation of smog chamber results has
recently been developed1"0 and applied (i) to determine the predicted effect
of carbon monoxide on the ozone levels in photochemical smog systems1"1 and
(ii) to determine the relative importance of the various intermediate species
in the olefin removal reactions in photochemical smog10^. The main features
of the effect of CO on the rate of 03 formation in NO-polluted atmospheres were
interpreted in terms of the following simplified reaction sequence:1°9
123
-------
2ND + Oa •*• 2N02 (1)
HaO + NO + N0a <=* 2HONO (2)
; HONQ + hv -•• HO + NO (3)
| HO + CO * H + C0a (4)
I HO + NO (+Na,0a) -»• HONO (+N2,0a) (5).
' HO + N0a (+Na,0a) -* HONOa (+Na,0a) (6)
i H + Oa (-HJa,Oa) -»• H0a (-HJa,Oa) (7)
; H0a + NO * HO + NO, ,(8)
i 2HOa -»• HaOa + Oa (9)
i N0a + hv * 0 + NO (10)
0 + Oa (-HJa,Oa) -*• Os (-Hla,0a) (11)
0, + NO -»• Oa + N0a (12)
03 + N0a * Oa + N03 (13)
i 03 +• hv -»• 0 + Oa (14)
NOS + NO •»• 2NOa (15)
NOS + N0a ?=* Na05 (16)
Na03 + HaO -*• 2HONOa (17)
Two recent developments prompt us to repeat and extend our computer simu
lations of this system. Firstly, reliable experimental data on reactions 5
6 indicate that the rate constants are significantly faster than formerly be-
lievedlfOj and secondly, experimental data on the photooxidation of NO in the
presence of CO have become available from smog chamber studiesl713'172.
III-B—1. Results and. Discussion
In view of the high rate constant-^'3 for reaction k and the relatively high
CO levels present in polluted atmospheres, two groups of scientists, Heicklen
et al., ' and Stedman et al. ,-^-°" came to the conclusion that the. rapid oxidation
of NO to NOg which occurs in the polluted atmosphere may be linked to the occur-
rence of reaction k coupled with the reactions 7 and 8:
HO + CO •»• C0a + H (4)
y H + Oa + M -f H0a + M (7)
i H0a + NO -»- HO + N0a (8)
This sequence provid.es a chain mechanism in which the formation of a single HO
radical could conceivably initiate the oxidation of many NO molecules. Although
absolute magnitude of kQ remains somewhat open to question, using our "best"
estimates of all of the rate constants for the reaction 1 - IT, we can simulate
the chemical changes which we might expect for the NO-NOa-CO-containing atmospheres
12k
-------
l6Q
and test our predictions with smog chamber data u;7.
The striking effect -of CO on the rate of NO oxidation in the NO-NO
system irradiated in air can be seen in Figs. 38A and 39A, obtained from the
smog chamber experiments of Wilson and Miller1'1, in which 100 and hOO ppm of
CO respectively were added to the systems. The results of our computer simu-
lation of this system based on the above mechanism are shown in Figs. 38B and 39B.
It can be seen that the simulations match the experimentally observed time de-
pendence of the measured reactants and products reasonably well. In addition
the concentration and time dependence of the minor products, COa, HONO, HONOa,
and HaOa, which we predict from our simulations, are shown in Figs. 380 and 39C,
but there are no experimental data with which to compare in these cases.
In view of the reasonably good match between experiment and simulation we
have attempted to extrapolate these observed effects of CO to the concentration
range of reactants commonly encountered in the polluted atmosphere. One such
extrapolation is shown in Fig. *K)C. The theoretical time dependence of the con-
centration of the reactants and products is shown for the sunlight-irradiated
mixtures with the initial concentrations of impurities: [NO] = 0.10, [N0a3 =
0.00, and [C0]° = 10 ppm, at a relative humidity, 50$ (25° C). Although we have
chosen conditions which minimize the chance for chemical change in the system,
there is a remarkable change expected in the rates of product formation on ad-
dition of CO; compare the plots of Figs kOA, ^OB, and hOC in which dry and moist
atmospheres contaminated with identical levels of NO and NOa are present with
and without cai-bon monoxide impurity. This effect is illustrated also in Figs.
klA and UlB; in this case the initial concentrations of the contaminants are:
[N0]° = 0.075, [N02]° = 0.025 ppm, and the relative humidity =5$. In Fig klE
the identical initial levels of NO, N02 and humidity are employed, but 10 ppm
of CO is added as well. The predicted role of carbon monoxide in enhancing
the NO oxidation to NOa through the reaction sequence k, 1 and 8 is apparent in
both comparisons. The decrease in the initial level of NOa observed in irradiated
dry and moist atmospheres contaminated with only NO and NOa is in striking con-
trast to the rapid increase in NOa which occurs after a small delay in the CO-
containing systems.
For given fixed NO and NOa impurity levels, the rate of NO oxidation in-
creases most noticeably with increasing CO level when the CO impurity is in
the .0-25 ppm range. Within this range the higher the initial [CO] the shorter
is the sunlight irradiation time necessary to reach the maximum [NOal and the
higher is that maximum. (See the data presented in columns 2 and 3 of Table XXXV.).
Thus at the initial reactant levels, [N02]° = 0.025, [N0]° = 0.075 ppm at the
relative humidity = 50$, the [NOa] maximizes at 3.5 pphm in 232 min for [CO] =
5 ppm, and at about 5.7 pphm at IkO min for [CO] = 25 pphm. It can be seen that
the [NOa]max and the time to reach this level are both rather insensitive to
increases in the concentration of CO when it is in the range 25-100 ppm. Ob-'
viously a major predicted effect of increasing the NOa level by the CO-sensi-
tized photooxidation is to increase the rate of 0(3P)-atom generation and the
ozone concentration in the system. The data of Table XXXV shows that the
addition of 5 ppm of CO to the moist atmosphere contaminated with NO and NOa
is predicted to cause more than a 10-fold increase in the ozone level reached
after exposure to sunlight for 152 min,,
The level of CO at which one expects significant effects on the rate of NO
photooxidation is somewhat higher than that which we suggested in our earlier
studies1"-1. The present estimates are based on more accurate recent rate constant
125
-------
0 60 120 180 240 300
IRRADIATION TIME, MIN
Fig.38. Comparison of experimental and computer simulated chemical
changes in NO-N02-CO mixtures irradiated in moist air; Fig.ssa exper-
imental data of Wilson and Miller (ref.lTi/ initial concentrations,
[NO]0 = 51 pphm; [N02]° = 10 pphro; [CO]0 = 100 ppm; relative humidity
about 13% at 90°F; Fig. 38b. computer simulation for the experimental
conditions employed in Pig. 38a Fig. 38c computer simulation of the
expected time dependence of the minor products for the conditions
employed in Fig.SSa experimental analysis for these products was not
made.
126
-------
62.5
60 120 180 MO 300
IRRADIATION TIME, WIN
Fig. 3g. Comparison of experimental and computer simulated chemical
changes in NO-N02-CO mixtures irradiated in moist air; Fig. 393
experimental data of Wilson and Miller (ref 171 initial concentra-
tions, [NO]0 = 50 pphm; [N021° = 10.5 pphm; [CO]0 = 400 ppm; relative
humidity about 13% at 90°F; Fig. 39b computer simulation for the
conditions used in Fig .39 a; Fig. 39 C simulation of the expected time
dependence of the minor products for the conditions employed in
Fig.39a experimental analysis for these products was not made.
127
-------
Q.
Q.
z00
O
IT
s
NO
60 120 180 240 300 360 420 480
IRRADIATION TIME, MIN
Fig. 40. The expected effects of water vapor and carbon monoxide addition
on the products of the NO photooxidation in air; comparison of the
computer simulated chemical changes in three sunlight irradiated (z =
40°), NO-polluted atmospheres; Fig.40a dry atmosphere; Fig.4Ob moist
atmosphere (50% relative humidity, 25°C); Fig.4OC moist atmosphere
containing carbon monoxide, [CO]0 = 10 ppm; in each case the initial
concentrations are [N0]° = 10 pphm, [N02]° = 0.0 pphm.
128
-------
30 45 60 75 90
IRRADIATION TIME, MIN
105 120
Fig.41. The NO-NC>2-CO-polluted atmosphere; simulation of the effects
of added carbon monoxide on the time dependence of the major products
formed in sunlight-irradiated (z = 40°), N0x-polluted atmospheres;
the initial concentrations in each case are: [NO]° = 7.5 pphm;
[Nd2]° » 2.5 pphm; relative humidity, 50% (25°C); Fig. 41 a [00]° -
0.0 ppm; Fig.4lb [CO]0 = 10 ppm.
129
-------
TABLE XXXV.
The Effect of Carbon Monoxide Level on the Concentrations (pphm) of Products Formed in the Sunlight-
Irradiated Moist Atmospheres (25°C) Containing Initially [N0a]° - 2.50; [N0]e - 7.50 pphm and [CO]0
as Indicated: z •* 40° Assumed.
Relative
Humidity,
% (25°C)
Ppm
ppm
min
Concentrations, pphm, at t "151.6 min
[N0a] [NO] [0,]
[NaO,] [HNOa] [HNOS] [HaOa] [C0a]
x 10* x 102
100
0
1
5
10
25
50
75
100
2.50
2.50
3.37
4.41
5.50
6.07
6.30
6.42
0
0
148
117
102
94
92
91
0.64
1.30
3.37
4.32
5.20
5.65
5.83
5.94
7.98
6.77
3.39
2.32
1.81
1.68
1.64
1.62
0.17 .
0.40
2.05
3.82
5.88
6.89
7.28
7.49
0.00
0.04
1.7
6.9
17.9
25.6
29.2
31.2
0.34
0.41
0.33
0.25
0.22
0.21
0.21
0.21
1.0
1.5
2.9
3.1
2.8
2.5
2.3
2.2
0.03
0.46
19.1
61.7
145
203
229
244
0
1.7
8.2
12.6
18.0
20.9
22.3
22.9
50
0
1
5
10
25
50
75
100
2.
2.
3.
4.
5.
6.
6.
6.
50
50
46
56
70
28
52
65
0
0
232
176
140
128
125
123
0.90
1.44
3.32
4.53
5.69
6.24
6.46
6.58
8.00
7.20
4.61
3.22
2.32
2.07
2.00
1.96
0.24
0.42
1.49
2.90
5.03
6.17
6.62
6.88
0.01
0.04
1.0
5.0
18.0
28.8
33.9
36.9
0.26
0.29
0.26
0.21
0.16
0.15
0.15
0.14
0.84,
1.1
1.8
2.0
1.8
1.5
1.4
1.3
0.01
0.19
6.5
24.7
72.6
110
128
138
0
1.1
5.4
9.0
13.7,
16.3
17.4
18.0
-------
determinations including that for the reaction. HO + NOa + M -> HON02 + M (6).
It is important to note from the data of Table XXXV that the major effects
of added CO impurity in an NO-NOa- contaminated system are expected to be seen
at relatively low concentrations of CO; the most dramatic increase in [os] oc-
curs in the range of [CO] = 0 to 5 PP^- ' A significant but smaller effect is
seen also with f CO] increased from 5 to 25 ppm, while further increase in [CO]
to 100 ppm has little further effect on the ozone level. Of course at very high
CO levels the duration of the CO enhancement of NO photooxidation would be pro-
longed to much greater times for a given rate of dilution in the atmosphere.
Recent rate data have been obtained by Dodge and Bufalini from a similar
smog chamber study of the NO-NOa-CO-HaO-containing system. These results are
shown in Figs. 42A and UjA. Although the effects of CO addition are similar to
those observed by Wilson and Miller1?1, there are some significant differences.
In fact the same set of rate constants which we have employed with some success
in the previous work fails to fit well these CO-containing runs. The reaction
volumes are similar for the reaction vessels employed by the Wilson and Bufalini
groups, and hence the surface-to-volume ratio is probably not very different..
There are two major differences which one should recognize. First the black
lamps employed by the Dodge and Bufalini group probably have a significantly
different wavelength distribution from the selection of sun-lamps employed in
the Wilson and Miller study. Although the rate of NOa photo-decomposition has
been measured in both studies (klo) and is a parameter which is not variable
in the simulations, the rate of nitrous acid photolysis may be significantly
larger in the study of Wilson and Miller, since there is probably a larger con-
tribution from the shorter wavelength region of light (~ JIQO A) -with the lamps
which they employed. There is an obvious second difference between the two
studies: the gaseous mixture was circulated mechanically only in the study of
Wilson and Miller. Presumably this circulation would allow a faster replenish-
ment of the HONO if there were a significant heterogeneous component to the rate
of reaction in the system of Wilson and Miller. Either one or both of these
effects may be important. Since it is really impossible to judge correctly from
the limited data at hand, we have arbitrarily chosen the first effect alone in
an attempt to fit better the Dodge and Bufalini results. Using the value of
'*26a = 8.8 x 10~3 instead of our adjusted sunlight value of 8.8 x 10 ~2 min"1, we
find the simulations shown in Figs. ^2B and 1*3B. The fit is considerably better
for this choice of rate data, but some real problems remain. Particularly, the
ozone levels observed by Dodge and Bufalini are much lower than those predicted
in the simulations. The explanation of this difference is not clear. Our
mechanism may be incomplete in some as yet undetermined way, or conceivably
there may be analytical problems in the data of Dodge and Bufalini. For example,
the thermal reaction, Q3 + NO -» Oa + NOa, roay occur significantly in the sampling
line leading to the analytical instruments for 03 and NOa- This would result
in higher NOa and lower Os levels than those actually present in the chamber.
The error would be present until such times that the NO is largely converted to
N02 °r other products. This problem has been observed recently in smog chamber
experiments1?^. As an example of the magnitude of . this possible effect for the
present case, consider the products in Fig. ^3B at 100 minutes irradiation time.
Note that a delay of only 1.6 min in transporting the NO, NOa, 03 mixture in the
dark to the analysis systems results in lowering the ozone to one-half of its
value when it was present in the chamber. Further speculation on this interesting
NO-NOa-CO-HaO system must await more complete experimental data over a more
complete range of experimental parameters .
131
-------
2.0
S
a.
£1.6
O
{E 1.2
z
0
§ 0.8
o
ff
z a4
0
i • i 1 i
-A-
— . A
A^
- *>
._.__
- A» _
/ A H^^^^^^"""""^^^
-If ^jm"^V^
A/ ^L^f^
/* ^^^*
'&
1 1 1 1
I I
• 5.3 ppm
T 5.0 ppm
0 4.6 ppm
A 4.9 ppm
_ ^
1 1
1
NOX
NOX-53%RH
NOx-20OOppmCO _
NO -ZOOOppmCO
56% RH
• -
T T~
-
—
1
a. cu
a.
o (P.-
at
Z •!.
ui —
8 g.
-B-
345
IRRADIATION TIME, HOURS
Fig. 42. The photooxidation of NO in CO-containing mixtures; comparison
of experimental and computer simulated chemical changes in N0-N02-C0
mixtures irradiated in relatively dry and moist air; Fig.42a exper-
imental data of Dodge and Bufalini (ref. ITS initial conditions as
shown; Fig. 42b computer simulation of the product concentration data
for the conditions employed in Fig. 42a.
From the available information on the CO-containing system and our simu-
lation studies, we conclude that the atmospheric scientists should give careful
consideration to the predicted enhancement of ozone levels by relatively small
amounts of CO in an NOX-polluted, but hydrocarbon free, atmosphere. In fact
these unexpected effects should be considered in the future development of more
detailed air quality standards. In view of the technological difficulties in
removing WOX and CO from auto exhaust and NOX from stack gases, there is a
reasonable possibility that significant ozone levels may continue to plague
many urban areas even though a near total removal of the reactive hydrocarbons
might be effected.
132
-------
0.2.4
I 0.20
a.
Q 0.16
ac.
£0.12
u
8 0.08
0.04
0
I I
A 0.21 ppm NOx and 100 ppm CO
• 0.22 ppm NOx Qnd 20 PPm co
70% RELATIVE HUMIDITY
I I I I I I I
Sj
c>
8.
!S_,
cM
Sj
ti
50 100 150 200 250 300 350 400 450 500 550 600
IRRADIATION TIME, MIN
Fig., 43. The photooxidation of NO in CO-containing mixtures; comparison
of the experimental and computer simulated chemical changes in NO-N02-CO
mixtures irradiated in moist air; Fig.4.aa experimental data of Dodge
and Bufalini (ref 172 initial conditions as shown; Fig^ab com-
puter simulation of the experiment shown in Fig. 43a-
III-C. Computer Simulation of the Rates and Mechanisms of
Photochemical Smog Formation^-75.
Today the phenomenon of photochemical smog formation, once considered
peculiar to the Los Angeles area, occurs to some extent in most large cities
throughout the world. In our rapidly growing urban areas, a high density of
automobile traffic, nearby power generation, and high commercial and industrial
activity are becoming more commonplace. When the relatively high ambient con-
centrations of NO, NOs, and hydrocarbon emissions from these geographically con-
centrated sources are coupled with bright sunshine and poor atmospheric ventilation,
photochemical smog formation is inevitable.
The unpleasant phenomenon of photochemical smog is characterized by the build
up in the urban atmosphere of significant levels of ozone, peroxyacyl nitrates,
and many other compounds such as the aldehydes, acids, etc., which arise from
133
-------
the partial oxidation of the hydrocarbons1^,176. The resident of a smog-bound
community may experience eye irritation from compounds not yet clearly defined.
As the usual smoggy day^develops, haze formation in the atmosphere may greatly
restrict his visibility of mountains, buildings, and other objects of some distance
away. Many agricultural crops and decorative plantings may show significant
deterioration when grown in an area plagued by smog. More importantly photo-
chemical smog represents a potential threat to the health of the urban dweller.
In view of the information available in 1971 relevant to the effects of. the common
air pollutants on human health, the Administrator of the U. S. Environmental
Protection Agency recommended to the U. S. Congress the adoption of national
ambient air quality standards for several common air pollutants: N02, CO, SOg,
hydrocarbons, photochemical oxidant, and particulates1''. Among the standards
adopted, and one with which we will be particularly concerned in our consider-
ations here, was that for photochemical oxidant (ozone): 0.08 ppm or 160 (J/g-rtf3
for the maximum 1 h concentration which should not be exceeded more than one
hour per year. The development of realistic control strategies which will ensure
the achievement of the ambient air quality standards is an active goal of the
E.P.A. However there is a special difficulty associated with the plans to
control the ozone level which provides the kineticist with an important op-
portunity to apply his trade. SOg, hydrocarbons, CO, and other primary pollutants
are in large part directly introduced into the atmosphere from sources which in
principle can be controlled to the desired degree. Ozone is a secondary pollutant,
generated in urban atmospheres through the occurrence of a complicated series of
interrelated, sunlight-induced, chemical changes which are not entirely understood,
but which are known to involve NO, WOg, hydrocarbons, CO, aldehydes, and perhaps
other reactants. The complex nature of the system and the extremely short time
base for the attainment of controls have stimulated a flurry of both Edisonian
and fundamental research efforts to solve the problem. However it seems probable
that any lasting control strategy for ozone must be built upon a solid scientific
knowledge of the chemical and transport mechanisms which determine the ozone level.
Such a data base is not now available in the chemical field and the same is likely
the case for the meterological input. In this paper we will face only some of
the chemical kinetic and mechanistic problems related to these goals.
Since the early 1950s the kineticists have had an active role in elucidating
the elementary reactions in smog formation. Leighton was the first to make a
comprehensive quantitative attempt to evaluate the alternative reaction paths
which may be operative in photochemical smog-'-39. Sophistication of the kinetics
treatment of these systems has increased in recent years as our knowledge related
to the various elementary reactions has grown, and computer techniques have become
available for the practical solution of complex series of interrelated differential
rate equations. However the research which bears on the mechanism of photochemical
smog formation has left unanswered many critical questions. When one appreciates
the full complexity of these systems the relatively slow progress is understandable.
Kineticists soon realized that a full consideration of'the many thousands of
reactions between the many hundreds of reactants in the real urban polluted atmos-
heres was an overly ambitious plan. We and many.others feel that the definition
and characterization of the reactions which are important in the real atmosphere
lies in the consideration of the simpler, well controlled, simulated polluted
atmospheres of the "smog chamber". Even these systems are very complex1^ and
difficult to treat quantitatively since many reactions which appear to be im-
portant in theory have not ..been studied in detail, and theoretical estimates of
rate constants must be made in desperation; the thermochemical-kinetic techniques
developed by Benson1?" have given us some reasonable guidance in this phase of
our work.
13k
-------
In this article we will analyze the mechanism and kinetics of photochemical
smog formation using an updated and expanded version of the chemical reaction
mechanism developed some years ago by Demerjian, Kerr, and Calvert1°0516l,l62,127 .
All of the results reported here have been calculated assuming that the rate
constants for the reactions, N205 + IfeO -* 2HON02 (8), NO + N02 + IfeO -» 2HONO
(l6a) and 2HONO -> NO + N02 + HgO (l6b) are zero. There remains some uncertainty
as to the degree of involvement of these reactions in the real atmosphere-^?.
Recent studies of Cox and Atkins-*-"" show clearly that kisa and k16v are very
small. Morris and Niki-'-"9 have reported that k8 ^ 1.9 x 10"5 ppm min"1 [liter/
(mol-s) x 2.U5 x 10~6 = ppm"1 min"1]. In any case the alternative choices of
these rate constants have only a very minor influence on most kinetic properties
of the systems-^. The rate constants and reaction mechanisms outlined by
Demerjian et al.-^f have been used, in most cases. The only significant rate
constant change is that for the reaction (17) H02 + N02 -» HONO + 02; here we
used k17 = 2.9 x 101 ppm"1 min"1, based on work of Simonaitis and Heicklen1^0,
esoEpt where specific notation is given. Reference to the extensive review of
Demerjian et al.12? should be made for reaction details not outlined here.
The new rate constants which we have employed in the SOg-containing systems
are outlined in Table XLIV.
Our attempts at computer modeling of photochemical smog mixtures represent one
of many efforts in this area^-79-loT. Our chemical system is somewhat more com-
plete than others in order to satisfy several unique demands of the kineticists.
It should serve as a useful predictive tool which will allow a resonable selection
of key chemical reactions needed for the highly abbreviated mechanisms used in
the comprehensive urban air shed models. It should provide some insight into
the nature and extent of some potentially important compounds as yet unidentified
in the real atmosphere but expected in theory. However the major role which we
hope our study can play is in the allerting of the kineticists to those reactions
which are seemingly important but for which little or no reliable kinetic data
now exist.
Before we consider the results of our simulations, let us. understand the
limitations of such considerations. In the simulations of the polluted atmosphere-
like systems described in this work, we have assumed that there is no atmospheric
dilution of products and reactants and that the sunlight is of fixed intensity.
All photochemical rates have been estimated for a solar zenith angle of k(f , a
value near the average encountered during a typical day in the United States.
Estimates of the actinic irradiance were taken from Leighton-^-39; these are theo-
retical estimates which should apply approximately for representative atmospheric
conditions near sea level on a clear day. Suitable modifications which incor-
porate diffusion and regular change in solar zenith angle can be made readily,
but they offer a degree of sophistication at this point in time which our current
chemical knowledge does not warrant, and they actually mask some of the chemical
features we are trying to identify. Obviously this choice of model will not
allow us to describe well the product rates expected for positions near point
or line sources such as smoke stacks or freeways where large fluctuations in
the input pollutants levels occur . The "box" models such as we employ here
should give reasonably good answers to chemical questions related to urban atmos-
phere somewhat removed from major pollutant sources and for conditions of strong
atmospheric temperature inversion. Obviously the coupling of such a reaction
scheme, suitably pruned to acceptable complexity, to atmospheric diffusion models
and local emission patterns must be made to achieve the full predictive potential
135
-------
of the comprehensive air pollution,models Which are necessary for local air
pollution control strategies180'162>l83b>l85,187.
III-C-1. Results and Discussion
Computer Analysis of the Chemistry of Smog formation _in ?a Simulated NOx-Hydrocar-
bon Polluted Atmosphere
For our first considerations we have chosen a relatively simple mixture of
several important classes of reactants; to the mixture of NO at 0.075 ppm and
WQz at 0.025 ppm, we nave added the background level of CEi, 1.5 ppm; a typical
level of CO, 10 ppm; typical levels of the total olefinic hydrocarbon, repre-
sented by trans-2-butene at 0.10 ppm; and aldehydes, CHgO at 0.10 ppm; and
CHSCHO, repsentative of all of the higher aldehydes, at 0.06 ppm. The effects
of n-butane addition at 0.10 ppm, representing the saturated paraffinic hydro-
carbons, and 0.10 ppm of SOg, a molecule which is often present today in urban
atmospheres, will be considered later. We will not consider the effects of the
addition of the aromatic hydrocarbons since the authors are not aware of meaning-
ful experimental evidence relating to their decay paths. We have assumed that
the relative hunidity is at 5$ and the temperature of the atmosphere is 25° C.
Through the computer simulation of this system we have calculated'the theoretical
time dependence of the products expected when this mixture is irradiated in
sunlight (z = hO°}. The concentration-time profiles for a few of the major
reactants and products are shown in Figure kk for two different starting con-
ditions: aldehydes are initially present for the solid curves, and aldehydes
are absent initially for the dashed curves. Without the aldehydes present
initially a small induction period appears in the rates of alkene removal, and
NOs, OQ, and peroxyacetyl nitrate (PAN) formation. Also recognize from a com-
parison of the final levels of the dashed and solid curves for the various
products that the concentrations of such problem compounds as Os and PAN are
raised when aldehydes are present initially. We will consider the effects of
aldehydes in more detail in the next section. Note in Figure hh the general
features which characterize the chemistry of a smoggy atmosphere: NO is converted
to NOa as the alkene is removed, and ozone and PAN build up with time. Present
knowledge of these systems points to only three reactions as the major ones
which control the ozone concentration^-39 3lo5,loo,191,1655192.
(1) NO2 + MX < 4300 A) ->• 0(3P) + NO
(2) 0(3P) + 02 (+ M) -» O3 (+ M)
(3) 03 + NO ->• 02 + N02
To a first approximation the concentration of the ozone which builds is related
to the [NOal/CNO] ratio and the intensity of solar radiation absorbed by NOg at
the effective wavelength region (\ < kj>00 A).
[N02] fc,
(I) [O3] =
() [N01 k,
136
-------
0.2 i
o.
Q.
O.I
bl
U
O
u
30 60 90
IRRADIATION TIME, mil)
120
I Figure ^».Theoretical rates of product formation in a computer simulated, sunlight-
l irradiated (z = 40°) NO^-hydrocarbon-aldehyde-polluted atmosphere; initial concentra-
i tiotis (ppm) for the solid curves: [NO,]0 = 0.025; [NO]0 = 0.075; [C4H,]° = 0.10;
[CH3O]° = 0.10; [CHjCHO]0 = 0.06; [CO] = 10; [CHJ = 1.5; relative humidity,
50% (25°C); dashed curves same but aldehyde concentrations are zero initially.
ki represents the apparent first-order rate constant for 0(3P) atom formation in
reaction (l) in the urban atmosphere. It is related to the integral over the
wavelength range 2900 < X < 4 300 A to the product of the solar irradiance in the
lower atmosphere at a given wavelength times the absorption coefficient of NOa
at that X, times the quantum yield of 0(3P) formation from NOs at X . We have
estimated for a solar zenith angle z = U0° and typical atmospheric conditions
near sea level, ki/k3 = 0.021 ppmf. Thus from relation (l) we expect only
very low levels of 03 (< 0.01 ppm) to build up in an atmosphere which is loaded
with a mixture of the oxides of nitrogen which is typical of the early morning
hours: [NOal/TNO] = O.J>. For reasons which we will explore in detail later, the
;N02]/[NO~! ratio increases dramatically as the sunlight irradiation of the pol-
lutant mixture continues, and the [03] is expected in theory to follow this
change approximately in accord with relation (l). Thus for the simulation shown
in Figure UU (solid curves) the values of the ratio [OsXNOl/CNOs1 are 0.0188,
0.0191, 0.019*^ 0.0197, 0.0193 Ppm at 10, 30, 60, 90, and 120 min, respectively,
very near the theoretical magnitude of ki/k3 for the conditions chosen. Since
the NO, NOa, arid 03 comprise a chemical system which is continually undergoing
change, the [03] at any time lags about a minute or two behind the value expected
for a true photostat ionary state. Although subsequent improvements in the theo-
retical and experimental estimates of the ratio ki/k3 are necessary and expected,
it appears from all available data that 03 development occurs largely through the
simple sequence of reactions (1-3) as Leighton suggested years
The real problem in the kinetic treatment of the ozone build up lies in the
realistic description of the processes which control the NO to NOs conversion
rate. The reaction (U ) can be important in generating a small level of NOs (up
137
-------
to 25$ of ttie NOX) during 'the early stages of the dilution of the exhaust of the
automobile, the power plant, or other sources which often contain 500 ppm of NO
initially.
(4) 2 NO + O2 -» 2 NO2
However reaction (k} is much too slow to account for any significant fraction of
the NO -* NOg conversion observed in the real atmosphere with typical ambient
levels of NO, 0.05-0.5 ppm. Thus in the simulation shown in Figure kk , the
observed rate of NO -* NOa conversion is l.&f x 10~3 ppm-min"1 at 10 min while
the rate of NOa formation in reaction (k) at this time is only 0.0046 x 10~3
ppm-min'1 .
It may be somewhat surprising to those unfamiliar with atmospheric reactions
to find that most of the chemistry that occurs in a sunlight- irradiated urban
atmosphere, including that responsible for the NO -* NO'a conversion, involves the
interaction of a vast variety of unstable, excited molecules, atoms, and free
radicals. The origin of a few of these potentially important species can be
appreciated from a consideration of a few of their major sources.
Ozone may react with NOa to create the symmetrical N03 radical in (5 ) :
i(5) O3 + NO2 -» NO3 + O2
The N03 species forms N20s by reaction with NO^ or oxidizes NO to
(6) (+ M) NO3 + NO2 -» N2OS (+ M)
(7) NO3 + NO -* 2 NO2
may dissociate to reform N03 and NOa or possibly react with water to
generate nitric acid:
(8) (+ M) N2OS -> NO3 + NO2 (+ M)
(9) N2OS + H2O ->• 2 HONO2
The two electronically excited and chemically reactive singlet molecular oxygen
species, Os^Ag) and 02(1£g+), are produced in the atmosphere by direct ab-
sorption of sunlight in reactions (lOa) and (lOb), by electronic energy transfer
from electronically excited NOa molecules formed by sunlight absorption by NOa
at wavelengths greater than 4000 A which provide insufficient energy to dis-
sociate NOa in reaction (l), or by ozone photolysis in sunlight, reaction (12):
. (10a) 1 12,700 A-»02('Ap
02 + sunlight <
(lOb) ( 7,600 A
(11) N02* +02 ^N02 +
(12a)
(12b) 03 + sunlight
02c)
2900-3060 A^02(1A^) + 0(1D)
2900-3500 A-* O2 + O(1D): or O(3 P) + O2(' Ag,
4500-7000 A -> O2 + O(3P)
138
-------
The 0(1D) species formed in (12a) and (12b) is much more reactive than the 0(3P)
ground state atom; it reacts efficiently when it collides with a water molecule
to form an important transient in smog, the hydroxyl radical:
(13) 0(1D) + H2O-" 2 HO
HO may react with NO to give nitrous acid in (lU) or with NOg to give nitric
acid in (15):
(14) HO + NO (+ M) -»• HONO (+ M)
(15) HO + NO2 (+ M) -» HONO2 (+ M)
A small concentration of HONO is expected, to appear in the polluted atmosphere
as result of reactions (l£)-(17) as well as (14), and it continues to supply
another source of the HO radical in (18):
(16a) H2O + NO + NO2 -»• 2 HONO
(16b) 2 HONO -» H2O + NO + NO2
•
(IT) HO2 + NO2 -»• HONO + O2
(18) HONO + /iK2900-4100 A) -»• HO + NO
However the homogeneous components of the rates of the reactions (l6a) and. (l6b)
are uncertain (see footnote in the introduction to this section).
A careful review of the net results of the occurrence of reactions (l)-(l8)
which we have outlined here will reveal that these reactions alone cannot explain
the rapid conversion of NO to NOa observed in the atmosphere. In fact if these
reactions alone occurred, we would expect that the original supply of NOg in our
atmosphere to be depleted somewhat as irradiation with sunlight occurred, and a
small and nearly constant [03] would be created in a very few minutes. The key
to the observed NO to NOa conversion lies in a sequence of reactions between the
transient species which have been generated and the other reactive molecules such
as carbon monoxide, the hydrocarbons, and the aldehydes present in the polluted
atmosphere. One such rational sequence of reactions was first delineated in-
dependently by two groups of scientists, Heicklen and coworkers!93 and Weinstock
and coworkerslo^. They noted that a chain reaction involved the HO and
radicals and CO may be important in driving NO to NOg in the atmosphere:
(19) HO + CO-» H + CO2
(20) H + O2 (+ M) -» HO2 (+ M)
(21) HO2 +NO-* N02 + HO
139
-------
Many cycles of the sequence of ((19), (20), (21), (19)....) may occur and many
molecules of NO may be oxidized to NOa for each HO radical that is formed
originally. •
Several other molecules present in the polluted atmosphere can assume the
potentially important role suggested for CO. Thus the aldehydes and hydrocarbons
may carry the chain and reform HOg radicals from HO radicals. With formaldehyde
the path is fairly simple:
(22) HO + CH2O-» HjO + HCO
(23) HCO + 02 -»• H02 + CO
(21) HO2 +NO-" HO + NO2
The aldehydes in the polluted atmosphere can function in another fashion to
generate the seemingly important HOg radical-^ . Thus formaldehyde is decom-
posed by sunlight according to the reactions (24a) and
(24a) CH20 + sunlight (3700-2900 A) -* H + HCO
i
(24b) -» H2 + CO
Both the formyl radical and the H atom formed in (2Ua) will react in air largely
to form H02 through (20) and (23).
A direct clue to additional steps in the mechanism of smog formation is had
from smog chamber studies of hydrocarbon-NO-NOa mixtures and the atmospheric
studies of the changes in the relative composition of the hydrocarbon pollutants
as the day progresses. For example, Stephens and Burlesonl95 found that the
complex mixture of hydrocarbon pollutants sampled in the atmosphere of the early
morning, contained a much greater fraction of olefinic hydrocarbons than a
similar sample taken in the late afternoon. Similar direct observations were
made on the composition of products of trapped auto exhaust-polluted air before
and after irradiation with ultraviolet lamps in the laboratory-^!? -19° . The
clear implication of these results is that the chemical reactions initiated in
the complex NO-NOg-hydrocarbon mixtures by the action of sunlight remove the
olefinic hydrocarbons at a much higher rate than the paraffinic hydrocarbons.
A great variety of excellent work in. both industrial and government laboratories
has helped establish various reactivity scales for hydrocarbons based on the
relative ability of the hydrocarbons to generate NOa from NO in chamber experi-
ments-^?. From these experiments and many others, scientists have concluded
that one or more intermediate species present in smog must react with the olefins
and ultimately creat HOa radical and organic peroxy radicals which stimulate the
observed NO to NOa conversion. Of course the olefins are transparent to the
sunlight within the lower atmosphere so they themselves are not acted upon by
the sunlight. However, when an olefin is added to an irradiated mix of the
oxides of nitrogen in air, we expect its removal to be relatively rapid through
the attack on it by the various reactive intermediates which we have discussed.
Thus in the case of the alkene, propylene, the following primary reactions may
occuir with these species:
-------
(2Sa) 0 + CH3CH=CHj -»• C3 H6O
(25b) -» HO + CH2=CHCH2
(26) 03 + CH3CH=CH2 -»• [CHjCHCHjOOQ]
(27a) HO + CH3CH=CH2 -» H2O + CH2=CHCH2
(27b) .->• CH3CHCH2OH
(21 c) -" CH3CHOHCH2
(28a) H02 + CH3CH=CH2 -»• CH3CHCH2O2H
(28c) -»• H2O2 + CH2=CHCH2
(29) NO3 + CH2CH=CH2 -> HONO2 + CH2=CHCH2
(30) O2(' Ag) + CH3CH=CH2 -»• CH2=CHCH2O2H
A great variety of chemical reactions, many of which remain unclear, follow
the formation of the initial products of reactions (25)-(jO); even the extent
of the partitioning of (27) between the several modes which are alternate in
theory, is not welL established today1^ . we will analyze the nature and the
extent of all of these reactions following the somewhat arbitrary but reasonable
scheme of Demerjian et al. Regardless of the details of these schemes, one
important result of all those proposed to date is the ultimate generation of
some number (a) of EOs, radicals, some number (p) of alkylperoxy (EO^) or acyl-.
peroxy (RCOOa) radicals, and a variety of different molecules which are largely
aldehydes, ketones, and acids:
HO (or other intermediate + RH (alkene, alkane, aromatic -*•
(31) in the mixture) hydrocarbon, aldehyde, etc.)
a HO2 + 0 RO2 (or RCOO2) + other products
The HOg radicals formed as a result of HO-alkene interactions in (31) may react
as before to cause the NO to NOs conversion in (21). Presumably the alkylperoxy
and acylperoxy radicals formed in (31) may oxidize NO as well in reactions (32)
and (33)5 "the analogues to (21) involving HOa radicals:
(32) RO2 + NO -> RO + NO2
(33) RCOO2 + NO -* RCO2 + NO2
The notorious peroxyacyl nitrates, of which peroxyacetyl nitrate is the
most common (R = CH3), will form in the atmosphere on the association of acyl-
peroxy radicals with N02-*-98:
(34) RCOO2 + NO2 ->• RCOO2NO2
There has been much speculation and a large uncertainty as to the relative
importance of the various atmospheric intermediates in the attack on the olefin
-------
hydrocarbons in smog.. Leighton1^ pointed to the probable importance in photo-
chemical smog of 0-atora and Os-molecule reactions with the olefinic hydrocarbons.
From the data available to him in 1961 he was unable to evaluate the relative
importance of these reactions compared, to those for the transients such as HO,
RQs., NOa, alkoxy, and alkylperoxy radicals, and singlet oxygen, 02(1A ) arid
02( £e+)' Since that time several investigators have given further insight into
the probable importance of some of these possible reactions. Thus it appears to
be generally accepted now that the theoretically calculated combined rate of
attack of 0(3P) atoms and 03 molecules on olefinic hydrocarbon molecules in the
photooxidation of NO-olefin systems in air may be significantly less than the
experimentally observed rate of olefin loss in such systems-1-99. Recently
Heicklen-^J and. Weinstock ° and their coworkers have presented, evidence that
the HO radical plays a major role in both the NO-CO and the NO-olefin photo-
oxidation chains in photochemical smog simulation studies. Also Bayes200-202^
Pitts^°5, Kummler \ and Berry2 5 and their research groups have stimulated
interest in reactions of singlet oxygen, OgC^g), in the atmosphere. Among the
many other possible reactions, the participation of this species in the olefinic
hydrocarbon removal reactions in photochemical smog has been suggested. The
extent of this involvement has remained untested. Recently Stephens and Price^l
" "I ^1 £> yOr*!
Wilson et al. •* , and Louw et al. have speculated on the possible significant
participation of NOa and NgOs in the chemistry of polluted air. Demerjian, Kerr,
and Calvert-^ considered this problem in their study completed in 1972. Using
an updated kinetic data base we will consider this system here.
The data of Figure U5 provide our best estimates of the rates of attack of
the various reactant species on the alkene for conditions used .in Figure kh
(dashed curves). The height of the ordlnant between the two curves defining
the rate for a given species should be compared. Obviously the kinetic data
suggest that the HO radical is the dominant olefin reactant for irradiation
times up to 120 min; ozone becomes increasingly important as its concentration
climbs. The rates of HOa radical attack appear.to be somewhat less significant
at all times. Note that although 0(3P) atom formation in reaction (l) is by
far the fastest photochemical reaction in the simulated smog mixture (d[o(3P)]/
dt = 2.98 x 10~2 ppm-min"1 at JO min, aldehyde-free system), the reaction of
0(3P) with Oa is so much faster than those with the alkene and the aldehydes
that 0(3P) is only a minor source of reactions which drive NO to NOz, The
rates of attack on alkene, for the CH30, NOs, and OZ^&P), presumed to be present
in the simulated smog mixture, are also unimportant in this regard, 9-7 x 10~7,
6.k x 10"8, and 3 x 10"9 ppm min'1, respectively, at JO min. When one compares
the rate of the attack on the aldehydes by the transient species, a much greater
discrimination which favors the HO reactions is seen than in the case of the
alkene reactions. For the simulations shown in Figure hh, about 99$ of the total
rate of the H-abstraction reactions from the aldehydes occurs by the HO radical
in reaction (22) and its analogues.
Now examine in greater detail the rates of the various reactions which are
responsible for the net NO -» NOg conversion at the shorter times. These data
are given in Table XXXVI as derived from the computer simulation corresponding
to Figure kk (solid curves). Note that at 10 min into the simulation NOa is
being destroyed by photodissociation at a rate which is 268.J x 10~4 ppm-min"1.
Several other loss reactions shown contribute another 9-3 x 10~4 ppm-min"1 to
the total loss rate of 277-6 x 10~4 ppm-min"1. You will note that N02 is being
formed at 10 min largely by reactions (j), (21), and (32),
-------
0(SP)
0 30 60 90 120
IRRADIATION TIME, min
Figure 4STrheoretical rates of reaction of various free radical species with
alkene in the simulated sunlight-irradiated, polluted atmosphere of Figure44(no
initial aldehydes).
(3)
(21)
(32)
O3 + NO -*• NO2 + O2
HO2 + NO ->• NO2 + HO
RO2 + NO ->• NO2 + RO
for which the rates are 21*0.7 x 10~4, 17.0 x 10"4, and 35-1 x 10
ppm-min"1,
respectively. The other reactions shown contribute 3-5 x 10~4 ppm-min"1 to the
total rate of formation at 10 rain, 296.3 x 10~4 ppm-min"1. Thus the net rate
of N02 formation given either by the tangent to the [N02]-time plot at 10 min or
by the summation of the individual reaction rates, is (296.3-277.6) x 10~4 =
18-7 x 10~4 ppm-min"1. The situation changes at longer times as the supply of
many of the reactants is exhausted; the rates of formation and destruction become
equal as [NOg] maximizes near 30 min, and there is a net rate of loss of N02 at
60 min. Thus the detailed mechanism predicts that the oxidation of NO to N02
is really promoted by H02 and R02 radicals; it is suggested that the alkyl
peroxyl radicals are about twice as important as the H02 species in the early
stages of the reaction.
From the rate data summarized in Tables XXXVTI to XLIII we can trace the
origin of the H02 and R02 radicals to derive the theoretical relative importance
of the various reactants and reactions as the sources of these important species.
In Table XXXVIl it can be seen that HONO photolysis is the major primary source
of the HO radical, while the formaldehyde and acetaldehyde photolyses are the
major primary sources of the H02 and R02 radicals. Note that these sunlight
induced rates of generation of the radicals H02 (6.11 x 10~4) and R02 (l*-33 x 10~4
1U3
-------
TABLE XXXVI. Rates of the Major N02 Formation and Decay Reaction in a
Simulated Smoggy Atmosphere.8
Reaction
---- d[N02]/dt,
At 10 min
ppn-min"1 x 10*
At JO min
At 60 min
(a) Loss reactions:
+ NO
N02 + hv •+ 0(3P)
N02 + N03 •» N20s
N02 + CH30 -» CHaONOs
H02 + N02 -+ HONO + 02
N02 + CH3C002 -» CHgCOOeNOa
N02 + HO (+M) -» HON02 (+M)
N02 + 03 -» N03 + Qs
N02 + HC002 -* (HCOOaNOa) •*
HONOa + C02
Other loss reactions
Total rate of all loss reactions
(b) Formation reactions:
-268.3
-1.0
- 0.0?
-3.5
- 1. 2
- 1. 1
-1.2
- L2
- 0.0k
-277.6
-323.2
- 9.6
- 0.07
-5.3
- 1. 9
- 1. 1
4.0
- 1.8
-'0.08
-3^7. 1
-285.1
-22.7
- 0.05
-5.0
- 2. 3
- 1. 0
- 5.8
- 1.9
- 0.08
-323. 9
Os + NO -» N02 + 02
ROs + NO -» RO + N02
H02 + NO -4 HO + N02
NjQs -» N02 + N03
N03 + NO •« 2N02
CH3C002N02 + NO -» 2HOg + CHgC02
Other formation reactions
Total rate of all formation
reactions
Net rate of formation of N02
21*0.7
35.1
17.0
1.0
2.3
0.1
0.1
296.3
+18.7
295.0
2l*.6
9.3
9.6
7,8
0.2
0.1
3^6.6
zM
265.0
15.6
5.*
22.7
llA
0.3
o.3
320.lt
il
alnitial concentrations (ppm) of reactants: [NO2]°, 0.025; [NO]0, 0.075; [trans-2-
C4 H8 ] ° , 0.10; [CH,O] ° , 0.10; |CH, CHO] ° , 0.06; [CO] ° , 10; [CH4 ] ° , 1.5; relative humidity,
50% (298°K); solar zenith angle, 40°.
ppm-min"1 .at 10 min) are not sufficient to account for the rate of the reactions
H02 + NO -» RO + N02 (35-1 x 10~4 ppm-min~4). Obviously there is a chain re-
action involving these species and CO, CHsO, CH3CHO, and C4H8 molecules which
develops the "extra" radicals which help drive WO to N02. For the reactant
conditions which we have chosen in the simulation here, the chain length for
the various species can be estimated. In Table XXXVIII we see the accounting
of the rates of the major photochemical and thermal reactions leading to HO
formation and destruction. Similar bookkeeping is shown for the H02 and the
R02 radicals in Tables XXXIX and XL, respectively. The chain lengths for the
various species calculated from these data are summarized in Table XLI. Thus
at 10 min into the simulated sunlight irradiation, chain -lengths are of the
order of 8 for the HO and R02 species and about U for the H02 radical. These
lengths decrease at longer times as the olefin is consumed. Termination of
the chains occurs by a variety of reactions involving all the radical species;
the major processes are shown in Table XLII. It is seen that the termination
of chains by H02-H02 interaction and H02-N02 disproportionation are highly
favored over reactions involving HO, RO and R02 radicals. This accounts for
the shorter chain lengths involving the H02 radical. We see as expected that
the rate of radical removal in the termination reactions is equal within the
-------
Table XXXVII. The Theoretical Primary Rates of the Major Radical
Generation Reactions in the Simulated Polluted Atmos-
phere .a
Reaction
Rate, ppm-min"1 x 10*
At 10 min
At 30 min
a) HO formation:
MONO + hv -« HO + NO
03 + hv -» 0('D) ... (+HsO) -» 2HO
HeOa + hv -» 2HO
No2,0a + hv -* o(3p) ... (+RH) -» HO
Total primary rate of HO formation
b) VOs formation:
2.1*5
0.07
0.06
0.005
+ hv -» H + HCO
H + 02 -» H02
HCO + 02 -» H02 + CO
CHaCHO + hv -» CHa + HCO
HCO + 02 -» H02 + CO
H02 . . .
(2.13)
2.13
L52
(1.85)
L32
Total primary rate of HOg formation
c) R02 formation:
6.11
0.19
0.25
0.007
(2.23)
2.23
1.59
(2.37)
1.69
2.12
L63.
At 60
5.27
0.31
0.62
0.007
6.21
(2.20)
2.20
1.57
(a 70)
1.93
CH^ + hv -» H + HCO
HCO + 02 -» HCOOj?
CHgCHO + hv -. CHa + HCO
HCO + 02 -» HC002
CHa + 02 -* CHa02
No2,o3 + hv -» o(3p) ... (+RH) -» R02
Oa + C«Ha •» R02 ...
Total primary rate of KOg formation
Total primary rate of formation of
all free radj cals
(2.13)
0.61
(1.85)
0.53
1.85
0.20
l.ll»
^53
13.Q
(2. 23)
0.61*
(2.37)
0.68
2.37
0.18
2.12
5.99
22ti
(£20)
0.63
(2.70)
0.77
2.70
0.09
1.89
6.08
19.9
Initial concentrations of reactants as specified in Table XXXVI.
precision of our bookkeeping to the primary rate of radical generation.
Now we are in a position to investigate the net effect of each type of
rea.ctant on the rate of the NO conversion to NOa according to the chemical
scheme adopted here. During the important early period of the irradiation in
which the oxidation of NO is most rapid, we have estimated the percentage of
the total rate of NO oxidation (by H02 and ROa) which results directly from the
participation of each of the major reactants. This breakdown is shown in
Table XLIII. Clearly the results of this theoretical model confirm the common
belief that the olefin is the major source of the H02 and R02 reactants (about
70$ at 10 min) which convert to NO to N02 and hence establish the 03 level
through relation (l). The two aldehydes considered in our synthetic smog
mixture contribute only about 12$> each, while an even smaller amount, about 5T°3
of the NO -» N02 conversion is a direct result of the CO reactions for the typical
impurity concentrations chosen.
1U5
-------
Table XXXVIII. Theoretical rates of the
Reactions in the
Major HO Formation and Loss
Simulated Polluted
Atmosphere.
Reaction
a) Formation reactions:
All photochemical sources (Table II )
HOg + NO -» HO + K02 •
CHgCHOeH -» CHaCHO + HO
H02 + 03 -> HO + 202
Total rate of HO formation from
all sources shown
; b) Loss reactions:
'< HO + C4Ha -* HOC-tHe I
: HO + C^He •* HaO + CHaCH=CHCH3 J
HO + CHaO -» HeO + HCO
HO + CHsCHO -» HsO + CHaCO
HO + CO -» H + CO?
1 HO + NO (+M) -» HONO (+M)
' HO + H0£ (+M) -• HON02 (+M)
l
i Total rate of HO destruction from
i all sources shown
__-.__ Pn-fo
At 10 ndn
2.58
17.00
1.47
0,09
21.1
11.28
2.88
aos
3.20
0.60
1.07
21.1
g
Initial concentrations of reactants as
ppm-min"1 x 10
At 50 min
5.1A
9-32
1.'32
0.31
16A
7.10
2.68
2.33
2.85
0.23
1.15
16.3
4
At 60 min
6.21
5.36
0.77
0.54
12.o
3.80
2. 5^
2.6l
2.75
0.12
0.98
12.8
specified in Table XXXVI.
Table XXXIX. Theoretical Rates of the Major HOa formation and Loss
Reactions in the Simulated Polluted Atmosphere.8
React xon
a) Formation reactions:
H + 02 (+M) •» H02 (+M)
(H from HO + CO -» H + C02
CKzfl + hv -» H + HCO)
HCO + 02 -» H02 + CO
(HCO from HO + OfeO -» T&) + HCO
CH3CH(6)CHO -» CHsCHO + HCO
CHaO + hv -• H + HCO
CH3CHO + hv -» HCO + CHa)
CHgO + Os •* H02 + CHsO
CH3CH(6)(6) + 02 -» H02 + CH3C02
(radical from 03-C4Hs reaction)
HC02 + Oa -» H02 + C02
Total rate of formation of H02
b) Destruction reactions:
H02 + NO -» HO + NOS
H0£ + 03 -* HO + 20a
H02 + N02 -• HONO + 02
H02 + C^Hs -> HOgC4He
2H02 -» Hs02 + 02
CHa02 + HOa -»-CH302H + 02
Total rate of loss of HDs
At 10 min
5.33
(3.20)
(2.13)
7.18
(2.17)
(2. 4l)
(1.61)
(0.99)
11.81
1.63
27.1
17.00
0.09
3.50
1.47
4.94
0.09
SLi
.x
ppro— roin~ x 1O
At 30 min
5.08
(2.85)
(2.23)
6.60
(1.88)
(2.00)
(1.56)
(1.16)
9.79
2.12
0.84
2U.lt
9.32
0. 31 '
5.34
L32
7.9U
0.22
24.4
At 60 min
4.95
(2.75)
(2.20)
6.09
(1.64)
(1.77)
(1.42)
(1.24)
T.Wt
1.89
0.55
20.9
5-36
o. 5U
U.99
0.78
8.90
0.33
£0.9
Initial concentrations of reactants as specified in Table XXXVI.
-------
Table XL. Theoretical Rates of R02 Formation and NO Oxidation to
by the Major ROa Radicals Formed in the Simulated Polluted
Atmosphere.8
Radical
Primary sources of the radical Rate, ppm-min^xlO*
' At 10 min At 30 min At 60 min
CHaOa CHa + Os;
a) CHa from CHaCHO photolysis
b) CHaCH(OH)6 •* CHa + HC02H
(radical from HO addn C^)
c) CHaCOa -» CHa + C02
[CHaCH(6)(6) + Os •« CHaCOa
+ HOai radical from 03 +
1.85
8.03
2.00
(1.13)
2.33
5.05
2.59
(2.12)
2.70
2.71
2.24
(1.88)
[CHaCOOa + NO •» CHaCOa + N0e;
radical from HO attack on
CHaCHO] (0.86)
HCOOE
CHaCOOs
CH3CH(OH)02
CH3CH(OH)CH(62)CH3
i — O-O - ,
CH2CH(62)CHCH3
HCO + Os; HCO from ClfeO and
CHaCHO photolysis and HO
attack on CHaO
CHgCO + 02; CH3CO from HO
attack on CHaCHO
Formed in reaction sequence
following HO addn. to C4Ha
Formed in reaction sequence
following H-abstraction from
C*He by HO and other radicals
0.63
0.87
8.03
8.03
3.21
CHaCH(HOa)CH(6s)CH3
CH3CH(CHaO)CH(6a)CH3
Formed in sequence following
HOa addn. to
Formed in sequence following
addn. to
0.02
(0.47)
0.84
0.47
5.05
5.04
2.02
Total rate of NO oxidation b^ all ROg1 s shown
Tapproximately equal to rate of ROs
formation)
(0.36)
0.54
0.36
2.70
2.70
0.76
0.00
1-31
o.oi
24.6
'initial concentrations of reactants as specified in Table XXXVI.
The Theoretical Concentrations of.HO and HOg _vn the Simulated Smoggy Atmospheres
and Related Data from Real Atmospheres.
It is of some interest to the kineticists and spectroscopists to observe
the concentrations of the HO and HOg species which are predicted to exist in
the simulated atmosphere. These values are shown in Figure 46 for the conditions
employed in the experiments used to construct Figure hk. The dashed curves and
the solid curves, respectively, correspond to the simulations with and without
aldehydes present initially. Values for [HOg] climb to about 7 x 109 molec-cm~3
in the run with aldehyde present initially and to about 5.k x 1C9 molec-cm~3 in
the runs without initial aldehyde. The theoretical HO levels are much lower
than those of H02s in the range of (2-7 ± 0.3) x 10s molec-crrf3 for these con-
ditions. The concentration range of these transients is neither sensitive to
alternative choices of most of the kinetic parameters employed nor to the initial
concentrations of the impurities in our synthetic polluted atmosphere. The first
of these points can be illustrated using the data of Figure Vf. The rate constant
-------
Table XLI. The Theoretical Chain Lengths for the HO, H02, and R02 Radical
Reactions in the Simulated Polluted Atmosphere.8
Radical
HO
H02
R02
Primary rate of generation,
ppm-min"1 x 104
Primary rate of generation,
ppm-min"1 x 104
Total rate all reactions,
ppm-min -1 x 104
Chain length
Primary rate of generation,
ppm-min"1 x 104
Total rate all reactions,
ppm-min"1 x 104
Chain length
Primary rate of generation,
ppm-min"1 x 104
Total rate all reactions,
ppm-min"1 x 104
Chain length
At 10 min
a 58
21.1
8.2
6.11
27.1
h.k
^33
35.1
8.1
At 30 min
5.«A
16.3
M
7.63
2k. k
2il
5.99
2k. 6
k.i
At 60 min
6.21
12.8
2.1
7.59
20.9
2.8 .
6.08
15.6
2.6
Initial concentrations of reactants as specified in Table XXXVI.
Table XLII. The Theoretical Rates of the Major Chain Termination Re-
actions in the Simulated Polluted Atmosphere.3
Reaction Rate of radical removal,
At 10 min At 3° min
2H02 -» H^g + 02
HO -y NO (+M) -. HONO (+M)
HO + N02 (+M) -» HON02 (+M)
H02 + N02 -» HOHO -I- 02
CH3Oa + H02 -» CH302H + 02
CHaO + NO -» CH3ONO
CHaO + N02 -* CH3OH02
CH30 + N02 -» CHjjO + HONO
CH3C002 + N02 -» CHsC002fI02
HCOOg + N02 -» (HC002N02) -» HON02 + C02
N03 + H02 -* HON02 + 02
Total rate of HO, HOH, RO, and R02
removal from above reactions
k.9k
0.60
1.07
3.50
0.09
0.0?
0.07
0.01
1.21
1.25
0.00
12.8
7-9)*
o. 23
1.15
5. 31*
0. 22
0.01
0.07
0. 01
1.85
1.80
0. 01
18.6
ppm-min"1 x 104
At 60 min
8.90
0.12
0.98
^•99
0.33
0.00
0.05
0.01
2.26
1.88
0. 04
19^6
Total primary rate of HO, H02, RO,
and R02 fteneration T/Table IlT~ 13.0 19.1 19-9
3.
Initial concentrations of reactants as specified in Table XXXVI.
-------
Table XLIII. The Fractional NO -» N02 Conversion Rate Due to the Various
Reactants in the Simulated Polluted Atmosphere at 10 min
into the Reaction.8
Radical
Total
-- Rate of NO -* NOs conversion, ppn-min'1 x 104
CO CHaO CHsCHO
H02
R02
5.20
Total 3. 20
0.00
U.67(hv)
2. 10(HO)
6/rr
0. 36 (hv)
0. 49(HO)
3.o6(hv)
0.86(HO)
5.92
2.07(hv)
1.73 (HO)
2.96(HO abst. )
8. 03 (HO addn. )
2.27r03,0{3P)]
13.26
2^.09 (HO addn. )
3. 75 (HO abst.)
1.13C03, 0(3P)1
1. !»7(H02 addn. )
o.oo 0.85
0.02(CH30 addn. )
5.80 30.1*6
NO -» HOg convers ion
directly attribut-
able to reactant 5.1
12.2
12.4
Initial concentrations of reactants as specified in Table XXXVI.
for the important H02 termination reaction (17), H02 + N02 -» HONO + 02, was
"
~2
"1
mn"1 here12^ rather th
"1
used as 5-0 x 10 ~ ppm
the value we now accept as
much more reliable, k17 = 2.9 x 101 ppm"1 min'1. Compare the curve repre-
senting the [alkene]0 = 0.1 ppm in Figure 1+7 with the solid curve for [HO] in
Figure U6. Although the shape of the profile has changed significantly, the
predicted levels of HO are of the same general magnitude, (2.0 ± O.U) x 10s
molec-cnf3, as those observed in Figure h6 for the same initial concentrations
of reactants, (2.7 x O.U) x 10s molec-cm"3. The relative insensitivity of [HO]
to alkene level is seen in Figure Vf. In these runs the [alkene]0 was varied
from 0.01 to O.I+O ppm while keeping all other reactants fixed as in the Figure 4
simulations. The highest levels for the HO radical are achieved at some inter-
mediates value of the added alkene near 0.05 ppm. Increase or decrease of the
alkene from this value lowers the [H0]-time profile somewhat. Of course the
decrease at high [alkene] reflects in part the greater rate of its removal re-
action involving the olefin. However the magnitude of the [HO] stays within the
range ( 2 ± 0.7) x 10s molec-cm"3 over a wide range of olefin levels.
In Figure U8 compare the [H0]-time profiles expected in theory for the
simulated smog-like atmospheres with, varied [N0]° from O.OOJ to 1.20 ppm. The
rate of HO generation is very low for the mixtures at low [NOX]; the HO removal
rates are high, largely a result of HO attack on the alkene; this results in a
rather low steady state concentrations of HO; e.g., [HOJ = 6 x 10s molec-cm"3
for the [NO]0 =* 0.01 case. At somewhat higher NO levels the chain reaction
involving H02 + WO -» HO + N02 becomes more important and HO levels rise. At
very high NO concentrations the suppression of [HO] occurs in part through the
reaction. HO + NOa (+ M) •* HONCfe (+ M) and the other chain termination reactions
-------
30 60 90
IRRADIATION TIME, mill
i
o
4 =!
o
ta
o
s
120
Figure46.Theoretical HO and HO3 concentrations as a function of irradiation
time in the simulated sunlight-irradiated, polluted atmospheres of Figure 44dashed
curves, aldehydes present initially; solid curves, aldehydes absent initially.
which involve NOX; see Table XLII. Note again however that with variation of
the [NO]0 from O.OU to 1.2 ppm the [HO] remains in the range (2 ± l) x 10s molec-
cm~3 over much of the time period shown.
The theoretical time profiles for the concentrations of HO and H02 in the
simulated smog runs with varied CO are shown in Figure k$. Observe that the
more significant effect of CO addition is on the [HO]• At [CO] = 50 ppm the
reaction with CO becomes a dominant loss mechanism for HO. Nevertheless the
range of [HO], (2 ± l) x 10s molec-cm~3 is still not large. The changes in
the rHOgl are much less pronounced as [ CO] is increased from 1 to 5° PPm- A
partial explanation of this effect which is expected in theory is that the H02
does not react significantly .with CO. and its rates of generation and destruc-
tion (largely 2 H02 "* H202 + 02 and H02 + NO -» HO + N02) remain about the same.
There is no direct measurement of the [HOal in "the real atmosphere of our
smoggy communities, but we can make some qualitative estimates using the analyti-
cal data for H202 determined by Gay and Bufalini20?. They estimated the con-
centrations of HgOa in the smoggy atmospheres of several urban communities.
The chemical kinetic and mechanistic data available to us today suggest that the
major source of Es02 in these atmospheres is the reaction. 2 H02 -» Hs02 + 02.
150
-------
7 a
E
M
U
_»
U
U
o
40
60
TIME, nin
120
Figure *»TThe theoretical effects of varied alkene concentration on the HO concentra-
tion in the simulated sunlight-irradiated, polluted atmospheres; concentrations of
impuritjes as in Figure44 except [alkene] is as shown (ppm); *,, = 5.0 X 10~2
pprrr* min"' is assumed in this case.
Our simulations show that the reactions such as EOs. + RCHO -» E^Og + RCO are
negligible in rate for the normal pollutant concentrations encountered. The
rates of radical attack on H^Oa and HaOa photolysis in sunlight are very slow
and can be beglected here. HeOp gay be destroyed in the real atmospheres by a
heterogeneous reaction with NOp. , but this is probably not fast. In Gay
and Bufalini's experiments the [HsQal built up to values as high as 18 pphm in
about 4-5 h period during a day of very severe smog in Riverside, California.
More typical levels of EsOs were those of k pphm found in the early afternoon
hours during moderate photochemical smog episodes in Hoboken, N. J., and River-
side, California. If we assume a uniform rate of formation of Hp.0a during the
4 or 5 h period of buildup to the k and 18 pphm levels attained then from the
known rate constant for the 2 HOa ~* Ha02 + HaO reaction (5-3 x 103 ppm"1 min"1)
we estimate that the average [HO?.] during these measurements was in the range
from k x 109 to 9 x 109 molec-cm"3. These estimates correspond well to the
range of levels estimated from our simulated polluted atmospheres in Figures k6
and U9; these varied from about 3 x 109 to about 7 x 109 molec-cm"3 for ir-
radiation times longer than about 30 min, and they were relatively insensitive
to the reactant concentrations over a rather wide range. Thus the limited
available data on E02 seem to be in good accord with the expectations of our
simulation. We may hope that the sensitivity of the new unambiguous cpectroscopic
151
-------
4|—
n
I
i
u
I
to
u
-I
9
U
w
o
0.30
40
80
TIME , min
120
Figuie48. The theoretical effects of varied [NO]0 on the HO concentration in the
simulated sunlight-irradiated polluted atmospheres; concentrations of impurities as is
shown in Figure44except [NO]0 is as shown (ppm); [NOJ0/[NO2]° = 3.0;*,, =>-
] 5.0 X IQ-* ppm
min"1 is assumed in this case.
techniques for HOa detection will be extended to allow direct observations of
the HOa concentrations in sunlight-irradiated, polluted urban atmospheres in
thp VPSTS ahpar? •'J -^ and a much mr>re critical test of reaction mechanisms
the years ahead
can be had.
Direct measurements of the [HO] in airohave been reported recently by
Wang and Davis^H. They detected the .3090 A fluorescence of the HO radical
excited with a tunable laser of output near 2825.8 A. The monitored sample
of air was brought into the apparatus from outside the Ford Motor Company
research laboratories near Dearborn, Michigan with a transport time of about
6 s. The estimated concentrations of HO varied from 1.5 x 10s molec-cm~3 in
the early afternoon (~ 2:00 P.M.) to 1.6 x 107 molec-cm~3 in the early evening
(~ 6:00 P.M.). These experimental [HO] estimates are about a factor of 50
times higher than the concentrations predicted from all of our simulations;
see Figures k6 UT, kQ, and 49- There seems to be little question that HO was
the species seen in the experiments of Wang and Davis in view of its unique
excitation spectrum, its rapid destruction when the incoming air was mixed with
butane, and the careful detail given to the elimination of HO production by
the laser beam itself. However in view of our simulations and other experimental
results which we will consider, we question the accuracy of the absolute magni-
tude of the Wang and Davis estimate. We cannot visualize any other major sources
152
-------
I
o
X 2
(0
Ul
3
o
O I
0 40 80 120
TIME, min
Figure49. The theoretical effects of varied [CO]° on the HO and HO2 concentrations
in simulated sunlight-irradiated, polluted atmospheres; initial concentrations as in
Figure44aldehydes present); [CO]0 (ppm) is as shown on curves; fc,, = 5.0 X 10";
ppm" min"1 is assumed in this case.
of HO than those included in our simulation, although this may indicate a weakness
in the present authors rather than a problem in the results of Wang and Davis.
Large variations in the proportions of the various impurity reactants was seen
to have little effect on the [HO] in our simulations, so the difference cannot
have its origin in some special blend of impurities in the Ford laboratories.
It is important to realize that if the concentration of HO which was piped
into the laboratory was at the 1.5 x 10s molec-cm~3 level after a 6 s transport
time, it would have to be even larger in the sunlight irradiated atmosphere
itself; the decay of HO in the dark is reasonably rapid for typical impurity
levels encountered. For HO at the concentrations reported by Wang and Davis
the rates of impurity reactant removal in the polluted atmosphere would be un-
realistically large: NOa would be removed at a rate greater than ~ h to $$> per
min. depending on our choice of rate constant for reaction (15)1^?212; the
alkenes in this hypothetical atmosphere would be removed remarkably fast as
well, greater than 15$ per min for C3H6, j8$> per min for iso-C4H8, and 6h% per
min for tra_ns_-2-C4H8. No such rates of removal of the impurity components of
smog has ever been observed in irradiated atmospheric s ample s •'•95 jJ-90 and in the
host of smog chamber experiments which have been reported through the years-*--?' .
New direct measurements of the concentration of HO in the real urban atmos-
pheres should be made to resolve this problem. When this is done again, the
153
-------
concentrations and the nature of the major impurities present in the air studied
should "be established within the same time period, and the dark decay time of
the HO species determined. Such measurements would provid.e a much less ambiguous
proof of the identification of HO, the modes of HO generation and destruction in
the real atmosphere, and help to establish reasonably accurate simulation models.
Theoretical Mass Balance of the Nitrogen-Containing Compounds Formed in the
Simulated Polluted. AtmospKere
The nature of the nitrogen-containing products formed in photochemical
smog has been a matter of considerable interest among scientists. It may be
instructive to note the distribution of the products predicted by the present
model. In Figure 50 the percentage of the total nitrogen which is present in a
given compound is plotted as a function of the sunlight irradiation time for
the synthetic polluted atmosphere of the composition employed in the mixture of
Figure 44 for initially aldehyde-free reactants. We see that the conversion
of NO to NOa is followed by a continuing transformation of NO and NOg into two
major products, nitric acid and PAN. A small level of nitrous acid build.s up
rather quickly as a result of the very fast reactions, HO + NO (+ M) -* HONO (+ M)
and HQa + NOs -» HONO + Q^, but its reasonably rapid photochemical decay, HONO +
hv -» HO + NOs prevents its growth to substantial levels in the atmosphere.
Methyl nitrate, like nitric acid., is a very weak absorber of sunlight in the
lower atmosphere, and it is reasonably stable chemically, so it is expected to
accumulate; however, the CH30 radical which is its source, is at very low con-
centrations. Thus very little CH3ON02 forms; 0.1$ of the total nitrogen is in
this compound at 120 min. Methyl nitrite is also formed, but its rate of photo-
chemical decomposition at the solar wavelengths below 4100 A in the lower atmos-
phere restricts its growth; only 0.002$ of the nitrogen is expected to be in
the CH3ONO. NaOs, not shown in Figure 50, also grows somewhat to account for
another 0.24f> of the nitrogen at 120 min. However it is highly speculative
whether this compound will accumulate to this extent or be removed as nitric
acid by reaction at moist aerosol surfaces. The transients N03 and, HNO remain
very low, amounting to only 7.4 x 10~3'and 5.7 x lO"9^, respectively, of the
total nitrogen at 120 min.
There are some data with which we can check the predictions of the model.
Both PAN and nitric acid and its salts have been observed in the atmosphere of
some cities, but generally PAW is in excess of the acid and its salts. In the
real urban atmosphere nitric acid does not seem to build up to the extent that
we predict from the present reactions scheme. Thus Miller and Spicer213a re-
ported 24 h average HONOa concentrations in the Los Angeles area of about 3 ppb
and 1-h average maximum values of 10 ppb. For the simulation shown in Figure 44,
HONOa levels reach about l8ppb after 2 h of irradia'tion. Presumably the nitric
acid created in the real atmosphere reacts with certain basic impurities which
are present also, e.g., NH3, the oxides of the metals, etc. The quantitative
treatment of such heterogeneous reactions is not possible at this time, and no
attempt has been made to include them in our mechanism. However, they must
occur in the real atmosphere, and nitric acid, may be converted efficiently to
ammonium nitrate if the concentration of ammonia impurity is sufficiently high.
Ammonium nitrate comprises approximately 10-15$ of the total airborne particles
in composite samples collected in the Los Angeles area over the year 1971-1972213°.
One other factor may have contributed to an overestimation of the HON02 formed
in our experiments. We used the Demerjian, Kerr, and Calvert12' estimate of the
rate constant for reaction (15). An equally reliable estimate of Tsang212 is a
factor of 3 lower for this constant. Thus if the Tsang estimate of k15 is used
151*
-------
60
70
60
O
Q.
O
O
50
u
(9
O
£40
4
o 30
20
10
NO
MONO,
—. HONO
30 60 90
IRRADIATION TIME , mln
CHSON02
BgureSO. The time dependence of the theoretical composition of the nitrogen-
containing compounds formed in the simulated, sunlight-irradiated polluted atmosphere
of Figure44(no initial aldehydes).
in simulations like those in Figure hk, only 13 ppb of HON02 forms after 120
min. It follows that the amount of PAN and HONOg observed in our simulation
could be more nearly equal with other reasonable choices of rate data.
The Effects o_f the Variation of_ the Concentrations of the Impurities o_n Pro sue:
Formation in the Simulated Sunlight-Irradiated Polluted Atmosphere
The saturated "reactive" hydrocarbons. When n-butane is added to the
components of the smog mixture at the 0.10 ppm level, there is little effect
on the NOa oxidation rate, maximum Os reached etc. HO radicals do attack the
alkanes at a moderate rate to abstract H atoms and generate alkyl percxy arid
EOs radicals. However total rates of radical production are altered by only a
few percent-^. Thus with paraffin hydrocarbon additions at the levels near
those of the olefins, the situation which exists in the real auto-exhaust-pol-
luted atmosphere in the early morning hours-^c^ v;e can expect only a relatively
small perturbation of the general reaction scheme outlined here for the simpler
compositions.
The effects of variation _of the olefinic hydrocarbon concentrations. Several
observations- have been made on the relationship of certain smog manifestations
on the concentrations of hydrocarbons and the oxides of nitrogen. Thus Faith,
155
-------
Renzetti, and Rogers 5 derived the relation of eye irritation to NOX and hydro-
carbon levels in smog chamber studies shown in Figure 51- In Figure 52 the
relationship between the maximum daily 1-h average oxidant levels and the 6-9 AM
average concentration of non-methane hydrocarbons is shown for several cities
as presented by Schuck, Altshuller, Earth, and Morton^ . This is the re-
lationship which was the basis of much of the original planning for the extent
of control of hydrocarbon emissions to ensure that the ambient air standard for
ozone would not be exceeded. Note the maximum values of oxidant observed de-
fine a curve which falls with decrease in hydrocarbon level, but the shape of
the curve is obscure at low hydrocarbon values, the region of greatest interest.
It is instructive to see similar predicted, correlations with our model of the
simulated atmosphere, recognizing the serious limitations which any such theoreti-
cal treatment must have with our present state of knowledge. The time dependence
of ozone as a function of alkene (trans-2-butene) level was determined in com-
puter simulations. In two series of experiments the initial concentrations of
the oxides of nitrogen were fixed at [N0]° = 0.075, and [N02]0 = 0.025 pprn, and
the alkene was varied; in one series the initial aldehyde levels were taken at
zero and in the other initial aldehyde levels were: [CH20]0 = 0.10, [CH3CHO]° =
0.06 ppm.. The effects are summarized in Figure 53- Here the 8-h integrals of
the [03] vs. time and the [PAN] vs. time plots are shown as a function of the
olefin concentration. Such dosage sources are of particular interest in health
related considerations. The dashed curves and solid curves, respectively, cor-
respond to runs with and without aldehydes initially present. The important
point to observe is that the variation of the integrated oxidant and PAN levels
with the initial olefinic hydrocarbon concentration, so important in our de-
terminations of standards, are altered dramatically by the initial presence of
aldehyd.es. A reference value of the J480 [03]dt data is had by the horizontal
dashed line drawn at 38.6 ppm-min, the0value of this integral if the maximum
allowable 1-h average [03] in the ambient air quality standards of the EPA
were maintained for the 8 h. It appears from these data that the 03 standard
could not be met if the aldehydes remained high, [ CE20'] = 0.10, [CH3CHO] = 0.06
ppm, even if nearly all of the olefinic hydrocarbon were removed. Also observe
the theoretically expected large enhancement of the integrated. [PAN]-time levels
which results when the aldehydes are present; acetaldehyde is the main cause of
this effect. We should learn from these data that the "true" relationship
between nonmethane hydrocarbons and maximum 1-h oxidant at low hydrocarbon
levels could be a critical function of a variable which is not routinely
measured, now, namely the concentration of the impurity aldehydes.
The effects _o_f variation of the concentrations of the nitrogen oxides.
It has been recognized for years from the results of smog chamber experiments
that there is an inhibiting effect on product rates and certain smog mani-
festations seen at very high nitric oxide concentrations. Note for example
ir. Figure 51 that if one considered the variation in eye irritation with runs
at a fixed total hydrocarbon concentration of about k.5 ppm, then the expected
eye irritation passes through a maximum level and decreases again as runs at
increasing [NOX] are considered, The inhibiting effect of high [NO"1 has been
observed in product rate data determined by Tuesday and coworkers in smog chamber
experiments using alkene-NOx mixtures^9Ta. We have applied, our smog simulation
model to this case as well. We calculated the time dependence of the theoreti-
cally expected ozone and PAN concentrations for runs at fixed hydrocarbon levels:
rC4H8]0 _ o.io ppm, [aldehyd.es]0 = 0; [00]° = 10 ppm; rCH4]° = 1.5 ppm; 5$
relative humidity; the initial ratio [NO]°/[N02]° was held constant at 3.0. The
significance of these calculations is seen in the data of Figure 54. The 8-h
156
-------
.Relation of eye irritation to hydrocarbon and oxides of nitrogen levels in
smog chamber experiments from Faith, Renzetti, and Rogers
0.30 r-
0.25
a
a
I
0)
O>
to
<
£
0-20
0.15
0.10
0.05
Los Angeles «
Los Angeles f+
Washington *^«**"** Denver
^' * Los Angeles
x-*^'Philadelphia
^' Los Angeles
Philadelphia * S
Philadelphia J
Washington^" /
Washington^ Philadelphia
Washington » » * » / » •
»/ »
A » V «»»
Washington ^ ^ ^ ^t
— ' «»»
* \
_L
_L
J_
J_
0 0.5 1.0 1.5 2.0 2.5
6-9 a.m. Average Nonmethane Hydrocarbon Concentration (ppmC)
Figure 52.Relationship between the maximum daily 1-h average oxidant levels and the
6-y AM average concentration of nonmethane hydrocarbons derived from the data from
several cities; from Schuck, Altshuller, Barth, and Morgan216-
157
-------
f*our i
'Jo I 5Jdt VALUE "
[os] • 0.08 pp« for Oh
1
10
20 SO
[C4 HB]° pphn
Figure 5 3, The theoretical effect of the reactive hydrocarbon concentration on the
8-h integral of the [O3] and [PAN] vs. time data derived from the simulated sunlight-
irradiated, polluted atmospheres; reactant conditions as in Figure44dashed curves,
aldehyde present; solid curves, aldehydes absent initially.
lOOOp-
c
6 100
I
E
ex
a.
g 10
-PAN-
I I
100
10
o
-------
integral of [o3]- and. [PAN]-time data are given as a function of the initial
[NO]. As one progresses from the very low level of [N0]° = 0.15 to 15 pphm
there is a gradual increase in the magnitude of the integrals, but further in-
creases of [NO]0 to JO, 60, and 120 pphm cause a suppression in the dosage
curves. The maximum in the integrals occurs near the stoichiometric mixture
of the alkene and NOX- These data should not be interpreted to imply that if
unrestricted emissions of NOX were allowed, our smog problem would be solved.
Of course one must be concerned about the health effects of elevated NOX levels
themselves. However it seems the problems of large 03 and PAN concentrations
would be delayed in time to some extent.
In this regard compare the [03] vs. time data given in Figure 55. The
initial concentrations (ppm) of the reactants in the simulation were: [alkene] =
0.10; [CO] = 10; [CHaO]0 = 0.10; J"CH3CHO] = 0.06; [CH4]° = 1.5; the relative
humidity was chosen to be 5$ (25 C); [NO]O/[N021° = 3-0. The initial level of
NO was taken as 0.15 PPffl in one case and 0.375 PPm i° the other. Note that at
the lower NO level the ozone climbs above 0.1 ppm after about U5 min of ir-
radiation. However about two hours are required for the mixture at the higher
initial [NO] to climb above 0.1 ppm level of 03, although the final level
reached after 6 h is not much lower than that of the mixture with [N0]° =0.15
ppm. The delay in ozone growth in the run at [N0]° = 0.375 ppm results from
the additional time necessary for the NO -» N02 conversion reactions to transform
the larger quantity of NO with basically the same radical supply, and to es-
tablish a suitable high ratio of [N02]/[NO] necessary to produce a significant
concentration of ozone.
One other factor which bears on the problem of smog suppression through
regulation of the NO pollutant levels can be illustrated in Figure 56. This
was constructed.from simulations in which the ratio of impurities was held
constant; [C0]° : [C4H8]° : [N0^° : [N02] : [OfeO^0:: 100: 1: 0.75: 0.25: 1:
0.6, but the concentration was changed over a 100-fold range: [alkene] = 0.01
to 1 ppm. The solid curves are for mixtures in which aldehydes are not present
initially and the dashed curves are for runs with aldehydes. Note from these
data that the dosages of ozone and PAN are not simple linear functions of the
impurity concentrations. Thus if a 10-fold dilution of the highly polluted
atmosphere occurs as an air mass moves across an air basin, the simulation sug-
gests that as little as a 3-8-fold reduction in the potential 03 8-h dosage
may result. With a 100-fold dilution as little as a 23-fold reduction in the
ozone dosage could result. Recent experience in the Los Angeles area probably
relates to the two phenomena described here. The ozone levels appear to be
down somewhat in downtown Los Angeles compared to previous years, probably as
a result of the [NO] increase which has occured in recent years; yet very high
levels of 03 and other smog products are seen in the Riverside area after some
modification of the Los Angeles mixture has occurred in route and the air mass
has been transported to Riverside by the prevailing atmospheric motion.
Effects of_ variation in_ the initial carbon monoxide concentrations. At
relatively low levels of CO in N0x-hydrocarbon polluted atmospheres there is
no dramatic effect seen in the rates of products formed, in simulation runs;
see Figure 57- An increase in the initial concentration of CO from 10 to 50
ppm decreases somewhat the 03 and PAN concentrations for a given sunlight
exposure time. As we have discussed previously, most of the attack on the
hydrocarbons and aldehydes occurs by HO radicals which are generated in the
smog system. The reaction of HO with CO becomes competitive with that of the
159
-------
0.2
O.I
[NO]°« 0.15 ppm
60
120 ISO 240
IRRADIATION TIME, mln
300
360
Figure 55. Theoretical effect of increased [NO]0 on the [OJ-time profile in
simulated sunlight-irradiated, polluted atmospheres; impurity concentrations as in Figure
44 '(aldehydes present initially), except for the oxides of nitrogen; [NO] °/[NO, ] ° = 3.0;
[NO]0 as shown on the curves.
,
/
10
io P-
480 -in
100
ppm— mln
1000
E
a
a
O
0.01
10,000
I I I I I
T'T
O.I
10
480 Kin
PAN] dl, ppm- mill
100
Figure 56. The theoretical effect of the dilution of a highly polluted atmosphere on
tne 8-h integrals of the [O3] and [PAN] vs. time curves obtained in simulated sunlight-
irradiated, polluted atmospheres; initial pollutant concentration ratios held constant:
CO:C4H, :NO:N02:CH,O:CH3CHO: 100:1:0.75:0.25:1:0.6 for dashed curves; aldehydes
were absent initially in the solid curves.
160
-------
20
ie
16
14
S"
W
o
o
o
' I ,2,000 I
20
60 80
TIME(mln)
100
120
Figure B7. The theoretical effect of variation of the initial concentration of the
carbon monoxide on the O3 and PAN formation in a simulated sunlight-irradiated,
polluted atmosphere; concentrations as in Figure44(aldehydes present initially) and
[CO]0 as shown with the curves; fc,, = 5 X 10'J assumed in this case.
alkenes and aldehydes at the higher CO concentrations, and an increasing share
of the chain regeneration of the HOa radicals from HO is then born by the
sequence (19)-(21). Lowered attack of HO on aldehyde and alkene occurs with
resulting lowered rates of alkylperoxy and acetylperoxy radical formation.
Thus at higher [CO] we tend to lose some of the boost in the NO to NOa conversion
associated with the alternative chain processes involving the hydrocarbons and
aldehydes. As a result of 03 concentrations are lowered somewhat. The lowered
rate of CH3C002 radical generation results in large part from the less important
attack of HO on acetaldehyde and alkene, and this is reflected in a lowered PAN
concentration since PAN comes largely for the following sequence:
(35)
(36)
(37)
CH3CHO + HO -* CH3CO + H2O
CH3CO + O2 -» CH3COO2
CH3COO2 + NO2 -» CH3COO2NO2
We see in Figure 57 that at extremely high levels of CO (2000 ppm), PAW formation
is practically eliminated since attack of HO on acetaldehyde and alkene is no
longer competitive with the CO reaction (19). Obviously we would not suggest
161
-------
that PAN formation be reduced in our polluted atmospheres "by removing controls
on CO emissions; the toxic properties of this compound alone then outweigh the
useful influence gained in theory by PAW reduction. Some free radical scavengers
have been considered by Heicklen and coworkers as possible inhibitors of smog
formation217. However the chain lengths in these systems are not very great and
relatively large amounts of even the most effective additives would be necessary
to have a significant influence. There are few, if any, compounds which would
not introduce a health hazard, as serious as the smog when present at useful smog
inhibiting levels.
One other previously unexpected aspect of the chemistry of CO-containing
atmospheres should be reviewed briefly here. If the hydrocarbon and aldehyde
impurities were entirely removed from the atmosphere and CO allowed to rise
along with the oxides of nitrogen, then CO may be a relatively effective reactant
to pump NO to N02, and significant ozone levels may be achieved1"". In this case
the only driving force for the reaction is the generation of HO radicals from
nitrous acid photolysis in (18). If the reactions of nitrous acid formation and
destruction (l6a, l6b) occur in the atmosphere, either homogeneously or hetero-
geneously, at rates comparable to those observed in chambers, then the CO-effect
can be significant for relatively small ambient levels of NOX and CO. In theory
sunlight-irradiated, moist air contaminated with only NO, N02, and CO at 0.075,
0.025, and 50 ppm, respectively, could generate ozone concentrations approaching
the 1-h maximum level of'0.08 ppm of the air quality standards after about two
and one-half hours of sunlight irradiation. One expects in theory that the same
kind -of influence noted here for CO, in mixtures of the reactive alkanes with
NOX, but free from alkenes.
Effects of S0g Addition on the Reactions in the Simulated Polluted Atmospheres
The available quantitative information related to the chemistry of SOg in
the urban atmosphere leaves many unanswered questions. However interest in the
sulfur dioxide removal mechanisms in the urban atmosphere remains high among
atmospheric scientists. The potentially harmful health effects associated with
urban atmospheres containing moderately low levels of SOa and. the apparent en-
hancement of these effects when there is a significant concentration of sus-
pended particulates and a relatively high humidity, have stimulated this interest.
The special problem of aerosols containing high sulfate levels has focused new
concern on atmospheric SOe levels since the recent CHESS report-*-. However there
is now little evidence as to the chemical nature of the active species responsible
for these observed effects. The detailed mechanism of sulfur dioxide oxidation
in the urban atmosphere remains unclear. The majority of existing evidence sug-
gests that a major fraction of the sulfur dioxide is ultimately converted to
sulfuric acid and sulfate salts, but the intermediate species involved and the
reaction paths which lead to these products are open to question.
The relatively slow rates of photooxidation of sulfur dioxide in air exposed
to sunlight""", and the demonstrated catalytic influence of certain solids and
moisture on the rate of S02 oxidation9~15 have led to the common belief that
heterogeneous paths for SOg oxidation probably far outweigh the homogeneous modes.
Although--this conclusion is probably true for the overcast, high humidity condi-
tions which prevailed during the Donora and London episodes in 19)48 and 1952,
respectively, it is by no means established that the conclusion is true for the
sunny, NO-NOs-hydrocarbon-polluted atmospheres often encountered today. There
is a real question as to the availability of sufficient reactive metallic oxide,
catalyst particles, and acid-neutralizing compounds in many atmospheres to promote
162
-------
removal by heterogeneous reactions at the observed rates. The most compelling
argument which has favored the importance of the heterogeneous removal processes
has been the apparent lack of alternative homogeneous reaction paths of sufficient
rate which might be involved. This situation has changed some in recent years.
Many research groups, including our own, have suggested some possible paths, and
these are summarized in Table XL IV. Shown are the approximate enthalpy changes
Table XLIV. Estimated Enthalpy Changes, Rate Constants of Possible
Homogeneous Elementary Reaction Paths for SOa in a Pol-
luted Urban Atmosphere.
Reaction
j kcal/mol
ka
Reference
(38) S02 + hv (02) -« S03 Sk
(39) o(3p) + so2 (+M) -• so3 (HM)
(1*0) 03 + S02 ->
(1*1) N02 + S02
(1*2) N03 + S02
(1*3) NaOs + SOs
(1.1*) <$£&2 +
(1*5) •CH200. +
CH2 =0*0 •
(1*6) H02 + S02
(1*7) H02 + S02
(1*8) CK302 + SO
(1*9) CHs02 + SO
. S03 + 02
-« S03 + NO
-» S03 + N02
• -» S03 + Na04
S02 -• S03 + 2CHS0
S02 -* S03 + CHaO
t S02 -* S03 + CHaO
-» HO + S03
- HOsSOs
a -» CH30 + S03
2 . ClfaOaSOa
(50) HO + so2 (+M) -» Hoso2 (+M)
(51) CH.O.SO.
-* CH3OS02
83
58
10
33
21*
8l
117
85
19
>25
30
>25
-37
.30
*so3 = 5 x 10~4 218
2.7 x Id9
<5 x ID"3
5.3 x 10-9
<1*.2
<2. 5 x I0"e
7
7
7
1.8 x 10s
7
-1.8 x 10s
?
3. !* x 10s
2.1* x 108
-1.8 x 106
240
128
222
128
128
221,223
221,223
230
(estd. )
231
232
(estd. )
aUnits on rate constants, 1 mol"'s"' for all reactions but 38 (dimensionlcss) and 39
(I'-mol-V).
for the reactions and the rate constants where they are known or can be estimated
approximately. All of the reactions shown are rather exothermic and are favored
from the viewpoint of thermodynamics. However the reactions (38)-(^3)> including
the photooxidation of S02, and oxidation by 0(3P), 03, NOg- N03, and HgOs, appear
to be severely rate limited. If one couples the quantum yield data of Cox21o for
the photooxidation of S02 in air with our estimated solar absorption rates of
S02 at z = U0030; the rates of S02 removal by the as yet unresolved detailed
mechanism of photooxidation, are very small. The Cox estimate of $gQ is some-
what lower than that of Allen, McQuigg, and Cadle^°, and it is considerably
greater than that of Friend, et al.2-"-9. However use of any of these rate data
does not alter the conclusion that the overall reaction (38) is relatively slow
163
-------
in the lower atmosphere. The 0(3P) reaction (39) is also slow but not negligible;
we will include this in our simulation. The homogeneous rate constant for (^0)?0
is reocrted by several workers to be too low to measure; for example see Cadle
Dauber.diek and Calvert12^ recently derived the upper limit for k4O given in
Table XLIV. Cox and Penkett found a measurable rate of reaction (kO ) cor-
responding to as much as 0.07^ conversion of S02 per hour, but it seems to us
that this may result from heterogeneous reactions. On the basis of our estimate
of k40 we conclude that (itO) is unimportant as a homogeneous removal path of S02.
The rate data for reaction (4l) were obtained by extrapolation to 298° K of the
data of Boreskov and Illarionov222 obtained in experiments at hjk-^Ok0 K; its
significant participation in the S02 removal paths in the atmosphere is also
excluded. The reactions (U2) and (^3) have been considered and envoked to ex-
plain Quantitatively product trends observed in some previous smog chamber
work^5-Lj206j but our recent estimates^- also rule out their significant con-
tribution in the urban atmosphere.
The reactions (kk) and (^5) of Table XLIV have been suggested recently by
Cox and Penkett 221,2 2J from kinetic studies of synthetic olef in-ozone-S02
mixtures at low concentration. They found that S02 is oxidized at an appreciable
rate by some reactive species formed in the dark on reaction between ozone and
alkene. The rate of ^° S02 removal per hour was observed when the alkene was
cis-2-pentene, while with propylene a rate of Q.U% per hour was seen. They sug-
gested that two likely candidates for the specific oxidizing agent in these ex-
periments are the ozonide of the alkene shown for that of ethylene in reaction
(kh] of Table XLIV, .'and the so-called "zwitterion" intermediate formed in the
Os-alkene reaction. Ripperton et al.22^ also suggested this role for the
zwitterion. Wilson et al. '. favor this interpretation in their computer
simulation of the Cox and Penkett reaction rate from cis-2-butene-ozone-S02
mixtures. The structure and reactivity of the zwitterionic species can be
considered to be diradical in character (as shown for the intermediate resulting
from ethylene in the first reaction (^4-5) shown in Table XLIV) or it may be like
an aldehyde oxide in structure (as shown for the second reaction (^5))- In
either case there seems to be little doubt that such transients would have very
high reactivity. However that one or both of these species may be ohe agent
which oxidizes S02 in these systems is open to question. For example if the
•CKsOO* entity is generated in the 02-rich atmosphere, it has several paths of
reaction which appear to be rapid in theory. We have speculated that the fol-
lowing sequence will generate the -OCHsO' radical-^? .
(52) -CH2OO + O2 -» -OOCHjOO-
/°-°x
(53) -OOCH200- •* 0 0 -» Oa + 'OCl^O-
^Ctts
O'Neal and Blumstein22" postulate that the same intermediate 'OCHaO- will be
generated quickly in the sequence (5*0, presumably in the absence of 02.
' (54) 0
->
0
In any case the «OCHaO* species might be expected to form very quickly in air
from the initial fragmentation product, 'CH200'. We have pictured it reacting
further with 02 to generate other radical species; e.g., -OCHsO- + 02 -» HC02 +
EOzi KC02 + 02 -» C02 + H02, both very exothermic reactions. O'Neal and Blumstein
suggested that the destruction of the 'OCJ^O* species (again presumably in an
Os-free system) occurs through H-atom transfer to form formic acid. If new
161*
-------
transients are sought to oxidize S02 in dilute 03-alkene-air mixtures it seems
to us that the •OCHaO" species (or its analogues in the higher alkenes) or other
radical species derived from it. might be a better choice than the original di-
radical fragment •CH200° (the so-called zwitterionic species). There is abundant
evidence that the fragmentation of the initial ozonide formed from the simple
alkenes creates highly excited free radicals including HO and other molecular
species22?. The chemiluminescence from such excited .species of the C2H4-03
reaction observed even at atmospheric pressure is sufficiently intense to allow
its use as a detection system for ozone today. In the Demerjian Kerr, and
Calvert mechanism of the Os-alkene reactions which we employed in this study,
the HC>2; HC02, CH3C025 CH3C>2, and other radicals are presumed to be formed in
subsequent reaction of the initial diradical species with oxygen12?. O'Neal
and Blumstein22° have suggested several other paths of ozonide decomposition
which better explain some of the chemiluminescent products observed in Os-
alkene reactions22?. The reactions of HO and HOg with S02 are fast and it is
reasonable to expect that other free radicals not yet studied may react rapidly
with S02 as well. One cannot at this point in time assess the rate constants
for reactions such as (44) and (45) since the observed reactions may be due to
the other reactive free radicals formed in these reactions. Thus we have not
included reactions (44) and (45) as elementary reactions in our simulations.
The H02 and HO radical reactions (46) and (50) have been studied recently,
and present kinetic data indicate their potential importance2?0"2?2. We have
included these reactions in our simulations. The analogous reactions of CH302
(reaction (48)) and CH30 (reaction (51)) would appear to be potentially signifi-
cant as well, although there are no experimental estimates of these rate con-
stants available;, our rough "theoretical" estimates are shown in Table XLIV.
•is-
In deriving our estimates given in Table XLIV we have made the assumption that
^46 = ^46- °ne would expect in theory that the preexponential factor for
these reactions would be nearly the same, and the activation energy for re-
action (48) may be somewhat lower than that for (46), since the overall re-
action (48) is somewhat, more exothermic. The value for k^ was estimated
assuming the reactivity of CH30 may be approximated by that for CF3. The
rate constant for HO addition to C2H4 is the same order of magnitude as that
for HO addition to S02- Thus the ratio of the measured rate constants for
CF3 and HO addition to the jt-system of C2H4 (6.J x 10~3)22° was used to
estimate k^ - 6.3 x 10*3 k^o. These are obviously only rough approximations.
There are no experimental kinetic data now available on the potentially
significant H02 and CH302 addition reactions (47) and (49). Our estimates12?
for the analogous addition reactions of these radicals with alkenes, based
largely on the considerations of Lloyd229j suggest that these reactions have '
considerably smaller rste constants (1/10,000th) than those of the HO addition.
If this relationship is maintained for the S02 system then the reactions (47)
and (49) would remove S02 at a rate of less than -0.1$ per h for a typical smog
mixture. We will neglect these reactions in our simulation although their true
importance remains uncertain.
Little is now known about the ultimate fate in the urban atmosphere for the
initial free radical products of the radical addition reactions ((47;, (49), (50),
and (51)). although we have speculated about the nature of these reactions233,93,^34,
Presumably radical addition to S02 would be followed by subsequent steps which
will lead to sulfuric acid, peroxysulfuric acid, alkyl sulfates. and various
165
-------
other theoretically possible precursors to sulfuric acid, nitric acid, and the
salts of these acids. All of the compounds should ultimately give sulfuric
acid, sulfate. peroxysulfate, nitrate, or nitrite salts as a further modification
of the primary products would occur, probably in a dynamic liquid phase on an
aerosol particle. One such reaction sequence which we felt may be important
in urban atmospheres and which will illustrate the possible mechanism for HO
attack on SOa in the following:
(50) HO + S02 (+ M) -> HOS02 (+ M)
(55) HOSOj + 02 -> HOS0202
(56) HOSO2O2 + NO ->• HOS02O + NO2
(57) HOS02O2 + N02 ->• HOS0202N02
(58) HOSO2O2 + HO2 -> HOSO2O2H + 02
(59) HOSO2O + HO2 ->• HOS02OH + O2
(60) HOSO2O + NO -» HOS02ONO
(61) HOSO2O + N02 -> HOS02ON02
(62a) HOS020 + RH -» HOS02OH + R
(62b) HOS02O + RH(alkene) ->• HOSO2ORH
(63) H2SO4 + H2O(NH3, CH2O, CnH2n, etc.) -»• Bowing aerosol
(64) HOSO2ONO2 + H2O -* H2SO4 + HONO2 (H^O) growing aerosol
(65) HOSO2ONO + sunlight -»• HOSO2O + NO
(66) HOS02ONO + H2O -» H2S04 + HONO ^ growing aerosol
Some of the species expected as products of the HO radical addition to SQ2 should
be highly reactive. Especially interesting is the presumed formation of an in-
organic analogue to the acylperoxy nitrates, HOSOsOgNOs, in reaction (57). Note
that free radical addition to SOs in a smog mixture need not result directly in
chain termination reactions. By analogy with the HOa radical reaction (21).
presumably the HOSOaOa radical would normally oxidize NO to HOg in reaction (56).
The reactivity of the HOS020 free radical may be similar to those of HO end RO
radicals. It should abstract H-atoms from alkene, alkane, or aldehyde (reaction
62a) and add to alkenes (reaction 62b) as. well as terminate chains through re-
actions (60) and (6l). The chain oxidation of the SQ2 molecules through the
regeneration of the HO radical in (67) is not likely since -this reaction is
very endothermic (AH = ^3 kcal-mole"1), and it could not compete with the
alternative reactions of this species.
(67) HOSO2O - SO3 + HO
166
-------
Any mechanism which we consider seriously should satisfy the few require-
ments which seem to evolve from the limited chamber data for WOx-hydrocarbon-S02
mixtures in a ir-^^3} 235 39. These results show: (l) the presence of S02 does
not alter significantly the ultimate ozone level reached or the rate of olefin
removal in the system. (2) The removal rate for S02 shows an induction period
characteristic of the time delay expected for the buildup of the reactive radi-
cal species (HO, H02, etc.) which are generated in the chain reactions involving
the alkenes. With the proper choice of rate constants for the sequence of
reactions (50), (55)-(67), one can easily satisfy these limited requirements.
At this point in time it will be sufficient to assume that the primary radical
reactions with S02 are rate determining in the removal of SOg, and that none of
the primary products regenerate S02.
The primary reaction scheme outlined for the SOa reactions in Table XLIV
was added to the general smog reaction scheme and simulations made. The results
of one such experiment are shown in Figure 58- The initial concentrations of
the impurities have been picked as follows: [S02]° = 0.10; [NOJ° = 0.15; [N02]° =
0-05; [C4Ha]° = 0.10: [CO]0 = 10; [CH4]° = 1.5; relative humidity, 5$ (25° C).
The length of the ordinate between curves defining each species represents the
theoretical rate of S02 removal by this species. We observe a near equal im-
portance of the H02 and HO reactions with S02 (U6) and (50)- Much smaller, but
much less certain, estimated contributions come from the CHs02 reaction (48),
the CH30 addition in (51), and o(3P) addition in (39)- S02 removal in these
reactions corresponds to a maximum rate of conversion of about !.!$> per h. In
similar situations with aldehydes present initially ([CffeO]0 = 0.10, [CH3CHO]° =
0.06), rates of S02 removal are a maximum of about 1.5$ per h. Obviously the
rates of homogeneous removal of S02 in a smoggy atmosphere may be significant.
When one recognizes that there are probably other processes analogous to (U8)
and (50) which may occur involving other R02 and RO species, and the likelihood
of the occurrence of other homogeneous reactions not yet identified, the gas
phase kineticist is assured of an important task in the ultimate characterization
and control of the S02- contain ing smog mixtures.
Summary
The apparent success today of computer modeling of photochemical smog
formation in both the smog chambers and the real urban atmospheres must be at
least in part related to the few demands which the present incomplete product
analysis data place on the modeler and his reaction schemes. Any computer
simulation of a complex system can only be good as the input parameters will
allow. One must not take too seriously the results from complex simulations
at this stage of our knowledge. In particular the development of sound re-
action schemes and realistic smog models require much more quantitative kinetic
information as to the detailed reaction paths which appear to be important in
photochemical smog. Specifically we feel that the most important needs today
are kinetic and mechanism studies related to the reactions of 03, HO. and H02
with the alkenes and aromatic hydrocarbons at low concentration in air at 1 atm
pressure. With the increasing use of high sulfur fuels it is imperative that
we delineate the S02 removal modes which are operative in the N0x-hydrocarbon
polluted urban atmosphere. There is no question that as such information be-
comes available, present models will require substantial changes. In this light
it should be apparent to those who use chemical models to simulate the photo-
chemical smog formation that there are potentially serious deficiencies in
present systems which must be adequately tested for and removed. One should not
167
-------
CHjO
30 60 90
IRRADIATION TIME , min
120
Figure 58. The theoretical rate of attack of various free radical species on SO, for a
simulated sunlight-irradiated (z = 40°), polluted atmosphere, initial concentrations
(ppm): [SO,]0 = 0.10; [NO]0 = 0.15; [N0,)° = 0.05; [C4HJ° = 0.10; (CO]° =
10; (CH4 ]° = 1.5; relative humidity, 50% (25°C); the total maximum rate of SO2
removal from the reactants shown is about l.lw per h.
at this stage of our knowledge of smog systems bet our lives on the development
of air pollution control programs based on these necessarily speculative schemes,
However it is our belief that computer modeling of photochemical smog will con-
tinue to be a very useful guide to the experimentalist, and it will ultimately
serve us well in comprehensive air shed models which are based on sound kir.etic
schemes.
168
-------
References
(l) Community Health and Environmental Surveillance System Committee of the
Environmental Protection Agency Report, "Health Consequences of Sulfur
Oxides," EPA-650/1-74-004, May, 1974.
(2) a) A. F. Altshuller and J. J. Bulfalini, Fhotochem. Fhotobiol, 14-, 97,
(1965); b) Ibid., Environ. Sci. Technol., 5, 39 U97TT
(3) R. D. Cadle'and E. R. Allen, Sci., Ib7, 243 (1970).
(4) M. Katz, Can. J. Chem. Eng., 48, 3 (1970).
(5) P. Urone and W. H. Schroeder, Environ. Sci. Technol., 3, 436 (1969).
(6) T. C. Hall, Jr., "Photochemical Studies of Nitrogen Dioxide and Sulfur
Dioxide," Ph.D. Thesis, Univ. Calif., Los Angeles, 1953-
(7) E. R. Gerhard and H. F. Johnstone, Ind. Eng. Chem., hi, 972 (1958).
(8) R. A. Cox and S. A. Penkett, Atmos. Environ., 4, 425 (1970).
(9) P. M. Foster, Atmos. Environ., 3, 157 (1969).
(10) H. F. Johnstone and D. R. Coughanowr, Ind. Eng. Chem., 50, 1169 (1958).
(11) H. F. Johnstone and A. J. Moll, Ihiol., 52, 86l (i960).
(12) C. E. Junge and T. G. Ryan, Quart. J. Roy. Meteor. Soc., 84, 46 (1958).
(13) M. J. Matteson, W. Stober, and H. Luther, Ind. Eng. Chem. Fundam., 8, 677
(1969).
(14) A. P. van den Heuval, B. J. Mason, Quart. J. Roy. Meteorol. Soc., 89,
271 (1963).
(15) M. Corn and R. T. Cheng, J. Air Polln. Control Assocn., 22, 871 (1972).
(16) This portion of the work under this grant was published during the
first year of our present study: F. B. Wampler, J. G. Calvert, and
E. K. Damon, Int. J. Chem. Kinet., 5, 107-117 (1973).
(17) S. Okuda, T. N. Rao, D. H. Slater, and J. G. Calvert, J. Fhys. Chem.,
73, 4412 (1969).
(18) T. N. Rao, S. S. Collier, and J. G. Calvert, J. Amer. Chem. Soc. , 91,
1609 (1969).
169
-------
References (continued
(19) C. C. Badcock, H. W. Sidebottom, J. G. Calvert, G. W. Reinhardt, and
E. K. Damon, J. Amer. Chem. Soc., 93, 3115 (19T1)-
(20) H. W. Sidebottom, C. C. Badcock, J. G. Calvert, B. R. Rate, and E. K.
Damon, J. Amer. Chem. Soc., 93, 3121 (1971)-
(21) E. Cehelnik, C. W. Spicer, and J. Heicklen, J. Amer. Chem. Soc., 93,
5371 (1971).
(22) F. B. Wampler, A. Horowitz, and J. G. Calvert, J. Amer. Chem. Soc., 9^>
5523 (1972).
(23) T. N. Rao and J. G. Calvert, J. Phys. Chem., Ik, 681 (1970).
(2U) A. Horowitz and J. G. Calvert, Int. J. Chem. Kinet., 14-, 175 (1972).
(25) A. Horowitz and J. G. Calvert, Int. J. Chem. Kinet., h, 191 (1972).
(26) H. W. Sidebottom, C. C. Badcock, J. G. Calvert, G. W. Reinhardt,
B. R. Rabe, and E. K. Damon, J. Amer. Chem. Soc., 93, 2587 (1971).
(27) H. W. Sidebottom, J. G. Calvert, B. R. Rabe, and E. K. Damon, to be
published.
(28) H. D. Mettee, J. Phys. Chem., 73, 1071 (1969).
(29) T. N. Rao, S. S. Collier, and J. G. Calvert, J. Amer. Chem. Soc., 91,
1616 (1969).
(30) H. W. Sidebottom, C. C. Badcock, G. E. Jackson, J. G. Calvert, G. W.
Reinhardt, and E. K. Damon, Enivon. Sci. Technol., 6, 72 (1972).
(31) This portion of the work under this grant was published during the first
year of our present study: A. Horowitz and J. G. Calvert, Int. J. Chem.
Kinet., 5, 2^3 (1973).
(32) G. E. Jackson and J. G. Calvert, J. Amer. Chem. Soc., 93, 2593 (1971).
(33) A. Horowitz and J. G. Calvert, Int. J. Chem. Kinet., h, 207 (1972).
(3U) H. D. Mettee, J. Chem. Fhys., ^9, 1784 (1968).
(35) H. D. Mettee, J. Phys. Chem., 73, 1071 (1969).
(36) S. J. Strickler and D. B. Howell, J. Chem. Phys. \9, 19^7 (1968).
(37) T. N. Rao, S. S. Collier, and J. G. Calvert, J. Amer. Chem. Soc., 91,
1609 (1969).
(38) T. N. Rao and J. G. Calvert, J. Phys. Chem., 7^, 681 (1970).
(39) A. Horowitz and J. G. Calvert, Int. J. Chem. Kinet., 4, 175 (1972).
170
-------
References (continued)
(40) A. Horowitz and J. G. Calvert, Int. J. Chem. Kinet., 4, 191 (1972).
(4l) H. W. Sidebottom, C. C. Badcock, G. E. Jackson, J. G. Calvert, G. W.
Reinhardt, and E. K. Damon, Environ. Sci. Technol., 6, 72 (1972).
(42) H. W. Sidebottom, C. C. Badcock, J. G. Calvert, B. R. Rabe, and E. K.
Damon, J. Amer. Chem. Soc., 93, 3121 (19T1).
(43) A. Horowitz and J. G. Calvert, unpublished results.
(44) a) G. M. Almy and P. R. Gillette, J. Chem. Phys., 11, 188 (1943);
b) S. S. Collier, D. H. Slater, and J. G. Calvert, Photochem. Photobiol.,
7, 737 (1968).
(45) A. Fatta and J. Heicklen, communication to the authors.
(46) F. B. Wampler, A. Horowitz, and J. G. Calvert. J. Amer. Chem. Soc., 94,
5523 (1972).
(47) This portion of the work under this grant was published during the first
year of our present study: F. B. Wampler, K. Otsuka, J. G. Calvert, and
E. K. Damon, Int. J. Chem. Kinet., 5, 669 (1973).
(48) E. R. Allen, R. D. McQuigg, and R. D. Cadle, Chemosphere, 1, 25 (1972).
(49) J. A. Davidson and E. W. Abrahamson, Photochem. Photobiol., 15, 403
(1972).
(50) S. S. Collier, A. Morikawa, D. H. Slater, J. G. Calvert, G. W. Reinhardt,
and E. K. Damon, J. Amer. Chem. Soc., 92, 217 (1970).
(51) K. Otsuka and J. G. Calvert, J. Amer. Chem. Soc., 93, 2581 (1971).
(52) J. C. D. Brand, V. T. Jones, and C. di Lauro, J. Molec. Spectrosc.,
40, 6l6 (1971).
(53) I- H. Hillier and V. R. Saunders, Molec. Phys. 22, 193 (1971).
(54) K. J. Chung and C. R. Clayton, Department of Chemistry, The Ohio State
University and the Battelle Memorial Institute, Columbus, Ohio 43210,
unpublished calculations.
(55) a) G. Herzberg, "Molecular Spectra and Molecular Structure. III.
Electronic Spectra and Electronic Structure of Polyatomic Molecules,"
Van Nostrand, Princeton, N.J., 1966, p. 605; b) J. C. D. Brand and
K. Srikameswaran, Chem. Phys. Lett., 15, 130 (1972); c) J. C. D. Brand
and R. Wanes, J. Mol. Spectrosc., in press.
(56) E. Cehelnik, C. W. Spicer, and J. Heicklen, J. Amer. Chem. Soc., 93,
5371 (1971).
(57) S. E. Braslavsky, and J. Heicklen, J. Amer. Chem. Soc., 94, 4864 (1972).
171
-------
References (continued).
(58) F. S. Dainton and K. J. Ivin, Trans. Faraday Soc. , h6, 37^ (1950).
(59) R. B. Timmons, Photochem. Photdbiol. , 12, 219 (19TO).
(60) K. L. Demerjian, Ph.D. Dissertation, The Ohio State University, Chemistry
Department, Columbus, Ohio, Winter 1973? to be published.
(6l) C. C. Badcock and J. G. Calvert, to be published.
(62) R. B. Cundall and T. F. Palmer, Trans. Faraday Soc., 56, 1211 (1960).
(63) a) R. J. Cvetanovic, Can._J. Chem. , 38, 1678 (i960); b) R. J. Cvetanovic
and R. S. Irwin, J. Chem. Phys., U6, 1702 (1967).
(6*0 This portion of the work under this grant was published during the
second year of our present study: K. L. Demerjian, J. G. Calvert, and
D. L. Thorsell, Int. J. Chem. Kinet. , 6, 829 (197*0-
(65) G. M. Bristow and F. S. Dainton, Proc . Roy. Soc. A, 229, 525 (1955).
(66) R. D. Penzhorn . and H. Gusten, Z. Naturforsch. , 27a, 1^01 (1972).
(67) R. A. Cox, J. Photochem. , 2, 1 (1973-710.
(68) P. W. Jones and A. H. Adelman, Environ . Sci . Techaol. , 6, 933 (1972).
(69) J. G. Calvert and J. N. Pitts, Jr., "Photochemistry," New York, 1967,
p.
(70) G. E. Jackson and J. G. Calvert, J. Amer. Chem. Soc., 93, 2593 (1971).
(71) C. G. Hatchard and C. A. Parker, Proc. Roy. Soc. A, 235, 5lB (1956).
(72) D. J. Brookman and D. T. Sawyer, Anal. Chem., ^0, 106 (1968).
(73) D. M. Golden, K. W. Egger, and S. W. Benson, J. Amer. Chem. Soc. , 86,
(196*0.
S. Sato, K. Kikuchi, and M. Tanaka, J. Chem. Phys., 39, 239 (1963).
(75) M. Tanaka, T. Terumi, and S. Sato, Bull. Chem. Soc. Japan, 38, 16^5
(1965).
(76) G. A. Haninger, Jr., and E. K. C. Lee, J. Phys. Chem., 71, 310^ (1967).
(77) E. K. C. Lee, H. 0. Denschlag, and G. A. Haninger, J. Chem. Phys., '48,
(1968). ~
(78) K. Fukano and S. Sato, Kogyu Kogaku Zasshi, 72, 213 (1969).
(79) Y. G. Shekk and M. V. Alfimov, Opt. Spect . , 32, 335 (1972).
172
-------
References (continued)
(80) F. J. Wagner and G. S. Hammond. Adv. Photochem., 5, 21 (1968).
(8l) R. E. Rebbert and F. Ausloos, J. Amer. Chem. Soc., 87, 55^9 (1965).
(82) W. A. Noyes, Jr., G. B. Porter, and J. E. Jolley, Chem. Rev., 56, 1+9
(1956).
(83) J. W. Gall, D. H. Slater, and J. G. Calvert, in "Chemical Reactions in the
Urban Atmospheres," American Elsevier, New York, 1971, p. 133-
(8U) J. H. Moore, Jr., J. Fhys. Chem., 76, 1130 (1972).
(85) W. Flicker, 0. Mosher, and A. Kupperman, to be published.
(86) R. A. Caldwell and S. F. James, J. Amer. Chem. Soc., 91, 518^ (1969).
(87) I. E. Kochevar and P. J. Wagner, J. Amer. Chem. Soc., 9k, 3859 (1972).
(88) L. Stockburger, III, S. Braslavsky, and J. Heicklen, "Photolysis of S02
in the Presence of Foreign Gases. III.—Quenching of Emission by Foreign
Gases," preprint of work provided by Heicklen to the authors.
(89) E. Cehelnik, J. Heicklen, S. Braslavsky, L. Stockburger, III, and E.
Mathias, "Photolysis of S02 in the Presence of Foreign Gases. IV.—
Wavelength and Temperature Effects with CO," Center for Air Environment
Studies Publ. 2*1-1-72, Pennsylvania State University (Oct. 1972).
(90) A. M. Fatta, E. Mathias, J. Heicklen, L. Stockburger, III, and S. Braslavsky,
"Photolysis of S02 in the Presence of Foreign Gases. V.—Sensitized
Phosphorescence of Biacetyl," Center for Air Environment Studies, Publ.
217-71, Pennsylvania State University, (Oct. 1972).
(91) See section I-F; also Ph.D. dissertation of K. J. Chung, Ohio State
University, Columbus, (197*0-
(92) J. D. Brand, V. T. Jones, and C. di Lauro, J. Molec. Spectro., kO, 6l6
(1971).
(93) J. G. Calvert, "Interactions of Air Pollutants," in Proc. Conf. on Health
Effects of Air Pollutants, pp. 19-101, National Academy of Sciences,
October 3-5, 1973, Stock 5270-02105, U.S. Government Printing Office,
Washington, D.C., 20^02.
(9*0 The work reported in this section was completed during the present study
and was published during the present year of the study: K. L. Demerjian
and J. G. Calvert, Int. J. Chem. Kinet., 7, ^5 (1975).
(95) a) Y. Hamada and A. J. Merer, presented at the 29th Symp. on Molecular
Structure and Spectroscopy, Ohio State University, Columbus, June Ik, 197^,
paper FB3; b) D. R. Humphrey, J. L. Hardwick, J. C. D. Brand, Y. Hamada,
and A. J. Merer, presented at the 29th Symp. on Molecular Structure and
Spectroscopy, Ohio State University, Columbus, June 1^, 197*t, paper FBI.
173
-------
References (continued)
(96) J. R. McDonald and L. E. Brus, presented at the 29th Symp. on Molecular
Structure and Spectroscopy, Ohio State University, Columbus, June 14,
1974, paper FB2.
(97) H. W. Sidebottom, J. G. Calvert, K. Chung, and E. K. Damon, in
preparation.
(98) H. W. Sidebottom, K. Otsuka, A. Horowitz, J. G. Calvert, B. R. Rabe, and
E. K. Damon, Chem. Fhys. Lett., 13, 337 (1972).
(99) J. G. Calvert, Chem. Phys. Lett., 20, 484 (1973)-
(100) H. D. Mettee, J. Phys. Chem., 73, 1071 (1969)-
(101) The work reported in this section was completed during the present study
and was published during the present year: K. Chung, J. G. Calvert, and
J. W. Bottenheim, Int. J. Chem. Kinet., 7, l6l (1975).
(102) R. A. Cox, J. Phys. Chem., 76, 8l4 (1972).
(103) E. R. Allen'-and J. E. Bonelli, Abstracts, 6th Informal Photochemistry
Conference-, Oklahoma State Univ., June, 1972; we thank Dr. Allen for
providing us with the quantum yield data obtained in these experiments
before publication of this work.
(104) H. N. S. Schafer, Anal. Chem., 39, 1719 (1967).
(105) J. W. Laxton and P. J. Jackson, J. Inst. Fuel, 12 (1964).
(106) P. W. Schenk and H. Platz, Z. Anorg. u. Allg. Chem., 222, 177 (1935).
(107) A. Kurtenacker and E. Furstenan, Z. Anorg. u. Allg. Chem., 215, 257 (1933)
(108) G. Nickless, "Inorganic Sulfur Chemistry," Elsevier, New York, 1968,
ch. 14.
(109) R. G. W. Worrish and G. A. Oldershaw, Proc. Roy. Soc. A, 249, 498 (1959).
(110) N. Basco and R. D. Morse, Chem. Phys. Lett., 20, 557 (1973).
(ill) F. C. James, J. A. Kerr, and J. P. Simons, Chem. Fhys. Lett., 25, 431
(1974); the authors are grateful to Dr. Kerr for a preprint of this
work given to them before publication.
(112) D. S. Sethi, J. Air.. Folln. Control Assocn. , 21, 4lB (1971).
(113) P. J. Warneck, G. C. A., Technology Div., Belford, Mass., personal
communication to one of the authors (J.G.C.).
(114) E. R. Allen, private communication to one of the authors (J.G.C.).
(115) R- L. Daubendiek and J. G. Calvert, paper in preparation.
174
-------
References (continued)
(llo) K. Schofield, J. Fhys. Chem.. Ref. Data, 2, 25 (1973).
(117) C. J. Halstead and B. A. Thrush, Fhotochem. Fhotobiol., k, 1007 (1965).
(118) K. H. Hoyermann, G. Wagner, and J. Wolfrum, Ber. Bunsenges. Physik. Chem.,
71, 605 (1967).
(119) R. J. Donovan, D. Husain, and P. T. Jackson, Trans. Faraday Soc., 65,
2930 (1969).
(120) W. H. Breckenridge and T. A. Miller, J. Chem. Phys., 56, 465 (1972).
(121) J. G. Calvert, Chem. Phys. Lett., 20, l&k (1973).
(122) L. E. Brus and J. R. McDonald, Chem. Phys. Lett., 21, 283 (1973).
(123) H. W. Sidebottom, J. G. Calvert, K. J. Chung, and E. K. Damon, to be
published.
(12^) This portion of the work under this grant was"published during the study
period; see reference 121.
(125) K. F. Greenough and A. B. F. Duncan, J. Am. Chem. Soc., 83, 555 (1961).
(126) H. D. Mettee and T. Fisher, private communication to the author.
(127) K. L. Demerjian, J. A. Kerr, and J. G. Calvert, Environ. Sci. Technol.,
U, 1 (197*0.
(128) This portion of the work under this grant was completed during the project
period and published during the final year of the grant: R. L. Daubendiek
and J. G. Calvert, Environ. Letters, 8, 10J (1975).
(129) C. F. H. Tipper and R. K. Williams, Trans. Faraday Soc., 57, 79 (I96l).
(130) C. F. Cullis, R. M. Henson, and D. L. Trimm, Froc. Roy. Soc. (London),
A295, 72 (1966).
(131) For recent examples see: a) W. E. Wilson, Jr., A. Levy, and D. B. Wimmer,
J. Air Polln. Control Assocn., 22, 27 (1970); b) E. R. Stephens and M.
A. Price, J. Coll. and Interface Science, 39, 272 (1972).
(132) R. L. Daubendiek and J. G. Calvert, "A Photochemical Study Involving S02,
02, S03, and 03" presented at the 165th National Meeting of the American
Chemical Society, Dallas, Texas, April, 19735 paper in preparation for
publication.
(133) See for example, H. S. Johnston, "Gas Phase Reaction Rate Theory,"
Ronald Press, New York, 1966, p. 1^-32.
(13*0 N. S. Grunhut, M. Goldfrank, M. L. Gushing, and G. V. Caesar, in Inorg.
Sjn. Vol. 3, Editor, L. F. Audrieth, McGraw-Hill, 1950, P. 78.
175
-------
References (continued)
,135) G. Schott and N. Davidson, J. Amer. Chem. Soc. , 80, l84l (1958).
(136) Environmental Protection Agency, National Primary and Secondary Ambient
Air Quality Standards, Part II, Federal Register, Vol. 36, No. 8^, 8196,
1971."
(137) R. H. Pierson, A. N. Fletcher, and E. St.Clair Gantz, Anal. Chem. , 28,
1218 (1956).
(138) E. J. Jones and 0. R. Wulf, J. Chem. Phys., 5, 873 (1937).
(139) P. A. Leighton, "Photochemistry of Air Pollution," Academic Press, New
York, 1961, p. 29-
(l40) This is the total pressure of N02 and N204. At the temperature of 30°
the calculated partial pressures of N02 and N204 are 8.68 and 0.^7 Torr,
respectively.
(lUl) P. Urone, W. H. Schroeder, and S. R. Miller, "Reactions of Sulfur
Dioxide in Air," in "Proceedings of the Clean Air Conference," Washington,
D.C. , 1970.
(1^2) R. C. Paul, C. L. Arora, and K. C. Malhotra, Indian J. Chem. , 10, 9U
(1972) .
(lU3) For some -recent examples of (NO)2S2Cv preparation see: a) K. Stopperka
and F. Kilz, Z. Chem., 8, ij-35 (1968); prepared from N203 and fuming
sulfuric acid; b) K. Stopperka, F. Wolf, and G. Suess, Z. Anorg . Allg.
Chem., 359, Ik (1968); prepared from gaseous S02 and N02 at 225-230°C;
c) R. Dejaeger, B. Vandorpe, and J. Heubel, Rev. Chim. Miner. , k, 195
(1967); prepared from NOC1S03 on heating; d) R. DeJaeger, P. Legrand,
and J. Heubel, Compt. Rend. Ac ad. Sci. Paris, Ser. C, 265 , 1117 (1967);
prepared from NOS03C1 reaction with N203 or N204.
For some recent examples of (N02)2S207 preparation see: a) B. Vandorpe
and J. Heubel, Compt. Rend. , 260, 6619 (1965); formed from N205 in liquid
S03; b) K. Stopperka and V. Grove, Z. Chem. , 5, 111 (1965); formed from
N02 and S02C12 at 70-125°C.
E. I. DuPont de Nemours and Co. (inc.), product literature on Teflon FEF
film available from the Plastics Department, Fluorocarbon Divison,
Wilmington, Delaware, 19898, 1973-
(146) P. M. McClough and B. A. Thrush, Chem. Ind. (London), 1966, Nov. 19,
P- 1971.
S. W. Benson and A. E. Axworthy, Jr., J. Chem. Phys., 26, 17 IB (1957).
(11*8) F. J. H. Woltz and E. A. Jones, J. Chem. Phys., 17, 502 (19^9)-
A. H. Nielsen, T. G. Burke, P. J. H. Woltz, and E. A. Jones, J. Chem.
Phys., 20, 596 (1952).
176
-------
References (continued)
(150) R. H. Fierson, A. N. Fletcher, and E. St.Clair-Gantz, Anal. Chem. , 28,
1218 (1956).
(151) E. A. Jones, J. S. Kirby-Smith, P. J. H. Woltz, and A. Nielsen, J. Chem.
Phys., 19, 242 (1951).
(152) D. Saunders and J. Heicklen, J. Amer. Chem. Soc. , 87, 2088 (1965). It
appears to us that Saunders and Heicklen underestimated their CF20 con-
centrations. They '^measured" CF20 by introducing it into a g.c. column,
and then determined the amount of CC-2 produced from an assumed stoichi-
ometry for the reaction between CF20 and the silica gel column packing.
Actually the conversion may have been incomplete. The effect of the
presence of impurities in the sample would lead also to an overestimation
of the extinction coefficient.
(153) J- Heicklen in "Advances in Photochemistry," Vol. 7, Wiley- Interscience,
New York, N. Y. , 1969, p. 79-
R. W. Murray, Accounts Chem. Res., 1, 313 (1968).
(155) J. Heicklen, J. Phys. Chem., 70, 477 (1966).
(156) Y. K. Wei and-R. J. Cvetanovic, Can. J. Chem., 4l, 913 (1963).
(157) P. G. Stecher, Editor, "The Merck Index," 8th Edition, Merck and
Company, Inc., Rahway, N. J., 1968, p. 209.
(158) This portion of the work under the present grant was carried out and
published during the second year of the grant period: R. L. Daubendiek
and J. G. Calvert, Environ. Letters, 6, 253 (197*0.
(159) This portion of the work under the present grant was carried out and
published during the first year of the grant period: J. G. Calvert,
K. L. Demerjian, and J. A. Kerr, Environ. Letters, 4, 123 (1973)-
(160) a) K. L. Demerjian, Masters Thesis, The Ohio State University, Columbus,
Ohio, 1970; b) K. L. Demerjian, J. A. Kerr, and J. G. Calvert, ref. 127;
c) J. A. Kerr, J. G. Calvert, and K. L. Demerjian, Chem. Brit . , 8, 252
(1972).
(l6l) K. L. Demerjian, J. A. Kerr, and J. G. Calvert, Environ. Letters, 3,
73 (1972).
(162) K. L. Demerjian, J. A. Kerr, and J. G. Calvert, Environ. Letters, 3,
137(1972).
(163) L. G. Wayne and D. M. Yost, J. Chem. Phys . , 19, 4l (1951).
(164) S. Jaffe and H. W. Ford, J. Phys. Chem., 71, 1832 (1967).
(165) J. Heicklen, K. Westberg, and N. Cohen, Publications No. 115-69 (Center
for Air Environmental Studies, University Park, Pa., 1969); K. Westberg,
H. Cohen and K. W. Wilson, Science, 171, 1013 (1971).
177
-------
References (continued)
(166) H. Stedman, E. D. Morris, Jr., E. E. Daby, H. Niki, and B. Weinstock,
paper in the Division of Water, Air, and Waste Chemistry, l60th National
American Chemical Society Meeting, Chicago, 111., Sept. 1970.
(167) E. R. Stephens and F. R. Burleson, J. Air Pollution Control Assoc., 17,
11*7 (1967).
(168) This portion of the work under this grant was carried out and published
during the first year of the grant period: J. G. Calvert, K. L.
Demerjian, and J, A. Kerr, Environ. Letters, 4, 28l (1973).
(l£9) The reaction scheme shown involves only 17 of the most important of the
total of 58 chemical reactions considered in our more complete reaction
scheme employed in the simulations of Figs. 38-^3; the complete mechanism
and the rate constant data selection is published elsewhere.127
(170) C. Morley and I. W. M. Smith, J. Chem. Soc. Faraday Trans. II, 68, 1016
(1972).
(171) W. E. Wilson, Jr. and D. Miller, unpublished results from studies made
in the large smog chamber at Battelle Memorial Institute, Columbus,
Ohio; we are grateful for the use of these data before publication.
(172) M. C. Dodge and J. J. Bufalini, private communication; the authors are
grateful for the use of a preprint of this work before publication.
(173) G. Dixon-Lewis, W. E. Wilson and A. A. Westenberg, J. Chem. Fhys., kk,
2877 (1966).
(174) K. Westberg, N. Cohen and K. W. Wilson, Science, 171, 1013 (1971).
(175) This portion of the work under the grant was carried out and published
during the final year of the grant period: J. G. Calvert and R. D.
McQuigg, Int. J. Chem. Kinet., Symp. 1, 113 (1975).
(176) A. Stern, "Air Pollution," 2nd Ed., Academic Press, New York, 1968.
(177) Fed. Reg., 36, 8186 (1971).
(178) S. W. Benson, "Thermochemical Kinetics," John Wiley and Sons, Inc., New
York, 1968.
(179) K. Westberg and N. Cohen, "The Chemical Kinetics of Photochemical Smog
as Analyzed by Computer," Report No. ATR-70(8l07)-l, Aerospace Corporation,
El Secundo, Calif., Dec., 1969.
(iBO) G. Hilst, "A Coupled Two Dimensional Diffusion and Chemistry Model for
Turbulent and Inhomogeneously Mixed Reaction Systems," Report from
Aeronautical Research Associates of Princeton, Inc., Princeton, N.J.,
1973s Contract RU-73-016C to the Office of Research and Monitoring, U.S.
Environmental Protection Agency.
178
-------
References (continued)
(l8l) a) A. Q. Eschenroeder and J. R. Martinez, Final Report. Contract No.
CPA-22-69-127, General Research Corporation, Santa Barbara, Calif., Nov.
1969; b) A. Q. Eschenroeder and J. R. Martinez, Advan. Chem. Ser . , 113,
101 (1972).
(182) S. K. Friedlander and J. H. Seinfeld, Environ. Sci. Techno!., 3, 1175
(1969).
(183) a) T. A. Hecht and J. H. Seinfeld, Environ. Sci. Technol. , 6, 47 (1972);
b) J. H. Seinfeld, S. D. Reynolds, and F. M. Roth, Advan. Chem. Ser. ,
113, 581 (1972).
L. G. Wayne and J. E. Ernest, "Photochemical Smog, Simulated by Computer,"
presented at the 62nd Annual Meeting of the Air Pollution Control
Association, New York, June
(185) L. G. Wayne, M. Weisburd, R. Danchick, and A. Kokin, "Final Report —
Development of a Simulation Model for Estimating Ground Level Concentrations
of Photochemical Pollutants," System Development Corp., Santa Monica,
Calif., 1971.
(186) H. Niki, E. E. Daby, and B. Weinstock, Advan. Chem. Ser., 113, 16 (1972).
(187) T. A. Hecht, J. H. Seinfeld, and M. C. Dodge, Curr. Res., 8, 327 (197*0.
(l88) R. A. Cox and R. Atkins, Atomic Energy Research Establishment, Harwell,
personal communication to J. G. C. from Dr. A. E. J. Eggleton.
(189) E. D. Morris, Jr., and H. Niki, J. Phys. Chem., 77, 1929 (1973).
(190) R. Simonaitis and J. Heicklen, J. Fhys. Chem., 78, 653 (1974).
(191) D. H. Stedman and H. Niki, Environ. Sci. Technol. , 7, 735 (1973).
(192) D. H. Stedman and J. 0. Jackson, paper in "Symposium on Chemical Kinetics
Data for the Lower and Upper Atmosphere," Air lie House, Va. , Sept., 1974.
(193) J. Heicklen, K. Westberg, and N. Cohen, Publication No. 115-69,
Center for Air Environmental Studies, University of Pennsylvania,
University Park, Pa., 1969.
(194) J. G. Calvert, J. A. Kerr, K. L. Demerjian, and R. D. McQuigg, Science,
175, 751 (1972).
(195) E. R. Stephens and F. R. Burleson, J. Air Follut. Contr. Assoc. , 19,
929 (1969).
(196) E. R. Stephens and F. R. Burleson, J. Air Follut. Contr. As soc . , 17,
147 (1967).
179
-------
References (continued)
(197) a) C. S. Tuesday in "Chemical Reactions in the Urban Atmosphere," R. D.
Cadle, Ed., Interscience Publishers, New York, 1961, p. 15; b) E. A.
Schuck, G. J. Doyle and N. Endow, Report 31, Air Pollution Foundation,
San Marino, Calif., Dec. I960; c) A. P. Altshuller, S. L. Kopczynski,
W. .A. Lonneman, T. L. Becker, and R. Slater, Environ. Sci. Technol., 1,
899 (1967); d) W. E. Wilson, Jr., and A. Levy, J. Air Follut. Contr. Assoc.,
20, 385 (1970); e) J. J. Bufalini and K. L. Brubaker, in "Chemical
Reactions in Urban Atmospheres," C. S. Tuesday, Ed., Elsevier Publishing
Co., New.York, 1971, p. 225; f) B. Dimitriades and T. C. Wesson, J. Air
Follut. Contr. Assoc., 22, 33 (1972); g) A. P. Altshuller and J. J.
Bufalini, Environ. Sci. Technol., 5, 39 (1971).
(198) E. R. Stephens, Advan. Environ. Sci., 1, 73 (1969).
(199) a) E. A. Schuck and G. J. Doyle, Report 29, Air Pollution Foundation, San
Marino, Calif., 1959; b) A. P. Altshuller and J. J. Bufalini, Photochem.
Photobiol.,4, 97 (1965), and ref. 29g; c) E. R. Stephens in "Chemical
Reactions in Urban Atmospheres," C. S. Tuesday, Ed., Elsevier Publishing
Co., New York, 1971, p. ^5; d) B. Dimitriades and M. L. Whisman, "Chemical
Reactions in Urban Atmospheres," C. S. Tuesday, Ed., Elsevier Publishing
Co., New York, 1971,.p. 89; e) see also ref. 1S6.
(200) K. D. Bayes, presented at the 6th Informal Photochemistry Conference,
University of California, Davis, June 196*1-.
(201) T. N. Jones and K. D. Bayes, Chem. Phys. Lett., 11, 163 (1971).
(202) A. M. Winter and K. D. Bayes, J. Phys. Chem., 70, 302 (1966).
(203) A. U. Khan, J. N. Pitts, Jr., and E. B. Smith, Environ. Sci. Technol., 1.,
656 (1967).
(20U) R. H. Kummler, M. H. Bortner, and T. Baurer, Environ. Sci. Technol., 3,
248 (1969).
(205) T. Frankiewicz and R. S. Berry, Environ. Sci. Technol., 6, 365 (1972).
(206) R. Louw, J. van Ham, and H. Nieboer, J. Air Pollut. Contr. Assoc., 23,
716 (1973).
(207) B. W. Gay, Jr., and J. J. Bufalini, Advan. Chem. Ser., 113, 255 (1972).
(208) D. Gray, E. Lissi, and J. Heicklen, J. Phys. Chem., 76, 1919 (1972).
(209) H. E. Hunziker and H. R. Wendt, J. Chem. Fhys. 60, 4622 (1974).
(210) H. E. Radford, K. M. Evenson, and C. J. Howard, J. Chem. Phys., 60, 3178
(1974).
(211) C. C. Wang and L. I. Davis, Jr., Fhys. Rev. Lett., 32, 349 (1974).
180
-------
References (continued)
(212) W. Tsang, in "Chemical Kinetic Data Survey VI: Photochemical and Rate
Data for Twelve Gas Phase Reactions of Interest for Atmospheric Chemistry,"
N.B.S.I.R., 73-207, R. F. Hampton, Ed., National Bureau of Standards,
Aug., 1973.
(213) a) D. F. Miller and C. W. Spicer,'"A Continuous Analyzer for Detecting
Nitric Acid," paper 7^-17 presented at the 67th Meeting of the Air
Pollution Control Assoc., Denver, Colorado, June, 197^; b) R- J. Gordon
and R. J. Bryan, Environ. Sci. Technol., 7, 6*t-5 (1973).
(21k) E. R. Stephens, Advan. Environ. Sci., 1, 119 (1969).
(215) W. L. Faith, N.A. Renzetti, and L. H. Rogers, Report 22, Air Pollution
Foundation, San Marino, Calif., March 1958.
(216) E. A. Schuck, A. P. Altshuller, D. S. Earth, and G. B. Morgan, J. Air
Pollut. Contr. Assoc., 20, 297 (1970).
(217) a) A. Gitchell, R. Simonaitis, and J. Heicklen, J. Air Pollut. Contr.
Assoc., 2k, 357 (191k)-, b) J. Air Pollut. Contr. Assoc., 24, 772 (197*0.
(218) R. A. Cox, J. Phys. Chem., 76, Qlk (1972).
(219) J. P. Friend, R. Leifer, and M. Trichon, J. Atmos. Sci., 30, 14-65 (1973).
(220) R. D. Cadle in "Air Pollution Handbook," P. L. Magill, F. R. Holden, and
C. Ackley, Eds., McGraw-Hill, New York, 1956, p. 3-
(221) R. A. Cox and S. A. Penkett, J. Chem. Soc. Faraday Trans. I, 68, 1735
(1972).
(222) G. K. Boreskov and V. V. Illarionov, J_-. Phys. Chem. Moscow, Ik, lU28
(19^0).
(223) a) R. A. Cox and S. A. Penkett, Nature, 229, ^86 (1971); b) Nature, 230,
321 (1971).
(224) L. A. Ripperton, H. E. Jefferies, and 0. White, Advan. Chem. Ser., 113,
219 (1972).
(225) W. E. Wilson, M. C. Dodge, D. N. McNelis, and J. Overton, paper presented
before Division of Environmental Chemistry, American Chemical Society,
Los Angeles, April 197^.
(226) H. E. O'Neal and C. Blumstein, Int. J. Chem. Kinet., 5, 397 (1973).
(227) a) R. Atkinson, B. J. Finlayson, and J. N. Pitts, Jr., J. Amer. Chem. Soc.,
95, 7592 (1973); b) B. J. Finlayson, J. N. Pitts, Jr., and H. Akimoto,
Chem. Phys. Lett., 12, 1+95 (1972); c) W. A. Kummer, J. N. Pitts, Jr.,
and R. P. Steer, Environ. Sci. Technol., 10, 104 5 (1971); d) J. N. Pitts, Jr.,
R. P. Steer, and B. J. Finlayson, Advan. Chem. Ser., 113, 2k6 (1972).
181
-------
References (continued)
(228)' J. A. Kerr and M. J. Parsonage, "Evaluated Kinetic Data on Gas Phase
Addition Reactions," Butterworths, University of Birmingham, England,
1972.
(229) A. C. Lloyd, "Evaluated and Estimated Kinetic Data for the Gas Phase
Reactions of the Hydroperoxyl Radical," National Bureau of Standards
(U.S.), report 10 Wf (1970).
(230) D. D. Davis, W. A. Payne, and L. J. Stief, Science, 179, 280 (1973).
(231) R- A. Cox, J. Photochem, in press.
(232) A. W. Castleman, Jr., R. E. Davis, H. R. Munkelwitz, I. N. Tang, and
W. P. Wood, paper in "Symposium on Chemical Kinetics Data for the Lower
and Upper Atmosphere," Airlie House, Va. , Sept.
(233) J. G. Calvert, D. H. Slater, and J. W. Gall, in "Chemical Reactions in
Urban Atmospheres," C. S. Tuesday, Ed., Elsevier Publishing Co., New
York, 1971, p. 133-
(23*0 J. G. Calvert; "Modes of Formation of the Salts of Sulfur and Nitrogen
in an N0x-S02 -Hydrocarbon-Polluted Atmosphere," paper at Conference of
Atmospheric Salts and Gases of Sulfur and Nitrogen in Association with
Photochemical Oxidant, University of California, Irvine, Jan. 197^;
State of California Air Resources Board Report, ARE Contract 3-197,
T. T. Crocker, Ed.
(235) N. A. Renzetti and G. J. Doyle, Int. J. Air Pollut . , 2, 327 (i960).
(236) A. P. Altshuller, S. J. Kopczynski, W. A. Lonneman, T. L. Becker, and
D. L. Wilson, Environ. Sci. Technol. , 2, 696 (1968).
(237) W. E. Wilson, Jr., and A. Levy, J. Air Follut. Contr. Assoc. , 20, 385
(1970).
(238) W. E. Wilson, Jr., and A. Levy, Curr. Res., 6, U23 (1972).
(239) J- P. Smith and P. Urone, Environ. Sci. Technol. , 8, 7it-2 (197*0.
(2^0) M. F. R. Mulcahy, J. R. Stephen, and J. C. Ward, J. Phys. Chem. , 71,
(1967).
182
-------
CONCLUSIONS
In this work several significant new observations have been made related to
the chemical reactions which occur in sunlight-irradiated, N0x-hydrocarbon-aldehyde-
atmospheres .
!_._ In the first part of this study we have investigated many of the primary
reactions necessary for the quantitative evaluation of the mechanism of photo-
oxidation of S02 in the atmosphere. The rate constants were determined for the
reactions of the excited states of B02 with many of the compounds which occur in
the polluted lower atmosphere. In view of these results several conclusions can
be made which bear on the SO^ atmospheric removal reactions.
l) The theoretical maximum rates of SOa-photooxidation in sunlight-
irradiated, SOg- containing atmospheres is 1.9$ per hour at a solar zenith angle
of 20°; see Sections I-A-2, p. 11, and I-B-2, p. 2k. In view of the high propor-
tion of chemically unreactive quenching collisions between excited SOg and Og,
it is probable that the actual rate of S02 photooxidation is 0.2$ per hour or
less. This rate of S02 conversion is small compared to the rates of other removal
reactions which occur in a highly polluted atmosphere. It may be significant in
atmospheres in which S02 is the dominant pollutant present.
2) The reaction of excited sulfur dioxide molecules with the alkene
hydrocarbons has been found to be extremely rapid. The rate constants are near
equal to the collision number for these systems. However the major result of
this interaction is the isomerization of the olefinic hydrocarbon; the extent of
incorporation of S02 into the carbon-containing products is small. See Section
I-D. p. 5U. We conclude that aerosol formation in urban atmospheres probably
does not involve excited S02-alkene reactions to any significant extent.
5) The mechanism and quantum efficiency of the S03 formation in ir-
radiated SOs- contain ing mixtures were determined in a flow system with various
added reactants. See Section I-F. p. 69-86. The apparently divergent results
from previous steady-state measurements of other workers and our results from
flow systems are' rationized well in terms of the significant occurrence of the
reaction SO + S03 -» 2S02 in the steady-state systems. In this work the first
unambiguous and interpretable kinetic study of the rates of S03 formation has
been made. The results point to the significantly greater importance of the S03
product in the irradiated S02- containing systems. Previous steady state measure-
ments of S03 rates in these systems probably do not apply to the atmospheric
system in view of the neglect of S03 destruction steps in the interpretation of
these experiments.
II. In the second phase of this study several rate constants were estimated
which are of immediate value in estimation of S02 conversion rates in the pol-
luted urban atmosphere.
1) We have established that the rate constant for the gas phase
reaction, 03 + SOg ~* S03 + Og, is: k< 5 x 10~3 liter mole'1 sec"1 at room tempera-
ture; see Section II-A, p. 93- This estimate is very much smaller than the pre-
viously published upper limits for this rate constant. There remains no doubt
that the rate of the homogeneous oxidation of SOa by 03 in urban atmospheres is
insignificant.
2) We have estimated the rate constants for the homogeneous gas phase
reactions, N03 + S02 -» N02 + S03, and N205 + S02 -* N204 + S03; these are k < U.2
liter mole"1 sec"1 and k< 2.5 x 10~2 liter mole"1 sec"1, respectively. See
183
-------
Section II-A-2, p. 97- These results when coupled with computer simulation
studies show that the rate of N03 .and N205 oxidation of S02 cannot be important
paths in the polluted atmospheres.
3) Preliminary observations in this work point to a previously un-
expected result which deserves further study. The gaseous reaction of S03 with
NOa appears to be extremely fast, and a solid product forms which may be nitrosyl
nitryl pyrosulfate: (NO)S20y(W02). See-Section II-A-2, p. 98. It is not possible
to say at this time how important the S03-NC>2 reaction observed here might be
relative to other SOs reactions in the polluted atmosphere. The significance of
the S03-N02 add.uct as a participant in urban, aerosol formation is an intriguing
possibility which cannot be evaluated without further quantitative work.
h) Our work has shown that ozone reacts with Teflon to form CF2Q
among other products. See Section II-B, p. 99- Workers who utilize Teflon
fixtures and Teflon-coated chambers or bags in the study of atmospheric reactions
should be alerted to this reaction and the possible contamination which will re-
sult from this practice.
III. In the third part of this work, an evaluation was made of the photo-
chemical smog mechanisms using the computer to simulate the rates of change in
various polluted atmospheres. Several important features of special interest in
the development of control strategy were observed.
l) Our results show that the dominant reactive entity which chemically
removes the olefinic and paraffinic hydrocarbons and aldehydes in smog is the HO-
radical. HOa, 03, and 0(3p) reactions with the olefinic hydrocarbons can be
significant for certain conditions as well. See Sections III-A. p. 112 and III-C-
1, p. 136.
2) The effect of carbon monoxide on the rate of 03 development in
NOx-CO-polluted atmospheres has been investigated. From this work we conclude
that atmospheric scientists should give careful consideration to the predicted
enhancement of ozone levels by relatively small amounts of CO in an N0x-polluted,
but hydrocarbon-free, atmosphere. In view of the technological difficulties
associated with the removal of NOx and CO from exhaust gases, there is a reasonable
possibility that significant ozone levels (0.08 ppm) may continue to plague many
urban areas even though a near total removal of the reactive hydrocarbons might
be effected.
3) Theoretical estimates based on computer simulation of reactions in
a smog system show that the concentrations of the dominant chain carrier species,
the HO and E02 radicals, are not strongly affected by variations in pollutant con-
centrations over a rather wide range. See Section III-C-1, p. 1^7- These data
should prove very useful in guiding the development of direct experimentation for
the. direct detection of these reactive species in the urban atmospheres.
4) Theoretical estimates of the nitrogen mass balance during reactions
in a simulated polluted atmospheres show that nitric acid and PAN are expected to
be the dominant first nitrogen-containing products in smog mixtures, and these
compounds may be formed in near equal amounts. See Section III-C-1. p. 15^•
5) Calculations based on computer simulations of smog systems show
that the initial presence of ald.ehydes in the reactant mixtures is expected to
accelerate hydrocarbon and NO oxidations and ozone formation. It appears from
-------
these data that the E.P.A. National photochemical oxidant standard could not be
met if the aldehydes remained high, (CH3CHO) ^ 0.06, (CKgQ) ^ 0.10 ppm, even if
nearly all of the olefinic hydrocarbons were removed. We should learn from these
"udies th&t the "true" relationship between non-methane hydrocarbons and maximum
I-h oxidant at low hydrocarbon levels could be a critical function of a variable
which is not now measured routinely, namely the concentration of impurity aldehydes.
6) The predicted effects of variation of the concentration of NOX and
hydrocarbons on the integrated (os)-time and (PAN)-time data for computer simulated
atmospheres show that the control strategy for photochemical oxidant based on NOx
and RH regulation should be reasonable sound provided that impurity aldehydes and
CO are controlled as well.
7) Finally we have studied through computer simulation the potentially
important reactions which control S02 conversion in an NOX-RH-S02 contaminated
atmosphere. It is concluded that the rate of oxidation of S02 to S03, HaSO.^, and
the other sulfate-containing species will be promoted by homogeneous reactions at
a rate ^ 1.1-1.5$ per hour. The dominant rate determining steps appear to be:
HO + S02 -» HOS02 and H02 + S02 -» HO + S03; see Section III-C-l, p. 162. It is
important to note that several previously unidentified yet potentially important
secondary products of S02 oxidation may form as a precursor to sulfuric acid and
sulfate salts; namely we would anticipate the formation of an inorganic analogue
to the notorious peroxyacylnitrates, HOS0202N02. as well as peroxysulfuric acid
(Caro's acid) and other seemingly important reactive intermediate species. These
should be considered as potential members of the ill-defined "sulfate" fraction
of the urban aerosol which correlates well with respiratory problems in humans.
185
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing}
1. REPORT NO.
EPA-600/3-76-070
2.
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
MECHANISM OF PHOTOCHEMICALLY INITIATED OXIDATIONS
5. REPORT DATE
June 1976
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
Jack G. Calvert
9.
RFQRMING ORGANIZATION NAME AND ADDRESS
stry Department
The Ohio State University
140 West 18th Avenue
Columbus, Ohio 43210
10. PROGRAM ELEMENT NO.
1A1008
11. CONTRACT/GRANT NO.
R800398
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Sciences Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final 1/73 - 12/75
14. SPONSORING AGENCY CODE
EPA-ORD
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Several significant new observations have been made relative to .chemical reaction!
that occur in sunlight-irradiated NOx/hydrocarbon/aldehyde/CO/S02 polluted atmosphctes.
Many of the primary reactions that are needed to quantitatively evaluate the photo-
oxidation mechanisms of S02 in the atmosphere were measured. Rate constants for the
reactions of the excited S02(3B-|) state of S0£ with various atmospheric gases, alkanes
alkenes, NO, CO, etc., were determined. In view of these results, the rate of S02
photooxidation in the atmosphere is estimated, and the possible role of excited-
S02/alkene interactions that generate aerosols is evaluated. Rate constants for the
homogeneous reaction of SO? with 03, NOa, and N20s were also estimated. All of these
reactions are relatively slow for conditions that usually exist in polluted atmosphere^.
The unusual reaction of S03 with N02 was observed, although Its importance in the
atmosphere cannot be evaluated accurately from the existing data. An evaluation was
made of the photochemical smog mechanisms using a computer to simulate the rates of
change in various polluted atmospheres. Several important features of special
interest in developing control strategies were observed.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COS AT I Field/Group
*A1r pollution
Tests
*0x1dat1on
*Photochemical reactions
*Reaction kinetics
*Sulfur dioxide
*N1trogen oxides
*A1kanes
*Alkene hydro-
carbons
Computerized
simulation
13B
14B
07B
07C
07E
07D
09B
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (ThisReport)
UNCLASSIFIED
21. NO. OF PAGES
200
20. SECURITY CLASS (This page)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
186
-------
INSTRUCTIONS
1. REPORT NUMBER
Insert the EPA report number as it appears on the cover of the publication..
2. LEAVE BLANK
3. RECIPIENTS ACCESSION NUMBER
Reserved for use by each report recipient.
4. TITLE AND SUBTITLE
Title should indicate clearly and briefly the subject coverage of the report, and be displayed prominently. Set subtitle, if used,, in smaller
type or otherwise subordinate it to main title. When a report is prepared in more than one volume, repeat the primary title, add volume
number and include subtitle for the specific title.
5. REPORT DATE • - . "
Each report shall carry a date indicating at least month and year. Indicate the basis on which it was selected- (e.g., date of issue,, date of
approval, date of preparation, etc.).
6. PERFORMING ORGANIZATION CODE
Leave blank. ••••••;'
7. AUTHOBJS) ~, , , t '•'.,,« ,.'•-.
Give name®«»conventional order (John R. Doe, J. Robert Doe, etc.). List author's affiliation If it differs from the performing organi-
zation. • - ' - •-.-.••,'•
8. PERFpRMING pRGANIZATION REPORT NUMBER
Insert if performing organization" Wishes to assign this number.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Give name, street, city, state, and ZIP code. List no more than two levels of an organizational hirearchy.
10. PROGRAM ELEMENT NUMBER
Use the program element number under which the report was prepared. Subordinate numbers may be included in parentheses.
-. r .-.•.,.-.- • • •
11. CONTRACT/GRANT NUMBER .
"'. ' Insert contract of grant tiiimber under which report was prepared:
.-., . . -• ...- -,,r ... , ..!..,;. ... •,.-... ., . .
12. SPONSORING AGENCY NAME AND ADDRESS
Include ZIP code. ;
13. TYPE OF REPORT AND PERIOD COVERED
Indicate interim final, etc., and if applicable,'dates covered.
14. SPONSORING AGENCY CODE
' Leave'blank. ' -.-•
IS. SUPPLEMENTARY NpTES
Enter information hot included elsewhere but useful, such as: Prepared in cooperation with, Translation of, Presented at conference* or;;
To be published in, Supersedes, Supplements, etc.
16. ABSTRACT . .
• Include a brief (200 words of less) factual summary of the most significant information contained-in .the report. If the. repoiUcontains a .
significant bibliography or literature survey, mention it here.
17. KEY VVORDS AND DOCUMENT ANALYSIS
(a) DESCRIPTORS - Select from the Thesaurus of Engineering and Scientific Terms the proper authorized terms that identify the major
concept of the research and are sufficiently specific and precise to be used as index entries for cataloging.
(b) IDENTIFIERS AND OPEN-ENDED TERMS - Use identifiers for project names, code names, equipment designators, etc. Use open-
ended terms written in descriptor form for those subjects for which no descriptor exists.
(c) COSATI FIELD GROUP - Field and group assignments are to be taken from the 1965 COSATI Subject Category List. Since the ma-.
jority of documents are multidisciplinary in nature, the Primary Field/Group assignment(s) will be specific discipline, area of human
endeavor, or type of physical object. The application(s) will be cross-referenced with secondary Field/Group assignments that will follow
the primary posting(s).
18. DISTRIBUTION^STATEMENT
Denote releasabjlit^ to the public or limitation for reasons other than security for example -"Release Unlimited." Cite any availability to
the public, witn address and price. /
19.&20. SECURITY.ICLASSIFICATION
DO NOT submit Classified reports to the National Technical Information service.
21. NUMBER OF PAGES
Insert the total number of pages, including this one and unnumbered pages, but exclude distribution list, if any.
22. PRICE
Insert the price set by the National Technical Information Service or the Government Printing Office, if known.
EPA Form 2220-1 r>73) (Revwia)
------- |