University of Rhode Island University of Rhode Island
DigitalCommons@URI DigitalCommons@URI
Graduate School of Oceanography Faculty
Publications
Graduate School of Oceanography
1990
The geochemical cycling of reactive chlorine through the marine The geochemical cycling of reactive chlorine through the marine
troposphere troposphere
William C. Keene
Alexander A. P. Pszenny
Daniel J. Jacob
Robert A. Duce
University of Rhode Island
James N. Galloway
See next page for additional authors
Follow this and additional works at: https://digitalcommons.uri.edu/gsofacpubs
Terms of Use
All rights reserved under copyright.
Citation/Publisher Attribution Citation/Publisher Attribution
Keene, W. C., A. A. P. Pszenny, D. J. Jacob, R. A. Duce, J. N. Galloway, J. J. Schultz‐Tokos, H. Sievering, and
J. F. Boatman (1990), The geochemical cycling of reactive chlorine through the marine troposphere,
Global Biogeochem. Cycles
, 4(4), 407–430, doi: 10.1029/GB004i004p00407.
Available at: http://dx.doi.org/10.1029/GB004i004p00407
This Article is brought to you for free and open access by the Graduate School of Oceanography at
DigitalCommons@URI. It has been accepted for inclusion in Graduate School of Oceanography Faculty
Publications by an authorized administrator of DigitalCommons@URI. For more information, please contact
digitalcommons@etal.uri.edu.
GLOBAL BIOGEOCHEMICAL CYCLES, VOL. 4, NO. 4, PAGES 407--430, DECEMBER 1990
THE GEOCHEMICAL CYCLING OF REACTIVE
CHLORINE THROUGH THE MARINE TROPOSPHERE
William C. Keene,1
2 Daniel J. Jacob, 3
Alexander A. P. Pszenny,
Robert A. Duce, 4 James N. Galloway, 1
Joseph J. Schultz-Tokos, 4'5
Herman Sievering, 6'7 and Joe F. Boatman 6
Abstract. Heterogeneous reactions
involving sea-salt aerosol in the marine
troposphere are the major global source
for volatile inorganic chlorine. We meas-
ured reactant and product species hypothe-
sized to be associated with these chemical
transformations as a function of phase,
particle size, and altitude over the North
Atlantic Ocean during the summer of 1988.
Concentrations of HC1 were typically less
than 1.0 ppbv near the sea surface and
--iDepartment of Environmental Sciences,
University of Virginia, Charlottesville.
2Ocean Chemistry Division, NOAA Atlantic
Oceanographic and I/eteorologica1 Labora-
tory, I/iami, Florida.
3Division of Applied Sciences and
Department of Earth and Planetary Sciences,
Harvard University, Cambridge,
I/assachusetts.
4Graduate School of Oceanography, Uni-
versity of Rhode Island, Narragansett.
5Now at Institut fuer I/eereskunde an der
Un}versitaet Kiel, Kiel, Germany.
VAeroso1 Research Section, NOAA Air
Resources Laboratory, Boulder, Colorado.
7Also at Center for Environmental Sci-
ences, University of Colorado at Denver.
Copyright 1990
by the American Geophysical Union.
Paper number 90GB02425.
0886-6236 / 90 / 90GB-02425 •10 . 00
decreased with altitude and with distance
from the U.S. east coast. Concentrations
of C1 volatilized from aerosols were gen-
erally equivalent to the corresponding
concentrations of HC1 and ranged fro le
than detection limits to 125 nmol m -• ss STP.
Highest absolute and percentage losses of
particulate C1 were typically associated
with elevated concentrations of anthropo-
genic combustion p_roducts. Concentrations
of product uss S04z-and NO 3- in coarse
aerosol fractions indicate that on average
only 38% of measured C1- deficits could be
accounted for by the combined effects of
acid-base desorption and reactions involv-
ing nonacidic N gases. We hypothesize a
mechanism for the C1 loss initiated by
reaction of 03 at sea-salt aerosol sur-
faces, generating C12, followed by rapid
photochemical conversion of C12 to HC1 via
C1 atoms (C1') and eventual recapture of
HC1 by the aerosol. Simulations with a
zero-dimension (0-D) photochemical model
suggest that oxidation by C1' may be an
important tropospheric sink for dimethyl
sulfide and hydrocarbons. Under low-NO x
conditions, the rapid cycling of reactive
C1 would provide a catalytic loss mecha-
nism for 03 , which would possibly explain
the low 03 concentrations often observed
above the world's oceans.
1. INTRODUCTION
The injection of sea-salt aerosol gen-
erated by breaking waves on the ocean's
408 l/eene et al.' Geochemical Cycling of Reactive Chlorine
surface is the major global source for
atmospheric C1 [Erickson and Duce, 1988].
Most of this C1 remains in the aerosol and
is returned to the ocean surface via dry
and wet deposition, but important frac-
tions, ranging from 3 to 20% on average,
are released from the aerosol as inorganic
ß
C1 vapor (Cllg) [e.g., Cicerone, 1981].
Although this particle-to-gas conversion
is by far the major source for gaseous C1
in the global troposphere [e.g., Friend,
1989], decades of research have failed to
demonstrate conclusively the primary mech-
anism(s) involved. This is due in part to
the fact that few studies have measured
principal reactant and product species
simultaneously as a function of phase and
particle size in marine regions remote
from continental influences.
HC1 is generally r. ecognized as the
major fraction of C1 l_ in the marine
troposphere [Ryan andgMukherjee, 1974;
Wofsy and McElroy, 1974]. Thermodynamic
considerations coupled with field measure-
ments of aerosol chemistry and Clig sug-
gest that most of the HC1 may originate
with direct volatilization from sea-salt
aerosol which is acidified to low pH (•3)
by the incorporation of HNO 3 and H2SO 4
(Eriksson, [1959], Duce, [1969], Martens
et al. [1973], Berg and Winchester [1977],
Kritz and Rancher [1980], Brimblecombe and
Clegg [1988], Legrand and Delmas [1988],
Wall et al. [1988], among many others) as
follows:
NNO 3 ( g ) +NaC1 (p)-- >itC1 ( g ) +NaNO 3 (p) ( 1 )
H2SO4(p)+2 NaCl(p)-->2 HCl(g)+Na2SO4(p)
(2)
Although exchange. between particulate-
phase C1- and Cll has been demonstrated
g
clearly by extensive field measurementsß
the importance of the acid-base desorption
mechanism as the principal driver for C1
phase change in the marine troposphere
remains open to question [e.g.ß Ciceroneß
1981i Friendß 1989].
Alternative mechanisms involving reac-
tions of various N gases with sea-salt
aerosol have also been suggested as
sources for chemically active halogen
gases in the marine troposphere. Reaction
of NO 2 with sea-salt aerosol has been
hypothesized [A1tshuller, 1958.] and demon-
stoated to be a source for C11 [Schroeder
and Urone, 1974; Finlayson-Pit•s, 1983] as
follows:
2 NO2(g ) + NaCl(p) --} NOCI(g) + NaNO3(p)
(3)
NOC1 has short lifetimes against photo-
lysis (5-30 rain) and hydrolysis (0.03 s to
45 rain). Photolysis of NOC1 genezates C1
atoms (C1') which may then initiate oxida-
tion of hydrocarbons to produce HC1.
Hydrolysis of NOC1 generates HONO- a
source of OH radical (OH') and H• )ß
ß . In
both casesß the final products of these
rapid reaction sequences are HC1 and
NaNO 3, the same products expected from an
acid-base desorption involving HNO 3 (reac-
tion (1)). Finlayson-Pitts et al. [1989]
recently reported that CiNO 3 and N205
react wi. th NaC1 aerosol to generate reac-
tive Cll as follows'
ClNO3(g ) + NaCl(p) --} C12(g ) + NaNO3(p)
(4)
N205(g ) + NaCl-(p) --} C1NO2(g ) + NaNO3(p)
Rapid photolysis of product C12 and CINO 2
generates C1' and ultimately HC1 through
subsequent reactions. Again, the final
products are HC1 and NaNO 3. Considerable
uncertainties exist in assessing the
potential for a significant influence of
reactions (3), (4)ß and (5) in the remote
marine troposphere, but given the expected
and observed concentrations of reactant N
gases [e.g.ß Levy and Moxim, 1989], it has
been suggested that such transformations
will be an important source for C1 i only
in more polluted regions [e.g.ß Sinlh and
/asting, 1988].
The nature of heterogeneous reactions
which generate C1 i have iuiportant impli-
cations for marinegtropospheric chemistry.
For instanceß transformations involving
NO 2, CINO 3, and N205 generate highly
reactive C1 compounds which can initiate
photochemical reactions in an analogous
manner to 'OH, whereas acid-base desorp-
tion generates relatively unreactive HC1.
The modeling investigation of Singh and
Kasting [1988] suggestsß heavevet, that if
ppbv concentrations of HC1 are generated
by any mechanismsß reaction with 'OH can
produce sufficient CI' to photooxidize a
significant fraction (20 to 40%) of non-
methane alkanes in the marine troposphere.
It is clearly essential that the compounds
involved, mechanisms of emissionß and
rates of reaction be identified unequivo-
cally if we are to understand major pro-
cesses in the chemical cycling of S, N, C,
odd O, odd H, and C1 through this dynamic
system.
The present study was designed to meas-
ure major reactant and product species
Keene et al.: Geochemical Cycling of Reactive Chlorine 409
which are thought to be involve. d in the
heterogeneous generation of C1 •- from
g
reactions involving sea-salt aerosol in
the marine troposphere. Hypothesized
sources are assessed in light of these
data.
2. METNODS
Bulk- and size-segregated samples of
atmospheric aerosol, some with simultane-
ous samples of alkaline reactive C1, N,
and S gases (hereinafter referred to as
HC1, HNO 3, and SO 2, respectively), were
collected from an aircraft and a ship over
the North Atlantic Ocean (NAO) during the
summer of 1988 as part of the Global
Change Expedition, Coordinated Air-Sea
Experiment, and Western Atlantic Oceav
Experiment (GCE/CASE/WATOX) [Pszenny et
al., 1990a]. Major features of sampling
systems which generated data evaluated in
this study are summarized in Table 1. The
University of Virginia (UVA) and the Air
Ouality Group (AQG; now referred to as the
Aerosol Research Section) from the
National Oceanic and Atmospheric Adminis-
tration (NOAA) sampled the western NAO
boundary layer (BL) and free troposphere
(FT) with a high-flow dichotomous filter
pack system mounted on the NOAA King Air
research aircraft [Bardwell et al., 1990].
Air was sampled within 250 km of the U.S.
mid-Atlantic coast and in the vicinity of
Bermuda.
Atmospheric samples were also collected
with a variety of systems throughout the
NAO from a 10-m bow tower on the NOAA ship
Mt. Mitchell. UVA/AQG sampled major
particulate- and vapor-phase species with
a bulk filter pack system similar in
design and operation to the bulk filter
pack component of the dichotomous system
deployed on the aircraft [Bardwell et al.,
1990]. Intercomparison with data from
other measurement systems on the ship
indicates that the open-face inlet on this
shipboard sampler selectively excluded
larger particles resulting in an approxi-
mate 20% underestimate of sea-salt
species.
NOAA's Atlantic Oceanographic and
Meteorological Laboratory (AOML) sampled
BL air from the ship with a filter pack
system, a high-volume aerosol sampler
(hive1), and a six-stage cascade impactor
[Pszenny et al., 1990b]. In addition, the
University of Rhode Island (URI) sampled
atmospheric aerosol with a seven-stage
cascade impactor [Pszenny et al., 1989].
The AOML filter pack was designed to sam-
pie fine aerosol and alkaline reactive N
and S gases preferentially. Intercompari-
son with other data sets for shipboard
collections suggests that the inlet for
this sampler selectively excluded approxi-
mately two thirds of the sea-salt aerosol
mass. Relationships between wind velocity
and sea-salt aerosol concentrations meas-
ured with the AOML hivol were similar to
those observed some years ago in the NAO
with an isokinetic sampler [Lovett, 1978]
suggesting that the hivol collected repre-
sentative samples of sea-salt aerosol
mass.
A number of potential artifacts could
bias data for size-segregated aerosol
generated with cascade impactors in marine
regions. These include internal losses on
slot throats of
significant retention of HC1 by Whatman 41
substrates for contact times longer that
about i ms [Fogg, 1986], and inefficient
retention of submicron aerosol by
Whatman 41 backup filters [Lodge, 1986].
Although internal losses may have intro-
duced modest negative bias for impactor
data reported in this paper, other arti-
facts were probably not important. Calcu-
lated contact times for the cascade
samplers were typically less than i ms,
suggesting that retention of HC1, and by
analogy HNO 3, was probably minimal under
our sampling conditions. In addition, a
large body of information indicates that
Whatman 41 filters collect representative
samples of submicron aerosol
Lowenthal and Rahn, 1987: Watts et al.,
1987; Kitto and Anderson, 1988].
The qualities of most data sets evalu-
ated in this study have been assessed by
thorough in-house testing and by intercom-
parison [e.g., Bardwell et al., 1990;
Boatman et al.
1990b]. There were, however, no independ-
ent measurements of HC1 by other groups to
intercompare with measurements by UVA and
AQG. Given this lack of independent
information, and in light of the impor-
tance of these data for the present study,
we include in the appendix a brief assess-
ment for the quality of HC1 and non-sea-
salt (nss) C1-data generated during the
experiment. Results indicate that
particle-to-particle and gas-to-particle
reactions on bulk aerosol prefilters may
cause large (factor of 2) positive bias in
measurement of HC1 and negative bias in
measurement of particulate nss C1-. As
such, we recommend that published data for
these species which were generated using
bulk prefilters be viewed with caution.