-------






X
0
z
^-x.
z
0
CO
C£

1 1 1


UJ
t£.
&.
o
u.
oo
0
l~-t

1 —
t_H
o

c5

c

u
cl


c
o
4.)
•o
c
o
u
*IO
£
c





1



*


c

£
J

^



i.
+
cc


^~


X


*
1
•e
«o


c
0)
*>
JC
4J
UJ
*

u
4-)
i=

<
c
H-

 c
> «—

gj
r c


o

X
o>


0
o

n
0


CM
o
0
c


GO
cn
o



s
m
,—




CM
o


CM
0
,_
f^.




s


r-
00

ro

CM
O
UJ

1 1 1 1 1
o o o o o

X X X X X
O O" Ci O"> C*


CC CO CC CO CO
o o o o o
O 0 O O 0

r"> m n m ro
0 O 0 0 0


CM >n 01 •—
O O r- O O
o o o o o
0 0 O 0 0


v^ in cy» f~ n
O O O «— CM



ro i^ CM u^ ^i
in ^j f^ ^ CO
m vn 10 erv CO
*~ ^ ^*




O> CM r- O» CO
fo *ff o CF cy*


r-x in in o\ CM
r*. to CM in CM
CO CO O f- f"-
CO CO rs. to to




O f*- *— g; m



CM tO Cf> V

o» CT* o o in

5n to cn S ?
CM CM CM CM CM
1 1 1 1 1
UJ UJ UJ UJ UJ
f» n <*i
i i i
222


*o *o to
^J pj ^J


§00 CO
0 0
0 O 0

ro m n
o o o


CO •—
o ^- o
o o o
O 0 0


SCO •—
CM r**
*- CM •—



f*1 •— CM
**i co cy>

"~ m !n




tO CM M
co cn 9
CM ^ CM

r* in CM
to 10 in
m m c*>




O V CM
r*. CM to


in n m

CO r*. CO
CM o» CM

CM r** m
* •<»• *
CM CM CM
1 1 1
UJ UJ Ul

1
0

X
c*


g
o

f*)
o


o
o


CM
0
0



0
c*
m





CM
m


,_
c*
CO




2




irt
00

to
CM
1
UJ
7
2


iO
^J


s
0

f>
o


CM
g
O


r-»
-



0
CM
m
ro




en
00
CM

O
m
CO




r*.




^
to

S
t
UJ



















uT
£
U
"o
1
6
£
*•
1
«
c
o
1
u

1
1
CD
t/t
,_
ID
"^
fO
f
g .
IQ >/»
|t
ft* L>
• E

f^1 -5?

01 CM tf»
0) ••,-
ex S §
0. CD *
* •»-


































Ol
4-1
CM
C
k.
4-*
C
0)

o
c
. S
!;
X C

£ IA
•o ^
c >»
c!
9J "
5 i
" «
244

-------
experiments was for nitrate formation via the ROX + NO
reaction.  For the high paraffin mix (EC-232, EC-233, and
EC-246) a value of 1000 ppm" min"  was used, but in the other
e'ight simulations a value of 500 ppm" min"  was used.
Initial values for the CBM species were determined as follows:
Ethylene was treated as ethylene, OLE was set equal to the
reported molecular propylene concentration, CARB was twice
the reported molecular concentration of trans-2-butene plus
any reported initial carbonyls, and PAR was the sum of four
times the reported butane concentration, six times the
reported 2,3-dimethyl butane concentration, the reported
propylene concentration, twice the trans-2-butane concentra-
tion, the toluene concentration, and twice the xylene concen-
tration.  ARO was three times the molecular concentrations
of toluene and xylene.  The only rate constant involving the
initial hydrocarbons that had to be determined was for ARO +
                                  4    -1   -1
OH-, for which a value of 2.4 x 10  ppm  min   was used.
This value is approximately the weighted root-mean-square
between the rate constants for the reactions of m-xylene
and toluene with OH-.  The ARO + OH* reaction, as discussed
above, was assumed to yield H0| approximately two-thirds
of the time.

A photolysis compatibility problem is demonstrated in the
simulations of mixtures containing seven hydrocarbons.  As
shown in Table 38, we had to use almost twice the calculated
aldehyde photolysis  rate constant for the four high-aromatic
experiments.  Figures 62 and 63 show sample simulations of
low aromatic concentrations (EC-233) and high aromatic con-
centrations (EC-245).  Table 36 shows the one-hour-average
N02 and 03 maxima.
                            245

-------
Q.
Q.
LU
     O.Mr
     OU30
     0.20
     0.10
     0.00
               SO     100
150     200    250

  TIME, minutes

     (a)   0,
                                                   300     350     400
    0.100
    0.075
    0.050
 Q-
 O.
UJ
g  0.025
    0.000
                SO      100     150     200     250

                               TIME, minutes

                           (b)  PAN, N02,  and NO
                     300    350    400
        Figure 62.    Simulation results  of a UCR seven-hydrocarbon
                      experiment (EC-233) with the Carbon-Bond
                      Mechanism (low  aromatic mixture)
                                 246

-------
   0.045
CL
Q.
   0.030
LU
g 0.015
   0.000
               50     100     150     200     250

                               TIME, minutes


                               (c)   Olefins
300
350
400
 a.
 CL
    9.00
    6.00
    7.00
LU
O

o
<->  6.00
    5.00
               50     100     ISO     200    250

                               TIME,  minutes


                               (d)   Paraffins



                          Figure 62  (Continued)
300
350
400
                                 247

-------
   0.20
   0.15
   0.10
UJ
o
O 0.05
   0.00.
50     100
                            ISO     200     250
                             TIME,  minutes
                             (e)  Aromatics
300    350     400
    0.48
 E  0.96
 o.
 o.
    0.24
 a:
 8 0.12
    0.00
              50     100     150     ZOO    250
                              TIME, minutes
                            (f)  Formaldehyde

                          Figure 62  (Continued)
                                     300     350    400
                                   248

-------
    0.28
I.  °-21
D.
O
^-H

5  O.U


I—
z
LU
O

O
<->  0.07
    0.00.
_j	i

 50      100
                                     J	i
                             150     200    250
                               TIME, minutes


                               (g)   Ethylene
300     350     400
                           Figure  62 (Concluded)
                                   249

-------
 i.20r
0.00.
100
                        150     ZOO     250

                         TIME, minutes

                      (a)  N02, NO, and 0
                            300
350
400
o.oo.
          50
100
                        ISO    200    250
                         TIME, minutes

                           (b)  PAN
                            300
350
400
  Figure 63.   Simulation results of a  UCR  seven-hydrocarbon
               experiment (EC-245) with the Carbon-Bond
               Mechanism (high aromatic mixture)
                           250

-------
   0.12
E  0.09
Q.
CX
   0.06
UJ
o


S 0.03
   0.00
              50      100     ISO    200     250

                              TIME, minutes


                               (c)  Olefins
300
350
400
   4.80
 E 4.20
 Q.
 Q.
   3.60
S 3.00
   2.40
              50     100    ISC    200     250

                             TIME, minutes

                             (d)   Paraffins
300
350
400
                         Figure  63  (Continued)
                                 251

-------
    2.00i
Q.
Q.
    J.50
<   1.001
OL
    0.501
    o.ooj
               50     100    150    ZOO     250
                              TIME, minutes


                              (e)  Aromatics
300    350
400
    2.00rv
    0.501
    0.00'
               50     100    150    200     250

                               TIME, minutes

                               (f)  Ethylene



                           Figure 63  (Continued)
300    350
400
                                 252

-------
    i.60r
E   1.20
Q.
Q.
   0.80
o


S  0.40
   0.00
0      50      100
                            150     200    250

                              TIME, minutes


                            (g)   Formaldehyde
300    350    400
                          Figure 63  (Concluded)
                                  253

-------
                                SECTION 7
                 SIMULATION OF PROPYLENE/NOX EXPERIMENTS
                          IN SEVERAL SMOG CHAMBERS
INTRODUCTION
     A major part of the current contract calls for studying the simulations
of similar propylene/NO  experiments performed in different smog chambers.
                       A
The reasons for such a study are twofold:

     >  To elucidate chamber effects.
     >  To expand the data base used to validate explicit and
        generalized kinetic mechanisms.

As defined here, chamber effects include any effect that would produce
different results from similar experiments in different smog chambers or
from a hypothetical well-mixed air parcel in the open air compared with
a smog chamber.

     Although the study will continue for another year, the results at this
time indicate four areas of chamber effects:  analytical, wall, temperature,
and lighting.  The analytical effects lead to experimental uncertainties
rather than to the difference between smog formation in chambers compared
with the open atmosphere.  Such effects would be common to virtually any
comparison of similar experiments performed using different equipment.  How-
ever, these effects are discussed because the uncertainties in the measure-
ments are often large enough to mask the other differences among smog chambers
Wall effects may be the most commonly considered type of chamber effects.
Indeed, one might argue that wall effects constitute the only real difference
between a smog chamber and the open atmosphere (at least for small, well-
mixed air parcels).  However, the preliminary findings of the present study
                                    254

-------
indicate that wall effects play only a minor role in most smog  chamber
experiments.  Refinement of kinetic mechanisms  has greatly reduced  the need
to hypothesize radical  sources or sinks on chamber walls  to obtain  simula-
tions in "agreement with measurements.

     The latest mechanisms show a very high sensitivity to spectral  distri-
bution (Whitten and Hogo, 1977).  Hence, application of these mechanisms
to the atmosphere or to various chambers requires a careful assessment of the
spectral distribution.   Although present kinetic mechanisms are highly sen-
sitive to spectral effects, they do not respond to changes in temperature
adequately enough to simulate similar smog chamber experiments  performed
at different temperatures.  A case can still be made that wall  effects
explain some of the temperature effects.  Lowering the wall temperature might
condense radicals or aldehydes, thereby slowing smog formation, and  raising
the temperature might condense fewer radicals or aldehydes.  A  higher tem-
perature might also cause radical precursors such as aldehydes  form  previous
experiments to "boil off" the walls and to accelerate smog formation.

     If chamber effects can be elucidated sufficiently, the expanded data
base to be used for developing and validating smog mechanisms will  provide
a more rational basis for acceptance of generalized mechanisms  for  modeling
the atmosphere.  In Section 3 we described the overall mechanism as  composed
of four stages:  inorganics, single-carbon-atom species,  higher aldehydes
and other partially oxidized hydrocarbons, and initial hydrocarbons—paraffins,
olefins, and aromatics.  Presumably, the chamber effects  would  mostly be
common to the hierarchical levels below hydrocarbons.  Hence, the study of
chamber effects for propylene/NO  experiments should be applicable  to smog
                                /\
chamber experiments using other initial hydrocarbons.  Of course, hydrocarbon-
specific chamber effects may occur; an example is toluene since poor account-
ing for carbon mass is common to smog chamber experiments using toluene.

DEVELOPMENT OF A DATA BASE

     Several factors must be considered in the development of a data base
for the chamber effects study:

                                    255

-------
Initial pollutant concentration
-  [CH^ - 10
-  [NO 1  <.  2 ppm.
 -     A
-  Data from experiments at many different initial concen-
   trations should be supplied if possible.
Light source
-  Type (e.g., blacklights or xenon arc).
-  Spectrum (preferably measured—otherwise  manufacturer's
   specifications).
-  Intensity in chamber (k,, k., or other  photometric
   measurements, including the most recent measurement
   reported before each experiment).
-  Age of light sources before each experiment  (average  age
   and spread in ages for multiple sources).
-  Operating temperatures of fluorescent light  sources.
Analytical methods
-  Listing of instrumentation used for the actual  experiments
   reported.
-  User's estimates of accuracy in the measurements.
-  Corrections applied for interferences with measurement
   methods, if any.
-  Documentation of the most recent calibration before each
   experiment, the calibration method, the number of points
   on the calibration curve, and the concentration at each
   calibration point.
Chamber cleaning and seasoning
-  Method of cleaning.
-  Deliberate seasoning method, if any.
-  Number and general type of experiments  since the last
   cleaning (to evaluate the possible action of the walls
   as contaminant sources).
-  Most recent "light" and "dark" ozone decay data before
   each experiment.
                            256

-------
     >  Mixing time (should be determined with lights  on  if
        possible—otherwise, use mixing data  without lights).
     >  Chemical data (data for all  chemical  species measured
        during each experiment, preferably in tabulated form,
        or else graphical presentations).
     >  Miscellaneous data (variations during the run  when
        known—otherwise use of chamber operating temperatures,
        humidity (RH or dewpoint),  and dilution rate).

     One of the decision factors in  the choice of the  data  base is the
availability of the data.  Much of  the data were reported in graphical
form, and chamber characteristics were often  obscure.  Furthermore, uncer-
tainty ranges were not reported for some of the chamber runs.  The following
runs were chosen based on availability and the other factors mentioned above:

     >  UCR-EC (Runs 121, 177).
     >  UNC (runs performed on 9 August 1975, 5 November  1976,
        and 8 August 1977).
     >  Research Triangle Institute  (RTI) (runs performed on
        11 October 1976).
     >  Battelle (S-019, S-114, S-115).
     >  National Air Pollution Control Administration  (NAPCA)
        (156, 164, 172).
     >  UCR-AGC (runs performed on  24 February 1973  and
        5 March 1973).
     >  CALSPAN (11, 15, 16).
     >  Lockheed (40, 41, 42, 43).

     Of the eight facilities listed  above, UCR provided the most  detailed
discussion of experimental conditions.  The 5775-liter UCR  chamber is an
evacuable cylinder (EC) coated with  FEP Teflon 3.66  m  long  and 1.37 m in diameter.
There are quartz windows on each end of the chamber.   The surface-to-volume
ratio is 0.054 cnf .  The irradiation source  is a xenon short arc lamp
                                   257

-------
that uses 25 kilowatts (Pitts et al.,  1977).   Two  repeated  runs  (EC-121
and EC-177) were chosen from nine UCR runs with the same  initial  concen-
trations because the uncertainty of  the reported light intensities  seemed
to be the" lowest for those experiments.  (The light source  had been
replaced just before each of those experiments.)  Sampling  techniques
should be more certain in those runs  than in  earlier runs,  mainly because
of the experience gained from previous experiments.

     The University of North Carolina and Research Triangle Institute  both
have outdoor smog chamber facilities.   The UNC facility consists  of two
compartments (red side and blue side), so that two experiments can  be  per-
formed simultaneously.  Each side is  an A-shape frame (9.14 m wide, 6.1 m
high at the peak, and 12.19 m long)  with walls made of FEP  Teflon film
(Jeffries, Fox, and Kamens, 1975).  The volume of the chamber is  3.1 x 10
liters.  The surface-to-volume ratio is 0.013 cm"  .  Of the alternatives
for which detailed data were available, five  are propylene/NO experiments.
                                                             X
One set of experiments, performed on 16 August 1975, had  continuous injection
of propylene and NO in the blue side of the facility but  not the  red side.
This set of experiments will be investigated  in the coming  year.  For the
current study, we chose four experiments (one on 9 August 1975,  two on
5 November 1976, and one on 8 August 1977) as part of the data base for the
chamber effects study.  The use of natural sunlight as the  irradiation source
increases the need for spectral measurements  of the sunlight during the day.
Jeffries, Fox, and Kamens (1975) reported an empirical relationship between
the N02 photolysis rate and the total solar radiation.  From the total solar
radiation measurements by UNC for the days of the experiments, we were able
to estimate the N0~ photolysis rate  constant and all other  photolysis rate
constants required in the propylene/NO  kinetic mechanism.
                                      A

     The  RTI  facility  is  similar  to the  UNC  facility.  RTI  has four separate
chambers  that  allow four  experiments  to be carried out simultaneously.  The
surface-to-volume  ratio  for  each  chamber  is  0.019  cm"  .  We received data
on two sets of four experiments performed at RTI on 6 October and  11 October
1976.   These  experiments  were carried  out for  35 hours.  The data  from
                                    258

-------
experiments on 6 October 1976 are questionable because a correction  factor
for the ozone measurements was needed to compensate for insufficient sample
flow rates.  In the set of experiments performed on 11 October 1976, no
ozone was formed in Chamber 1.  Therefore, only three experiments  could  be
used as part of the data base for our chamber studies.  For the RTI  experi-
ments, we assumed the same total solar radiation as for the UNC experiments,
since the facilities are located in the same area.

     The Battelle chamber is constructed of aluminum and is Teflon-coated
(Scofield, Levy, and Miller, 1969).  The  18,272-liter smog chamber has a
surface-to-volume ratio of 0.026 cm"  and is irradiated with fluorescent
blacklights.  Data for only three propylene/NO  experiments were received
                                              y\
from Battelle.  The reported experimental  conditions are detailed  enough
that we were able to perform simulations of all three experiments, but we
had to assume a blacklight spectrum.

     The National Air Pollution Control  Administration chamber was a 9500-
liter chamber made of aluminum with Mylar windows (Korth, 1963).   It was
irradiated with fluorescent blacklights.  Data for  six propylene/NO   experi-
                                                                  A
ments performed at NAPCA were available to us.  The only intensity spectrum
for the light source available to us was reported by Korth, Rose,  and
Stahirvm (1964).  The propylene/NO  experiments were performed in 1965, one
                                 A
year later.  Thus, the uncertainty in the light spectrum is great.

     The 6370-liter UCR all-glass chamber (AGC) is  made of Pyrex glass and
is similar in size to the evacuable chamber (Pitts  et al., 1977).   Its
surface-to-volume ratio is 0.0324 cm~ .   The glass  chamber is irradiated with
fluorescent blacklights.  Two runs from the UCR glass chamber were chosen
solely on the basis of data availability.  Since the available data were
limited, we had to assume a blacklamp spectrum for these experiments
based on the reported k. value.

     The smog chamber at CALSPAN is a cylindrical chamber 9.14 m in
diameter and 9.14 m high, with a surface-to-volume ratio of 0.0066 cm"1
                                   259

-------
(Kocmond et al., 1973).   The walls of the chamber are  coated with  a  fluoro-
epoxy urethane having surface energy and reactivity properties  similar  to
FEP Teflon.  The light source consists of 24 lighting  modules,  each  con-
taining two 40-watt sunlamps, eight 85-watt high  output  blacklamps,  and
two 215-watt specially produced blacklamps.  The  measured  N00 decay  rate
                 _1                                        C-
(k ,) was 0.35 min" .

     Of the seven propylene/NO  experiments performed  at CALSPAN from
                              /\
21 October 1974 to 3  January 1975, raw data for two experiments (Nos. 15
and 16) were available to us.  Initial conditions and  concentration  versus
time plots were reported for Experiments 11 and 14. Only  concentration
versus time profiles  were reported for the other  three experiments (Nos. 10,
12, and 13).  For the computer simulations, we chose Experiments 15  and 16
because of the detailed data on them.  We chose Experiment 11 but  not
Experiment 14 because a new reactive HC analyzer  was used  with  Experiment 14.
We feel that the errors associated with the use of a new instrument  may be
greater in Experiment 14 than the errors associated with an instrument
already used in the earlier Experiment 11.

      The  1866-liter  hexagonal  smog  chamber at  Lockheed  (Jaffe  and Last,
 1974) is  made  of  six flat  side  panels with Teflon-coated aluminum frames.
 The surface-to-volume ratio varies  depending on  the material  used and  the
 initial conditions of the  experiments;  the surface-to-volume was 0.047 cnf
 when Teflon was  used.  The light source in the Lockheed chamber is  a xenon
 arc lamp  with  a  spectrum reported by Jaffe and Last (1974).  Propylene/NO
                                                                         3\
 experiments under various  chamber conditions were  performed, including
 different wall  materials,  surface-to-volume ratios, and cut spectrum (280
 to 350 nm light removed).   We chose four experiments  (Nos. 40, 41,  42, 43),
 all performed  with Teflon-coated walls  and a  surface-to-volume ratio of
 0.043 cm.   Both conditions are similar to the  UCR evacuable  chamber.   Two
 of the Lockheed experiments (Nos.  40 and 41) were  performed with a  cut
 spectrum  ranging  from 350  to over 500 nm.   (The  full  spectrum  ranges from
                                     260

-------
280 nm to over 500 nm.)  These experiments provide comparison runs  for exam-
ining the effects of aldehyde photolysis and ozone photolysis in the propylene
kinetic mechanism because both photolysis reactions are significant only  in
the 280 to 340 nm range.

     Thus, the data base tor the chamber study consists of at least two to
three experiments in each of eight smog chambers.  For many of the  cham-
bers, data were available only for the two or three experiments chosen.
The most important uncertainty in nearly all of these experiments is
the spectrum of the light source.  Since we did not have light spectra
for the Battelle, CALSPAN, RTI, and UCR glass chambers, we assumed
representative spectra.  Table 39 summarizes the data base for the
chamber study.

THEORETICAL ANALYSIS OF PARTICLE FLOW IN THE SMOG CHAMBER

     Before performing computer simulations for each of the experiments
listed above, we investigated theoretical  aspects of particle flow  in smog
chambers and particle collision frequencies with chamber walls.   The pur-
pose of this effort was to develop a deeper understanding of wall effects.
After the effects of chamber geometries and stirring procedures  are isolated,
the way will be cleared for investigating the role of light sources and the
chemistry of various reactive intermediates.

     We begin by deriving a picture of the transport of materials to the  chamber
walls, and then we use this description as a first step in assessing the
relative importance of wall effects on gas-phase chemical reactions.
Ideally, such an assessment would be based on knowledge of an experimentally
determined decay constant for each of the i reacting species.  Detailed
measurements of the transport properties (temperature, concentration, and
velocity profiles) would then provide input to some transport description
obtained by complete solution of the coupled equations of continuity,
motion, and energy for the entire system.   In practice, these ideal condi-
tions are compromised in one or both of the following ways:
                                     261

-------
      TABLE  39.    SUMMARY  OF DATA BASE  FOR CHAMBER EFFECTS STUDY
Chamber
UCR-EC

UNC blue
UNC blue
UNC red
RTI-2
RT1-3
RTI-4
Battelle


NAPCA


Lockheed



CALSPAN


UCR-AGC

Initial concentration (ppm) ,
Run no. 	 "^ — ^ lmin~ 1
or date Propylene NO, *d ' " '
121
177
8/9/76
11/5/76
11/5/76
10/11/76
10/11/76
10/11/76
S-019
S-114
S-115
156
164
172
40
41
42
43
11
15
16
2/23/73
3/5/73
0.483
0.493
0.66
1.15
0.43
0.786
0.421
0.389
1.52
0.95
0.97
2.06
1.85
2.04
3.0
3.0
3.0
3.0
1.0
1.0
1.0
0.52
0.50
0.51
0.463
0.41
0.554
0.526
1.57
0.519
0.468
0.552
0.449
0.525
2.06
0.99
0.99
1.5
1.5
1.5
1.5
0.59
0.50
0.50
0.30
0.255
0.3*
0.33*
Variable
Variable
Variable
Variable
Variable
Variable
0.3*
0.38*
0.38*
0.4
0.4
0.4
0.3
0.3
0.3
0.3
0.35
0.35
0.35
0.35
0.35
Light source
25 kU solar
simulator
25 kW solar
simulator
Sunlight
Sunlight
Sunlight
Sunlight
Sunlight
Sunlight
Blacklight
Blacklight
Blacklight
Blacklight
Blacklight
Blacklight
Xenon arc
Xenon arc
Xenon arc
Xenon arc
Blacklioht
(mainly)
Blacklight
(mainly)
Blackllaht
(mainly)
Blacklight
Blacklight
Chamber
material
Teflon
Teflon
Teflon
Teflon
Teflon
Teflon
Teflon
Teflon
Aluminum and
Teflon
Aluminum and
Teflon
Aluminum and
Teflon
Aluminum and
plastic film
windows
Aluminum and
plastic film
windows
Aluminum and
plastic film
windows
Teflon
Teflon
Teflon
Teflon
Fluoroepoxy
urethane
Fluoroepoxy
urethane
Fluoroepoxy
urethane
Glass
Glass
Surface-to-
volume ratic
(cm'1)
0.034
0.034
0.013
0.013
0.013
0.019
0.019
0.019
0.026
0.026
0.026
--
—
--
0.043
0.043
0.043
0.043
0.0066
0.0066
0.0066
0.0324
0.0324
Ozone measuring
' instrument or
method
Dasibi 1070
Dasibi 1070
Chemi luminescent
Chemi luminescent
Chemi luminescent



Chemi luminescent
Chemi luminescent
Chemi luminescent
Chemi luminescent
Chemi luminescent
Chemi luminescent
McMillan 1100
McMillan 1100
McMillan 1100
McMillan 1100
Chemiluminescent
(Bendix 8002)
Chemiluminescent
(Bendix 8002)
Chemiluminescent
(Bendix 8002)
Dasibi 1003
Dasibi 1003
* Value of k, rather than
                                     262

-------
      >  With few exceptions,  no data  on the concentrations of
        reactive intermediates  are available.   In  particular,
        there is a paucity of information  on the decay  rates
        of any  individual  species  in  unirradiated  chambers.
      >  The transport properties of most chambers  are inade-
        quately characterized,  especially  in terms of temperature
        and velocity profiles,  to permit input to  a  sophis-
        ticated transport description.

     In view of the points above, we  devised the hiqhly simplified  picture
of transport within chambers  described  below,  which  has as inputs quantities
that are either already available or  easily measurable.   The validation of
this picture for the transport  of ozone to the walls and the results of
applying it to the eight smog chambers  are discussed later.

Reactant Transport Inside Smog  Chambers

     The objective of the transport description is to relate the observed
concentration c. of reacting species  i  to  its  rate of removal  by the cham-
ber walls.  The approach consists of assuming  the  existence of a stagnant
boundary layer of thickness 6 at the  chamber walls;  since the  layer is
associated with the convective motion of the air,  6  depends on the  local
air velocity; and consequently, at a  distance  of 6 or greater  from  the
wall the velocity component parallel  to the wall is  assumed to be V ,
                                                                   co
which is to be determined experimentally or deduced from mixing time
data.  Transport to the wall  is considered as  though the wall  were  a semi-
infinite flat surface, with the x-direction parallel to the surface.  Some
physical grounds exist for this approach (see  Figure 64). The solution
for 5(x) (Bird, Stewart, and Lightfoot, 1960)  is:

                        5(x)  =  4.64 (vx/V  )1/2    ,                  (97)
                                        CO
                                                          2     1
where v is the kinematic viscosity of air  (taken as  0.15 cm  sec ).
                                   263

-------
  VELOCITY  IN X-DIRECTION  =  V
                    BOUNDARY  LAYER  6(X)
                    VELOCITY  IN X-DIRECTION AT SURFACE  = 0
 (a)   Boundary  layer around a  thin,  semi-infinite  flat surface
                                           BOUNDARY  LAYER  fi(x)
I I t I I I I I I I I I I I I I I I I I I I I  I  I  I  I  I  I  I  I  I  I
                                                     SMOG  CHAMBER  WALL
(b)   Boundary layer caused by circulation  of air,near  the  smog
     chamber wall
                Figure 64.    Simplified boundary  layers
                               264

-------
   A velocity profile,
                    \r= 2
(98)
where y is the distance above the plate, was  assumed  in  calculating
the solution above.   Computation of 6(x) thus  reduces  to finding an
expression for x/V^.   We chose x to be  a characteristic  length whose
physical significance is shown in Figure 65.   We approximate V  by
                              x'/t .
                              * /T-mix
(99)
 where  t .   is  the  mixing  time  in the chamber and x1 is the length
        mix
 within which mixing is  presumed to occur  (i.e., the longest dimension
 of the chamber).  We set x' = x (Figure 65), yielding
                                    J/2
                                                                   (100)
where <5 is in cm and t .  is in seconds.  It is then straightforward to
                      III I /\
describe the transport and wall reaction of species i by using the coordi
nate system of Figure 66.
                                       SMOG CHAMBER
        Figure 65.   Characteristic lengths  in  smog  chambers.
                     Arrows  represent air circulation from
                     convection or  stirring.
                                   265

-------
                                i  =  ci
                                                         z = 0
               BOUNDARY
                LAYER
     V = 0
                               C.  -
              I  I  I  I  I   I   I   I   I   I   I   I   I   I
Z = 6
              Figure  66%   Transport  to  the chamber walls

    In the discussion below, we  adopt the  following notation:
         N.  =  flux of species  i  in  the  z-direction,  in molecules
          1      -2   -1
              cm  sec  ,
                                                     _3
         c  =  total  gas concentration,  in molecules  cm   ,
         c.  =  local  concentration of  species  i,
         c,  =  concentration  of species  i for  z <_Q,  i.e., bulk
              concentration,
                                                             2    -1
         D  =  diffusion coefficient of  species i  in  air, in cm sec  ,
         K  =  rate constant, in  sec"  ,  for  bulk  concentration change
              due to wall  reaction:   dc,/dt = -Kc, *
         K1  =  rate constant,in cm sec"  , for removal  of species i
              in terms of surface flux,
          S  =  chamber surface  area,
          V  =  chamber volume.
     A mass balance at steady state gives  dN./dz  =  0.   Differentiating
Eq. (100) and applying the boundary conditions  c. = c,  at z  =  0  and
c. = N.J/K' at z = 6 leads directly to the  solution

                           cD
                 N. = -   ~ - £n f-^-\   '               (101)
                           1 1 '
                            Snix

-------
From Eq.  (101)»  for small  c,,  we  have

                                   DC
           1
           2
N,« 	^	   .                      (102)
                                  t1/
                                    '1/2
If the reaction product is not adsorbed irreversibly  but  is released
as species j (i.e., if i  v^i?  j),  then  N.  =  -N.,  and  Pick's Law
                                        '      J
becomes
                                        +  '    =  -D-    '         (103)
which upon integration gives Eq.  (102)  exactly.

     Note that N. and c, are related through the surface-to-volume  ratio;
Solving for N., substituting into Eq. (102), and solving for K1  gives
                                     w
                           i/i _,      \ mix/	
                     _
where K' is in cm sec  .  Equation (105) is the desired relation between
K1 and K .  Calculation of a few test cases indicates that, for the
chambers listed in Table 40, the mixing time effect contributes a cor-
rection on the order of 20 percent or less to K'; indeed, for the limiting
case where [D/(tmix)1/2] • [S/fVKj] » 1, KW a K'(S/V).   Clearly, mixing
times are most important in  relatively  quiescent  chambers.
                                   267

-------
TABLE 40.   SURFACE-RELATED OZONE DECAY PARAMETERS FOR SELECTED SMOG CHAMBERS
Chamber
UCR-EC









UNC red
UNC blue


UNC red

UNC blue

RTI 1
RTI 2
RTI 3§ §
RTI 4
RTI 1
RTI 2
RTI 3
RTI 4
Battelle



Date of run
3-12-74
3-12-1',
4-19-/4
4-26-74
5-31-74
5-31-74
6-3-74
6-7-74
7-6-76
7-6/7-76
ll-4-73:
12-6-735
11-4-731
12-6-73*
11-4/5-73-
12-6-737
ll-4/5-73:
12-6-73'
8-14-75**
8-14-75**
8-14-75**
8-14-75**
8-14-75**
8-14-75**
8-14-75"
8-14-75**




Surface-to-
volume
ratio
Material (cm )
Teflon
Teflon
Teflon
Teflon
Teflon
Teflon
Teflon
Teflon
Teflon
Teflon
Teflon
Teflon
Teflon
Tenon
Teflon
Teflon
Teflon
Teflon
Teflon
Teflon
Teflon
Teflon
Teflon
Teflon
Teflon
Teflon
Aluminum-
Teflon
Aluminum-
Teflon
0.034
0.034
0.034
0.034
0.034
0.034
0.034
0.034
0.034
0.034
0.013
0.013
0.013
0.013
0.013
J.013
0.013
0.013
0.019
0.019
0.019
0.019
0.019
0.019
0.019
0.019
0.026

0.026

k
w
Lights on (106sec
No 36.4
Yes* 46.6
No 10.1
No 10.3
Yes 46.5
No 28.4
No 19.6
No 17.0
Yes 33.0
No 12.0
Yes 9.4
Yes 11.5
Yes 9.6
Yes 10.0
No 4.0
No 3.1
No 3.8
No 2.6
YesTtt
Yesm
Yesm 19
Yesm
No 8.8
No 7.2
No 8.4
No 6.7
No 27

Yes 55

K' km Km
"') (lO^cmsec"1) (106sec"1) (104cm sec"1 1
14.2
14.8
3.2
3.2
14 7
10.3
6.6
5.7
10.3
3.8
7.18
8.80
7.34
7.65
3.06
2.36
2.9
2.0
(
)
± 10% ) 10
[
4.7
3.8
4.4
3.5
11.0

21.0

..
-
..
..
..
-
..
..
-
..
..
—
..
..
..
..
..
--
..
—
-
--
--
..
..
--
..

..


                                                               (continued)
                                      268

-------
                                      TABLE   40  (Concluded)
Surface-to-
volume
ratio
Chamber Date of run Material (cm )
f * OU
Lockheed - °vrex
Aluminum
" -- Aluminum
" -- Pyrex
Pyrex
Teflon
Teflon
Stainless
" -- Stainless
" — Pyrex
Aluminum
" -- Aluminum
" -- Pyrex
" -- Pyrex
Teflon
Teflon
Stainless
Stainless
Mean value*+*
'°°/_i./~~
\mean )
0.047
0.090
0.136
0.090
0.136
0.090
0.136
0.090
0.136
0.047
0.090
0.136
0.090
0.136
0.090
0.136
0.090
0.136
-
--

Lights on
No
No
No
NO
No
No
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes

--

w
(106sec"')
26
34
43
32
34
39
33
72
61
64
54
55
42
38
42
57
96
115
37
34

K'
(104cm sec"1)
6.2***
3.8
3.1
3.8
2.4
3.7
2.9
5.3
3.7
14.0
6.0
4.0
4.7
2.8
4.8
4.3
11.0
8.5
34.0
36.0

km "m
(lO^ec"1) (104cm sec"1)
__
7.2
16.0
7.2
5.3
6.2
12.0
34.0
45.0
--
--
--
--
—
—
--
32.0
51.0
77.0
89.0

__
1.7
1.8
1.7
0.6
1.4
1 t.
7.9
5.1
--
—
-.
--
--
-
--
7.4
5.8
2.7
92.0

 * 20 kW xenon  arc  lamp filtered through new  Pyrex pane.
 + Run carried  out  from 1028 to 1404 EST.
 1 Run carried  out  from 1301 to 1501 EST.
 ,• Run carried  out  from 2102 to 0500 EST
 -. Run carried  out  from 1901 to 2358 EST
 •• Dates not given.   Data were taken from an  RTI progress report dated  14 August 1975.
 «« Natural sunlight through Teflon walls.
 §§ Stirring data indicated  V.  - 800 fps.  We used this value and x  as  the chamber circumference In the
   expression « • 4.64(vx/V.)l/2  and used that value of 6  in Eq.  (105).
 44  Values for the chamber In the presence of  inserts fabricated from the materials shown. .
 77  Base  chamber (empty):  KW = Kg.
"•  r calculated using the  relation KW= K'(S/V).
'"  Lockheed chamber only.
                                                      269

-------
Validation of the Transport Picture

     We calculated first-order light and dark decay  constants  (K1)  for
ozone bas_ed on measured ozone half-lives in the RTI,  UCR-EC, UNC, Battelle,
and Lockheed chambers.   The results of the calculations  are presented
in Table 40.  All the light and dark decay constants  for ozone were averaged
to generate the statistical information in Table 41.

     The main conclusion to be drawn from Table 41  is that standard
deviations amongst the  chambers in the apparent rate  of  ozone  decay
(K ) are substantially  reduced under both light and  dark irradiation
  w
conditions when individual  chamber geometries (surface-to-volume  ratios
and mixing times) are taken into account by calculating  K1.

Materials Effects—The  Lockheed Data

     The runs in the Lockheed chamber (Jaffe and Last, 1974) provide some
opportunity to examine  the effects of construction material on the  value
of K1.   Jaffe and Last  obtained ozone decay times under  both light  and
dark conditions when samples of different materials  were inserted into
the chamber shell.  An  attempt was made to estimate  the  influence of
materials on ozone decay by assuming all decay processes to  be independent
and first order.  If KD and 1C are the volume and the surface  decay rate
                      D      S3
constants,  respectively, for the base chamber, then one can  associate
with each construction material the analogous rate constants  Km and K^,
where

                        Km = Kw -  KB = K         •               (106)
and K  and K' are the volume and surface decay rate constants, respectively,
     m      rn
for each of the different materials inserted into the base chamber.
                                    270

-------
      TABLE 41.    INFLUENCE OF CHAMBER GEOMETRY  ON  THE  ESTIMATION  OF
                  SURFACE-RELATED OZONE DESTRUCTION*

.







Rate
10
10
1
1
0
0
6
4
6
4
X
X
X
X






constant
K,
w
K1,
K ,
w
K1,
light
light
dark
dark
19
9
8
3


Value



.6 ± 10
.6 ±
.1 ±
.9 ±
1
5
1



.8
.8
.3
.1
Coefficient of
deviation"*"
(percent)
55%
18
65
29

       * Materials  effects  are not considered.
       t Coefficients of deviation (i.e.,  standard  deviation
         value)  are expressed as  percentages  of  mean  values.
v mean
     Clearly, the effectiveness of this type of analysis  is  greatest when
K  » K' .   However, as the data in Table 40 show,  the K 's for the Lock-
 rn     D                                               ni
heed chamber are generally less than 1C; moreover, the dark  decay  data
of Table 40 show anomalous values for Pyrex and stainless steel.   In the
runs with lower surface-to-volume ratios, ozone was apparently destroyed
more rapidly than in the runs with higher surface-to-volume  ratios, which
would imply negative K 's.  In addition, during photolysis,  ozone  was
removed less rapidly in the presence of material other than  Pyrex.  Ozone
was removed more rapidly in the presence of stainless steel  than  in the
presence of Pyrex.  As a consequence, it was generally impossible  to
associate  surface rate constants with different materials during irradia-
tions of ozone in the Lockheed chamber.  In the dark, values of K'  were
in good agreement for Teflon and for aluminum,  but not for Pyrex or stain-
less steel.   (For a given material, values of K'  in the light and  dark
should be  identical.)  Thus, from the Lockheed  chamber data, the following
conclusions  can be drawn:
                                   271

-------
     >  The effects of chamber geometry are outweighed by the
        effects of different materials.  Note that no improve-
        ment in the coefficient of deviation resulted when surface
        ozone decay rates were expressed in terms of K'  rather
        than K or in terms of K' rather than K .
     >  Surface-related effects for Pyrex and for stainless steel
        were possibly dominated by history-dependent phenomena,
        in view of the decrease in K1 and (for dark reactions)
        K' with increasing surface-to-volume ratio.  For stainless
        steel, K1 was little affected whether the light source was
                ill
        on or off, suggesting that surface effects may have been
        associated more with changes in surface reactivity than
        with contamination found on the sample materials.
     >  Tne apparent lack of dependence of K^ on area for aluminum
        and for Teflon raises the possibility that geometric effects
        in chambers constructed of these materials could be part-
        ially accounted for by simple corrections for the surface-
        to- volume ratios for mixing time.

Wall Reactions of Species Other Than Ozone

     Upper limits were estimated for the removal  of species other than
ozone at the walls of the UCR chamber* by using the expression:
where K  i is the rate constant for removal  of species i  at the wall,  DI
is the diffusion coefficient of species i  in air, and the rest of the  terms
are as defined earlier.  The reliability of these constants depends primarily
on the choice of D. and t .  ; D. was usually estimated by using Graham's  law
                   I      nil X   1
* The UCR chamber was chosen for simulation only because of our previous
  experience with an explicit propylene mechanism tested with smog chamber
  data from that facility.  No conclusion regarding the actual reactivity
  of the UCR chamber walls should be inferred from this study.  In fact,
  Eq. (107) is based on the assumption of an infinitely reactive surface.
                                    272

-------
and measured diffusion coefficients for gases of similar molecular weight.
More  refined estimates would have to be based on Lennard-Jones parameters,
which might be obtained from viscosity data or critical point data, neither
of which seem to be available for reactive intermediates.  We believe that
                                                                        1/2
the accuracy limits are governed by the uncertainty in the use of (t .  )
as an approximation to the boundary layer thickness.  [We point out here
                                           1 /o
that  the dimension of the expression (tmip) '   is centimeters; strictly
speaking the boundary layer thickness 6 should be expressed as 6 = C(tmi-n)l/2,
where C is a proportionality constant having a value of unity when 6 and
t  .   are expressed in cgs units.]

     Table 42 gives the species examined and their rate constants.   Figures
67, 68, and 69 show the influence of the reactions that affected the
concentrations of 03, NOos and propylene, respectively.   Under the initial
conditions chosen, wall  reactions of OH-, 0 atoms, N03, and RCOX had no
effect on the concentrations of the three species examined; all  three,  however,
were sensitive to HONO.   The propylene simulations were unaffected by ozone-
wall reactions, even though the maximum possible rate of ozone loss to  the
walls was sufficient to suppress completely the buildup of ozone.   We empha-
size that the simulations above are valid only for the initial conditions
chosen; they should not be extended to other systems.

RESULTS OF THE CHAMBER EFFECTS STUDIES

     Much of the work for the eight different chambers listed above was done
with the explicit kinetic mechanism for propylene discussed in Section  5
of this report.

Results of the Study of the UCR Experiments

     The simulation results  of the propylene experiments  are discussed  earlier
(see Section 5).   In this section, we summarize some of the conclusions
reached from investigation of UCR Runs EC-121  and EC-177.

     Although the light source was replaced just before each of these runs,
we found that even using a higher than calculated photolysis rate  constant

                                   273

-------
TABLE 42.   MAXIMUM RATE CONSTANTS FOR HYPOTHETICAL
            WALL REACTIONS IN THE UCR CHAMBER

Species Kw^min '
CH3CH(02)CHOH
ti0'2
0
HONO
N03
PAN
CH3°2
Ozonide
CH3C(0)0;>
OH-
O
0.47
1.4
2.2
0.71
0.54
0.18
1.2
0.3
0.84
2.2
0.62
                        274

-------
  o.to
  0.45
            	 BASE CASE
            	HOMO
            	PAN, CH, OZONIDE
a 0.30
   0.15
             50
                     100       150      ZOO       250      300

                                 TIME, minutes
                                                              350
                                                                      400
      Figure 67.    Maximum influence  of wall reactions  on ozone
                    concentrations during the NO/propylene irradia-
                    tions.  Curves represent the simulated ozone
                    concentrations when species shown are presumed
                    to react with walls at the maximum rate con-
                    stants given in Table 42.
                                   275

-------
   o.«r
   0.45
                                                       	 BASE CASE
                                                       	MONO
                                                       	HOj
   0.30 -
   0.15
              SO
                      100
                              150
                                      200
                                   TIME, minutes
                                              ZSO
                                                      300
                                                              ISO
                                                                      400
Figure 68.   Maximum influence of wall  reactions on N02 concentrations
             during the simulated NO/propylene irradiations.  An
             initial MONO concentration of 30 ppb was assumed.  Curves
             represent simulated N02 concentrations when species  shown
             are  presumed to react with walls at the maximum rate
             constants given in Table 42.
                                   276

-------
      0.60
                                                          	 BASE CASE

                                                          	MONO
      0.45

   £  0.30

                                    \
      0.15
                 50
                        100
                                 150
                                         ZOO

                                      TIME, minutes
                                                 250
                                                         300
                                                                 350
                                                                         400
Figure 69.   Maximum influence of wall  reactions on propylene concentrations
             during the NO/propylene  irradiations.  Curves  represent the
             propylene simulated concentrations when species  shown are
             presumed to react with walls  at the maximum  rate constants
             given in Table 42.
                                      277

-------
we were not able to follow the propylene decay; thus, the ozone behavior was
overpredicted.  We chose EC-121  and EC-177 for this study because the
reported light spectrum should be accurate, and deterioration effects
would not have occurred.  Yet, there may be some problems in the data
documentation because we can simulate the propylene decay much better in
other UCR propylene experiments  with the same initial conditions.  In
future studies of the UCR experiments, we will investigate the possible
chemical reactions that may affect the propylene decay and the effect of
chamber cleaning on some of the  UCR experiments.  Prior to the EC-121
experiment, the UCR chamber was  cleaned by "boiling off" any chemical
species that may have adsorbed onto the walls.  If the chamber was not
"seasoned" after this cleaning period, then ozone or other species may have
been affected more in EC-121 than in other UCR experiments.

Results of the Study of the UNC Outdoor Smog Chamber

     The second chamber investigated was the outdoor smog chamber at the
University of North Carolina.  The main differences between this chamber
and  the UCR evacuable chamber are the natural irradiation and the tempera-
ture in the UNC chamber.  The physical characteristics and operating condi-
tions of the  chamber were discussed by Jeffries, Fox, and Kamens (1975).
Of the  reactions listed in Table 1 (see Section 4), only two reactions
(besides the  photolysis reactions) are important in chamber effects studies:
(1)  03 +wall and (2)  N2C>5 + H20.  Jeffries, Fox, and Kamens (1975) found
that the nighttime Cu loss rate in the UNC chamber yielded an 0, half-life
between 48 and 70 hours.  We estimate the rate constant for the 00 ->• wall
                                     -4    -1
reaction to be approximately 2.2 x 10   min   .

     Jeffries, Fox, and Kamens (1975) also reported a rate constant of
5.6  x 10"  ppnf min"  for the N90r + H,.0 reaction.  This value is a factor
                                   -6-1   -1
of 9 lower than the value of 5 x 10   ppm  min   used in the UCR chamber
simulations described in Section 5.  Since effects of the N205 + h^O reaction
tend to occur after the N02 peak and since the UNC NOp measurements do not
include PAN separately, we were not able to derive a better estimate
                                    278

-------
of the rate of the N?0r + H?0 reaction in the UNC chamber.   We tentatively
subtracted the simulated concentration of PAN from the measured N0? con-
centration to obtain an estimate of the actual  NCL concentration and,  hence,
the effects of the Np05 + hLO reaction.

     Figure 70 shows a typical solar radiation  profile for  the UNC experi-
ments.  The chamber experiments begin early in  the morning  (usually before
sunrise).  The speed of the overall chemistry increases slowly as the  sun
rises and then begins to decrease in the afternoon as the sun starts to
set.   In constant-light experiments, the speed also varies as propylene
is consumed.  Thus, the combination of light-induced acceleration and
chemically induced acceleration can test the mechanism in a novel manner.
We incorporated a variable radiation and temperature algorithm into our
kinetic program using a simple linear interpolation scheme.

     The run of 9 August 1975 was affected by clouds, causing "choppiness"
in the solar radiation profile.  Jeffries, Fox, and Kamens  (1976) sug-
gested that the "choppiness" may cause an increase in 0., production.
Their conclusions were based on experiments with the same initial concen-
trations of hydrocarbons and NO  performed on two consecutive days:  a
                               X
clear sunny day and a partly cloudy day.  More  0- was measured in the
chamber on the partly cloudy day.

     We performed three simulations of the 9 August 1975 run varying only
the "choppiness" of the solar radiation profile.  In these  simulations we
used the propylene mechanism presented by Whitten and Hogo  (1977).  Figures
71 and 72 show the solar radiation profiles used.  Curve 1  in Figure 71
represents our estimate of the solar radiation  profile on a clear day. Curve
2 was estimated by averaging through the peaks  in the observed solar radia-
tion profile.  Figures 72(a) and 72(b) show solar radiation profiles based
on input data every 15 minutes and every 5 minutes.   Figure 72(b) is the
closest representation of the solar radiation profile presented in Figure  70.
Computer simulations with these four profiles show the same results for
oxidant formation (Figure 73), except for Curve 1.   Thus, our old mechan-
ism does not show any effect due to "choppiness" at 5-minute or 15-minute
intervals.
                                   279

-------
2.Or
                         -§—'—rtr
                            HOURS,  EOT
Source: Jeffries (1976).
   Figure 70.   Observed diurnal  variation  in  solar intensity  at  the UNC
                chamber on 9 August 1975
                                280

-------
                          10     K     12    13     14     15
15     i7
Figure 71.   Approximations to the observed  solar intensity at UNC
             on 9 August 1975
                           281

-------
             on.     inn.
                           27"      3(.n,     460.
                              TIME, minutes
                                                G40.     MO.     720.
                     (a)  15~minutes intervals
Figure 72.   Calculated diurnal  variation of  the  N0£ photolysis  rate
             constant (k,)  in  the UNC chamber on  9 August 1975
                                282

-------
0.60*
   I
   I
                  inn.      2TO.     0*0,     450.
                               TIME, minutes
                                                 MA.      »:in.     730.
                  (b)  5-minute  intervals


                     Figure  72  (Concluded)
                              283

-------
                     U   CURVt 1

                   	 15-MINUTE INTERVAL

                   	 5-MIHUTE IMTERVAL

                   	 CURVE Z
 • .to
    •      *•.     IB*.
                                                ooooooooo
0000000
Figure  73.    Effect  of different  N02 photolysis rate constants 0'
              ozone concentrations in UNC  blue chamber experimen'
              on 9 August 1975.  Asterisks  indicate ozone measu''
              ments in the chamber.
                                 284

-------
     Temperature, with  a  diurnal  variation similar to that of solar
radiation, may also  have  a  large  effect on the overall chemistry.  In
our first simulation of the 9 August 1975 run, we varied the tempera-
ture as observed.  We then  performed the same simulation with a  constant
average temperature.  The temperature for the 9 August 1975 run  averaged
over 11 hours is 299K.   In  both runs, we used the Curve 2 solar  radia-
tion profile in Figure  71.   Figure 74 shows the 0- predictions from  the
two simulations.  As one  can see, (L is produced at nearly the same  rates
during the early parts  of the simulations, but not as much 0- is produced
at the end of the simulation at a constant temperature.
        1.2*
        ».»»
 *  EXPERIMENT RESULTS
 0  VARIABLE TEMPERATURE SIMULATION
	CONSTANT TEMPERATURE SIMULATION
                         1M.
                                   TIME, minute*
                                                                   TM.
         Figure 74.   Effects of  different  temperature profiles on
                      simulated ozone  concentrations
      The final  simulations were performed  on  four UNC chamber runs (one on
 9  August 1975,  two simultaneously on 5 November 1976, and one on 8 August
 1977) with the  propylene mechanism presented  in Section 5.  The simulation
 results are presented in Volume II of this report.
                                     285

-------
     An interesting observation is  the difference  between  summer  and
late fall  radiation intensities.   During the fall  months,  the total  radia-
tion is lower and the ratio of aldehyde to N0? photolysis  rate constants
is different.  It is also known that this ratio is not constant during  the
course of a day.

     In simulations of the 5 November 1976 runs, we had to lower  the
aldehyde/NOp photolysis rate constant ratio by 40 percent to fit  the
observed data.  Two UNC runs with different initial hydrocarbon con-
centrations were performed simultaneously.  We are able to simulate
both runs fairly closely up to the 03 peak (Figures 75 and 76).  For
the blue chamber run, we could not simulate the 0- concentrations after
the Oo peak.  For the red chamber run, we were not able to simulate NO  as
     O                                                                A
closely as 03 and propylene.  However, the use of a significantly lower
photolysis ratio for these November runs probably accounts more for the
colder temperature  then than the actual change in the photolysis  ratio.

Results of the RTI  Chamber Study

     The Research Triangle Institute  outdoor  smog chambers are similar to
the UNC chamber  and are  located in  the  same area.  Therefore, we assumed
the same solar radiation profile for  RTI  as for the UNC chamber.   Of the
experimental  data available, we investigated  the  propylene/NO  experiments
                                                             /\
performed  in October 1976  because this  set of data appeared  to have the
least  experimental  errors  in  the 03 values.   The  rate constants of the
0-  -»• wall  reaction  and the N90C + HpO reaction were changed  in the mechan-
               -4     -1           -e.     -1    -1
ism to 5.1 x 10   min    and 5  x 10  °  ppm  min  ,  respectively.

      Although RTI performed runs for  35  hours, we simulated  only the first
18  hours of  each run.  The simulation results for Chambers 2,  3, and 4 are
presented  in Volume II (no 03 was observed in Chamber 1).  As shown by
these  results, we are able to  simulate  the data fairly well  for the first
day,  except  for  Chamber  2  (Figure 77).   The NOp data  shown have been corrected
for PAN by subtracting the simulated  PAN  concentrations from the N0? data
as  was done  with the UNC data.
                                   286

-------
o.oo
    0      70     140     210     280    350


                         TIME,  minutes



                    (a)   N09,  NO and 07
                           
-------
    o.i6r
E   O.L2
Q.
Q-
    0.08
o:
o
z
8   0.04
    0.00
              70     140     210     280     350

                             TIME, minutes


                               (a)  0,
420    490
560
    0.48r
    0>°°0      70     140     210     280     350     420     490    560

                             TIME, minutes

                     (b)   Propylene, N02> and NO


   Figure 76.   Simulation results of a UNC propylene experiment on
                5 November 1976 (red side)
                                  288
-------
   0.16
 £ 0.12
 o.
 Q.
   0.08
O
•z.

3 0.04
   0.00
             150    300    450    600     750


                            TIME, minutes



                              (a)  0,
                                              goo
1050    1200
   1.20
 Q.
 o_
   0.90
-------
     The 03 data reported for Chamber 2 are suspicious.   At the NCL peak,
reported CL concentrations are approximately a factor of 10 lower than
the amounts predicted by steady-state calculations.   The 03 values re-
ported for Chambers 3 and 4 at the NCL peak were low by  only a factor of
about 3.  Our simulated 0- values are high for Chamber 2 by a factor of 3
                         O
but are fairly close for Chamber 3.  One aspect of the RTI data that we
have not taken into account is the effect the fan has on the species behav-
ior.  The fan was turned off in Chamber 3 only.  When the fan is left
running, we expect a greater wall effect owing to the turbulence caused by
the fan.  This might account for the overprediction of ozone in the simula-
tion of Chamber 2.

     The ozone was also overpredicted in Chamber 4,  perhaps also because
of the turbulence caused by the fan.  Another possible explanation is that
the chamber has not been "seasoned" long enough.  Simulations with a higher
03 -»• wall reaction (at 2.5 x 10~  min~ ) resulted in better fits of the
ozone concentrations in Chamber 4, but had little effect on the results of
Chamber 2.  Figure 7Dshows the simulation results of this run.  Thus, we
feel that a longer "seasoning" period may be required for Chamber 4.  Note
that in Figure 78 the predicted ozone in Chamber 4 decays rapidly after
the ozone peak while the observational data remains constant.  If the chamber
is not well seasoned before the experiment, it becomes a little more seasoned
during the experiment.  Thus, the 03 ->• wall reaction rate may decrease with
time, and the observed ozone would remain constant.

     During the next contract year, we hope to investigate the turbulence
effect caused by the fans and the need to have the chamber "seasoned"
sufficiently in the RTI experiments.

Results of the Study of the Battelle and NAPCA Chambers

     Both the Battelle and NAPCA chambers use constant irradiation sources
similar to that in the UCR chamber.  Both groups reported values for the NO,,
decay rate.  Battelle reported an actual k, (N0~ photolysis rate), which was
used in the computer simulations.

                                    290
-------
CL
Q.
O

o
O
    0.16
    0.12
2   0.08
fee
0.04
    0.00
              150     300    450    600    750
                              TIME, minutes

                                (a)   Ozone
                                              900
1050    1200
 CL
 Q.
o
    0.90r
    0.30 -
    0.00
    0.20 -
          150    300    450    600     750
                          TIME, minutes

                       (b)   N02,  NO, and
                                                  900
1050    1200
     Figure 78.   Simulation  results of the ozone behavior in RTI
                  Chambers  2  and 4 with the ozone wall  reaction
                  at  2.5  x  10~3 ppm~lmin~l
                                291
-------
     We investigated three Battelle runs (S-019,  S-114,  and  S-115)
performed with blacklight irradiation (k,  equal  to 0.3 min"   for S-019
             1                          i
and 0.16 min   for S-114).  Initial simulations  using  the  blacklight
spectral .distribution reported for the NAPCA chamber showed  high ozone
for S-019 but low ozone predictions for Runs S-114 and S-115.   However,
the fit was improved if the NO, photolysis rate  constant was raised from
         1            I       £
0.16 min"  to 0.2 min  .   [This change of 25 percent is  within  the
estimated limit of 30 percent uncertainty reported by  Wu and Niki
(1975)].  The simulation results for Battelle Runs S-019,  S-114,  and
S-115 are shown in Volume II.
     Of the five propylene/NO  experiments available to us  from NAPCA,
                             /\
we chose three runs with similar operating conditions (Runs 156, 164,
and 172).  Each run was statically charged and run in Chamber 2.  Run
156 was performed at full light intensity, and Runs 164 and 172 were per-
formed at one-third of the full light intensity.   The only  light intensity
spectrum available was reported by Korth, Rose, and Stahman (1964).  The
k. value used was 0.4 min"  (Kuntz, Kopczynski, and Bufalini, 1973).  From
estimates of k, from a k. value due to blacklamps, we estimated the  k,
               I         Q                                -I            I
value for the NAPCA chamber to be approximately 0.27 min~ .

     In our initial simulations of NAPCA Runs 156, 164, and 172, the pro-
pylene decay rate was too rapid and the N02 induction period too short.
These simulations are not shown.  Because the reported light spectrum was
taken in 1963 and the propylene experiments were performed  in 1965,  we
suspect that the light source may have deteriorated.  We applied short
wavelength attenuation to the reported light spectrum of Korth, Rose,  and
Stahman, (1974) and calculated a new set of photolysis ratios (Table 43).
Simulations with the shifted spectrum produced better fits  (see Volume  II)
                                   292
-------
   TABLE 43.    PHOTOLYSIS RATE CONSTANTS (RELATIVE  TO Iq  =  1)  USED  IN
               COMPUTER SIMULATIONS OF THE NAPCA RUNS*


03 + hv -»•
03 + hv ->
HONO + hv
H202 + hv
HCHO + hv
HCHO + hv
CH3CHO +
CH3CH2CHO
Original rate
Reaction constant"*"
0(1D) 0.0257
0(3P) 0.034
-* NO + OH- 0.426
-*• 20H- 0.0037
+ H02 + HCO- 0.0065
•* H2 + CO 0.0065
02
hv •* CH302 + HCO- 0.0065
°2
+ hv -> CH3CH20^ + HCO- 0.007
Attenuated rate
constant5
0.0052
0.03
0.375
0.0014
0.0038
0.0038
0.0038
0.0038

* The light intensity was cut to one-third the full  intensity for
  Runs 164 and 172.   Therefore, we lowered the aldehyde photolysis  by
  one-half to simulate this condition.

t Calculated from data reported by Korth,  Rose and Stahman  (1964).

§ Calculated for the 1965 runs assuming deterioration of short-
  wavelength intensity of the light source from the 1964 data.
                                    293
-------
Results of the Study of the UCR Glass Chamber

     One of the advantages of using a glass wall  smog  chamber for  hydro-
carbon/NO  experiments may be the low adsorption  of NO  to the glass walls
        •-X                                            X
compared with the adsorption by walls constructed of Teflon.   We used  the
rather low value of 8 x 10"  ppnf min"  for the NoOr + H90 reaction rate
                                                 £ 0    c
constant.

     We were able to obtain only limited information on two UCR glass
experiments performed in 1973.   Since the reported k .  for these experiments
           -1                            -1
is 0.35 min  , we used a value of 0.3 min   for k, in  the computer simula-
tions.  The simulated predictions of the NO , propylene and ozone  behavior
are shown in Figures 79 and SO.  Further investigations using higher alde-
hyde photolysis show an effect on the ozone induction  period with  little
effect on the ozone maximum (see Figure 81).  Another  investigation using
a higher k, value (0.5 min~ ) showed that by raising the N02 photolysis
rate constant we were able to predict the ozone behavior without affecting
the NO  and propylene predictions (see Figure 82).  Although this  last
set of simulations shows the closest agreement for ozone maxima between
simulations and the UCR glass experiments, we are skeptical about  the  high
NOp photolysis rate constant used in the simulations.

     During the next contract year, we hope to obtain  more information
concerning these experiments and perhaps to elucidate  the apparently high
concentrations of ozone found in these experiments.

Results of the Study of the CALSPAN Chamber

     For the CALSPAN chamber, we investigated three propylene/NO   experiments
                                                               J\
(Runs 11, 15, and 16).  Of these, only Run 11 did not  contain any  hydro-
carbon concentration versus time profile.  The computer simulations of Runs
11, 15, and 16 are shown in Figures 83, 84, and 85.  In the simulations of
Runs 15 and 16 (Figures 84 and 85) the propylene decay was simulated,  but
not the ozone behavior.  In the simulations of Run 16  (Figure 85), the simu-
lations have a shorter time to NO  crossover compared  with the actual  data;
                                 A
                                   294
-------
Q.
Q.
«c
o:
UJ
t_>
o
C_5
    0.14 -
    0.07 -
    o.oo.
               50
100
150     ZOO     250
  TIME,  minutes
300
350
                                                                  400
                              (a)  N02 and
                      NO
    o.eor
    0.60 -
Q.
Q.
h-  0.40 -
o

§  0.20
    0.00.
                      100
       ISO    200     250

          TIME,  minutes

      (b)  Propylene  and
                      300
       350
       400
   Figure  79.    Simulation results  of the UCR glass chamber experiment
                 performed on 23  February 1973
                                295
-------
   O.ZSp
    0.00.
              50
100
150     7.00    250

  TIME, minutes

 (a)  N02 and NO
300
350
400
Q.
Q.
QL
    0.80
    0.60
    0.40
o

R  0.20
    0.00
              50     100    ISO    ?00     250     300    350    400
                               TIME,  minutes

                            (b)   Propylene and Oo


    Figure 80.   Simulation results  of an UCR glass chamber experiment
                 performed  on  5  March 1973
                               296
-------
CL
O_
   0.28r
   0.21  -
   0.14
O 0.07 -
   0.00
             50      100
              150     ZOO    250

               TIME, minutes
                                                 300    350    400
                           (a)   N02 and
                          NO
   0.80
 a.
 Q-
   o.eo
o

o
   0.40
   0.20
   0.00
50     100    150    200    250

                TIME, minutes


           (b)   Propylene and
                                                 300    350    400
 Figure 81.  Simulation results of an UCR  glass chamber experiment
             with  the formaldehyde photolysis  at 3 x 10-3 min-l
                            297
-------
   0.28r
   -0.21 -
D.
a.
H-  0.14 -
o
z
o
O
0.07 -
    0.00
50     100
                            ISO     200     250

                              TIME, minutes
                             (a)   N02 and
                                       NO
                                             300
350
400
Q_
Q.
    o.eor
    0.60
    0.40
o


I  0-20
    0.00
              50     100     150     200    250

                              TIME, minutes


                         (b)  Propylene and
                                             300
                                                        350
       400
  Figure  82.  Simulation results of an  UCR glass chamber experiment
               with the N02 photolysis at  0.5 min-1
                               298
-------
Q.
Q.
   0.80
   0.60
   0.40
g  0.20
   0.00
             40      80
120     160    200

 TIME,  minutes
240    280    3ZO
   Figure  83.  Simulation results of propylene/NO  Experiment 11
               performed in the CALSPAN chamber  x
                                299
-------
    0.80
E  -0.60
Q.
Q.
UJ
o
<->   0.20
    0.00
               40     80     120    160    ZOO

                              TIME, minutes


                          (a)   N09> NO, and 0,
240    280
320
 Q.
 Q.
     1.04
    0.76
    O.S2
 UJ
 o   0.26
     0.00
               40      80     120     160     200

                            TIME, minutes


                            (b)  Propylene
240     280
320
  Figure 84.   Simulation results of  propylene/NOx Experiment  15
               performed in the CALSPAN  chamber
                               300
-------
   o.72r
o
   0.00
                     80
 120     160     ZOO
  TIME, minutes

(a)   N02, NO, and
       240
280
320
   1.04
 E J.7B
 Q.
 Q_
-------
     Using the reported light spectra, we calculate photolysis  ratios  for
both the cut and uncut spectra (Table 44).  Jaffe and Last (1974)  reported
a kd of 0.3 min"  for both spectra.   Wu and Niki  (1975)  provided the
following, equation to calculate k-|  from data on the decay of N02 in  N?:

                    n 5          kdLM]   k [M]     [NOJ,
            |/  =    U' J      I 4.  *+       3    o n     f- _ I
             1   Tt- t \        k       k        fNO l
             i   v »»p   '•i /        ^q      ^q       L'^^oJo
                                                                   (108)
where [NCLL and [N0p]2 are the concentrations of N02 at times  t,  and t^,
[NO ]Q = [N02] + [NO], and k-, k., and kr are the rate constants  for the
reactions of triplet oxygen atoms with N0?, N02+M, and NO,  respectively.
This equation should be used when [MO]/[N02] >. 0.5.   Data taken from Jaffe
and Last (1974) indicate that [NO^  = 0.58 ppm, [N02J2 = 0.415 ppm,
[NO ] = 1.6 ppm, t. = 3 minutes, and t? = 4 minutes.   By eq.  (108),  these
   A               I                   £-
data and the rate constants recommended by Hampson and Garvin (1978) indicate
that k, = 0.33 min" .  Use of the value of 0.33 min"   in the  computer simu-
lations resulted in prediction of an early NO  crossover and  fast propylene
                                             X
decay.  Simulation results for Runs 40 and 42 are presented in  Figures 86
and 87; simulation results for Runs 41 and 43 are presented in  Volume II
of this report.  To help simulate the high amounts of N02 in  the data, we
included a reaction that simulates N09 desorption off the walls at a rate
         -3        -1
of 2 x 10   ppm min  .  But even with this rate, we were not  able to
predict the N02 peak concentrations.

     In the cut spectrum Run 40 (Figure 86) we could  not follow the ozone
behavior or the propylene decay.  Note in Figure 86 that the  propylene
decay has two distinctive slopes:  It is fairly slow until  approximately
200 minutes and then becomes much faster.  This period (200 minutes) also
marks the point in which the ozone begins to influence the propylene decay.
In earlier simulations with the propylene mechanism presented by Whitten
                                   302
-------
    TABLE 44.    PHOTOLYSIS RATE CONSTANTS (RELATIVE TO k-j  = 1)  USED IN
                SIMULATIONS OF LOCKHEED CHAMBER RUNS


03 + hv -»•
03 + hv -»•
MONO + hv
H000 + hv
2 2

HCHO + hv
HCHO + hv

CH3CHO +

CH3CH2CHO
Reaction
O^D)
0(3P)
-»• NO + OH-
-* 20H-

02
-»• HO^ + HCO-
•* H2 + CO
02
hv -»• CH30^ + HCO-
02
+ hv -> CH3CH20^ + HCO-
Full spectrum
0.0064
0.0127
0.18
0.001


0.0018
0.0018

0.0018*

0.0018*
Cut spectrum
0
0.14
0.18
0.00044


0.00015
0.00015

0

0

* Quantum yield equals one.
                                   303
-------
   2.00r-
E -"i.50
Q.
Q.
   1.00
o

S  0.50
   0.00
                                       X X
0      40      80      120    .160     200     240     280

                      TIME, minutes


                 (a)  N02, NO,  and 03
                                                               320
   3.20r
 E 2.40
 o.
 o.
   1.60
3 o.eo
   0.00
              J	1	I	I	1       I       I    	I
              40      80
                     120     160     200

                     TIME,  minutes

                     (b)   Propylene
240    280    320
  Figure 86.    Simulation results  of propylene/NO  Experiment 40
                performed at Lockheed using a cut Spectrum
                               304
-------
E  .i.SO

Q.
   1.00
O
<-> 0.50
   0.00
                            1ZO     160     200
                             TIME, minutes

                          (a)  N02,  NO,  and
                     210    280     320
   3.20
E 2.40

Q.
    1.60
o

O
<-> 0.80
   0.00
              40      80
120     160    200

 TIME,  minutes

 (b)   Propylene
240    280    320
   Figure 87.    Simulation results  of propylene/NOx  Experiment 42
                 performed at Lockheed using a full spectrum
                               305
-------
and Hogo (1977), we found that the major source of radicals  is  the  ozone-
olefin reaction.  Since the present propylene mechanism does not  have  a
high radical yield from the ozone-olefin reaction, the simulations  do  not
follow the propylene decay as well.  In future work the ozone-olefin
reaction will be investigated further to determine, if possible,  the amount
of radicals required to simulate the Lockheed experiments.   Also  we will
examine the extent of NO  desorption from the walls and will investigate
                        X
other experiments using different materials to see whether  or not NO
                                                                   /v
desorption occurs extensively.

CONCLUDING REMARKS

     Our work to date seems to be leading toward the conclusion that the
same basic propylene mechanism is adequate for simulating experiments
in most chambers.  The most important chamber-specific effects  seem to be
the overall light intensity and the spectral distribution of the  light.
Wall effects appear to play a small but important role in the overall
chemistry.  At this time, the effects from uncertainties in analytical
data and light source data seem to be larger than the effects due to the
walls.
     We have received more propylene/NO  runs from UNC, which we are currently
                                       X
simulating with the hope of elucidating certain wall effects, such as NO
                                                                        X
and formaldehyde adsorption/desorption in the UNC chamber.   For the RTI
chamber, we are evaluating the effect of turbulence due to  the fan.  In
the  CALSPAN and Lockheed chambers there seem to be strong NO  desorption
                                                            X
effects, since the total NO  concentration increases in some of these
                           X
experiments.  We plan to estimate what effects NO  coming off the wall has
                                                 X
in all of the chambers.

     During the coming year, we also plan to investigate the effect of
temperature on the mechanism using the results of a temperature study
performed at UCR.
                                   306
-------
                               REFERENCES
Anderson, G. E., et al. (1977), "Air Quality in the Denver Metropolitan
     Region 1974-2000," EF77-222, EPA-908/1-77-002, Environmental
     Protection Agency Region VIII, Denver, Colorado.

Atkinson, R., and R. J. Cvetanovic (1971),  "Determination  of  the Absolute
     Values of the Rate Constants of the Reactions  of  0(3P) Atoms  with
     Alkenes by a Modulation Technique," J. Chem. Phys., Vol.  55,  pp. 659-663.

Atkinson, R. and J. N. Pitts, Jr. (1.978), J. Chem.  Phys.,  Vol.  68, p. 3581.

Baldwin, A. C., and D. M.  Golden (1978), "Application  of RRKM Theory Using
     a Hindered Rotational Gorin Model  Transition State to the  Reaction
     H02N02 + N2 £ H02 + N02 + Ng," in press.

Barker, J. R., and D. M. Golden (1977), "Measurement of Rate  Constants of
     Importance in Smog,"  Quarterly Progress Report, SRI International,
     Menlo Park, California.

Barker, J. R., S. W. Benson, and D. M.  Golden (1977),  "The Decomposition
     of Dimethyl Peroxide and the Rate Constant for CH30 + 02 •* Ch^O +
     H02," Int. J. Chem. Kinet.. Vol.  9, pp. 31-53.

Barker, J. R., et al.  (1977), "Measurement of Rate  Constants  of Importance
     in Smog," EPA-600/3-77-110, Environmental Protection  Agency,  Research
     Triangle Park, North Carolina.

Bird, B., W. E. Stewart, and E. N. Lightfoot (1960), Transport Phenomena
     (John Wiley & Sons, New York, New York).

Bufalini, J. J., B. W. Gay, Jr., and K. L.  Brubaker (1972), "Hydrogen
     Peroxide Formation from Formaldehyde Photooxidation and  Its Presence
     in Urban Atmospheres," Environ. Sci. Techno!., Vol. 6, pp. 816-821.

Calvert, J. G., and J. N.  Pitts, Jr. (1966), Photochemistry (John  Wiley
     & Sons, New York, New York).

Carter, W.P.L., et al.  (1978), "Computer Modeling of Smog  Chamber  Data:
     Progress in Validation of a Detailed Mechanism for  the Photooxidation
     of Propene and n-Butane in Photochemical Smog," Int.  J.  Chem. Kinet.,
     accepted for publication May 1978.

	 (1976), "Evidence for Alkoxy Radical Isomerization in'C^-Cs
     Alkanes in N0x-Air Systems," Extended Abstracts,  12th Informal
     Conference on Photochemistry, National Bureau  of  Standards, Washington,
     D.C., pp. N4-1 to N4-5.


                                   307
-------
Cox, R.  A., and R.  G.  Derwent (1976),  "The Ultra-Violet  Absorption  Spectrum
     of Gaseous Nitrous Acid,"  J.  Photochem.,  Vol.  6,  pp.  23-34.

           (1975),  "Kinetics of the Reaction  of HO^ with Nitric Oxide  and
     Nitrogen Dioxide," J.  Photochem..  Vol.  4,  pp.  139-153.

Cox, R.  A., and M.  J.  Roffey (1977),  "Thermal  Decomposition  of  Peroxyacetyl-
     nitrate in the Presence of Nitric  Oxide,"  Environ.  Sci.  Technol..
     Vol.  11, pp.  900-906.

Cox, R.  A., R. G.  Derwent,  and P.  M.  Holt (1976),  "Relative  Rate  Constants
     for the Reactions of OH Radicals with H2.CH4,  CO, NO, and  HONO  at
     Atmospheric Pressure and 296K,"  J.  Chem.  Soc., Farad. Trans., Vol.  72,
     pp. 2031-2043.

Chan, W. H., et al. (1977), "The Pressure Dependence of  the  Rate  Constant
     for the Reaction:  HO + CO + H + C02," Chem.  Phys.  Lett.,  Vol.  45,
     pp. 240-244.

Darnall, K. R., et al. (1976a), "Importance of R02  + NO  in Alkyl  Nitrate
     Formation from 04-65 Alkane Photooxidation Under Simulated Atmo-
     spheric Conditions," J. Phys. Chem., Vol.  80,  pp. 1948-1950.

           (1976), "Reactivity Scale  for Atmospheric Hydrocarbons  Based
     on Reaction with Hydroxyl  Radical," Environ.  Sci.  Techno!.,  Vol.  10,
     pp. 692-696.

Davis, D. D. (1976), private communication with Dr.  M.  Dodge of EPA.

Davis, D. D., et al. (1975), "A Kinetics Study of the Reaction of OH
     Radicals with Two Co Hydrocarbons:  C2H4 and C2H2," J.  Chem. Phys.,
     Vol. 63, pp. 1707-1712.

           (1972), "Absolute Rate Constants for the Reaction of Atomic
     Oxygen with Ethylene over the Temperature Range 232, - 500K," J_.
     Chem. Phys., Vol. 56, pp. 4868-4876.

Demerjian, K. L., A. J. Kerr, and J. G. Calvert (1974), "The Mechanism
     of Photochemical Smog Formation," in Advances in Environmental Science
     Technology, J. N. Pitts and R. L. Metcalf, eds. (John Wiley & Sons,
     New York, New York).

Dodge, M. C., and R. R. Arnts (1978), "A New Mechanism for the Reaction
     of Ozone with Olefins," submitted to Int. J. Chem. Kinet.

Durbin, P. A., T. A. Hecht, and G.  Z. Whitten  (1975), "Mathematical Modeling
     of Simulated Photochemical Smog," EPA-650/4-75-026, Systems Applications,
     Incorporated, San Rafael, California.

Graham, R.  (1975), private communication.


                                  308
-------
Graham, R. A., and H. S. Johnston (1978), "The Photochemistry of N03 and
     the Kinetics of the ^Oc-Oo System," J.  Phys.  Chem..  Vol.  82,
     pp. 254-268.

Hampson, R. F., Jr., and D. Garvin (1978), "Reaction Rate  and Photochemical
     Data for Atmospheric Chemistry-1977," NBS Special Publication  513,
     National Bureau of Standards, Washington, D.C.

Hecht, T. A., and J. H. Seinfeld (1972), "Development and  Validation of  a
     Generalized Mechanism for Photochemical  Smog," Environ.  Sci. Techno!.,
     Vol. 6, pp. 47-57.

Hecht, T. A., M. K. Liu, and D. C. Whitney (1974),  "Mathematical Simulation
     of Smog Chamber Photochemical Experiments," EPA-650/4-74-040,  Systems
     Applications, Incorporated, San Rafael,  California.

Hecht, T. A., J. H. Seinfeld, and M. C. Dodge (1974), "Further Development
     of a Generalized Mechanism for Photochemical  Smog,"  Environ. Sci.
     Technol., Vol. 8, pp. 327-339.

Hendry, D. G., and R. A. Kenley (1977), "Generation of Peroxy Radicals
     from Pernitrates (RC^NO?).  Decomposition of Peracylnitrates,"
     J. Amer. Chem. Soc., Vol. 99, p. 3198.

Hendry, D. G., et a!. (1977), "Computer Modeling of Simulated Photo-
     chemical Smog," Progress Narrative 8, SRI International, Menlo
     Park, California.

Herron, G. T., and R. E. Huie (1974), "Rate Constants for  the Reactions
     of Atomic Oxygen (03P) with Organic Compounds  in the  Gas Phase,"
     J. Phys. and Chem. Ref.  Data, Vol. 2, pp. 467-518.

Herron, J., and R. E. Huie (1977), "Stopped-Flow Studies  of the Mechanisms
     of Ozone-Alkene Reactions in the Gas Phase.  Ethylene,"  J. Amer. Chem.
     Soc., Vol. 99, p. 5430.

Howard, C. J. (1977), "Kinetics of the Reaction of  H02 with NO?," J. Chem.
     Phys., Vol. 67, pp. 5258-5263.

           (1976), "Rate Constants for the Gas-Phase Reactions  of OH
     Radicals with Ethylene and Halogenated Ethylene Compounds,"  J.  Chem.
     Phys.. Vol. 65, pp. 4771-4777.

Howard, C. J., and K. M. Evenson 0977), "Kinetics of the Reaction of H02
     with N02," Geophys. Res. Lett.. Vol. 4, pp.  437-440.

Jaffe, R. J., and K. W. Last (1974), "Study of Factors Affecting  Reactions
     in Environmental Chambers," Report No. LMSC-D406484, Lockheed Missiles
     and Space Company, Sunnyvale,  California.

Japar, S. M., and H. Niki  (1975),  "The Gas-Phase  Reactions  of the Nitrate
     Radical  with Olefins," J.  Phys.  Chem., Vol.  79, pp.  1629-1632.


                                 309
-------
Japar, S. M., C.  H.  Wu, and H.  Niki  (1974),  "Rate  Constants  for  the
     Reaction of Ozone with Olefins  in the Gas  Phase,"  J.  Phys.  Chem.,
     Vol. 78, pp. 2318-2320.

Jeffries, H. (1976), Progress Report to Dr.  M.  C.  Dodge of the EPA from  the
     University of North Carolina,  dated November  1976.

Jeffries, H., D.  Fox, and R.  Kamens  (1976),  "Outdoor Smog  Chamber  Studies:
     Light Effects Relative to Indoor Chambers," Environ.  Sci. Technol.,
     Vol. 10, pp. 1006-1011.

           (1975), "Outdoor Smog Chamber Studies:   Effect  of Hydrocarbon
     Reduction on Nitrogen Dioxide," EPA-650/3-75-011,  University  of
     North Carolina, Chapel Hill, North Carolina.

Kaiser, E. W., and C. H. Wu (1977), "A Kinetic Study of the Gas  Phase
     Formation and Decomposition Reactions of Nitrous  Acid," J.  Phys.
     Chem., Vol. 81, pp. 1701-1706.

Kajimoto, 0., and R. J. Cvetanovic (1976), "Temperature Dependence of
     0(1D) Production in the Photolysis of Ozone at 313 nm," Chem. Phys.
     Lett., Vol. 37, pp. 533-536.

Kocmond, W. C., et al.  (1973), "Determination of the Formation Mechanisms
     and Composition of Photochemical Aerosols," EPA-650/3-73-002,
     CALSPAN Corporation, Buffalo, New York.

Korth, M. W. (1963), "Dynamic Irradiation Chamber Tests of Automotive
     Exhaust," Publication No. 999-AP-5, Robert A. Taft Sanitary Engineer-
     ing Center, U.S. Public Health Service,  Cincinnati, Ohio.

Korth, M. W., A. H. Rose, Jr., and R. C. Stahman (1964), "Effects  of Hydro-
     carbon to Oxides of Nitrogen Ratios on Irradiated Auto Exhaust, Part  1,"
     J. Air Poll. Contr. Assoc.. Vol. 14, pp. 168-175.

Kuntz, R. L., S. L. Kopczynski, and J. S. Bufalini (1973), "Photochemical
     Reactivity of Benzaldehyde-N0x and Benzaldehyde-Hydrocarbon-N0x Mixtures,"
     Environ. Sci. Techno!., Vol. 7, pp. 1119-1123.

Leighton, P. A. (1961), Photochemistry of Air Pollution (Academic  Press,
     New York, New York).

Levine, S. Z., et al.  (1977), Chem. Phys. Lett.. Vol.  48, p. 528.

Lin, C. L., N. Rohatgi, and W. B. DeMore (1978), "The Absorption Cross
     Sections of Hydrogen Peroxide," 13th Informal Conference on Photo-
     chemistry, Clearwater Beach, Florida.

Lloyd, A. C., et al.  (1976), "Relative Rate Constants for Reactions of
     the Hydroxyl Radical with a Series of Alkanes, Alkenes, and Aromatic
     Hydrocarbons," J.  Phys. Chem., Vol. 80,  pp. 789-794.
                                    310
-------
MacCracken, M. C., and G. D. Sauter, eds.  (1975),  "Development of an Air
     Pollution Model for the San Francisco Bay Area," UCRL-51920, Vol.  1,
     Lawrence Livermore Laboratory, Livermore, California.

Meagher, J. F., and J. Heicklen (1976), "Reaction  of HO with C2H4,"  J.  Phys
     Chem., Vol. 80, pp. 1645-1652.

Moortgat, G. K., et al. (1978), "Wavelength Dependence of Relative and
     Absolute Quantum Yields in the Near UV-Photolysis of Formaldehyde
     at 25°C," 13th Informal Conference on Photochemistry,  Clearwater.
     Beach, Florida.

Morris, E. D., Jr., and H. Niki (1973), "Reaction  of Dinitrogen Pentoxide
     with Water," J. Phys. Chem.. Vol. 77, pp. 1929-1932.

            (1971 a), "Mass Spectrometric Study of the Reaction of Hydroxyl
     Radical with Formaldehyde," J. Chem. Phys.. Vol.  55, pp.  1991-1992.

	  (1971b), "Reactivity of Hydroxyl  Radicals with Olefins,"  J.
     Phys.  Chem.. Vol. 75, pp. 3640-3641.

Niki, H., (1978), "Reactions of Olefins with Hydroxyl  Radical  and with
     Ozone," presented at EPA/NBS Workshop on Chemical  Kinetic Data  Needs
     for Modeling the Lower Troposphere, 15-17 May 1978,  Reston,  Virginia.

Niki, H., E. Daby, and B. Weinstock (1972),  "Mechanisms of Smog Reactions,"
     in Photochemical Smog and Ozone Reactions,  Advances  in Chemistry
     Series, R. F. Gould, ed., Vol. 113, pp. 16-57 (American Chemical
     Society, Washington, D.C.).

	  (1969), Proc. 12th Combustion Symposium,  Poitiers,  France, p.  277,
Niki, H. et al. (1977), Chem. Phys. Lett.. Vol. 45, p. 564.

Pastrana, A. V., and R. W. Carr, Jr.  (1975), "Kinetics of the Reaction  of
     Hydroxyl  Radicals with Ethylene, Propylene, 1-Butene, and trans-2-
     Butene,"  J. Phys. Chem.. Vol.  79,  pp. 765-770.

Philen, D. L., R.  T. Watson, and D. D.  Davis (1977), "A Quantum Yield
     Determination of 0('D) Production  from Ozone via Laser Flash  Photo-
     lysis," J. Chem. Phys., Vol. 67, pp.  3316-3321.

Pitts, J. N.,  Jr., et al.  (1977), "Mechanisms of Photochemical Reactions
     in Urban  Air.  Volume II.   Chamber Studies," EPA-600/3-77-014b,
     State Air Pollution Research Center,  University of California at
     Riverside, Riverside, California.

Reynolds, S. D. et al., (1974), "Mathematical Modeling of Photochemical
     Air Pollution—III.  Evaluation of the Model," Atmos. Environ.,
     Vol. 8, pp. 563-596.
                                   311
-------
Scofield, F., A. Levy, and S. E. Miller (1969), "Design and Validation of
     a Smog Chamber," Scientific Circular No. 79, National  Paint,  Varnish,
     and Lacquer Association, Washington, D.C.

Sie, B.K.T., R. Simonaitis, and J. Heicklen (1976), "The Reaction  of OH
     with CO," Inter. 0. Chem. Kinet., Vol. 8, pp. 85-98.

Simonaitis,  R., and J.  I.  Heicklen (1974),  "Reactions  of OhO^ with  NO and
     N02,"  J.  Phys. Chem.,  Vol.  78,  pp.  2417-2421.

Spicer, C.  W.,  et al. (1973), "The Reactions  of Methylperoxy  Radicals  with
     NO and N02," J. Amer.  Chem.  Soc.,  Vol.  95, p.  13.

Stedman, D. H.  (1977), "Large Scale, High Levels of Ambient Ozone  in
     Lower Michigan," First Annual Report to National  Science Foundation,
     Grant No. ATM  76-03793.

Tesche, T. W., and  C. S. Burton (1978), "Simulated Impact of Alternative
     Emissions Control Strategies on Photochemical Oxidants in Los Angeles:
     Volume I--Preliminary Discussion of Results," EF78-22R,  Systems
     Applications,  Incorporated, San Rafael, California.

Tuesday, C.  S.  (1961), Chemical  Reactions in the Lower and  Upper Atmosphere,
     R. D;  Cadle, ed., pp.  1-49 (Interscience Publications, New  York,  New York)

Walter, T. A., J. J. Bufalini, and B. W. Gay, Jr. (1977), "Mechanism for
     Olefin-Ozone Reactions," Environ. Sci. Technol..  Vol.  11, pp. 382-386.

Westenberg, A. A.,  and N. de Haas (1969), Proc. 12th  Combustion Symposium.
     Poitiers, France, p. 289.

Whitbeck, M. R., et al. (1976), "A Kinetic Study of CH302 and (CH3)3C02
     Radical Reactions by Kinetic Flash Spectroscopy," Extended Abstracts,
     12th Informal  Conference on Photochemistry, National Bureau of Standards,
     Washington, D.C., pp. Kl-1 to Kl-5.

Whitten, G. Z.,  and M. C. Dodge (1976), "Aldehydes and Photochemical Smog,"
     171st National  Meeting, American Chemical Society, April 1976, New
     York, New  York.

Whitten, G. Z., and  H. Hogo  (1977), "Mathematical Modeling  of Simulated
     Photochemical  Smog," EPA-600/3-77-011, Systems Applications,  Incor-
     porated, San Rafael, California.

Whitten, G. Z.,  and J. P. Meyer (.1975), "CHEMK:  A Computer Modeling Scheme
     for Chemical  Kinetics," CS75-70R2, Systems Applications, Incorporated,
     San Rafael, California.
                                    312
-------
Wilson, W. E., Jr. (1972), "A Critical  Review of the  Gas-Phase  Reaction
     Kinetics of the Hydroxyl Radical," J.  Phys. and  Chem.  Ref.  Data,
     Vol. 1, pp. 535-573.

Wu, C. H.~, and H. Niki (1975), "Methods for Measuring N02  Photodissociation
     Rate," Environ. Sci. Technol..  Vol. 9, pp.  46-52.

Wu, C. H., S. H. Japar, and H. Niki  (1976), "Relative Reactivities  of  HO-
     Hydrocarbon Reactions from Smog Reactor Studies,"  J.  Environ.  Sci.
     and Health, Vol. A-ll, pp. 191-200.
                                   313
-------
                                  TECHNICAL REPORT DATA
                           (Please read Instructions on the reverse before completing)
1  REPORT NO.
                                                          3. RECIPIENT'S ACCESSION"NO.
   EPA-600/3-79-Q01a	|_	
4. TITLE AND SUBTITLE
 MODELING  OF  SIMULATED PHOTOCHEMICAL SMOG WITH  KINETIC
 MECHANISMS    Volume 1.  Interim Report
                                                      5. REPORT DATE
                                                       January  1979
                                                      6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
 G.I.  Whitten,  H.  Hogo, M.J. Meldgin, J.P.  Killus,
 and  P.  J.  Bekowies
                                                          8. PERFORMING ORGANIZATION REPORT NO.
                                                          EF78-121A
9. PERFORMING ORGANIZATION NAME AND ADDRESS
  Systems Applications, Incorporated
  950 Northgate  Drive
  San Rafael,  California  94903
                                                      10. PROGRAM ELEMENT NO.

                                                        1AA603   AC-19  (FY-78)
                                                      11. CONTRACT/GRANT NO.

                                                        Contract  No.  68-02-2428
12. SPONSORING AGENCY NAME AND ADDRESS
                                                          13. TYPE OF REPORT AND PERIOD COVERED
  Environmental  Sciences Research Laboratory-RTP,  NC
  Office  of  Research and Development
  U.S.  Environmental Protection Agency
  Research Triangle Park, N.C.  27711	
                                                        Interim   7/76-7/78
                                                      14. SPONSORING AGENCY CODE
                                                        EPA/600/09
15. SUPPLEMENTARY NOTES
  Volume  2.   Appendix   EPA-600/3-79-001b
16. ABSTRACT
       Computer modeling of smog chamber data  is  discussed in three parts.   First,
 a  series  of detailed chemical mechanisms were developed to describe the photo-
 chemical  formation of ozone from nitrogen  oxides and the following organic
 compounds (alone and in various combinations):   formaldehyde, acetaldehyde,,
 ethylene, propylene, butane, 1-butene, trans-2-butene, and 2,3-dimethylbutane.
 Second,  a generalized kinetic scheme  intended for use in models simulating
 the  formation of ozone in urban atmospheres  was refined.  The generalized
 mechanism includes a condensed version of  the detailed mechanisms developed
 in the first part plus a semi-empirical  scheme  to describe the oxidation of
 aromatic  hydrocarbons.  Third, the  effects of smog chambers on ozone formation
 were examined.  For this part of the  study,  similar experiments using  nitrogen
 oxides and propylene in eight different  smog chambers were simulated using
 the  detailed propylene mechanism.   The main  chamber effects identified thus
 far  are apparently due to nitrogen  oxides  degassing from the walls during
 experiments and differences between chambers in the spectral distribution
 of ultraviolet irradiation..
       Volume 1 contains all textual  material.  Volume 2 contains graphs of  measured and
 simulated pollutant conr.pnt.rati'nn<;  fnr many  <;mng chamber experiments 	
17.
                               KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                         b.IDENTIFIERS/OPEN ENDED TERMS  C.  COSATI Field/Group
  *
  *
Air pollution
Reaction kinetics
Photochemical  reactions
Test chambers
Mathematical models
Computerized simulation
13B
07D
07 E
14B
12A
09B
13. DISTRIBUTION STATEMENT
  RELEASE TO PUBLIC
                                              19. SECURITY CLASS (ThisReport/
                                               UNCLASSIFIED
                                                                    21. NO. OF PAGES
                                                                       332
                                                   CLASS (This page)
                                                                    22. PRICE
EPA Form 2220-1 (9-73)
                                            314
-------











m
-o
en
O
O

oo
i
O
0
1— '









*8*
O "^ X
^ ^ o

Qj •» CX.
5! c^
~*i >•• ^
Co ^J r>*
* ^ ^
^~ ^
33 ^
PI
Qj t3
> C
o ^
^s ^
ll
^~ 3
§ ^
> ^
"^ Q)
^ s
*-•+
f^. 3
O C"!
^ ^


c? -*>
Qj ^
"^ O
o^ c
-"^ Q)
Q) a
a 'x
Q- 5
 On
? 5
-i 3
0 0
o
3- ^
Cr ""*
ft) re
Ql Q)
Q- fb
i" °
£ a
"CD
0
3.
H)
o-
o
s
0)
to-
rt)
Z 13

m 5
0 -n -n

  0 o
c < c
z > w

  m
    rn
m
^S
"0 -
r (f>
O u
< S
m °
  S1 O m

  !' f? <
  3 CD 33

g» I a I

2 — CD m
-3 -n C/) -^
2 « ? 5
           cc -j
           QJ O
           -? zr
       Tl
       33
3" Q)

ias
5 o  m
3 CD  O
o> <  _i
   O q ^
   CD = >
   -n CD ^
   IE 8
       z
       o
-------