EPA-460/3-76-005
April 1975
      OZONE AND PARTICULATE
                       FORMATION
             IN PHOTOCHEMICAL
                       REACTIONS
      l.S. ENVIRONMENTAL PROTECTION AGENCY
          Office of Air and Waste Management
       Office of Mobile Source Air Pollution Control
         Emission Control Technology Division
            Ann Arbor. Michigan 18105

-------
                              EPA-460/3-76-005
OZONE AND PARTICULATE
           FORMATION
     IN PHOTOCHEMICAL
           REACTIONS
         Chi-Hunp Shcn and George S. Springer

            The University of Michigan
            Ann Arbor, Michigan 48104

              Grant No. R-801476
         EPA Project Officer: Gordon J. Kennedy
                Prepared for

      l'.S. EMIRONMENTAL PROTECTION AGENCY
          Office of Air and Waste Manpemcnt
       Office of Mobile Source Air Pollution Control
         Kmi.-.-ion Control Technology Division
            Vim Arbor. Michigan 18101^

                 Xpril I«»7S

-------
This report is issued by the Environmental Protection Agency to report
technical data of interest to a limited number of readers.  Copies are
available free of charge to Federal employees,  current contractors and
grantees, and nonprofit organizations - as supplies permit - from the
Air Pollution  Technical Information Center, Environmental Protection
Agency,  Research Triangle Park, North Carolina 27711; or, for a fee,
from the National Technical Information Service,  5285 Port Royal Road,
Springfield, Virginia 22161.
This report was furnished to the Environmental Protection Agency by
The University of Michigan, Ann Arbor, Michigan 48104, in fulfillment
of Grant No. R-801476.  The contents of this report are reproduced herein
as received from The University of Michigan.  The opinions, findings,
and conclusions expressed are those of the author and not necessarily
those of the Environmental Protection Agency.  Mention of company or
product names is not to be considered as an endorsement by the Environmen-
tal Protection Agency.
                 Publication No. EPA-460/3-76-005
                                    11

-------
                             ABSTRACT

     Experiments were performed to investigate the photochemical  forma-
tion of nuclei and aerosols in sulfur dioxide-air mixtures and the photo-
chemical formation of ozone, nuclei, and aerosols in cyclohexene-nitrogen
dioxide-air mixtures.  The mixtures were irradiated by ultraviolet fluor-
escent lamps while flowing at a steady speed through a 9.12 m long and
15.2 cm i.d. Pyrex tube.  The amounts of ozone, nuclei, and aerosols  gen-
erated in the gas-air mixtures were measured at various points along  the
tube by a chemiluminescent analyzer, a condensation nuclei counter, and
an aerosol counter, respectively.
     Photooxidation experiments with sulfur dioxide-air mixtures  were
performed for five different initial sulfur dioxide concentrations (0.2,
0.3, 0.5, 0.7, and 1 ppm) and for four relative humidities (25, 50, 75,
and 100%).  The data showed that the maximum number of nuclei  produced
is directly proportional to the initial  sulfur dioxide concentration  and
to the square of the relative humidity.   There were no measurable amounts
of aerosols observed during the tests.
     Experiments with cyclohexene-nitrogen dioxide-air mixtures were
performed with the initial cyclohexene concentration ranging from 0.5
to 50 ppm and the initial nitrogen dioxide concentration from 0.6 to
10 ppm.  The relative humidity was kept constant at 50%.  The amount  of
ozone produced depended on the initial concentrations of cyclohexene
and nitrogen dioxide as well as on the light intensity.  On the basis
of the present measurements and the data obtained by previous  investiga-
tors, the maximum amount of ozone was found to be given by the expressions
                               m

-------
and
Here, k, represents the nitrogen dioxide photodissociation rate constant
in s   and is proportional to the light intensity.   R is the concentration
ratio of the hydrocarbon to nitrogen dioxide at which the ozone content
is maximum.  The value of R was found to be related to the relative
                       A             nr*
hydroxyl rate constant kg by R = 5.1/|kg .
     The maximum number of nuclei produced varied directly with the
square root of both the cyclohexene concentration and the nitrogen di-
oxide concentration.
     The maximum number of aerosols produced appeared to depend only on
the concentration ratio of cyclohexene to nitrogen dioxide and seemed to
be insensitive to the initial cyclohexene and nitrogen dioxide concen-
trations.  The aerosol size distributions were also measured and followed
closely Friedlander's universal self-preserving curve.

-------
                         TABLE OF CONTENTS

                                                                   Page



LIST OF FIGURES	     vi

L 1ST OF TABL ES	     i x

LIST OF APPENDICES	      x

CHAPTER

     I.  INTRODUCTION	      1

         1.1  Literature Survey	      2

         1.2  Objective	      5

    II.  EXPERIMENTAL APPARATUS	      7

         2.t  Test Section	      7

         2,2.  Gas Supply System	     14

         2.3  Measuring Instruments	     18

   III.  EXPERIMENTAL PROCEDURE	     28

    IV.  PHOTOCHEMICAL PARTICULATE FORMATION IN SULFUR

         DIOXIDE-AIR MIXTURES	     40

     V.  OZONE AND PARTICULATE FORMATION IN CYCLOHEXENE-

         NITROGEN DIOXIDE-AIR MIXTURES	     52

         5.1  The Effects of Mixture Composition on Ozone
              Formation	     53

         5.2  The effects of Mixture Composition on Particulate
              Formation	     77

    VI.  CONCLUDING REMARKS	     94

APPENDIX	     96

REFERENCES	    122

-------
                          LIST OF FIGURES
Figure                                                             Page
  2.1      Schematic of experimental  apparatus	      8
  2.2      Schemati c of the sampl i ng  probes	      9
  2.3      Spectral  energy of the  ultraviolet  light  as  specified
          by the manufacturer	     13
  2.4      Correspondence between  the aerosol  diameter  and  the
          channel  number on the multichannel  analyzer	     20
  2.5      Typical  aerosol  size distribution in  room air and
          i n the tunnel	     21
  2.6      Size distribution of aerosols  in  room air	     25
  3.1      Variation of ozone and  nuclei  contents with  time
          along the tunnel	     29
  3.2      Variation of ozone, nuclei,  and aerosol contents
          with time along the tunnel	     30
  3.3      Spectral  energy of the  light inside the tunnel	     32
  3.4      Ozone content along the tunnel in various nitrogen
          di oxi de-ai r mi xtures	     35
  3.5      Relationship between the photostationary  content of
          ozone and the initial concentration of nitrogen
          dioxide	     36
  4.1      Number of nuclei formed as a function of  irradiation
          time in  sulfur dioxide-air mixtures	     41
  4.2      Number of nuclei formed in sulfur dioxide-air mixtures.     43
  4.3      Number of nuclei formed in sulfur dioxide-air mixtures.     44
  4.4      Number of nuclei formed in sulfur dioxide-air mixtures.
          Correlation of the data for  various initial  sulfur
          dioxide  concentrations  [S02]0  and initial  relative
          humidities   0	     46
  4.5      Formation of sulfuric acid in  sulfur  dioxide-air
          mi xtures	     49
  4.6      Amount of light scattered  from sulfur dioxide-air
          mi xtures	     51
                                VI

-------
                    LIST OF FIGURES  (continued)

Figure                                                             Page

 5.1       Typical  variations of ozone,  nuclei,  and  aerosol
          contents along the tunnel  in  cyclohexene-nitrogen
          dioxide-air mixtures	     54

 5.2       Ozone content in  cyclohexene-nitrogen  dioxide-air
          mi xtures	     58

 5.3       Ozone content in  cyclohexene-nitrogen  dioxide-air
          mixtures	     59

 5.4       Normalized ozone  content in cyclohexene-nitrogen dioxide-
          air mixtures	     64

 5.5       Normalized ozone  content in hydrocarbon-nitrogen oxides-
          ai r mi xtures	     65

 5.6       Maximum  amount of ozone M  as  a  function of  the  con-
          centration ratio  R at which M occurs	     69

 5.7       Maximum  amount of ozone as a  function  of  the  initial
          hydrocarbon concentration	     70

 5.8       Maximum  amount of ozone as a  function  of  the  initial
          nitrogen oxides concentration	     72

 5.9       Variation in the  concentration  ratio  R (at  which the
          ozone content is  maximum)  with  the  reactivity kg 	     75

 5.10     Ozone content in  cyclohexene-nitrogen  dioxide-air
          mi xtures	     76

 5.11      (a)   Oxidant content  in the atmosphere, and (b) ozone
          content  in propylene-nitric oxide-air  mixtures	     78

 5.12     Nuclei content in cyclohexene-nitrogen dioxide-air
          mixtures	     80

 5.13     Nuclei content as a function  of the initial concentra-
          tion ratio in cyclohexene-nitrogen  dioxide-air  mixture.     82

 5.14     Aerosol  content in cyclohexene-nitrogen dioxide-air
          mixtures	     85

 5.15     Aerosol  content as a  function of the  initial  concen-
          tration  ratio in  cyclohexene-nitrogen  dioxide-air
          mixtures	     86
                                vii

-------
                    LIST OF FIGURES  (continued)

Figure                                                             Page

 5.16     Light scattering from aerosols  produced  in  cyclo-
          hexene-nitrogen oxides-air mixtures	    90

 5.17     Aerosol  size distribution  obtained  in  the tunnel for
          cyclohexene-nitrogen dioxide-air  mixtures	    92
                               vm

-------
                          LIST OF TABLES
Table                                                              Page
 2.1      Flowrates of the FTowmeters used in the Experiments	   16
 3.1      Rates of Photodissociation of Nitrogen Dioxide in the
          Lower Atmosphere in the Wavelength Range 29° to 430 nm...   38
 3.2      Summary of Values of kj used in Differerent Experiments..   39
 4.1      Summary of Experimental Conditions used by Different
          Investigators	   47
 5.1      The Reaction Mechanism Proposed by Eschenroeder and
          Martinez [19]	   56
                                                  A
 5.2      Reactivities of Different Hydrocarbons (k = k/kc3H6)	   74

-------
                        LIST OF APPENDICES



Appendix                                                           Page



  A.      LITERATURE SURVEY	     96



  B.      CERTIFIED CONCENTRATION OF TEST GASES	    106



  C.      ESTIMATION OF PARTICLE LOSSES DUE TO SETTLING,  COAGULA-



          TION, AND DIFFUSION	    107



          C.I  Loss due to Settling	    107



          C.2  Loss due to Coagulation	    109



          C.3  Loss due to Diffusion	    109



  D.      TABULATION OF EXPERIMENTAL DATA	    112



          D.I  Ozone Content along the Tunnel	    112



          D.2.  Nuclei and Aerosol Contents along the Tunnel	    116



  E.      ESTIMATED PEAK VALUES OF OZONE, NUCLEI..  AND AEROSOLS



          FOR DIFFERENT INITIAL MIXTURE COMPOSITIONS	    120

-------
                                -1-

                             CHAPTER I
                           INTRODUCTION

     It is well recognized that ozone and particulate matter contribute
to the problem of air pollution.  Excessive amounts of ozone in the atmos-
phere accelerate the degradation processes of elastomers and textiles,  and
adversely affect the growth of plants [69].  High concentrations of par-
ticulate matter in the atmosphere reduce visibility [66].  Ozone and par-
ticulate matter are also major constituents of smog [66,69].  In order
to determine the amounts of ozone and particulates  in the atmosphere,
and to establish effective means for controlling the production of these
pollutants, it is necessary to understand in detail the processes by
which ozone and particulates are generated..  This study is addressed to
one aspect of the problem, namely, the formation of ozone and particu-
lates due to photochemical reactions.
     Ozone is formed photochemically in the atmosphere when a mixture
of nitrogen oxides (nitrogen dioxide or nitric oxide) and certain types
of hydrocarbons is irradiated by sunlight [3,8,31,53,69].  Nitrogen oxides
and hydrocarbons are usually introduced into the atmosphere by automo-
tivevehicles [68,70].  Hence, large amounts of ozone are produced in
areas where the traffic is heavy and where the weather conditions are
such that there is much sunshine but little chance  for the automobile
exhaust to disperse rapidly.
     Particulates  are formed photochemically when  air containing sulfur
dioxide, or nitrogen oxides and hydrocarbons is irradiated [3,8,28,53,63,66].
     In reference [66], HEW has suggested that "particle"  or "particu-
     late" be defined as any dispersed matter, solid or liquid (except
     uncombined water), in which the individual  aggregates are larger
     than single molecules (  0.2 n'"dia.) but smaller than about 500 yw-m
     in diameter.                                                    '

-------
                                 -2-
Since automotive vehicles may emit considerable quantities of sulfur
dioxide, nitrogen oxides and hydrocarbons,  photochemical  particulates
may form whenever there is heavy traffic.   Sulfur dioxide is also in-
troduced into the atmosphere by power plants burning high sulfur content
fuels [67].  In the vicinity of such plants, the amount of particulates
produced due to photooxidation of sulfur dioxide may also be significant.
     In view of the importance of the ozone and the particulates formed
by photochemical reactions, the mechanisms  by which they are produced
need to be understood and the amounts generated need to he kno/\m.  Al-
though this probleir has been investigated widely, many facets of the
problem still remain unclear.  The overall  objective of this study was
to perform experiments with the aim of obtaining information which will
contribute to the general understanding of  the photochemical formation
of ozone and particulates, and provide quantitative data on the amounts
of ozone and particulate matter formed.  The specific goals of this
study are listed at the end of this section.  First, however, a brief
survey is given of the previous relevant investigations.  A detailed
survey of past research efforts is presented in Appendix A.

1.1  Literature Survey
     In most previous experiments studying  photochemical  reactions, the
procedure normally employed was the following.  The gas-air mixture to
be studied was introduced into a chamber.   The chamber was then irrad-
iated by sunlight or artificial ultraviolet light, and selected reaction
products were monitored.
     The production of ozone during the photooxidation of hydrocarbon-
nitrogen oxides-air mixtures has been studied by several investigators

-------
                                -3-
[2,5,6,7,10,25,31,41,50,51,58].  The data are generally presented in one
of three ways:  1) the initial ozone formation rate [25], 2) the ozone
dosage (i.e. the integral of the ozone content versus time curve [6,7,10],
and 3) the amount of ozone produced [2,5,31,41,50,51].  Of these three
types of measurements, the last one is of the greatest interest to us,
because only this type of measurement provides data which gives directly
the maximum amount of the ozone generated.
     The production of ozone in the presence of nitrogen oxides and
several different types of hydrocarbons has been studied.  Ozone mea-
surements with paraffins were reported in references [6,31,50], with ole-
fins in references [2,5,10,25,31,51,58], with aromatics in references
[7,25,31], and with aldehydes in reference [41].  Unfortunately, the in-
formation provided by many of the available data were limited,  because
most measurements were made only for a few selected initial concentra-
tions of the hydrocarbon and the nitrogen oxides.  Consequently, it is
difficult to extend the applicability of the data to conditions beyond
which they were obtained, to correlate the results of the various ex-
periments, and to draw general conclusions from the existing data.  In
order to provide additional insight into the photochemical formation
of ozone, a series of tests would be needed in which the initial con-
centrations of the hydrocarbon and the nitrogen oxides were varied sys-
tematically over a wide range.
     The particulate formation in hydrocarbon-nitrogen oxides-air mix-
tures and hydrocarbon-nitrogen oxides-sulfur dioxide-air mixtures has
also been investigated [27,29,47,49,53,59].  In most of the previous
particulate experiments, the initial concentrations of the hydrocarbon
and the nitrogen oxides were kept constant and only the type of hydro-

-------
                                 -4-
carbon was  varied.  Experiments have been performed with the following
types of hydrocarbons:  paraffins [29,47,49], olefins [27,29,47,49,53,59],
cyclo-olefins [28,29,47,49,59], diolefins [29,47,49,53], and aromatics
[28,29,49].  The results of these investigations indicate that certain
types of hydrocarbons produce particles under certain experimental con-
ditions.  The results, however, do not provide information for those
concentrations which have not been tested.  In order to obtain a better
understanding of the particulate formation process, a comprehensive ex-
periment would be necessary in which the initial concentrations of the
hydrocarbon and the nitrogen oxides were varied over a range.
     Several experiments have been performed in the past for measuring
particulates in sulfur dioxide-air mixtures [13,14,17,24,40,48,49].
The  major goal of these experiments was to determine the effects of the
initial sulfur dioxide concentration and the relative humidity on the
number of particles formed.  Most results indicate that the number of
particles formed increased with increasing amounts of initial sulfur
dioxide concentration.  However, the relationship between the amount of
particulate matter and the initial sulfur dioxide concentration is not
yet  well established.  The effect of humidity on the amount of particu-
late matter formed is also subject to debate.  For example, Gerhard and
Johnstone [24] and Quon et. al. [48] found that the humidity had no
effect on the amount of particles formed; Renzetti and Doyle [49], demon-
strated that humidity influenced significantly the formation of par-
ticles.  Thus, a need for further study of the effects of the initial
sulfur dioxide concentration and the relative humidity on particulate
formation is evident.

-------
                                -5-
1.2  Objective
     The foregoing survey indicates the need for further work on the
photochemical formation of ozone and particulates in mixtures of hydro-
carbon-nitrogen oxides-air mixtures in which the initial concentrations
of the hydrocarbon and the nitrogen oxides are varied over a wide range.
Particulate formation in sulfur dioxide-air mixtures also require fur-
ther investigation.  Therefore, the major goals of this study were:
1) to design and build an apparatus suitable for studying photochemical
reactions in gas mixtures, and 2) to determine the amounts of ozone  and
particulates formed in selected mixtures.  In particular, experiments
were performed to investigate the following problems.
     1)  Pa-rbiculate formation in sulfur dioxide-air mixtures and the
evaluation of the amount of particulates produced as a function of
a) the length of time the mixture is irradiated, b) the initial sulfur
dioxide concentration, and c) the relative humidity.
     2)  Ozone formation and particulate formation in a hydrocrabon-
nitrogen dioxide-air mixture, and the determination of the amounts of
ozone and particulates formed for different initial concentrations of
the hydrocarbon and the nitrogen dioxide.  In the present study,
cyclohexene was selected as the hydrocarbon because it has been shown
to produce large quantities of particulates in photochemical reactions
[29,47,49,59].
     The ozone concentration was measured by a chemiluminescent analyzer.
The particles were examined in two different categories.  The first
category contained those particles whose diameters were between 2 nm
and 0.1 yu m.  These particles were designated as condensation nuclei and
were measured by a condensation nuclei counter.  In the second category

-------
                                -6-
were those particles whose diameters were greater than about 0.3^-m.
These particles were designated as aerosols,  and were measured by a
Royco Particle Counter.  The number concentrations (no.  per cur)  of
the condensation nuclei and of the aerosols,  as  well  as  the size  dis-
tribution of the aerosols were all measured in the present experiment.

-------
                                -7-

                             CHAPTER II
                      EXPERIMENTAL APPARATUS
     The experiment was performed in a flow type apparatus which con-
sisted of:  1) the test section, 2) the gas supply system, and 3) various
measuring instruments (Fig. 2.1).

2.1  Test Section
     The test section had three major components, the irradiation tunnel,
the gas mixing chamber, and the lighting system (Fig. 2.1).
     The photochemical reactions were generated in the irradiation
tunnel.  This tunnel was made from six identical Pyrex glass tubes joined
together.  The tubes were arranged horizontally in a straight line.
Pyrex glass was selected because of its excellent chemical resistance
and constant ultraviolet transmissivity [39,58].  Each tube was 1.52 m
long and 15.2 cm i.d. with a wall thickness of 0.84 cm.  The tubes were
connected by compression couplings lined with Teflon gaskets.  The six
sections together were 9.12 m long and had a total volume of approxi-
mately 0.166 m3.  Two openings were made in the wall of each tube,
through which the sampling probes were inserted.  These openings were
located on top of the irradiation tunnel.  The first opening was 23 cm
from the entrance of the tunnel.  The distance between any two successive
openings was 76 cm.  A 5 cm long standard 10/30 outer ground glass joint
was welded to each opening (Fig. 2.2).
     Two types of sampling probes were used in the experiments, one for
sampling particulates and one for sampling ozone (Fig. 2.2).  The probes
for the particulates were made from 0.636 cm o.d. stainless steel
tubing with a wall thickness of 0.075 cm.  Each probe was 30 cm long.

-------
                         Mixer
                                                  Fluorescent
                                                  Lamps   t-Photodiode
Manometer [

        Flowmeters
	-*To Exhaust     ^- Shell
i—Filter
         ^- Humidifier
                                                                  To
                                                                  Exhaust
• Thermocouples
o Sampling Holes
                   Test Gases
                                 Filter
        — Air Inlet
    Compressor
Storage Tank
                                          00
         Fig. 
-------
                         (a)
(b)
 Stoinless
 Steel Probe
 Sampling
 Holes
 Pyrex
 Tube
 Tygon Connecting
 Tube
Outer
Shell
 Teflon
 Joint
              Polyethylene
               Connecting
                     Tube
                     Outer
                     Shell


                    Teflon
                     Joint
                   Needle
                    Pyrex
                     Tube
                                                                                          
-------
                               -10-
One end of each probe was sealed.  Twelve 0.223 cm diameter holes were
drilled into each probe.  The holes were arranged in pairs, six holes
being on one side of the probe and six on the opposite side.   The first
two holes were 1.27 cm from the sealed tip of the probe.   The distance
between each subsequent pair of holes was 2.54 cm.   The probes were in-
serted into the irradiation tunnel through the ground glass openings,
and were held in place by standard 10/18 Teflon adaptor joints fitted
with Viton-A 0-rings.  The probes were mounted vertically so that the
closed tips were touching the bottom of the tunnel.   The axes of the
holes were normal tj the direction of the flow, i.e., the holes faced
the walls of the tunnel.  The open end of each probe was  outside the
tunnel and was connected to a straight bore Teflon  stopcock by a 15 cm
long tygon tube.
     The ozone sampling probes were 26 gauge (0.02  cm i.d.  and 1.27 cm
                                                             2
long) hypodermic needles.  The needles were connected to 1  cm  nominal
capacity syringes (9 cm long and 0.48 cm i.d.).  The stems  of the sy-
ringes were mounted in the glass joints using sv.andard 10/18 Teflon
adaptors fitted with Viton-A 0-rings.  The needles  were placed ver-
tically, with the openings about 4 cm deep inside the irradiation tunnel
The plungers inside the syringes were removed, and  one end  of a 60 cm
long polyethylene intramedic tubing (0.376 cm i.d.)  was inserted into
each syringe.  The other end of the tubing was either connected dir-
ectly to the ozone analyzer (during measurements) or was  sealed with
a C-clamp.
     The mixing chamber, made from a 30.4 cm long,  15.2 cm  i.d. and
15.46 cm o.d. stainless steel tube, was connected to the upstream end
of the irradiation tunnel (Fig. 2.1).  The upstream end of  the chamber

-------
                               -11-
was covered by a stainless, steel plate fitted with three 0.636 cm Swage-
lok connectors.  One of the connectors was connected to the gas supply
system and one to a mercury manometer.  The third connector served as a
thermocouple feedthrough for a copper-constantan thermocouple placed
about 2 cm inside the mixing chamber.
     The gas-air mixture inside the chamber was stirred by a 12.7 cm
diameter fan rotating at 500 rpm.  The fan was located 6 cm from the
cover plate with its axis coinciding with the axis of the chamber.  The
fan was driven by a variable speed motor placed outside the mixing
chamber and connected to the fan shaft by a rotary feedthrough.
     Six stainless steel screens, serving as flow straighteners, were
placed downstream of the fan.  These screens were made of 0.08 cm thick
perforated plates.  The perforations were 0.318 cm diameter holes ar-
ranged uniformly in square arrays 0.48 cm apart from each other.  The
plates were cut into 15.2 cm diameter circles, and were fitted inside
the mixing chamber so that the holes faced the flow.  The screens were
placed side by side with 2.5 cm between each screen.  The last screen
was at the end of the chamber.
     The downstream end of the irradiation tunnel was connected to a
15.2 cm long hollow stainless steel  cone (15.2 cm i.d. at upstream end,
10.2 cm i.d. at downstream end and 0.13 cm wall thickness).  The down-
stream end of the cone was covered by a stainless steel plate, which
had three openings in it.  The first was 2.54 cm in diameter and was
connected either to an exhaust vent or a vacuum pump.  The second served
as a feedthrough for a copper-constantan thermocouple probe placed 5 cm
inside the tunnel.  The third was used to insert a light intensity de-
tector into the tunnel.

-------
                               -12-
     The ultraviolet radiation was produced by forty-eight F40BL (General
Electric) fluorescent lamps.  The initial  spectral  distribution of the
lamps, provided by the manufacturer, is shown in Fig.  2.3.  Each lamp
was 1.22 m long and 3.8 cm in diameter.  The lamps  were arranged in six
clusters along the tunnel, each cluster illuminating one section of the
tunnel.  Each cluster contained eight lamps placed  around the tunnel
systematically on a 31.7 cm diameter circle.  The lamps in each cluster
were mounted on two semi-circular (18.5 cm radius)  galvanized metal
shells.  The insides of the shells were covered with aluminum foils.
The shells thus se;ved both as light fixtures and reflectors.  The shells
also provided ducting for the cooling air as explained below.  The
shells were hinged together and could be opened to  install the lamps
and the sampling probes.  There were small holes along the upper edges
of the shells to allow the sampling probes to pass  through.
     The ultraviolet lamps generated a large amount of heat.   To mini-
mize the temperature rise of the gas-air mixture inside the tunnel [9],
room air was blown through the annul us between the  irradiation tunnel
and the sheTT.  This air was blown by a 45 cm fan in the direction oppo-
site to the flow inside the tunnel.  The cooling air flowed along the
entire length of the tunnel and was then discharged into an exhaust
vent.  The temperature of the cooling air was monitored by four copper-
constantan thermocouples placed inside the annulus  at equal intervals
along the tunnel.  The maximum temperature rise of  the cooling air was
7°C.  The temperature rise of the gas-air mixture in the tunnel (i.e.
the temperature difference between the gas-air mixture in the mixing
chamber and at the exhaust port) was always less than 4°C.

-------
                          -13-
    1.5
 E
 c
O

~  i.o


o
LJ
<0.5

I-
o
UJ
0.
                                              n
     300
400           500

 WAVELENGTH, nm
600
   Fig. 2.3.  Spectral energy of the ultraviolet light as specified
            by the manufacturer (General  Electric Fluorescent Lamp
            F40BL)

-------
                              -14-
2.2  Gas Supply System
     The air used in the experiments was supplied from an  air-conditioned
room kept at a constant temperature (20°C)  and at a constant  relative
humidity (75%).  The air was compressed by  an oil-free,  diaphragm type
compressor (Sears-Model No.  15122)  into a 0.03 m3 storage  tank.   The air
                                                        2              2
pressure inside the tank was maintained between 110 kN m   and 165 kN m
gauge pressures.  The air stream from the tank passed  through four traps
connected in series.  The first trap, consisting of 0.45 kg of activated
carbon in a Plexiglas container, was installed to remove tract; gases pre-
sent in the air.  The second trap was a 1.5 m long, 1.9  cm i.d.  tygon
tubing filled with anhydrous calcium chloride.   The third  trap was 0.45kg
of "indicating" anhydrous calcium sulfate packed in a  glass jar.   These
latter two traps were intended to absorb the water vapor in the air
stream.  A change in color in the calcium sulfate from blue to pink in-
dicated that the calcium chloride had been  saturated and needed replace-
ment.  The fourth trap was a Gel man type A  glass fiber filter for re-
moving atmospheric aerosols  from the air stream.  After these four stages
of purification, the air was branched into  two streams.  One  stream,
designated as "dry air", passed through a flowmeter and  a  needle valve.
The other stream, which was  to be saturated with water vapor, passed
through a flowmeter, a needle valve, and a  humidif;er.  The humidifier
was made from a 19 cm high and 12.7 cm i.d. cylindrical  Plexiglas chamber.
The chamber was packed three-quarters full  with 0.4 cm diameter glass
spheres and filled with double-distilled water so that the water level
was about 1 cm above the glass spheres.  The temperature of the humidi-
fier was kept constant (~'250C) by  electrical heating  tapes surrounded

-------
                               -15-
by thermal insulation.  The air was bubbled through the humidifier.   The
air, thus saturated with water vapor (100% relative humidity), was led
through a water trap consisting of a 10 cm high and 4 cm i.d.  cylindrical
Plexiglas chamber.  After passing through the trap and a needle valve,
the saturated air was reunited with the dry air stream.  The relative
humidity of the test air was controlled by adjusting the flowrates of the
two streams of air.  For example, equal flowrates of the dry and sat-
urated stream would result in an air of 50% relative humidity.  The air
stream was led through a 0.636 cm o.d., 0.486 cm i.d., and 60 cm long
stainless steel tube until it was mixed with the test gases.
     The test gases were supplied as special blends and were pre-diluted
with pure dry nitrogen to approximately 200 ppm by the manufacturers.
Nitrogen dioxide was obtained from the Air Products and Chemicals Com-
pany.  Sulfur dioxide and cyclohexene were supplied by the Matheson Gas
Company.  The composition of each gas was individually analyzed and
certified by the supplier (see Appendix B).  The supply line of the
test gases were made of 0.636 cm o.d. and 0.486 cm i.d. stainless steel
tubing.  The flowrates of each gas were regulated by stainless steel
needle valves and were measured by two flowmeters connected in parallel.
These flowmeters operated at different ranges of flowrates (see Table
2.1), allowing accurate measurements of the flowrates (and consequently
the concentration of the test gases) over a wide range of values.  After
the flowmeters, the gas lines were joined with the line containing the
humidity-controlled air.  The entire gas-air mixture then passed through
another Gelman type A glass fiber filter and a three-way stainless
steel ball valve.  This valve was used to admit the mixture either
directly into the mixing chamber or into an exhaust vent.

-------
                             -16-
Table 2.1    Flowrates of the Flowmeters Used in the Experiments
           (all flowmeters were manufactured by Kontes Glass Company)
                                  Flowmeter

    Gas

Dry Air
Moist Air
Sulfur Dioxide
Nitrogen Dioxide
Cyclohexene
Size No.
1221
1221
1123
1123
1423
1123
1423
Flowrate, cnrs
12-130
12-130
0.08-1.75
0.08-1.75
0.33-17.9
0.08-1.75
0.33-17.9

-------
                                -17-
      It is noted here that the lengths of the gas supply lines from the
 tanks to the flowmeters were about 4 m, from the flowmeters to the con-
 nection with the air 0.6 m,  and from this connection to the mixing cham-
 ber 1.5 m.
      After the apparatus was assembled, the downstream end of the tunnel
 was connected to a mechanical  vacuum pump, and the entire system was
 tested for leakage.   The lowest pressure obtained in the system (in-
 cluding the gas lines,  the mixing chamber, and the irradiation tunnel)
            2                                          21
 was 66 N m  .  The leak rate of the system was 3 mN m  s  .
      During the experiments, the tunnel was not connected to the vacuum
 pump, but to an exhaust vent.   The pressure in the vent was atmospheric.
 The pressure in the air supply line ahead of the traps was kept at 35
      2
 kN m   gauge.  The pressure  decreased considerably along the supply line
 till  it became practically atmospheric in the mixing chamber, as indi-
 cated by the mercury manometer.  The pressure drop inside the tunnel
                        2
 was less than 0.7 mN m   , so that the pressure throughout the tunnel
 was practically atmospheric  during the tests.
      The flowrate in the tunnel could be varied from 46 cnrs   to
 184 crrr^s  .   The average velocities corresponding to these flowrates
 were 0.26 cm s   and 1.04 cm s"1.   The Reynolds numbers, based on these
 velocities,  were approximately 24 and 96, respectively.  At such low
•Reynolds numbers the flow was  expected to be laminar in the tunnel.
      An estimate of the number of particles lost along the tunnel  due
 to various mechanisms is presented in Appendix C.   It was found that,
 in general,  the loss of particles in the tunnel was negligible except
 when the particle diameter was either smaller than 0.01 M m or greater
 than 1  u m.

-------
                              -18-
2.3  Measuring Instruments
     In this investigation, measurements of the following parameters were
of major concern:  nuclei concentration, aerosol  concentration, ozone
concentration, total light intensity, and  spectral  distribution of the
light.
     The concentration of the condensation nuclei was measured by a
"Small-Particle Detector" manufactured by Gardner Associates,  Inc.  In
this instrument, the sample was drawn into a chamber lined with wet
blotting paper.  The pressure was then suddenly reduced in the chamber,
resulting in water condensing on the nuclei present  in the sample.  A
light beam was directed through the sample, and the  decrease in light
intensity (wftfch is proportional to the number of nuclei) was  measured
by a photomultiplier.  The output of the photomultiplier was displayed
on a meter.  The reading on this meter, together with a calibration chart
supplied by the manufacturer, gave the number of nuclei per cm .  The
instrument could measure nuclei whose diameters were between 2 nm and
0.1 yum.  The measurable concentration range was  from 200 to 1x10
nuclei cm~ .  The fluctuation of the instrument was  within -10% when
the nuclei count was above 1000 per cm3.  The fluctuation became higher
(±20%) when the nuclei count was smaller than 1000 per cm3.  The cali-
bration chart provided by the manufacturer is accurate within  -20%.
Thus the maximum percentage error in the nuclei measurement was about
- 40%.  All condensation nuclei measurements presented in sections IV
and V were the averages of at least three readings.
     The concentrations of the aerosols (i.e. particles having diameters
greater than 0.3 um) were measured by a Royco Particle Counter (Model
225) coupled with a Royco plug-in unit (Module 507).  This instrument

-------
                               -19-
operated as follows:  The sample was drawn into a chamber and a light
was passed through it.  The light scattered in the near-forward direc-
tion by the aerosols in the sample was measured by a photomultiplier.
The output of this photomultiplier was calibrated by the manufacturer
to indicate directly the number of aerosols in a certain size range.
Although the unit could be used to measure the aerosol  content in five
different size ranges, in the present experiment, it was always set to
measure the concentration of all the aerosols larger than 0.3 urn in
diameter.  The measured particle concentration was displayed digitally
and was also recorded by a five-digit printer.
     In order to measure the size distribution of the aerosols, a multi-
channel (128 channels) analyzer (Nuclear Data Inc., Model ND 555) was
attached to the Royco Particle Counter.  The output of the Royco Instru-
ments was fed directly into the analyzer with each channel corresponding
to a given particle size. The particle size corresponding to a parti-
cular channel was obtained by comparing the particle count on the
Royco Instrument with the particle count given by the analyzer.  This
calibration was performed at the five particle size ranges available  on
the Royco Instruments (particle diameter greater than 0.3, 0.7, 1.4,  3,
and 5yum) (Fig. 2.4).  The calibration could not be done accurately  for
particles between 0.7 and 1.4 Mm because, in this size range, the output
of the Royco Instruments varied with the particle size in a complex
manner.
     During the test, the output of the analyzer was displayed on a
Tektronix oscilloscope (type 565 Dual-Beam).  A typical result obtained
with room air is shown in Fig. 2.5.  Each bright dot in the picture

-------
                          -20-
  10.0
   5.0
o:
LJ
h-
LJ
5
O
CO
o
cr
UJ
 r 2.0
   1.0
   0.5
   0.2
                        I
1
1
     20      40      60      80      100

                     CHANNEL NUMBER
                 120
 Fie. .-.4.  Corresponaence between the aerosol  diameter and the  channel
          number on tne n.ultichannel analyzer. O calibration  data,
          —- fit to data

-------
                                -21-
                               (b)
Fig.  2.5.   Typical  aerosol size distributions in a) room air and b)  in
           the  tunnel.  Vertical axis represents the voltage output  from
           the  multichannel analyzer (0.1 volt per division):   Voltage  is
           proportional to the number of particles (eq.  2.1).   Horizontal
           •axis represents channel number and, correspondingly, particle
           size

-------
                                -22-
corresponds to a channel, with the first dot on the right hand side
corresponding to channel #128.  The correspondence between the channel
number and the aerosol size is given in Fig. 2.4.   The vertical  axis in
Fig. 2.5 has the units of volts and is proportional to the number of
aerosols counted in a particular channel (i.e. in  a given size range)
during a given period of time.  The relationship between the voltage
and the number of aerosols counted was determined  by a selector switch
on the multichannel analyzer.  The selector switch has 8 positions.
At positions 10 , 10  , 10 , or 10  the output of  1 volt per channel
corresponded to 10 , 10 , 10 , or 10  counts, respectively.   When the
selector switch was at positions 4x10 , 4x10^, or  4x10 , the output of
                                          23         4
2.5 volts per channel corresponded to 4x10 , 4x10  , or 4x10  counts,
respectively.  For these seven selector switch settings, the voltage
was linearly proportional to the number of aerosols counted.  There was
an eighth setting in which the output voltage for  each channel,  v, was
related to the number of aerosols counted, n, by the expression
               v - c  log1Qn +  VQ                                 (2.1)

where c and v  were undetermined constants.  These constants were evalu-
ated by sampling a mixture, and by measuring the number of aerosols in
a particular channel using one of the seven settings for which n was
linearly proportional to v.  The selector switch was then turned to the
logarithmic scale and the voltage was measured for the same channel.
Repeating this procedure twice, two sets of v and  n values for the log-
arithmic setting were obtained.  Substituting these values into  eq. 2.1,
c and v  could be calculated.  In practice, the procedure was repeated
for several sets of v and n values in order to obtain representative

-------
                               -23-
average values for c and v .  The c and v  values were found to be 0.14
and 0.06 volts, respectively.
     From the results displayed in pictures such as Fig. 2.5, the number
of aerosols per channel could be determined to within 20% accuracy.  How-
ever, the inaccuracy in the aerosol results (both the size distribution
and the total number count) was considerably larger due to the statis-
tical nature of the measurement.  The actual number of aerosols fluc-
tuated and this was manifested in the fluctuations in the outputs of the
Royco Particle Counter and the multichannel analyzer.  Characteristically,
the fluctuation was of the order of ±30%.  All the results presented in
Section V for the total number of aerosol count and for the size dis-
tribution were the averages of at least three measurements.
     The Royco Particle Counter could be operated at sampling rates of
     ol           O    1
47 cm  s   and 4.7 cm  s   .  The lower sampling rate (which was about
one tenth of the flowrate in the tunnel) was used in the experiments.
At this flowrate the lowest and highest detectable aerosol concentrations
                 o              3
were 0 and 3.5x10  particles cm  , respectively.  It is noted here, that
neither the-Small-Particle Detector nor the Royco Particle Counter was
calibrated in our laboratory.  The results reported were based on the
calibrations provided by the maunfacturers.
     In order to test the instruments, the size distribution of the at-
mospheric aerosols in the room were determined.  The data were presented
in terms of the dimensionless parameters
                   A N/Ar
                                                                 (2.3)
                  N/(V/N)1/3

-------
                                -24-
 AN  is  the  number per  unit volume of aerosols whose radii lie within the
 range of r and  r + Ar.  N is the total number of the aerosols per unit
 volume
                                                                  (2.4)
and V  is the total aerosol volume per unit sample volume
                              N                                   (2.5)
The  parameters^ and •& were used in presenting the results because, as
has  been shown by  Friedlander [21], and by Liu and Whitby [45], with
the  use of these parameters, all the data reduce to a common curve.
Moreover, it" toss- been demonstrated both experimentally [11,22,38,45] and
analytically  [45]  that in the size range where the aerosol is in a state
of dynamic equilibrium (i.e. where the rate of gain of aerosols due to
coagulation is equal to the rate of loss of aerosols due to coagulation,
diffusion and sedimentation) \{r varies inversely with the fourth power
of n  ,  i.e.
                                                                 (2.6)
     The results of the measurements are given in Fig. 2.6.  The data
 follow quite closely a line whose slope equals -4.  The present results
 also fall within the size range observed by other investigators [11,22,
 38,45].
     The ozone content was measured by a Thermo Electron Corporation
 Model 12A Chemiluminescent Analyzer.  This instrument operated on the
 principle that light was produced during the reaction of ozone with
 nitric oxide [56],  Therefore, the reaction

-------
                               -25-
       10
o
10°
ID'1

-2
10


.o3

•€
y \
v\
•\
" \ \ "
\ \
\\
0 1
\ -\
\ • \
\ v \
\B
0 \
\-. \
__ Line of \ V • \ 	
\ \ %
Slope -4— — Ji \ ° \
\ * 0 \
\ \ .-\
\-\











33 O.I 1.0 10
r


T\~ ». .,*m

/ v V'**
IN")
Fig. £.b.  Size distribution of aerosols in room air.  o   present data,
           	range of data from Cartwright et.  al.  [H],  Friedlanrier
           and Pasceri [22], Junge [38], and Liu and Whitby  j.*&J

-------
                               -26-
                      NO
was generated inside the instrument and the emitted light was measured by
a photomultiplier.  The photo current produced by the photomultiplier was
a measure of the ozone content in the sample.   A linear,  full scale meter
on the instrument indicated directly the ozone concentration in the
sampled gas stream.   Eight concentration ranges were provided with the
instrument:  0-0.01, 0-0.025, 0-0.1, 0-0.25, 0-1, 0-10,  0-100, and 0-1000
ppm.  Within each of these ranges the ozone concentration  could be assessed
to within two percent precision, while the fluctuations  in the measure-
ments were less than ±5%.
     The total intensity of the ultraviolet light inside  the irradiation
tunnel was monitored by a silicon Schottky  PIN-5 photodiode (United
Detector Technology, Inc.).  This photodiode was 0.915 cm in diameter
and had an active area of 0.04 cm2.  The photodiode was  cemented onto
one end of a 40 cm long, 0.636 cm o.d. stainless steel tube.  The tube
(with the photodiode at its tip) was then inserted about  30 cm deep inside
the irradiation tunnel through one of the Swagelok connectors at the
downstream end of the tunnel.  The electrical  leads of the photodiode
were led through the tube to the outside of the tunnel.   The signal from
the photodiode was processed by an electrical  circuit recommended by the
manufacturer, and was displayed on a digital voltmeter.   The relation-
ship between the output voltage and the light intensity  was not de-
termined directly.  Nevertheless, the voltage output indicated any
change in the light intensity in the tunnel, and any possible deteriora-
tion in the intensity of the fluorescent lamps.  It is noted, however,

-------
                               -27-
that a measure of the total light intensity in the tunnel was obtained
as described in Section III.
     The spectral distribution of the ultraviolet radiation was measured
with a CGA/McPherson Monochromator and an RCA photomultiplier tube
1P28-A.  The light was collected by a 30 cm long, 2.5 cm i.d. and 3 cm
o.d. brass tube which passed through the shell surrounding the tunnel.
One end of the tube was attached to the monochromator.  The other end,
machined to the shape of the outer surface of the tunnel, was attached
to the outer wall of the tunnel.  All openings were carefully sealed so
that only light from the tunnel  could reach the monochromator.  The
spectral distribution of the ultraviolet light, as given by the output
of the monochromator, was recorded on a strip chart recorder.  The
output was not calibrated to give the value of the light intensity at
different wavelengths.  However, the output indicated the magnitudes of
the light intensities and showed any change in these magnitudes.  There-
fore, this measurement sufficed  in the present experiments where the
only concern was to ascertain that the spectral distribution remained
constant throughout the tests.

-------
                               -28-

                            CHAPTER III
                      EXPERIMENTAL PROCEDURE

     The experimental procedures  used in studying the photochemical reac-
tions in sulfur dioxide-air and in cyclohexene-nitrogen dioxide-air mix-
tures were similar.
     At the start of each experiment, the ultraviolet lamps were turned
on, the flow of the cooling air was started, and 50% relative humidity
air (without any of the test gases) was fed through the system contin-
                             0  _1
uously at a flowrate of 46 cnr s   .  The temperatures inside the tunnel
and in the cooling air were recorded.  The system required three to four
hours to reach thermal equilibrium.  The particulate contents were also
measured along the entire length  of the tunnel.  This procedure contin-
ued until both the nuclei and aerosol counters indicated zero readings
(i.e., a "clean" system) at every sampling station.  This condition was
usually reached in four to seven  days.  Once the system was thus
"cleaned", the required gas-air mixture was introduced into the tunnel.
After about four hours, both the  ozone and particulate contents reached
a constant value ("steady state") in the entire tunnel.  This is demon-
strated in Figs. 3.1 and 3.2 where the ozone, nuclei and aerosol con-
tents are shown as a function of  time at different sampling stations.
When the conditions had reached steady state, the ozone and particulate
contents were measured, and the gas-air mixture was changed to a new
value.
     The composition of the gas-air mixture was not measured along the
tunnel.  The mixture composition  was known only at the entrance of
the tunnel.  Therefore, in the subsequent discussions, the specified

-------
 0.6
                                          Stotion 10
                                          Station I
                       -a.
                                                                                               ro
                                                                                               vo
                   •8   8    8    a.
                                          Station 3
                                          Station 9
                                    I
                                    I
                         8         12         16

                              TIME, HOUR
                                             20
24
Fig.  3.1.
Variation of ozone  and  nuclei  contents with time along the tunnel.
Mixture composition:  cyclohexerie  1.25 ppni, nitrogen dioxide 0.5 ppm,
relative humidity 50%.   Mean  flow  velocity 0.255 cm s-1

-------
NUCLEI AND
AEROSOLS OZONE
CONTENT, no cm3 CONTENT, ppm
I.O
1.2
0.8
0.4
0
I03
IOE
1 1 1 1 1 1
o o o
^^~** w o
_ y^ Station 10 _
-If Station 1
f 1 1 1 1 1

1 1 1 1 _ 1

ff " Nuclei \ /
If Station 3 /
_ . , /Stations. __
Aerosol / \
^^o 	 « 	 a 	 O/ a\ P
468

     TIME,  HOUR
                                                 10
                                                                       12
                                                                                     OJ
                                                                                     o
Fig.  3.
Variation  of ozone, nuclei, and aerosol  contents  with  time along the tunnel
Mixture composition:  cyclohexene 1U ppm, nitrogen  dioxide i; ppn,, relative
humidity bU%.   Mean flow velocity 0.255  cm s~l

-------
                               -31-
test gas concentration refers to the amount of test gas in the air
prior to exposure to the ultraviolet light.  This condition is denoted
by the subscript o.
     The spectral distribution of the ultraviolet light and the total
light intensity was monitored throughout the tests.  The spectral dis-
tribution of the light did not vary regardless of the compositions or
concentrations of" the mixture in the tunnel.  Typical spectra taken with
just air in the tunnel (i.e. no photochemical reactions) and with a cyclo-
hexene-nitrogen dioxide-air mixture (in which photochemical reactions
were taking place) are shown in Fig. 3.3.
     The total light intensity also remained constant throughout the
tests, as indicated by the steady voltage output of the photodiode.
As mentioned fn Section II, the output of the photodiode was not cali-
brated.   Thus, while the output indicated that the intensity remained
constant, it did not provide the actual  value of the light intensity
in the tunnel.
     The light intensity,  J, may be related to the photolysis rate con-
stant (k-j) of nitrogen dioxide by the expression [43]
                kl-Z 2.303 £^JA                             (3-D

where 
-------
                         -32-
O
or —
Ul »
«w ^-

il
ui io
             AIR


         (50% Relative Humidity)
         C6H|0-N02-AIR MIXTURE
         (5O%Relative Humidity)  I
             325         350         375

                    WAVE LENGTH,  nm
400
   Ma.  3.3.  jpecLral  Lnercy of the liciht inside the tunnel.

             Top:  enertjy in air.  BotLon:   energy in cycloiiexene-

             nitroqen  dioxide-air mixture

-------
                              -33-
     The dominant reactions at the initial stage of the photochemical
reactions in a nitrogen dioxide-dry air mixture are [8,43,54,57]

                           ki                                    (3 2)
               N02 + \\V - *-»• NO + 0                           l   '

                           k
              0 + 02 + M - -- + Og + M                          (3.3)

                           k
                0  + NO  - l - + N02 + 02                       (3.4)

where k^ , k2 and k3 are the respective rate constants.  The rate equa-
tion for the nitrogen dioxide may be written as

               ~^~ = -kjNO^ k^M                     (3'5)

A nitrogen mass balance at any instant gives

                                                                 (3.6)
where [N02J0 is the initial nitrogen dioxide concentration (i.e.  the
concentration supplied to the system).  Some time after the start of
the photochemical reaction, the "photostationary state" is reached in
the system, at which the concentrations of nitrogen dioxide, nitric
oxide, ozone and oxygen atom do not change with time, i.e.,

                                              • o
Depending upon the initial nitrogen dioxide concentration, this state
was reached in the tunnel in 120-300 seconds.  Equations (3.5) and (3.7)
yield

                                                                 (3.8)

-------
                             -34-
where the subscript pss denotes photostationary state.   For low initial
nitrogen dioxide concentrations ([NOg]  less  than 2 ppm) , [N0j
                                      v                       TjS 5
may be approximated by [03]     [54,57]
                                                                  (3.9)

From eqs.  (3.6),  (3.8)  and (3.9),  we  obtain

                 3lpss
               l°3lpss  -  iElo  "Npss                (3.10)
According to the above equation,  a plot of [°3]LS  against  •[ [N02]   -
[03]    \  should result in a straight line,  the value of k^/k-  being
given by the slope of this line.   Knowing the slope of the  [OslLc
versus  {[NQ^L - ^3^DSS}  curve and tne value of  k3, the  parameter k,
can be calculated readily.  The value of k3  has been determined by  sev-
eral investigators,  the reported values of  k3 ranging from 0.35 to 1.23
ppm"  s   [12,20,37,52].  The value k3 = 0.425 ppm    s   has been
adopted here as recommended by Stedman and Niki [57].
     In order to generate the [°3]DSS versus   {
curve, the ozone concentration was measured  at different points inside
the tunnel using dry air and three different initial  nitrogen dioxide
concentrations.  Since the flowrate in the tunnel was constant, the
distance along the tunnel could be related to a time scale, as  shown
on the top of Fig. 3.4.  The maximum values  of the  ozone concentration
curves were taken as the photostationary state values as these  maximum
points satisfy the condition  d[03]/dt = 0 (see eq. 3.7).   The  data thus
generated was used to plot a [03]pSS versus   f [N02]Q - [03]DSSJ  curve
(Fig. 3.5).  From the slope of this plot, and for k3 = 0.425 ppm"  s~  ,
the value of k, was found to be 6.33 ks  .  Typical values  of k1  for the

-------
 0.16-
                10
   TIME, MINUTE

20          30
40
                    23456

                   DISTANCE  ALONG TUNNEL,  m
50
                      2ppm Initial NO?  Concentration
                                                                                            CO
                                                                                            in
                                                                                            i
Fig.  3.4.  Ozone content along the tunnel in various  nitrogen dioxide-air mixtures.
          Top axis  represents irradiation time

-------
     0035
     O.O3O
     0.025
                              -36-
CSJ
     0.020
 o.
   
 O
      O.OIO
      0.005
                     0.5
1.0
                        [N02]o- [03]oss, ppm
    I

   1.5

'pss
2.0
2.5
        Fig.  3.5.  Relationship between the photostationary content of
                  ozone and  the initial  concentration of nitrogen
                  dioxide,  o data,	fit to data

-------
                               -37-
photodissociation of nitrogen dioxide in the lower atmosphere are pre-
sented in Table 3.1.  As can be seen, the k1 values of the present ex-
periment compare well with those which exist in the atmosphere.  It is
noted that k-, values obtained in other experimental apparatus range from
0.78 ks"1 to 9.17 ks"1 (Table 3.2).  The present k] = 6.33 ks"1 falls in
this range.

-------
                          -38-
Table 3.1   Rates of Photodissociation of Nitrogen Dioxide in
           the Lower Atmosphere in the Wavelength Range 290 to  430  nm
Solar zenith angle                  Photodissociation  Rate
     Z, degree              k,  = £  2.303 4>  <.   J.   , ks"1
                             1    X        'A    A    A
0
20
40
60
80
9.51
9.17
8.02
5.60
1.68
Values of <>  from Table A-II, reference [15]
Values of 
-------
                    Table  3.2   Summary of Values  of k-.  Used in Different Experiments
Volume, nr
0.166
0.15
0.044
0.022
0.141
8.49
1.87
0.09
1.811
9.48
0.09
0.2

Material
Pyrex tube
Teflon bag
Pyrex jar
Pyrex flask
Pyrex tube
Stainless steel chamber
Aluminum, Pyrex, Teflon and
stainless steel removable surfaces
Pyrex flask
Aluminum chamber, Teflon window
Aluminum chamber, poly vinyl fluoride
wi ndow
Mylar and Teflon bags
Pyrex jar
3_m HA CD n^fh miiltinlo K*of 1 or t i nn
in L/uoc pa in multiple i c i ICI*LIUM
infrared cell
Type of Experiment
F
S
S
S
F
S
S
S
S
F
S
F
S
«J
k^ ks'1
6.33
0.78
2.33
3.67
3.83
4.17
5.0
5.33
6.33
6.67
6.67 (Teflon)
5.0 (Mylar)
8.33
Q 17
y • i i
Reference
Present Investigation
Lilian [44]
Stedman & Niki [57]
Laity [42]
Stephens & Price [58]
Groblicki & Nebel [29]
Jaffe & Smith [36]
Holmes et. al . [34]
Dimitriades [16]
Altshuller et. al . [5]
Altshuller & Cohen [1]
Wilson et. al . [65]
TiiP^dav FfiOl
i ucoviujr L^^J
                                                                                                                oo
                                                                                                                vo
                                                                                                                 i
S   Static type experiment

F   Flow type experiment

-------
                               -40-

                            CHAPTER IV
       PHOTOCHEMICAL PARTICULATE FORMATION IN SOg-AIR MIXTURES

     It has been demonstrated by several investigators that participates
form when a mixture of sulfur dioxide and air is exposed to ultraviolet
light [13,17,24,48,49].  However, the number of particulates formed and
the dependence of this number on the initial sulfur dioxide concentra-
tion and the humidity of the air are not well established.   The objec-
tives of the present experiments were, therefore, to measure the number
of particulates formed in sulfur dioxide-air mixtures, and  to determine
the relationships between the number of particulates formed, the initial
sulfur dioxide concentration and the humidity of the air.
     When a sulfur dioxide-air mixture was exposed in the tunnel to
ultraviolet radiation, the number of particulates formed increased with
time, reaching a maximum value.  The number of particulates then re-
mained nearly constant for some time at the maximum value (see Fig. 4.1)
before it decreased due to coagulation, diffusion, and deposition on
the walls.  Here we were interested in the maximum number of particulates
formed in the mixture.  In order to determine the exposure  time required
to reach the maximum number of particulates, the number of  nuclei and
aerosols were measured at sampling station #1 at different  flowrates
                  3-1          3-1
ranging from 46 cm  s   to 184 cm  s  .  Since sampling station #1 was
23 cm from the entrance to the tunnel, the exposure times corresponding
to these flowrates ranged from 90 s to 22.7 s.  The experiments were
repeated for different sulfur dioxide concentrations and different
humidities (Fig. 4.1).  There were no measurable amounts of aerosols in
any of the experiments.  There were, however, nuclei formed in the mix-

-------
   10
UJ
_i
o
D
UJ
m
I     I
                           -41-
     I     I
I     I     I
                                   .Oppm, 50%
                                   0.5ppm,50%
                                   0.2ppm,50%
              0.2ppm,25%
                                Initial S02

                                Concentration, ppm

                                Relative Humidity, %
1111    I     I
                                                   I
     0    10   20   30   40   50   60  70   80  90  100

                   IRRADIATION TIME, sec
     Fiq. 4.1.  Number of nuclei formed as a function of irradiation
              time in sulfur dioxide-air mixtures

-------
                               -42-

ture.  The amount of nuclei  formed depended on the exposure time,  as
shown in Fig.  4.1.   The time required to reach the maximum number  of
nuclei increased as both the initial  sulfur dioxide concentration  and
the humidity in the air decreased.
     The lowest initial sulfur dioxide concentration used in the ex-
periment was 0.2 ppm, and the lowest  relative humidity was 25%.  For
this condition, the tests showed that the time required to form the
maximum number of nuclei was about 90 s.  Therefore, to ensure that the
maximum number of nuclei formed was achieved in all the tests, the 90 s
exposure time was used throughout the subsequent experiments.   The
number of nuclei per cm  given below correspond to the maximum value
obtained in the manner just  discussed.
     The effects of the initial sulfur dioxide concentration and the
relative humidity on the nuclei formation were measured for five differ-
ent sulfur dioxide concentrations ([S09]  = 0.2, 0.3, 0.5, 0.7,  and 1 ppir),
                                      £ o
and four different relative  humidities ([^]Q = 25, 50, 75, and 100%).
The data are presented in Fig. 4.2, and indicats that the number of nuclei
formed vary linearly with the initial sulfur dioxide concentration in
the air.
     The results in Fig. 4.2 were cross plotted in order to obtain dir-
ectly the variation of the number of nuclei with relative humidity.  As
shown in Fig.  4.3, the number of nuclei varies with the square of  the
relative humidity.  Thus, according to the results of Fig. 4.2 and 4.3,
the maximum number of nuclei formed may be expressed as
                N =  ACS0][                                    (4.!)

-------
                         -43-
5x10*
u
 •»

Ul

CJ
u.
o
QC
Id
00
I  I
          oDATA

         — FIT TO DATA
            (Slope =1)
           Relative
           Humidity
              25%
         i  i i I
            I	I
                              I
            O.I                        1.0          3.0

       INITIAL SULFUR DIOXIDE CONCENTRATION, ppm
    Fig. 4.2.  number of nuclei formed in sulfur dioxide-air mixtures

-------
                          -44-
5x10'
UJ
_l
(J
u.
o
a:
UJ
CD
        i  i  i i
      i     i    i   i  i  i  i i
            Initial S02

            Concentration
        i  i
                        0.75
      0.5
0.25 ppm
      I	I
              10                          100

                  RELATIVE HUMIDITY, %
     Fia. 4.3.  Number of nuclei formed  in sulfur dioxide-air mixtures

-------
                                -45-
where A is a constant.  In order to test further the validity of eq.  (4.1)
and to obtain the value of A, all the data were compiled on an N versus
                  ^Fl^' 4-4).  The data correlated well  with the above
expression for A = 3.33x10  no. cm   ppm  .   Thus, the maximum number of
nuclei is given by the expression
where [S02]Q is in ppm and []_ is in percent.
     It should be noted that eq. (4.2) was determined from the data
obtained under specific experimental  conditions.    As the light inten-
sity inside the tunnel was not varied during these experiments, eq. (4.2)
does not indicate the effect of the light intensity on nuclei  production.
However, Groblicki and Nebel [29] have found that the maximum  amount of
aerosols produced in a propylene-nitric oxide-sulfur dioxide-air mixture
was not affected by a change in the light intensity.  The light intensity
merely affected the time required to produce the maximum number of nuclei.
It is expected, therefore, that a change in the light intensity would
not alter eq. (4.2), but would change the irradiation time required to
produce the maximum number of nuclei.
     It would be of interest now to compare the present data with re-
sults of other investigations.  Results of direct measurements of the
number of nuclei formed in sulfur dioxide-air mixtures were reported in
references [13,14,40,48,49].  The conditions of these experiments are
listed in Table 4.1.  The present results can be compared with those
obtained by Renzetti and Doyle [49] and Kocmond et. al . [40].   The
number of nuclei obtained by Renzetti and Doyle for 0.24 ppm sulfur
                                                 45          -3
dioxide at 50% relative humidity ranged from 5x10  to 8x10  nuclei cm  .

-------
                             -46-
    io7cr
'S    6
  .  10
UJ
Li.
O
cr
LU    5
£  10
    ICT
       I  I  I  I I 11       I    I  I   I I  I I  I
                         FIT TO DATA
                         N = 3-33xl06[S02]0[*]|
i    i   i  i  i  i 111	i   i  i   i i i i  i
     0.01
               O.I
1.0   2.0
      Fig. 4.4.  Number of nuclei formed in sulfur dioxide-air mixtures.
               Correlation of the data for various initial  sulfur
               dioxide concentrations [SO^j  and initial  relative
               humidities fl 0             °

-------
              Table 4.1  Summary of Experimental Conditions Used by Different Investigators
Reference
Volume, m
                                          3
       Experimental Apparatus

Material                Light Source
Type
Cox and Penkett [13]
Cox [14]
Gerhard and Johnstone [24]
Kocmond et. al. [40]

Quon et. al . [48]
Renzetti and Doyle [49]
0.216
0.005
0.008
0.25
1.7
600
0.014
0.05
Aluminum Chamber
Pyrex Flask
Lucite Chamber
Teflon Bags
Steel Chamber
Saran Bags
Pyrex Flask
Sunlight
Mercury Lamp
Sun Lamp
Fluorescent Lamps
Fluorescent Lamps
Fluorescent Lamps
Mercury Lamps
S
S
S
S
S
S
F
S    Static type experiment
F    Flow type experiment

-------
                                -48-
For this sulfur dioxide concentration and relative humidity,  the number
of nuclei calculated from eq. (4.2) is 2x10  nuclei  cm" .   Hence,  the
value obtained in the present study lies within the  spread of Renzetti
and Doyle's data.  The amounts of nuclei measured by Kocmond  et.  al.  were
two to four times higher than the values obtained in this  investigation.
The reason for this is unclear.   The discrepancy between the  two re-
sults might possibly be due to the fact that the present data were ob-
tained from flow type experiments while Kocmond et.  al. performed  their
measurements in static type experiments.
     The results of Cox and Penkett [13], Quon et.  al.  [48],  and Cox  [14]
cannot be compared with the present data.  Cox and Penkett did not re-
port the humidity in their results; Quon et. al. performed measurements
before the maximum amounts of nuclei were produced.   The data obtained  by
Cox were obtained for much higher initial sulfur dioxide concentrations
(5 to 1000 ppm) than those used in the present study (0.2  to  1  ppm).
     Some indirect comparisons with other existing data can be made to
confirm the present results.  Gerhard and Johnstone  [24] performed static
type experiments in a 0.008 m  Lucite chamber irradiated by a sun  lamp,
and collected and measured the amount of sulfuric acid  formed in sulfur
dioxide-air mixtures.  If it is  assumed that most of the nuclei  formed
in the present study were composed of sulfuric acid, then  the number  of
nuclei should be proportional to the amount of sulfuric acid  collected
by Gerhard and Johnstone.  The variation of sulfuric acid  with the ini-
tial sulfur dioxide concentration, as reported in reference [24],  is
presented in Fig. 4.5.  Similar to the number of nuclei formed (see
eq. 4.2), the amount of sulfuric acid collected varies  linearly with  the
initial sulfur dioxide concentration.

-------
                        -49-
10
I
 E    |

 o»

 E
    0.5
o

(T
   o.2
    O.I
            o DATA

           — FIT TO DATA (Slope = I)
                 Irradiation Time I20min
  0.05
      2            5        10       20         50


    INITIAL SULFUR DIOXIDE CONCENTRATION, ppm
Fie. 4.3.  Formation of iulfuric acid in sulfur dioxide-air mixtures.

         Ucta fron, Gernard and Johnstone [24]

-------
                               -50-
     Renzetti and Doyle [49] measured the amount of light scattered from
a sulfur dioxide-air mixture undergoing photochemical  reaction.   Their
data are reproduced in Fig. 4.6.   Although there is considerable spread
in Renzetti and Doyle's data, the amount of light scattered (which is
proportional to the number of nuclei formed) appears to vary linearly
with the initial sulfur dioxide concentration, further supporting the
validity of eq. (4.2).
     The information available on the effect of the relative humidity on
the number of nuclei formed is conflicting.  Renzetti  and Doy'e  observed
that the amount of light scattered increased with relative humidity.
Cox [14] also reported that the number of nuclei formed increased with
humidity.  Katz and Gale [39] observed an increase in the conversion
rate of sulfur dioxide to sulfur trioxide (or su'lfuric acid) as  the
humidity was increased.  These results thus support the present  data
which show an increase in nucleation rate as well as the number  of nuclei
formed with relative humidity.  In contrast to these results, Gerhard
and Johnstone [24] found that the sulfur dioxide conversion rate did  not
change with relative humidity.  The results of Quon et. al. [48] indicated
that the number of nuclei formed was independent of relative humidity.
The data obtained by Kocmond et.  al. [40] did not show any dependence
of the conversion rate of sulfur dioxide on relative humidity.   However,
Kocmond et. al. found that the rate of change of the total volume of
the aerosols depended on relative humidity.  The effect of humidity on
the nuclei formation in sulfur dioxide-air mixtures would thus merit
further studies.

-------
                          -01-
  100
o:
<
cr
H
CD
o  10
a:
UJ
u

X
Ij
                           o DATA
                         	FIT TO DATA (Slope«I)
                                         I   1   I   I  I I I
    0.2                   1.0                            10
      INITIAL SULFUR DIOXIDE CONCENTRATION, ppm
   Fia.  4.6.  Amount of light scattered from sulfur dioxide-air
             mixtures.  Data from Renzetti and Doyle 1.49]

-------
                               -52-
                             CHAPTER V
      OZONE AND PARTICULATE" FORMATION IN CgH10-N02-AIR MIXTURES

     Photochemical  reactions in a cyclohexene-nitrogen dioxide-air mix-
ture depend on several  parameters, including a)  the initial  concentra-
tion of cyclohexene, b) the initial  concentration of nitrogen dioxide,
c) the ultraviolet  light intensity,  and d) the length of time the mix-
ture is irradiated.   The major goal  of this study was to evaluate the
first two of these  effects, namely,  the influence of the initial  con-
centrations of cyc'ohexene and nitrogen dioxide  on the amounts of ozone,
condensation nuclei  and aerosols formed.  Therefore, throughout the tests,
the light intensity was kept at the  constant value of k-,=6.33 ks   ,
                                        o   ]
and the flowrate was maintained at 46 cm  s  .  All the results dis-
cussed below apply  to these conditions.
     One might also expect the relative humidity to affect the formation
of ozone, nuclei and aerosols.  Bufalini and Altshuller [10] and  Stephens
and Price [58] have found, however,  that the amount of ozone formed was
not influenced by the relative humidity.  The effects of the humidity on
nuclei and aerosol  formation have not yet been established.   In the pre-
sent study, a relative humidity of 50% was used.  According  to the re-
sults of Bufalini and Altshuller and of Stepehns and Price,  the use of
a constant relative humidity would not affect the amount of  ozone formed.
The relative humidity might influence the nuclei and aerosol formation
but this effect was not explored here.
     The initial concentrations of cyclohexene and nitrogen  dioxide were
varied in a systematic manner.  Initially, five  concentrations of cyclo-
hexene (0.5, 1, 2,  3, and 5 ppm) and five concentrations of  nitrogen

-------
                               -53-
dioxide  (0.6, 1, 2, 3, and 5 ppm) were permutated to provide twenty-five
different mixture composition combinations.  As the measurements progressed,
it became evident that additional mixture compositions were needed to pro-
vide more information in certain areas of particular interest.  Hence,
measurements were also made with additional mixtures.  A complete set of
the experimental conditions and of the data obtained are given in Appen-
dix D.  The data from a typical series of tests are presented in Fig.
5.1.  The results in this figure show the amounts of ozone, nuclei and
aerosols formed along the irradiation tunnel.  Since the flowrate in the
tunnel was constant, the horizontal axis in Fig. 5.1 represents both the
distance along the tunnel and the length of time the mixture has been
irradiated.
     As can be seen from Fig. 5.1, the amounts of ozone, nuclei and aero-
sols first increase, reach a peak and then decrease along the tunnel.
Here, we shall be concerned with the "peaks" which represent the maximum
amounts of ozone, nuclei and aerosols formed.  Our interest in the maxi-
mum values is motivated by the general concern about the maximum amounts
of ozone, nuclei and aerosols formed in the atmosphere due to photochem-
ical reactions.   In all subsequent discussions, the term "amount" was
used to denote the peak values, which were determined by plotting all the
data in a manner similar to those shown in Fig. 5.1, and by estimating
the peaks from these plots.  The results thus obtained are tabulated in
Appendix E.

5.1  The Effects of Mixture Composition on Ozone Formation
     The typical behavior of ozone along the tunnel is illustrated in
Fi;. 5.1.  Although this figure represents only one set of data, similar

-------
             TIME, mmult
        10   ZO   30   40
90
      TIME, minute
IP   20   30   4O   50   60  |08
   O     2     4      6    8

     DISTANCE ALONG TUNNEL.  m
                    O      2      4     6     6

                         DISTANCE ALONG TUNNEL, m
                         (O
Fig. b.l.  Typical  variations of ozone, nuclei, and  aerosol  contents along the  tunnel  in
           cyclohexene-nitrogen dioxide-air mixtures.   Initial nitrogen dioxide concentration:
           3 ppm.   Relative humidity:  50%.  Top  axis  represents irradiation  time

-------
                                -55-

trends were observed with other initial concentrations of cyclohexene and
nitrogen idoxide.  For all mixture combinations, the amount of ozone along
the tunnel first increased, reached a peak and then decreased.  In order
to describe these results quantitatively, it would be necessary to know
in detail the reaction mechanism in the mixture and the corresponding re-
action equations.  To date, many different sets of reaction equations have
been proposed [15,19,32,46,61,62], but the validity and accuracy of the
various sets are still of considerable debate.  Nevertheless, it is
possible to explain, at least qualitatively, the observed behavior of the
ozone formation by the relatively simple kinetic scheme introduced by
Eschenroeder and Martinez [19].  These investigators  suggested twelve
reactions for describing the photochemical processes occurring in hydro-
carbon-nitrogen oxides-air mixtures.   The reaction equations proposed by
Eschenroeder and Martinez are listed  in Table 5.1 (eqs. 5.1-5.12).
     According to eqs. (5.1) and (5.2), ozone and nitric oxide are formed
by the photodissociation of nitrogen  dioxide.  The ozone reacts with the
nitric oxide and the hydrocarbon (eqs. 5.3 and 5.10).  These reactions
tend to deplete the ozone.  However,  simultaneously, the nitric oxide
reacts with various free radicals(denoted by R02, eq. 5.11), and this re-
action proceeds at a much faster rate than the reactions of ozone with
the nitric oxide and the hydrocarbon.  The net result is an excess amount
of ozone in the system; hence, the observed increase in the ozone con-
tent in the first part of the tunnel.  As the photochemical reaction pro-
gresses, the nitrogen dioxide is depleted by reacting with the free
radicals (eqs. 5.4 and 5.12).  As the amount of nitrogen dioxide dim-
inishes, so does the amount of ozone  formed resulting in an eventual de-
crease in the ozone content along the tunnel.

-------
                          -56-

     Table 5.1   The Reaction Mechanism Proposed by
                 Eschenroeder and Martinez [19]
OH
1MU- T \\f 	 " HVJ T \J
+ n a. M > n a. M
U_ f 1"! - U— T I'l
_ + WU 	 •*•••• *• NU- •+ U_
k4
\tn i jjri ^ t~ ouMn
wu T wu_ ,. ;i ^ zin«u_
?. H20 2
HMO _L v> ±y - — — — *• ni i • 4- MO
iiiiU— T n " • uji T ri\j
k?
:. J. tin -k M -'- - *• HMO 4- M
ka
ko
nu. -L. 11^" 7 ^- •u /nrv ^
Ull- T Ho ^ Dp^«*J_;
• T no ^ u»j^inj_y
R02 + NO 	 ^> N02 + d(OH)
DA -L. MA 	 _Jl£_» M^T>AM^
^.A;
(5.2)
(5.3)
(5.4)
(5.5)
(5.6)
(5.7)
(5.8)
(5.9)
(5.10)
(5.11)
fc 10\

-------
                               -57-
     It is now possible to evaluate the maximum amount (referred to sim-
ply as the "amount") of ozone produced for different initial  concentra-
tions of cyclohexene ([C5H]Q]0) and nitrogen dioxide ([N02]0)«  The amount
of ozone produced as a function of the initial concentrations of cyclo-
hexene and nitrogen dioxide is shown in Figs. 5.2 and 5.3.   For a given
initial nitrogen dioxide concentration, the amount of ozone first in-
creased with increasing cyclohexene concentration.  An increase in the
cyclohexene concentration beyond a certain value resulted in  a decrease
in the amount of ozone (Fig. 5.2).  Similar behavior was observed when
the initial cyclohexene concentration was kept constant and the concen-
tration of nitrogen dioxide was varied (Fig. 5.3).  Such an interrelation
between the amount of ozone formed and the initial concentrations of the
hydrocarbon and the nitrogen oxides has also been noticed by  other in-
vestigators [2,5,6,7,25,41,50], and plausible qualitative explanations
of these phenomena have been offered in references [25,50,55].   A brief
explanation of these phenomena may be given by referring again to eqs.
(5.1)-(5.12).  First, we consider the case when the initial nitrogen
dioxide concentration is kept constant and the initial  concentration of
cyclohexene is varied (Fig. 5.2).  As discussed above,  the  photodissocia-
tion of nitrogen dioxide yields ozone and nitric oxide (eqs.  5.1 and 5.2).
Ozone is depleted by reactions with the nitric oxide and the  hydrocarbon
(eqs. 5.3 and 5.10).  In addition to reacting with ozone, the nitric
oxide also reacts with the free radicals (eq. 5.11), and this reaction
proceeds faster than the reactions depleting ozone.  Thus,  an increase in
the hydrocarbon concentration results in a decrease in the  amount of nit-
ric oxide available for the depletion of ozone.  Consequently,  the amount

-------
  5.0
E
a.
o.


._• 1.0
UJ
o
o
o
IM
O
                     I   I
I  I I  I
I   I  1  ITT T
                                            Initial N02

                                            Concentration


                                               5ppm
   0.1

    0.3




  Fig. 5.
         I   I  I  ! I  I I
                     I    I  I  I  I  I I
                   I    I  I  I I  I I
       1.0                         10                         100

  INITIAL CYCLOHEXENE CONCENTRATION (C6H|0]0 ,  ppm


Ozone content in cyclohexene-nitrogen dioxide-air mixtures.  Relative

humidity:  50%.  Light intensity  Iq = 6.33 ks-1

-------
                             -59-
  5.0
i
   1.0
o
o

UJ
z
o
N
O
  O.I
             I  I  I  I
T    I   I  I  I  i I
                                        2ppm


                                     0 ^      Initial C6HIO

                                             Concentration
         i   i  i  i  i i
    "0               1.0                             10

         INITIAL NITROGEN DIOXIDE CONC. [N02]Q  , ppm
                         20
 Fig.  b.3.  Ozone content in cyclohexene-nitrogen dioxide-air mixtures.
           Relative humidity:   bOi.  Light  intensity k] = 6.33 ks'l

-------
                               -60-
of ozone formed increases with initial  increase in the hydrocarbon concen-
tration.  However, once there is a sufficient amount of hydrocarbon pre-
sent in the mixture to react with all  the available nitric oxide,  then
any additional hydrocarbon will react  with ozone.   Under these condi-
tions, an increase in the hydrocarbon  concentration results in a decrease
in the ozone content.
     The formation of ozone can be explained similarly when the initial
concentration of the hydrocarbon is kept constant while the initial con-
centration of nitrogen dioxide is varied (Fig. 5.3).  Ar the nitrogen
dioxide concentration is increased, both the amounts of ozone and nitric
oxide are Increased (eqs. 5.1 and 5.2).   Because the reaction of nitric
oxide with- the free radicals (eq. 5.11)  proceeds at a faster rate than
the reactions depleting ozone (eqs. 5.3 and 5.10), ozone accumulates in
the mixture.  This explains the increase in the ozone content with the
initial increase in the nitrogen dioxide concentration.  When the ini-
tial nitrogen dioxide concentration is increased tc the point where all
the hydrocarbon present in the mixture reacts with all the nitric oxide
produced by the nitrogen dioxide, then any further increase in the nit-
rogen dioxide concentration provides some excess nitric oxide to react
with the ozone.  More ozone is depleted due to this reaction than is
formed due to the increase in the nitrogen dioxide concentration.   Hence,
the ozone content is reduced when the  nitrogen dioxide concentration is
increased.
     The results presented in Fig. 5.2 and 5.3 would only apply to the
formation of ozone in cyclohexene-nitrogen dioxide-air mixtures.  An
attempt is now made to generalize these results so as to gain additional
information regarding the ozone production by other types of hydrocarbons.

-------
                                -61-
 It  is  recognized  [15,46,55]  that,  in a  reaction described by N reaction
 equations,  the  amount of  ozone  formed depends on:  a) the initial con-
 centration  of nitrogen  dioxide  [N02] ,  b) the initial concentration of
 the hydrocarbon [HC]  ,  and c) the  rate  constants k.(i=l to N), i.e.

                fo3>   f(fN02!0, rHC]0, kj_  , k2 .....kj, )         (5.13)

 The terms in eq.  (5.13) may  be  nondimensionalized, and eq. (5.13) may be
 expressed as
                                     *    *         *
                                                                  (5.14)
where  k-j ,  k2 ...... k^_,  represent N-l dimensionless groups formed from the
different  rate constants, and  [63]  is some characteristic ozone concen-
tration  in the mixture.  To select  [03lc, it is recalled that in a nitro-
gen  dioxide-air mixture  the amount  of ozone formed is (Section III)

                                                                  (3.10)
 In most  practical situations,  [Os]-.. is negligible compared to [N02]0>
 in which case eq. (3.10) becomes
                l°3JPss  -   ^l"Vo                        (5.15)

 Representing  [03]  by  this expression, eq.  (5.14) can be rewritten as

                                                  	k*_<  )   (5-16)
              V  O
 For  a  hydrocarbon^nitrogen dioxide-air mixture which contains just one
 type of  hydrocarbon,  the  values of the k^s are constants, and
\lIT  [N02L  depends only on  the  ratio of the initial concentrations of
 the  hydrocarbon  and the  nitrogen  dioxide, i.e.,

-------
                               -62-
                     Coj
               /JL
               V   k3
     Equation (5.17) relates the amount of ozone  formed  to  the  initial
nitrogen dioxide concentration and the initial  concentration  ratio.
     In many previous experiments, instead of nitrogen dioxide,  nitric
oxide was used in the initial  gas-air mixture.  With  nitric oxide  in  the
initial gas-air mixture, the production of ozone  occurred at  a  later  time
than with nitrogen dioxide,  but the maximum amount  of oznne generated was
found to be the sar.e for both  nitric oxide and nitrogen  dioxide.   It  is
noted, further, that some investigators reported  measurements of "oxidant"
instead of ozone.  The term oxidant designates certain pollutants  formed
during the photochemical reaction.  These pollutants  are so named  be-
cause they will oxidize some specific chemical  reagents  (e.g. a  solution
of neutral-phosphate-buffered  potassium iodide) which are not readily
oxidized by oxygen [69].  Although the oxidant measured  includes other
chemical compounds as well as  ozone (e.g. peroxyacyl  nitrates),  ozone
is generally the largest amount (by volume) present in the  measured oxi-
dant, and the data reported for ozone and oxidant were usually  very sim-
ilar [8,69].
     For the reasons outlined  above, results obtained with  nitrogen
dioxide or with nitric oxide,  and results reported  in terms of  ozone  or
oxidant may be discussed together, and compared directly.
     According to eq. (5.17),  for a given type of hydrocarbon and  fixed
initial concentrations of the  hydrocarbon and the nitrogen  dioxide, the
amount of ozone varies directly with the square root  of  k,
                                                                  (5-18)

-------
                               -63-
It is recalled from Section III that k, is proportional  to the light in-
tensity (eq. 3.1).  Hence, the ozone content varies with the square root
of the light intensity.  Harton and Bolze [31], Altshuller et. al.  [5],
and Groblicki and Nebel [29] measured the effect of light intensity on
the amount of ozone (or oxidant) formed in various hydrocarbon-nitrogen
oxide-air mixtures.  The results reported by these investigators also
confirm that the amount of ozone (or oxidant) varies with the square root
of the light intensity.  The computer solution obtained  by Niki  et. al.
[46] for the photochemical reactions in a propylene-nitrogen oxide-air
mixture also confirms the validity of eq. (5.18).
     Equation (5.17) also shows that, for a given type of hydrocarbon,
                                                         Hrj
the data points should fall on a common curve on a [Q^]/ / jj—• [N02]Q
versus [HCL/[N09]  plot.  Such a plot for the present data is given in
           0    ' o
Fig. 5.4 .   it is seen that, within experimental error,  the data follows
a single curve.  This lends support to the validity of eq.  (5.17).   In
order to test further the validity of this equation, the data reported
by other investigators have been plotted in a graph similar to Fig. 5.4.
These results, presented in Fig. 5.5, were for eight different types of
hydrocarbons.  The data given in Fig. 5.5 were all obtained in chambers
irradiated artificially by either fluorescent lights [2,5,6,41,58]  or by
mercury lamps [50].  The measurements were made either in a static  type
chamber [2,6,41,50] or in a flow type tube [5,58].  Scott [50] irradiated
3-methylheptane-nitrogen dioxide-air mixtures in a long  path infrared ab-
sorption cell.  Altshuller and Cohen [2] tested ethylene-nitrie oxide-
air mixtures in Teflon bags.  Altshuller et.  al. [6] and Kopczynski et.
al. [41] used the same aluminum irradiation chamber and  tested n-butane-
nitric oxide-air mixtures [6] and aldehyde-nitric oxide-air mixtures [41].

-------
    10
     8
-I 1*1
I    I   I  I  I  I I
i    i   i  i  i  i i
I     II   I  I  I  I I  |
                                            o DATA

                                               FIT TO DATA
                                        i    ill
                             I      I   i
     0.2                  1.0                           10

          INITIAL CONCENTRATION  RATIO,  "[ C6Hlo]o/[ N0z]o

   Fig. b.4.  Normalized ozone content in cyclohexene-nitrogon dioxide-air mixtures.
            Relative humidity: 50%.  Light  intensity kj  = 6.33 ks'1
                                                                50

-------
                             -6b-
       2

       0

      10


^    5
 X
O
-I 10
 H«
 0

10
       0

      10
       O.I
Fig. b.b.
      fo I
                     (a)
                           (b)
                           (c)
                           (d)
         1.0
10
100
                                      10
                                0

                                10
                0
               10
                  (e)
                                   (g)
10
100
            INITIAL CONCENTRATION RATIO, [HC]0/tN°x]o
     Normalized  ozone content in hydrocarbon-nitrogen oxides-
     air mixtures,   (a) Ethylene, Altshuller and  Cohen  [2].
     (b) 3-methylheptane, Scott [50].  (c)  Cis-2-butene,
     Stephens  and  Price [58].  (d) Propionaldehyde,  Kopczynski,
     et. al.  [41].   (e)  Acetaldehyde, Kopczynski, et.  al. [41].
     (f)  Propylene, Altshuller et. al. [5].  (g) Formaldehyde,
     Kopczynski, et. al. [41].  (h) N-butane, Altshuller  et. al.
     [6]

-------
                               -66-
Flow type experiments were performed by Altshuller et.  al.  [5]  for propy-
lene-nitric oxide-air mixtures in an aluminum chamber,  and  by Stephens
and Price [58] for cis-2-butene-nitric oxide-air mixtures  in a  Pyrex
tube.  Generally, fewer data points were taken in the previous  experi-
ments than in the present study.   Nevertheless,  the previous data, sum-
marized in Fig. 5.5 also indicate that the data  can be  correlated ac-
cording to eq. (5.17).
     In addition to the investigations quoted above, ozone  (or  oxidant)
measurements were reported in [7,10,25,31,50,51].  The  results  of these
investigations [7/0,25,31,50,51] could not be included in  Fig.  5.5 be-
cause either there was not enough information given to  plot the graph
[10,31], or the amount of ozone (or oxidant) measured was  not given
directly [7,25,50,51].  The amount of ozone (or oxidant) was measured by
Bufalini and Altshuller [10] and  by Harton and Bolze [31].   However,
Bufalini and Altshuller reported  the nitrogen dioxide dosage instead of
the initial concentration of nitrogen dioxide.  Harton  and  Bolze did not
report the maximum amount of the  oxidant generated.  The amount of ozone
(or oxidant) was reported indirectly by Altshuller et.  al.  [7],  Glasson
and Tuesday [25], Haagen-Smit [50], and by Romanovsky et.  al. [51].
Altshuller et. al. reported the oxidant dosage,  Glasson and Tuesday the
initial formation rate of ozone.   Haagen-Smit measured  the  crack depths
of rubber strips.  Romanovsky et. al. presented  their results in the
form of ozone contour maps.
     An attempt can now be made to estimate  a)  the maximum possible
amount of ozone formed in a hydrocarbon-nitrogen dioxide-air mixture,
and b) the concentration ratio at which this maximum occurs. Let the

-------
                               -67-
                      ki
                      -   [N02]Q  versus the concentration ratio curve be
maximum of the
M, and the concentration ratio corresponding to this point be R (see
Fig. 5.4 and 5.5), i.e.
                         Co.]
                             'max
                  M =
and
                  R  s.
                          2J0
                                                                  (5.19)
                                           at  f°31
                                                                  (5.20)
In principle, the values of M and R can be obtained from the appropriate
reaction equations (e.g. eqs. 5.1-5.12).  In practice, this is a formid-
able, if not an impossible task owing to the complexity of the equations.
The data summarized here, however, provide some information about M and
R.  Equations (5.19) and (5.20) may be rearranged in the form

                                                                  (5.21)
                                                                  (5-22)
                          M
     The maximum amounts of ozone formed ([03!,,,) in Figs. 5.2 and 5.3
                                              met A
are represented by the envelopes of the ozone versus cyclohexene and the
ozone versus nitrogen dioxide curves (dotted lines in Figs. 5.2 and 5.3).
From Figs. 5.2 and 5.3, the equations of these curves (envelopes) may be
determined, and are
                                                                  (5.23)
                                                                  (5.24)

-------
                                -68-

                                                                     1/2
 Comparing eqs.  (5.21) and (5.23), and noting that/k^/k-j'  =  0.122  ppm
(Section  III),   the following expression is obtained
                  M-5.3/T                                     (5.25)

     Tne above  relationship between M and R was derived  entirely from  the
 present data using cyclohexene.  In order to estimate the  relationship
 between M and R for other types of hydrocarbons, M was plotted  as  a  func-
 tion of R for the nine different hydrocarbons for which  values  of  M  and
 R  could be estimated (Figs. 5.4 and 5.5).  The results,  given in Fig.  5.6,
 show that in all c
-------
     70
  x
O
     10
>
          I   i  i  i i
                                             I   I  TIT
           I   I  I  I I
                                      I    I   I  I  I  I  I
                                                                                             I
                                                                                             er>
       0.4
                     I.O
                                R =
                                      [HC](
10
30
Fig. 5.6.   Maximum amount of ozone  M  as  a  function of the concentration ratio R
           at which M occurs.   0 cyclohexene,  present study;   o ethylene, Altshullpr
           and Cohen [2];  a  3-methylheptane,  Scott [50];  t> Cis-2-butene, Stephens
           and Price [58];   7  propionaldehyde, Kopczynski et. al. [41];  <  acetaldehyde,
           Kopczynski  et. al.  [41];   *   propylene, Altshuller et. al. [5]; A  formaldehyde,
           Kopczynski  et. al.  [41];   O   n-butane, Altshuller et. al. f6l

-------
                           -70-
CM
  Fig. 5.7.
INITIAL  HYDROCARBON CONG. [HC]O , ppm

 Maximum amount of ozone as a function of the initial
 hydrocarbon concentration. 	 present  result;
  «» ethylene, Altshuller and Cohen [2];  
-------
                                -71-
were also included.  These atmospheric measurements were reported by the
U.S. Department of Health, Education and Welfare [70] and gave the oxi-
dant content as a function of non-methane hydrocarbon concentration in
ppm C.  In order to include the atmospheric data in Fig. 5.7, an average
carbon number of five was assumed for these non-methane hydrocarbon mix-
tures and k, was taken to be 8 ks   (see Table 3.1).  It is interesting
to note that eq. (5.27), which was based on the data obtained from mix-
tures containing only one type of hydrocarbon, also yields reasonable
results for mixtures containing several  types of hydrocarbons such as
those present in the atmosphere.
     The maximum amount of ozone that can be formed for a- given concen-
tration of rrftrogen dioxide depends on M (eq. 5.22).  Since it is easier
to determine.R than M, eq. (5.22) is written as (see eqs. 5.22 and 5.25)

                           8.13 /k/TyrN0l                     (5.28)
or
                 fo.l
                    'max
                    ,r          w   J2'0                           (5.29)

Therefore, the maximum amount of ozone depends on R which is different
for different types of hydrocarbons.  However, the ratio [03]max//k-|R
is independent of the type of the hydrocarbon and depends only on the
initial concentration of the nitrogen dioxide.  Equation (5.29) is pre-
sented in Fig. 5.8 together with the available data deduced from Figs.
5.4 and 5.5.  It is seen that there is reasonable agreement between the
curve given by eq. (5.29) and the existing data.
     The foregoing discussion illustrated that in order to determine the
maximum amounts of ozone produced in a given hydrocarbon-nitrogen dioxide-

-------
                          -72-

-------
                                -73-
 air mixture,  the  values of~M and R for that hydrocarbon are needed.  It
 would  be desirable, therefore, to be able to estimate the values of M and
 R  for  different types of hydrocarbons.  The values of M and R depend on
 all  the rate  constants in a complex manner.  An inspection of the reac-
 tion equations (eqs. 5.1 to 5.12), indicated that the rate constants which
 would  most likely influence M and R were kg, kg, and k1Q.  In addition,
 M  and  R might also be related to the conversion rate, kNQ, of nitric oxide
 to nitrogen dioxide [1,3,4,26,33].  Therefore, an attempt was made to
 correlate R with each of these rate constants.  To facilitate this cor-
 relation, the suggestion of Glasson and Tuesday [26] was followed and a
                                                          A
 dimensionless relative rate constant called "reactivity", k, is defined
as
                     *
                     k =
                                                                  (5.30)
where ^CHK 1S tne rate constant of propylene.  The results, presented
in Table 5.2 show only weak correlations between R and the reactivities
A   A        A                                                    A
kg, k,Q, and kNQ, but indicate a strong correlation between R and kg.
                               A
Therefore, a graph of R versus kg was plotted (Fig. 5.9), and the follow-
                                           A
ing approximate relationship between R and kg was obtained

                    R =    FIT                                   (5-31)
                           / K9
The value of R can be found from this expression for any given hydrocar-
bon.  This R value may then be used in eqs. (5.26) and (5.28) to calcu-
late the maximum amount of ozone formed.
     Finally, in Fig. 5.10, an ozone contour map is presented on the
basis of the cyclohexene data.  The general features of this map were
similar to those reported for  propylene-nitrie oxide-air mixtures [51]

-------
                               -74-
  Table 5.2  Reactivities of Different Hydrocarbons (k = k/kC3Hg)
Hydrocarbon
          Initial
Reactivity

n-Butane
Formal dehyde
Propylene
Acetal dehyde
Propional dehyde
Cycl ohexene
Cis-2-Butene
3-Methyl heptane
Ethyl ene
Concentration
Ratio, Ra
10.5
8
5
4
3.5
2.4
2.1
2.0
1.3, 17f
A
k8
0.008C
0.05C
1.0
0.15C
0.2C
	
3.6e
0.08e
0.2C
— - - -^
A
kg
0.24C
0.9C
1.0
0.9C
1.8C
	
	
	
0.1C
/y . A
I, b i,
K10 KNO
	 0.2C
.... 0.7C
1.0 1.0
— - 0.7C
	 2.0C
5.2 1.0d
2.0 2.0d
— - 0.27d
0.33 0.3C
a
b
c
d
e
f
From Figs. 5.4 and 5.5
From Table 46 reference [43]
From reference [46]
From reference [26]
From reference [54]
From reference [25]

-------
          30
           10
    o
 x
O
                 I   I  I  I I
          1.0
           0.4
                                  T	1—J   I  I  I  I
                        FIT TO DATA
                        R=5.l /
        I   I  I  I I
_LL
                                                                                                 on
                                                                                                 t
                 O.I
   1.0
4.0
                 REACTIVITY  BASED ON  HYDROXYL RATE  CONSTANT,
                                                                            9
Fig.  5.9.  Variation ij the concentration ratio R (at which the ozone content  is maximum) with the
          reactivity kg.  <* ethylene, Altshuller and Cohen [2];  O n-butane, Altshuller et. al. [6];
           A formaldehyde, Kopczynski et.  al. [41]; x propylene, Altshuller  et. al.  [5];
           « acetaldehyde, Kopczynski et.  al. [41]; 7 propionaldehyde, Kopczynski, et. al. [41]

-------
 E

 S 10

-------
                               -77-
and for mixtures of non-methane hydrocarbons in the atmosphere [70] (Figs,
5.10 and 5.11).  There was one notable difference between the maps ob-
tained with one single hydrocarbon (e.g. cyclohexene or propylene) and
with mixtures of hydrocarbons.  When only a single hydrocarbon was used,
each curve had reasonably well defined "knolls", which could be connected
by a straight line.  For the atmospheric data,  a straight line could not
be drawn through the knolls either because the  measurements were inac-
curate, or because the many different types of  hydrocarbons in the at-
mosphere distorted the slope of the curve.

5.2  The Effects of Mixture Composition on Particulate Formation
     The nuclei and the aerosol contents in the tunnel are given in
Fig. 5.1 for s typical set of initial cyclohexene and nitrogen dioxide
concentrations.  For other initial concentrations of nitrogen dioxide and
cyclohexene, the number distribution of the nuclei and the aerosols along
the tunnel exhibited similar trends, i.e. the numbers first increased,
reached maximum values and then decreased slightly.  Thus, in general,
the nuclei and the aerosol formation followed a pattern similar to the
process of ozone formation.  However, the nuclei and aerosol  contents
changed at slower rates than the ozone content; in fact, the number of
nuclei and aerosols varied only slightly past the third sampling station
which was 1.75 m from the entrance of the tunnel.  Since the nuclei and
the aerosol counts at the tunnel entrance were  nearly zero, most of the
nuclei and the aerosols were formed in the region between the entrance
and the third sampling station. According to the data in Fig. 5.1, ozone
was also generated at the fastest rate in this  region.

-------
E
Q.
Q.
o
o
en
UJ
o
X
o
UJ
o
o

-------
                              -79-
     Im'tially, nuclei were produced by the interaction of some of the
gaseous reactants which were the byproducts of the photochemical  reac-
tions [18,49,53,63].  The nuclei then interacted both with gaseous mole-
cules and with each other, resulting in an increase in the size of the
nuclei.  Eventually, the nuclei reached a size which became detectable
on the aerosol counter.  These "nuclei" were then designated as aerosols.
The number of nuclei and aerosols decreased along the tunnel due to co-
agulation, diffusion, and deposition.  Since the number of nuclei and
aerosols remained reasonably constant along the tunnel, it appeared that
in the present system the formation and the depletion of nuclei and aero-
sols proceeded at nearly the same rates.  Here, we were not concerned
with the details of these two processes but were only interested in the
maximum number of nuclei and aerosols generated for each combination of
nitrogen dioxide and cyclohexene concentrations.  The maximum numbers of
nuclei  and aerosols deduced from all  the data are listed in Appendix E.
     The maximum number (referred to simply as the "number") of nuclei
produced for different initial  concentrations of cyclohexene and nitro-
gen dioxide is given in Fig.  5.12.   In comparing this figure with Fig.
5.2, it can be seen that the variation in the nuclei content with the
initial  concentrations of nitrogen dioxide and cyclohexene is remarkably
similar to the variation of the ozone content with the initial  concen-
trations of nitrogen dioxide and cyclohexene.  For example, similar to
ozone,  for a given initial nitrogen dioxide concentration, the number
of nuclei produced first increased with the cyclohexene concentration,
reached a peak and then decreased with further increase in the initial
cyclohexene concentration.  Moreover, for a given initial  nitrogen diox-
ide concentration, the amounts of the ozone and the nuclei produced

-------
   I04
10
'E
 u
LJ
_l
O
or
u
m
                  [N]mox= 540
                                                     I     T
                                                                  I  IT
                                                              lOppm
                                                              5ppm
                                              O.Sppm
                    Initial  N02  Concentration
         i  i i  i  i i
I
                               i    i  i  i  i i  i
I   i   i  i  i i i
    0.3            I                         10

      INITIAL CYCLOHEXENE CONCENTRATION
                                                       , ppm
 Fig.  5.12.  Nuclei content in cyclohexene-nitrogen dioxide-air mixtures
            humidity:  50%.  Light intensity kj = 6.33ks-l
                                                                      100
                                                                     Relative

-------
                                -81-
reached their respective maximum values at approximately the same initial
cyclohexene concentration (e.g. for [N02] =2 ppm, [03]    occurs at
                                         U            Illu A
[C6H10]0=4'6 ppm and the maximum number of nuclei [N]max at [CsHio]0 =
4.4 ppm).
     The similarities in the results of the nuclei and the ozone become
even more evident if the number of nuclei is plotted against the ini-
tial concentration ratio ([C6Hio]0/[N02]0) (Fig. 5.13).  It is seen that
the maximum number of nuclei always occurs at a concentration ratio of
about 2.4, which is the same ratio at which the ozone content is a maxi-
mum in the mixture (Fig. 5.4).
     The maximum number of nuclei produced for a given initial cyclohex-
ene concentration is represented by the dotted line in Fig. 5.12.  The
equation of this line is
                                Hlo                               (5.32)
                                                      -3    -1/2
A, is a  numerical constant having the value of 540 cm   ppm  '  .   Again,
similar to ozone, a) the maximum number of nuclei may be expressed inde-
pendently of the initial nitrogen dioxide concentration, and b)  the maxi-
mum number of nuclei varies with the square root of the initial  cyclo-
hexene concentration.  The maximum number of nuclei produced for a given
initial nitrogen dioxide concentration may also be evaluated from the
data (Fig. 5.12).  The results show that
                                  o                               (5.33)
                 -3    -1/2
where A~ = 835 cm   ppnf   .   As in the case of ozone (see eq.  5.24),
the maximum number of nuclei  varies directly with the square root of the
initial nitrogen dioxide concentration.

-------
       ICT
   10
   I
    E
    u
    UJ
    or
    UJ
    O>
       10'
       i i  i
                               Initial

                               Concentration


                                10 ppm
                                                 5ppm
I   I  I  i I  I
I
i   i  i  i i  i
                                                                                                 CO
                                                                                                 .N5
                                                                                                 I
        0.3             I                          10

             INITIAL CONCENTRATION RATIO, [C6H,0]0/ [N02]0
                                                               100
Fig.  5.13.  Nuclei content as  a  function of the  initial concentration  ratio in c/clohexene-
           nitrogen dioxide-air mixtures.   Relative humidity:   50%.   Light intensity
           k  = 6.33ks-l

-------
                               -83-
     It should be noted that eqs. (5.32) and (5.33) were determined from
the data obtained under specific experimental  conditions, i.e.,  k^=6.33
ks   and ^ =50%.  The constants A, and Ap, therefore, may depend on the
light intensity and the relative humidity.  Data are as yet unavailable
which would indicate the dependence of A-,  and  Ap on the light intensity
and the relative humidity.
     It would be of interest now to assess whether or not eqs.  (5.32)  and
(5.33) could be applied to other types of hydrocarbons.  In order to ac-
complish these goals, data such as reported here, but obtained with
different types of hydrocarbons, would be needed.  Unfortunately, such
data are unavailable at the present time.   Only in a few investigations
were the nuclei number counts measured, and most of the existing data
were obtained under selected conditions.  For example, in most previous
experiments, the initial hydrocarbon and nitrogen oxide concentrations
were fixed and only the type of the hydrocarbon used in the mixture was
varied.  Consequently, the results of these tests do not give a  range  of
concentration ratios but apply only to one specific value.  Furthermore,
in most of the past tests the number of nuclei was not determined dir-
ectly; instead, the relative number concentration of the nuclei  (or
aerosols) was evaluated by measuring the light scattered from a  light
beam passing through the mixture.  Hence,  we were unable to fit  similar
expressions like eqs. (5.32) and (5.33) to the data obtained with other
types of hydrocarbons.
     There appears to be only one nuclei measurement which can be com-
pared directly with the present results.  Stevenson et. al. [59] per-
formed flow type experiments in a 0.125 m3 Pyrex tube which was  irradiated

-------
                               -84-
by fluorescent lamps.  They reported that for a mixture containing 2 ppm
of nitrogen dioxide and 3.5 ppm of cyclohexene, the number of nuclei pro-
duced was about 700 cm" .   They did not report either the light intensity
or the relative humidity inside their Pyrex tube.   However, for the cor-
responding mixture composition, the present data indicate a nuclei count
          •3
of 800 cm   (Fig.  5.12) which is in good agreement with the value obtained
by Stevenson, et.  al.
     The number of aerosols as a function of the initial  cyclohexene and
nitrogen dioxide concentrations is shown in Fig. 5.14.   For a given ini-
tial  nitrogen dioxide concentration, the aerosol content at first in-
creased sharply with increasing concentrations of cyclohexene.   At about
                   •3
400 aerosols per cm , the aerosol  content gradually levelled off with
further increase in the initial cyclohexene concentration.  Figure 5.14
does  not provide much information  on the number of aerosols formed.  When
the aerosol content is plotted against the initial concentration ratio
([CsHio] /[N02]Q), the data points all collapse onto one common curve
(Fig. 5.15).  In contrast to the nuclei (Fig. 5.13) and ozone (Fig. 5.4)
formation, the number of aerosols  produced appear to depend only on the
initial concentration ratio.  The  effects of the initial  nitrogen dioxide
concentration on the aerosol content seem to be small.   The apparent lack
of influence of the nitrogen dioxide concentration on the aerosol pro-
duction may be due to the fact that the fluctuations in the aerosol mea-
surements (-30%) may be larger than the effects of the changes  in the
nitrogen dioxide concentrations on the number of aerosols produced.
     It is interesting to note that the aerosol content becomes appre-
ciable only when the initial concentration ratio ([C6HioL/[N02]0) is
larger than about 2.4, the same value at which the amounts of ozone and

-------
    1000
  to
  I
  u
  en
  O
  
  o
      100
  oc
  §
Initial N02
Concentration  0.6 ppm
                           1ppm
       10
        0.3
 i   i  i i i i  I
I
                                    i   i
                                                    I   I   i  Mi1-
                                                                                      c:
                                                                                      in
                                                                                       i
                                                        lOppm
                                                               l    i  i  i i  i i
           1.0                         10
       INITIAL CYCUOHEXENE CONCENTRATION, ppm
                                             100
Fig.  5.14.  Aerosol  content in cyclohexene-nitrogen dioxide-air mixtures,
           humidity: 50%.  Light  intensity k]  = 6.33 ks-1
                                                          Relative

-------
                              -86-
   10"
10
 u


 3
o
o:
UJ
<
u.
o
or
UJ
CD
IO
    10
                                                     I    I   T_
                                  Initial N02

                                  Concentration  ppm


                                  x  0.6
                                  o  i

                                  *  2

                                  a  3
                                  A  5

                                  010
               o


               1
                                                 J	I
                                     10                      60

        INITIAL CONCENTRATION  RATIO,  [C6H|0]0/[N02]0
   Fig.  5.15.  Aerosol  content as  a function of the initial concentration

              ratio in cyclohexene-nitrogen dioxide-air mixtures.

              Relative humidity:  50%.  Light intensity kj = 6.33 ks'1

-------
                               -87-
nuclei reach their maximums (see Figs. 5.4 and 5.13).  When the ratio is
below 2.4, only about 50 aerosols cm   are generated in the mixture.
This is a small number compared to the nearly 600 aerosols cm   generated
when the concentration ratio is ~10.
     It would be desirable now to compare the present aerosol data with
the results of other investigators.  Direct comparison between the re-
sults of the present and previous investigations is not possible because
different light intensities and different relative humidities were used
in different experiments.  The effects of the light intensity and the
relative humidity on aerosol formation in hydrocarbon-nitrogen oxides-
air mixtures have not been studied in the past.   Some indications of  these
two effects may be deduced from the aerosol studies performed with hydro-
carbon-nitrogen oxides-sulfur dioxide-air mixtures.  Groblicki and Nebel
[29] investigated the effect of the light intensity on aerosol formation
in propylene-nitric oxide-sulfur dioxide-air mixtures.  They found that
the light intensity had no effect on the maximum amount of aerosols pro-
duced.  The light intensity merely affected the  time required to produce
the maximum amount of aerosols.
     The effect of the relative humidity on the  amount of aerosols gen-
erated in hydrocarbon-nitrogen oxides-sulfur dioxide-air mixtures has been
studied by Groblicki and Nebel [29], Harkins and Nicksic [30], Prager et.
al. [47], Schuck et. al. [53], and by Wilson and Levy [64].  Some of  the
results showed an increase in the amount of aerosols with increasing
humidity [53], while some showed a decrease [29,30,64].  Furthermore,
some of the data showed no apparent relationship between the amount of
aerosols generated and the relative humidity [47].

-------
                              -88-
     Comparisons between the present and existing aerosol  data are not
possible not only because the light intensities and the relative humi-
dities were different in the tests, but also because the number count of
the aerosols was not reported in most previous investigations.
     In most of the past measurements, the light scattering method was
employed which provided the amount of light scattered by the aerosols
in some arbitrary unit, but did not yield directly the number of aerosols
per unit volume [29,47,49,59].   One notable exception was  the test by
Renzetti and Doyle [49] who performed flow type experiments in a 0.05 m
Pyrex flask irradiated by mercury arc lamps.  They recorded the number
                                 _3
of aerosols produced to be 290 cm   in a mixture containing 1 ppm nitric
oxide and 2 ppnr cyclohexene at a relative humidity of 50%.  The present
data, shown in Fig. 5.15, indicate that the number of aerosols is in the
range of 100 to 300 cm"3.  Renzetti and Doyle's data fall  within the
spread of the present results.
     A direct quantitative comparison is not feasible between the pre-
sent data and the data obtained by light scattering methods.  A quali-
tative comparison between the present results and the data obtained by
light scattering methods is possible, because the amount of light
scattered is proportional to the number of aerosols in the mixture.  Mea-
surements of light scattering in hydrocarbon-nitrogen oxides-air mixtures
were reported in references [29,47,49,59].  Renzetti and Doyle [49]
                                           o
performed flow type experiments in a 0.05 m  Pyrex flask irradiated by
mercury lamps, using 3 ppm hydrocarbon, 1 ppm nitric oxide, and 50%
relative humidity in all their tests.  Prager et. al. [47] performed
static type experiments in a long-path infrared absorption cell irradiated
by mercury lamps.  Prager et. al. kept the mixture compositions at

-------
                               -89-
10 ppm hydrocarbon, 5 ppm nitric oxide, and 100 ppm water vapor (~  0.3%
relative humidity).  Groblicki and Nebel [29] tested gas-air mixtures in
a 8.5 nr* stainless steel chamber irradiated by fluorescent lamps.   The
mixture composition was fixed at 4 ppm hydrocarbon and 2 ppm nitric oxide.
Stevenson et. al. [59] performed flow type experiments in a 0.125  m
Pyrex tube irradiated by fluorescent lamps and measured the aerosol  light
scattering in cyclohexene-nitrogen dioxide-air mixtures (at 3.5 ppm cyclo-
hexene and 2 ppm nitrogen dioxide, and at 1.83 ppm cyclohexene and 1 ppm
nitrogen dioxide).  Altogether, over 40 different types of hydrocarbons
were tested by these previous investigators [29,47,49,59].  However, in
each experiment, only one set of initial hydrocarbon and nitrogen  oxides
concentrations, was. tested.  Hence, an aerosol light scattering versus
initial  concentration ratio curve could not be plotted for any of  these
hydrocarbons, with the exception of cyclohexene for which light scatter-
ing was measured by four different investigators [29,47,49,59]. For cy-
clohexene, the aerosol light scattering against the concentration  ratio
could be plotted.  This plot, given in Fig. 5.16, may be compared  with the
present data shown in Fig. 5.15.  It is seen that these two figures  ex-
hibit certain common features.  First, both the aerosol light scattering
and the aerosol number count indicate that when the initial concentration
ratio is increased from 2 to 5 the amount of aerosols produced is  in-
creased by a factor of ten (i.e. the aerosol content is increased  from
40 to 400 cm" ; the light scattering from 3 to 30).  Second, the light
scattering measurements support the present observations that:  a) few
aerosols are produced when the concentration ratio is below 2.4, and
b) the number of aerosols produced defends mainly on the concentration
ratio and  very  little on  the  initial  nitrogen dioxide  concentration.

-------
                       -90-
       100
     o

     !o
         10
     o:
     LJ
     <
     o
III  I  I L-
                           I    I   I  I  I  I  1
           I
            10
                      INITIAL

             CONCENTRATION  RATIO

                [C6HJ0/[NOX]0

Fig. 5.16.  Light scattering from aerosols produced in cyclohexene-
          nitrogen oxides-air mixtures, a Stevenson et. al. [59];
          4 Groblicki and fJebel [29];  S Prager et. al. [47];
          { Renzetti  and Doyle [49]

-------
                               -91-
     In addition to the total number of aerosols, the size distributions
of the aerosols were also determined for several  mixture compositions.
The results of these aerosol size measurements obtained at two different
sampling stations (#3 and #11) are shown in Fig.  5.17.  The data are
expressed in terms of the dimensionless parameters (see Section II)
              7
J,
'
                       A NMr
                                                                   (2.3)
With the use of these paramp.ters, all aerosol  size measurements should
fall on one common curve [21, 45].  Moreover,  in the size range where
the aerosol is irt a ai.ate of dynamic equilibrium (i.e.  the rate of gain
of aerosols by ro&gulation is equal to the rate of loss of aerosols due
to coagulation, diffusion and sedimentation), ^- should  vary inversely
with the fourth power of *) [11,22,38,45], i.e.,

                  Y* <^  n-4                                     (2.6)
     The results in Fig. 5.17 show that, on a  graph of/" versus ^ , the
present data follow a common curve and the slope of this curve is nearly
-4.  The slope of the curve becomes somewhat steeper for larger values
of H (i.e., larger aerosols).  These results agree very closely with the
results of Husar and Whitby  [35] (solid line  in Fig. 5.17) who monitored
the growth of aerosols due to photochemical reactions in atmospheric air
contained in a 77.5 m  polyethylene bag.  Husar and Whitby did not re-
port the composition of the atmospheric air.

-------
                             -92-
           iov
           10
  4
                   o   In Tunnel
                 	Room Air

                 	Atm.Air
                                              \
                                  A
Present
00,0

Husor a   \  \
Whllby    V\
                                                 •\
                                                  •  \
             0.03     0.1
                            1.0
                     10
Aerosol  size distribution obtained in  the tunnel for
cyclohexene-nitrogen  dioxide-air mixtures,  o  in the  tunnel,
present  study;	room air, present study;	atmospheric
air, Husar and Whitby [35]

-------
                               -93-
     The size distribution of the aerosols in the room where the pre-
sent tests were performed (see Fig. 2.6) is also included in Fig.  5.17
(dotted line).  It appears that the size distributions (expressed  in
terms of ijr and n ) of the aerosols produced by the photochemical reac-
tions, either in a tunnel or in a bag, do not differ markedly from the
size distribution of the aerosols in the natural atomsphere.  The  aero-
sol size distributions in the present tunnel, in Husar and Whitby's test,
and in the atmosphere all agree very well with each other when presented
on a\t- versus rj plot.
     The size distributions (expressed in terms of ^ and rj ) seem  to be
unaffected by the mixture composition, the light intensity or the  rela-
tive humidity.

-------
                               -94-
                            CHAPTER VI
                        CONCLUDING REMARKS
     The experiments performed in this  investigation  provide  new  informa-
tion on the participate formation in sulfur dioxide-air mixtures  and on  the
ozone and particulate formation in cyclohexene-nitrogen dioxide-air mix-
tures.
     The results obtained during the photooxidation of sulfur dioxide-air
mixtures showed that, for initial sulfur dioxide concentrations ranging
from 0.2 to 1  ppm and relative humidities ranging from 25  to  100%, the
maximum number of nuclei varied directly with the sulfur dioxide  concen-
tration and with tie square of the relative humidity.   The present results
obtained at sulfur dioxide concentrations less than 1  ppm  are in  general
agreement with the data reported by earlier investigators.
     Aerosol measurements were also made in sulfur dioxide-air mixtures;
the amounts of aerosols produced in the mixture were  found to be  negligible.
     The maximum amount of ozone formed in  cyclohexene-nitrogen dioxide-
air mixtures was shown to be directly proportional to a) the  square  root
of the initial cyclohexene concentration, b) the square root  of the  ini-
tial nitrogen dioxide concentration, and c) the square root of the light
intensity.  Using the results of the present study and the data obtained
by previous investigators for different types of hydrocarbons, expres-
sions were developed which indicated quantitatively  the maximum amount
of ozone which can be formed with different types of  hydrocarbons.
     Nuclei measurements in cyclohexene-nitrogen dioxide-air  mixtures
showed that the maximum number of nuclei varied directly with the square
root of both the cyclohexene and the nitrogen dioxide concentrations.
     The production of aerosols in cyclohexene-nitrogen dioxide-air  mix-
tures appeared to depend only on the ratio of the initial  concentrations

-------
                               -95-
of cyclohexene and nitrogen dioxide.  Within the accuracy of the measure-
ments, the amount of aerosols formed seemed to be insensitive to the nit-
rogen dioxide concentration.
     The aerosol size distributions in cyclohexene-nitrogen dioxide-air
mixtures were also measured for several mixture compositions.  The size
distributions were found to follow closely Friedlander's universal self-
preserving curve.
     During the experiments with cyclohexene-nitrogen dioxide-air mixtures,
the light intensity and the relative humidity were kept constant.  The
effects of these two parameters on the amount of ozone formed could be
assessed by utilizing the results of previous investigations.  The effects
of the light intensity and the humidity on nuclei and aerosol formation
could not be evaluated because of the lack of relevant data.   Hence, the
influence of light intensity and relative humidity on particulate forma-
tion merit further investigations.
     Finally, it is noted that static type experiments were also performed
in  a 0.6 m i.d. and 1.2 m tall Plexiglas irradiation chamber.  The re-
sults obtained from these tests were unrepeatable and were scattered widely.
The exact reasons for the scatter in the data are unknown.  It is strongly
suggested, however, that the scatter was caused by the fact that the cham-
ber could not be cleaned thoroughly before each test.  Although the chamber
was flushed with clean air several times prior to each test,  such flushing
did not seem to clean the system adequately.  This conclusion is supported
by the observation that in the flow type experiments, up to seven days of
continuous flushing was needed to achieve a "clean" chamber (Section III).

-------
                             APPENDIX A
                         LITERATURE SURVEY

     In the experiments performed in the past to study photochemical  re-
actions, the following procedure was generally used.   The  gas  mixture was
introduced into a container.   The mixture was then irradiated  and  some
products of the reaction were monitored.  The source  of irradiation was
either natural  sunlight or artificial ultraviolet light.   The  gas  mix-
ture either stayed inside the container during the entire  experiment
(static type experiment) or flowed through the container at a  steady
rate (flow type experiment).
     In this survey, only those experiments are summarized in  which the
measurements of ozone (or oxidant, see below) and particulates were of
primary interest.  Accordingly, this survey is divided into two parts.
The first part describes experiments in which the measurement  of ozone
(or oxidant) was the major concern.  The second part  describes experi-
ments in which the production of particulates wpre investigated.  The
results of the various studies are presented in chronological  order.
Ozone
     Here the results of those previous investigations are included in
which the amount of either ozone or oxidant was measured.   This is be-
cause the formation processes of the oxidant and of the ozone  are very
similar [8,69].  Moreover, in a major portion of the  existing  literature
measurements of the oxidant and not ozone were reported.   It is, there-
fore, both desirable and advantageous to include these reports in the
survey.

                                 -96-

-------
                                -97-
     Haagen-Smit [50] was among the first to study the photochemical
reaction in hydrocarbon-nitrogen oxides-air mixtures.   He investigated
the formation of ozone in a series of static type experiments  in which
                                       3
the gas mixture was placed in a 0.005 m  Pyrex flask irradiated by flu-
orescent lights.  The gas mixture tested was 3-methyheptane and nitro-
gen dioxide diluted in dry oxygen.  Haagen-Smit varied the initial  con-
centrations of 3-methyheptane (0.1 to 10,000 ppm) and  of nitrogen dioxide
(0.04 to 400 ppm).  In the report, he also presented measurements in
which natural sunlight and air were used instead of the fluorescent
lights and oxygen.  In all these experiments, the amounts of ozone gen-
erated were inferred from the crack depths of the rubber strips placed
inside the flask and irradiated for ten hours.
     Scott [47] performed static type experiments in which the gas mix-
ture was placed in a long-path infrared absorption cell  irradiated by  a
mercury lamp.  The experiments were performed for two  initial  concentra-
tions of 3-methyheptane (3 and 40 ppm).  The initial nitrogen  dioxide
concentration was varied from 0.1 to 8 ppm.  For each  mixture  composi-
tion, the amount of ozone generated was monitored as a function of the
irradiation time (up to six hours).
     Harton and Bolze [29] performed a series of static type experiments
in which the hydrocarbon-nitrogen dioxide-air mixture  was placed in a
62 m  glass chamber irradiated by natural sunlight.  The initial  con-
centrations were usually 0.3 and 0.6 ppm for the hydrocarbon and 0.6 and
1.2 ppm for the nitrogen dioxide.  Harton and Bolze tested over thirty
different types of hydrocarbons which included paraffins, straight and
branched olefins, cyclo-olefins, diolefins, aromatics  as well  as commer-

-------
                                -98-
cial solvents.  Besides measuring the oxidant and the nitrogen oxides
concentrations during the tests, eye irritation measurements were also
taken.
     Altshuller and Cohen [2] performed static type experiments in 0.09
 3
m  Teflon bags irradiated by fluorescent lamps.  They studied the photo-
chemical reactions in ethylene-nitrie oxide-air mixtures by varying the
initial concentrations of ethylene (0.1 to 5 ppm) and of nitric oxide
(0.2 to 10 ppm).  In addition to measuring the amount of oxidant gen-
erated, Altshuller and Cohen also monitored the concentrations of ethy-
lene, nitrogen dioxide, carbon monoxide, and formaldehyde during the
tests.
     Glasson and Tuesday [25] performed static type experiments.  The
hydrocarbon-nitric oxide-air mixture was placed in a long-path infrared
absorption cell irradiated by fluorescent lights.  Glasson and Tuesday
varied the initial concentration of nitric oxide from 0.05 to 3 ppm.
For the five hydrocarbon tested (propylene, ethylene, trans-2-butene,
isobutene, and m-xylene), the initial concentrations most often used
were 1 and 2 ppm.  In their experiments, Glasson and Tuesday measured the
formation rate of ozone, the oxidation rate of hydrocarbon and the con-
centration of peroxyacetyl nitrate.
     Romanovsky et. al. [51] performed static type experiments in a
32 m  glass aluminum chamber irradiated by fluorescent lamps.  The mix-
ture investigated was propylene and nitric oxide in air.  The initial
concentration of propylene was varied from 0 to 8 ppm and that of nitric
oxide was varied from 0 to 4 ppm.  During the experiments, the amounts
of oxidant, nitrogen oxides, propylene, and formaldehyde were monitored
and eye irritation was also studied.

-------
                                -99-
     Altshuller et. al. [5] performed both static type and flow type
                     3
experiments in a 95 m  aluminum chamber irradiated by fluorescent lamps.
In the propylene-nitric oxide-air mixture studied, the initial  concen-
tration of propylene was varied from 0.25 to 3 ppm and the nitric oxide
initial concentration was varied from 0.125 to 4 ppm.  The concentra-
tions of oxidant, nitrogen dioxide, carbon monoxide, methyl  nitrate, and
the aldehydes were all monitored.
     Stephens and Price [58] performed flow type experiments in a 0.141
m  Pyrex tube irradiated by fluorescent lamps.  The initial  concentration
of cis-2-butene was varied from 0 to 2.4 ppm while the nitric oxide con-
centrations used were 0.05, 0.125, 0.25, and 0.5 ppm.  Stephens and Price
measured the amounts of oxidant, nitrogen dioxide, methyl  nitrate, and
peroxyacetyl nitrate during these tests.
     Bufalini and Altshuller [10] performed static type experiments in
                                                                    3
which hydrocarbon-nitric oxide-air mixtures were placed in a 0.072 m
Pyrex flask irradiated by fluorescent lamps.  The hydrocarbons  tested
were 2-3-dimetnyl-2-butene, 1-butene, n-butane, and 1,3,5-trimethylben-
zene.  The initial concentrations of the hydrocarbon were  varied such that
the ratio of the hydrocarbon concentration to the nitric oxide  concen-
tration ranged from 0 to 5.  The parameters measured included the amount
of oxidant formed, the percent of hydrocarbon reacted, and the  nitrogen
dioxide dosage produced.
     Altshuller et. al. [6] performed static type experiments in which
paraffinic hydrocarbon-nitric oxide-air mixtures were placed in a 95 m
aluminum chamber irradiated by fluorescent lamps.  Of the  eight paraffins
studied, n-butane was most frequently tested.  The initial concentra-
tion of n-butane was kept at 3 and 6 ppm while the nitric  oxide concen-

-------
                                -100-
tration was varied from 0 to 2.4 ppm.   In their experiments Altshuller
et. al. monitored the amounts of oxidant, nitrogen dioxide, aldehydes,
peroxyacetyl nitrate, and methyl nitrate, and also studied the eye irri-
tation effect as well as the damage  incurred in Pinto Beam plants.
     Using the apparatus in reference [6], Altshuller et.  al.  [7] also
studied the photochemical reaction in aromatic hydrocarbon-nitric oxide-
air mixtures.  The initial concentrations of the two aromatics studied
(toluene and m-xylene) were kept at 1.5, 3, and 4.5 ppm while  the nitric
oxide initial concentration was varied from 0.2 to 1.5 ppm. Altshuller
et. al. measured the dosages of the oxidant, nitrogen dioxide  and per-
oxyacetyl nitrate and evaluated eye irritation effects.
     Kopczynski et. al. [41] used the same apparatus as Altshuller et. al.
[6,7].  Kopczynski et. al. investigated the photochemical  reaction in
aldehyde-nitric oxide-air mixtures.  The initial concentrations of the
three aldehydes tested (formaldehyde,  acetaldehyde and propionaldehyde)
were kept at 3 ppm while the initial  nitric oxide concentration was  varied
from 0.1 to 7.6 ppm.  In this investigation the concentrations of the
oxidant, hydrocarbon, nitrogen dioxide and peroxyacetyl nitrate were
monitored and the eye irritation effect was studied.
     In general, the results of all the previous investigations showed
that, for a given initial hydrocarbon concentration, the amount of ozone
produced first increased with the initial nitrogen oxides  concentration,
reached a peak, and then decreased with further increase in the nitro-
gen oxides concentration.  When the initial nitrogen oxides concentra-
tion was kept fixed and the initial hydrocarbon concentration  was varied,
the amount of ozone produced also first increased with the initial

-------
                               -101-
hydrocarbon concentration, reached a peak, and then decreased with
further increase in the initial  hydrocarbon concentration.   Compre-
hensive quantitative comparisons and correlations of the results were
not attempted in these investigations.
Particulates
     Particulate formation has been studied in many different types of
gas-air mixtures and the particles formed have been generally classi-
fied under two categories.  "Condensation nuclei" are particles whose
diameters lie between 2 nm and 0.1 Mm and "aerosols" are particles
whose diameters are greater than 0.3yum.  Condensation nuclei were
usually detected and measured by condensation nuclei counters (e.g.
General Electrtc Condensation Nuclei Counter).  Aerosols were frequently
detected and" measured by aerosol light scattering photometers (e.g.
Royco Particle Counter).
     The formation of particles  during the photooxidation of sulfur
dioxide-air mixtures were studied by several  investigators  [13,14,17,
24,40,48,49].  Gerhard and Johnstone [24] performed static  type experi-
ments in a 0.008 cnr* Lucite chamber irradiated by a sun lamp.  The ini-
tial sulfur dioxide concentrations were 5, 10, 20, and 30 ppm.  The
relative humidity was varied from 32 to 91% and the length  of irradia-
tion time from 15 minutes to 6 hours.   Gerhard and Johnstone measured
the amount of sulfuric acid drops collected on a high velocity impactor.
     Dunham [17] performed two static  type tests to demonstrate the
pohtooxidation mechanism in sulfur dioxide-air mixtures.  In the first
test, sulfur dioxide was introduced into a quartz tube containing ozone-
rich air, while the ultraviolet  light  was turned off.  In the second
test, a sulfur dioxide-air mixture in  the quartz tube was irradiated by

-------
                               -102-
an ultraviolet light.  Dunham found that nuclei was formed only in the

second test and concluded that nuclei formation was due to the photo-

oxidation of sulfur dioxide.

     Renzetti and Doyle [49] performed flow type experiments in a 0.05

m  Pyrex flask irradiated by mercury lamps.  The initial  sulfur dioxide

concentration was varied from 0 to 0.5 ppm.  Relative humidities of 0

and 50% were used.  The residence time of the gas-air mixture inside the

flask was kept at either 17 or 27 minutes.   Particulate formation was

measured by an aerosol counter photometer and a nuclei counter.

     Quon et. al. [48] performed a series of static type  experiments in

0.014 m  Saran bags irradiated by fluorescent lamps.   The initial  sulfur

dioxide concentration was varied from 0.202 to 0.647  ppm, the relative

humidity from 13 to 77%, and the length of the irradiation time from 30

to 240 seconds.  The amount of nuclei formed was measured by a conden-

sation nuclei counter.
                                                                        T
     Cox and Penkett [13] performed static type experiments in a 0.216 m

aluminum chamber irradiated by natural sunlight.  Sulfur  dioxide at var-

ious initial concentrations (0.08 to 0.76 ppm) were tested.  Cox and

Penkett did not report the  relative humidity in the  tests.
                                                          o
     Cox [14] performed flow type experiments in a 0.005  m  Pyrex flask

irradiated by a mercury lamp.  The initial  sulfur dioxide concentration

was varied from 5 to 1000 ppm.  The relative humidities tested were 1,

5, 24, and 80%.  The residence time of the gas mixture inside the flask

was fixed at 150 seconds.  The number of nuclei formed was measured

during the tests.

-------
                                -103-
     Kocmond et. al.  [40] performed a series of static type experi-
ments in three different chambers irradiated by fluorescent lamps.   Two
of the chambers were 0.25 and 1.7 m  Teflon bags, the third one was a
600 m  steel chamber whose interior surfaces were lined with a spec-
ially formulated fluoro-epoxy polymer.  Various initial  sulfur dioxide
concentrations (from 0.049 to 2.88 ppm) and relative humidities (from 11
to 77%) were used in the experiments.  The particulates produced were
measured by a Whitby Aerosol Analyzer and a General  Electric Condensa-
tion Nuclei Counter.
     The results from the investigations described above all agreed that
the number of nuclei  produced increased with the initial sulfur dioxide
concentration.  However, the exact relationship between the number  of
nuclei produced and the initial  sulfur dioxide concentration still  re-
mains unclear.  There is also disagreement among the reports on the
effect of the relative humidity on the production of the condensation
nuclei.  Some studies showed that the number of nuclei produced in-
creased with humidity [14,49]; others found that there was no humidity
effect [24,48].
     Experiments on particulate formation in hydrocarbon-nitrogen
oxides-air mixtures were reported in references [29,47,49,59].  Using
the apparatus described before,  Renzetti and Doyle [49]   studied aerosol
production in 18 different types of hydrocarbons.  The same initial
concentration (3 ppm hydrocarbon, 1  ppm nitric oxide and 50% relative
humidity) was used in each test for each hydrocarbon.  Renzetti and
Doyle used an aerosol counter photometer to measure  the amount of light
scattered by the aerosols.

-------
                                -104-
     Prager et. al.  [47] performed static type experiments in a long-
path infrared absorption cell irradiated by mercury lamps, with 10 ppm
hydrocarbon, 5 ppm nitrogen dioxide and 100 ppm water vapor.   Altogether
Prager et. al. tested 17 different types of hydrocarbons for aerosol
production.  They used a Sinclair-Phoenix smoke photometer to measure
the amount of light scattered by the aerosols.
     Stevenson et. al.  [59] performed flow type experiments in a 0.125 m3
Pyrex tube irradiated by fluorescent lamps.  Selected initial concentra-
tions of five hydrocarbons and nitrogen dioxide were tested.   The
amounts of particulates formed were measured by a nuclei counter and  a
Sinclair-Phoenix smoke photometer.
     Groblicki and Nebel [29] performed static type experiments in a
     3
8.5 m  stainless steel  chamber irradiated by fluorescent lamps.  Five
different hydrocarbons were used in the hydrocarbon-nitric oxide-air
mixture for the production of particulates.  The initial hydrocarbon
and nitric oxide concentrations were always 4 ppm and 2 ppm respectively.
Groblicki and Nebel  used a condensation nuclei counter and a Sinclair-
Phoenix smoke photometer in their particulate measurements.
     A total of over 40 different types of hydrocarbons were studied  for
particulate formation in the investigations listed above [29,47,49,59].
The general conclusion reached by the previous studies was that only
highly branched olefins, cyclo-olefins and diolefins were capable of
producing large quantities of particles.
     Investigations were made to study the effect of adding sulfur diox-
ide to hydrocarbon-nitrogen oxides-air mixtures [27,29,47,49,53,59].
Groblicki and Nebel  [29], Prager et. al. [47], Renzetti and Doyle [49],
and Stevenson et.al. [59] all introduced known concentrations of sulfur

-------
                                -105-
dioxide into the hydrocarbon-nitrogen oxides-air mixtures and measured
the amounts of particulates formed.  Schuck et.  al.  [53] performed static
                            3
type experiments in a 14.7 m  glass chamber irradiated by a combination
of mercury lamps and fluorescent lamps.   Schuck  et.  al.  measured the
aerosol light scattering produced in an  air mixture  containing 3 ppm
2-methyl-2-butene, 1 ppm nitric oxide, and 0.1  ppm sulfur dioxide at
various relative humidities.  Other types of hydrocarbons (ethylene,
acetylene, isobutylene, and formaldehyde) were also  tested for particu-
late formation.  Goetz and Pueschel [27] performed  flow type experi-
ments in which 1-octene-nitrogen dioxide-sulfur  dioxide-air mixtures
                        •5
were placed in a 0.019 m  glass tube irradiated  by  fluorescent lamps..
The initial concentrations of 1-octene and nitrogen  dioxide were kept at
80 and 60 ppm respectively.  The initial sulfur  dioxide concentrations
used were i* 15, and 60 ppm and the relative humidity was changed from
15 to 70%.  Goetz and Pueschel measured  the aerosols by means of a
Royco Particle Counter and an aerosol spectrometer.
     The results of these investigations showed  that, in general, sulfur
dioxide would enhance the production of  particulates in a hydrocarbon-
nitrogen oxides-air mixture.
     There have been very few experiments performed  to determine the
chemical compositions of the photochemical particles.  The particulates
formed in sulfur dioxide-air mixtures have been  found to be sulfuric
acid [13,14,24,48].  The exact chemical  compositions of the particulates
formed in hydrocarbon-nitrogen oxides-air mixtures  and in hydrocarbon-
nitrogen oxides-sulfur dioxide-air mixtures are  not  well known, although
a few measurements have indicated that the particulates are predominantly
nitrates and sulfates [18,29,47].

-------
                              -106-


                            APPENDIX B

               CERTIFIED CONCENTRATION OF TEST GASES



    Test Gas	Concentration,  ppm	Supplier	

Nitrogen Dioxide                 205         Air Products & Chemicals
                                                     Company


Cyclohexene                      219         Matheson Gas Company


Sulfur Dioxide                   208         Matheson Gas Company

-------
                              -107-

                            APPENDIX C
               ESTIMATION OF PARTICLE LOSSES DUE TO
                SETTLING, COAGULATION AND DIFFUSION

     Particles may be lost along the irradiation tunnel and along the
sampling line between the sampling stations and the particle counters.
Table C.I lists the flow conditions in the tunnel and in the sampling
line.  The "residence time", (defined as the average time during which
the gas-air mixture stays in the tunnel or in the sampling line) is ob-
tained by dividing the length of the tunnel or the sampling line by the
corresponding mean flow velocity.  The Reynolds numbers in both flow
situations are low and the flows are essentially laminar (Table C.I).

C.I  Loss Due to Settling
     The settling velocities of very small particles settling under the
action of gravity are given in reference [23].   The distance travelled
by a particle falling under the action of gravity in a certain time is
equal to the product of the settling velocity and the time.  Table C.2
lists the settling velocities and the vertical  distances travelled by
different size particles during a time interval equal to the residence
time both in the tunnel  and in the sampling line.  Particles deposit on
the lower part of the inside surface of the tube when the distance fallen
is larger than the inside diameter of the tube.  The inside diameters of
the tunnel  and the sampling line are 15.2 and 0.635 cm respectively.
These distances are larger than the distances travelled by particles
less than ~ 1 yu m in diameter.   Thus, losses due to settling is neglig-
ible for particles less  than -I  *-ni in diameter.

-------
                               -108-
 Table C.I    Flow Conditions in the Tunnel  and in the Sampling  Line

Tunnel
Sampling
Line
Diameter
cm
15.2

0.635
Length
cm
912

120
Flowrate
cnrs"'
46

4.72
Mean Flow
Velocity
cm s-T
0.255

14.9
Residence
Time
3,598

8.1
Reynolds
Number
24

59
Table C.2   Estimation of the Vertical  Distance Travelled by Different
            Size Particles Due to Gravity During Residence Time in the
            Tunnel and in the Sampling  Line
Particle
Diameter
;A m
2xlO"3
4xlO"3
IxlO"2
Settling
Velocity*
nm s-1
1.31x10
2.62x10
6.63x10
Distance
Tunnel

4.7xlO"3
9.4x!0"3
2.4xlO"2
Travelled
Sampling

IxlO"5
2xlO"5
5x1 O"5
2x10

1x10

2x10

1

2
-2

-1
-1
1.37x10'

8.64xl02

2.24xlO:

3.47x10^

1.28x10*
4.9x10"

3.1x10"

8.1x10"

1.2x10

4.6x10
1.1x10"
    _4
                                                      7x10
                                                      1.8x10"

                                                      2.8x10"

                                                      IxlO"1
                                   Tunnel Diameter=15.2 cm
                                   Sampling Line Diameter=0.635 cm
   From Table 13 in reference [23]

-------
                               -109-
C.2  Loss Due to Coagulation
     The number of particles also decreases due to coagulation with the
passage of time.  For particles initially of the same size, it can be
shown that the number of particles n, after time t, is related to the
initial number nQ by the expression

                ~n^ = ~"l + KnQt                                  (C.I)
where K is the coagulation constant [23].  Table C.3 lists the coagula-
tion constants for different particle sizes.  In order to compute the
value n/n0, t is assumed to be equal  to the residence time and n0 is
                •s              3
assumed to be 10  particles cm  .  It is found that, for every particle
size, the value n/n  calculated from eq. (C.I) is very close to one.
Hence, the Toss due to coagulation is negligible for all sizes of par-
ticles that may be present in the gas -air mixture.

C.3  Loss Due to Diffusion
     Particles may also be lost due to diffusion to the inside surface
of a tube.  An expression was given in reference [23] for the diffusion
of particles in a laminar flow inside a circular tube.  If nQ is the
initial particle concentration, n the particle concentration after passing
through a tube of radius R and length L, U the mean flow velocity in the
tube, and D the diffusion coefficient of the particle, then n/nQ is
given by
                --  = 1  -  2.56u.2/3 +  1.?, + 0.177v1/3

                              yu.    + 1.?^, + 0.177/v            (c.2)
where JU is a dimensionless parameter equal  to DL/UR^ .   Calculations
have been made for various particle sizes using eq.  (C.2)  and the re-

-------
                               -110-
    Table C.3   Calculation of Particle Loss  Due to Coagulation
                       3        3
                
-------
suits are presented in Table C.4.  It is seen that the loss of aerosols
(particle diameter > 0.3 urn) due to diffusion is negligible.   However,
for the very small particles (particle diameter < 0.01 yum)  which fall
into the condensation nuclei category, the loss due to diffusion is ap-
preciable.  Nearly all the particles whose diameters are about 0.002yum
appear to be lost due to diffusion.  A correction for this  effect was
not made to the present data since the size distribution of the conden-
sation nuclei was not measured.   The measurable range of the condensation
nuclei counter is from 0.002 u m to 0.1  M-m.   The loss due  to  diffusion
of the larger size particles recorded by the nuclei counter is seen to
be small (Table C.4).

-------
Appendix D.I  Ozone Content Along the Tunnel
[N02]
ppm
0.6
0.6
0.6
0.6
0.6
0.6
0.6
0.6
1.0
1.0
1.0
1.0
1.0
1.0
1.0
jre Composition
[C6H10]0
ppm
0.5
1.0
1.25
2.0
3.0
5.0
10.0
20.0
0.5
1.0
2.0
3.0
5.0
10.0
20.0
1
0.09
0.155
0.22
0.23
0.265
0.2
0.13
0.072
0.12
0.15
0.22
0.375
0.29
0.185
0.105
2
0.185
0.345
0.62
0.47
0.46
0.325
0.188
0.102
0.17
0.32
0.76
1.0
0.67
0.37
0.188
Ozone Content (ppm) at Different Sampling Stations
4 6 8 10 12
0.255
0.49
0.70
0.575
0.54
0.31
0.14
0.068
0.157
0.36
0.825
1.0
0.605
0.265
0.124
0^32
0.585
0.75
0.66
0.60
0.305
0.115
0.504
0.157
0.395
0.91
1.0
0.635
0.215
0.088
0.33
0.595
0.71
0.67
0.59
0.28
0.088
0.041
0.145
0.37
0.87
1.0
0.62
0.17
0.058
0.41
0.695
0.82
0.81
0.73
0.33
0.086
0.039
0.158
0.395
0.98
1.1
0.77
0.19
0.058
0.385
0.61
0.42
0.715
0.66
0.30
0.083
0.037
0.145
0.315
0.86
1.0
0.70
0.175
0.055
                                                                                   ro
                                                                                    i

-------
                                        Appendix 0.1   (continued)


 Mixture Composition                              Ozone  Content  (ppm)  at  Different  Sampling  Stations
ppm           ppm            1            2            4            6            8            10           12
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
0.5
1.0
2.0
3.0
5.0
10.0
20.0
50.0
0.5
1.0
2.0
3.0
5.0
7.5
10.0
20.0
30.0
0.115
0.115
0.27
0.41
0.65
0.38
0.215
0.078
0.112
0.118
0.168
0.28
0.55
0.85
0.65
0.35
0.24
0.115
0.15
0.585
0.98
1.3
0.81
0.415
0.133
0.081
0.092
0.167
0.52
1.2
1.65
1.3
0.71
0.44
0.071
0.11
0.545
0.91
1.2
0.69
0.29
0.09
0.041
0.047
0.088
0.38
1.05
1.4
1.15
0.56
0.32
0.054
0.092
0.50
0.90
1.2
0.695
0.225
0.068
0.029
0.032
0.062
0.26
0.95
1.2
1.2
0.5
0.23
0.041
0.070
0.41
0.8
1.1
0.66
0.16
0.046
0.0215
0.026
0.046
0.16
0.85
1.0
1.1
0.42
0.15
0.042
0.072
0.41
0.86
1.2
0.82
0.165
0.043
0.022
0:026
0.044
0.13
0.77
1.0
1.05
0.47
0.145
0.037
0.063
0.33
0.73
1.1
0.71
0.135
0.032
0.017
0.02
0.033
0.09
0.73
0.60
1.0
0.32
0.095
                                                                                                               U)
                                                                                                               I

-------
                                         Appendix D.I  (continued)




Mixture  Composition                                  Ozone Content (ppm) at Different Sampling Stations
  ppm          ppm            1            2           4           6           8           10          12
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
0.5
1.0
2.0
3.0
5.0
10.0
12.5
20.0
30.0
50.0
0.07
0.12
0.143
0.17
0.34
0.8
1.8
0.56
0.4
0.23
0.035
0.061
0.074
0.106
0.35
1.75
1.9
1.4
0.84
0.42
0.015
0.027
0.035
0.049
0.148
1.55
1.55
1.25
0.67
0.33
0.01
0.018
0.024
0.034
0.095
1.45
1.45
1.25
0.59
0.25
0.008
0.013
0.018
0.024
0.068
1.3
1.3
1.2
0.46
0.17
0.007
0.012
0.017
0.025
0.068
1.4
1.4
1.35
0.50
0.15
0.007
0.009
0.012
0.018
0.042
0.90
0.80
0.90
0.30
0.09
                                                                                                                 -p.
                                                                                                                 I

-------
                                       Appendix D.I   (continued)



Mixture Composition                               Ozone Content  (ppm)  at  Different  Sampling  Stations
L ~£JQ
ppm
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
L O 1 \J J/v
ppm
0.5
1.0
2.0
3.0
5.0
10.0
15.0
20.0
25.0
30.0
50.0
1
0.054
0.039
0.039
0.036
0.037
0.125
0.35
0.70
0.95
0.70
0.50
2
0.014
0.01
0.009
0.015
0.01
0.055
0.68
1.50
1.70
1.60
1.05
4.
0.0055
0.004
0.004
0.009
0.004
0.027
0.62
1.25
1.4
1.5
0.97
6
0.0035
0.001
0.002
0.007
0.002
0.022
0.53
1.05
1.1
1.4
0.95
8
0.003
0.0
0.001
0.006
0.001
0.017
0.51
0.85
0.9
1.3
0.87
10
0.0025
0.0
0.001
0.005
0.001
0.017
0.55
0.90
0.95
1.4
0.90
12
0.002
0.0
0.001
0.005
0.001
0.012
0.37
0.60
0.70
0.90
0.60

-------
                                                                      Q
           Appendix 0.2   Nuclei  (N) and Aerosol  (A)  Contents  (no.  cm  )  Along  the  Tunnel
Mixture Composition
Sampling Station
[N02]
ppm
0.6
0.6
0.6
0.6
0.6
0.6
0.6
0.6
1.0
1.0
1.0
1.0
1.0
1.0
1.0
ppm
0.5
1.0
1.25
2.0
3.0
5.0
10.0
20.0
0.5
1.0
2.0
3.0
5.0
10.0
20.0
3
N
250
350
650
400
450
400
< 200
< 200
< 200
250
480
550
450
400
300
A
0.0
0.0
0.0
17
95
170
230
152
0.0
0.0
1
159
389
512
424
5
N
250
400
650
500
500
400
350
^200
<200
300
650
700
550
400
300
A
0.0
0.0
0.0
71
240
356
374
272
0.0
0.0
3
243
555
664
565
7
N
250
550
600
700
550
450
300
<200
<200
350
700
750
520
350
250
A
0.0
0.0
0.0
92
314
424
388
300
0.0
0.0
6
216
572
505
424
9
N
250
550
500
650
650
400
250
<200
300
300
600
650
450
350
350
A
0.0
0.0
0.0
99
339
424
360
293
0.0
0.0
7
219
551
576
530
11
N
300
600
800
600
600
500
250
250
<200
<200
700
700
450
400
300
A
0.0
0.0
0.0
85
293
424
353
279
0.0
0.0
14
184
459
530
671

-------
                                      Appendix D.2  (continued)
Mixture Composition
Sampling Station
[N02]Q
ppm
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
[C6H10]0
ppm
0.5
1.0
2.0
3.0
5.0
10.0
20.0
50.0
0.5
1.0
2.0
3.0
5.0
7.5
10.0
20.0
30.0
3
N
< 200

-------
Appendix D.2  (continued)
Mixture
[N02]0
ppm
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
Composition
[C6H10]0
ppm
0.5
1.0
2.0
3.0
5.0
10.0
12.5
20.0
30.0
50.0
3
< 200
* 200
<200
250
500
1400
2000
1600
1000
650
A
0.0
0.0
0.0
0.0
0.0
155
152
290
300
212
5
N
<2QQ
<200
<200
300
400
1300
1500
1500
800
705
A
0.0
0.0
0.0
0.0
0.0
166
145
240
261
237
*>OIII|J 1 II
N
<200
< 200
<200
300
500
1400
1200
1100
800
680
iy j ia i. n
A
0.0
0.0
0.0
0.0
0.0
124
95
120
198
198
9
N
<200
< 200
<200
300
400
1500
1200
900
850
500
A
0.0
0.0
0.0
0.0
0.0
138
99
134
346
565
11
N
<200
<200
<200
300
500
1600
800
1200
900
600
A
0.0
0.0
0.0
0.0
0.0
124
81
124
636
813
                                                                        00

-------
                                         Appendix 0.2  (continued)
Mixture Composition
[N02]Q   [C6Hlo]0
 ppm      ppm
Sampling Station
               9
   A        N
11
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
0.5
1.0
2.0
3.0
5.0
10.0
15.0
20.0
25.0
30.0
50.0
<200
< 200
< 200
< 200
< 200
500
1500
2000
2600
2200
1400
0.0
0.0
0.0
0.0
0.0
0.0
0.0
81
60
187
212
<200
<200
<200
<200
<200
500
1300
1600
2400
2200
1200
0.0
0.0
0.0
0.0
0.0
0.0
0.0
74
53
130
191
<200
<200
<200
<200
<200
350
1400
1600
2000
1700
1000
0.0
0.0
0.0
0.0
0.0
0.0
0.0
49
34
53
95
<200
<200
^200
<200
<200
250
1300
1500
2100
1700
1000
0.0
0.0
0.0
0.0
0.0
0.0
0.0
56
35
64
173
<200
< 200
<200
<200
<200
400
1200
1400
2200
1600
1300
0.0
0.0
0.0
0.0
0.0
0.0
0.0
49
31
60
460
                                                                                                                  I
                                                                                                                  «J
                                                                                                                  vo

-------
                           -120-
Appendix E   Estimated Peak Values of Ozone, Nuclei and
             Aerosols for Different Initial Mixture Compositions
[N02]Q
ppm
0.6
0.6
0.6
0.6
0.6
0.6
0.6
0.6
1.0
1.0
1.0
1.0
1.0
1.0
1.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
ppm
0.5
1.0
1.25
2.0
3.0
5.0
10.0
20.0
0.5
1.0
2.0
3.0
5.0
10.0
20.0
0.5
1.0
2.0
3.0
5.0
10.0
20.0
50.0
[03]
ppm
0.3
0.6
0.7
0.65
0.50
0.33
0.185
0.1
0.17
0.375
0.9
1.0
0.68
0.38
0.19
0.0
0.15
0.585
0.98
1.3
0.81
0.42
0.16
Nuclei
_3
no. cm
250
550
650
580
450
350
250
<200
< 200
350
700
750
550
400
300
<200
< 200
450
700
900
650
450
<: 200
Aerosols
no. cm
0.0
0.0
0.0
99
318
424
389
300
0.0
0.0
14
247
565
636
565
0.0
0.0
0.0
2.0
166
495
565
353

-------
           -121-
Appendix E  (continued)
ppm °
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
3.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
ppm °
0.5
1.0
2.0
3.0
5.0
7.5
10.0
20.0
30.0
0.5
1.0
2.0
3.0
5.0
10.0
12.5
20.0
30.0
50.0
0.5
1.0
2.0
3.0
5.0
10.0
15.0
20.0
25.0
30.0
50.0
[03]
ppm
0.0
0.0
0.2
0.52
1.2
1.67
1.33
0.72
0.46
0.0
0.0
0.0
0.0
0.48
1.87
2.0
1.42
0.88
0.46
0.0
0.0
0.0
0.0
0.0
0.0
0.70
1.55
1.8
1.6
1.05
Nuclei
no. cm"-'
<200
< 200
350
600
1000
1200
1100
600
500
< 200
< 200
< 200
400
700
1500
1800
1600
1000
650
< 200
<200
< 200
< 200
400
800
1500
2200
2600
2200
1400
Aerosols
no. cm- 3
0.0
0.0
0.0
0.0
10
46
343
424
406
0.0
0.0
0.0
0.0
0.0
166
152
290
636
707
0.0
0.0
0.0
0.0
0.0
0.0
0.0
81
60
187
240

-------
                               -122-
                             REFERENCES

 1.   Altshuller,  A.P.,  and  Cohen,  I.R.,   "Structural Effects on the
     Rate of Nitrogen Dioxide  Formation  in the Photo-Oxidation of Or-
     ganic Compound-Nitric  Oxide Mixtures in Air,"  International
     Journal  of Air  and Water  Pollution,  ]_, 787-797 (1963).

 2.   Altshuller,  A.P.,  and  Cohen,  I.R.,   "Atmospheric Photo-oxidation
     of the Ethylene-Nitric Oxide  System,"  International Journal of
     Air and Water Pollution,   8,  611-632 (1964).

 3.   Altshuller,  A.P.,  Klosterman, D.L.,  Leach, P.W., Hindawi, I.J.,
     and Sigsby,  J.E.,  Jr.,  "Products and Biological Effects from
     Irradiation  of  Nitrogen Oxides with  Hydrocarbons or Aldehydes
     Under Dynamic Conditions,"  International Journal of Air and
     Water Pollution, 1_0, 81-98 (1966).

 4.   Altshuller,  A.P.,  "Reactivity of Organic Substances in Atmospheric
     Photooxidatior1  Reactions,"  International Journal of Air and Water
     Pollution, 1£,  713-733 (1966).

 5.   Altshuller,  A.P.,  Kopczynski, S.L.,  Lonneman, W.A., Becker, T.L.,
     and Slater,  R., "Chemical Aspects of the Photooxidation of the
     Propylene-Nitrogen Oxide  System,"  Environmental Science and
     Technology,  1,  899-914 (1967).

 6.   Altshuller,  A.P.,  Kopczynski, S.L.,  Wilson, D., Lonneman, W.,
     and Sutterfield, F.D.,  "Photochemical Reactivities of N-Butane
     and Other Paraffinic Hydrocarbons,"  Journal of the Air Pollution
     Control  Association,   J9.'  787-790 (1969).

 7.   Altshuller,  A.P.,  Kopczynski, S.L.,  Lonneman, W.A., Sutterfield,
     F.D., and Wilson,  D.L.,   "Photochemical Reactivities of Aromatic
     Hydrocarbon-Nitrogen Oxide and Related Systems,"  Environmental
     Science and  Technology, 4_, 44-49  (1970).

 8.   Altshuller,  A.P.,  and  Bufalini, J.J.,  "Photochemical Aspects of
     Air Pollution:  A  Review,"  Environmental Science and Technology,
     5_, 39-64 (1971).

 9.   Bufalini, J.J., and Altshuller, A.P.,  "The Effect of Temperature
     on Photochemical Smog  Reactions,"   International Journal of Air
     and Water Pollution, £,  769-771 (1963).

10.   Bufalini, J.J,, and Altshuller, A.P.,  "Oxidation of Nitric Oxide
     in the Presence of Ultraviolet Light and Hydrocarbons,"  Environ-
     mental Science  and Technology, 3, 469-472 (1969).

11.   Cartwright,  J., Nagelschmidt, G., and Skidmore, J.W.,  "The Study
     of Air Pollution with  the Electron Microscope,"  Quarterly Journal
     of the Royal Meteorology  Society, 82., 82-86 (1956).

-------
                                -123-
12.  Clyne, M.A.A., Thrush, B.A., and Nayne, R.P.,   "Kinetics  of  the
     Chemiluminescent Reaction between Nitric Oxide  and  Ozone,"   Trans-
     action of the Faraday Society, 60, 359-370 (1964).

13.  Cox, R.A., and Penkett, S.A.,  "The Photooxidation  of Sulfur Diox-
     ide in Sunlight,"  Atmospheric Environment, 4^  425-433 (1970).

14.  Cox, R.A.,  "Particle Formation from Homogeneous  Reactions of Sulfur-
     Dioxide and Nitrogen Dioxide,"  Tullus, 26, 235-240 (1974).

lb.  Demerjian, K.L., Kerr, J.A., and Calvert, J.G.,  "The Mechanism of
     Photochemical Smog Formation,"  in Advances in  Environmental  Science
     and Technology, edited by Pitts, J.N., Metcalf, R.L., and Lloyd,
     A.C.,  4_, 1-262 (1974), John Wiley & Sons,  New  York.

16.  Dimitriades, B.,  "Methodology in Air Pollution Studies Using Ir-
     radiation Chambers,"  Journal  of the Air Pollution  Control Assoc-
     iation, ]_7, 460-466 (1967).

17.  Dunham, S.B.,  "Detection of Photochemical  Oxidation  of Sulfur
     Dioxide by Condensation Nuclei Techniques," Nature,  188, 51-52
     (1960).

18.  Endow, tCr Doyle, G.J., and  Jones, J.L.,  "The  Nature of  Some Model
     Photochemical Aerosols,"  Journal of the Air Pollution Control
     Association, ]_3, 141-147 (1963).

19.  Eschenroeder, A.Q., Martinez, J.R.,  "Concepts  and  Applications
     of Photochemical Smog Models,"  in "Photochemical Smog and Ozone
     Reactions," edited by Gould, R.F., Advances in  Chemistry  Series.
     113, 107-168 (1972), American Chemical Society, Washington,  D.C.

20.  Ford, H.W., Doyle, G.J., and Endon, N.,  "Rate  Constants  at  Low
     Concentrations. II Reaction  between Nitric  Oxide  and  Ozone in Air
     at Room Temperature,"  Journal of Chemical  Physics, .26, 1337  (1957).

21.  Friedlander, S.K.,  "Similarity Considerations  for  the Particle-
     Size Spectrum of a Coagulating, Sedimenting Aerosol," Journal of
     Meterology, V7, 479-483 (I960).

22.  Friedlander, S.K., and Pasceri, R.E.,  "Measurement of the Particle
     Size Distribution of the Atmospheric Aerosol: I.  Introduction and
     Experimental Methods,"  Journal of Atmospheric  Science, 22,  571-
     576 (1965).

23.  Fuchs, N.A.,  "The Mechanics of Aerosols,"   The Macmillan Company,
     New York (1964).

24.  Gerhard, E.R., and Johnstone, H.F.,  "Photochemical Oxidation of
     Sulfur Dioxide in Air,"  Industrial and Engineering Chemistry,
     47, 972-976 (1955).

-------
                               -124-
25.  Glasson, W.A., and Tuesday,  C.S.,   "Inhibition of  the Atmospheric
     Photooxidation of Hydrocarbons  by  Nitric  Oxide,"   Research Publi-
     cation GMR-475, General  Motors  Research Laboratories, Warren,
     Michigan (1965).

26.  Glasson, W.A., and Tuesday,  C.S.,   "Hydrocarbon Reactivities in the
     Atmospheric Photooxidation of Nitric Oxide,"  Environmental Science
     and Technology, 4_, 916-924  (1970).

27.  Goetz, A., and Pueschel, R., "Basic Mechanisms of Photochemical
     Aerosol Formation,"  Atmospheric Environment, ]_, 287-306  (1967).

28.  Goetz, A., and Klejnst,  O.J.,  "Formation and Degradation of Aero-
     colloids by Ultraviolet  Radiation,"  Environmental  Science and
     Technology, 6_, 143-151  (1972).

29.  Groblicki, P.O.,  and Nebel,  G.J.,   "The Photochemical Formation of
     Aerosols in Urban Atmospheres," in "Chemical Reactions in Urban
     Atmospheres,"  edited by  Tuesday, C.S., 241-264  (1971), American
     Elsevier Publishing Company, New York.

30.  Harkins, J., and  Nicksic, S.W., "Studies  on the Role of Sulfur
     Dioxide in Visibility Reduction,"   Journal of the  Air Pollution
     Control Association, 1_5, 218-221 (1965).

31.  Harton, E.E.,  Jr., and Bolze, C.C.,  "Eye Irritation from Solar
     Radiatton of Organic Compounds  and Nitrogen Dioxide,"  Report No.
     23, Air Pollution Foundation, San  Marino, California (1958).

32.  Hecht, T.A., and  Seinfeld, J.H.,  "Development and Validation of a
     Generalized Mechanism for Photochemical Smog,"  Environmental Science
     and Technology, 6, 47-57 (1972).

33.  Heuss, J.M., and  Glasson, W.A., "Hydrocarbon Reactivity  and Eye
     Irritation,"  Environmental  Science and Technology, 2, 1109-1116
     (1968).

34.  Holmes, J.R.,  O'Brien, R.J., Crabtree, J.H., Hecht, T.A., and
     Seinfeld, J.H.,  "Measurement of Ultraviolet Radiation Intensity
     in Photochemical  Smog Studies," Environmental Science and Technology,
     7_, b!9-523 (1973).

35.  Husar, R.B., and  Whitby, K.T.,   "Growth Mechanisms  and Size Spectra
     of Photochemical  Aerosols,"  Environmental  Science  and Technology,
     7., 241-247 (1973).

36.  Jaffe, R.J., and  Smith,  F.C., Jr.,  "Factors Affecting Reactions in
     Smog Chambers,"  Lockheed Missiles & Space Company, Inc., Sunnyvale,
     California.  Presented at'the 67th Annual  Meeting  of the Air Pollu-
     tion Control Association, Denver,  Colorado, June 13, 1974.

37.  Johnston, ri.S., and Crosby,  H.J.,   "Kinetics of the Fast  Gas Phase
     Reaction between  Ozone and Nitric  Oxide," Journal  of Chemical
     Physics, 22, 689-692 (1954).

-------
                                -125-
 36.   Junge,  C.E.,   "Atomspheric  Chemistry,"   in Advances In Geophysics.
      edited  by  Lansberg,  H.E., and  Van Kieghem, J., 4_, 1-101 (1958),
      Academic Press,  Mew  York.

 39.   Katz, M.,  and  Gale,  S.B.,   "f'.echanism of Photooxidation of Sulfur
      Dioxide in  Atmosphere,"  Proceedings of the Second International
      Clean Air  Congress,  edited  by  Englund, H.f'., and Beery, W.T.,
      336-343 (1971),  Academic Press, New York.

 40.   Kocmond, W.C., Kittelson, D.B., Yang, J.Y., arid Denerjian, K.L.,
      "Determination of the  Formation Mechanisms and Compositions of
      Photochemical  Aerosols,"  Calspan Corporation, Buffalo, Hew York.
      First Annual Summary Report, 31 August 1973.  Prepared for Environ-
      nental  Protection Agency, Durham, North Carolina and Coordinating
      Research Council, Mew  York, NewYork.

 41.   Kopczynski, S.L., Altshuller,  A.P., and Sutterfield, F.U.,  "Photo-
      chemical Reactivities  of Aldehyde - Nitrogen Oxide Systems,"
      Environmental  Science  anc Technology, J3, 909-918 (1974).

 42.   Laity,  J.L.,   "A Smog  Chamber  Study Comparing Blocklight Fluroes-
      cent Lamps with Natural Sunlight,"  Environmental  Science and
      Technology, b_, 1218-1220 (1971).

 43.   Leighton, P.A.,  "Photochemistry of Air Pollution,"  Academic
      Press,  New York  (19bl).

 44.   Lillian, D.,  "Formation and Destruction of Ozone in a Simulated
      natural System (Nitrogen Dioxide + «* -Pinene + hv),"  in  "Photo-
      chemical Smog and Ozone Reactions," edited by Gould, R.F., Advances
      in Chemistry Series. 113, 211-218 (1972), American Chemical  Society,
      Washington, D.C.

 45.   Liu, B.Y.H., and Whitby, K.T.,  "Dynamic Equilibrium in Self-Pre-
      serving Aerosols,"  Journal  of Colloid and Interface Science, 26,
      lfal-165 (1968).                                               ~"~

 4b.   iiiki, H., Daby, E.E., and Weinstock, B.,  "Mechanisms  of  Smog
      Reactions,"  in "Photochemical  Smog and Ozone Reactions,"  edited
      by Gould, R.F., Advances in  Chemistry Series. 113, 16-57  (1972),
      American Chemical Society, Washington, D.C.

47.   Prager, M.J., Stephens, E.R., and Scott, W.E.,   "Aerosol  Formation
      from Gaseous Air Pollutants,"  Industrial  and Engineering  Chem-
      istry, 52_, 521-524 (1960).

48.   Quon, J.E., Siegel,  R.P., and Hulburt, H.M.,   "Particle Formation
      from Photooxidation  of Sulfur Dioxide in Air,"   Proceedings  of the
      Second  International  Clean Air Congress, edited by Englund,  H.M.,
      and Keery, W.T.,  330-335 (1971), Academic Press, Mew York.

4$.   Renzetti, N.A., and  Doyle,  G.J.,  "Photochemical Aerosol  Formation
      in Sulfur Dioxide-Hydrocarbon Systems,"  International  Journal  of
      Air Pollution,  2, 327-345 (1960).

-------
                               -126-
bG.  Rogers, L.H., ed. ,  "Proceedings of the Conference on Chemical
     Reactions in Urban Atmospheres,"  Report No.  lb,  Air Pollution
     Foundation, Los Angeles, California, 1956.

bl.  Romanovsky, J.C.,  Ingels, R.M. ,  and Gordon, R.J.,  "Estimation  of
     Smog Effects in the Hydrocarbon-Nitric Oxide  Systeir,"  Journal  of
     the Air Pollution  Control Association, 1_7_,  454-45S (1967).

62.  Schofield, K. ,  "An Evaluation  of Kinetic Rate Uata for Reactions
     of Neutrals of Atmospheric Interest,"  Planetary  and Space  Science,
     ]5., 643-670 (1967).

b3.  Schuck, E.A. , Doyle, G.J., and  Endow, N. ,  "A Progress Report on
     the Photochemistry of Polluted  Atmospheres,"   Report No.  31,  Air
     Pollution Foundation, San Marino, California, 1960.

b4.  Schuck, E.A. , and  Stephens, E.R.,  "Oxides  of Nitrogen,"   in
     Advances in Environmental Science and Technology, edited  by Pitts,
     J.i, ..and Metcalf,  R.L.,  1, 73-117 (1969),  John Wiley i Sons,
     Mew York.

b5.  Seinfeld, J.H., Hecht, T.A. , and Roth, P.M.,   "Existing fieeds in
     the Experimental and Observational  Study of Atmospheric Chen.ical
     Reactions," Report Mo. EPA-R4-73-031, Systems Applications, Inc.,
     deverly hills, California, Prepared for U.S.  Environmental  Protec-
     tion Agency, 1973.

56.  Stedman, D.H. , Daby, E.E., Stuhl , F., and Niki, H.,  "Analysis  of
     Ozone and Nitric Oxide by a Cliemi luminescent  Method in Laboratory
     and Atmospheric Studies of Photochemical Smog,"  Journal  of the Air
     Pollution Control  Association,  22_,  260-263  (1972).

b7.  Stedman, D.H. , and l.iki, H. ,  "Photolysis of  NO?  in Air as  Measure-
     ment Method for Light Intensity,"  Environmental  Science  and
     Technology, 7., 735-739 (1973).

58.  Stephens, E.R. , and Price, H.A.,  "Atmospheric Photochemical  Reac-
     tions in a Tube Flow Reactor,"   Atmospheric Environment,  3, 573-
         (1969).
b9.  Stevenson, H.J.R. , Sanderson, D.E., and Altshuller,  A. P.,   "Forma-
     tion of Photochemical  Aerosols,"  International  Journal  of Air and
     Water Pollution, £, 3C7-375 (1965).

60.  Tuesday, C.S.,  "The Atmosphere Photooxidation of Trans-Butene-2
     and Nitric Oxide,"  in "Chemical Reactions in the Lower  and Upper
     Atmosphere," edited by Cadle, R.D. , 15-49 (1961), Interscience,
     Mew York.

Ol.  Wayne, L.G. , Weisburd, II. ,  Danchick, R., and Kokin,  A.,   "Final
     Report- Development of a Simulation  Model for Estinating  Grounc
     Level Concentrations of Photochemical  Pollutants,"  Technical
     Memorandum, System Development Corporation, Santa Monica,  California,
     (1971).

-------
                               -127-
62.  Westberg, K., and Cohen, N.,  "The Chemical  Kinetics  of Photo-
     chemical Smog as Analyzed by Computer,"  ATR-70(8107)-1, The Aero-
     space Corporation, LI Segundo, California (1969).

03.  Wilson, Wm. E., Jr., Merryrcan, E.L.,  and Levy,  A.,   "A Litera-
     ture Survey of Aerosol  Formation and  Visibility Reduction  in Photo-
     chemical Smog,"  Battelle Memorial Institute,  Columbus, Ohio,
     Project EF-2, August 1, 1969, Prepared for the  American Petroleum
     Institute, Committee for Air and Water Conservation.

64.  Wilson, W.E., Jr., Levy, A., and Wimmer, D.B.,   "A  Study of Sulfur
     Dioxide in Photochemical Smog II.   Effect of Sulfur Dioxide on
     Oxidant Formation in Photochemical Smog,"  Journal  of the  Air
     Pollution Control Association, 22_, 27-32 (1972).

65.  Wilson, W.E., Jr., Merryman, h.L., Levy, A., and Taliaferro, H.R.,
     "Aerosol Formation in Photochemical Smog I.  Effect of Stirring,"
     Journal of the Air Pollution Control  Association, 21, 128-132
     (1971).

66.  "Air Quality Criteria for Particulate Matter,"  National Air Pollu-
     tion Control Administration, Publication No. AP-49, 1969,  U.S.
     Department of Health, Education, and  Welfare, Washington,  D.C.

67.  "Air Quality Criteria for Sulfur Oxides,"  National  Air Pollution
     Control Administration  Publication No.  AP-50, January 1969,
     U.S. Department of Health, Education, and Welfare,  Washington, D.C.

68.  "Air Quality Criteria for iJitrogen Oxides,"  Pollution  Control
     Uffice Publication ho.  AP-G4, 1971, U.S.  Environmental  Protection
     Agency, Washington, D.C.

69.  "Air Quality Criteria for Photochemical  Oxidants,"  National Air
     Pollution Control Administration Publication No. AP-63, 1970, U.S.
     Department of Health, Education  and Welfare, Washington, D.C.

70.  "Air Quality Criteria for Hydrocarbons,"   National  Air  Pollution
     Control Administration  Publication No. AP-64, 1970,  U.S. Depart-
     ment of Health, Education and Welfare, Washington,  D.C.

-------
                                            128
                                  TECHNICAL REPORT DATA
                           (Please read Instructions on the reverse before completing)
1  REPORT NO
 EPA-460/3-76-Q05
                                                          3 RECIPIENT'S ACCESSION-NO
4 TITLE AND SUBTITLE

 Ozone and Particulate  Formation  In Photochemical

 Reactions                        	
             5 REPORT DATE

              April  1Q75
             6 PERFORMING ORGANIZATION CODE
7 AUTHOR(S)
 Chi-Hung Shen
 George S. Springer
                                                          8 PERFORMING ORGANIZATION REPORT NO
9 PERFORMING ORGANIZATION NAME AND ADDRESS

 The University of Michigan
 Ann Arbor, Michigan   48104
                                                          10 PROGRAM ELEMENT NO.
             11 CONTRACT/GRANT NO


              R-801476
12 SPONSORING AGENCY NAME AND ADDRESS
  U.S.  Environmental  Protection Agency
  Office  of Air and Waste Management
  Office  of Mobile Source Air Pollution  Control
  2565  Plymouth Road
  Ann Arbor.  Michigan  48105	
             13 TYPE OF REPORT AND PERIOD COVERED

              TOPIC Final	
             14 SPONSORING AGENCY CODE
15 SUPPLEMENTARY NOTES
16. ABSTRACT
       Experiments were performed to  investigate the photochemical formation  of
  nuclei  and aerosols in sulfur dioxide  and  air mixtures; and also the photo-
  chemical  formation of ozona,. nuclei, and aerosols  in cyclohexene, nitrogen
  dioxide,  and air mixtures.  These mixtures were irradiated with ultraviolet
  radiation while they flowed at a constant  speed through a pyrex tubulation
  measuring 15.2 cm inside diameter by 9.12  meters long.   The quantities of ozone,
  nuclei,  and aerosols generated in the  gas  -  air mixtures were measured at
  various  points along the tubulation by means  of a  chemiluminescent analyzer, a
  condensation nuclei counter, and an aerosol  counter.
       Photo-oxidation experiments with  sulfur  dioxide and air mixtures were
  performed for five different initial sulfur dioxide concentrations of 0.2, 0.3,
  0.5,  0.7, and 1.0 ppm; and for four different relative  humidities of 25, 50, 75,
  and  100%.  Test results showed that the maximum number  of nuclei produced was
  directly  proportional  to the initial sulfur dioxide concentration, and to the
  square of the relative humidity.  There were  no measurable amounts of aerosols
  observed  during the tests.
       Experiments with  cyclohexene, nitrogen dioxide,  and air mixtures were
  performed with the initial cyclohexene concentration  ranging from 0.5 to 50 ppm;
  and  the  initial  nitrogen dioxide concentratinn fmm n fi tn in   m	
17
                               KEY WORDS AND DOCUMENT ANALYSIS
                 DESCRIPTORS
 Air pollution
 Particulates
 Ozone
 Photochemical
 Aerosols
 Nuclei
                                             b IDENTIFIERS/OPEN ENDED TERMS
                                                                       L  COSATI I ield/Gr. up
                                13B
n DISTRIBUTION STATEM&M
 Unlimited
19 SECURITY CLASS (//in Kipar
 Unclassified
20 SECURITY CLASS \flinpaKf)

 Unclassified
                                                                        21 NO OF PAGES
                                                                            135
                                                                        22 PRICE
EPA Form 2220-1 (9-73)

-------