The role of soot and primary oxidants in atmospheric chemistry

The Science of the Total Environment, 36 (1984) 1--10 Elsevier Science Publishers B.V., A m s t e r d a m -- Printed in The Netherlands

±

THE ROLE OF SOOT AND PRIMARY OXIDANTS IN ATMOSPHERIC CHEMISTRY*

T. NOVAKOV Applied Science Division, Lawrence Berkeley Laboratory, University of California, Berkeley, California 94720, USA

ABSTRACT This paper discusses the role of soot in atmospheric sulfur chemistry. A methodology for estimating the primary and secondary particulate carbon concentrations is outlined. The relationship between ambient S02, sulfate, and primary carbon is examined; and certain regularities in their behavior are defined. An explanation of these regularities is proposed, based on a mechanism involving oxygen chemisorbed on combustion-generated carbon particles and its subsequent reaction with aqueous sulfite.

INTRODUCTION Carbon is the largest elemental fraction of atmospheric aerosol particles and is found in many different chemical and physical forms.

These carbonaceous par-

ticles play important roles in atmospheric chemistry and physics and are crucially involved in a number of ecological and environmental effects, including human health.

These may be broadly classified according to their origin or

structure, using the following definitions: Total carbon, C - Total particulate carbonaceous material. Black carbon, C b - Combustion-produced black particulate carbon having a graphitic microstructure. Primary carbon, C'

Particulate carbon derived directly from sources, being

the sum of primary organic species, C'o, and black carbon.

The term "soot," Cs,

is synonymous with primary carbon derived from combustion. Secondary carbon, C" - Organic particulate carbon formed by atmospheric reactions from gaseous precursors, defined as C" = C - C'. The units of all these quantities are micrograms of carbon, either per cubic meter of air or per square centimeter of filter. This paper discusses i) methods for determining the concentrations of primary and secondary carbonaceous material, 2) an examination of evidence that suggests a chemical link between the carbonaceous fraction and other aerosol species such as sulfates, and 3) the role of primary carbonaceous species in the atmospheric *This work was supported by the Director, Office of Energy Research, Office of Health and Environmental Research, Physical and Technological Research Division of the U.S. Department of Energy under contract DE-ACO3-76SFO0098 and by the National Science Foundation under contract ATM 82-10343.

oxidation of sulfur dioxide to particulate mostly to polluted

atmospheres

direct source of carbonaceous

DETE~IINATION

is the principal

the relative abundances

will be discussed.

(black carbon/total (ref.l).

carbon)

Cb. and compares the ambientCb/C

to the ratio Cb/C for sources and source-dominated

(EGA) to obtain characteristic

secondary carbon components

the black carbon concentration has been shown to be linear, (ref.4).

C is determined

of the optical

absorption.

The method

for ambient

attenuation

ATN = OCb,

of black carbon on a filter.

by a conventional

combustion

The total

technique.

contain black carbon and primary organics

and have a specific attenuation of Z' = O.Cb/(Cb+C,o). contribute

produce black carbon.

to measure

Z = ATN/C = O.Cb/C gives the ratio of black to total

Primary particulates

sion mechanisms

o£

of primary and

(LTM)(ref.3)

and free of interferences

It gives a measurement

Specific attenuation

method

on a filter by its optical

sensitive,

to the surface concentration

carbon concentration

"signatures"

(ref.2).

The first appraoch uses the laser transmission

carbon.

of primary and secondary

The second approach relies on the more elaborate methodology

evolved gas analysis

samples

direct or in-

The first uses a fast and simple

to measure the black carbon component,

proportional

are limited

aerosol particles.

for determining

carbon in ambient particles

sites

where combustion

The discussions

OF PRIMARY AND SECONDARY CARBON

Two methods

technique

sulfate.

secondary carbon,

The specific

sample is then Z = O'Cb/(Cb+C'o+C,,).

Gas-to-particle

(C'o) conver-

C", to the particles but cannot

attenuation

of such a mixed carbon aerosol

The inclusion of C" in the denominator

re-

duces Z, so that the fraction of primary carbon may be determined by the relation

[C'/(C'+C")]

= Z/Z'

The validity of this approach depends on the constan-

cy of Z' for primary ambient carbonaceous

Manhattan, t

--

'

jj-~~l"

particles.

In Fig. 1 we show the

NY

---7

~

i[ 2'057 + . ~RangeforLsources Fig. I. Monthly average specific attenuation values, Z, for Manhattan. ~' is the adopted value for primary carbon. The range of Z values for automotive sources is also shown.

_

~

1979 DJ

I

1980

J

D

range of ~ w~lues for samples from automotive values for ambient samples collected heavily impacted by vehicular

and local

pect

to be dominantly

its carbonaceous

aerosol

show this constancy, ~alue for sources.

without

seasonal

carbon,

them

California).

stationary

and calculated

New York. sources,

This location

Fig.

5.7 * 0.5, will be adopted

2 shows monthl'," average E/r'

primary carbon data For samples from Anaheim

also tend to show lower values sults are consistent

production

with lowest

California (ref.5).

samples

These re-

of secondary carbonaceous

CA

]

..4j

0.5

Figure 2. bbnthIy average Z/Z' ratios, total carbon C, and primary carbon C' concentrations for Anaheim, California.

0 30

b d

(sou-

matter.

J

1.0

IE

Time-resolved

in the daytime than at night

with the photochemical

Anaheim, 1.5

ratios,

It is clear that the Z values for Anaheim are significantly

(i.e., highest C"/C) during the summer.

particulate

ex

The data

effects and with a mean similar to the mean

lower than for New York and that they" show a seasona[ variability values

is

and we therefore

primary" in character.

The mean Z value for Manhattan,

as the E' value for this discussion. total

sources and the monthly average

in Hanhattan,

20

lO

1977 D J

1978 D J

The method described of samples,

above is suitable

with the objective

For more detailed analysis

1979 D,

for the rapid study of large numbers

of determining

studies of the composition

(EGA) can be used

(refs.2,6,7).

average compositional of individual

In this method,

from a quartz fiber filter sample is heated at a constant 102) or neutral catalyst

(N2) atmosphere.

samples, evolved gas a small punch taken rate in an oxidizing

The evolved gases are passed over an oxidizing

to convert all carbonaceous

CO 2 analyzer.

properties.

The carbon evolution

ature gives the carbon "thermogram."

gases to CO2, which

is then detected by a

rate plotted as a function of sample temperBy simultaneous

measurement

of the optical

attenuation, a high-temperature peak corresponding to the decomposition o£ the b~ack carbon component can be identified.

The EGA-based approach to the differ-

entiation of primary and secondary fractions depends on the construction of an average thermogram for primary carbonaceous material.

After normalization o£

the black carbon peak (because this is unchanged by secondary mechanisms), the difference between the thermogram of an ambient sample and the primary thermogram is due to the secondary carbon.

This method was used to estimate the pri-

mary and secondary carbon in a number of samples from the Los Angeles area, collected during intense smog episodes when the highest concentrations of secondary carbon are expected (ref.8).

The samples were collected simultaneously at

a receptor site (Duarte) and a source-dominated site (Lennox). The potential of the EGA method is illustrated in Fig. 3, with thermograms of a daytime and a nighttime sample from the receptor site.

The common features of

both are the black carbon peak, 5, and the group of volatile organic species, ~. However, only the daytime sample shows the presence of B and y peaks.

Because

thermograms of the nighttime samples from this location were found to be similar to those at the source-dominated site, the B and y peaks have been identified as belonging to secondary carbon.

Detailed descriptions of the procedure and dis-

cussions of the chemical nature of secondary carbon are presented in ref. 8. The results of these analyses agree well with theoretical source apportionment calculations

II

(ref.9).

~ ! I

1 ~1

I

"(3

Daytime~

0

~NJqhttime-~I/~

I

I

I

Fig. 3. Carbon thermograms for a daytime and a nighttime sample collected at a receptor site (Duarte) in southern Califor-

!

11 l I i I f 100 200 300 400 500

nia.~represenvtoslatislpeecieBand s, y have been assigned to secondary carbon, and ~ is the black carbon peak.

600

Temperature °C

FIlE ROLE OF SOOT AND PRIMARY OXIDANTS IN SO2 OXIDATION Correlations tration fornia

(ref.lI),

is st r i k i n g understand ative

between sulfate

concentration

have heen o b s e r v e d f i r s t

to find such similar

and I , j u b l j a n a ,

were a n a l y z e d

for total

at

(ref.14).

carbon,

will

a n a l y z e d by" FGA.

The s a m p l e s from L j u b l j a n a

and b l i l l i p o r e

present

In l , j u b l j a n a ,

It

To b e t t e r

several

collabor-

t h e most p e r t i n e n t (ref.13);

a m b i e n t s a m p l e s were c o l filters.

These samples

and p a r t i c u l a t e

are particularly

of pollutants

sulfur.

A number

a v e r a g e 2 4 - h o u r SO2 c o n c e n -

interesting

during winter,

c a r b o n i s p r i m a r y and p h o t o c h e m i c a l p r o c e s s e s of particulate

shown i n F i g s . tions

so d i f f e r e n t .

Anaheim, ( : a l i f o r n i a

Twenty-four-hour

black carbon,

in (Tall fref.12).

were a l s o d e t e r m i n e d .

high concentrations

tions

smog c o n d i t i o n s

locations

(ref.13);

fiber

carbon concen-

and s u b s e q u e n t l y

we h a v e i n i t i a t e d

This discussion

,17-ram q u a r t z

o f s a m p l e s were a l s o trations

behavior

i n New York C i t y

Yugoslavia

on o p e n - f a c e

photochemical

correlations,

sampling programs.

from s t u d i e s

lected

intense

the causes of these

field

results

even d u r i n g

and s o o t o r t o t a l

i n Sweden ( r e f . 1 0 )

exhibit

4a, b,

sulfur

(mostly present

a similar

pattern,

suggesting

are negligible.

as s ulfa te ),

and c, w h i c h show c l e a r l y

because there

when a l m o s t a l ]

that

are

the particulate

Daily concentra-

carbon,

and SO2 a r e

SO2 :md c a r b o n c o n c e n t r a -

common s o u r c e s .

It

is also

evident

/~g/m 3

20 l-"

I

1

nll-

I

.hi

0'3

150

--

b

I

-

b F i g . 4. T w e n t y - f o u r - h o u r a v e r a g e conc e n t r a t i o n s o f a) p a r t i c u l a t e s u l f u r S, b) p r i m a r y c a r b o n C ' , and SO2. S/(I' r a t i o s a r e a l s o shown.

50O ff

250 0

O.4O 0,30 Or)

0.20 0.10 0

October November December

1981

6 that the variations

of particulate

ine these similarities C = C', ref.14),

sulfur and carbon are similar.

more quantitatively

by means of the S/C' ratios

centration

reaches ~ 30 ~g/m 3.

the average

Chere,

which are shown in Fig. 4d.

Figure 4d shows that the S/C' ratio is relatively

by approximately

We may exam-

constant after the C' con-

At lower C' concentrations,

a factor of 2.

This behavior

this ratio is higher

is shown in Table I, which gives

Ljubljana S/C' ratio for several C' concentrations.

Average S/C'

TABLE l Average sulfur to primary carbon ratios.

C' (ug/m 3)

LJB a

<5 20

~.23±o.o3

6030

t~ 0 . 1 4 ± 0 . 0 5

>60

S/C' ANAc

NYCb

f

o.19±o.o7

BRWd

SWEe

0.5±0.1

fo 8 o,o

0.40

•

0.14±0.03

aLJB - Ljubljana; 70 samples taken Sept., Oct., Nov., Dec. 1981. bNYC - New York City; 56 samples taken Feb., March, April, May, .June 1979. CANA - Anaheim; 48 samples taken Sept., Oct., Dec. 1977. dBRW - Barrow; 12 samples taken Jan., Feb., March 1983. eSWE - Sweden; samples taken Oct., Nov., Dec. 1974, Jan. Feb., March 1975. fFrom ref. 16. ratios corresponding

to different

cations are also presented. heim

(ref.13),

and Barrow,

taken from ref. 16.

C' concentration

Alaska

(ref. 15).

For Manhattan,

from several other lo(ref.13),

Ana-

Swedish data for S/soot ratios are

it was assumed that C = C', while Anaheim

and Barrow data use values of C' calculated ratios

ranges

Included are our data for Manhattan

from these entirely different

by methods described

locations

vior with respect to C' concentrations

show remarkably

earlier.

The

consistent beha-

(see Fig. 5).

Detailed analysis of these data shows the following general regularities: o For the same range of C' concentration, the same at all locations

the S/C' ratio is approximately

studied.

o The S/C' ratio increases at low concentrations. o Only a fraction of the available o Once particulate

sulfur is converted

sulfur has formed,

to sulfate.

there is no evidence

for significant

further oxidation of SO 2 to sulfate. Most atmospheric

oxidation processes dependent

lead to concentrations conditions emission, sulfur,

of S uncorrelated

of meteorology,

stagnation,

on SO 2 concentrations

with those of C', depending

solar radiation,

if strictly defined as the high-temperature

is inadequate

in magnitude

sion to explain the data.

and insufficiently

etc.

would

instead on

Primary sulfate

oxidation product of fuel correlated

The remainder of this paper outlines

with soot emisa theory that

0.8 0.7

I

I

I

I

I

I

L

~z B R W A SWE [] ANA • LJB o NYC

0.6 0.5 0.4

_

F i g . 5. Average S / £ ' r a t i o s f o r d i f f e r e n t C' c o n c e n t r a t i o n s , q'he d a t a a r e f o r b I a n h a t t u n (NYC); L j u b l j a n a , Y u g o s l a v i a (LJBi; Anaheim, C a l i f o r n i a (ANA); Sweden (%Ix'E)(from r e f . 1(~); and B a r r o w , A l a s k a (BRI~')( f r o m r e f . . . . . .

0.3

0.2

It

0.15

I 0

I

I

I

I

I

10 20 30 40 50 60 70 80 9 0 1 0 0

C' ~g/m 3) can s e m i q u a n t i t a t i v e l y fate.

This theory

explain

involves

the observed behavior

two m e c h a n i s m s :

o x y g e n c h e m i s o r b e d on w a t e r - c o v e r e d rite

studied

of soot in the oxidation

in o u r l a b o r a t o r y

carbon

( r e f . 17).

trated

in F i g .

of time after solution. in Fig.

that

of tMs

linear

that ([I

sulfate

illus-

c a r b o n i s added t o a d i l u t e i n two s t a g e s .

of sulfite

in Fig.

sulfitc

The f i r s t

(I

and t h e f o r m a t i o n o f s u l -

6) i s a s l o w e r p r o c e s s ,

to s ulfa te .

have been studied

cannot fully

are

as a func tion

c o u l d n o t be f o l l o w e d by t h e a n a l y t i c a l

conversion of sulfite

in a t m ospheric

process

concentrations

proceeds

disappearance

process

II process

second stage process

has been extensively

oxidation

and s u l f a t e

the oxidation

The s e c o n d s t a g e

of the stage

nificance

features

a known amount o f d r y a c t i v a t e d

used.

in a g r a d u a l tics

The p r i n c i p a l

This is a very fast

techniques

with of sul-

peroxides.

o f SO2 t o s u l f a t e

6, w h i c h shows s u l f i t e

It is clear

of sulfite

and 2) t h e r e a c t i o n

sul-

u s i n g b o t h c o m b u s t i o n - p r o d u c e d s o o t s and a c t i v a t e d

6) c a u s e s t h e i n i t i a l

fate.

1) t h e r e a c t i o n

soot particles

with primary combustion-generated The r o l e

of ambient particulate

The mechanism and k i n e -

in d e t a i l

(ref.

f o r m a t i o n h a s been a s s e s s e d e x p l a i n by i t s e l f

resulting

17),

and i t s

sig-

(refs.18,19).

This

the observed regularities

for ambient sulfate. The f i r s t those

stage process,

inferred

amount o f s u l f i t e bon;

Fig.

oxidized

6c shows t h a t

produced saturates process

however, has features

from t h e a n a l y s i s is

linearly

proportional

S/C'

despite

ratio,

Fig.

a considerable

to the of car-

t h e amount o f s u l f a t e

excess of aw~ilable in proportion

and would consume o n l y p a r t

similar

6b shows t h a t

to the c o n c e n t r a t i o n

for a fixed carbon concentration,

in t h e a t m o s p h e r e would produce s u l f a t e

a constant

qualitatively

of ambient samples.

of the total

sulfite. to carbon, available

Such a i.e., SO,.

8

>.

I

I

P

i

T

"~

f

¢..

--

/~omplete

4

2

l

/

6 0.. ~ 0 I 0

f

C

b

Of

e l~t

O~ ,4 0

x"-'

4

• L 2

//~

2

•

j 1

I 3

Figure 6. a) Sootcatalyzed SO 2 oxidation, showing the fast (stage I) and slower (stage II) processes. b) Initial drop in sulfite concentration vs. carbon for constant sulfite concentration, c) Initial drop in sulfite concentration vs. sulfite concentration for constant carbon concentration.

/

/

t

f

-

//~conversion

0 _ ::~--e] 4

2

i

I

I

I

4

6

8

10

(10-4M)

,.01/

10

20 30 T i m e (rain)

0

50

Based on the known mechanism of carbon-catalyzed lowing scenario

is proposed

onto the surface and becomes cools,

porarily

sulfate.

are exposed to air, oxygen is chemisorbed

"activated"

by the adsorption process.

When this aqueous

active oxygen at the solid-liquid The product

the fol-

As the com-

a film of liquid water will condense on the particles,

in which SO 2 can dissolve.

sulfate.

(ref.17),

to explain the formation of particulate

When freshly formed soot particles

bustion effluent

reaction

interface,

sulfate dissolves

free of oxygen.

sulfite gets in contact with the it will immediately react to form

in the water,

This constitutes

leaving the surface tem-

the rapid stage I process.

Dissolved

oxygen from the water will slowly diffuse to the carbon surface to f o r m a s i m i l a r surface oxygen complex, cess is diffusion

which will also react with aqueous sulfite.

controlled

and is responsible

for the slower

This pro-

(stage II) cata-

lytic oxidation process.

Because these processes occur on the liquid-solid

terface,

of condensed

a few monolayers

water are sufficient.

[) process will lead to sulfate formation the condensation

The rapid

in-

(stage

in the plume as soon as conditions

of liquid water and solution of SO 2 have been reached.

for

Parti-

culate sulfate will be formed by this mechanism relatively close to the source and could therefore be termed quasi-primary Soots produced by combustion

of propane,

soot have been found to behave qualitatively

sulfate. natural

gas, and automotive

like activated

like the one in Fig. 6b, S/C' ratios for the fast oxidation for different and natural

carbons:

0.004,

gas, respectively.

of S/C' ratios determined

carbon.

(tunnel)

From graphs

stage are derived

0.05, and O.1 for activated carbon,

tunnel

soot,

The latter two agree well with the lower ranges

from ambient measurements

(c.f. Fig. 4).

The following arguments

show the consistency of these results with the pro-

posed surface chemical mechanism.

Assuming that each sulfite ion occupies

area of 3 x i0 -15 cm 2 on the surface at the time of reaction, vated carbon must have a solid-liquid in order to accommodate

interfacial

0.004 grams of sulfur.

surface area obtained by nitrogen adsorption. most of the gas-solid available

an

one gram of acti-

area of approximately

22 m 2

This area is about 5% of the This result seems correct because

surface area is the internal pore area that may not be

to reactions

of the liquid-solid

face areas calculated

for tunnel

560 m2/g C, which are reasonable mation have diameters

interface.

soot and natural

The liquid-solid

inter-

gas soot are 280 m2/g C and

because soot particles

immediately

from about 50 to a few thousand ~ngstr~ms.

after for-

Such particles

have surface areas of hundreds of meters square per gram. During the short time when sulfate formation occurs, surface area is available

for reaction.

most of the particle

After residing

in the atmosphere,

surfaces become covered with sulfate as well as other species, surface that is chemically and physically Recent research in contrast oxidation

different

in a

less active.

in our laboratory has shown that aged ambient aerosol particles,

to source particles,

are practically

inactive as catalysts

for SOo

(ref.20).

From the foregoing discussion, particles

it is plausible

can produce enough particulate

er limit of the S/C' ratios observed possible explanations

that surface reactions

sulfur to account

in ambient air.

for the higher S/C' ratios.

lack of size segregation

There are at least two

more likely to be present at high carbon concentrations, surface areas.

be produced after longer reaction

cess, which could also be enhanced by nitrogen oxides oxidation mechanisms

involving peroxide-type

The concept of primary oxidants new because secondary

organic peroxides incomplete

However,

(ref.21),

in the context of atmospheric

it is known

(ref.22)

or by the SO 2

chemistry

is

including peroxides,

are

reactions, and under conditions of

strated that propane flames and internal combustion

We have recently demonengines emit substantial

amounts of hydrogen peroxide and organic hydroperoxides, in water

pro-

that hydrogen peroxide and

could appear in the effluent.

sulfite to sulfate when dissolved

sulfur can

primary oxidants.

are formed in precombustion

combustion,

particulate

(stage II) oxidation

it is generally assumed that all oxidants,

in origin.

Large particles,

have lower S/C' ratios

Additional

times by the slower

systematic

could be related to the

of the samples used in our studies.

because they have low specific

on soot

for at least the low-

The apparently

increase of the S/C' ratio at low carbon concentrations

(ref.23).

unit of fuel consumed depend greatly on the combustion a very turbulent propane flame produces more peroxide flame.

resulting

and generally

the

which rapidly oxidize

Their emission rates per conditions.

For example,

than a steady diffusion

This in turn produces more than a premixed propane-air

flame,

and a

I0 premixed propane-oxygen flame produces the least.

These results imply that more

primary peroxides will be produced by sooty flames than by nonsooty ones.

The

sulfate produced by the reaction of primary peroxides would therefore generally follow the soot concentrations,

thereby retaining the S-C' correlation.

The

reactions involving primary oxidants, in addition to forming particulate sulfates, may play a role in cloud and fog chemistry, and thus contribute to the overall atmospheric sulfate burden.

REFERENCES

1

2 3 4 5

6 7

8 9 I0 ii

12 13 14 15 16

17 18 19 20 21 22 23

T. Novakov, in H. Malissa, M. Grasserbauer and R. Belcher (Eds.), Nature, Aim and Methods of ~icrochemistry, Springer-Verlag, New York, 1981, pp. 141165. E.C. Ellis and T. Novakov, Sci. of Total Environ., 23 (1982) 227-238. H. Rosen and T. Novakov, Appl. Opt., 22 (1983) 1265-1267. L.A. Gundel, R.L. Dod, H. Rosen and T. Novakov, these Proceedings. A.D.A. Hansen, H. Rosen, R.L. Dod and T. Novakov, in T. Novakov (Ed.), Proceedings, Conference on Carbonaceous Particles in the Atmosphere, Lawrence Berkeley Laboratory report LBL-9037, Berkeley, 1979, pp. I16-121. H. Malissa, H. Puxbaum and E. Pell, Z. Anal. Chem., 282 (1976) i09-i13. R.L. Dod, H. Rosen and T. Novakov, in Atmospheric Aerosol Research Annual Report 1977-78, Lawrence Berkeley Laboratory report LBL-8696, Berkeley, 1979, pp. 2-I0. E.C. Ellis, M. Zeldin and T. Novakov, these Proceedings. S.K. Friedlander, Ibid. C. Brosset, Ambio, 2 (1973) 2-9. T. Novakov, in Proceedings, Second Joint Conference on Sensing of Environmental Pollutants, Instrument Society of America, Pittsburgh, 1973, pp. 197204; and T. Novakov, S.C. Chang and A.B. Harker, Science, 186 (1974) 259-261. B.R. Appel, E.L. Kothny, E.M. Hoffer, G.M. Hidy and J. Wesolowski, Environ. Sci. Technol., 12 (1978) 418-425. T. Novakov, unpublished data. M. Bizjak, V. Hudnik, S. Comiscek, A.D.A. Hansen and T. Novakov, these Proceedings. H. Rosen, unpublished data. C. Brosset, in T. Novakov (Ed.), Proceedings, Conference on Carbonaceous Particles in the Atmosphere, Lawrence Berkeley Laboratory report LBL-9037, Berkeley, 1979, pp. 95-I01. R. Brodzinsky, S.G. Chang, S.S. Markowitz and T. Novakov, J. Phys. Chem., 84 (1980) 3354-3358. P. Middleton, C.S. Kiang and V.A. Mohnen, Atmos. Environ., 14 (1980) 463-472. S.C. Chang, R. Toossi and T. Novakov, Atmos. Environ., 15 (1981) 1287-1292. W.H. Benner and T. Novakov, in Environmental Research Program Annual Report FY 1982, Lawrence Berkeley Laboratory report LBL-15298 (in press). W.R. Corer III, D.R. Schryer and R.S. Rogowski, Atmos. Environ., 15 (1981) 1281-1280. See, for example, D.J. Patterson and N.A. Henein, Emissions from Combustion Engines and Their Control, Ann Arbor Science, Ann Arbor, 1972, 347 pp. M. McKinney, W.H. Benner and T. Novakov, in Environmental Research Program Annual Report FY 1982, Lawrence Berkeley Laboratory Report LBL-15298 (in press).