PB 194 060
ATMOSPHERIC REACTION STUDIES IN THE LOS ANGELES BASIN,
DATA ANALYSIS AND METHODS IMPROVEMENT
Scott Research Laboratories, Incorporated
Plumsteadville, Pennsylvania
7 March 1970
PHASE II
INFORMATION SERVICE
Distributed . . •'to foster, serve and promote the
nation's economic development
and technological advancement.'
U.S. DEPARTMENT OF COMMERCE
Tfcis
has bMi approved fjtf public release and sale.
-------
REACTION "STUDIES IN
B^^^lf^^^
PHASE. 11 .1 ?•„..-'J
DATA ANALYSIS AND . •'~'
" METHODS IMPROVEMENT
' *%•-.
COORDINATING RESEARCH COUNCIL, NC.
.IAPMC Projoet No CAPA-7-08)
POLLUTION CONTROL ADMlSfilRAtHrff"
--i
(KAPCA Coatrcct Re. CPA 22-69-19)
r?** ^ -c-s. - .
RESEARCH LABORATORIEST^C
a ft n**v n±M ^ "
P O
PUAtSTEACMLLE P&MSVUMMA
NATIONAL TECHNICAL
iMORjsuntoH »vica
STANDARD TITLE PACE
FOR TECWBCAL REPORTS
Atmospheric Reaction Studies in the Los Angeles Basin
'.".'ML " I ."t Ji T.- '..run in lii'iTi Phase II
9 Perfo-ni.-^ OrtaRiuIwi Name and Address
Scott Research Laboratories, Inc.
P. 0. Sox D-ll
Pluasteadville,-Pennsylvania 18949
12 "Sponsorint Ajrvy Name and Address
Coordinating Research Council, Inc.
30 Rockefeller Plaza
New York. N. Y. 10020
National Air Pollution
Control Administration
411 Chapel Hill Street
Durham, North Carolina
Report One
July 28, 1969
6 Perlornmi Organization CMC
> Performing Organiiatlon'Rept No~
1116
10. Pioiecl/Taik/work Unit No
CAPA-7-68/22-69-19
IT Conrict7fi'ram"No " ~
NAPCA-22-6S-19
CRC-APRAC-CAPA- 7-68-1
rypeo
Atmospheric
August 22-November 14,
14. Sponsoring Agency Code""
1)68
16 Abstram
4 comprehensive survey of the atmosphere of the Los Angeles Basin was conducted during
period August 22 to November 14, 1968. A large number of contaminants were determined
continuously and periodically, and readings of pertinent meteorological factors i
at two ground level sites and in an aircraft at several altitudes above the grow
Mobile trailer-type laboratories were used in this study. The trailers contain in:
ation for continuous monitoring of a number of contaminants, photochemical product'
meteorological parameters.
The pollutants measured included carbon monoxide, oxidant, ozone, nitno-exide.V
dioxide, total hydrocarbons, individual Ci to CIQ hydrocarbons, peroxjgoSjarl ni;
formaldehyde, acrolein and total aliphatic aldehydes. The meteorolog^aUlemertCs
monitored included temperature, relative humidity, wind speed and diiflttlWi,
violet radiation intensity. The most significant advance was the ' '" ~
determination of more than 100 individual hydrocarbons.
>gen
17 Key fora am Oocianmt Analysis (a) Descriptors
Photochemical Smog
Chemical and Meteorological Parameters
Hygro-thermograph
Carbon Monoxide
Oxicants
Ozone
Nitric Oxide
Nitrogen Dioxide
Total Hydrocarbons
Flame loaization
176 laemilierfape»Enoed Terra
Acrolein
Total Aliphatic
Tedlar Bags
NOx
Vertical Diffusi&O
Olefins
Peroxyacetyl Nitrtffel
Aldehydes :
Ultraviolet Radiatioi
Oxidation
T
17c
U DisMMtoaS
19 Secirity ClasslThis Repon)
UHCLASSIFIEO
21. No of Pages
/U.Secuity Class (Tins Pate)
UNCLASSIFIED
FORM MBX-eail-791
-------
SRL Project No. 1133
March 7, 1970
Table of contents
SUMMARY
1. INTBODUCTION Z. 1— i
Atmospheric Reaction Studies in
the Los Angeles Sea m 2. ANALYSIS OF DATA PEON 1968 LOS ANGELES BASIN SURVEY -————— 2— 1
Phase II 2.1 GENERAL APPNGACH TO THE ATMOSPHERIC SYSTEM - — 2 1
Data Analysis and 2.2 CONTAMINANT CONCENTRATIONS, DIURNAL PATTERNS AND
Methods Improvement
INTERRELATIONSHIPS — —- 2-3
2.3 FORMATION RATE OF NITROGEN DIOXIDE — 2—21
2.4 FORMATION OF OZONE AND OXIDANT 2-36
for
I 2.5 FORMATION OF ALDENYDES —--—— ———— 2—48
2.6 PHOTOCHEMICAL AGE OP THE ATMOSPHERE ——--— 2 —19
2.7 TRANSPO R Y OF POLIAflANTS — — . - . — 2-.53
Coordinating Research Council, Inc.
I 2.8 AIRBORNE ZCASUREMENTS — —————————— — 2—55
and
National Air Pollution Control Administration
3. IMPROVEMENT OF ANALYTICAL METHODS FOR ATMOSPHERIC
(Contract No. CPA 22-69—19)
POLLUIANTS - — — 3—1
3.1 OROAN IC NITROGEN COMPOUNDS — 3.4
by 3.2 HYDROCARBONS — 3..3
3.3 NITROGEN OXIDES ————— -‘———— 3—6
SCOTT RESEARCh LABORATORIES, INC. • 3.4 OXIDANTS AND OZONE — —.- 3..7
P. 0. BOX D-ll
Pitensteadvi lie, Pennsylvania 18949 3.5 MISCELLANEOUS 3—8
-------
H
SWQ4ARV
Table of Contents I
Thcserometric data collected by Scott in the Los Angeles Basin
4. DISCUSsION AND RECOI(M lDflxONs — — ——— ——————————— 4l during the Fall of 1968 was analyzed to study the relationships among the
chemical and meteorological parameters measured. Emphasis was placed on
5. REERZNCSS 5— 1 defining the formation rates of the products of atmospheric chemical
end photochemical reactions especially nitrogen dioxide and ozone. The
data analysis was also performed to aid in the design of the 1969 field
program.
The study wee concentrated on the empirical relationships which
connect the outputs of the atmospheric system, i.e. the photochemical prod .cts,
to the inputs represented by pollutant emieeions and ultraviolet radiation.
This approach offered the possibility of greatly facilitating the derivation
of relationships which could be used to evaluate the effects of controlling
the system inputs.
The concentrations of many contaminants foll.owed similar diurnal
patterns. Good correlation was found among carbon monoxide, acetylene and
total C 5 to Cie hydrocarbons. A close relationship was also found among methane,
ethane and propane, but theme hydrocarbons did not correlate well with total C6
to C 10 hydrocarbons. Different primary sources were indicated.
The concentrations of gasoline range olefins found in the atmosphere
before photochemical reactions started were only half of those expected based
on analyses of gasolines in use during the period.
— — The ratio of carbon monoxide to nitrogen oxides was sub]ect to
wide variation from day to day. Nowever, the hourly averages for the entire
period showed little change with time of day. The average carbon monoxice to
F nitrogen oxides ratio was 30 which was the same as reported for exhaust fron
-------
The oxidant appeared to increase at a faster rate ourinç, transport t tan • i c r
the air mass remained in the Huntington Park area.
The’eirborne measurements revealed nuch infornation about the thret-
aimensional aspects of the Los Angeles Basin. The vertical rrcf ties s io ’c-
that tee base of the temperature inversion was at an approxindte eltit -de
of 700 feet above sea level during the morning. The fresn inputs to in c
atnospoere were confined below this level. Oxidant was higner within inc
inversion tnan below it during the early morning despite the lo ier concen-
trations of reactants present. As a typical day progressed, o u.ant incraase
rapioly in the air below the inversion. The inversion base began to rise anca
the sea breeze started. The smog front was visible from aloft as it m e
towara t e 1 Monte site. By mid-afternoon the inversion either broke u:
completely or rose to an altitude above 1500 feet. ‘the composition of
pollutarts in the atmosphere was then quite uniform between ground levcl anc.
1500 to :soo feet above sea level.
The work dealing with improvement of methods applicable to atl’cs
pheric aralysis concentrated on techniques for atmospheric nitrogen con c .nc.s,
hydrocarz’ons, nitrogen oxides and oxidants Among the achievenents of tnis
effort were automation of the dcc’ ron capture chromatograph usec for cernxy-
acetyl ‘ttrate and development of an irçroved technique and calioration starc.aro
for use with it. The C i to C i i, hydrocarbon analysis was also iipro:e .
revisirg tie chronatograph flow systen, developing a purifier for the carr:cr
gas at .dertifying additional chromatograth c peaks T te stattlity of - i to-
carborc ann nitrogen oxides in Tedler nacs was founo to oe sat_sfactor for
use in airrorne work. Other instrumentation for chemical a-io neteoroloti;a
parameters was checked out, revisea or repaireo wnere nccesssr? snc. r’s e
operational for the start of the 1969 field aurvey.
I • INTRODUCTION
This final report describes the work performed on Phase II of
APRAC Project CA n t— i— GB, “Atmospheric Reaction Studies in the Los Angeles
Basin” (Contract No. CPA—22-69-l9). Phase II consisted of two major
activities: analysis of the aercmetric data collected by Scott during the
fall of 1968 and development of improved methods of chemical analysis
applicable to the 1969 field survey and other future aerometric studies.
The purpose of the data analysis was to study the relationships
among the chemical and meteorological parameters measured in the 1968 field
survey of the Los Angeles Basin 11 ’ 31 in order to define the formation rates
of various photochemical products, especially nitrogen dioxide, oxidant and
ozone in terms of these parameters. Additional relationshipa were studied
to shed lig tt on emissions sources, reaction times and the overall behavior
of the atmospheric system. The data analysis was also conducted to aid in
the design of the 1969 field program by providing a firmer basis for
evaluating the most important factors to be measured, the critical periods
during which measurementa should be made and the location of sampling points.
The methods development was undertaken to improve analytical
techniques which did not produce satisfactory results in the 1968 field
survey or devise alternate methods, to expand other procedures to sake them
capable of providing additional useful data, and to validate the performance
of the instrumentation and procedures employed in the field work. Emphasis
was placed on analysis of nitrogen compounds and individual Ci to C hydro-
carbons. Most of the work was performed in Scott’s Plunsteadville laboratory
using the same inatrumentation and equipment that had been located in the
1 — 1.
-------
1—2 2 — 1
scott mobile laboratory used to obtain data at the El Monte site in the 2. ANALYSIS OF DATA F M 1968 LOS ANGELES BASIN SURVEY
1968 las Angeles Basin aurvay. I
2.1 GENERAL APPsoACH TO THE ATICSPNERIC SYSTEM
The study described in this report was limited to en empirical
analysis of aerometric data collected by Scott in the 1968 survey of the
Los Angeles Basin. Since nearly all of the 1968 data were collected at two
fixed sites, the enalysis was limited to the sub—systems representing the
atmosphere in the imsediate vicinity of the ssmplsng points rather then
to the entire Los Angeles Basin atmosphere. The relatlonship between the
sub—systems at the two sites was also investigated. while no attempt was
made to model the Basin atmospheric system, a generalired concept of the
atmospheric system was desirable as an aid in searching for re lstionships
among the variables measured.
There are three distinct aspects which must be considered in
defining the changes in composition end concentration which take place in
an atmospheric system, namely input of pollutants into the system, chemical
reactions and physical processes within the system, and transport of reactants
and products out of the system. All variations in atmospheric concentrations
of the pollutants measured an the field survey were combined effects of
these three aspects, so that while the data analysis deals essentially with
atmospheric chemical reactions, the true rate of each reaction cannot he
determined without a knowledge of pollutant inputs and transport charecter—
istics.
The concentration of each of the literelly hundreds of individual
pollutants present in the atmosphere is constantly changing. Prsah material
is continually being added to the atmosphere but its composition and rate
-------
2—2
2—3
of addition are subject to variation as the relative strength of each
pollutant source follows its diurnal pattern. The rate of transport out of
the system is likewise changing as wind speed, inversion height and other
meteorological factors vary, Finally, the rats of each of the hundreds of
separate chemical reactions changes as concentrations, radiation intensity
and temperature change. When the additional complexities of vertical and
horizontal non—homogeneity are considered, the system becomes too chaotic
to be treated in its entirety. Therefore, the system must be segmented and
eiisplif ted.
Study was concentrated on the empirical relationships which connect
the outputs of the system, i.e. the pbotochemical products, to the inputs
represented by contaminant emissions and ultraviolet radiation. This
approach made it possible to bypass the kinetics of individual reactions end
disregard the involv nt of such intermediates am zwittariona and free
radicals. This approach offers the possibility of greatly facilitating the
derivation of relationships which could be used to evaluate the effects of
controlling the various system inputs.
It was decided that the data were best analyzed by following the
short term changes in the atmosphere rather than daily or aeaaonal avereges.
The latter identify the meteorological conditions which favor heavy smog
formation but tend to mask the effects of individual contaminants which are
of greater interest because they can be potentially controlled. The half
hourly averages presented in the data file were inadequate to define rates
of reaction during periods of rapid change, and it was necessary to refer to
the continuous strip charts to obtein the desired precision.
The overall strategy was thus to develop pertinent relationships
which defined the formation rate of each photochemical product and the extent
to which it accimsilated in the atmosphere. These individual relationships
could then be tied together into sub-models and models in future work.
2.2 CT)NThJIINMT NCENTRATIONS, DIURNAL PADIEKNS AND INTERRELATIONSHIPS
The pollutants present in the 1cm Angeles Basin atmosphere can be
divided into tic categories, those emanating directly from emission sources
and those formed through chemical raactiona occurring in the atmosphere.
This section considers those contaminants in the first category which were
measured in the 196$ field survey. They include carbon monoxide, individual
C i to C 1 0 h drocarbona, nitrogen oxides and aldehydes. The reaction products
such as ozone, oxidanta, and nitrogen dioxide and aldehydes produced by
atmospheric reactions are discussed in subsequent asctions.
The comprehensiveness of the aerometric data collected at the two
ground sites permits a detailed look into contaminant concentrations, their
day-to—day end within dayveriations, and their sources. While the data apply
only to the sits at which they were obtained, the relative remoteness of both
sites from known high essiemiona sources makes it reasonable to assume that the
atmosphere at each site was twice! of that in a relatively wide area around
the site.
The sources of contaminants present in the air during the early
morning period are of considerable interest because it is these contaminants
which are trapped by inversions and irradiated to produce the peak concen-
trations of orona and other photochemical producte noted in the afternoon.
On the other hand, previous studies have shown that afternoon emissions
cause only a slight rise in contaminant concentrations and usually play a
-------
2-4 - -
minor role in photochemacal processes.
CA 0N ICEJIDE
Carbon “ de dcae not par icipeta in’photdcheaical r& tiàns.
For this reason end because it has been estimated that as much as 95% of
the carbñ_n monoxide in the Los Angeles Basin atmosphere comes from automotive
exhaust emissions, carbon monoxide appeared potentially valuable as an
a spberic tracer for auto exhaust.
In order to tast the above sssuszption, the relationship between
carbon monoxide and acetylene, a photocbenically mart hydrocarbon also
believed to be contributed almost exclusively by auto exhaust, was invent—
gated using data from the huntington Park site. Linear regression analysis
of 97 eimultaneoua measurements of carbon monoxide end acetylene concentrations
yielded the following relationship:
(ppm) — 113 C 5 B 5 (Ppm) — 0.06
Correlation Coefficient — 0.967
Standard Error of Estimate — 1.98 pps
The carbon monoxide to acetylene ratio of 113 is close to the value of 120
generally accepted for exhaust from pre—1966 cars (based on 3% carbon monoxide
and 250 ppm acetylene).
The high correlation between carbon monoxide and acetylene concen-
trations and the fact that they were present Sn essontially the sees ratio
as found in auto exhaust eupporte the sseumption that esbaus ie the major
source ci these contaminants. However, the quantities of carbon monoxide
and acetylene esattad to the atmosphere from other combustion processes of
fuels must be detarnisod before the relative contributions of vehicular end
2—5
other coerces can be firmly eetebliehed.
- - En the oorrolotton otudy of carbon monoxide and acetylene the
correlation was not nearly no good when th e carbon monoxide values wâu
obteined by interpolating the half hour svermgme to the tine that the acetylene
n le was oollectad. The good correlation, found above resulted from reading
the carbon sonoxida concentrstion from tbs strip chart records. Subsequent
studies of other parameters confirmed that when one parameter was determined
using an instantaneous sample, much better correlation vas obtained when the
second parameter was reed sialtaneouely. Thie finding points out that
values for comtinuotmly recorded paraweters should he reported’ for intervals -
shorter than the thirty minutes used for the 1968 dsta.
The concentration of carbon monoxide at Huntington Psrk ranged
from <1 to 37 m. The highest concentrations usually occurred between 0630
and 0730 PST although high overnight concentrations were frequently found in
Not. -
The highest average concentration of carbon monoxide over one day
The averagen for individual days of the weak were:
Sunday 4.4
monday 3.3
Tuesday 4.0
Wednesday 5.2
Thursday 4.7
Friday 2.8
Saturday 22
Avg. 4.2 0 0.79
was 12 .
I ’
-------
-6
2—7
Pr evious monitoring ptogr had few %hat carbon monoxide was
1 ir ts±iô 4ay fld than o’ zce s of th lower traffic
t ThO pieooiat d tào noi utsi ’h In a Ôre srno janif iáãt
dtffereaâ ;in the avoragoo for different days of the week. This means either
dwt woekyend traffic wolt in tho iton Park area was not signif i—
1e ier tlan thAi on wookdays or that daily carbon sonoxfde concentrations
ware rere a functien of meteorological conditions than of traffic density.
The higbeot carbon monoxide concentration measured at El Monte
umo 14 Wa. This noxisesm was found within one-half hour of the period high
- -
of 37 ñm at Huntington Perk. The carbon monoxide at El Monte was generally
ft c one—third to ono—heif that at Huntington Park during the corning (0900 to
flCO PSTI. The El Monte concentration decreased steadily during this period
then siwed a sharp mpsunge whoa the eoa brooso brought in the pollutod
air the southwest. This eooend . int 1 which usually occurred between
1200 to 1400 PST, seldom aceoded the early morning maximum for that day.
t e are insufficient data points for El Monte to develop additional averages
or trends.
w CA ms
Bydrocarbono era major participate in the chemical reactions
ooc*arieg in the we Angolee Basin e phere. The availability of data for
approximately 100 individual hydrocarbono covering the range from Ci to Cie,
nearly all of which are pbotochenically reective to oone degree, provided
sufficient information for a study of sources end reaction rates • This
section i.e limited to e study of composition end eourcee while the investigation
of reaction gates is coveted in e later seotism.
thrcategre *iSe colais were used for hydrocarbons, one for
Ci to C ., and the other for C 5 to C 10 . The correlation between n—butane
which was determined with the first column, and n-pegtane, which was determined
wl tti the second column, was investigated to verify that the data from the two
columns were compatible. The relationship found using 101 data points from
mtntinflon Park was,
n—Cemio (ppb) — 2.18 n—Ceflia (ppb) + 10.9
Correlation Coefficient — 0.950
Standard Error of Estimate — 16.2 ppb
The data from the two columns are thus reasonably compatible. The concen-
trations of n-butane era comparable to those found in downtcwn Los Angeles
in 1967 end the rstio of pentane to butane is in agreement with results
reported by Stephens for samplee collected in the Los Angelas Basin and
other sections of California. The intercept probably represents background
butans emanating from a second source containing etch lees pentane then the
primary eource.
The data for total concentretion of C to Cie hydrocarbons are less
accurate at low levels than at high levels because ths peaks of some minor
coaponente ware too email to be counted by the integrator and occasional
epurtoua peaks were counted. It was hypothaeixsd that the eum of a few
hydrocarbons, normally present in high concentration end representative of
the various carbon numbers and classes, would provide a better seanure of
the total C 5 to Cio hydrocarbons then would the aim of all the detected
compounds. The six hydrocarbons selected were n—pentane, 2—asthylpentene,
2—methylbexane, methylcyclohexsne, tolusne and s + p—xylcne. The relationship
obtained was,
r
-------
2—8
2—9
Total HC (ppb) — 2.06 E 6HC (ppb) + 8.9
Cor! ela ion Coetfic&ent — 0.993
Standard Error of Estimate — 37.2 ppb
A study of she outliers showed that all were the result of either spurious
or uncounted peaks. This demonstrited that the sum of the six hydrocarbons
was a slightly more reliable measure of total Cs to Cio hydrocarbons than
was the sum of all the compounds listed in the data printouts.
Individual correlation of acetylene and carbon monoxide with the
U..-
total Cg to Cie hydrocarbons end with the sum of the above six hydrocarbons
gave the following relationahipsz
C all 3 (ppb) a 0.188 Total HC (ppb) + 2.81
C 5 11 2 (ppb) — 0.581 Z 6 SC (ppb) + 3.98
co (ppm) — 22.0 Total MC (ppm) — 0.09
0 ( ) a 68.0 £ 6 MC (ppm) + 0.04
The high correlation between carbon monoxide and the total Co to C io
hydrocarbons does out necessarily prove that they have e comecn source, nemely
auto exhaust. First, the degree of correlation attributable to the atmospheric
dispersion rate, which tends to cause all contaminants to increase end decrease
at the same rate, must be factored out. Second, detailed data on exhaust
composition must be studied to determine if the ratios of carbon monoxide to
the various hydrocarbons with low photochemical reactivity are similar for
eabauet and the atmosphere. Adequate date era sot available so no conclusions
can be drawn. However, the day to day variations in atmospheric cosposition
noted both with regard to hydrocarbon class and molecular weight range suggest
one majdr turce of Cs to C, o hydrocarbons. If vehicular exhaust
were the only major contributor, the hydrocarbon composition woulc not change
with time axcept through photochemical reactions. This is true because the
large nurtet of vehiclss ointributing to each sample would be expected to
morealise any variations due to vehicle or fuel differences.
The concentrations of methane, ethena and prcpane did not exhibit
nearly as good correlation with the total Ca to Cio hydrocarbons as was
Lobed for acetylene. tn Figures 2—1 through 2—4, the data for these four
hydrocarbons are plotted against the sum of the 6 hydrocarbons, which were
found to represent the total Ce to C 10 hydrocarbons. The much larger scatter
for methane, cthane and propane than for acetylene is quite evident. The
linear relationship between methane and athens and propane is illustrated
in Figures 2—5 and 2—6. This suggests a con scurce for these three
contaminants. The methane concentration of approximately 1.4 ppm known to
be presant in the natural atmospheric beckground is evident from the intercepts -
of the methane plots.
These findings are in line with those from previous studies it -
has been proposed by Stephens ‘ 1 that these light peraffins result primarily
from natural gas leaks end from evaporative processes. Still, the concen-
trations of propane were higher than he could readily explain because the
atmospheric propane concentration was approximately half the ethane while in
natural gas it is only one—fifth the ethana. It is apparent that additional
source data are needed to clear up the rjstery of the exce s propane. However,
the relatively low reactivity of propane minimizea the importance of solving
Correlation
Coefficient
0.936
0.936
0.961
0.963
Standard
Error of
Eatinata
23.2 ppb
23.1 ppb
2.06 ppm
2.03 pp.
-------
if,
• -. E.; •
I,
/
I.. •
.%I
. ,..
•1.
a.
Figure 2—2
Sum of 6 Papr.u.ntative C - Cio Hydrocarbons
vs
Metham
Humtingtcn Park
I,
— .
4
I..’ - ..
,
/ •
.••p .
-S..
I I I
I- I I I
1000 2000 3000 4000
I I I I
5000 6000 7000 8000
Methane (ppb)
40 1.20 200 210
_______________ £o.tll.TIs (ppb)
440 SZO 600
S
00• -
-------
,
/
,
,
Is__s
-
I
/
/
/
• .
‘7
—.7..
‘I
--p -
— . . ,- ‘ .
- -,• • ,. ;
,
/
Ithane
Huntington
I I I I I I V
40 130 200 280 360 440 520 600
Ethan (p b)
- -
‘ 7’
• . .i•.•
so
100
I I I
150 200 250
Propan. (ppb)
300
-I
‘ 7:
/
7
,
I-.
U)
- -I
Ptq rs2—3
Sta of 6 1 pr...ntativs C. - Hdroc b a
vs
‘0
0
0
0
i0
500
LOS-
Park
—t
2-4
S of 6 JIsprsantati e C - C 1 flydrocaxbon .
vs
Prspan.
Siatthqton Pa*
t .
-------
9000
7000 -
ooo-
e0o-
)000
2000
1000
- _____
I :
;1 I
I.
‘S
I.
.
— I.
/
/
S / •
I ,
- I •7 ••
5 1
-t -z
- 7
at. --1-
a.
1’
figure 2—5
Netbans vs Propan.
Huntington Park
0000
7000 -
6000
5000 - -
a
5
3000 - -
2000 - -
boo - -
/
‘a
5/
2’..
Figure 2—6
Methane vs Ethan.
Huntington Park
./
/
5/
• •/
2—15
/
/
I.
,.
/
I I I
100 200 300 400
/
500 600
I I I I I
100 200 300 400 500 600
$th. (pb)
-------
2-1*
2—if
this problem.
The law concentrations of fuel rsnge-clafinp found in early nornii
- saisplts taken bets pbotocheaicfl reactions 2 Et1dtedhy solar -
radiation are difficult to explain. Samples of gasolines in use during the
test ps Sod were analyzed for individual ) drocarjions as part of a separate
- - - - , --- - ..1_
program. The average olefin content reportedby the seven articipeting
laboratories was 7.3 wt.%, naarly 90% of which was made ug of Ce and Cs
hydrocarbons. The concentration of 2-nethyl-2-butene reported was 1.1 wt. C
for both premium and regular fuel • The concentrations of isopentane and
i gentkte he and U% gnlar fuel wire 7.6% and 4.9%
respectively. The irasored isopentane to n-pantone ratio of 1.7 wee close
to that calculated from the fuel analyses. The expected 2—nethyl—2-butene
to s-pentana ratio was 0.22, but the actual early corning ratios found
ranged from 0.08 to 0.12. Other fuel range olefins were also found is rich
lOwer concentrations than anticipated. Several possible explanations for
this arei 1. Internally bonded fuel range olef ins are present in exhaust
at lever relative concentrations compared to the fuel than are saturates 1
2. Internally bonded olaf ins react to a significant degree in the absence
of ultraviolet radiation and photochemical reactions, 3. There is a
significant source of Ca saturates which contains little or no C 5 olefine.
In any event the concentrations of fuel range olefina present at the start
of photochesical activity were considerably less thn expected. The
importance of thie finding is shown an eubeequsnt sections which deal eith
chemical reaction rates.
HITlCGflI OXIDES
the photochenical reactions in the a spbere of the Los Angeles Basin.
These nitrogen oxides are produced in all combustion processes both vehicular
- izat cases the portion -mff. edsa siithcjkiAe is ii i ‘ -
excess of 90%. The data for Huntington Park show that the nitric oxide
represented up to 00% of the total nitrogen oxides in eemples collected before
- uz&iss.
The ratio of carbon conoxids to nitrogen oxides is of interest in
estimating source contributions and in dstexnining the extent to which nitrogen
oxides are lost frost the stcosphere through reaction. The correlation found
between carbon rcnoxide aid total nitrogen o dei, nitric oxide Aind nitrogen
dioxide for 73 Huntington Perk data points before 0900 Fr I along with the
corresponding correlation coefficients end standard errors of estimate war
C (ppm) - 37.7 NO (ppm) - 1.43
Ct (ppm) — 39.4 HO (ppm) + 1.86
CO (ppm) 51.7 N0 (ppm) + 3.30
The poor correlation and high standard error of estimate indicate
that at the ties photochea4cal reactions are just starting, the ratio between
carbon conoxide and nitrogen oxides is subject to wide variation.
Another factor to be studied is the variation an CU/NOx ratio with
time of day. Hourly averages for the carbon amnoxida to nitrogen oxides ratio
for 194 Huntington Park data points are plotted in Figure 2-7. It can be
seen that the ratio passed through a sinisun around 0830 PS?. This was
caused by the corning nitrogen oxidsa peak occurring consistently later then
the carbon nonoride peak. The cause is not readily apparent. Changes in the
Correlation
Coefficient
0.80
0.74
0.41
Standard
Error of
Estimate
3.81 ppm
4.32 ppm
5.83 ppm
Both nitric oxide end nitrogen dioxide are of major importance to
-------
2-18
- o e XQQQQ
. N C C v n
m N m N N
7 ? i i i i i
o
‘- —4 C C
I & 00.
C d C
C
14 00
I. NW U
D 0 > 5 m
D lx 50
0
o
o
0
- -.4
r
I n
0
0
— 0 .
-
0
0
• 0
0
0
0 ’
I -.
C l
(fl/NO 5 ratio in auto exhaust as driving patterns change during the corning
tush hour is a possible cause. However, the day to day variations in ratio
for ea4 tan period.j r(consi6erebly greater thahthe time *riat.tons
within day. This eujg j Sat thbltipla sourceS are the primary cause of tht
poor correlation between carbon senoxide and nitrogen oxides. The relatively
stant ratio from LOQO 1ST through 1330 PaT is in strong diaegreeeient with
laboratory tharter stedies where the nit ogan oxides are rapidly depleted as
pbotockemical reactions proceed. A possible explanation is that loss of nitrogen
oxides through atmospheric reactions is substantially slower than in laboratory
- syste m end/or secondary sources emitting no carbon sononde replace the
nitrogen oxides as rapidly as they are reseved by reaction.
The average carbon monoxide to nitrogen oxides ratio for the 194 points
shove was spproxiisetsly 30. The accuracy of carbon monoxide concentretions
below S __ set as good as at higher values because of the zero drift
experienced with the non-dispersive infrared enalyrer. In order to verify the
validity of the overall average ratio, those data points where the carbon
monoxide was equal td or greater than S ppm and the nitrogen oxides were
equal to or greater than 15 pphm ware etudied. The distribution of the 58
points smeting these criteria was skewed to the high side em the median was
determined rather than the mean. The median ratio was 32 with 70% of the 58
ratios between 20 end 40. Thus, part of the scatter in the 194 data points
was due to poor accuracy at low concentrations. However the enrage was not
distorted.
The carbon monoxide to nitrogen oxides ratio for exhaust from uncon—
trolleci (pre-1946) cars wee found to be 30 based on road tests at en average
route speed of 25 wiles per hour in urban areas (e) • The growing population
.3
0
- S
-l
- ;. - . --; t-; -’- -
- - -- C • - - - ; ..
-------
2 —20
2 -2L
of device equipped care, which t eubetantielly less carbon monoxide but
at least as moch nitrogen oxides as uncontrolled care, would be expected to
ie s6 to a ratio lower than 30. stationary aouroe etflseiona. which are bilieved
to oontaTin more nitrogen oxides than carbon môncsid, should alSo reduce the
atmospheric carbon ecnoxide to nitrogen oxides ratio. -
.fle fact that the obeerved ratio was higher then expected might be
explained by a loss of nitrogen oxides through chemical reaction, however the
ratio did not decrease during the middle of the day when chemical reaction
rates were greatest. It is more likely that area driving patterns resulted
in lower vehicular nitrogen oxides emissions then the 25 mile per hour road
teat. The average carbon monoxide to nitrogen oxides ratio found for both
the C rce and El Monte eitee in the 1969 survey was between 20 and 25.
This ratio appears to reflect the influence of vehicles with carbon monoxide
ountrole.
Aldehydes are emitted by both vehicular and stationary sources,
end they are also products of photochenical reactions. ta for formaldehyde,
acrolein and total aliphatic aldehydes were oollected an the field program.
The relationships among these three paramatera were investigated
by Linear regression analysis. The results for 179 data points at Huntington
Park were:
Form (ppha) • 0.67 Tot A Id (pphs) — 0.19
Form tpphm) • 2.57 Acr (ppiun) + 1.99
of the date at the very low levels normally present. The formaldehyde to
acrolein rat to approached 10 at the sore reliable higher concentrations.
Attuepte to correlate aldehydes with carbon monoi jde-ant nitrojen
oxides during venous time periods jitoved fruitless as the coiteia ioA never
exceeded 0.50. Aldehydes seemed to vary independently of other oontaninants
even before photochemicsl reactions were initiated.
The average and maximum concentrations of various aldahydes found
at the two sites were:
Huntington
park
El
Monte
Max
Ave
Max
Ave
Aliphatic Aldehydee
17.3
4.5
14.8
5.1
Formaldehyde
13.6
2.8
9.0
2.5
Acrolein
1.0
0.3
0.8
0.2
The rate at which nitric oxide is converted to nitrogen dioxide
is of greet importance not only because the nitrogen dioxide formed is in
itself ob)ectioneble, but to a greeter extent because other noxious nog
products such as osone and peroxyacyl nitrates are formed only after most
of the nitric oxide has been converted to nitrogen dioxide.
— — — - - - WhiiCIäã of the nitrogen dioxide present in the atmosphere is
emitted as such from its source or is formed by the thermal reaction of
nitric oxide end oxygen, the major portion present during g apisodes is
Concentration (pphm )
2.3 FORMATION mATE OP NITI GEN DrOXIDE
BAQcG30UND
Correlation
coefficient
0.88
0.40
Standard
Error of
Estimate
1.02 pphe
1.98 pphs
hs poor correlation with acrolain appears to result f row the limited accuracy
-------
2—22
produced by photochesical reactions involving hydrocarbons. nitric oxide.
nitrogen dioxide arid ultraviolet radiation. Laboratory experiments have
ehowre- that..the initial nitic oxide conversion rate is dependent on the -,
- 4 !’1 Ion nd .iesctivity of the hydrocarbons • tha concentrations of
nitric oxide and nitrogen dioxids, the ratio of hydrocarbon to nitrogen
oxides, ultraviolet radiation i ntensity, temperature and humidity. Attempting
to dei?ive an expression reLating the rate to all of these parameters in an
initial step appeared unwise because their effects, except for ultraviolet
radiation, have been shown to be non-linear Therefore, eech parameter
was treated separately.
ATIrSPUERIC R ES OF 1 s FOWATION
The study of NO 3 formation rates from the data collected in the
1968 field survey was limited by the number of data points available for
nitrogen oxides concentrations and ultraviolet radiation intensity. Failure
to obtain delivery of the continuous nitric oxide — nitrogen dioxide instrument
left only hourly readings from manual analyses for thase two contszainsflts.
while the availability of only one ultraviolet instrument for the ground
sites reduced the number of data points recorded for radiation intensity.
Analyses at the Huntington Park site were made from 0530 pr
until early afternoon, and thus completely covered the 0700 to 0930 period
during which the bulk of the photochssical conversion of nitric oxide to
nitrogen dioxide occurred. The El Monte site data was less useful because
the conversion was usually well under way before the first analysis in the
enrning was made. The atmospheric formation rates of nitrogen dioxide were
calculated from the differences between successive hourly measurements of
nitric oxide and nitrogen dioxide. In making the calculation it was assumed
.4
that the change in rate between the two successive points was linear and that
the total nitrogen oxides c1 anged only by asphsric dilution. Thsse
assusvtions wet-a nacSsary-bâause f thi limited number of data points
available and the lack of data for syetem input rates, dispersion rates and
reaction rates for a spher1.c nitrates. The relatively nell variation
with time noted for tie co, x ratio in Section 2.2 of this report appears
to validate the assimiption that atmospheric dilution was the ma)Or cause
of the variations in total nitrogen oxides concentration.
The average observed rate of nitrogen dioxide formation at Huntington
Park was 0.7 ppb/mth. The retes ranged from <0.5 ppb/min on low pollution
days to a maximum of 2.2 ppb/min.
EFFECT OF NITRIC OXIDE cowCarTRsnoN
A regression enalysis of nitrogen dioxide formation rate and nitric
oxide concentration employing 46 data points from Huntington Park gave the
following expression for the rate given in ppb/sxn.
d p ) — 9.59 x l0 (No) 1 ’ 3 °
Correlation Coefficient = 0.77
Standard Error of Estimate = 0.96 ppb sin’
Standard Error of Corrsl. Coeff. — 0.17
The NO coefficient of 1.35 was higher then the values found for
most hydrocarbons in laboratory studies at and General Ittors
Research ). Typical laboratory values for the nitric oxide concentration
range found in the a sphere included 0.4 for propylene, 0.6 for meeitylene
end -0.3 for n—henna. The correlation coefficient is not especially high
but the standard error of the nitric oxide coefficient was only 0.17, so a
-------
consid rabl. difference between laboratory and athospheric ayst is indicated. I
This relationship is useful Sn determining the value of controlling
ox4des cf nitrogen emissions from motor vehicles and other sources; 0nfo -
- S. - •-. • . 1
tunately, en additional evidence can be gained from the 1968 data alone. A
study of other factors which say have influenced the observed relationship
points out the edditional information required to reach a Liner conclusion.
The addition of fresh emissions to the atmosphere has a greater effect on days
when atmeapheric concentrations are low. The emission rates are relatively
constant from day to day (excluding weekends) while the dispersion rates vary
considerably because of varying meteorological conditions. Ibis emsns thst
the air near ground level will contain a greater proportion of fresh emissions
on low concentration days. Since the major sources of nitrogen oxides emissions
contain a iamb higher proportion 180 to 95%) of nitric oxide than nitrogen
dioxide, the addition of these emissions can greatly reduce the observed
epperent rete of nitrogen dioxide formation. This effect is greatest late
in the conversion period when the contaminant concentrations are almost
always lover than at the start of conversion end the diaparity between the
relative arints of nitric oxide Sn the system inputs aid the atmosphere is
gr eatest.
An apparent negative conversion rate was noted on several occasions
which illustrated the effect of unusual amounts of new emissions being added
to the system. On October 16 the nitrogen dioxide decreased substantially
aid the nithic oxide increased from 0900 to 1005 PSI. The analyst, noting
this unempected occurrence, went outdoors and found thst a large slow-moving
truck was being used to paint lines on the ad acant street. On amother
occasion when a negative rsts was noted • spikes were found on the carbon
monoxide trace during the samp Ling jeriod. suggesting that a vehicle had
been left running near the sampling point.
Other factors whidrcl anged with t3sie, Such as bfdrorbon reactivity -
end ultraviolet irradiation, mast also be exsatined to determine their affect
at tines when the nitric oxide conceetratioe tended to be high as opposed
to wham it was low. This should be losaible using the data obtained in the
1969 survey.
Line? or HYD CARD0N WNCER IRATION 52W sEACTIV ITY
The rates at which nitric oxide is converted to nitrogen dioxide
in the grasance of a large ra er of individual hydrocarbons have been
determined by Glasson and Tuesday’° • Each hydrocarbon was irrsdiated at
1 ppm with 0.38 ppm nitric oxide end 0.02 ppm nitrogen dioxide. The initial
rates of nitrogen dioxide formation were cosçuted from the time required to
convert half of the nitric oxide to nitrogen dioxide. itional experinenta
involving the irradiation of three gasolines under ths same conditions
showed that good agreement was obtained between the measured conversion
rete and that calculated from the sum of tha products of the mole fraction
of each individual hydrocarbon saltiplied by its rate of conversion as found —
in the previous tests with single hydrocarbon.
TM seam type of calculation used by Glasson and Tuasday was epplied
to a nunter of enalyses for individual Ci to a , 0 hydrocarbons made during the
0700 to 0930 PS? period. and tha results were compared to observed rates of
nitrogen dioxids fo tion. TM rate of nitrogen dioxide formation caused
by each hydrocarbon was calculated as the product of its concentration (pps)
sulfiplied by the assigned nitrogen dioxide formation rate (ppb 80 5 /ninjppm
hydrocarbon). Some hydrocarbons ware grouped into classes having similar
I —
-------
2— 27
-a’- —3- qce ’ a via - - a
fIJ a. .a gjj d.u4 j
o
• aiL J Onwoon au,-..Io nua,mua mu
0 .iDa 00 uDqa, NunauI..u usua l. •
mu a c’a aq..uoau,. .eo ’a coo’s •
— uiu 1 0000,.. Con,. —,
I I ’ ..-u 0I nO I - a r - c a,
ii usr flu-io u ,. u —or..., encu., u-u
o cgt Oem..u,.. a,unvu ...a, I-
8 3 1 11°.91’ °0- 00
Sm uscousu’u.u -ueQsu uflouso 0
j. at, d4s.. Ouuur
a, F- n .,.., u
a. —
0
I ,
C ai.u.u.ui ‘aC unufl)uq C.tuusqu a,..uu.,n .
us . oaei v.uutuuusul. I-u- l u -u n uuu.eu’a u,
a, S US. ] ..I.-unnNJ’a nt — ac One,. —.
S — 0000u-flfl 000 -I QQ .i ui
-.au. . . - . .
C. - a,
2 °
X ; 00CC uflrur.q ru
Lu tuqI ster-nu’uo a,a,a —
00 0 cat uann.a,u- runo cnn ,- c
051 0000,n 0000 0000 .fl
PM u _ - u - . - - - - - -
-; U
3 WUia,usupr. uuuun a, uS-.no 0
H “ ::: :
C—u us ..oIsi f in e u sSue aSm m n.cu’a —
C Cu a C%J -u..uou-la,un u,uu.-.-u, onvumu a,
CI —u umØ 0000..in 000 - I Ococ .
- nut - u - -. . .. -
u— U
-I
u — i CouPus •o,-’a u- i
uu usu o-u.a,o. no-un a,a,.,... a,
o 551 on..u.cq nuuugufl 0- ma
831 °.9°.9’ri 0,9 A 0•0 0 9
a.
a.
.uM slauu’.mnn 000 no
>3 fine,-. i - wa , u-en
PU 00000 ,na, . ,u.,
— Ca
mc
M m
•
55 ?°
M M
em S C m c
c m cu. t.5.u
SC a, —
5u4 5 p -u
U frm ac t _u SM
h H i -
uAt fLu fl ta
raactivitiee to expedite the calculation. The average rates for each clans
e calculated from the average rater reactivitiea for that class in the three
!°‘ ! C J duby Ins and Tuesday. The gasoline compositions were uZuomparec
to other published data and the distribution within each cuss was found to
be typical. Methane and acetylene were assigned zero reactivity rates. The
atthaptu.eric data for terminal butenes was not reliabla because of baseline
shifts in the chramstograms so a figure of one—half the propylene concentration
was assigned, based on exhaust data reported by Papa (10) and atmospheric data
reported by Stephens and Gordon et al .
- The chromatographic peaks for so of the fuel range clef ins were
obscured by the much larger saturate peaks. Zn addition, sosia peak areas
ware below the minimius level counted by the integrator. For this reason the
olefina clearly separated were calculated manually by direct measurement of
the chromatograms, and the total found in this manner was adjusted for those
olefins which were obscured by the saturates, ma correction was based on
published fuel analyaea. This calculation was conaidared valid because the
major fuel range olefins, such as 2-methyl-2-butene and trana—2-pentene, were
clearly separated.
TM calculated nitrogen dioxide formation rates for nine days are
shown in Table 2-1. The calculated rates are compared with obsarved values
in Table 2—2. The two agree within a factor of 2. Differences can be
attrIbuted to affects of factors such as hydrocarbon to nitrogen oxides ratio
and ultraviolet radiation intensity. This indicatea that the atmospheric
formation rate of nitrogen dioxide can be accounted for by photochamacal
reactions involving hydrocarbons, nitrogen oxides and ultraviolet radiation
I I
2
f t
N
0
0
0
I
ft
a
a
a ,
a
a
U
a.
U
Lu
0
a
a.
a
S
S
t.
I
e
c
a
j
-------
lab Ia 2-1
Continued
Nitrogen Dioxide Formation Rates
Calculated free Atioepheric Rnalyuw
Oct. 22, 1968 0600 78? Oct. 23, 1968 0852 79? OcE.’28, 1969 0735 IS?
Calculated Calculated - Ca lculated
Reactivity Rate of NO 3 Rate of O Rate of IC 3
teepound Factor Conc. Formation S ¶‘otal Conc. Foenat .on I total Conc. Formation I total
or Claus ( pt/nin/ HCJ J 1p /ot L Rate jp j L och/sin) Rate ( thfliSn ) ‘ Rate
Btna 0.25 .406 .1015 4.6 .572 .1430 5.7 .105 0263 41
PreAeea 0.39 .166 .0647 3.1 .283 .1104 4.4 .125 .0189 7.6.
zutan 0.62 .034 .0211 1.1 .062 çojas 1.5 • .015 0093 l.5
n—$ftsne 0.75 .165 .1237 5.9 .254 t l905 7.6 .072 .0446 7.0
C 5 nd Higher Sat. 0.93 .561. .5225 .9 4M AL. ! .159 .UZ2 . .
Total Saturates 1.332 Th5i! 35.6 2.009 1,2624 50.5 .2769 42.2 ’
1 I .. .
rthylene 1.70 .1390 .2360 11.2 .1610 . .2735 ‘ 10.9 .0240 .0408 6.4,’
Propylene C t—Sutenee 3.50 .0475 .2365 11.2 .0750 .2825 lOb S . .0465 .1629 25.6
,‘ed Higher OLefins 6.90 .0413 .3680 17.5 .0147 4 ft jtj .0033 .0294
Tota l ‘Olefine .2476 .6405 39.9 5ThI .6669 26.7 17S .2330 36.3
9*jene 0.33 .0321 .0106 0.5 .0501 .0165 0.7 .0111 .0037 0.8
to iuene 1.30 .0764 .0992 4.7 .1031 .1346 5.4 .0167. .0243 3.1’
Cd end Higher Arcs. 1.50 .1292 .3230 A!Jj .1674 .J1!! IL! .I2 ! 16 1.:’
1&el Arosatica .2377 .4328 20.5 .3209 5696 22.6 .0712 .1315
Total 1.8175 2.1068 100.0 2.5806 2.4968 100.0 .6210 .6414 10019.
Glaeeon and Tuesday 1 1 for eten with 1 ppo hydrocerbon, 0.38 ppm NO, 0.02 ppm HOC
-------
F * -;;;. t
! s- ,I
C—.0t—fla .‘001n on_cm c
a I
oooo-.r. oo 8 — coo.. u
—1 —j%42q9
Li !° —. .°‘l’
pal iqn..m..&n .ai .,4...
Nfl .I It
0.
U I
8
9A* k
• -
ocr - I -
C P uaI oe..r.r.m ttN rn _ cnn
00 5 API WI-MWIa0 ri’a..IO S tal i n
Let 0 .COm
H If
H .
n.IOt ; flulI0 a 0.a
9
0
C c...c,4 nma.mm, CAb S I-inON
s -0501 wfl..nrnw Cl-c. OOAal
C A c..J Ocoor- a l c m o s crime
U .I 000000 0000 0000
4. ._. omoim .raflq A
U P A l nonner- aoo _csnn o
o p. n o.-.r.n 0_cop. 0 0_cr.
LEt ‘ °.°°. ‘ °A9° o o o o
C M
— N
F - 0
0
I
F.
F . -C
I
a
-C
N
a
A
3-
C
ur..M A ‘fl
0
g
2—30
Table 2—2
•
Sanr la
Date flaa(gSr)
oct. 17 0847
Oct.. 18 0830
Calculated and Observed Nitrogen Dioxsde
flornation Rates at H uitington Park
NQp Foreattt Reactive
Calculated a,aervad Hydrocarbons
C b abr ) ( t azC ) (p cj
- 0.48 0.9 0.45
0.57 1.1 0.50
Nitrogen
Oxides
(j,st)
0.22
0.24
Ratio
HCj 0 1
2 O
2.1
oct. 21 0750
1.18 0.6 1.01
0.23
4.6
oct. 22 0800
2.11 1.0 1.82
0.30
6.1
oct. 23 0852
Oct. 28 0735
2.50 1.6 2.58
0.64 0.6 0.62
0.51
0.20
5.1
3.1
oct. 3* 0647
0.23 <0.2 0.19
0.07
2.7
Nov. 5 0745
1.31
0.8 1.01
0.42
2.6
-
4 . mmn ..n
za
00000
r a
-------
- 1 - 3 2
at the levels present in the atan J re . This does not rule out the
1 cseibility that ether macbaniesm say contribute to nitrogen dioxide
€ , a - -
r. c :b - -
The relitive cântribution of each hydrocarbon class to nitrogen
dioxide formation as determined from the fraction it contributed to the
calcul%tad nitrogen dioxide forsation rate is sofl interesting. Tale J—l
shows that the foal range olefins accounted for 0 to 20% of the total.
Calculations of edditional chromatcgrsnw, snclitng those given in Table 2-4,
showed that these olefins contributed from 15 to 20% at the start of
pbotoc)4aicsi nitric oxide conversion (0700 1ST) end from 0 to 5% at the -
time nitric oxide conversion vax nearly complete and ozone formation began.
The cracked olefins (ethylene, propylene and terminal butenes) contributed-
approximately 25% throughout the period, the eromatics contributed 20 to 25%
and the saturates 40 to 50%.
The proportions of the calculated nitrogen dioxide formation rate
essigned to eat oIias ii i Tible 2-1 are substantially diE ferejtt than those
calculated using soat reactivity scales. This difference is caused by the
fact that in using the reactivity ecelee, it in assused that reactivity based
on initial reaction rates is applicable to the entire reaction period. In
reality the very reactive hydrocarbons disappear rapidly as they are consumed
in the reactinna so that their contribution likewise decreases. On the other
hand, the lean reactive hydrocarbons disappear slowly end their overall
reactivity remains relatively constant throughout the reaction period. A sore
realistic reactivity sc ls would be based on the average rate over the entire
atsoapheric reaction period (at least several hours) end not just the initial
rate which say be valid for only tha first 15 to 30 minutes of reaction.
SE lECT 0? yo canow tO Nfl’ECG OXIDES RATIO
The work of Glasson and Tuesdsy(e showed that increasing the
hydrocarbon to nitrogetoxide rato y inoráeing the hydrocarbon ooncen-
- - -1.. •- -
tration above 1 ppm ithile holding the nitric oxide constant at 0.38 ppm
always resulted in an increased nitrogen dioxide formation rate, but the
percent incieaee in the rate vat last than the percent increase in hydrocarbon.
This slow rate increase wax sost noticeable for saturates and aroma tics.
Increasing bexane from 1 to 5 ppm increased the rate by 73%, and increasing
o—xylene from 1 to 5 ppm increased the rate by 62%. A similar 400% increase
in propylene increased the rate ‘w 200%. Decreasing the ratio by lowering
the hydrocarbon concentrations by 50% (fron 1.0 to 0.5 ppm) resulted in a
formation rate decrease of 31% for hexane, 31% for o—xylene and 40% for
propylene. The results obtained by changing the hydrocarbon to nitrogen
oxides ratio by varying tha nitric oxide and keeping the hydrocarbon concen-
tration at 1 ppm shoved similar trends.
These laboratory results desonatrata that the nitrogen dioxide
formation rate increases as the ratio is increased but the rate change is
proportionately much lass than the ratio change. In other words, Rate a
k x Ra ibn where n is approximately 0.4 for hexane and o-xylene and 0.7 for
propylene. Limited deta indicated that changing the reactant concentrations
while keeping the ratio constant also resulted in a rate change proportionately
slower then the change in concentration. For propylene the relationship was
Rats — k x ConcentrationO.SC and for smaitylene it was Rate a k x Concentratione. ?.
This seans that theoretically the reactivity factore used to calculate the
nitrogen dioxide formation rate are applicable only at a total hydrocarbon
concentration of 1 ppm end that they are less when the hydrocarbon in greater
than 1 ppm and vice versa.
-------
-33
1-
The atmospheric data shown in Table 2-2 agree with the laboratory
findings in that the calculations predict a higher rate than observed at high
ratios (4.6 to 6.1) and pre iot a lower rate than c&served at low ratios
F - ,
1.9 to 2.1). The effect of total concentration c m molar reactivity is
difficult to separate from the ratio effect. Effects due to other parameters,
e.g. radiation intensity and temperature, are responsible for part of the
differences between the calculated end observed ratea. Theae, however,
cannot acoount for the magnitude of the ratio effects noted.
Although there are not enough data points to validate the conclusions
statistically, there is strong evidgnc&tbot hy4ocetbon - nitrogen oxide
ratios play a similar role in the atmosphere as in laboratory systems.
OF ULTRAVIOLET RADIATION INTMISXTT
Laboratory experinsnta 1 have shown that the rate of nitrogen
d 4n i e formation is directly proportional to radiation intensity for a
nomber of hydrocarbons of different classes. Attempts to verify this by
mol t iple regression analyses of 46 data pointa from Huntington Park,
attituting the ultraviolet maasurementa at El monte for the points at
shids there were no data at Hwitiagton Park, resulted in a regression
coefficient of 0.25 and a correlation coefficient of -0.13. The plete
lack of correlation found does not necessarily nean that the laboratory
relationship is not valid hut rather that the day to day variations in
nitrogen dioxide formation rate were prinarily caused by changes in
t 4 naq concentration end ratio, and theae peranetere exhibited greater
day to day variation than did radiation intensity. However, it was cuLts
apparent that nitrogen dioxide formation was delayed on daye where fog
persisted to mid-earning and greatly reduced on cloudy days.
3-34
EFFECFS OF TEMPZMTUU AND HUM IDITY
Tempereti e is known to increase the rates of certain reections
i.nsolved in the conversion of nittic oxide to nitrogen dioxide and thereby --
— ‘— ,i. — - — -
increüe the overali rats. ‘The effect of humidity is subject to debate.
While some investigators have reported that low humidity leeds to decreased
nitrogen 4lnwjAnntrmatii rates, others hevé 1oyed dry systems without -
at fect. - gissidity affects were noted at Scott (7 ) after prolonged use of a
dry- e,fstta is the study of lees reactive hydrocarhone. The effects were
believed dma to reactions on the chamber well because they were noted only
sitar dry gas had been passed through the chamber for several days.
It was difficulet to assess the effect of temperature on the
atmospheric nitrogen dioxide formation rate because the temperature during
the conversion period changed little from day to day. There were three
days in October 1968 on which the temperature wes substantially above normal.
Data for two of these, October 17 and 18, are compared to date for three other
daya Sit Table 2—3. All data were adjusted to 0830 flt, the approximate
mid—point of the conversion period. November 6 offers the beet comparison
because the concentrations and ratios were nearly identical to those on
october 17 and 18 • The rate on Nove m ber 6 is approximately 40% lower than
on both October 17 and 1.8, but part of this can be accounted for by a 20%
reduction in radiation intensity. The lover rates on October 21 and 22 are
partially due to the ratio effects discussed previously. These comparisons
suggest that t aratUre baa a small but real effect on the nitrogen dioxide
formation rate, but there is no evidence that lower humidity reduces the rate.
-------
2—36
I
2.4 m18 ATLc* or OZONE sao oxxDafl
table 2-3
ozone M other oxidanta begin to accumulate in the atmosphere
after most of the nitric oxide has been converted to nitrogen dioxide. The
term oxidant ia used to rsfer to any compound which wall generate iodine when
passed through a buffered potassium iodide solution. Ozone reacts rapidly
with the Iodide ion so that its full concentration is recorded. On the
other hand, compounds such as nitrogen dioxide react more slowly end less
completely so that their response is much less than unity. since the tam
from sbsorption of the ound until the eeasnres*nt is made • whether by
colorisetry or coulometry , varies from instrument to instrument, the
response factors for oxidsnts other than ozone usist be determined for each
instrument or procedurç. For example, nitrogen dioxide has a response
factor of 0.08 with the Nest Ozone Recorder and to twice that aunt with
the mackeen Acralyzer.
Nearly all laboratory studies have been concerned with the total
oxidant present D i tha test ctember after en arbitrary irradiation period
rather than the rate at which oxidant forum once the nitric oxide has
been converted to nitrogen dioxide. the irradiation times have been shorter
than the actual atmospheric reaction period, end serious losses of nitrogen
dioxide have frequently occurred by the time oxidant began to form. These
bases have been explained as due to nitric acid formation on tha wells of
the chamber. Therefore, laboratory oxidant formation data must be treated
as qualitative rather than quantitative when making comparisons with the
atmospheric system.
Date
oct.h
Oct.18
oct.21
Oct. 22
Nev. 6
• Caictiated and Obaans Nitropea Dioxide
Porgatian Pates at Humtisgtón Park-at 0830 -Pfl; - -:
- •- - -. •c) - a N —- _ ; - -t , . i
I d tthtitjast Nets - - ile:tiv;
Calculated Observed Temp. Humidity - Ratio
( veb siC ’) ( ccb aia 1 .C L. ( ‘ I - ftC/t IO
0.61 . - 4.1 - 74 •• ,3$ . 2.1
0.57 - - 1. ) 74 15 2.1
1.01 0.7 57 92 3:7
1.99 0.9 60 70 5.6
0.59 0,7 .: - 74 2 ,l
-1
-------
Z 38
s — i :...
q..n.. p. eo oJn eoit 1o 0•
H ‘ li d.o °‘e
C l
‘I
o II
— to
sac
• omel o-’ no.i isenin .n
!
g
3 sIr (IC’ ¶I O P ¶Pt°.° 0
,. nn 4naa.’n Q.b_I.l or.r.s d
a n ln — -In 2
I.
d i
II II
o
di I; fljj I t1I U!
3 II
001 n..iQ.4’Ofl .IoOl l 004fl ‘
n h Ul • 4. • •
-
. 5
2 < 3j r.eeews’ ri.nio oe.. r. q
C iflfl ’IF 9 ’Z °‘ g
2.1 —
• %. 3 i $sIOICO’D ifl4$’D 0
x i
t ,I1 OC4’SnO nP- - It0
o wia-IWeW qnor
8 i °.° .° .‘t” 0000 0000 P
‘I
ns,ntn
‘3°( lPt t
ni, 00000
11
C
2 ,37:
The atmospheric reactions tab convert oxygen to ozone are believed
to be aneXogoue to those idtich convert nitric oxide to nitrogen dioxide.
.r.__._--. _ - . -
Thus, St %was ?131e ad- a carbon zeactivittea
- r.- • ; , :t,$t ’ . ---- -- - -r -
used in studying nit oqan dioxide fonintion should also be applicable to
ozope and caidanforstation.
Cotinuope for ozone and oiidant were available for both sitee.
Ozone was daterisiüd by difterence between total ozidant and oxidant after
r vingthe dzone by ttane—2—butene reaction according to the procedure
developed by Euffalini • The rates of formation of ozone and oxidant were
read, from the eloiie of the recoz4er chart, This was., found to y4eld more
,,C, s ,.$-- - - . -
precise rates than use of the half hour avenges.
VSC ? OF REACPARZ wIawznTxo s aND NTD QflON flCTIVITT
The calculated nitrogen dioxide formation rates of samples analyzed
jsst bifore the onset of cane Station orduring the first hour after its
onset for nine seperate days are dtowa in Table 2—4, The calculations were
sede in the em anner as those in Section 2.3. The seen contribution by
clauses for the nine aerplee was: saturate ,, 54%; aroaatics, 26%, cracked
olefi.ns, 19%: end fuel range olefina, 1%. Nearly all of the fuel range
olefins ware consi d in nitrogen dioxide formation and thus not available
to produce ozone.
The calculated nitrogen dioxide formation retes along with the
observed rates of oxidant formation, total reactive hydrocarbons, nitrogen
oxides and hydrocarbon to nitrogen oxides ratios for the nine days are
mnrazed in Table 2—5. On the three days for which the samples illustrated
were collected before the start of ozone formation, the results have been
interpolated to one—half hour after ozone was first noted. This permits a
I I
( I
0
0
0
n O O
( I ( IV,
CC
ec
Ca
. . at I
0 - I
so
-------
• From damson and Tuesday for system watli I ppm hydrocarbon, 0.38 ppm NO, 0.02 NO 5
Table 2—4
Continued
Nitrogen Dioxide Formation mates
Calculated from Atmospheric Analyeee
Table 2-4
Continued
Nitrogen Dioxide Formation Rat
Calculated from Atmoepherlo Ana3 ee
Reactivity
Compound Factor
or Wese ) p*imin/ppe NC )
0.25
ft ppane 0.39
‘ ltbuta2e 0.62
o mEotane 0.75
Ce and Higher Sat. 0.93
TQtai Saturates
Ethylene 1.70
Stopyieee • t—Sutenee 3.50
051 and Higher Otefins 8.90
total OJ,ef ins
Senaeno 0.33
Toiusne 1.30
Ce and Haghor Arom. 2 50
Total Aromatics
Total
Oct .
24, 1968 1015
PST
Oct.
fi, 0968 1004 PET
Calculated
Calculated
Rate of NO 5
Pate of 8O -
Conc.
is!.
Formation
( ‘minJ
8 Total
gte
Conc.
1 L
formation
( /etn)
.049
.0123
3.4
.180
“ .0450
.018
.0070
1.9
.0 61
, .0316
.006
.0037
1.0
.033
.0143
.044
.0330
9.1
.105
.0787
.285
.1560
.2020
j
58.3
.69
,
.5306
.008
.0136
3.7
.0760
.1292
.009
.0315
8.7
.0375
.1313
a
a
a
a
a
.017
.0451
12.4
.1195
.3139
I Total
Pate,, ,
4.1
2.8
1.3
7.2
as
4&7
11.9
12.1,
4.9’
20.8
‘4..
o.
5 5
16.3’
22. 5 /
09.26, 1968 1110 96!
- Calob istad
Pato f 5 0 s ,
Conc, Fprmation I 7otal
( r ob/mini ___
.0400 5 4
.0369
.oss. ? .oon
.oea “ .oew d.s
a - ‘ 26.8
.626 ,,,,3, .4133 8610
‘1 ,
.0490)1,, .0833
.0180?( .0630 W.5
fl ut
.0069, ‘ .0062 0 .6
.038 ç” .0503 6.8
i iA
.1026 .5695 ‘ 1
.7180 i.oa;;.
.0047 .0015 0.4 .0231 .0076
.0300 .0390 10.7 .0464 .0602
.0265 .0663 18.2 .0708 .1770
.0612 .1088 29.4 .1403 .2448
.3632 .3639 100.0 .8568 0.0890 100.0
Oct.
28, 1968 0955
nr
Oct.
31, 196b 1115
9Sf
Nov., 1, 0968 0950
9Sf
Caiculated
Calculated
Calcu lated
Reactivity
C ound pace.ore
or Claea (meb/mirQpcm MC)
Ethene 0.25
9r ne 0.39
isáotane 0.62
n-Rateme 0.75
Ce S NigSr Sat. 0.93
Total Satiwatee
• ,
fthylene 1.70
fliylene • t—autenaa 3.50
Ce1 Higher define 8.90
Total Olefina
j j
.131
.123
.018
.084
.158
.515
.036
.016
.002
.056
Pate of NO 5
flrmfl ion
(roWmin)
.0328
.0480
.0117
.0630
.1470
.3025
.0612
.0630
.0178
.1420
% Total
mate
5.7
8.3
2.0
10.9
25.4
52.2
10.6
10.9
3.0
24.5
Conc •
jp
.0320
.0110
.0020
.0000
. a
.0782
.004
.002
.000
Th i
Pate of mo 5
Formation 8 Total •
(j iein ) Pate
.ooao 11.0
.0430 5.9’
.0012 1.7
.0075 10,3
-
.0426 58.7 r
,
.0068 9.4
.0070 L6 •
a ‘
.0138 - 19.0 ,
,
a
Rate Of NOs
Conc. , formation
JaaL ’ (a db/min)
.0260 4.0065
.0100 ‘ .0039
.0020 , , 0012
.0080 i .0046
. ‘f 3z
.1048 ,:.0721 ,
,‘
.004 .0068
.002’ .0070
.000
.006 .0138
‘
•‘
.
8 ToteS
Rate
2.9 t.
.9
5.0 ’
iLL ”
53. 2
..
s .°U
.2t ”
-
l5 7 . -
‘‘
k’
Densene 0.33
loluene 1.30
Ce end Higher Axon. 2.50
Total Aroeatice
.0155
.0319
.0353
.0827
.0051
.0415
.0863
.1349
.9
7.2
isa
23.3
.0069
.0033
.0036
.0158
.0029
.0043
.0162
4.0
5.9 ,
124
22.3
.0104 • .0034
.0089 .0116
£2.&i !i a M A
.0330 .0493
2.5,
8.6’,
3Aoi
36.5 -
Total
.6517
.5794
100.0
.1000
.0726
100.0
.1438 .1352
lth.O
From Ctaeaon and Tuesday e l for system vtth 1 ppm hydrocarbon, 0.39 ppm mc, 0.02 gs a NO
-------
2—42
Table 2-5
Calculated titrogen Dioxide Formation Rates
eM C*fs&wt4 O*$4iptjtxg,Afloz Rates. attau8 gton Park
direct comparison between calculated and observed rates, The observed
oxidant formation rates are approximately twice the calculated nitrogen
dioxide rpSs’txcept on October 22 where a high MC/KOx ratio pereista4
throujhout the day end on Octob er 31 w)nth was a cloudy day with low con-
taminant concentrations. Eight of the nine observed oxidant formation ratee
were within ±0.32 ppb air’ of twice the calculated nitrogen dioxide rates.
The oxidant rates also appeared to correlate well with nitrogen dioxide
concentrations. Unfortunately, there were not enough days with high oxidant
formation rates to derive a mathematical relationship.
FECT OF TEMPERATURE AND HUMIDITY
The data for the nine days listed tn Table 2—S are revised to
include tasparature, hmsidity end wind data in Table 2-6. There were no
apparent effects of any of these parameters on oxidant rate.
ACCUIU IaTIUM OF OXIDANTS AND OZONE
The total amount of oxidant and ozone that accumulates in the
atmosphere i.e of even greater interest than the initial rate at which it
forms. In the case of nitrogen dioxide the maximum atmospheric concentration
i.e limited by the total nitric oxide and nitrogen dioxide emitted to the
atmosphere. However, there is no such constraint on ozone since it is
formed from oxygen which is in abundant supply.
A study of the continuous strip chart records for oxidant end
ozone showed that the initial rate of ozone formation could be decreased
by only three factors. They were a decrease in contaminant concentration,
a decrease in ultraviolet radiation, and an increase in the proportion of
new asissions. When none of these occurred, ozone continued to accumulate
to high levels. The reaction rate of ozone with other contaminants increases
-
Calculated
Rate of HOC
Formation
T R e Ss fltfl l (e j% l3
-. -
-Oct fl 1310 0.52
Observed
Rate of
Oxidant
Formatipn
( b misC ’)
.
1.3
Totel
Reactive
Hydrocarbons
(cul ) .
-
0.57
Total
Nitrogen
Oxides - - Mtzo
(ç;n) C IC / S I c.
.-ç -
6.20 2.8
•
Oct. !aa 0956
1.66
1.6
1.71
0.33 . 5.2
Oct. 23 1055
0.93
1.6
0.93
0.37 2.5
Oct. 24 1015
0.36
0.1
0.36
0.1.3 2.8
Oct. 25C 1110
0.69
1.7
0.65
0.23 2.8
Oct. 26 1110
0.74
1.8
0.80
— —
Oct.28’ 1210
0.57
1.0
0.65
0.17
Oct. 31 UtS
0.07
0.0
0.10
0.05
Now. 1 1030
0.14
0.2
0.14
0.06
* Valnas thtatnea by interpolation
to the time
indicated.
1
3.8
2.0
2.3
-------
2-44
- Table 2—6
, 4Q fldant rotDetlg,p *$
¶‘ t. i-n. a.tTh r
- bbssrved
calculated Rate of
Rate of °s Oxidant
rotation Formation Th9p
- ( PPb mtn ) Sb eiC 1 ) fl
0.52 1.3 73
1.66 1.6 71
0.93 1.6 86
- c - 0.36 0.7 89
— 0.69 1:7 83
0.74 1.8 78
0.57 1.0 69
0.07 0.0 65
0 .14 0.2 67
as the ozone increases so that eventually this becoeea a controlling factor
an limiting the ozone concentration, but this factor bad no noticeable -
effect at etthar ground cite.
I-- -. -
This helps to explain the accumulation of high concentrations of
ozone in resete ground sites and aloft. The amount of inhibition due to
nirQijtde fresh emissions is negligible at these jioints so that ozone
continuso to increase as long as the sun provides ultraviolet radiation.
The rate of increase is limited by the low reactant concentrations, but
the radiation period may be as long as 10 to 12 hours.
J IL l ERI EXPRESSION FOR O86NE
The equilibrium relationship for ozone postulated oy Stephens
is:
ka(UV) (NOs )
(0 ) — ke(NO)
The rates of the reactions involved in this equilibrium are so rapid that
only a few seconds en required to regain equilibrium if one of the variables
changes in magnitude. In order to facilitate testing the equation was solved
for k:
( t )(Op )
(DV) ( IC) are
Calculations of k from Huntington Park and El Monte date (referring to the
strip charts for ozone concentrations) for a number of high oxidant days are
shown in Tables 2—7 and 2—8. When the values of k were plotted against the
inverse of the absolute t arature in Figure 2—8, it was observed that the
data points for the - sites belonged to the sate population. The equilib-
rium expression appeared reasonably valid with I c being temperature dependent.
The value of I c doubled for each 10°F increase in temperature. This means that
Data
Oct. 21 1110
Oct. 22 0956
Oct. 23 1055
Oct. 24 - 1015
Oct. 25 1110
Oct. 26 1110
Oct. 28 1210
Oct.31 1115
Nov. 1 “ l03D
Relative
Ntflidity
t-t)
28
33
6
4
19
28
59
62
48
W IM
speed Direction
t 4 *i. True )
2 197
2 220
2 291
2
3 185
3 220
3 177
2 86
4 55
-------
Tahle 2-7 Ozone Equtlibrita Data — mttington flfl
fl ee O s P C . I N
2 s 2 ). ( she) ( 1a 1 ( ts3 ( v/ r n 3 ) K .C!L rr’ ,
10/21 1145 9.0 1.3 8.4 20.7 0.0673. 76.5 .001 964
1255 8.5 3.2 12.2 19.2 0.1161 76.0 .001866
1358 6.5 1.3 11.2 13.3 0.0567 72.0 .001880
10/22 1.115 11.0 2.6 19.5 21.15 0.0693 80 .ooit Si
1240 12.0 1.3. 6.7 21.0 0.0938 80.5 .001.8$Q’
10/23 1208 16.0 2.2 12.6 21.2 0.1318 90.5 .001816
1330 8.8 0.6 7.6 16.8 0.0403 82.5 .001839
10/24 1250 13.0 0.9 5.2 19.65 0.1145 96 .001198
1400 11.0 1.0 9.5 13.1 0.0884 92 .001811
U I
10/25 1133 8.5 2.3 16.7 19.5 0.0600 85 .001634
1245 12.0 6.8 6.0 15.25 0.8918 82 . .001841
10/28 1200 4.5 1.1 15.7 18.2 0.0205 71 .00 18 3
1324 11.0 0.9 14.7 15.4 0.0437 75 .001883
11/6 1122 2.0 1.2 8.7 18.7 0.0148 71 .00 1883
1233 4.0 1.5 8.7 16.8 0.0410 74.5 .0018 ,1
Table 2—8 Owne Equilibrium D na — El Monte
O P lO P C 2 I S V r
2 S ! 2 Zk. ( ppbr) ( bm) ( pphm) ( v/ n 3 ) . . _____
10/10 1312 5.0 0.7 6.6 14.36 0.037 71 .001887
I ”
10/18 1400 8.6 0.9 10.0 6.7 ;0.118 84 .001842 .
10/21 1135 8.0 1.0 12.3 20.36 0.032 77 .ooth s.
1242 16.0 0.6 6.4 19.32 0.078 79 .i, .001819,,
1350 16.0 0.5 5.9 13.80 0.098 81.5 .00 185b!
10/22 1134 10 1.6 13.8 21.5 0.054 87.5 .001830
1235 22 1.1 14.2 22.7 0.075 87.5 .001830
1338 30 0.6 10.4 21.4 0.108 88 .001828
1442 20.5 0.4 6.9 21.6 0.055 86.5 .001833
10/23 1130 11.5 1.0 2.8 22.9 0.1793 92.5 .001813
1235 12 0.9 2.7 20.14 0.1986 96.5 .OO lBOp
1340 15.8 0.7 5.8 14.6 0.1306 97 .001799
1445 24 0.5 11.8 8.4 0.1211 95.5 .001833
10/24 1328 9.5 0.7 3.0 16.2 0.1368 100
1435 10 0.8 5.2 11.2 0.1374 98.5 . .001794
10/25 1305 17 0.9 6.5 15.1 0.156 90 .001821
1410 14 1.1 10.0 10.7 0.1439 89 .001825
10/28 1212 8 1.3 15.7 16.5 0.03 5 8 74.5 .001874
1315 14 1.1 13.0 15.9 0.0745 37 .001866
1420 it 1.4 11.6 10.0 0.2319 73.5 .G01B1
10/29 1155 9.5 1.0 9.6 18.8 0.0516 69 .001894
1303 6 0.6 10.4 16.1 0.0215 68 .001898
-------
Pigure 2-8 Ozone Equilthrius Constant (k) vs
ci vocal of &b o1uta 7 St IX
k — ( Os) (NO )
WV) (NOa)
— I — ,—,-- .—- — — — a —
2—4*
if the tiV and N0 2 /NO ratio remain constant, the ozone concentration will be
twice as large at 90 0 F as at 000? and four times as large at 100°F as at 800?.
2.5 F0 9 TI(i OF ALE HYDES
Various aldehydes and ketones are produced by photochemical reactions
involving hydroc.xbona and nitrogen oxides. ?orma1d.I yde is formed jo moch
greater quantities than other aldehydea. Laboratory studies have shown that
the specific aldebyde. s.itt.d and the yield of each was a function of the
hydrocarbon structure and th. experimental conditions. The yield of formal-
dehyde per ea]a of drocaxbon reacted varied o r the range frc less than
0.1 to 1.0. An sssid msan of 0.5 seeme to be of the right order of
magnitude. If it is further ass ed that each mole of hydrocarbon which is
const d results in the conversion of two moles of nitric oxide to nitrogen
dioxide (1 , the .st meted rate of formaldehyde formation would be one—fourth
of the rate of nitrogen dioxid. formation.
It is difficult to determine precise formation rates of formaldehyde
from the 1968 data because of the low concentrations and contrthutions from
mon-photochemical o ces. 8 vsr, it appears valid to state that the rate
was within a factor of 2 of the predicted rate. On days with a nitrogen
dioxide formation rate of 6 to 10 pphm/hr the formaldehyde rate was approximately
2 pp /hr. The formaldehyde rate also appears related to reactive hydrocarbon
concentration. Higher rates mere noted on days with high hydrocarbon - nitrogen
oxide ratios as co ared to low ratio days with a similar nitrogen dioxide
formation rate. On days with a nitrogen dioxide formation rate of less than
4 pp /hr the formaldehyde rate did not exceed 1 pphs/hr.
The rates of formation of total aliphatic aldehydes mere in the same
.7;
.c
—
-
—
—
—
— —
—-
. . .
.
Fi
L4i
. a- - - .— - -- — — — — — — -
- .l 1 -i- -. . ..
4O
I 1
20-
IG-
— — — —
.06-
.05. . I — .
tt
04 - .- - —- - - - —a- . — — - - -- —
-
— - , ..- — - —a
r LI
.03 .. — - - —___- - - —t. -.. _. — .— -— — . — .— — — —
.01
z
ti±
-U
J..Li
th±
±b-
1_LU.
Liii.
thrfh
_L
LW. J_ ..U.
t
tttt
-! LU,
J J. £J..U.
t . tf
.1 -14.
i_LU.
. .- — r—
tiHH i:
4 —i_LL 1
.U..LL L. .i..... .i ._ ....
7. -
.:
.
..._.
0017
.0019
.0018
Tl(°F1 .
-------
1 49
tango doe fogmoldehydo, but sore erratic. This msy be die racy
of tho daterninoti en o cAn oe there was no general diurnal pattern td the j - ,
dtflureSeo &t n te d1 i i*a T tjZ-’ I I
‘
2.6 EEOTOCHEIIICAL AGE 07 Nt tzwean - - .
Ebr *2w purpoos of th&o analysis the $otobbthgbatage ‘ISf an air
- . 1 -
porcol will be defined as the average period of time that the reactive
ttcarbone in the parcel have participated in pbotocheeical reactions.
This sledge is required to assess the relative contributions of individual
hydrocarbons and specific sources of. hydrocarbons to smog formation. The
core reactive hydrocarbons contribute significantly during the initial period
of Stochemicai activity, but they are ceneuned in the reactions and their
utportanca disinichee with tine. The less reactive hydrocarbone are present
in higher concentrations and their relative contribution increases as the
reaction time increases.
The extent to which hydrocarbons have reacted can be measured by
use of ratios between reactive and unreactive hydrocarbons. Stephens baa
used the propylene to acetylene ratio to demonstrate this with samples
collected at Rivereide and other locations. Unfortunately, this ratio is
not always walid because of apparent additional sources of propylene which
do ndt contain acetylene. The data from the Runtington Park site showed a
rodent decrease in propylene—acetylene ratio from 0.3 to 0.2 on high oxidant
days, bet no consistent change on other days even when the Cs to Ce terminal
olaf ins with similar reactivities were substantially reduced. The 1961
data collected by NSPCA and the Cali rnia An 3 1 show no change in propylene-
acetylene ratio with time for either downtown 1.0 5 Angeles or Muse.
- . dn for individual C to C o hydrocarbons was
otudied to find othar rot butt een reactive and relatively unreactive •
edsieh ad rv o4 4ure .the Mtent w rbon
- ct r’$. -tv-v p •,- ‘ - - •- —-
leaction. The criteria uàe th A i the r&active compound diaeppeaked at a
significant rate yetcsaoereble eemwtts remained at all times, stile the
nsr 9 %iW -r -.w.n4 , distppeore - giiidb rs oloitly The ratios between the
bydrocarboas during the early morning period before photochamical activity
started most he the a from day to day. Visual examination of the data
indicated that 1,2,4—trimetbylbenrene (1,2,4—NE) should be an ideal reactive
hydrocarbon and that n-pentsna, tojuene and ethylbeneene were prime choices
for the relatively unreactive hydrocarbon. Pbs correlatione obtained using
39 data points taken at Huntington Park before 0800 PEP with the ccrreaponding
correlation coefficients and standard errors of estimate were:
1,2,4— INS (pit) — 0.339 n—pentane (ppb) + 0.08
1,2,4—Tn (ppb) a 0.206 toluene (pph) + 1.62
1,2,4—Tn (ppb) a 1.55 ethylbenrens (pit) — 0.59
1,2,4—Tn (pph) — 0.128 (n—pentane + toluene +
ethylbenz.ne (ppb) — 0.03
Ethylbenrene and the sun of n—pentane, toluene and ethylbenrene, which showe’3
the beat correlation, were used to measure the extent to which the 1,2,4—
trimethylbenrefla bad reacted as the day proceeded.
Correlations of the ratios of 1,2,4—Tn to ethylbenrene and ,f
1,2,4—Tn to the ens of n—pentane, toluene and ethylbensene with time after
0800 PST are s rired in Table 2—9. Correlation coefficiente were higher
I .
Correlation
Coefficient
0.917
0.908
0.959
Standard
Error of
Estimate
3.76 pph
3.94 ppb
2.61 pjt
0.952 2.89 ppb
-------
• p5V
Onrrolatbn of 1, 2,4—Trimsthylbenaena (1, 2,4—t3)
ujth’Tnr$oS Parameters -
Dna 3 . Pa t h
D a ahL
1,2,4-Tn - n-Pentans’ . - 0.92
1,2,4—Tn Toluona .0.9 1
L,2, 6-SS Ethylbensqne(Et5) - - 04&.
- 1a2,4 —ZS S 3.If acne - - : oss - -
Data Dance: 52 Hydrocarbon AnalySe at Huntington Park after 0800 PS? -
• tsvndrt Variable lMepw.Sant Variable(s) Correlation dofttcient
Time 1,2,4— T n / StE —0.25
Time 1,2 ,4—mB/S 3 MC - rO. 52 - . -
time 0xi nt - - o : 1s
Time l,2,4-T5a/EtB, Oxidant 0.58
Time 1,2,4—Tea/S 3 IC, Oxidant 0.64
Time 1,2,4-nIB/Eta, Oxidant, 5 3 IC 0.69
Tine 1,2,4—TIC/S 3 BC, Oxidant, 5 3 IC 0.72
Data Source: 44 Hydrocarbon Analyses at Huntington Perk between 0800 and 1300 PST
D ieSent Variable Independent Variable(s) Correlation Coefficiant
Tins 1,2,4-TIC/Et a -0.31
Time 1,2,4—mB/S 3 HC —0.61
Time Oxidant - 0.53
Time 1,2,4-TIIH/EtH, Oxidant 0.57
Time 1,2,4—TIC/S 3 MC, Oxidant 0.68
Time 1,2,4—7MB/Eta, Oxidant 5 3 BC 0.70
T i me 1,2,4—TMB/S 3 BC, Oxidant, £ HO 0.76
Data Source: 32 Hydrocarbon Analyses at Huntington Park between 0800 and 1300 PS?
Total Ce - Cie Hydrocarbon >100 ppb
Dependent Variable Independent Variable(s) Correlation Coefficient
Ties 1,2,4—malt 3 IC 0.76
Tame i,:,:.’aa,z 3 HO, Oxidant, 5 3 IC 0.86
1
2.52.
then the data po iAte bethoen 0800 and -1300 PS? were tested than when all data
after 0800 mere ened. This appeared to be caused by a decrease in photochenicel.
r rr ” Correlation gas also better t$en the.
sum of t)iiee hydrocnxbonb usa eployed instead of ethylbenxene alone.
The highest coefficient was 0.61. .. date points with high residuals (outliers)
usuaLly occtrS on tys.uith low ultraviolet radiation due to heavy clouds or
high dispersion sates due to persistent high winds. Significant photochesical
reactions prior to 0800 PS? ware noted occaaionally.
Hultiple linear regressions involving total oxidant end hydrocarbon
concentration in addition to the above parameters produced correlation
coefficients as high as 0.76. Finally, restricting the data to the 32 points
where the total Ca to C ,o hydrocarbons exceeded 100 ppb reaulted in the
highest correlation. The relationship was:
Time (hr) a —23.l(l,2,4—TMB/53HC) + 0.11 Ox (pphs) — 0.0042 531C (ppb) + 3.42
Correlation Coefficient a 0.86
Standard Error of Estimate a 0.70 hr
The coefficient of 0.11 for oxidant means that the apparent photochesical
age was 0.11 hour greater for each 1 pphm increase in oxidant. The relation-
ship also indicates that at any given time the average percent of l,2,4—TMB
reacted was 3% greater for each pphm of oxidant present. This evidence
that on high oxidant days hydrocarbons and other contaminants reaain in the
lower atsosphere for approximately one hour longer for each 10 pphs of
additional oxidant provides quantitative confirmation of a generally
accepted concept of erg formation.
-------
24
2.7
Bean atrosplwre -
Sr lsontal ,_ted for many years since it was fo,und
— - — : _c : . — — -
that peak oxidant Qonc 1 tr+1n 4nelir occurred in locations downwiid
- -‘ •. -r- •I
on &l b5er ,
-, are frequebtly transportod’-ss far as ‘Riversidi -san
lee ettntdnfl-4et& oejn zc Aloe nrty iles be lit zonitorin
etatione at whidi tkidei t concentrations exceed 15 pphs the ‘greet et length
: 1 ,, - . fr .e..a,-. -
t zt ? t4” t 1n ’$ ‘;
10 to 20 miles from dSuntc sn I Angeles aid all li i ,the axrthaast qiedrant.
‘D JagtI teld yt tIIe ia gicalph o on whiCh isa
primarily responsible for horizontal transport on high oxidant daye wee the
sea breóme which started during Lt€e Srning. The Se breeid iei generated
W ‘maya, hea€ing of load end water. Its direction alweye lay in i ke southwest
quadrant, end its typical mid-afternoon speed was 7 to 8 mph. After unset
the wind speed decreased to 3 mph or less, end its direction shifted to the
mortheest or sout east quadants and rained there overnight. During the
morning traffic peak (0600 to 0800 flfl the wind speed was generally 2 mph
or lees end its direction was variable. Little horizontal transport occurred
and the pollutants remained in the vicinity of their sources • While other
wind patterns euch as Santa Ass conditions occurred on low oxidant days,
the pattern deecribed above sea typical of all high oxidant days in the 1968
field survey.
El Monte was eelected ae the site for one of the leboratory trailers
because it was located in the normal sea breeze trajectory from the primary
&
. 1
aite at Runtington Park end in the generel area shore bagheat cxident levels
- I - - - - C
were’ frequently recorded. Some ozone wee groddeed t El Monte by photochssacal
* _ - ‘ -
“pt 3nvolviay flutante qittted in tht,öiidiat UitinL yi ut the’
C r -J •
tptal oxidant prior to the eea breeze nevef exceeded 14 pphm end was usually
less_tI an l0 ,pphm. OnOctcbert2 l when’ the etyt of àie sea breeze yes de1aygd, .
—— , — • r—r- — — I - I - - -
p .oxida9 cd cetration t II’ Mont Ia i41 U* 2 to
I I ’ f’ ’’! -,
3 hours before passage of the g front This 4p ared to be the naxinum
- I . ‘C f
could be “produced frcm)Siesiosc at l iel.presmnt in the El
Monte area, - - K -
r. ‘The pesasgeaufuthe eeog front ’ hrdugh.qiinn eho iy’ ‘at tar the
- 4 t3, — - ‘ I • “- — ‘ -r’ ‘hz - .
onset of the eea breeze was clearly evident, It was eiire obvious from the
repid oxidant increase ehcwn on the continuous strip charte than from the half
hour oxidant averagee. The passage of the ezog front, representing the
boundary of the highly polluted air ease frami the dcwntown area, wan marked
by a rapid increase in oxidant of 10 to 25 bm. Other contaminants including
hydrocarbons and carbon zoicxide also increased sharply. It is unfortunate
that continuous data for these parameters were not available so that the rate
and magnitude of increase of contaminants could be compared to that for
oxidant and ozone, However, thase data were obtained in the 1969 survey.
The oxidant maximum was reached from one to two hours after the
front passed. The relatively rapid decrease in oxidant after the maxinun
ie generally attributed to a decrease in overall pollution levels after the
main polluted air mess has paaesd on to the east. This seams to explain the
decrease during the first half hour to hour after the maximim, bet the large
d,ecreaee during the period from 1400 to 1600 PS? is more likely the result
of the reduction of ultraviolet radiation intensity by a factor of three and
-------
a
• 1
: b 0u1 tc tho nitric eaido im’irgab aissio tab increaoe en the - .,
aftorfloon traffic donoity iporoaase. -
:3 t at 31 of a o of I
tho estd front wre- 3w ftan fàund diring the early lorning peek. The
airborlp datd iedid2ztcd that t vertical distriniation gsa quite uniform
sa _ o(w S 999 a 1 Oea broese, but tbet thert uere 1prge
gruwid Xivsl ncentrations end those 1oft during ths
A c ariocn of contaminant concentrations at Huntington Perk at
WL P 5 I b witb thea und
front pooeed shoved a alight decrease in contaminant concentrations but a
large increase in oxidant. The oxidant increase was greater then would be
expected if the air mass had remained at Huntington Park.
2.8 AIRUOWIE 8SASU ftS
The airborne maasur nt5 cede during the 1968 field survey
represented a pioneering effort in the exploration of the three-dinensa.onsl
appecte of the Ice Angeles Boom at aphsre. A considerable emount of new
ledga shout vertical pollutant patterns was gained, but as is the case
in cost initial efforts, e number of shortcomings in the flight plans were
found during the data analysis. The experience gained in the 1968 program
provided a sound basis for designing the airborne phase of the 1969 program
which yielded a wealth of information on the complex three-dimsnsional
structure of a 20 mjle cross-section of the Basin atmosphere. In order to
derive the cost information from the 1968 data, reference baa been made to
the 1969 data in areas which were cot covered in 1968. This permitted a
sore complete pea
- The vextiol profilps showed thit en ,4lvaraion,e2 Xstsd during tie
jsj A ji ,
o i4i $ay the 1q y
of tt frum attie r ‘ roflusjtely ‘709 feii ai ve iea level shoved 1’
relatively so.mtant !ertiedl tov qratuxe. above this usa a anne where the
tasperatte 4nc thatuitii i4tita. This vert lay t fr. a e :
- .9.4-__ - 2. - ‘
and the coetandnnatn tted at grqwtd level mixed only to t Im base of the -
inversion. o, in the jowest layp fêish amiesions were mixed and irrddiatad,
and photochemical reactions initiated. mae is ach clearer in the 1969
data where air aat p1es care collected in the lower ‘layer, whereas in 1968 - - .
most morning samples were collected above the inversion base.
The air within the inversion from perhaps 700 to 2500 feet ICL
contained contaminants moat of which were probably emitted during the previous
day. Thus, while tha concentrations were quite low, the photochemical age of
the air mass was sufficient for oxidanta to begin to forth alonst ionediately
epon irradiation. This resulted an higher esrly coming oxidant levels above
the inversion base than below it where the nitric oxide in fresh emissions
inhibited oxidant formation until mid-eoxnieg or later. The analyses of the
bag samples collected above the inversion base frequently showed substantial
amounts of hydrocarbons from the exhaust of piston engsne aircraft. This is
typefied by the high concentrations of Ce to Co branched-chain paraff ins and
C 7 and Cs aromatics. The 1969 data show that this hydrocarbon conposition is
cot representative of the asphere in general • Aircraft exhaust was found
in the 1968 samples because the patterns floun to col’l’ect bubbler sanples for
nitrogen oxides determination resulted in flights in areas frequented by
other aircraft.
-------
As a typical day progroeoed, oxidant increased r4idlp in th gQouM
the inveroiea boos oterted to rime. The oee breeae then began and jt
I “ . .t’ .” € ’i 2
tr pertod the polluted air4 * fln*ar - ¶ ,4apg -
• - ‘: - z-. r; .±y4 d ’t -. -.
clearly vioible from aloft. Sewg fronts were traversed on several
rw.e . jt wa found that oxidants vera higher at the leading edge than
abs clout. Vinibie veido cont ainii)g very low oxi 1 twsre älao
..A aikhia tho clout. The rapid easing uhich took plate at this time
. — °-- the ocatsmalntion of aircraft exhaust, and the chromatograms of
arpIe . keo at 1500 feet were barely distinguishable from those of ground
1 air. 1st the inversion disappeared p1etely, the teçeraturs
- Siabatic lapse rate, and the oxidant changed little from ground
l 1 G feet. At other times the inversion remained at high attitudes
and a uboip dscreaae in oxidant was noted between 1500 and 2000 feet. At
in 1ta the combination of updrafta containing high oxidant (%30 pphm)
ofts containing low oxidant (‘.10 pphm) kept the needle of tha oxidant
coving continuously.
The substantial differences in afternoon profiles for pointa
diffcr st diatanoao from the ocean and the extreme complexity of the overall
symt at dtnotrated teach sure clearly in the 1969 data • The 1968 data is
basicolly valid, but it does cot cover e broad enough area to permit an
uodarxr sng of its total meaning. Nevertheless, the background knowledge
that it provided was absolutely essential to the success of the 1969 airborne
p r og
-------
3. 1tW 1 OU - LITTCAL OTEODS 908
Tin tzzk to jacpz,n,j naolflcnl eethoda weed ;
flsdige j diriMad t tj idt’üon of organic n$t
aocwtona. nitto tides, ozi4a ta c M ozone. The
to *tprowe tothniMuco thick did not produce satisfactory reeulé’s or
tuba ‘alternstc otandi MSto- nS other procazços & th jan44 pJZJ,
jt4wI weaful dote. -
3.1 OsoANIC aI
Thu organic nitrupan ounds known to be produced in photochasicel
ooctionn belong to t classes. e.lkyt Ritretee and peronyeeyl MtrqI aj -
— —.——-—,—
bccia wee placed swon de+ains -tiou of peroryacyl nitrates which ore
proaent in objectionable conoontrationu. while thece coagoundo eke’ present
in low pefle per billion concentrations in the utsoephere, they are of great
totoroot bncaueo of the eanifeetetionn eneociated with t . Porossyucetpl
nitrate (PAS) in present in greeter concentrations than its higher 1 cge,
c M it in sasuetty the only porozyacyl nitrate deta1n4. . a,R.ever, its
hecologn, incleding porozyproptonyl nitrate (fits) end perosybutyryl nitrate
(950) have bat o n to he anre powerful eye irritants end plwtoiceete
than PAN • The recontly diocovered perorybensoyl nitrate (fleN ) bee been
found to he 200 times c c potent en eye irritant as formaldehyde
Tho tbree cajor cub-teak. reletcd to organic nitrogen caipound
enalysin cure iqsrovocant and expansion of the analytical technique employing
a gas dsx togrepb oqui pad with an electron capture datocthr, dsnlo at
of a stable calibration standard and enteantion of the analytical eyata.
PJaLTflCAL PW 1 P Ie
Difficultiea care omperianced in the 1968 field e wvoy in obtaining
a response to lo!7cnecentratieto Of N uais4 the verien Lerogroph Ny F t LU
Chzo8ncogrePhe. The col so had ,tQon fojutd rt.to factory on another i.netrusee4
- . - •,-a --r•’ •- ..br, 4 ---, .r
labnflory ta tporfd4p Sr s,e4a& orn. : 0c s
- • ‘ 5 y - - • ,fl ’ - _, %e - . ‘.fr i t :
checked obt in tiie loben r’uhing a chzanatogr fih free the ecgtt field -
trailer. PAN wee produced in a photochaicel reactor and transferred to
the 40 cater long path e4l it s Porkin.8lIeer Model 21 infrared spectr to ’ -
tartar. The concentration wac dotorained from the absorption peek at 12.6
al-crone ueing the solar cbcerptivity reported by Stephens ttes • samples
withdrawn in glees syringes, diluted to parts per billion concentrations and
jujocted into the cbroeatogre$1. The probl e m of lose of PAN withsn the
instruannt wee traced to the injection block. when the flow eyeten wee
ruvieed to bypeao the injection block so that the sample passed directly fron
the n ling loop to the cbrceetegrnpbic aeperation colunit, no losses of PAN
cure evident.
The overall oyetc wee optimized cal the following conditions
cure celected for fl-old usa .
Columns 23’ z l/S Tot ion packed with 5t Curbafex 400 on Disport S
Center gee. Nitrogen at 55 cchsin (80 peig)
Cois Temperaturei 3 5°C
Detector Temperatures 100° C
injection T orututes Aobieet
Sample Sizes S cc
At these conditions fiN von oleted isrndiatel& after PAN aid its concentration
could he determined if it ozceedcd one pert per billion. Unfortunately, no
nuitnblo cethod for low porta per billion concentrations of PSzN was developed
in tins for nsa in the 1969 field program.
-------
3.2 sY0 CA IS
The analytical procedure for individual C, to C 10 hydrocarbons
developed for use in the 1960 field survey eade possible for the first tian
the acquisition of data for the ccagilete range of hydrocarbons present in
$c.
ta — e” e. The.’jrdceduüS lcyede Perk.tq-Zlmer Nodal 900 Cbrceatogtapb
with deel fl ionisationkdete tors and a aub-enbient accessory A
is 1cw esl p ad at th Cv -t Ce ydrocarbons and a cjilsry
l*a’waveeS k4a er dte C to Cie c ounds concientretsoi oft ‘
lane than O.lEn could be; detected using a 25 ml sample. The problans .
sotarst 11 perSw..üig the tqdrocerhon enilyses i lisied interference
- !. 4- ! . — — ...,t . -
tetq drecer1ca ixpuri!ieo preeent la m the helium carrier gas, en uneta le
p .ip In .th pathe 3 olc during temperature progring end hang—up of
t .tJ - - -
high rl ln wOight cSi.—..entn in the inlet system of the chronatogreph.
psncanww
— -a -‘-- . ‘-. -‘‘- — . - -• -
I- r- tiz e; * - -’.- . -- . — .
catalytic purifier van designed and constructed to rave the
trace Wtuaflces from the carrier gee. The catalyst bed was cosposeâ of
40/00 sh ._ .r,ar oxide operated at 650 to 100° C. It yes followed by a
ealecelar eta trap to eater end carbon dioxide. This purifier
worked nfl Sn the laboratory with all grades of helium. Itelso raved
the atha from e gee mixture containing 2 ppm in air.
The problem of high leculer weight compound contaninetion was
traced to the Maya.regm end inserts in the flow controller and flow eater
..L...ma from the gee seiling velve. The all in-line molecular steve
trap hecema overloaded end the material bled through to the capillary column
and slated as a large broad poet. Frequent reactivation of the mo lecular
sieve trap by heating wee found to he anceseery to keep the undesired material
out of the system.
Ci to Ce INALYSIS
The Macline obtained with the Porepak Q - silica gel ooluma need
toceperete the Ci to Ce bydrocerbons wee frequently unstable to the point -
I :
OLUMTION S SP IRO POE fl20100N. CA fUR L T - - - -s. -
- •_4 44 - • - - -
response of the electron capturl detttor chengee from’ day to
çn tghetectqr Scntmo Sin ’cfl & ’ - ;
s--. .t -‘.-..- c-cf . -n .i.t. -. - ,
44 . .a4.djs needed for diibVedjusSjnt bu ms. i’fl hite le ii1tbrsseJ i a4
- -. - - I
I , -‘ p - (_ - --- 4
atainere • The alternative which wee explored wee-ic mmdc of ajcre stable
ound which wo Ud elute from the column in the viâfi1ity of I$ end- gr&hzce -. -
t_-d ;: — — - _4_ .-4-
e$pteetnr;reeponöc proportibnal to that of. PM tateie csen-
4,-. .. 7 ’i 3 -4 ._--.,-
from aUnt of a . dldate compounds end w4ec-t nfindd!? s 2$ad
at a concentretion of S pph. The ratio of 255 reejmonsn to própyl nitrate
- . - p -
epnn aa use the cane from day to day sad p?c trat otjprop44tratet
4s
aj6eiked smiffiOiintly otablo fdr field use. £ calibratiercurv e fc r iP 3ZI in
t of the prowl nitrate standard is required at the beginning of a test
program and rechecking is desirable occasionally thereafter, hut propyl
eitrate qp 4 gtiefactory for. dsjly calibration of the electrad capture
sea RthTIO OP -PS (Isovossiw
The Verien Aerograph Chroeatographs were automated by the addition
of .n-.sy . iate eolesotd velves end Liners • anelysee for PAN were then
perforead at 20 minute intervals on a ’ 24 ho ur basin. This se done to follow-
PS concentrations pest amen working hours since it had been reported
that PS persisted well into the night ’ . -
-------
3—5
3—6
that certain cospoundo could not be read accurately. Tine wee attributed to
physical chengee In the column material during temperature programning. A
nwter of alternete aubetratee Inclining carbon coated with Apteson.
Polypak-l and Porape3t Q proved unsatisfactory ten tested. Finally a column
packed with Baynal (colloidal alumina) produced the deeired aeperatlon. The
following conditione were found ca:tpetlble with the taiipereture program used
with the capillary column:
Column: 12’ x 1/B” e.e. pecked with 40/Ba mesh Baymal
Carrier: Helium B 75 peig
Fuel: Hydrogen 0 18 psig
Purge: Oxygen 0 40 peig
Sample Size: 25 ml
Temperature Programi Hold coiisei at —75° C f or 5 minutes, an3ect sample,
after one minute incraase temperature to _200 c end Itol.d for one minute,
then program at 6.50/ sin to 1200 C and hold. Beckflueh column after
pentana elutes.
HYDICCARBON ST 3ILITY D l TEDI.AR BAGS
The stability of hydrocarbons in Tedlar bags yea investigated in
order to determine if bags collected on high oxidant days could be held for
ees’erai days before analysts without significant changea in hydrocarbon
concentrationa. It was found that bags analyzed periodically over several
days exhibited only minor changes. with the lergeet losses occurring in
high molecular weight aroaatica.
work was also performed to identify additional peaks in the capillary
column chronatograns with emphasis on C 9 and C 10 aromatics and other high
molecular weight compounds.
3.3 NIT )GEN OXIDES
The obvious need for more frequent data points for nitrogen oxides
at both sites made it necessary to construct new sssltiple gas samplers. The
design of the original BEN easipler was modified to sccosndate a set of sixteen
bubblers, eight for nitrogen dioxide and eight for nitric oxide. Syringe
needles were used as orifices to control the sasple flow rates. twe or more
samples could be collected simultaneously with each of the completed samplers.
Through the use of overlapping cespling periods it was possible to obtain
data points at ten minute intervals tile still sampling for as long a period
as needed to develop nufficient color zmtenmity for eccurate results.
NITBOGDJ OXIDE STABILITY IN TEDLAR M
During the 1968 program nitrogen oxide samples were collected aloft
using a manual sampler similar to those employed on the ground. This
prodcedure was not satisfactory because pressure changes caused liquid flow
between bubblera at times, end the required 15 minute sampling ties seriously
hindered the profile exploration.
‘the stability of nitrogen oxides in Tedlar bags was investigated
with the idsa of extending their euccsssful use for hydrocarbons and carbon
mosnxida to nitrogen oxides. Several mixtures of nitric oxide and nitrogen
dioxide at concentrsttons in the range of 10 to 30 pphm were used to fill a
series of 6—litsr Tedlar bags. At each level, five duplicate bags were
analyzed at intervals fron 0 to 40 hours after filling, tosses averaged
0.1 pphm/hr for nitric oxide and 0.2 to 0.3 pphsfhr for nitrogen dioxide.
it was concluded that bags were applicable to nitrogen oxides sampling aloft,
out that all bags should be analysed on the day they are collected.
-------
3—7
BECEMAN ACRALYZEBE
The Beckman Acralysera • which had been delivered after the conclusion
of the 1966 field survey, were installed in the government—owned trailer. The
permanganete solution normally used to oxidize nitric oxide to nitrogen
dioxide was replaced with a mere efficient end reliable solid chronic oxide
oxidizer. A cold trap was placed ahead of the oxidizer to rerove aerosols
carried over from the first absorption column which would have shortened the
oxidizer’s life substantially. The scdified oxidizer unit performed quite
well, but the eaneitivo pressure balance of the liquid-gas system necessitated
additional smdificationa to achieve proper flows through the instrument.
3.4 OXIDANTS PRO OZONE
The analyses for oxidant and ozone in the 1968 programs lent sons
support to laboratory chaubar findings of oxidants other than ozone in photo-
chemical systems (1 ) • These include hydrogan peroxide and alkyl hydroperoxidea.
The methods reported in the literature for these compounds ‘Jars found
unsatisfactory for use in the 1969 field program because too imich tins would
have bean required each day to aat them up and perform them. The 1,2-di-
(4—pyridyl ) ethylene (DEE) method for ozone was found to be useful for
checking the Meat Oxidant Recorder calibration. Air was sampled from the
vicinity of a Xenon lamp to obtain the required ozone. The Mast inztrumsnt
was operated in parallel with the midget impingare used in the oar method.
The results from tha inatrument and the colorimetric procedure agreed within
1 to 2 pphm in the ranga from 20 to 40 pphm ozone.
The Mast oxidant analyzers used in the field program included
ole fin titration apparatus to determine ozone. The ozone was rarovad by
-------
3—B
4—1
addition of trane—2—butene in alternate 7 1/2 minute cycles. The concen-
tration of ozone was indicated by the difference between the oxidant reading
without addition of trana—2—butene and that with addition of trans—2—butene.
The olefin titration apparatus was rebuilt and the on—off cycles were increased
from 7 1/2 to 15 minutee each. This made it poasible to better observe the
oxidant fluctuations within each cycle and reaulted in lees total dead time
during cycle changea.
The Beckman Acralyzer for oxidant received after completion of the
1968 field work was installed in the Government-owned trailer. It served as
a supplement to the Mast Recorder.
3 • 5 MIS LLANEOUS
The substantial number of defects found in new equipment obtained
for the 19GB survey end the frequent failures experienced after a short term
of operation made it desirable to check out all new equipment as soon as it
was received. Some instrumentation used in the 1968 field work had been
obtained on a lean basin with new unite supplied when available. Other
equipment was reoeived from the sub—contractor, Bolt, aeranek and Newman at
the conclusion of their effort. The items which were checked included the
Eppley Radiation Photometers, Climet Temperature and Dew Point System, Bendix
Hydro—’therztographe, Gas Samplers, etc. The number of defects founo nade it
well worth the effort to carry out this task. Corrective action was taken
so that the equipment was operational by the start of the 1969 field program.
4. DISC1JSSIGI AND RECOMMENDATIONS
The analysis of the aerometric data collected in the Los Angeles
Basin in the Fall of 19GB has produced valuable information regarding the
composition of the atmosphere and the rates of the reactions which take place
in it. It was not possible to prove the validity of the hypothesized relation-
ships betwaen the syatem inputs and its outputs because of an insufficient number
of data pointa for certain inportant parametera. Nevertheless, the data
provide sufficient insight into atmospheric behavior to permit a better
understanding of the quantitative aspects of smog formation.
The data analysis suggeeted that a number of sources contributed
to the various contaminants found in the atmosphere. Detailed analyses of
the various sources 1 especially for Ci to C io hydrocarbons, is necessary in
order to firmly establish the quantity of contaminants emitted by each source.
This information is needed to determine the value of applying potential
eource controls.
The relatively good agreement between the observed rates of
formation of nitrogen dioxide and ozone and the values calculated from
laboratory reeu lte was encouraging. Mast data appeared to support the
findings of laboratory studies when the laboratory work employed teat chamber
systems which resembled the atmospheric system. The major difference noted
between the atmosphere and laboratory work was in the behavior of nitrogen
oxides. In the atmosphere the concentration of total nitrogen oxides changes
primarily by dilution as shown by the relatively constant within day CO/NOx
and NC/NOx ratios. On the other hand, laboratory systems frequently exhibit
large lomaea of nitrogen oxides with time, probably through reactions on
the chamber walls.
-------
4—2
5 — 1
5. REFER9 CES
1. “Final Report on Phase I—Atmospheric Reaction Studies in the Los Angeles
Basin, Vol* I,”
July 28, 1969.
2. “Final Report on Phase I—Atmospheric Reaction Studies in the Los Angeles
Basin, Volume II,”
July 28, 1969.
3. “Cs to Ce Hydrocarbons in the Los Angeles Atmosphere,”
Robert J. Gordon, Henry Msyrschn and Raymond 14. Ingele,
Environmental Sciesce and Technology,
Vol. 2, No. 12, December 1968.
4. “Sources and Reactivity of Light Hydrocarbons in Anbient Air,”
E. R. Stephens, I. F. Parley and F. R. Burleson,
Midyear Meeting of American Petroleum Institute’s Division of Refining,
May 16, 1967.
5. “Distribution of Light Hydrocarbons in Ambient Air,”
Edgar R. Stephens and Frank R. Surleson,
62nd Annual Meeting of Air Pollution Control Association,
June 25, 1969,
6. “Sunrary Report of Vehicle asissions and their Control, 0
A. H. Rose, Jr.,
USDIEW. Ptlic Health service, NAPCR, Cincinnati, Ohio,
October, 1965.
7. “Pbotochemical Reactivity of Hydrocerbons in the Presence of Nitric
Oxide in Air,”
L, a. Reckner and W. S. Scott,
American Chemical society Meeting,
April 1968.
8. “Hydrocarbon Reactivity end the Kinetics of the Atmospheric Phatooxidstion
of Nitric Oxide,”
William A. Glasson and Charles S. Tuesday,
August 15, 1966.
9. Hydrocarbon Raactivities in the Atmospheric Photooxidation of Hitric
Oxide,”
William A. Glasson and Charles S. Tuesday,
150th National Meeting of the American Chemical Society,
september 1965.
10. “Gas Chromatography - Measuring Ochauat Hydrocarbons Doxn to Parts Per
Billion,”
Louis 3. Papa,
Hid—Year Meeting society of Automotive Engineers,
May 15—19, 1967.
The formation rates of photochemical products vera favored by high
contaminant concentrations, It also appeared that the formation rates ware
inhibited by freoh emissions, If this was true, then the actual rates represent
the nat affect of the driving force of the reactants and the inhibiting force
of new emissions. This can account for the relatively high oxidant concen-
trations found in remote areas and aloft. The reactant concentrations are
lox but inhibition as negligible, and oxidant formation can proceed at a
moderete rats.
The low concentrations of fuel range olefins found at the start of
pootochenical activity are of considerable irportance. The calculation of
nitrogen dioxide formation rates suggests that these olaf ins are minor
contributors to photochemical products. If the assumptions made are valid,
proposed fuel modifications would not have a great affect on the rate of
chemical end photechemical reactions in the atmosphere.
The approach used in analysing the data, namely seeking empirical
relationships betwaen the inputs and outputs of the atmospheric system, gave
evidence of being very useful in developing the mathematical expressions
needed to evaluate proposed control measures applicable to the eye tam inputs.
In order to fully explore this approach, it is recosmended that aimultaneous,
comprehensive data be obtained for all of the pertinent parameters. The
parameters include l nitric oxide, nitrogen dioxide, individual hydrocarbons,
ozone, total oxidant, ultraviolet radiation intensity, temperature and humidity.
It is deairable that the data points be ae frequent as practical an that
accurate reaction rates can be determined. Additional vertical profile data
is also needed to obtain a three—dimensional view of the atmosphere and the
reactions and transport of pollutants which take piece in it.
-------
5—2
11. “Gas Phase Titration of Atmospheric Ozone,”
3. 3. suffalini,
April 1968.
12. “Chemistry of Atmospheric Oxidants,”
Edgar B. Stepheno,
61st Annual Meeting of the Air Pollution Control Aesoc3.etior.,
June 1968.
13. “Atmospheric Phctochesacsl Reactions in a Tube,”
C. R. Stephens and N. A. Price,
Presented at the National fleeting of the American Chemical society,
New York, September 1966.
14. “Hydrocarbon Reactivity and Eye Irritation,”
Jon M. Iieues and William A. Glasson,
lSSth National Meeting of the American Chemical Society.
April 2, 1968.
15. “Abeorptsvitiee 1 for Infrared Determination of Peroxyacyl Nitrates,”
C. R. Stephens,)
Analytical Chemistry, 36 , 928—29.
1964.
16. “Automat Ic Chroaatographic Measurement of PAN,”
0. C. Taylor, E. A. Stephens and E. A. Cardiff,
Air Pollution Control Aeeociation Meeting,
June 1968.
17. “Photooxidat on of Hydrocarbons in the Presence of Aliphatic Aldehydes,”
A. P. Altshuller
science, 156, 937
Hay 19, 1967.
is. “specific epectrophotometric oetersianation of Ozone in the Atmosphere
Using l,2—Di—(4—Pyridyl) Ethylene,”
T. R. Heuser and U. W. sradley,
Analytical Chemistry 38s 1529
October 1966.
------- |