GEOPHYSICAL RESEARCH LETTERS, VOL. 21, NO. 13, PAGES 1427-1430, JUNE 22, 1994
Chlorine chemistry and the potential for ozone depletion in the
arctic stratosphere in the winter of 1991/92
R. Mtiller,  Th. Peter,  P. J. Crutzen,  H. Oelhaf, 2G. P. Adrian 2 Th. v. Clarmann 2
A. Wegner, 2 U. Schmidt, 3 and D. Lary 4
Abstract. We present an analysis of chlorine chemistry in the duce the principal features the disturbed chlorine chemistry
Arctic stratosphere during the winter of 1991/92 and assess
its potential implications for ozone depletion. In accordance
with observations of total organic chlorine, C1ONO2 and
HC1, box model results indicate the following: (1) An al-
most complete activation of chlorine during the cold winter
period. (2) A possible contribution from the heterogeneous
reaction HOC1 + HC1 and the gas-phase reaction CHaO2 +
C10 to the complete conversion of HC1 to active chlorine.
(3) A strong buildup of C1ONO2 following PSC disappear-
in accordance with the observations, we further investigated
the potential consequences of the simulated high levels of
active chlorine for ozone loss.
Model description
We use a chemical box model [Crutzen et al., 1992] includ-
ing a complete set of gas-phase and heterogeneous reactions,
the latter both on liquid sulfuric acid and nitric acid trihy-
ance which remains the main chlorine reservoir for about a drate (NAT) surfaces (Table 1). In the model, NAT particles
month, after which HC1 becomes dominant. (4) Appreciable form at 3 K supercooling, as indicated by observations in the
chemical ozone loss in the lower stratosphere inside the polar Arctic [Schlager et al., 1990; Dye et al., 1992]. For the corn-
vortex is conceivable for the winter of 1991/92. putation of the photolysis rates, a scheme was adopted that
accurately describes photolysis at large solar zenith angles
Introduction [Lary and Pyle, 1991] and includes the temperature depen-
One objective of the European Arctic Stratospheric Ozone
Experiment (EASOE) was to provide a wide spatial and tem-
poral coverage of measurements of chemical species in the
polar stratosphere. Here, we use several of these measure-
ments in an attempt to provide a synoptic view of the chem- H 1
ical evolution ofchlorine and nitrogen species inthe Arctic H2
lower stratosphere during the winter of 1991/92. H3
Ground-based observations of total column C1ONO2 over H4
the entire winter [Adrian et al., 1994], along with balloon- H5
borne observations of C1ONO2 inside the vortex on January
13, 1992 and March 14, 1992 [v. Clarmann et al., 1993;
Oelhaf et al., 1994], indicate a sizable conversion of this R1
species to active chlorine during the coldest period from R2
the end of December to the end of January, followed by an R3
almost complete transformation of the available chlorine to R4
C1ONO2 during the beginning March. Moreover, ground- R5
based observations of very low column densities of HC1 R6
during the winter and spring of 1992 [Adrian et al., 1994; R7
Bell et al., 1994] suggest a substantial conversion of HC1 to R8
active chlorine. R9
In order to analyze the mechanisms responsible for the ex- R10
ceptional, unprecedented observations ofchlorine species we R11
conducted chemical box model calculations for air parcels R12
representative of the conditions in the Arctic vortex in R13
1991/92. Encouraged by the ability of our model to repro- R14
 MPI for Chemistry, Atmospheric Chemistry Dep. Mainz, Germany
2 Inst. Meteorel. Climate Res., KFK, Karlsruhe, Germany
3 KFA Jfilich, Jiilich, Germany
4 Dep. of Chemistry, University of Cambridge, Cambridge, UK
Copyright 1994 by the American Geophysical Union.
Paper number 94GL00465
0094-8534/94/94 GL-00465 $03.00
Table 1. Reactions important to this work
Reaction
N2 05 + H20 -- 2HNO3
C1ONO2 + H20 - HOC1 + HNO3
C1ONO2 + HC1 - C12 + HNO3
HOC1 + HC1 -+ C12 + H20
N205 + HC1 - C1NO2 + HNOa
0.0006
0.006
0.3
0.1
0.003
Rate const.
C1 + CH4 - HC1 + CH3 1.0(- 14)
C10 + NO2 + M -- C1ONO2 + M 1.6(-12)
C10 + He2 - HOC1 + O2 1.6(-11)
C1 + O3 - C10+ O2 7.9(-12)
C10 + NO -- C1 + NO2 2.7(-11)
ClO + 0 -- C1 + 02 4.3(-11)
HNO3 + h -- NO2 + OH 4.7(-8)
HNO3 + OH --> NO3 + H20 5.4(-13)
C1ONO2 + hp - C1 + NOa 1.8(-5)
C10+ CHaO2 - C1 +prod. 1.0(-12)
C10+C10+M - C1202 +M 2.7(-13)
C1202+M - C10+C10+M 3.9(-5)
C1202 + h - C1OO + C1 7.1(-4)
C10 + BrO -- C1 + Br + 02 8.7(-12)
For the heterogeneous reactions H1-H5, reaction prob-
abilities 7 for NAT [DeMore et al., 1992] are listed. On
sulfuric acid (with weight percentage W) only H 1 (7 = 0.1)
and H2 (7 = 101'86-ø'ø747'W) were included [Hanson and
Ravishankara, 1991 ]. For the gas-phase reactions R 1-R 14,
rate constants (incm a s- 1 molec- 1; except R12 in s- l) [De-
More et al., 1992] for 100 mbar and 200 K and photolysis
rates (in s-l) for 85 o zenith angle are given.
1427

1428 M()LLER ET AL.' CHLORINE CHEMISTRY AND OZONE DEPLETION
1.2
. ' NAT surface
,, _
_
area
._
0.8
0.6
0.4
0.2
0
210
205
200
195
190
185
15 3ec I Jan 15Jan IFeb 15Feb IMar 15 Mar
Day
Figure 1. The adopted temperature history and the NAT
surface area (in m2/cm a) over the model period December
15, 1991 to March 15, 1992.
dence of the UV-cross sections for HNOa [Burkholder et al.,
1993]. We use one quarter of the recommended upper limit
[DeMore et al., 1992] as reaction rate for R10. A rate of
this order has recently been substantiated by measurements
which further disclosed CHaOC1 as a product of R10 [Helleis
et al., 1993].
The temperature history (Fig. 1), of the air parcels inside
the polar vortex in the lower stratosphere during the winter
of 1991/92 was estimated from observations [Naujokat and
Labitzke, 1993; Waters et al., 1993]. Further, to simulate the
diabatic descent in the vortex indicated by observations of
inert tracers [Schmidt et al., 1993], they are assumed to sink
from 550K to 450K (22-17km) over the model period. Some
simulations have also been performed for different height
levels to estimate total column amounts. The air parcels are
either constrained to constant latitude (58øN-78øN), or si-
nusoidal variations in latitude (680 4-10øN) were prescribed.
In accordance with meteorological observations, the cooling
episodes and the southward excursions are assumed to take
place over northern Europe.
There are some limitations in using a box model in this,
highly idealized way, and, consequently, a detailed, quan-
titative comparison with measurements is not attempted.
Nonetheless, since the model assumptions are representative
of the situation inside the polar vortex in the winter 1991/92,
some conclusions about the chemical mechanisms at work
may be drawn on account of the model results.
Model results and observations
Chlorine and nitrogen species. In the model, NAT par-
ticles exist from the end of December to the end of January
(Fig. 1) in accordance with estimates based on a thorough
meteorological analysis [Newman et al., 1993] and lidar mea-
surements [Stein et al., 1993]. The surfaces of these particles
catalyze the heterogeneous reactions H3-H5 that activate
HC1. While H5 is unlikely to be very effective due to prior
removal of N205 by reaction on sulfate aerosol particles via
H 1, reactions H3 and H4 proceed rapidly, effectively titrating
the available C1ONO2 and HOC1 against HC1 on a timescale
of hours (Fig. 2). The further depletion in HC1 is much slower
and is controlled by the supply of reaction partners for HC1.
The most important partner under sunlight conditions and
when PSCs exist is HOC1 produced via R3. Further, when-
ever PSCs evaporate during warm periods (Fig. 1), HNOa
is released into the gas-phase and C1ONO2 recovers via R2
(Fig. 2), thereby using up the NOs produced by R7 and R8.
Thus, NO2 concentrations remain low. Similarly, in PSC-
free periods, HOC1 builds up via reaction R3 (Fig. 3). With
the PSCs reappearing during cold periods, HC1 is removed
via H3 and H4 with C1ONO2 and HOC1 being about equally
important. Since the buildup of HC1 during the warm pe-
riods is much slower than that of HOC1 and C1ONO2 (see
below), the recurrent evaporation of PSCs thus enhances the
depletion of HC1. In the model, HC1 is severely depleted in
the lower stratosphere, a finding supported by the very low
column amounts of HC1 measured atthe end of January. The
C1ONO2 concentrations remain extremely low in the model
throughout the period when PSCs exist, in good accordance
with the balloon-borne observation on January 13.
After the last PSCs disappear in the model simulation, the
active chlorine is almost completely converted to C1ONO2
2.5-
E.0
ClO
' 1.5 ..... : ..... : ..... : ....
1.0
0.5: ' -\" ' ' '\" \"' '
0.0
0 20 40 60
Day
Figure 2. The main chlorine reservoirs over the model
period: Upper panel shows the FTIR observations of
stratospheric HC1 column amounts over Kiruna (diamonds)
and Greenland (squares) and C1ONO2 total column amount
over Kiruna (triangles) (see Adrian et al. [1994] for details);
also column densities of HC1 (thick solid line) and C1ONO2
(thin solid line) derived from the model results are shown (all
in 1015 cm-2). Lower panel shows HC1 and C1ONO2 mix-
ing ratio in the lower stratosphere from the model results (at
68øN), dotted line shows simulated HC1 neglecting reactions
H4 and R10, dashed line the results allowing excursions in
latitude. Diamonds in the lower panel indicate the mixing
ratios of C1ONO2 on March 14, 1992, (about 17 km height)
and on January 13, 1992 (about 18 km height), both observed
by MIPAS-B.

MOLLER ET AL.: CHLORINE CHEMISTRY AND OZONE DEPLETION 1429
0.4
00
1.5 l
 clo ß
 1.0
_o 0.5
0.0 ,,,, , , lllllll
:3.5
 3.0
cf 2.5
2.0
0 20 40 60
Day
Figure 3. The mixing ratios of HOC1, C10 and ozone in the
lower stratosphere (at 68øN) over the model period. Dot-
ted line in lower panel indicates ozone simulated neglecting
reactions H4 and R10, dashed line the results allowing ex-
cursions in latitude.
via R2; the recovery of HC1 via R1 is much slower (Fig. 2).
This pattern is clearly borne out by measurements which
show (in agreement with the model) extremely high column
amounts of C1ONO2, yet only a slow recovery of HC1 to-
wards the end of winter (Fig. 2). It is further corroborated
by the balloon-borne observation on March 14 which shows
very high mixing ratios of C1ONO2 in the lower stratosphere.
The reason for the preferred production of C1ONO2 over
HC1 can be explained by the relative effectiveness of reac-
tions R1 and R2 for the removal of active chlorine:
d[HC1]/dt k [CH4] [C1] k [CHi]ks [NO]
d[ClONO2]/dt k2 [NO2] [C10] k2 [NO2]k4103]
Assuming NO/NO2  1 and mixing ratios of 03 and CH4
in the ppmv range, ct -, 10 -2, and thus practically all the
active chlorine will initially be converted into C1ONO2. The
rate of increase in C1ONO2, and thus the rate of chlorine
deactivation, is controlled by the rate of release of NO2 from
the HNOa via R7 and R8 after the evaporation of NAT.
The activation of chlorine is further reflected in the high
mixing ratios of C10 indicated by the model results (Fig. 3).
The simulated temporal variation of C10 is consistent with
aircraft [Crewell et al., 1994] and satellite observations [Wa-
ters et al., 1993].
Ozone. Previous modeling studies of Arctic ozone de-
pletion [McKenna et al., 1990; Brune et al., 1991; Chipper-
field et al., 1993] have simulated significant ozone loss for
the lower stratosphere for February 1989 and 1990. Here, we
focus on the entire winter of 1991/92. Our model results rep-
resentative of the polar vortex (78øN-68øN, corresponding
to a total of 300-670 hours of sunlight over the model pe-
riod) show an integral ozone destruction of 15%-38% in the
air parcel (Table 2) in accordance with observed ozone loss
rates for the period from mid January to mid February 1992
[Proffitt et al., 1993]. The results of model runs performed
for vortex conditions at lower latitudes (58øN and 63øN, Ta-
ble 2) as a sensitivity study, are consistent with the finding
of Brune et al. [ 1991] that the amount of ozone depletion
strongly increases with decreasing latitude. To demonstrate
the sensitivity of ozone depletion to the duration of PSC
existence, model runs were performed for an unrealistically
short (10 days) and long (50 days) PSC periods by shifting
temperatures up and down by 3 K (Table 2). However, in
1992, possibly also unprocessed air masses existed in the
vortex [Newman et al., 1993], for which the present analysis
is not applicable.
Furthermore, the suggestion by Crutzen et al. [1992], that
a cycle involving R10 and H4, could be of relevance to
Antarctic ozone depletion is extended here to the Arctic (Ta-
ble 2, Figs. 2 and 3). For the Arctic high latitude ozone loss,
this cycle may be of comparable importance as the BrO-C10
cycle (Table 2). Since it is strongly dependent on solar illu-
mination [Crutzen et al., 1992], it is triggered earlier during
southward excursions if variations in latitude are considered,
so that in this case, the HC1 activation, and thus the ozone
depletion, is slightly stronger (Table 2, Figs. 2,3).
In a study for the 1991/92 Arctic vortex [Salawitch et al.,
1993], the model was initialized with more C1ONO2 than
HC1 based on in-situ measurements of HC1 [Webster et al.,
1993]. Under these circumstances, the complete activation
of HC1 proceeds via H3 and requires no further explanation.
Thus, the impact of H4 and R10 would be reduced. Further,
if H2 does not occur in the stratosphere, as suggested by
Webster et al. [ 1993], the inorganic chlorine reservoir would
not be activated completely since any C1ONO2 in excess of
HC1 would remain unprocessed; consequently, less ozone de-
pletion than for the initialization with excess HC1 employed
here is simulated [Salawitch et al., 1993]. In contrast to this
view, the C1ONO profile measured on January 13, 1992 [v.
Clarmann et al., 1993; Oelhaf et al., 1994] indicates that
only very little C1ONO2 is present in the lower stratosphere
during the coldest period of the winter 1992.
Conclusions
Measurements and model results discussed here yield a
consistent picture. Specifically, the activation of chlorine
Table 2. Ozone depletion in percent
lat. T- 3K T T + 3K .,4 /3
58øN 81 60 39 64 45
63øN 70 56 33 49 42
68øN 49 38 23 28 27
73øN 27 23 18 13 15
78øN 17 15 14 7 9
68 ø 4- 10øN 53 42 24 29 30
Ozone depletion simulated in the model from 15.12.91
to 15.3.92 for temperature T (as in Fig. 1) and with the
temperature history shifted upward (T q- 3K) and downward
(T- 3K) by three Kelvin. Further, results are shown from
model runs neglecting ,,4: Reactions H4 and RI0 and/3:
neglecting bromine chemistry.

1430 MOLLER ET AL.: CHLORINE CHEMISTRY AND OZONE DEPLETION
during the cold winter period, reflected in balloon-borne
measurements of little C1ONO2 [v. Clarmann et al., 1993;
Oelhaf et al., 1994], aircraft measurements of high concen-
trations of CIO, low levels of HC1 [Crewell et al., 1994] and
large column amounts of OC10 [Brandtjen et al., 1994] and
ground-based observations of low column amounts of HC1
and C1ONO2 inside the polar vortex [Adrian et al., 1994; Bell
et al., 1994] is matched by the model results. After the final
disappearance of the PSCs, the model shows high column
amounts of C1ONO2 and an almost complete conversion of
active chlorine to C1ONO2 (Fig. 2), which remains the dom-
inant inorganic chlorine compound for more than a month,
all in accordance with the observations.
Our model results show that reactions H4 and R 10, hitherto
omitted in model studies of Arctic ozone loss, may possess
a substantial impact on halogen-catalyzed ozone destruction
(Table 2). The simulations performed for conditions repre-
sentative of the lower stratosphere inside the polar vortex for
the Arctic winter of 1991/92 indicate, in accordance with ob-
servations [Proffitt et al., 1993], that considerable chemical
ozone loss may have taken place.
Acknowledgments. We thank Th. Wawers, H. Gimm and
T.-O. Gunstr0m for computer support during the campaign and
M. Flender for help performing the model runs. Part of this work
was funded by the European Community and the German Ministry
for Research and Technology (BMFT).
References
Adrian, G.P. et al., First results of ground based FTIR measurements
of atmospheric trace gases in north Sweden and Greenland uring
EASOE, this issue.
Bell,W. et al., Groundbased measurements of stratospheric on-
stituents over ]ire, Sweden, this issue.
Brandtjen, R., Th. Kltipfel, D. Pemer, and B. Knudsen, Airbome
measurements during EASOE: Observations of OC10, this issue.
Brune, W.H. et al., The potential for ozone depletion in the Arctic
polar stratosphere, Science, 252, 1260-1266, 1991.
Burkholder, J.B., R.K. Talukdar, A.R. Ravishankara, and S.
Solomon, Temperature dependence of the HNO3 UV absorp-
tion cross sections, J. Geophys. Res., in press 1993.
Chipperfield, M.P. et al., A three-dimensional modeling study of
trace species in the arctic lower stratosphere during the winter
198% 1990, J. Geophys. Res., 98, 7199-7218, 1993.
v. Clarmann, T. et al., Retrieval of stratospheric 03, HNO3 and
C1ONO2 profiles from 1992 MIPAS-B limb emission spectra, J.
Geophys. Res., 98, 20495-20506, 1993.
Crewell, S., K. Ktinzi, H. Nett, and P. Hartogh, Aircraft measure-
ments of C10 and HC1 during EASOE 1991/92, this issue.
Crutzen, P.J., R. Mtiller, Ch. Brtihl, and Th. Peter, On the potential
importance of the gas phase reaction CH302 q- C10 -- C1OO q-
CH30 and the heterogeneous reaction HOClq-HC1  H2Oq-C12
in \"ozone hole\" chemistry, Geophys. Res. Lett., 19, 1113-1116,
1992.
DeMore, W.B. et al., Chemical kinetics and photochemical data for
use in stratospheric modeling, JPL Publ. 92-20, Pasadena, 185
pp., 1992.
Dye, J., D. Baumgardner, B.W. Ganrud, and R.G. Knollenberg,
Particle size distributions in Arctic polar stratospheric louds, J.
Geophys. Res., 97, 8015-8034, 1992.
Hanson, D.R., and A.R. Ravishankara, The reaction probabilities
of C1ONO2 and N2Os on 40 to 75% sulfuric acid solutions, J.
Geophys. Res., 96, 17307-17314, 1991.
Helleis, F., J.N. Crowley, and G.K. Moortgat, Temperature-
dependent rate constants and product branching ratios for the
gas-phase reaction between CH302 and C10, J. Phys. Chem.,
97, 11464-11473, 1993.
Lary, D.J., and J.A. Pyle, Diffuse radiation, twilight, and photo-
chemistry, J. Atmos. Chem., 13, 373-406, 1991.
McKenna, D.S. et al., Calculations of ozone destruction during the
1988/89 Arctic Winter, Geophys. Res. Lett., 17, 553-556, 1990.
Naujokat, B., and K. Labitzke, Collection of the reports on the
stratospheric irculation during the winters 1974/75-1991/92,
STEP, 301 pp., 1993.
Newman, P. et al., Stratospheric meteorological conditions in the
arctic polar vortex, 1991 to 1992, Science, 261, 1143-1146,1993.
Oelhaf, H. et al., Stratospheric C1ONO2, O3 and HNO3 profiles
inside the Arctic vortex from MIPAS-B limb emission spectra
obtained during EASOE, this issue.
Proffitt, M.H. et al., Ozone loss inside the northern polar vortex
during the 1991-1992 winter. Science, 261, 1150-1154, 1993.
Salawitch, R.J. et al., Chemical loss of ozone in the arctic polar
vortex in the winter 1991/92, Science, 261, 1146-1154, 1993.
Schlager, H., F. Amold, D.J. Hofmann, and T Deshler, Bal-
loon observations of nitric acid aerosol formation in the arctic
stratosphere, Geophys. Res. Lett., 17, 1275-1278, 1990.
Schmidt,U. et al., The variation of available chlorine ClOy in the
Arctic polar vortex during EASOE, this issue.
Stein, B. et al., Stratospheric aerosol sizedistributions from multi-
spectral idar measurements at Sodank15 during EASOE, Geo-
phys. Res. Lett., 20, this issue.
Waters, J.W. et al., Stratospheric C10 and ozone from the microwave
limb sounder on the upper atmosphere research satellite, Nature,
362, 597-602, 1993.
Webster, C.R. et al., Hydrochloric acid loss and chlorine chem-
istry on polar stratospheric loud particles in the Arctic winter,
Science, 261, 1130-1136, 1993.
P. J. Crutzen, R. Mailer, and Th. Peter, Max Planck Institute for
Chemistry, Dep. Atmospheric Chemistry, Postf. 3060, 55020
Mainz, Germany (e-mail: muller@nike.mpch-mainz.mpg.de)
Th. v. Clarmann, H. Oelhaf, G. P. Stiller (formerly Adrian), and
A. Wegner, Institute for Meteorology and Climate Research, KFK,
Postf. 3640, 76021 Karlsruhe, Germany
U. Schmidt, KFA, ICG-3, Postf. 1913, 52425 Jtilich, Germany
D. Lary, Dep. Chemistry, University of Cambridge, Lensfield
Rd, Cambridge CB2 1EW, UK
(received November 17, 1992; revised January 24, 1994;
accepted February 9, 1994.)