JOURNAL
OF GEOPHYSICAL
RESEARCH, VOL. 102, NO. D3, PAGES 3671-3682, FEBRUARY
20, 1997
Carbon aerosols and atmospheric photochemistry
D. J. Lary, 1 A.M. Lee,• R. Toumi,2 M. J. Newchurch,
3 M. Pirre,4 and J. B.
Renard 4
Abstract.
Carbon aerosolsare produced by all combustion processes. This
paper investigatessome possibleeffects of heterogeneousreduction of atmospheric
constituents on carbon aerosols. Reduction of HNO3, NO2, and 03 on carbon
aerosols may be an important effect of increased air traffic that has not been
consideredto date. It is shown that if HNO3, NO2 and 03 are heterogeneously
reduced on atmospheric amorphous carbon aerosols, then a significant, lower
stratospheric ozone loss mechanism could exist. This ozone loss mechanism is
almost independent of temperature and does not require the presenceof sunlight.
The mechanismcan operate at all latitudes where amorphouscarbon aerosolsare
present. The relative importance of the mechanismincreaseswith nightlength. The
reduction of HNO3 on carbon aerosolscould also be a significant renoxification
processwherevercarbon aerosolsare present. Owing to the very different soot levels
in the two hemispheres,this implies that there should be a hemisphericassymetry
in the role of these mechanisms. The renoxification lea,ds to simulated tropospheric
HNO3/NO• ratiosthat are closeto thoseobserved.In contrastto the stratospheric
response,the troposphericproduction of NO• due to the reduction of HNO3 would
lead to tropospheric ozone production.
Introduction
erally overestimate
the HNO3 abundance,particularly
The
recent
World
Meteorological
Organization
(WMO)
inthetroposphere
(e.g.
Chatfield,
1995);
third,
totroassessments
[WMO,
1992,
1994]
reported
thatforthe pospheric
ozone
production.
first time there were statistically significantdecreasesin
ozone in all seasons in both the northern
and southern
Heterogeneous Reactions on Carbon Aerosols
hemispheres
at mid latitudesand high latitudes during
Thlibi andPetit [1994]studiedthe interactionof soot
the 1980s,and that most of this decreaseoccurredin with NO v. They found that hydrogenatoms on the
the lowerstratosphere.This hasalsobeensupportedby soot surface play an important role in the interaction,
trendsderivedfrom ozonesondes
[Logan,1994]. If this HCN being observedin the desorbedproducts. Thlibi
ozone loss is due to in situ chemistry, the mechanism andPetit [1994]foundthat at the temperaturescloseto
involved must be able to operate at all temperatures.
thosefound in the atmospherethe gas/solidinteraction
In contrast to the ozone loss that occurs in the strato-
sphere,thereis an observedozoneincreasein the troposphere.This paperexaminesthe role of amorphouscarbon aerosols
in reducingatmosphericconstituents.This
is relevant to several issues,including the following:
first, midlatitude,lowerstratospheric
ozonelossand its
temperature, altitude, and seasonaldependence;second, to atmosphericrenoxification,wheremodelsgen• CentreFor AtmosphericScience,CambridgeUniversity,Cambridge, England.
quickly converted both NO2 and HNOa into NO. NO
was the main product formed, with CO2, CO, N2, and
N20 being minor products. (It wouldthereforebe interestingto seeif slightly larger N20 concentrationsare
observedin aircraftflight corridors.)NO reactedon the
soot at a rate which was at least i order of magnitude
slower than both NO2 and HNOa.
In the presenceof oxygen,NO is expectedto be par-
tially convertedinto NO2 [Thlibi and Petit, 1994]. It is
likely that three of the NOv/soot reactionsthat will be
important for the atmosphere are
2Departmentof Physics,Imperial College,London,England.
3Earth SystemScienceLaboratory,Universityof Alabama in
HNOa carb?n
NO2
Huntsville.
4Laboratoire de Phsyique et Chimie de l'Environnement.
CNRS, Orleans. France.
HNO3
0148-0227
/ 97/ 96JD-02969$09.00
•
NO
NO2 ca-L•
n NO
Copyright1997 by the AmericanGeophysicalUnion.
Paper number 96JD02969.
carbon
(1)
(2)
(3)
The exact fate of the oxygenand hydrogenatoms in
theseprocesses
is important and needsto be examined
3671
3672
further.
LARY
Direct
ET AL.: CARBON
AEROSOLS
AND ATMOSPHERIC
ozone loss can also occur on carbon
aerosols.
Oa
c•rbon
>
O2
(4)
PHOTOCHEMISTRY
The evaluationof DeMote et al. [1994]recommends
a reaction probability for Os on carbon/sootof 3 x
10-2. Fendeland Ott [1993]reportfast Os losson 10
to 100 nm solidcarbon agglomerates,with an estimated
Taboret al. [1993, 1994]measuredthe accommo- reactionprobabilitynear3 x 10-2. Fendelet al. [1995]
dation coefficientfor NO2 uptake on solid amorphous report that submicron carbon or iron aerosolparticles
carbon. Like Thlibi and Petit [1994],they foundthat destroy ozone efficiently;the sticking coefficientof Os
NO2 was reduced on the solid amorphouscarbon to
to the particlesis of the orderof 10-4 . They conclude
yieldNO. If atmospheric
NO2 is reducedin thiswayon that particles present in the stratospheremay repre-
solidamorphouscarbonthen it representsa significant sent a significantsink for Os. Smith et al. [1988]renighttimelossof Os asthe reductionof NO2 is actually port that the ozone/sootreactionis first orderin ozone,
with CO, CO2 and H20 the only stable gaseousprodthe rate limiting step of the catalytic cycle
ucts. Stephenset al. [1986]measuredCO, CO2 and
NO•
carbon
NO+Oa
NetOs
O2 as products with an O2 produced for each O3 re-
>
NO
>
NO•+O•
carbon
•
(5)
acted. Stephenset al. [1986]measureduptake coefficientswhichvariedfrom 10-s to 10-5 depending
on
02
the carbon sample and Os exposure. The Os reaction
probability on carbon aerosolsis clearly dependent on
This ozonelosscycleis unusualin that it can occur the carbon aerosol'ssurface history. Consequently,in
duringthe night. It doesnot requirethe presence
of this study,a stickingcoefficient
of 1 x 10-5 wasused
sunlight.However,the cyclealsoproceeds
duringthe for Os to O2 conversionas a lower limit, so that the
day. The cyclehas little effecton other NOv species effects on the ozone should be at least those reported
because as soon as the NO
is formed
it is almost
im-
mediately converted back in to NO2. This cycle should
play a role whenever there are carbon aerosols. There
is, of course,the possibility that the active sites for physisorption will become saturated and so will take no
further part in the heterogeneousprocess. If this is
the case, then desorptioncould be initiated by heating
and the presenceof UV light. Another important issue
is whether oxygen atoms are produced;if they are, this
may partially or completelynegate the ozonelossmechanism just described. These are all areas that need to
be further investigated.
here, and, in fact, will probably be greater.
Production,
Carbon
Transport, and Characteristics of
Aerosols
Atmospheric soot particles are produced by the incomplete combustion of fossil and other fuels. Combustiongeneratedaerosolscan affect the atmospherein
two main ways, via light absorption(e.g., Shaw and
Stamnes, 1980; Porch and MacCracken, 1982; Cess,
1983;Rosenand Hansen,1984) and via heterogeneous
reactionson their surfaces(e.g., Tabor et al., 1993,
Taboret al. [1993, 1994]report an accommodation1994). Heterogeneous
reductionof atmosphericcon-
coefficientfor NO2 uptake on solid amorphouscarbon stituents on carbon aerosols is a possibility that has
of (4.8 4- 0.6) x 10-2. They foundthat the onlyma- been largely overlookedup until now, but it could acjor product wasNO corresponding
to about 60% of the tually be very important.
NO2 adsorbed on the surface. Consequently, in this
The present subsonicair traffic occursmainly in the
study, a '/value for the heterogeneousreductionof NO2 northern hemisphere,with about 30% to 50% flying
to NO of 2.8 x 10-2 wasused. ThlibiandPetit [1994] abovethe tropopause[Schumann,1994]. Future highspeedcivil transport(HSCT) systemshavebeenpro-
found that the reduction of HNOs into NO also oc-
curson carbonaerosols.Rogaskiet al. [1996,alsoper- posed to fly in the middle and lower stratosphere. In
sonalcommunication,1996]measuredan uptake value 1990, about 176 Mt of aviation fuel was used. To put
of (3.8 4- 0.8) x 10-2. For the lengthof their exper- this into context,this aviationfuel constitutesabout 6%
iments(<45 min) they observedno time dependence.of all petrol productsand providesabout 3% of the CO2
They found that the major products of the heteroge- releasedby the burningof fossilfuels. Global fuel conneous interaction of HNOs with soot were H20, NO2, sumptiongrowsby about 3% per year, with a doubling
and NO. They lookedfor O2 and N2 productionbut did expectedwithin the next 18 to 25 years. Presentemisnot observeany. They determinedthe product yield for sionsby spaceflight are about a factor of 10,000times
each species.NO2 was the dominant NO• product. It smaller than those from aviation. The present aviawas a factor of 5 larger than the NO yield. They mea- tion traffic has already caused a considerable increase
suredthat, on average,for every three HNOa molecules in NO• concentrations
(between30% to 100%)in the
lost to the surface, two NOx moleculesare releasedto uppertroposphere
alongmain flight routes[Schumann,
the atmosphere.Consequently,in this study, a sticking
coefficientof 4.2 x 10-s wasusedfor HNOs to NO conGraphitic carbonparticlescan be transportedon the
versionanda stickingcoefficient
of 2.1 x 10-2 wasused global scaleand so are able to reach remote regions.
for HNOs to NO2 conversion.
Long-rangetransport of graphitic carbon particlesof
LARY
ET
AL.-
CARBON
AEROSOLS
AND
ATMOSPHERIC
PHOTOCHEMISTRY
3673
at least 2000 km from the closestsignificantsourcere- It is rather surprisingthat most of the soot aerosolsare
gionhasbeenobservedat groundlevel stationsthrough- not actually entrained in sulphate aerosols. Blake and
aerosols
out the westernArctic [Stonehouse,
1986;Rosenet al., Kato [1995]speculatedthat aircraft-generated
1981; Heintzenber9,1982; Rosenand Novakov,1983]. may constitute poor condensationnuclei. However,
Trace elementand meteorologicalanalysessuggestthat Kiircher et al. [1996]find strongevidencethat aircraft
even longer-rangetransport from midlatitudes between soot is responsiblefor the buildup of visible contrails.
5,000 and 10,000km awayhad taken place [Rahn and KSrcheret al. [1996]showthat the observationis consistent with model results under the assumptionthat
McCaffrey,1980;Bartie et al., 1981].
Horizontal profilesof graphitic carbon particles have soot getscoatedby a liquid H2SO4/H20 solution,posbeen provided by missionssuch as the National Oceano-
graphicand AtmosphericAdministration(NOAA) airplane sampling programme during March and April
1983 [Schnellet al., 1994; Rosenand Hansen, 1984;
Hansen and Rosen, i984]. it has been found that
graphitic carbon particles are presentthroughout the
siblyalsocontainingHNOa, andthat the sootcoretriggers heterogeneous
freezing of water ice during plume
cooling. They found that the observationscannot be explained by assumingdry soot upon which ice nucleates
results strongly point toward the fact that aircraft soot
arctic tropospherewith upper layers typically contain- leavesthe jet enginesalready entrainedin a liquid sulfuric acid solution. This is also very likely to be true in
ing more particles than at ground level.
Pusechelet al. [1992]reportedthat upper tropo-
cases where no visible ice contrails
form.
Even if soot
spheric aircraft emissionsof soot presently represent would be emitted without coating, quick coagulation
with in situ nucleatedH2SO4/H20 aerosols
approximately0.3% by massof the backgroundstrato- processes
sphericaerosols.Blake and Karo [1995]recentlypre- might entrain them into droplets. This would be simisented an overview of black carbon soot measurements
lar to the Kuwaiti oil fire plumesstudied by, for exammade in the upper troposphereand lower stratosphere, ple, Parun9o et al. [1992],where carbonhad become
k.-,+......
Ar•Oq •,•A ONONT nncl they annfirm
thi,q finclinp'.
entrained within liquid droplets. It could be that the
They alsopoint out that for volcanicallyquiescentperi- soot is entrained in dropletscloseto the aircraft, but in
ods, if surfaceareas are consideredinstead of mass, the the far-field situation as measured by Blake and Karo
surface area of carbon aerosolsis comparable to that of [1995],the sootis no longerentrainedwithin a droplet.
It seemsthat biomassburning does not contribute
sulphate aerosols. In fact, the surface area of carbon
aerosolsmay be more than that of sulphate aerosols significantly to the upper tropospheric, lower stratoduring volcanically quiescentperiods. This is because sphericsootloading. Blake and Karo [1995]foundthat
carbon aerosolsare typically not spherical, but rather the measured latitudinal distribution of black carbon
havea fractal geometry.Blakeand Karo [1995]consid- soot between 10 and 11 km covaried with commercial
ered two extreme cases and calculated the surface area
air traffic use, suggestingthat aircraft fuel combustion
of the carbon particles as if they were all spheresor as if is the principal source of soot at this altitude. In adthey had a purely fractal geometry.The fractal surface dition, they found that at latitudes where there is a
area calculationgavean area that was 30 times greater lot of commercial air traffic, significant levels of black
than the area calculated if all the carbon aerosols were
carbon soot were measuredeven at 20 km. This suggests that aircraft-generated soot injected just above
assumedto be 20 nm spheres.
Colbeck
and Nyeki[1992]presenteda reviewof fractal the tropopausemay be transported to higher altitudes.
structures as applied to environmental aerosols. They By assigning upper and lower estimates on the total
point out that the atmospheric lifetime of particles is fuel burned in the stratosphere by aircraft and compardependent on their terminal velocity, which is related to ing this to the measured soot concentrations, a black
the fractal dimensionof the particles. Many natural ob- carbon soot residence time of between 4 and 12 months
jects such as coastlines and clouds may be represented wasderivedby Blake and Karo [1995].
Baum#ardneret al. [1996]report resultsfrom a new
by fractal theory. A basic property of fractal objects is
instrument which can simultaneously measure aerosol
that they obey the scalingrelationship'
diameter between0.4 and 10 ttm and whichwasrecently
N ocBD
(6) flown on the NASA ER2 aircraft during a stratospheric
where N is the number of features, R is the resolution measurement campaign. The measured stratospheric
of measurements,and D is the fractal dimension of the refractive indicesdo not agreewell with theoretical preobject. Fractal clusterssedimentmore slowly than com- dictions, and vertical profiles suggestthe presenceof
pact spheresof the samemass.Berry [1989]calculated nonsphericalor absorbingparticlesin the altitude range
that clusters composedof 1000 individual spherulesof of 7 to 9 km. One possibility is that Baumgardner et
radius20 nm fall at approximately
100m yr-• if D:1.8 al. [1996]observedcarbonaerosols.
ascompared
with i km yr-• for solidclusters(D:3).
In the Arctic boundarylayer a low HNOa/NOx ratio
Blake and Kato [1995]state that the carbonaerosol has been inferred from measurements. Therefore, if the
distributions they presented did not account for soot soot reduction actually occurs, since the Arctic is where
that had been entrained within sulphuric acid droplets. soot accumulates,we would expect to seethe low values
3674
LARY ET AL.' CARBON
AEROSOLS AND ATMOSPHERIC
that are seen. The observationstherefore subjectively
profilesover Kiruna
support the reduction of HNO3 on soot.
These findings point to the potential importance of
NOu/amorphouscarboninteractionson the local and
iiiIillill
i,iiiiiilliiiiiiiiiii
t
10-
global scalein the stratosphereand the troposphere.If
the aviation fuel consumption is set to double in the
next 18 to 25 years, it is likely that the carbon surface
area will also double, while increasesof up to a factor
.,•!
of 10 couldoccurin flight corridors[WMO, 1994].This
/
lOO
opens the possibility that on the hemisphericscale the
carbon surface area could be double that
PHOTOCHEMISTRY
of the back-
-
groundsulphateaerosolsurfaceareaduringvolcanically
quiescentperiods, and 10 times that of the background
ß
AMON data
Simulation A
sulphate aerosolsurfacearea in flight corridors. Such a
........ Simulation B
Simulation C
large surface area could therefore be a potentially im1000
portant site for heterogeneousatmospheric chemistry,
0.001
0.01
0.1
1.0
10.0
which really should be thoroughly investigatedwith a
(ppbv)
senseof urgency.
It is not only nitrogen speciesthat could be reduced Figure 1. SecondEuropeanStratosphericArctic and
_
on carbon aerosols. For example, HOC1 could be reduced to HC1, and HOBr could be reducedto HBr. This
could be significantin polar regions,where the production of HC1 can becomethe rate limiting step for further
chlorine activation. These processesalso need detailed
examination.
Mid-latitudeExperiment(SESAME) campaigncomparisonfor the simulatedNO2 profilesover Kiruna on
February10, 1995. SimulationA (solidline) wasa control simulation
whichincludes
a standardstratospheric
chemistry
scheme.Simulation
B (dashedline) included
the reductionof HNO3 and NO2 on carbon aerosols,
both with a 7 valueof 2.8 x 10-2 and a carbonaerosol
surface
areaof 1 tzm2/cm
3. Simulation
C (dot-dashed
line) wasthe sameassimulationB, exceptthe 7 value
Comparison With Observations
for the reductionof HNO3 was reducedby a factor of
Since there is uncertainty associatedwith the rate at
which HNO3 and NO2 is reduced on carbon aerosols,
this sectioncomparessomeobservationswith model
10.
culations that include the reduction of HN03 and NO2
heterogeneouscarbon reactions. Simulation B includes
on carbon
aerosols.
the reductionof HNO3 and NO2 on carbonaerosols,
both with a 7 value of 2.8 x 10-2. SimulationC was
SESAME Comparison
the same as simulationB, except the 7 value for the
Figure 1 showsthree simulationsthat wereperformed reductionof HNO3 was reducedby a factor of 10 to
with a three-dimensional
chemicaltransportmodel[Chip- a value of 2.8 x 10-3. The simulations included the
perfieldet al., 1996],whichusedmeteorological
analy- hydrolysisof N205, C1ONO2,and BrONO2 on sulphate
sesto specifythe circulation. The model was integrated
for 12 days and usedresultsfrom a seasonalintegration,
which had been integrated from November 22, 1994, in
aerosols.
Figure 1 showsthat the NO2 observationsare brack-
gration had been initialized itself from a combinationof
etedby simulations
that includedheterogeneous
carbon
reactions(simulations
B andC). This is not conclusive
proofthat the proposedcarbonmechanismdoesoccur,
two-dimensional
but it does make it clear that without
order
to initialize
each simulation.
chemical
The
fields and data
seasonal
from
inte-
instru-
some kind of
ments on board the UARS satellite. This integration is
presentlybeing usedto investigatethe anomalouslylow
renoxificationmechanism,whether it is the mechanism
postulatedhere or someother mechanism,the vertical
03 amountsseenduringthe 1994/1995northernwinter profilesof NO2 arenot wellsimulatedwhenthe hydroly(as observedby Manne•let al. [1996]). The modelwas sis of N205 on sulphateaerosolsis included. The N20,
sampled during each simulation at the same time and hydrolysisused in this study is the temperature and
position relevant to measurementsof NO2 made by 'Ab- compositiondependentdata recommendedby DeMote
sorptionpar MinoritairesOzoneet NOx' (AMON) at et al. [1994]. A slowerN20, hydrolysison sulphate
night inside the polar vortex on February 10, 1995, over aerosolswould clearly also increasethe simulated NOx
Kiruna during the SecondEuropean StratosphericArc- concentration. The required renoxification mechanism
tic and Mid-latitude Experiment(SESAME)campaign needsto operate up to altitudes of around 25 to 30 km.
It is unclear what the transport mechanismwould be
[Hermannet al., 1996].
Simulation
A was a control simulation
that
includes
a standard stratosphericchemistry schemewithout any
for carbon aerosol to be found at these altitudes within
the polar vortex.
LARY ET AL'
CARBON
AEROSOLS
AND ATMOSPHERIC
ATMOS
Data •
0.2/.zm%m
• soot
3675
77.1oS, 147.1oE, 15:46 GMT, 9 November1994
44.6øN, 140.3øE, 7:17 GMT, 6 November1994
ß
.....
PHOTOCHEMISTRY
ß
.....
0 •m2/cm
• soot.............
0.1fi,m%m
• soot
0.5fi,m%m• soot..... I/•m2/cm
• soot
35•
ATMOSData
0.2/.zm%m
• soot
0 •m2/cm
• soot.............
0.1/•m2/cm
• soot
0.5fi,m2/cm
• soot..... I/.zm%m
• soot
35-
-35
-35
_
_
_
-30
I
• 30--
_
30
I
_
_
_
25
'O
_
::3
_
_
•
-
{
I
I
-2O
_
_
_
_
_
_•15
15-
L•5
15-_
_
_
_
•10
-10
• •{
0.03
0.06
100.01
0.1
HNOJNO•
o.o3
o.06
0.3
0.1
o.6
1.o
3.0
6 0 10.0
HNOJNO•
Figure 1. SecondEuropeanStratospheric
Arctic and Mid-latitudeExperiment(SESAME) campaign comparisonfor the simulatedNO2 profilesover Kiruna on February 10, 1995. Simulation
A (solidline) wasa controlsimulationwhichincludesa standardstratospheric
chemistryscheme.
SimulationB (dashedline) includedthe reductionof HNOa and NO2 on carbonaerosols,
both
with a 7 valueof 2.8 x 10-2 and a carbonaerosolsurfaceareaof 1 pm2/cm3. SimulationC
(dot-dashed
line) wasthe sameas simulationB, exceptthe 7 valuefor the reductionof HNOa
was reduced by a factor of 10.
ATMOS
ATLAS3
Comparison
Figure 2 showsthe effect of the reduction of HNO3
the northernhemispherehas a larger sootloadingthan
the southernhemisphere.The best agreementbetween
the model simulations
and the ATMOS
ATLAS3
ob-
and NOa on carbonaerosols(Table 1) on the vertical servationsin the northern hemispherefalls closeto the
profilesof NOa and the HNO3/NO•: ratio in the north- simulation
with 0.2/•m2/cm3. This is in generalagree-
ern and southern hemispherescompared to the Atmo-
sphericTraceMoleculeSpectroscopy
(ATMOS) Experiment ATLAS3 measurementsmade during November
1994.
The
model
used was the
AUTOCHEM
column
ment with the SESAME comparisondescribed above.
In contrast, the best agreementbetween the model simulations
and the ATMOS
ATLAS3
observations
in the
southern hemisphere falls close to the simulation with
modeldescribedby Lary [1996]and Lary et al. [1995, 0.1/•m2/cm3. This is consistent
with the fact that the
1996]. The columnmodel was run for 7 days starting southernhemispherehas a lowersootloading. Figure 2
from the ATMOS ATLAS3 vertical profiles of temperalsoshowshow sensitivethe HNO3/NOx ratio is to the
ature, 03, NO, NO2, N205, HNO3, HNO4, C1ONOa,
H20, CO, CO2, CH4 and NaO. Below 25 km the diurnal correction to the NO2 profiles is very important
[Newchurch
et al., 1996].
soot loading. An increaseof the soot loading by a fac-
tor of 10 reducesthe HNO3/NOx ratio by a factor of
approximately 100 in the lower stratosphere. Including the reduction of HNO3 on soot leads to a simulated
It can be seenfrom Figure 2 that above25 km there is HNOa/NOx ratio within the observed
rangeof the averexcellentagreementbetweenthe modeland the observa- age free troposphericratio of 1-9 reported by Chatfield
tions in both hemispheres.Below 25 km there is a slight [1995].
contrast between the hemispheres. As noted above,
• .... c comparisons• not conclusivelyprove that the
reduction on carbon aerosolsactually occurs but they
show that the reduction of NO2 and HNO3 is at least
Table
1.
Carbon
Reduction
Reactions
Included
in
This Study
sphericassymetryin the observedNO2 and HNOa/NO•
Reaction
HNO3
HNO3
NO2
03
consistent with observationsand can explain the hemi-
carbon
>
>
carbon
>
carbon
>
carbon
7 used
NO2
NO
NO
02
2.1 x 10-2
4.2 x 10-3
2.8 x 10-2
1 x 10-s
ratio vertical profiles. To say this another way, a renoxification mechanism, whether it is the one postulated
here or a completely different mechanism, can improve
the agreement between observationsand simulations of
the observedNOs partitioningif it is more effectivein
the northern hemispherethan the southern hemisphere.
Toumi et al. [1993]consideredthe ATMOS HNO3/
NO• ratio and concluded that the current recommen-
3676
LARY ET AL.: CARBON AEROSOLS AND ATMOSPHERIC PHOTOCHEMISTRY
dations for the sulphate N2Os lossare too fast to agree HNOa/NO• ratio [Chatfield,1995].Figure3 alsoshows
with observations. Considineet al. [1992], studying that the heterogeneousreduction of NO2 and Os on
limb infraredmonitorof the stratosphere
(LIMS) data, amorphouscarbon aerosolis quite rapid. The rate of
and McElroy et al. [1992]usingATMOS data, alsocon- Os reduction is likely to be faster than that displayed
cluded
that
the current
recommendations
for the sul-
in Figure 3, as a lowerlimit 7 valueof only 10-5 has
phate N2Os lossare too fast to agreewith observations. been used.
Figure 2 showsthat including heterogeneousreduction
Including the heterogeneousreduction of NO2 and
of HNO3 on carbon aerosolsimproves the agreement HNOa by amorphouscarbonaerosolsincreases
the night-
betweenthe simulatedand observedHNO3/NOx ratio
time NO concentration(Figure4). Without the hetero-
in the lower stratosphere of the northern hemisphere geneous
reductionof NO2 and HNOa modelspredictan
by providing a renoxificationmechanism.Now that we NO/NOy ratio that is almostzero.
have seen that the reduction of HNO3 and NO2 is at
least consistent with observations, the next section ex-
Figure3 shows
thatthemidnightrateoftheheteroge-
neousreductionof NO2 on amorphouscarbonaerosolis
amines in detail the sensitivity of the proposedmech- a functionof both temperatureandthe amorphous
carbon aerosol surface area. The rate varies between 10a
anisms to carbon surface area, temperature, altitude,
and day length.
molecules
cm-3 s-• for an amorphous
carbonaerosol
Sensitivity Experiments
surfaceareaof lessthan i/•m 2 cm-3 and 106molecules
cm-3 s-• for an amorphous
carbonaerosol
surfacearea
To examine the role of HNO3, NOa and 03 reduction
of 10/•m2 cm-3 at 200K. Sincethe midnight
rateof
reaction of NO2 with O3 closelyfollowsthe rate of re-
on amorphouscarbon aerosolsin the lowerstratosphere
and the upper troposphereat mid-latitudes a set of idealisedmodel simulationswere performed. The numerical model usedfor thesesimulationswas AUTOCHEM,
ductionofNO2to NO ontheamorphous
carbonaerosol,
it can be seenfrom Figures3 and 4 that the reduction
of NO2 to NO on the amorphouscarbonaerosolis the
rate-limitingstepof the catalyticozonelosscyclemen-
a modeldescribed
by Lary [1995]andLaw et al. [1995,
1996]andFisherandLary [1995].In the simulations
a
tioned above.
stationary air parcel at 45øN was considered.The sim-
ulationsincludedthe hydrolysisof NaOs, C1ONOaand
BrONOa on sulphate aerosols.
With a reactionprobabilityfor O3 of just 10-5, at
low carbon aerosolsurfaceareasthe direct lossof O3
on the carbonis likely to be faster than the catalytic
lossdueto the production
of NO (Figure3). Forhigh
carbonaerosolsurfaceareasthe direct lossof O3 on the
carbonis likely to be slowerthan the catalyticlossdue
to the productionof NO. By wayof a comparison,
under
In the first set of sensitivity experiments the tem- the sameconditions,the corresponding
ozonelossrate
perature of an air parcel at 70 mb at equinox was at noon due to the reactionof HO2 with 03, i.e., the
varied between 200 K and 240 K and the amorphous rate of the major ozonelosscycle,is approximately105
Carbon Surface Area and Temperature
Dependence
molecules
cm-3 s-• at 200K, risingto 106molecules
10/•m2 cm-3. Foreachof theseconditions
a 7-daysim- cm-3 s-• at 240 K. An ozonelossrate that can reach
carbon aerosol surface area was varied between 0 and
106molecules
cm-3 s-• istherefore
clearlya significant
ulation was performed.
Although the observedhemisphericaveragecarbon
aerosolsurfacearea is typicallyaround1/•m acm-3,
much higher areas do exist locally, for example, in aircraft wakes. In addition, becauseit is likely that the
carbon aerosol abundance
will double over the next 25
ozone loss rate.
Figure 3 showsthat the midnight rate of the heterogeneousreduction of HNO3 on amorphouscarbon
aerosolis a function of both temperature and the amorphous carbon aerosolsurface area. The rate varies be-
years,carbonaerosolareasup to 10/•macm-3 were tween3 x 103molecules
cm-3 s-• for an amorphous
considered.
carbonaerosolsurfacearea of lessthan i /•m2 cm-3
Figure 3 showsthat the heterogeneousreduction of up to 8 x 103molecules
cm-3 s-• for an amorphous
HNO3 on amorphouscarbon aerosol is a significant carbonaerosolsurfaceareaof 10/•m2 cm-3 at 240 K.
renoxificationprocessthan can obviouslycontinuethroug]This renoxificationby the reductionof HNO3 provides
out the day. For example, Figure 4 is for midnight. additional NOs which can take part in the catalytic
When no carbon aerosolsare present, then the simu- destruction of ozone.
Figure 4 showsthat the additional ozone losscaused
lation predictsthat approximately80% of NOy is in
the form of HNO3. However, with the assumed7 val- by the heterogeneous
reductionof HNO3, NO2 and O3
ues, increasingthe carbonarea to just 1 /•macm-3, on amorphouscarbon aerosolover the 7-day simulation
a value close to the current hemispheric average, re- periodreaches
nearly3% (80 ppbv)for an amorphous
ducesthe HNO3/NOy ratio to approximately0.5. Thus carbonaerosolsurfaceareaof 10 •um2 cm-3. It is noterenoxification on carbon aerosolswill also be important worthy that the additionalozonelosscausedby the hetin the troposphere and could explain the current dis- erogeneousreduction of NOu is almost independentof
crepancy between observed and modeled values of the temperature if the 7 value is not temperature depen-
LARY
ET AL.:
CARBON
AEROSOLS
AND
ATMOSPHERIC
PHOTOCHEMISTRY
Midnightrate of
Midnightrate of
HNO --)
0
I
2
3
4
240L i___
l • .••
5
H
6
7
8
9
10
o
I
240 I
I .j...I !. I...•.....
I...t..I...t....I.._[..
:.:•.-•':.:
:.-:,...
'-'•:..
;,.'•
3677
,.::-..:...,..-....-......:.
.....
..................................................................
•'-
2
--) NO
3
4
5
6
-'•....j:•?,'•x-:•
*:-:.
:':':':...
::•.....:.:•.:j:•.:.::
::
;::-'::":..-...:.
_.::::;"'"':•.:..:.•
.
•'"'":•""
'"':•..•':::i:::'
<"•••:'•"•""'"•"'''"
"•':"•:
'""'
"'"'
""":<•:•:•
:<"'
"":"'
......
••-
7
8
9
o
'' -•••'.:.-d
'
:::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
'""•"'"
========================================
'"'""'"'
:':
........................................
"••
---•'•••:i:;
"•:. :"
220
i•1t'
•• ...............
f•
::;::':½::;'::::::'•':•'R"\v'
====================================
35795"'"'"""•':•'"'•J!?
I• '•J:..'f•
- ._.".. •..•:,•;;:- --.--• ::.<::'
:...'::.'.•
•.•
.........
•=•:-:-.-•.--__-•-_-•:_•__-_:.
• .-..-.•-•----------•-•----.-.-.--•.•.-.•..•..•••...;:.:•
• i::.,.'.•
• :':':'
•-.
::'""- :•.•i
38928 :•..'•::;ii:
:'
"'
!•'
:'•"'::'•:•
a '-<':'
.....................
'"'•"••'
"--"--'
• :•, .
220
""-'-':'-'-'-'-'-"
28255
5651
.....
•..,•,::.:.:.,+::
5207
-'.
•,•
;•i
-..
210
26035
4919'"'"•
:<"½'•'
'":'
'
4285":'"'"•:"•iiii
'""
24598
21427
3057
'""':"<<'""
"::•
:'
15287
200
0
200
I
2
3
Carbon
4
5
Aerosol
6
7
Surface
8
9
IJiJ•JJJJ•
.................
0
10
I
2
':•"::::'•'•"••'•••••••
3
Carbon
Area
4
5
Aerosol
6
7
Surface
8
g
10
g
10
Area
Midnightrate of
Midnightrate of
-> NO
0
I
2
3
4
5
6
7
8
g
.................. *'"'•-:':••:•::-••••••••::'?-.:,<•
,e..-.-'-•
................................
•..
.....
153506 -.:..-:...-•.
..-....•
:..,
.':. ::'..::::
•
126362
'""';'"
'";'•'•:•!•}!
'-"
•
''"
,•••...
•,,,,,,,•<:,,•.,
...........
,..•.;..<.•.•,::•
230-
v
200326
"!i'i'i'!!:
?:i!:!':' •
::::::::::::::::::::::
'- .:.:.-..
:.x, ;:.:.:.:.:.:.:-:.'
l:::
78179,•,•...
-,...•.,-<,,....,:
57395
""::•
•.•
•::..
37138 '"'-"•..,
.•.
.'-:•'.:
ß
2
Carbon
3
4
5
6
7
5
6
7
8
220
E
ß
:::::::::::::::::::::::::::::::::
..... .::...::.:::::
::-::....:
....
........
:...::::..:::::..:--<......,..-.:::
:•,,:-:•:-:.--,
'-•
.....<•:•
--...'.-.
.':.
......
...:::....-:.-:
.......
-....................
..........
.......•....••...•
ß:::.:-:-<.::::':::
,•:••••••:....
::::::::::::::::::
'"•:'::::'::::::-:,,,
,•:,..'.:....•::
,,,,•:•G
,,-,
,,,.
:::.:...:.:..:..
::.::,::::
,-..,
.:..::.:::..:
.. 55256......
•J•V•
::.
:..
::::.:::•
g::::...........•:.:.:.:::
•.,:..:5•.:::•!
• g...'"•'"':•<;::::'•
":'
•'::::::
'f 48753
...
•:•..
:•:.::..-:.-::..:..'::...-.:.::':.:?-:•..':.•
•o,:.':,•!•
.'.:<.:.::.
,•-..:
....
..."...'•-.
e..:?:
•'
:.:::::::::.:
',,
42347
'"'
'•':"i
.::::::::::::::::
36038
•....••.
•-.••--.•••••:...-::
...........................
.::::::..-:::
.....::,-
i"""
mo-
• ::•::.:.....-:::::::
.-<....:::.'::•::
--':.
:.. ,,..
:e.••.-..":.•...'.:•
o,
o,-:-:,.
::--.•::..-:::
•, ::.........
:-,,,..:
...........
•'
'•'•••"'...........-....-.. '"'. '•..•":-•:
:"":'"""'
"'"'"'
'"'"::*
..........
.:::::::::.:.
.[-
200
8
9
10
Aerosol Surface Area
•"-'""'"'"'"""
':"•:••:•••iii'
"•..:-',..•_
_,•'.-'.-""•
':'"•••••••••1:
••ltl•''
.......
..""ßß
f
0
I
I
[
I
[
2
Carbon
I
3
i
I
4
i
I
5
Aerosol
•
"',. :
"""
"•i:
•.
:'<<'"
....
::::;:::.::::..
::• 23698'•"'":'½
:..,,..,
.:...,..:,:.....
:,..::
,:::
:,.::::..:
,•,. :.-...:.::
..:.•..
.,.,,
:..'.,,:"'
:•.
.:•:::...:
.:::.:,,
. :.200 V"':.......
i.......
:'"":i-':'";'
.....:-•"r'W'm'.....i'.....,' 'i ",
I
4
•:" ":'•:•:ii
"'""
'"
.............. -
...............
• •.,:
"'••':••"-•i
..................
:":••'*''
ß'
. ß,•••••.-..•: ,.:,
:-'-':..::-'-'.-:::'
......
0
3
•J•['"'"'"'"'""'"•:':•:•:•:•••:..
:iii
:•••JJ•'":•":':""•••:•'.•.
....
•":•"
'• •"<
•!::...:'::::i?:::.:
'•...-'••::•-"..•.,,,<....,.%..,.'..-..-.,•.
•-'.'-'--.....
:<-:<--:...,s•::•-::::....
- ..'I
• 230-•------'-----'--"---"---•"•"•'•••'•'••"'•;•••
-'•
,•".
•
2
240 -
ß'....
•--..-....-.._._.
•.•...•.....,,,••....•••••••• .-,<--.•
I
o
10
....'-;-'-.__.._
..... :...._:_-:-:-:_-.s."-'---'-'--'-'-•:-•'.:.. '::::'...:.-:::<• :.%4.4.-:.•:.<:,-:
:,-.:::=.•
..:.:•
•.......-.......::
:..::
:,: ....
I
6
•
':
I
7
Surface
[
I
[
8
I
9
• 11692
]
i10
Area
Figure 3. Resultsfor midnight at the end of a set of 7-day simulations. During the 7-day
simulationsthe air parcelwas kept at 70 mb, 45øN at equinox. The temperature and amorphous
carbon aerosolsurfacearea present were kept constantfor each air parcel through out the 7-day
period. The resultsare plottedas a functionof temperature.
(in Kelvin) and the amorphous
carbonaerosolsurfaceareain pm• cm-a. Note that eachplot hasa differentcontourinterval.
dent. This is in contrast, for example, to the hydrolysis sphere,irrespectiveof the temperature, and it is likely
of N20.s where althoughthe 7 value is only weakly tem- that it contributes to the observed ozone trends. This
peraturedependent,the N205 concentration
is strongly mechanismwill increasein importance with the increase
temperaturedependent.For amorphous
carbonaerosol
surface areas that are closeto the hemispheric average,
the additional ozone loss is of the order of 0.3-0.5% over
the period of the 1-weeksimulation.
in the abundance of carbon aerosol. It is therefore a
mechanism that must be accounted for when consider-
ing the impact of increasedair transport and all combustion processeson the atmosphere.
Figure4 showsthat the nighttimeozonelossdue
Altitude Dependence
to the heterogeneous
reductionof HNO3 and NO• has
In the secondset of sensitivity experiments the altisubstantiallyreducedthe nighttimeozonelifetimefrom
around 350 months without reduction on carbon to just
tude of the air parcel was varied between 10 and 25 km.
Figure 5 showsthe additional 03 losscausedby includ52 monthswith a carbonareaof only 1 Fm2 cm-3.
ing
the heterogeneousreduction of HNO3 and NO2 on
Heterogeneous
reduction
ofHNO3,NO2 and03 thereamorphous
carbon aerosol as a function of altitude and
forerepresents
a potentiallyimportantin situozoneloss
mechanismwhich can operate throughout the day at
all latitudes in the lower stratosphereand upper tropo-
carbon aerosolsurface area. It is immediately apparent
that in the stratosphere the enhancedlevelsof NOx lead
3678
LARY
ET AL.:
CARBON
AEROSOLS
AND
ATMOSPHERIC
PHOTOCHEMISTRY
Midnight
Midnightrate of
HNOJNO•
NO + O• --) NO• + O•
0
I
240 I
2
3
4
5
' i • • I. • I •
•---...:-:-:-:•.,--.-::::::::::
.......
230•
6
7
8
9
0
10
I
3
4
5
,..I .,...J • •
I • J • J
-. x,•
2
6
7
•
8
9
10
•. [.. •..[. ].....!....1.....
•.'"•'•---',•••••••••
""•'"'•'"
'"••-':'-"""-•:""•'"'
'-"'"':'
"•'"'•
ß
'
-,,',,
-.-••••••:.,
0.511--'-'-:---'"•-'-•.•ii
•230-:::::::::::::::::::::::::::::::::::::::::
0.421
'"
'"•"
:::"
';':'
--' ::'•
•.--x,•,x,•-,,-•-,.**;<....*•:•:._-..-•_::•_
...'-:.,...-,
-'.--
v
0.339
-•--•••••
••••:-..........................
--N"'"'•i
"':"
'"
'"'"'"
..........
•,.c<;•...-<..-..-.:..-<.--cc:•-•
-.:.c:.-.::•.::¾
.•
• :::::';:-'.':-*,
:•.:..-.:.:.:.:,<•:&':.::.::•:
:•.A•,:,
•,
,,,
13i8•8
102822
.............
'•":
'84783
..........................
••"••••••••
•'"'""•'"'•"
-"•.
'•••-..-,,•'
""':"•ßß--ß
-'"'"'""•••:':'••••••••••-.-:it'••"''"
'....
'
'•-'"'""
'"""'
......................
'"'••••••••!i•
'••'"
'
,• 220-
0.245
ß ::::::::::::::::::::::::::::::
I-- 210-'
...............
:-'.-:-'.•:'•
-'•i•."
••:':"'•':•••:•....
..................
E
0.147
o.128
' 7959
•_yu
I
'
0
I
I
'
I
2
'
I
•
3
I
4
•
I
5
.•
I
•
6
I
7
'
I
•
8
I
9
•
0.221
0.195'""(•.'•
:.:.:
:.,.-.....:.....---..
:.:..--:.:.:c.:.:.:-:.-:•..-<½.:..:A:•-
:--:-:
..:.:-:.:---:.
:-:-.---.•.-:.:
•
200
I
0
10
1
2
3
4
Midnight
2
3
4
5
7
8
9
10
Midnight
NOJNO,
I
6
Carbon Aerosol Surface Area
Carbon Aerosol Surface Area
0
5
6
NO/NO•
7
8
9
10
240
_
J,...:.:.:•...:•.:.:...•!i....,....:•.:....:.....:•....
;•..:l.:•:.:.i!:i
.....
J......,..L...,J:........L,,J..,,.-.•.
......
0.657
.........,,•••••••.•--•------------•
..•.,,,-..•.....••,•:.
,,,,
,,
,,.,•••-
,.-, 230-
.......................
...:..,-.....::::::,:
,; ........... ,:, ,. ,. ,..... ,.......... ,.
%N•'""l•,,
.........
'x----"':-•:'••.:..:.!•,,:-;.':'½•:'
'••'"""--":•••;••••
''•"
•'•'- -' .......
"•'"'
'•'"'•••0.491..........
••••••
•._
.-<•.-,,•,:.:••,,
...........
•; ..,,,,-••••
v
,;..
:-:-:.:....
:..y:.:::::::::
,,,
230
0.446 ....................
0.372
""i•
!i•i•i:'-'
0.333
A•
t• 22o-
r•
•..
I-- 210- .,...::,•-..-.
0.304
•.,,,.
x..•
0.272 '""•'•<.....:'J:•
...,-'
............
0.219""•-.'{i
":*:•:•"
'"'''
_
•...,,,,,,•:•:•...,•••!
................
,,•.. ,.•.•
...•..,,.•
,:•...,,•..•':-•'•i
•............
'•.
..............
..-•
'"•
220
__•
.':•;•,
"'
;",
..'.::.' '""'•••:.:i•:i
•
_• •:-'-'.•..-'.;;•:•
• -•.-'.:'•i,':'
-:•'---'-•'"•'"'"':"•--'
"%
"'"'"'""'•"%'
....... •-
::..:,.•
.•":•
'•
"'
4.50
E_
04-:•;;;i
'""-:
:-•:-•......-.
-----
E ,.:----,•,,,•r
-,ß, -"•,•%••
:::::::::::::::::::::::::::::::::
2.58E-04
:.:.:..:....:•
I-. ,,.,.,
,.,...•
•"•"•"':••••••'.-•••:
..................
•.":•" ':"
.'/•,::i;:!:iiiiiii:ii::!•:;;!!!:!i!ii!i!i:i!::i:i::
1.43E-04
L•'•-';•'..-.••..-..-••..:..:ii:::::::•...•:.::::•%¾:.?'!::•::•?:?.:::.::i:.i:.i:::.:.?:•i
•"'"•"'""'""'•"'"'•'••':-,,,,•
..........
0.162
"'""'
':':'""':•'
:"•:•
":""
-• ................
• ......
•....•....,,
....-.-..-,,•,_____-.•...•:••:.••.......
•======================================
!•--'"
"""'
"'-'••,.
'":•':•--'
"•"''
"''
'%••'" ••-:::'"::.::i:'-?-'":.;!::::??.
½
•....--••.....,....•
........................
.•••••••:•j•.:::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
200
!•""""•'"'"'"'
'-''" '•'"
0
I
2
3
Carbon
::,oo
-•.•`••`:•••.••••••••;•i!•:•:•:.•;•::•:•::•::i:•::::•.•:.•::•:•:•
'%11•"'""""'•'
................
"'
.........
:.;:-;:.-:-:.-
••"'"'"'"
'•"
'"'•:'••'•••••'""'"
:'""
':•:
':':
::
':::
i:"
':':':'i":::
':,'
•:-":';
'•'
:'½
4
5
Aerosol
6
7
Surface
8
9
10
0
I
Area
2
3
Carbon
4
5
Aerosol
6
7
Surface
8
9
10
9
10
Area
AdditionalOzone loss(%)
due to reactions on carbon
0
I
2
I
ß '
'
3
4
5
6
7
8
240
...................
......•.........,,,..•...:,•••;:•,•,%
..........
•..
¾:•
•:,'"'"'"'•"•'••••....'••,..•
••,••••x,,•
,,••.....
•!!!i;:•!?•i?:::•..-::.
•:::•::•!
....
:-:•:>..',"'"""'-'"
'
:•.• • x :.:;:;::•:;•
2.74:::::::::::::::::
::::::::::::::::::::::::::
:::::' %
';;-'•'"-;---
.•'::::::::.
"-..:,.
...':.:.'.
<•$.-• •
-...,,:,:.-,
............
2.48 =========================
.....
:..--.-..................
•.:.:..,.-,:'
.....'•..-.:.:..-.:.:.:*'-'::::-"-':-•'•.
• :•,:':':":.,:':-'".',.•-,•
....:::.:!i:::&,..
.........................
-:,;%•, -.,.
ß;-----..-,---•-,-....
:4..-.::..:.:44•.''"---"-'-'-'-'-'-'-'-'-•-'-:.•••:....
-,:.-::::.'.-:
• "•
'--------------"-"-'
..........................
;:s.::.-'.::::,•,.'.•::::::•
..............
• •.=================================
.:::•7.".'.-.'.'.'.'.'.'.'.'.'.'.'.
2.32
::':':'::':'":::•
....:"•'"'""•?',•,i
!:;?,',ii?,i!!!
2.•5':'"""'""-"'•-?-::•-:.-:
::::-.
,-.i':-,"•.
"t•••'•
..i:i:i:i:ii::!::..•---•
.,:..'.:.,--.-:;•-:.:
-'•:------'•'::•--'-::
:........................
-........................
':::"-.•"'."":
'"'"'•'*•'••••
"•'"'"'•'
...-.....-.-.-.-.-.-.._-.
842.6'
::•"•ii
'"•:"
'"'":::'
'
595.2 :::::::
.................
========================
471.1"-'•?¾<.•
i::i•,
'" ,•_•
.................
.•..•••..
-,•,
,:•:•:!
.......
• :,,-...'.-o-;.:j.;.-ii.:'::,•.•!•½',•i'.,•.,•-,-'•.•:•11;ii:-•
=======================
1.73:,•'"'.'
.........
•::•
,,, ••....,....:••:•••:•••••\•i•iiii::::ill
..............
• .. ::::::::::::::::::::::::::::::::::::::::
391.8
?,•
-•"'•,••••'•"•,•,:iiiii:•::iiiiiiii•-.
...............
' ''
":::•-':"':':-':-'::::::•':::::';•::•'::•::'::'::'
•':';;:;::• .........
333.9
259.9
"'":'"':'"•'•
":-.-'
233.2
0
I
2
Carbon
3
4
5
Aerosol
6
7
Surface
8
Area
9
10
..........................................................
-•:.-,:•,:.:•:.-.:.:,.-,;•:,-,.•.•.•.
........
1.94
220 -•,.,
x -. •........
......
,,,....•,....•?,•.-:•,,•
• ,,;....
• , ::::::::::::::::::::::::::::::::::::::
.......................
0
•.:.:.:.:.:.
:::::::::::::::::::::::::::::::
:.::..::.:.:
:_:_:. 0.
I
2
3
4
5
6
7
Carbon Aerosol Surface
8
9
10
Area
Figure 4. Additional resultsfor midnightat the end of a set of 7-day simulations.During the
sevenday simulationsthe air parcel was kept at 70 rob, 45øN at equinox. The temperature and
amorphouscarbon aerosolsurfacearea presentwere kept constantfor each air parcel through
out the 7-day period. The resultsare plotted as a functionof temperature(in Kelvin) and
the amorphouscarbonaerosolsurfacearea in •m • cm-3. Note that eachplot has a different
contourinterval. The NO/NO•, NO•/NO• and HNO3/NO• ratiosare dimensionless.
The 03
losstime-scaleat midnight is in units of days. The additional 03 losscausedby includingthe
heterogeneousreactions on amorphouscarbon aerosolover the 7 day simulation period is shown
as a percentage.
:::::::::::::::::::::
LARY ET AL.- CARBON
AEROSOLS
AND ATMOSPHERIC
PHOTOCHEMISTRY
AdditionalOzone loss (ppbv)
AdditionalOzone loss(%)
due to reactions on carbon
due to reactions on carbon
0
I
2
3
4
5
6
7
8
9
0
10
I
3
2
4
5
6
7
8
9
10
I • I I I i I I I i.. I ! .......
!.......
!.....
25.0:.:......<
•
•
:':..-.:-•-•
:'ß
3679
'"•:-':.....-'::'::-:•:•:::--i:i:
':!:•
:::
•:i:
•:i:
•::
:i:i:
::i:
•.i
:i:
i:i:i:i:i:
i:i:
i:i:i:i:
i:i-:-i:i
:::i:::i:i:i:
i...........
i:i:•:•:•:•:i:•:i:.
_
........
•:'"' ';'"'
½'"':':•"
':-:/'/';i:-i::iiiii:::•::i:-i::i!iiiii::i::i::i::i?:::111::11ii?•:'•
'::•?:ii:::•::i::?:::::!i!i:•::ii•i•:•ii::ii?:?:i;
231.40...........
4.32 ?:::::::::::::?:?:(::::::
:::::::::::::::::::::::::
3.42 ?:??:::?:?:?:?:?:?::
;:k:'""'½":::::::::•
.......
Zt•.,•O
. .===============
...........
.-•.._::.............................
•.-•••:...•••11:.!;....i:i.
•.5o
2.87':•'
'":•:"---------•---:•i
"'"":':':"'"'::"
2.39•':':"::':'
'""•
.08 .....................
.?
ix• '••••.
:•':•'•"••'
.....
'' ' '•' •"•.
E
.........................
•:"••'
'":
•••:••4;-•i
96.50
'"':'"':':"
"':':
"•
'
63.39
''""'"•'x'•"':•••;-!:':","
-
..............
•..•.•••:_•••••••!•,.•
•
:•,,
.::-.:
..:•
.:•
_
94• •:.•.
•__
:[''•""•••••....'•i
""'"'"
'
...............
•••••
-
100 iiiii!
..................
••:=•
......
0
•'"'•'••'"'•'•"•'••:'"'••'•••_'_.•
................. *"•'"""""
_
12.5
I
2
3
Carbon
4
5
'"'"
' ' ' '•"••"••••'""'"•!6
7
Aerosol Surface
8
g
o.oo
125
I:'••
"':'"
::""
100
0
0
I
3
2
4
5
6
7
8
9
10
Carbon Aerosol Surface Area
Area
Figure 5. The altitudeand areadependence
of 03 lossby the amorphous
carbonaerosol
mechanism.
Negative
valuescorrespond
to anozoneproduction.
(left)Additional03 losscaused
byincluding
theheterogeneous
reactions
onamorphous
carbonaerosol
(inpercent)
overthe7-day
simulation
period.(Right)Additional03 losscaused
by including
the heterogeneous
reactions
on amorphous
carbonaerosol(in ppbv)overthe 7-daysimulationperiod.
to ozone loss, whereas in the troposphere the enhanced
levels of NO• lead to ozone production.
As there is much more carbon
aerosol in the northern
hemisphere than there is in the southern hemisphere,
the heterogeneousreduction of HNO3, NO2, and O3
on amorphous carbon aerosol will give rise to different trends in the northern hemisphere and southern
hemisphere. In the troposphere the renoxification that
occurs on carbon
aerosols would tend to increase the
ozone concentration. If the carbon aerosolspresent are
entrained within droplets, then the behavior may be
quite different; this also needs to be examined.
Owingto the shapeof .theozoneand NO• profilesthe
largest additional Oa loss in terms of absolute magnitude is largest at higher altitudes. Therefore, if amorphous carbon aerosol can be transported from the aircraft flight corridorsup to 2• km, there will be a correspondingeffect on the O3 lossrate. Carbon aerosolsare
much smaller and lighter than sulphate aerosols,and if,
assuggested
by Blakeand Karo [1995],mostof the soot
is not within sulphate aerosols,there is the possibility
that the carbon aerosolresidencetime is long enough to
allow them to reach the midstratosphere. This may be,
in part, due to their fractal surface area, which means
that they sediment much slower than a sphere of the
same mass. If this is the case, they may be able to influence stratospheric chemistry, even up to 30 km. There
may be some indication of this in the comparison of
ATMOS data and simulationspresentedin Figure 2.
Figure 5 shows that increasing the carbon surface
area from just 0 /•m2 cm-a to 1 /•m2 cm-3 has a
marked effect on stratospheric ozone loss and tropospheric ozone production. Observations show that carbon aerosols are definitely present in the midlatitude
lower stratosphere between 12 and 20 km. In relative
terms, it can be seenfrom Figure 5 that the additional
O3 loss causedby including the heterogeneousreduction of HNO3 and NO2 on amorphouscarbon aerosolis
significantin this region.Solomonet al. [1996]showed
that when a two-dimensional
model is constrained with
time-varyingaerosolobservations
the shapeof the observed trends in ozone is reproduced, but their magni-
tude is about 50% larger than that which is observed.
The presenceof soot in the midlatitude lower stratospheremay help to explainpart of this discrepancybetween ozone observations
and simulations.
SincegasphaseHNO3 destructionis soslow,any processthat is a sink of HNOa is potentially significant.
Midnight
HNOJNO•
0
1
250
2
3
4
5
6
7
8
9
10
I
ß
•....•..•:...•..•.•••••:.
........................................
...................
•....':...•
..................................
•'::••••••
•'"'''
'"'"'•'"'•'' ' ' x
'
22.5
• 2o.o
.......
E •'"'""
'••
"-"
'"'•"-'"""•'•'""'•'•"-•••,•:.-':•
'•'%'' •'.•-,..-.,'.
•'•
•'"'
'"'"
'""
'"''
,.•:...,.,,¾-"'"'"'""'""'
•,•
-.--.•,•...-•.
'""'•'"'""-•----------•••..
-•
15.oo
i:K:!:i:i:i:i:!:!:!:.-':
10.00,>..:-•"•¾•
5.00
.,..,
:•,,:•.:.-:,.-.'""•"
'"":••"
'"":•••'
-"•-"..
'""'"
"'-"•.....•,
...........
'''' ß ..................
..'-%•.-.:::
'•
<r: 1•.o
'
'
2.00
1.00
0.75
0.50
-•••••••••__•',..'-•
......................................
-ß
0.30
.•..•..............•...
,,......._..,
ß
o.
"-""'"'"'"""'•""'""•'"'•"••:'•••••i:•
10.0 I
0
1
2
3
4
5
6
Aerosol Surface
7
8
9
10
Area
Figure 6. The simulatedmidlatitudeHNO3/NO•. ratio as a function of altitude
and carbon surface area.
3680
LARY
ET AL.'
CARBON
AEROSOLS
AND
ATMOSPHERIC
PHOTOCHEMISTRY
AdditionalOzone loss(ppbv)
AdditionalOzone loss (%)
due to reactions on carbon
due to reactions on carbon
0
I
2
3
4
5
6
7
8
9
10
0
I , I.....
!....
.!.
.......
!.....
J.....
,a•:!.
•:•.•>:•:.:.:.,
....
I , ...!
.........
,.......
I............
...
..::...:.
:::•
..........
I..........
I
•12.4 ...........................
........
.............................
....................
.............
'••••:•..
'"'"""'""'""•'•'
'•'•.'-•:::•&:•?a•:
12.8
I
2
3
4
5
6
7
8
9
10
13I::!i!ii?:::i;iiiii?:
I....................
:..•...•:.•••••••••
i•{ -.,.":
.........
'-'i•!•!i•i!i•ii!i!!!i!!•!
.........
:::::::::::::::::::::::::::::::::::::::::
• "•:.¾•::•,•..::i?:!::•::!i•::•::•::!:.i::ii:
"•....'...,
"-':<'-'-----'::-....
.'.....•.-.•.:.::.:.:
::::::::::::::::::::::::
• '-. .... :::::::::::::::::::::::::
ß
'•-."k• ii::i::iii::i::iii:::,•
:::::::::::::::::::::::::::::
$4.$
'"'"'"':'::-•:'
•, 11.a-:
..............
•::....:..,..•::
o•:•....•:.:...:
• •:•::•,.:
•.•:•,.:
....
•.•,..............
. ......:.:•:.:.:...::.•,•
........................ •, ii;:..i::?:ii?:i::',•-a.-'•,
.:-::jii!
•:::.•.::...:..-•:.•:.-!•,
:•:••'•-,,'.•
•'
'0
i!i::i::i!ii•
•-'•'••'••:•
'••'"
...........
'•""'
•' .:::::::::::.'
,o ":-•..:..:.':.'.:•:.:-?
::::::::::::::::::::::::::
':•'%•:'"'i•
:::!
.....................
'::•':•':':':'•':
' ..........
• ::: 1.$•
iiii:ii•!i!i-•
:iiiii: '0 11
.........
•:::::•........................
...:.:•r...<.:.:_,
.
....................................
• ß'•,•t
•.•
1õ.0
"-'-:'-':-•
-...'-.-'•.
ß
:.:.:.:.:.:.::::..:
-..............,.....•:._..••.::.>•:::•.•'•.•.,..:.•.•.,..>.•.
.......
. •:•,_:.:__...._.::.:.:.:._:.............:.:<..:•:<.-.:.:.:.:.:.:.:.:-:-.-.-.:.:.:.:.:.:.:.:.:-:.:-:-:.:.:.:.:
0
I
2
3
4
5
6
7
8
9
10
Carbon Aerosol Surface Area
0
2
3
4
5
6
7
8
9
10
Carbon Aerosol Surface Area
Figure 7. The seasonal
dependence
of 03 lossby the amorphous
carbonaerosolmechanism.
(Left)Additional03 losscaused
byincluding
theheterogeneous
reduction
ofNO2onamorphous
carbonaerosol(in percent)overa 7-daysimulation.(Right)Additional03 losscausedby includingthe heterogeneous
reduction
of NO2on amorphous
carbonaerosol
(in ppbv)overa 7-day
simulation.
Figure 6 showsthat increasingthe carbonsurfacearea
air traffic, then a significantrenoxification mechanism
fromjust 0 to 1/•m• cm-3 reduces
the HNO3/NOwra-
exists. This mechanism
leads to ozone loss in the strato-
tio by a factor of about 8. Chatfield[1995]statesthat sphere and ozone production in the troposphere. The
the averagefree troposphericHNO3/NOw ratio is be- stratospheric ozone loss mechanismis almost indepentween 1 and 9. As can be seen from Figure 6, this dent of temperature and does not require the presence
is entirely consistentwith these simulations. It is also of sunlight. The mechanismcan operate at all latitudes
clear that increasingthe carbonsurfacearea will sub- where amorphouscarbon aerosolsare present. The relstantially increaseNO• and the oxidizingcapacity of ative importance of the mechanism increasesslightly
with nightlength. Including the heterogeneousreducthe troposphere(Figure5).
tion of HNO• and NOa in model simulations predicts
Seasonal Dependence
a free troposphericand stratosphericHNO3[NOx ratio
To investigatethe seasonaldependence
of the mechanism,the third set of sensitivityexperimentsallowed
the time of year to vary betweenthe summersolstice
that is in good agreementwith observations.
Further laboratory studies are urgently required to
preciselyquantify the rate of reduction and investigate
other possibleheterogeneousreactionson atmospheric
amorphouscarbon aerosols.For example, chlorine and
bromine speciessuch as HOC1 and HOBr may also be
and the winter solstice. This corresponds to a day
lengthof between9.5 and 13 hoursfor an air parcel
at 45øN in the lower stratosphere.Figure 7 showsthat
for relatively high carbon surfaceareasthe additional
03 losscausedby including the heterogeneous
reduction of NO2on amorphouscarbon aerosolincreasesby
movingfrom the summersolsticeto the wintersolstice,
whereas at lower carbon surface areas there is little vari-
ationwith day length. Sothe relativeimportanceof the
mechanismtendsto increasewith the lengthof the night
for high carbonaerosolloadings.At higherlatitudesthe
mechanismmay lead to less ozone loss at some times,
as there will be more NO2 present to deactivate reactive
concentrations.
chlorine,this will, in turn, lead to higher C1ONO2
reduced on carbon aerosols. Further field measurements
are requiredto preciselyquantify the amount of amorphous carbon aerosolspresentin the atmosphere. An
assessmentof the impact of carbon entrained within
water or sulphuricacid droplets is also required.
Acknowledgments.
David Lary is a Royal SocietyUniversity Research Fellow and wishes to thank the Royal Society for its support wishes. He also thanks J.A. Pyle for
his support and Robert MacKenzie, and Dudley Shallcross
for very useful conversations. The Centre for Atmospheric
Science is a joint initiative of the Department of Chemistry
and the Department of Apphed Mathematics and Theoretical Physics. This work forms part of the NERC UK Universities Global Atmospheric Modelling Programme.
Summary
If HNO3 and NO2 are heterogeneouslyreduced on References
atmosphericamorphouscarbon aerosolsproducedby, Baumgardner, D., J. E. Dye, B. Gandrud, K. Barr, K. Kelly,
and K. R. Chan. Refractive, indexes of aerosols in the
for example, the combustionof fuel from commercial
LARY
ET AL.:
CARBON
AEROSOLS
AND
upper troposphereand lowerstratosphere,Geophys.Res.
Left., 23 (7), 749-752,1996.
Barrie, L. A., et al., The influenceof mid-latitudinal pol-
ATMOSPHERIC
PHOTOCHEMISTRY
3681
Arctic ozone depletion observedby U ARS MLS during the
1991/92 winter, Geophys.Res.Left., 23(1), 85-88, 1996.
McElroy, M. B., R. J. Salawitch and K. Minschwaner, The
lution sources on haze in the Canadian Arctic, A trnos.
Environ., 15, 1407-1419, 1981.
changingstratosphere,Planet. SpaceSci., J0(2-3), 373-
Berry,M. V., Fallingfractalflakes,PhysicaD, 38(1-3), 29-
Newchurch, M. J., et al., Stratospheric NO and NO2 abundances from ATMOS solar-occultation measurements,
31. 1989.
Blake, D. F., and K. Kato, Latitudinal distribution of black
carbon soot in the upper troposphere and lower strato-
sphere,J. Geophys.Res.,100 (D4), 7195-7202,1995.
Cess, R. D., Arctic aerosols- model estimates of interactive
influences upon the surface atmosphere clear-sky radiation budget, Atmos. Environ., 17, 2555-2564, 1983.
Chatfield, R. B., AnomalousHNOs/NOx ratio of remote
tropospheric air- Conversion of I-INOs to formic-acid and
NOx, Geophys.
Res.Left., 21 (24), 2705-2708,1995.
Chipperfield, M.P., M. L. Santee, L. Froidevaux, G. L. Manney, W. G. Read, J. W. Waters, A. E. Roche, and J. M.
Russell, Analysis of UARS data in the southern polar vortex in september 1992 using a chemical-transport model,
J. Geophys.Res.,101(D13),8861-18881,1996.
Colbeck, I., and S. Nyeki, Optical and dynamical investigations of fractal clusters, Sci. Prog. Oxford, 76, 149-166,
1992.
Considine, D. B., A. R. Douglass, and R. S. Stolarski, Heterogeneousconversionof N205 to HNOs on background
stratospheric aerosols- comparisonsof model results with
data, Geophys.
Res.Lett., 19 (4), 397-400,1992.
DeMore, W. B., et al., Chemical kinetics and photochemical
data for use in stratospheric modeling, Evaluation Number 10, Jet Propul. Lab., Pasedena, Calif., Publication
94-26, 1994.
Fendel, W., D. Matter, H. Burtscher, and A. Schmidtott,
Interaction between carbon or iron aerosol-particles and
ozone, Atrnos. Environ., 29, 967-973, 1995.
Fendel, W., A. S. Ott, J. Aerosol Sci., 2•, S317-S318, 1993.
Fisher, M., and D. J. Lary, Lagrangian four dimensional
variational data assimilation of chemical species, Q. J. R.
Meteorol.Soc.,121(527)Part A, 1681-1704,1995.
Heintzenberg, J., Size-segregatedmeasurements of particulate elemental carbon and aerosol light-absorption at
remote Arctic locations,Atmos. Environ., 16(10), 24612469, 1982.
Hansen, A.D. A.,and H. Rosen, Vertical distributions of
particulate carbon, sulfur, and bromine in the arctic haze
and comparison with ground-level measurements at Bar-
row, Alaska,Geophys.Res. Left., 11(5), 381-384,1984.
Hermann, O., et al., Correlative balloon measurementsof
the vertical distribution of N20, NO, NO2, NO•, HNOs,
N205, C1ONO2 and total reactive NOv inside the polar
vortex during SESAME, Proceedingsof the Third European Symposiumon StratosphericOzone, Schliersee, Germany, September 18-22, 1995, in press 1996.
K/•rcher, B., T. Peter, U.M. Biermann, and U. Schumann,
The initial compositionof jet condensationtrails, J. Atrnos. Sci., in press, 1996.
Lary, D. J., Gas phaseatmosphericbrominephotochemistry,
J. Geophys.Res., 101(D1), 1505-1516,1996.
Lary, D. J., M.P. Chipperfield, and R. Toumi, The potential
impact of the reaction OH+C10-+HCl+O2 on polar ozone
photochemistry,
J. Atrnos.Chern.,21(1), 61-79, 1995.
Lary, D. J., M.P. Chipperfield, R. Toumi, and T. M. Lenton,
Atmospheric heterogeneousbromine chemistry, J. Geo-
phys.Res., 101(D1), 1489-1504,1996.
Logan, J. A., Trendsin the vertical distribution of ozone: An
analysisof ozonesonde
data, J. Geophys.Res., 99(D12),
22553-25585, 1994.
Manney, G.L., L. Froidevaux, J. W. Waters, M. L. Santee,
W. G. Reed, D. A. Flower, R. F. Jarnot, and R. W. Zurek.
401, 1992.
Geophys.Res. Left., 23(17), 2373-2376,1996.
Porch, W. M., and M. C. MacCracken, Parametric study of
the effects of arctic soot on solar-radiation, Atmos. Environ., 16, 1365-1371, 1982.
Pusechel, R. F., D. F. Blake, K. G. Snetsinger, A.D.
A.
Hansen, S. Verma, and K. Kato, Black carbon (soot)
aerosol in the lower stratosphere and upper troposphere,
Geophys.Res. Left., 19(16), 1659-1662,1992.
Rahn, K. A., and R. McCaffrey, On the origin of the winter
arctic aerosol, Ann. N.Y. Acad. Sci., 338, 486-503, 1980.
Rogaski, C.A., D.M. Golden, and L.R. Williams, The heterogeneousreaction of HNOs on model soot compounds,
Abstr. Pap. Am. Chem.Soc.,211(1), 141-ANYL, 1996.
Rosen, H, and A.D.
A. Hansen, Role of combustiongenerated carbon particles in the absorption of solar-
radiation in the Arctic haze, Geophys.Res. Left., ••(5),
461-464, 1984.
Rosen, H., T. Novakov, and B. A. Bodhaine, Soot in the
Arctic, Atmos. Environ., 15, 1371-1374, 1981.
Rosen, H., and T. Novakov, Combustion-generatedcarbon
particlesin the Arctic atmosphere,Nature, 306 (5945),
768-770, 1983.
Schumann, U., Impact of emissionsfrom aircraft and spacecraft upon the atmosphere - An introduction, Paper presented at International Scientific Colloquium, Cologne,
Germany, April 18-20, 1994.
Shaw, G. E., and K. Stamnes, Arctic haze: Perturbation of
the polar radiation budget, Ann. N.Y. Acad. Sci., 338,
533-539, 1980.
Smith, D.M., W.F. Welch, J.A. Jassim, A.R. Chughtai,
and D.H. Stedman, Soot-ozone reaction kinetics - Spec-
troscopicand gravimetricstudies,Appl. Spectros.,,/2(8),
1473-1482, 1988.
Solomon,S., R. W. Portmann, R. R. Garcia, L. W. Thomason, L. R. Poole, and M.P. McCormick, The role of
aerosol variations in anthropogenic ozone depletion at
northernmid-latitudes,J. Geophys.
Res.,101(D3), 67136727, 1996.
Stephens,S., M.J. Rossi, and D.M. Golden, The heterogeneous reaction of ozone on carbonaceoussurfaces, Int. J.
Chem.Kinetics, 18(10), 1133-1149,1986.
Stonehouse,B., Arctic Air Pollution- Studies in Polar Research,Cambridge Univ. Press, 1986.
Tabor, K., L. Gutzwiller, and M. J. Rossi, The heterogeneous interaction of NO2 with amorphous-carbon, Geo-
phys.Res.Left., 20(14), 1431-1434,1993.
Tabor, K., L. Gutzwiller, and M. J. Rossi, The heterogeneous interaction of NO2 with amorphous-carbon at am-
bient temperature, J. Phys. Chem., 98(24), 6172-6186,
1994.
Thlibi, J., and J. C. Petit, A study of the NOv/Soot interaction in the temperature range 303-1223 K, Paper presented at International Scientific Colloquium, Cologne,
Germany, April 18-20, 1994.
Toumi, R.• S. Bekki and R. Cox, A model study of ATMOS
observations and the heterogeneousloss of N20• by the
sulfateaerosollayer, J. Atmos. Uhern.,16(2), 135-144,
1993.
World Meteorological Organisation, Scientific assessmentof
stratospheric ozone: 1991, Rep. 25, WMO Global Ozone
Res. and Monit. Proj., Geneva, 1992.
World Meteorological Organisation, Scientific assessmentof
3682
LARY ET AL.: CARBON
AEROSOLS
AND ATMOSPHERIC
stratosphericozone: 1994, Rep. 37, WMO Global Ozone
Res. and Monit. Proj., Geneva, 1994.
D. J. Lary and A.M.
Lee, Centre For Atmospheric
Science, Department of Chemistry, Cambridge University,
PHOTOCHEMISTRY
M. Pirre and J. B. Renard, Laboratoire de Phsyique et
Chimie de l'Environement. CNRS, Orleans, France.
R. Toumi, Department of Physics, Imperial College, London, SW7 2BZ, England.
LensfieldRoad, Cambridge,CB2 1EW, England. (e-mail:
[email protected])
M. J. Newchurch, Earth System ScienceLab., University
of Alabama in Huntsville, Huntsville, AL 35899, U.S.A.
(ReceivedNovember22, 1995; revisedAugust26, 1996;
acceptedAugust26, 1996.)