JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 102, NO. D17, PAGES 21,515-21,526, SEPTEMBER 20, 1997
Catalytic destruction of stratospheric ozone
D. J. Lary
Centre for Atmospheric Science, Cambridge University, Cambridge, England
Abstract. This paper presents a review of the main ozone destroying catalytic
cycles operating in the stratosphere. Particular attention is paid to the kinetic
aspects of these cycles such as the rate limiting step and chain length. Although it
is an important kinetic parameter, the chain length of the various cycles is seldom
considered when the various catalytic cycles are discussed. This survey highlights
that in the low stratosphere the cycles involving HO: and halogens (notably
bromine) are particularly important. In approximate order of effectiveness the
most important ozone loss cycles in the polar lower stratosphere are the BrO/C10,
HO2/BrO, and OH/HO2 cycles.
Introduction
The importance of atmospheric catalytic cycles was
first recognised by Bates and Nicolet [1950]. Since then,
it has become well established that the concentration
of stratospheric ozone is controlled by the balance be-
tween its production, and its destruction, and that the
destruction of ozone is mainly due to catalytic cycles
involving nitrogen, hydrogen, chlorine, and bromine
species. A comprehensive introduction to these cycles
is presented by, for example, Johnston and Podolske
[1978], Brasseur and Solomon [1986], Wayne [1991] and
the reports of the World Meteorological Organization
(WMO) [1986, 1990, 1992, 1994]. One of several use-
ful historical overviews has been presented by Schmidt
[1988].
The recent WMO assessments [WMO, 1992, 1994]
reported that for the first time there are statistically
significant decreases in ozone in all seasons in both the
northern and southern hemispheres at mid and high lat-
itudes during the 1980s, and that most of this decrease
is occurring in the lower stratosphere. This has also
been supported by trends derived from ozone sondes
[Logan, 1994].
The effectiveness of catalytic cycles in destroying
ozone is controlled by two factors, the chain length
of the catalytic cycles and the abundance of the rad-
ical which is the chain center. The chain length is the
number of times the catalytic cyclic is executed before
the reactive radical involved, the chain center, is de-
stroyed. To date, the chain length of catalytic ozone de-
struction cycles has received relatively little attention,
with emphasis being placed almost exclusively on the
abundance of the chain centers involved. It is therefore
valuable to systematically consider the effectiveness of
the ozone destruction cycles in the stratosphere. This
Copyright 1997 by the American Geophysical Union.
Paper number 97JD00912.
0148-0227/97/97JD-00912509.00
study examines the chain length and the rate of cat-
alytic cycles for the conditions typically encountered in
the stratosphere.
Chain Length and Effectiveness
The chain length of catalytic cycles is limited by ter-
mination steps which destroy the chain center (radical)
involved in the cycle. The chain length A/'is usually
defined as the rate of propagation (p) divided by the
rate of termination. The rate of propagation is the rate
of the rate limiting step. The rate of termination is the
rate of production or destruction (6) of the chain center.
A/' =-P (1)
If a particular radical is involved in a catalytic cycle
which has a very long chain length but is present in
only small concentrations, the effectiveness of the cycle
will be limited. It is therefore useful for us to define a
chain effectiveness R. This is particularly useful when
comparing different cycles involving radicals present in
very different amounts, for example, when comparing
HOx, ClOx, BrOx, and NOx catalytic cycles which de-
stroy 03.
= p (2)
Care needs to be taken when 5 becomes very small as
then very long chain lengths are calculated, which in
turn can lead to a large chain effectiveness even if the
rate of propagation p is tiny.
The following sections consider the chain length and
chain effectiveness of the various atmospheric catalytic
cycles as a function of altitude and latitude. The numer-
ical model used was the AUTOcHEM model described
by Lary et al. [1995, 1996], Lary [1996], and Fisher
and Lary [1995]. The version of the model used in
this study contains a total of 81 species. Of these,
74 species are integrated, namely: O(1D), O(3p), 03,
N, NO, NO2, NO3, N2Os, HONO, HNO3, HO2NO2,
21,515

21,516 LARY: CATALYTIC DESTRUCTION OF STRATOSPHERIC OZONE
CN, NCO, HCN, C1, C12, C10, C1OO, OC10, C1202,
C1NO2, C1ONO2, HC1, HOC1, CHaOC1, Br, Br2, BrO,
BrONO2, BrONO, HBr, HOBr, MeOBr, BrC1, H2,
H, OH, HO2, H202, CHa, CHaO, CHaO2, CHaOOH,
CHaONO2, CHaO2NO2, HCO, HCHO, CFa, CFaO,
CFaO2, CFaOOC1, CFaOH, CFaOOH, CFaOONO2, F,
F2, FO, FO2, F202, COF2, FCO, FCOO, FCOOH,
FC(O)O2, FNO, FONO, FO2NO2, HF, CH4, CHF3,
CHaBr, CF2C12, N20, and CO. The remaining seven
species are not integrated and not in photochemical
equilibrium, namely: CO2, H20, 02, N2, HCl(s), H20(s),
and HNO3(s). The model contains a total of 438 re-
actions, 287 bimolecular reactions, 43 trimolecular re-
actions, 65 photolysis processes, and 43 heterogeneous
reactions.
Hydrogen Catalytic Cycles
One of the early landmarks in the atmospheric chem-
istry of odd hydrogen radicals occurred in 1964 when
J. Hampson highlighted the fact that O(1D) produced
by O3 photolysis at wavelengths <310 nm led to the for-
mation of hydroxyl radicals due to the reaction of 0(1D)
with water vapor. This led to the socalled \"wet the-
ory\" of stratospheric ozone. In 1965, E. Hesstvedt and
P.L. Roney published the first papers which included
the effects of odd hydrogen radicals on the photochem-
istry of ozone. In 1969, P.J. Crutzen published a fur-
ther paper which considered the effect of odd hydrogen
radicals on ozone loss.
Figure 1 schematically shows the main HOx ozone
loss catalytic cycles. Catalytic cycles involving HO2 are
very important in the lower stratosphere. The fastest
of these cycles is
OH/HO2, HOx cycle 1
OH + O3 > HO2 q- 202 (3)
HO2 q- 03 > OH + 02 (4)
Net ß 203 > 202
This cycle has a rate of between 104 and 10 s molecules
cm -3 s -1 between 10 and 35 km and a chain length of
between 1 and 40 in the midlatitude lower stratosphere
(Figure 2). At midlatitudes it has a peak chain effec-
tiveness of 106 molecules cm -3 s -1 at around 20 km
(Figure 2). However, the cycle is also effective at de-
stroying ozone in the winter polar lower stratosphere
(Figure 3). At approximately 15 km and above the rate
limiting step is the reaction of O3 with H02 (i.e., the
rate of propagation is given by p- k4103][HO2]). Be-
low approximately 15 km the rate limiting step is the
reaction of OH with O3 (p- k3103][OH]). However,
both reactions (3) and (4) are fast.
h ClO
0 02
Figure 1. A schematic of the main HOx ozone loss
catalytic cycles.
The HO2 cycle involving C10 is also important.
HO2/C10, HOx cycle 2
C10 + HO2 > HOC1 + 02
HOCl+hv > Ci+OH
C1 q- 03 > C10 q- 02
OH+O3 > HO2+202
Net: 203 > 302
(5)
(6)
(7)
At midlatitudes this cycle has a peak rate of 10 4
molecules cm -3 s -1 and a chain length of between 1
and 25 in the lower stratosphere (Figure 2). It has a
peak chain effectiveness of 106 molecules cm -3 s -1 at
around 25 km (Figure 2). The cycle is very effective at
destroying ozone in the midlatitude lower stratosphere
and also plays a role in the sunlit polar lower strato-
sphere (Figure 3). The cycle is not effective in the dark
as throughout the stratosphere the rate limiting step is
the photolysis of HOC1 (p = jHOC] [HOC1]). The cy-
cle forms an important link between chlorine and odd-
hydrogen chemistry.
HO2/BrO, HOx cycle 3
BrO + HO2 > HOBr + 02
HOBr+hv > Br+OH
Br+03 > BrO+02
OH q- 03 > HO2 +202
Net :203 > 302
(8)
(9)
(10)
This cycle has a peak rate of 10 4 molecules cm -3
S --1 and a chain length of between 1 and over 1000 in
the stratosphere (Figure 2). It has a peak chain effec-
tiveness of almost 107 molecules cm -3 s - at around
20 km (Figures 2 and 3). Throughout he stratosphere
the rate limiting step is the photolysis of HOBr (p =

LARY: CATALYTIC DESTRUCTION OF STRATOSPHERIC OZONE 21,517
0.1
60
0.1
OH/HO2 (Cycle I)
OH/HO2 (Cyle V)
H/HO2
Local noon, mid-latitude at Equinox
1 10 100 1,000 ,-  o o,,, ,,,,,,,,, ,,, 
,, ,,,,,,I ,, ,,,,,,I ,, ,,,,,,I ,, ,,,,,, ,,,,, ,,,,,, ,,,,J ,,,,, ,,,,,,I ,,, ,,,,,a ,,,,, ,,,,, ,,,,,,
/
s
s
s
1 10 100 1,000
Chain Length
BrO/H02
CIO/H02
Chain Effectiveness
60
Rate of propogation
Local noon, mid-latitude at Equinox
0 0 0 0
0.1 1 10 100 1,000 ,-   o ,,, w ,,, ,,, o  o ,,, ,,,
60 ,,,, ,, ,1,,,I, 60 60
_
- \
40 40 -- , - 40  40
,
I II t 20  20 20  20
'\"' \"'\"'\"/ I '
0.1 1 10 100 1,000  o o o 7 ?  T  o o o  -  
 0      0  
Chain Length Chain Effectiveness Rate of propogation
Figure 2. The calculated chain effectiveness, rate, and chain length of various HOx ozone loss
catalytic cycles for local noon at midlatitudes at equinox.

21,518 LARY: CATALYTIC DESTRUCTION OF STRATOSPHERIC OZONE
OH/He2 (Cycle I)
Chain Effectiveness
t. atitude (degrees)
OH/He2 (Cyde V)
Chain Effectiveness
ß
;.t, .(XX),::i-;.,:.::.:..::::::.::..:::: ....;.;:1 e),.. ............................... - -..: 
atude (alegin.es)
.::.:.::, ;-;.. ....
HO2/CIO
Chain Effectiveness
HO2/BrO
Chain Effectiveness
Contour Interval ts 400.000
t$
'0.0 .0 .40 ' 0 20 40 0 eO TM
Laatude (degrees)
'. I!li::.:: .
o 2.4oo. ooo 4.eoo.ooo 7.oo. ooo ,.oo. ooo
H/He2
Chain Effectiveness
 35
'1o
Latitude (degrees)
! -'\".'. :. :::'\"'\" '\" '
4.E:,006 +(30 'L-E009 t009
BrO/ClO
Chain Effectiveness
Latitude (degrees)
.
.... i::;::  'i 
o ., .. 2,o..
Figure 3. The calculated variation in the chain effectiveness of various cycles for the winter
solstice as a function of altitude and latitude.

LARY: CATALYTIC DESTRUCTION OF STRATOSPHERIC OZONE 21,519
jHOBr [HOBr]). The cycle forms an important link be-
tween bromine and hydrogen chemistry and is particu-
larly effective at destroying ozone in the sunlit part of
the winter polar lower stratosphere (Figure 3).
H/OH/HO2, HOx cycle 4
M
H q- 02 > HO2 (11)
HO2 + O(aP) , OH+ O2 (12)
OH+O(3P) , H+O2 (13)
Net ' 20(3P) > 02
This is one of the most effective ozone loss cycles
in the upper stratosphere (Figures 2 and 3). This cy-
cle has a peak rate of l0 s molecules cm -s s - and
a chain length of between 1 and a few hundred in
the stratosphere. It has a peak chain effectiveness
of l09 molecules cm -3 s - at between 50 and 70 km
(Figure 2). Below approximately 35 km the rate lim-
iting step is the reaction of O(3p) with OH, reac-
tion (13) (p - k3[O(øP)J[OHJ). All three reactions
involved in this cycle are fast in the upper stratosphere.
Above approximately 35 km the rate limiting step is
either the reaction of O(3p) with HO2, reaction (12)
(p -- k210(3p)][H02]), or the reaction of H with 02
(p- k[O2][H]), reaction (11).
There is another OH/H02 cycle which is effective in
the mid to upper stratosphere but which is not as fast
HOx Cycle 1.
OH/HO2, HOx Cycle 5
OH + 03 > HO2 q- 202
H02 + O(3p) > OH q- 02
Net- O(3p) + 03 > 202
This cycle has a peak rate of approximately 106
molecules cm -3 s - between 30 and 40 km where it has
a chain length of between 1 and 12 (Figure 2). It has a
peak chain effectiveness of approximately 107 molecules
cm -3 s - at between 35 and 40 km (Figure 2). At
35 km and above the rate limiting step is the reaction
of 03 with OH (p - k3103][OH]); below this it is the
reaction of H02 with O(3p)(p = k210(3p)][H02]).
For all of the reactive hydrogen catalytic cycles just
considered the rate of destruction of the chain center
has been taken as the rate of formation of H2 and H2 O;
namely,  = k[H][H02] + k[OH] 2 + k[OH][H02]. Let
us now turn our attention to some chlorine cycles.
Chlorine Catalytic Cycles
The early landmark in the atmospheric chemistry of
reactive chlorine radicals occurred when Molina and
Rowland [1974] published their famous paper which
showed that the stratosphere was the only sink for chlo-
rofluoromethanes, and that ozone destruction would re-
suit as a consequence of the chlorine released. This was
closely followed by the work of $tolarski and Cicerone
[1974] and Rowland and Molina [1975]. Then Farman
et al. [1985] discovered the ozone hole and suggested
that it was due to the interaction of atmospheric chlo-
rine and nitrogen.
Figure 4 schematically shows the main C1Ox ozone
loss catalytic cycles.
C1/C10, C1Ox cycle 1
ClO + O(P) ,
C1 q- 03 
Net' O(3p) q- 03 )
C1 + 02
C10 q- 02
202
(14)
Between approximately 40 and 50 km this is one of
the most effective ozone loss cycles (Figures 5 and 6).
At midlatitudes it has a peak rate of 10  molecules cm -3
s - and a chain length of greater than 10 throughout
most of the stratosphere and of approximately 103 in
2. l_ ........ z_-2_--_L___ /'l-D:  .3 \ T4- ,.-,,
peak chain effectiveness ofapproximately 108 molecules
cm -a s - at between 35 and 45 km (Figure 5). The
rate limiting step is the reaction of O(aP) with C10
(p- k410(3p)][C10]).
C10/C10, C1Ox cycle 2
C10 q- C10 M > C1202
C1202 q- h > C1 q- C100
C100 M> C1 q- 02
2 x (el q- 03 , ClO q- 02)
Net ß 203 > 302
(15)
(16)
(17)
In the midlatitude lower stratosphere between ap-
proximately 20 and 25 km this cycle has a peak rate
of 104 molecules cm -3 s - (Figure 5). However, when
Figure 4. A schematic of the main C1Ox ozone loss
catalytic cycles.

21,520 LARY: CATALYTIC DESTRUCTION OF STRATOSPHERIC OZONE
CI/CIO
CIO/CIO
CIO/N02
0.1 1 10 100 1,000
_
_
_
40-- -
_
_ ß
- \
_ \
20-- , -
0.1 1 10 100 1,000
Chain Length
Br/BrO
BrO/CIO
BrO/NO2
Local noon, mid-latitude at Equinox
0 0 0 0
v- 0 0 0
v- _ _
o o +  + T
60
Chain Effectiveness Rate of propogation
Local noon, mid-latitude at Equinox
v- o o o
._o o + + +T  T o o
'- 0 W W '- 0 LLI W IJJ LLI IJJ IJJ v- 0 LLI W
Chain Length Chain Effectiveness Rate of propogation
Figure 5. The calculated chain effectiveness, rate, and chain length of various chlorine and
bromine ozone loss catalytic cycles for local noon at midlatitudes at equinox.

LARY: CATALYTIC DESTRUCTION OF STRATOSPHERIC OZONE 21,521
Br/BrO
Chain Effectiveness
Lalitude (degrees)
'Ti   :. : .............
CIO/N02
Chain Effectiveness
,.,.._..z,- - ..,.. -.,,..,. =.....-
:: ,.. :-  ../\" }
...... . .  ?-'::.% ............ . ......... : ,- ................  .................  ......  , ............... , ..........  
La'tude
BrO/N02
Chain Effectiveness
01tOlO
Chain Effectiveness
e .-.-- .......
el/NO2
Chain Effectiveness
.:
 ::.
'. ......................... , .......... .... ..........  ................................  .............. :..
Latitude (degrees)
ClOtCtO
Chain Effectiveness
...,,.........., .........  . ... .,,...... .............  . .. ,..,,,..,.. ........... ,,,,..... ....  ....... 
  ( ................. - .,
'\"'   m ß \"\",':' ',\" ' r:.:: \" ....:....:.:.... .
Figure 6. The calculated variation in the chain effectiveness of various cycles for the winter
solstice as a function of altitude and latitude.

21,522 LARY: CATALYTIC DESTRUCTION OF STRATOSPHERIC OZONE
cold aerosols and polar stratospheric louds (PSCs) are
absent, it is barely catalytic in this region (Figures 5
and 6). At midlatitudes it has a peak chain effectiveness
of approximately 10 s molecules cm -s s - at between
20 and 25 km (Figure 5) increasing to 2x104 molecules
cm -s s - in the polar lower stratosphere. In the strato-
sphere the rate limiting step is the photolysis of
(P -- jCI.O. [CleOn]).
C1/NO., C1Ox cycle 3
C10+NO > Cl+NO2
NO2+O(sP) > N0+02
Cl+Os > C10+02
Net'O(sP) + Os , 202
(18)
(xo)
This is the cycle that Farman et al. [1985] sug-
gested was involved with the formation of the ozone
hole. Although it is an effective ozone loss cycle over
a large altitude range, namely, between 15 and 50 km
(Figure 10), it is most effective at midlattitudes and
in the summer midstratosphere. It is not as effec-
tive in the spring polar lower stratosphere (Figure 6),
whereas the next cycle, the C10/NO. cycle is. It has
a peak rate of 106 molecules cm -s s - between 30
and 40 km. The cycle has a chain length of greater
than I throughout the stratosphere reaching approxi-
mately 1000 at around 30 km. At midlatitudes it has a
peak chain effectiveness ofapproximately 109 molecules
cm -s s - at between 30 and 40 km (Figure 10), but
reaches 10 ø molecules cm -s s - at high latitudes in
summer (Figure 6). At approximately 25 km and above
the rate limiting step is the reaction,, of C10 with NO
(p - ks[C10][NO]). Below this it is the reaction of
O(SP) with NO2 (p- kxg[O(SP)][N02]).
C10/NO2, C1Ox cycle 4
ClO + NO2 M (20) > ClON02
ClON02 + hv > C1 + NOs (21)
hv -- I-IO
 CI
Oa
Figure 7. A schematic of the main BrOx ozone loss
catalytic cycles.
BrO
h
Figure 8. A schematic of the main NOx ozone loss
catalytic cycles.
NOs+hv > N0+02
NO+Os > N02+02
Cl+Os > C10+02
Net :20s > 302
(22)
Between 15 and 40 km the C10/NO2 cycle is an ef-
fective ozone loss cycle (Figure 5). It has a peak rate
of 104 molecules cm -s s -1 between 20 and 25 km and
a chain length of between I and 10 between 15 and
35 km (Figures 5 and 6). It has a peak chain ef-
fectiveness of approximately 105 molecules cm -s s -1
at 20 km (Figure 5). Between approximately 45 and
10 km the rate limiting step is the photolysis of NOs
(p - jNO3 [NOs]). For most of the region above 45 km
the rate limiting step is the reaction of C10 with NO2
(p- k20[C10][NO2]).
For all of the reactive chlorine catalytic cycles just
considered the rate of destruction of the chain center
has been taken as the rate of formation of HC1; namely,
5- k[el][H2] + k[el][eH4] + k[el][H20] + k[C1][H02]
+ k[C][HCHO] + k[C][CHaOOH].
Let us now turn our attention to some bromine cat-
alytic cycles.
Bromine Catalytic Cycles
The atmospheric chemistry of bromine and its syn-
ergy with chlorine was first studied in detail by Wofsy et
al. [1975] and Yung et al. [1980]. Lary [1996] and Lary
et al. [1996] present a review of the atmospheric gas
phase and heterogeneous chemistry of bromine. Fig-
ure 7 schematically shows the main BrOx ozone loss
catalytic cycles.
Br/BrO, BrOx cycle 1
BrO + O(sP) > Br + 02 (23)

LARY: CATALYTIC DESTRUCTION OF STRATOSPHERIC OZONE 21,523
Br+Os , BrO+O2 (24)
Net ß O(sP) + Os , 202
This cycle has a peak rate of 10 4 molecules cm -s s -
at between 35 and 45 km and a chain length of greater
than 50 throughout the stratosphere (Figures 5 and 6),
with a chain length of greater than 10 4 at 40 kin. The
cycle has a peak chain effectiveness of approximately
108 molecules cm -s s - at between 35 and 45 km (Fig-
ure 5). At all altitudes the rate limiting step is the re-
action of O(sP) with BrO (p - k2s[O(sP)][BrO]). How-
ever, both reactions (23) and (10) are fast.
BrO/C10, BrOx cycle 2
BrO+C10 > Br+ClOO
ClO0 M> C1 q- 02
Cl+Os , C10+O2
Br+Os ) BrO+O2
Net :2Os ) 302.
(25)
Of all the bromine cycles this is the most important
for high-latitude, lower-stratosphere ozone loss with a
chain effectiveness which reaches 108 molecules cm -s
s - in the polar lower stratosphere (Figure 3).
The reaction of BrO with C10 has several channels.
The most effective ozone loss channel is the one yield-
ing C100. This cycle has a peak rate of almost 10 4
molecules cm -s s - at between 20 and 25 km and a
chain length of greater than i between 15 and 45 km
and is approximately 10 s in the lower stratosphere (Fig-
ure 5). Between 20 and 25 km the chain length ap-
proaches 2000. At midlatitudes the cycle has a peak
chain effectiveness of approximately 107 molecules cm -s
s - at between 20 and 25 km (Figure 5). At all altitudes
the rate limiting step is the reaction of BrO with C10
(p- k25[BrO][C10]).
(26)
BrO/N02, BrOx cycle 3
BrO + NO2 M> BrON02
BrONO2+hv > Br+NOs
NO3 q- hv > NO + O2
N0+03 , N02+02
Br+Os > BrO+02
Net: 20s > 302
This cycle will only be effective if BrONO2 photolyses
to give NOs [Lary et al., 1996]. If it does, as has been
assumed here, then this cycle has a peak rate of almost
104 molecules cm -s s - at 20 km and a chain length of
greater than 10 between 10 and 40 km (Figure 5). The
cycle has a peak chain length of greater than 1000 be-
tween 15 and 25 km (Figure 5). Between 15 and 35 km
the rate limiting step is the photolysis of BrONO2, and
below this it is the photolysis of NOs (p = jNOa [NOs]).
At 35 km and above, the rate limiting step is the reac-
tion of BrO with NO2 (p = k26[BrO][NO2]).
For all of the reactive bromine catalytic cycles just
considered the rate of destruction of the chain center
has been taken as the rate of formation of HBr; namely,
5- k[Br][H2] q- k[Br][HCHO] q- k[C1][H02].
Let us now turn our attention to the most important
nitrogen catalytic cycle.
Nitrogen Catalytic Cycle
In 1967 Bates and Hays published a paper on the
concentrations, sources and sinks for N20 in the atmo-
sphere. Then in 1970 and 1971 Crutzen and Johnston
introduced the consideration of the NOx radical into
the photochemical theory of stratospheric ozone. In
1971 Nicolet published a paper of the NO-O(D)-NO -
NO2 mechanism and Johnston pointed to the danger
of NOx emissions from stratospheric aircraft for strato-
spheric ozone. Figure 8 schematically shows the main
N Ox ozone loss catalytic cycles.
NO/NO2, NOx cycle 1
NO+Os } N02+02
NO2+O(aP) , N0+02
Net ' O(3p) + O3 > 202
(28)
This cycle has a peak rate of 106 molecules cm -3 s -
between 30 and 40 km and a chain length of greater
than 1 throughout the stratosphere reaching 105 in
the upper stratosphere (Figure 10). The cycle has a
chain effectiveness of greater than 104 molecules cm -3
s - throughout he stratosphere reaching a peak of
greater than 109 molecules cm -3 s - at 45 km (Fig-
ures 9 and 10). At 50 km and above, the rate limiting
step is the reaction of NO with 03 (p- k28[NO][03]).
NOIN02
Chain Effectiveness
Latitude (degrees)
- ::i .\"i .......
0 26+009 4E+009 6E+009 8E+009
Figure 9. The calculated variation in the chain effec-
tiveness of the NO/NO2 cycle for the winter solstice as
a function of altitude and latitude.

21,524 LARY: CATALYTIC DESTRUCTION OF STRATOSPHERIC OZONE
NO/NO2
CI/N02
O(3P)/O3
60
Local noon, mid-latitude at Equinox
0 0 0 0
-' 0 0 0 0 0 LU W W LU
0 - - - - - ',- ',- ',- ',- -
- 0 0
-- 0 LU LU - 0 LLI LLI W
60
40
2O
Chain Length
CF30/CF3OO
- )- F/FO
- F/FO/FO2
Chain Effectiveness Rate of propogation
Local noon, mid-latitude at Equinox
0.01
6O
40--
20--
0.1 1 100
I
0.1 1 10
...,
/ '\"'\"\"1 '\"'\"\"1 '\"'\"\"
0.1 0.01 0.1 1 1 10 100
1 10 100 1,000
_ _
_ _
_ _
_%
_ , _
-- I -
_
1 10 100 1,000
60
20
Chain Length Chain Effectiveness Rate of propogation
Figure 10. The calculated chain effectiveness, rate, and chain length of various nitrogen and
fluorine ozone loss catalytic cycles for local noon at midlatitudes at equinox.
Below 50 km the rate limiting step is the reaction of
NO2 with O(3p)(p = k2910(3P)][NO]).
For the reactive nitrogen catalytic cycles the rate of
destruction of the chain center has been taken as the
rate of formation of HNO3; namely, 5 - k[OH][NO2] +
k[H20][N2Os] + k[HO][C1ONO]. The last two terms
are the heterogeneous hydrolysis of N205 and C1ONO.
Finally, let us consider some fluorine catalytic cycles.
Fluorine Catalytic Cycles
In 1975 Zander, Roland and Delboville found an ab-
sorption line of HF in a solar spectrum which was due to
stratospheric absorption of HF. In 1978 Sze supported
their findings with model calculations.
There are two sets of catalytic ozone loss cycle in-
volving fluorine: those involving the CF3 group which

LARY: CATALYTIC DESTRUCTION OF STRATOSPHERIC OZONE 21,525
are terminated by the production of COF2 and those
involving F, FO, and FO2 which are terminated by the
formation of HF. This section examines the chain length
of the fastest of these cycles.
The fastest of all the fluorine catalytic cycles and the
only fluorine cycle considered to have a chain length of
greater than 1 is
CF30/CF3OO, CF3Ox cycle 1
03 -]- CF3OO > CF30 + 202
O3+CF30 > CF3OO+O2
Net :203 > 302
(3o)
(31)
Even so, under normal conditions in the midlatitude
lower stratosphere, this cycle is barely catalytic, having
a peak chain length of 2 at 20 km; that is the cycle is
only executed once before the CF3 group is destroyed
by the formation of COF2 (i.e., 5 - k[NO][CF30]). In
these calculations it has been assumed that as much
fluorine is in the atmosphere as chlorine. Even with
this assumption, the cycle proceeds at a maximum rate
of only 103 molecules cm -3 s - (Figure 10).
As examined by D. J. Lary et al. (Atmospheric fluo-
rine photochemistry, submitted to the Journal of Geo-
physical Research, 1996), the rate of termination by the
formation of COF2 due to the reaction of CF30 with
NO slows down as the temperature is increased as it has
a negative activation energy. In contrast, the reaction
of 03 with CF30 has a large positive activation energy
and is therefore faster at warmer temperatures. In ad-
dition, for high sulphate aerosol loadings less, NO will
be present in the atmosphere further slowing down the
rate of formation of COF2. Consequently, this can lead
to a longer chain length of approximately 40 for tem-
peratures around 240 K and sulphate aerosol loadings
of greater than 5 pm2cm -3.
The fastest F/FO/FO2 catalytic cycles are
F/FO, FOx cycle 1
cycles proceed at a rate of approximately 70 molecules
cm -3 s -1 (Figure 10). For both of these cycles the
rate limiting step is the reaction of FO with 03 (p =
k[FO][O3]), so both cycles have the same chain length.
Summary
Table 1 tabulates the approximate altitude at which
the rate of each cycle reaches a maximum. In line with
earlier work, such as that of Johnston and Podolske
[19781, Brasscur and Solomon [1986], Wayne et al.
[1991] and the reports of the WMO [1986, 1990, 1992,
1994], a few obvious points can be made in conclusion.
1. In the upper stratosphere, where the abundance
of 03 is relatively low, the most effective catalytic
cycles are those whose net reaction is 20(3P) 
02.
2. In the mid stratosphere the most effective cat-
alytic cycles are those whose net reaction is O(3p)
+ 03 - 202.
3. In the lower stratosphere, where the abundance of
O(3p) is relatively low, the most effective catalytic
cycles are those whose net reaction is 203 - 302.
4. Examining the rate limiting step alone does not
give any idea as to how many times a cycle can
proceed. Consequently, calculating the cycle's
chain length is valuable. This is particularly true
when comparing cycles which involve chain cen-
Table 1. Summary Table Showing the Approximate
Altitude in Kilometres at Which the Various Cat-
alytic Cycles are Fastest for Midlatitudes At Equinox.
For this Altitude the Approximate Chain Length,
is Also Given
Z
(km) Pmax J CYCLE
O3+F > FO+O. (32)
O3+FO  F+202 (33)
Net :203  302
F/FO/FO2, FOx cycle 2
03 - F > FO + 02
03 -k FO  FO2 + 02
FO2 M F + 02 (34)
Net: 203  302
Even so, in the calculations performed here, both of
these cycles are never catalytic, having a peak chain
length of 0.4 between approximately 30 and 40 km; that
is the cycle cannot proceed more than once as HF is
formed so rapidly by the reaction of F with CH4, H20,
and H2 (6 = k[F][H2] + k[F][CH4] + k[F][H20]). The
65 8x106 103 H/OH/HO2
40 2x106 103 C1/C10
lx106 106 0/03
2x104 104 Br/BrO
35 4x106 l0 s NO/NO2
3x106 103 C1/NO2
8x10 s 10 OH/HO2
30 8x104 80 OH/HO2
3x104 10 HO2/C10
25 9x104 10 C10/NO2
22 lx104 103 BrO/C10
6x103 103 BrO/NO2
2x103 2 C10/C10
20 5x103 103 HO2/BrO
10 .1 CF30/CF3OO
lxl02 .4 F/FO
1x102 .4 F/FO/FO2
HOx cycle 4
C1Ox cycle 1
BrOx cycle 1
NOx cycle 1
C1Ox cycle 3
HOx cycle 5
HOx cycle 1
HOx cycle 2
C1Ox cycle 4
BrOx cycle 2
BrOx cycle 3
C1Ox cycle 2
HOx cycle 3
CF3Ox cycle 1
FOx cycle 1
FOx cycle 2

21,$26 LARY: CATALYTIC DESTRUCTION OF STRATOSPHERIC OZONE
.
.
ters of very different abundance, such as the chlo-
rine and bromine cycles. In general, although
bromine is less abundant than chlorine, it is still
very effective in destroying ozone as the bromine
cycles tend to have very long chain lengths. In
this context, the chain effectiveness is useful as it
is the product of the cycle's rate and chain length.
A cycle with a chain effectiveness greater than
l0 s molecules cm -3 s -1 is likely to be an effective
ozone loss cycle.
In the low stratosphere the cycles involving HO2
are important [e.g. Brasseur and Solomon, 1986].
What has only relatively recently become appar-
ent is that the cycles involving HO2 and halogens
are particularly important. With increasing levels
of atmospheric bromine the HO2/BrO will assume
a greater role in lower stratospheric and upper
tropspheric ozone loss.
The most important ozone loss cycles in the po-
lar lower stratosphere in their approximate order
of importance are the BrO/C10, HO/BrO, and
OH/HO cycles, with chain effectivenesses reach-
ing 2x10 s, 9x106, and 4x106 molecules cm -3 s -1,
respectively, with the C10 dimer cycle having a
much lower chain effectiveness of around 3x104
molecules cm -s s -1.
7. The currently known fluorine cycles are barely
catalytic and experience rapid termination. They
are therefore not effective at destroying ozone.
Acknowledgments. David Lary is a Royal Society Uni-
versity Research Fellow and wishes to thank the Royal Soci-
ety for its support. He also thanks J.A. Pyle for his support,
and Robert MacKenzie, and Dudley Shallcross for very use-
ful conversations. The Centre for Atmospheric Science is a
joint initiative of the Department of Chemistry and the De-
partment of Applied Mathematics and Theoretical Physics.
This work forms part of the NERC UK Universities Global
Atmospheric Modelling Programme.
References
Bates, D. R., and M. Nicolet, The photochemistry of atmo-
spheric water vapor, J. Geophys. Res., 55, 301, 1950.
Brasseur, G., and S. Solomon, Aeronoray of the Middle At-
mosphere, 2nd Edition, D. Reidel, Norwell, Mass., 1986.
Farman, J. C., B. G. Gardiner, and J. D., Shanklin,
Large losses of total ozone in antarctica reveal seasonal
C1Ox/NOx interaction, Nature, 315 (6016), 207-210, 1985.
Fisher, M., and D. J. Lary, Lagrangian four dimensional
variational data assimilation of chemical species, Q. J. R..
Meteorol. Soc., Part A, 1œ1, (527), 1681-1704, 1995.
Johnston, H. S., and J. Podolske, Interpretation of strato-
spheric photochemistry, Rev. Geophys., 16, 491, 1978.
Lary, D. J., Gas phase atmospheric bromine photochemistry,
J. Geophys. Res., 101 (D1), 1505-1516, 1996.
Lary, D. J., M.P. Chipperfield, and R. Toumi, The potential
impact of the reaction OH+C10HCI+O2 on polar ozone
photochemistry, J. Atraos. Ghera., œI(1), 61-79, 1995.
Lary, D. J., M.P. Chipperfield, R. Toumi, and T. M. Lenton,
Atmospheric heterogeneous bromine chemistry, J. Geo-
phys. Res., 101(D1), 1489-1504, 1996.
Logan, J. A., Trends in the vertical distribution of ozone: An
analysis of ozonesonde data, J. Geophys. Res., 99, 25553-
25585, 1994.
Molina, M. J., and F. S. Roland, Stratospheric sink for chlo-
rofluoromethanes: Chlorine atom catalysed destruction of
ozone, Nature, 29, 810, 1974.
Roland, F. S., and M. J. Molina, Chlorofluoromethanes in
the environment, Rev. Geophys., 13, 1, 1975.
Schmidt, M., Pioneers of ozone research, Max-Planck Inst.,
Katlenburg-Lindau, Germany, 1988.
Stolarski, R. S., and R. J. Cicerone, Stratospheric chlorine:
A possible sink for ozone, Gan. J. Ghera., 5œ, 1610, 1974.
Wayne, R. P., The chemistry of atmospheres, 2nd Edition,
Oxford Univ. Press, New York, 1991.
Wofsy, S.C., M. B. McElroy, and Y. L. Yung, The chemistry
of atmospheric bromine, Geophys. Res. Left., 2, 215, 1975.
World Meteorological Organisation, Scientific assessment of
stratospheric ozone: 1985, WMO Global Ozone Research
and Monitoring Project, Rep. 16, Geneva, Switzerland,
1986.
World Meteorological Organisation, Scientific assessment of
stratospheric ozone: 1988, WMO Global Ozone Research
and Monitoring Project, Rep. 20, Geneva, Switzerland,
1990.
World Meteorological Organisation, Scientific assessment of
stratospheric ozone: 1991, WMO Global Ozone Research
and Monitoring Project, Rep. 25, Geneva, Switzerland,
1992.
World Meteorological Organization, Scientific assessment of
stratospheric ozone: 1994, WMO Global Ozone Research
and Monitoring Project, Rep. 37, Geneva, Switzerland,
1994.
Yung, Y. L., J.P. Pinto, R. T. Watson, and S. P. Sander, At-
mospheric bromine and ozone perturbations in the lower
stratosphere, J. Atmos. Sci., 37(2), 339, 1980.
D. J. Lary, Centre for Atmospheric Science, Department
of Chemistry, Cambridge University, Lensfield Road, Cam-
bridge, CB2 1EW, U.K. (emaihdavid@atm.cm.cam.ac.uk)
(Received June 20, 1996; revised March 18, 1997;
accepted March 24, 1997.)