Kinetics of HOBr uptake on NaBr and NaCl surfaces at varying relative humidity

Article abstract of DOI:10.1021/jp013594g

The uptake kinetics of HOBr on NaBr and NaCl particle-film surfaces has been studied in a flow reactor coupled with a differentially pumped quadrupole mass spectrometer. Spray-deposited and aerosol-deposited salt-particle films were used for the measurement of HOBr reactivity. A rapid gas-surface reaction was observed on the NaBr surface at 250 K, and the initial uptake coefficient ?3_w was determined to be 0.029 ?± 0.004 and 0.021 ?± 0.004 at relative humidities (RH) of 0.5 and 12.2%, respectively. ?3_w is lower at 260 K. For the reaction on NaCl particle surfaces, ?3_w decreases slightly from 2.0 ?± 0.3 ?? 10^-3 to 8.1 ?± 1.3 ?? 10^-4 at 250 K as RH increases from 1.5 to 18.2%. The results are interpreted in terms of the mechanism HOBr(ad) + NaX(s) a?? BrX + NaOH(s), which involves the adsorption of HOBr on the surface, with the reactions occurring mainly on a NaX surface domain without water coverage at low RH. The surface morphology of the particle films was characterized by scanning electron microscopy, and the tortuosity factor, ??, of the NaCl particle film was determined to be 2.18.?3_w was corrected for internal surface effects to provide true uptake coefficients ?3_t. Typical ?3_t values for HOBr on NaBr and NaCl at 250 K and RH of 10% are 2.5 ?? 10^-3 and 5 ?? 10^-5, respectively. Calculations suggest that the HOBr heterogeneous activation on NaBr may be competitive, if surface regeneration occurs.

Full text of DOI:10.1021/jp013594g

J. Phys. Chem. B 2002, 106, 5679-5688
5679
Kinetics of HOBr Uptake on NaBr and NaCl Surfaces at Varying Relative Humidity
Liang T. Chu,* Guowang Diao, and Liang Chu†
Wadsworth Center, NYS Department of Health and State UniVersity of New York, P.O. Box 509,
Albany, New York 12201-0509
ReceiVed: September 20, 2001; In Final Form: April 1, 2002
The uptake kinetics of HOBr on NaBr and NaCl particle-film surfaces has been studied in a flow reactor
coupled with a differentially pumped quadrupole mass spectrometer. Spray-deposited and aerosol-deposited
salt-particle films were used for the measurement of HOBr reactivity. A rapid gas-surface reaction was
observed on the NaBr surface at 250 K, and the initial uptake coefficient γ
.004 and 0.021 ( 0.004 at relative humidities (RH) of 0.5 and 12.2%, respectively. γ
w
was determined to be 0.029 (
0
w
is lower at 260 K.
-
3
-4
For the reaction on NaCl particle surfaces, γ
w
decreases slightly from 2.0 ( 0.3 × 10 to 8.1 ( 1.3 × 10
at 250 K as RH increases from 1.5 to 18.2%. The results are interpreted in terms of the mechanism HOBr(ad)
+
NaX(s) f BrX + NaOH(s), which involves the adsorption of HOBr on the surface, with the reactions
occurring mainly on a NaX surface domain without water coverage at low RH. The surface morphology of
the particle films was characterized by scanning electron microscopy, and the tortuosity factor, τ, of the NaCl
particle film was determined to be 2.18. γ
coefficients γ . Typical γ
values for HOBr on NaBr and NaCl at 250 K and RH of 10% are 2.5 × 10 and
× 10 , respectively. Calculations suggest that the HOBr heterogeneous activation on NaBr may be
competitive, if surface regeneration occurs.
w
was corrected for internal surface effects to provide “true” uptake
-
3
t
t
-
5
5
3
I. Introduction
cm . Mochida et al. showed that the initial uptake coefficient
-
4
of HOBr on NaCl powder was approximately 9 × 10 at
The dramatic loss of surface-layer ozone at sunrise has been
observed in the high Arctic over the past decade. High levels
of BrO (20-30 pptv) were found in the polar boundary layer,
and these events were always associated with the ozone loss.
The observation is still only partly explained in terms of halogen
11
3 18
[
HOBr] ) 1 × 10 molecules/cm . If we compare the uptake
1
coefficient of HOBr on NaCl powder with that on the deliques-
cent NaCl aerosol, it suggests that the uptake coefficient likely
depends on water content. On the other hand, studies have
shown that both the water vapor and adsorbed water had an
2
-5
chemistry, and the details are not completely understood.
1
1,12,24
impact on the uptake of HNO3 on NaCl surfaces.
Davids
More recently, even higher levels of BrO (mixing ratio ∼86
pptv) were reported from the Dead Sea basin, where complete
boundary-layer ozone destruction was associated with episodes
and Cox showed that the uptake coefficient γ of HNO3 on NaCl
increases with increasing relatively humidity (RH) (0-35%).11
They suggested that the observation was closely associated with
6
of high BrO. It is believed that Br species are activated via
2
5-27
the adsorption of water on NaCl surfaces.
heterogeneous processes and play a critical role in the boundary-
layer ozone destruction. Barrie et al. have suggested an ozone-
loss mechanism involving catalytic Br cycles. Br and Cl in
sea-salt aerosols were proposed to be potential sources for the
generation of photochemically active halogen species.
The uptake coefficient of N2O5 on (NH4)2SO4 aerosols
decreases from 0.05 to 0.017 as RH increases from 50% to
3.5% at 297 K. The uptake coefficient of N2O5 on sulfuric
acid aerosols decreases from 0.05 to 0.02 as RH increases from
.7 to 80%.
to be dependent on RH. However, the most intriguing aspect
of the result is that γ decreases slightly as RH increases. This
implies that water may not play a direct reactive role in the
rate-determining step of the hydrolysis reaction. Behnke et
al. found that the reaction of N2O5 on wet NaCl aerosols had
two different reaction pathways. One involves the hydrolysis
of N2O5, and the other is the formation of ClNO2. The ClNO2
yield was higher at a low RH. The experimental evidence leads
1
2
8
9
7
28-30
0
The hydrolysis of N2O5 on aerosols is expected
The activation of Br or Cl by reactions with nitrogen oxides
over sea salt (mainly NaCl and NaBr) has been extensively
studied.8-15
2
8
N O + NaCl(s) f ClNO + NaNO (s)
(1)
2
5
2
3
3
1
The activation of Br and Cl on sea-salt particles also extends
to reactions with Cl and Br reservoir compounds such as
32
1
6-21
ClONO2 and HOBr with NaCl and KBr.
The latter was
us to hypothesize that the uptake of HOBr on NaX may depend
either on the amount of water on the NaX surface or on the
RH.
suggested to play a role in the generation of active Br.2 Abbatt
and Waschewsky showed that the uptake coefficient, γ, of HOBr
over deliquescent NaCl aerosols depended on the aerosol
2,23
Under typical atmospheric conditions, atmospheric aerosols
are surrounded by water vapor. The physiochemical properties
of the aerosols are a function of water content. The effect of
RH on the uptake of HOBr on NaCl and NaBr was not
previously investigated. This fact motivated us to investigate
the heterogeneous reactions of HOBr on NaBr and NaCl aerosol
2
1
acidity. γ decreased from 0.2 for buffered aerosols to <1.5 ×
-3
12
13
10
for unbuffered aerosols at [HOBr] ) 10 -10 molecules/
3
3
*
To whom correspondence should be addressed. E-mail: lchu@albany.edu.
Current address: University of California, 1 Shields Ave., Davis, CA
†
9
5616-8627.
1
0.1021/jp013594g CCC: $22.00 © 2002 American Chemical Society
Published on Web 05/09/2002

5680 J. Phys. Chem. B, Vol. 106, No. 22, 2002
Chu et al.
particle surfaces as a function of RH and temperature. Our study
sheds light on the mechanism of Br and Cl activation on sea-
salt particle surfaces as well as on the role of adsorbed water in
these reactions. It also provides uptake coefficients as a function
of RH.
directly, while the detachable glass tube was heated by a heating
tape at 120 °C. NaBr was deposited on the wall surface, while
excess CH3OH evaporated rapidly at the warmer temperature.
The NaBr film was then dried in the oven for 1 h at 160 °C.
The amount of NaBr deposited was determined by an analytical
3
We report heterogeneous uptake of HOBr on both NaBr and
NaCl surfaces via the two following reactions:
balance (0.1 mg accuracy). From the bulk density (2.1 g/cm ;
see below for details) and the geometric surface area of the
detachable glass tube, the thickness of the film was determined.
The typical thickness of the NaBr particle film was 15 ( 2
µm.
HOBr + NaBr f Br + NaOH
(2)
(3)
2
HOBr + NaCl f BrCl + NaOH
NaCl Film Preparation. The solubility of NaCl in methanol
is lower than for NaBr, and it is difficult to prepare a uniform,
micron-sized NaCl film using the above method. A different
approach was used to prepare NaCl particle films. The detach-
able glass tube was heated by a heating tape and the temperature
was kept at approximately 150 °C. The heated tube was mounted
in a vertical position. NaCl aerosols were generated by a
homemade capillary atomizer from saturated NaCl-water
solutions. The size of the atomizer was smaller than the inner
diameter of the detachable glass tube. NaCl solutions were
pressurized by 1.5-atm N2 gas to generate micron-sized NaCl-
H2O aerosols through the atomizer. The atomizer was admitted
into the vertically mounted detachable glass tube and the
aerosols were sprayed on the inner wall of the tube. Water
moisture was evaporated rapidly at 150 °C, and NaCl was
deposited on the wall surface as micron-sized particles. As the
atomizer was slowly pulled downward at a constant speed, a
uniform fine NaCl particle film was formed on the glass wall.
The thickness of the film was determined by the amount of
NaCl deposited on the wall surface, which varied by the
deposition time. The detachable glass tube was finally heated
in the oven at 150-200 °C for 1 h to ensure that water moisture
was evaporated.
Our study was carried out in a fast flow reactor, and the loss of
HOBr and products was detected by a quadrupole mass
spectrometer (QMS). In the following sections, we will briefly
describe the experimental procedures. The initial uptake coef-
ficient as a function of RH and temperature for reactions 2 and
3
will be presented. The results were modeled by a gas-surface
reaction scheme. The NaX surface morphology was examined
using scanning electron microscope (SEM), and the effect of
surface porosity on the measured uptake coefficient was studied
in detail.
II. Experimental Section
Apparatus. The experiments were conducted in a fast flow
reactor coupled to a differentially pumped quadrupole mass
spectrometer. The reactor was similar to the one used in previous
studies.3
4,35
However, several major modifications were made
on the flow reactor which are described below. The inner
diameter of the Pyrex flow tube was 2.63 cm, and it was 35 cm
in length. A detachable cylindrical glass tube with outer diameter
of 2.55 cm and inner diameter of 2.20 cm was tightly fitted
inside the flow tube. The salt-particle film was prepared on the
inner wall of the detachable glass tube. HOBr was admitted
via a movable injector positioned at the center of the flow
reactor. The movable injector consisted of two concentric tubes;
the inner one was a small capillary and was used to admit HOBr
to the reactor. The downstream end of the outside tube was
sealed to the inner tube. Room-temperature dry air was purged
between these two tubes so that HOBr and water vapor would
not condense on the wall of the capillary at low temperature.
The temperature of the reactor was regulated by a liquid-
nitrogen-cooled methanol circulator (Neslab) and was measured
with two J-type thermocouples (Omega, DP-41) located in the
middle and at the downstream end of the flow tube. The
thermocouples were calibrated at the triple point of water (273.2
K) with an accuracy of (0.2 K. The temperature on the inner-
wall surface of the detachable glass tube was measured in a
separate experiment using a thin-gauge thermocouple wire under
the same experimental conditions. The results showed that the
temperature on the inner-wall surface of the detachable glass
tube was 0.2-0.3 K warmer than that measured on the flow-
tube surface. The temperatures reported in the paper were
corrected for this difference. The total pressure in the flow
reactor was controlled by a downstream throttle valve (MKS
Instrument, Model 651C), and the stability was better than 0.001
Torr. The pressure was measured using a high-precision Baraton
gauge (MKS, Model 690A) located at the downstream end of
the flow tube. The typical pressure used in this study was either
The advantage of this method is that both the amount of NaCl
deposited on the detachable glass tube surface and the particle
size distributions can be controlled. This allowed us to study
the effect of particle film morphology on the uptake coefficient.
Both NaCl and NaBr particle films were freshly prepared for
every experiment.
Measurement of the Bulk Density. The NaX particle film
was deposited onto a glass rod (2.8 mm in diameter and 19 cm
in length) using the methods described above. The film was
dried in the oven, and the mass was determined on the analytical
balance. The salt-deposited rod was then protected by an
ultrathin layer of an acrylic film. The volume of both the
protected salt glass rod and the bare glass rod was measured
by the displacement of water in a high-precision 2-ml pipet
(precision (0.005 mL). The volume of the protected salt film
was determined. The volume of the protective acrylic film was
approximately determined by the difference in volume between
the acrylic-protected glass rod and the bare glass rod. This
volume was subtracted from the volume of the protected salt
film, and the volume of the salt-particle film was thus obtained.
The bulk density of the NaBr and NaCl films was determined
-
3
to be 2.1 ( 0.3 and 1.1 ( 0.2 g‚cm , respectively.
Preparation and Calibration of HOBr. The HOBr solution
was prepared using the method described in our previous work.36
Briefly, 3.9 g of Ag2SO4 (Aldrich, 99%) were dissolved into
100 mL of 0.25 M H2SO4 in a glass flask cooled in an ice-
water bath. Br2 (Aldrich, 99.5%) was added to the solution drop
by drop, while the solution was stirred, until the orange color
of excess Br persisted. After the solution settled in the glass
flask, the solution was decanted into a separatory funnel. Carbon
tetrachloride was used to extract the excess Br2 in the solution,
0
.5 or 1.0 Torr.
Preparation of Salt Films. NaBr Film Preparation. The
detachable glass tube was cleaned with dichromate sulfuric acid
cleaning solutions and rinsed several times with distilled water.
The tube was then dried in an oven at 160 °C. Saturated NaBr-
CH3OH solutions were sprayed on the inner wall of the tube

Kinetics of HOBr Uptake on NaBr and NaCl Surfaces
J. Phys. Chem. B, Vol. 106, No. 22, 2002 5681
and the heavier CCl4 phase was removed from the funnel. A
slightly yellowish-clear HOBr solution was obtained and was
kept in a bubbler in the dark at 273.2 K.
HOBr vapor was bubbled into the movable injector using
He carrier gas. FEP tubing and Teflon Swagelok fittings were
used. The flow of HOBr was controlled by Halocarbon grease-
treated stainless-steel needle valve. The flow rate of the HOBr
and He mixture was measured by a mass flow meter (Hasting-
Teledyne), which was corrected for the water vapor. The typical
flow rate of the mixture was 10 sccm. Additional He was mixed
with the HOBr flow at the inlets of the injector to minimize
loss of HOBr. All flow rates were controlled by either the mass
flow controller or metering valves. The transfer lines were first
passivated by flowing HOBr until a stable signal was reached
as monitored by QMS.
Figure 1. Uptake of HOBr on a NaBr particle film surface. A plot of
log [HOBr] versus reaction time t showed that the pseudo-first-order
Impurities in the HOBr-He mixture stream were monitored
by QMS. We found that Br2 (m/e ) 180) was the major
impurity. Other QMS signals at m/e ) 196 (Br2O) and 80 were
small (<10%). The signal at m/e ) 80 could be HBr. In a
-
1
-6
rate k
s
was 186 s and γ
w
) 0.022 at 250.3 K and PHOBr ) 1.1 × 10
Torr. The formation of Br
determined to be 194 s . The real time is shown on the top axis. See
text for details.
2
was observed, and the rate constant was
-
1
-
6
separate experiment, we monitored the loss of HBr (PHBr∼10
Torr) onto the dry salt surface at room temperature and did not
minimize the loss of water vapor onto the FEP transfer line,
which was at room temperature. In a separate experiment, QMS
was used to monitor the water vapor signal at various water
flow rates. We found that the water signal varied linearly with
the flow rates of the water-He stream, suggesting that the water
vapor was in equilibrium with its liquid in the water bubbler
and the water/He mixing ratio was equal to their partial pressure
ratio. The total water vapor in the reactor was the sum of the
water vapor from both the water vapor-He mixture and the
water vapor in HOBr. The water vapor pressure in the HOBr
solution was approximately equal to saturation water vapor
pressure at 273 K. The water vapor pressure in the reactor could
be estimated from the flow rates and mixing ratios. However,
error was introduced at lower RH where the flow rate of the
water-He mixture was comparable with the HOBr flow rate
-
3
find any measurable loss. When there was water vapor (∼10
Torr) in transfer lines, HBr was lost on the lines. This suggests
that the HBr impurity, if any, should not affect the HOBr uptake
measurements. Gas-phase Br2 was from the HOBr solution and
from the decomposition of the HOBr vapor. The following
procedures were used to minimize Br2. Br2 accumulated during
the night in the HOBr bubbler was purged by He prior to
experiments. The needle valve, which controlled the flow of
HOBr, was cleaned frequently and was treated with Halocarbon
grease to minimize decomposition of HOBr on metal surfaces.
Both HOBr and Br2 signals were monitored using QMS until
their signal levels were stable prior to every uptake measure-
ment. The Br2 signal intensity was usually less than that of
HOBr (5-50%).
The [HOBr] was calibrated using the following experiment.36
An ice film was prepared on the wall of the flow tube at 190 K
(10 sccm). To eliminate the error, the water vapor pressure was
calibrated against the ice vapor pressure at 203-218 K (Pice )
(
1
without the detachable glass tube). A higher [HOBr] (∼5 ×
-3
37
1
.94 × 10 Torr at 203 K ) using the QMS under the same
-
6
0
Torr) was exposed to the ice surface for more than 30
total pressure and flow conditions. The estimated water vapor
pressure in the reactor was within 8% of the vapor pressure
measured by QMS. In this paper, the reported RH was measured
by QMS and the relative error was estimated to be <8%.
Measurement of the Uptake Coefficient. The uptake
coefficient of HOBr on salt-particle films was measured by
monitoring the gas-phase HOBr loss over the salt surface. Before
the measurement, the injector was placed in front of the salt-
particle film (the downstream end of the flow reactor), and the
initial HOBr signal was recorded by QMS at m/e ) 96. The
injector was pulled back over the salt surface 2 cm at a time.
The loss of HOBr was recorded as a function of the injector
position, z. The typical data-acquisition time for each point was
10 s. The decay of gas-phase HOBr on the salt surface can be
min so that the surface was nearly saturated. A lower [HCl]
-7
(
∼5 × 10 Torr) was allowed to react with the adsorbed HOBr
to produce BrCl. The BrCl signal was monitored by QMS at
m/e ) 114. The number of BrCl molecules formed was equal
-
to that of HCl. The BrCl sensitivity for the QMS (count/Torr
sec) was determined. Subsequently, we reversed the ratio of
[HOBr] to [HCl]. We determined the signal ratio of HOBr to
BrCl, and thus the HOBr sensitivity was determined.
The sensitivity of the QMS may change slightly over a period
of weeks. We also determined the signal ratio of HOBr to H2O
over the HOBr solution. Since HOBr solutions were very dilute,
the water vapor pressure over the solutions was approximately
constant at 273 K. The [HOBr]/[H2O] signal ratio provided an
excellent guideline to maintain constant [HOBr] throughout the
entire experiment.
3
5,38
described as
Relative Humidity. Water vapor, carried by He from a
temperature-controlled water bubbler, was admitted to the flow
reactor to control the RH over NaX surfaces, and the water vapor
was allowed to flow a sufficiently long time to achieve an
equilibrium with the surface. The flow rate of water-He mixture
was controlled by a metering valve and measured by the mass
flow meter (Hasting-Teledyne). A low flow rate (<100 sccm)
was chosen to ensure the equilibrium between water and its
vapor in the bubbler and to maintain a constant water vapor-
to-He mixing ratio. The water bubbler was kept at 278 K to
ln[HOBr]
z
) - k
s
(z/V) + ln[HOBr]
0
(4)
where [HOBr]z and [HOBr]0 are the concentrations of gas-phase
HOBr when the injector is at position z and that at the reference
point 0, respectively, V is the mean flow velocity, and ks is the
pseudo-first-order loss rate constant. ks was determined by a
plot of ln[HOBr]z versus the reaction time t ) z/V. A typical
experimental result for the HOBr loss on a NaBr film surface
-
1
is shown in Figure 1. ks was determined to be 186 s using
the data collected at t < 8 ms. After the gas-phase axial and

5682 J. Phys. Chem. B, Vol. 106, No. 22, 2002
Chu et al.
Figure 3. HOBr uptake on an aerosol-deposited fine NaCl particle
-
1
surface. A plot of log [HOBr] vs t showed k
1
s
to be 12.4 s and γ
.2 × 10 at 250.7 K. The formation of BrCl was detected by QMS
at m/e ) 114 and is shown on the right-hand axis of the plot. The rate
w
)
-
3
-
1
s
constant k was determined to be 13.5 s .
-
3
NaBr and NaCl films were measured to be 2.1 and 1.1 g‚cm ,
respectively. The true densities Ft of NaBr and NaCl are 3.20
-
3 37
and 2.17 g‚cm , respectively. The ratio of Fb/Ft for NaBr is
.66, which is close to 0.74, a value predicted by the pore-
diffusion model on the basis of the hexagonal close packing
0
4
2
(
HCP) of spheres. The density ratio for NaCl is 0.51, which
almost coincides with the value of 0.52 predicted by simple
42
cubic packing (SCP). The experimental data were treated using
two different granule-packing structures.
The “true” uptake coefficient, γt, can be calculated by eqs 7
4
2
and 8 for SCP and HCP, respectively:
Figure 2. Scanning electron micrographs of NaCl particle films as
O aerosols on a glass surface. (Top)
NaCl film is porous; (bottom) salt granules can be treated approximately
by a layered structure. A relationship between the film thickness and
prepared by depositing NaCl-H
2
γ
w
γt )
(7)
(8)
(
π/2)[1 + η(2N - 1)]
L
L
number of granule layers N can be determined from this micrograph.
Scale bars are in 10.0 µm (top) and 1.0 µm (bottom), respectively.
x
3γw
3
9
radial diffusion correction was applied using a standard method,
γ )
t
1
/2
-
1
π{1 + η[2(N - 1) + (3/2) ]}
the corrected rate constant, kw, was determined to be 232 s .
L
The diffusion coefficient of HOBr in He and water vapor was
calculated using
4
0
where η is the effectiveness factor defined by the fraction of
the film surface that participates in the reaction. Its determination
will be given in the Results section. NL is the number of granule
layers, which was determined by SEM.
-
1
xHe
HOBr
xwater
DHOBr
)
+
(5)
HOBr
(
)
D
D
He
Water
where xHe and xwater are the mole fractions of He and water,
Results
HOBr
HOBr
respectively. D
He and D
water are the diffusion coefficients
of HOBr in helium and water, respectively; they were estimated
2
-1
36,40
to be 0.340 and 0.121 cm ‚s at 250 K.
coefficient, γ, was calculated by38
The uptake
(6)
Uptake of HOBr on NaCl. Figure 3 shows the loss of HOBr
on a NaCl aerosol-deposited particle film at 250.7 K. The gas-
phase HOBr signal decreased with time t over the NaCl surface.
At the same time, the formation of BrCl was observed. A plot
of the BrCl signal versus t is shown on the right-hand axis of
Figure 3. The BrCl signal was undetectable by the QMS without
the presence of the NaCl sample in the flow reactor. This
suggests that reaction 3 was occurring on the NaCl surface. An
average [BrCl]/[HOBr] ratio was determined to be 0.62 ( 0.39
on the basis of a calibration of the BrCl/HOBr signal ratio from
the reaction of HOBr + HCl(s) f BrCl + H2O(s). A linear
plot of log [HOBr] versus t suggests the loss of HOBr on the
NaCl surface be to pseudo-first-order. The pseudo-first-order
γ ) 2rk /(ω + rk )
w
w
where r is the radius of the detachable glass tube (1.10 cm),
and ω is the mean molecular velocity for HOBr at 250 K. When
the HOBr uptake was measured on a freshly prepared clean
salt surface, we called the uptake coefficient to be the initial
uptake coefficient, γw.
The surface morphology of salt-particle films was examined
using both the optical microscope and SEM. The NaCl particle
films appeared “uniform” under a low-magnification optical
microscope; however, the fine salt-particle films used in this
study appeared porous under SEM in Figure 2. We adopted a
pore-diffusion model to correct for the internal surface areas of
-1
wall-loss rate constant was determined to be ks ) 12.4 s . After
the gas-phase diffusion corrections were taken into consider-
ation, kw and the initial uptake coefficient γw were determined
4
1,42
-1
-3
the particle film
The model has been successfully applied
to be 12.8 s and 1.2 × 10 , respectively. A rate constant ks
10,43
to treat sea salt and metal oxides.
The bulk densities Fb of
was also determined from the formation of BrCl. The BrCl data

Kinetics of HOBr Uptake on NaBr and NaCl Surfaces
J. Phys. Chem. B, Vol. 106, No. 22, 2002 5683
TABLE 1: Uptake of HOBr on NaCl Films at 250 Ka
V (m‚s-1)
k
(s-1
)
k
(s-1
)
γ
γ
b
RH%
T (K)
s
w
w
t
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
3
3
3
3
3
3
4
3
3
3
3
4
3
3
3
-5
-5
-5
-5
-5
-5
-5
-5
-5
-5
-5
-5
-5
-5
-5
1
1
4
5
6
8
9
9
.5 ( 0.2
.9 ( 0.2
.6 ( 0.4
.3 ( 0.5
.2 ( 0.6
.3 ( 0.9
.0 ( 1.2
.6 ( 0.8
250.1 ( 0.2
250.5 ( 0.4
250.4 ( 0.2
250.1 ( 0.2
250.6 ( 0.2
249.9 ( 0.3
250.3 ( 0.2
250.6 ( 0.2
250.2 ( 0.2
250.6 ( 0.4
250.4 ( 0.2
250.3 ( 0.2
250.4 ( 0.2
250.1 ( 0.3
250.2 ( 0.3
4.8
4.9
4.7
4.3
4.8
5.1
4.6
5.6
4.7
4.9
4.9
4.7
4.4
4.4
4.7
20.5 ( 2.5
15.5 ( 2.5
15.8 ( 6.0
12.3 ( 1.3
18.6 ( 2.2
12.0 ( 1.4
9.3 ( 2.5
18.2 ( 7.0
11.7 ( 0.7
10.9 ( 2.1
12.6 ( 1.8
8.4 ( 1.1
21.5 ( 2.2
16.0 ( 2.7
16.4 ( 6.5
12.7 ( 1.4
19.4 ( 2.4
12.3 ( 1.5
9.5 ( 2.6
19.0 ( 7.6
12.0 ( 0.7
11.2 ( 2.2
13.0 ( 1.9
8.5 ( 1.2
(2.0 ( 0.3) × 10
(1.5 ( 0.3) × 10
(1.5 ( 0.6) × 10
(1.2 ( 0.2) × 10
(1.8 ( 0.3) × 10
(1.2 ( 0.2) × 10
(8.9 ( 2.5) × 10
(1.8 ( 0.7) × 10
(1.1 ( 0.2) × 10
(1.1 ( 0.2) × 10
(1.2 ( 0.2) × 10
(8.1 ( 1.3) × 10
(2.1 ( 1.0) × 10
(1.3 ( 0.2) × 10
(2.2 ( 0.3) × 10
6.6 × 10
4.9 × 10
5.0 × 10
3.9 × 10
6.0 × 10
3.9 × 10
2.9 × 10
5.8 × 10
3.6 × 10
3.5 × 10
3.9 × 10
2.6 × 10
6.9 × 10
4.2 × 10
7.2 × 10
1
1
1
1
1
1
2
0.7 ( 0.9
3.6 ( 1.0
4.6 ( 1.1
8.2 ( 1.4
9.0 ( 1.5
9.8 ( 1.4
2.5 ( 2.0
21.5 ( 9.7
13.4 ( 1.6
22.3 ( 2.9
22.8 ( 10.8
13.9 ( 1.8
23.5 ( 3.2
a
HOBr ) (3.8 ( 0.3) × 10 Torr, Ptotal ) 1.00 Torr, thickness ) (10.0 ( 0.5) µm. b
-6
P
L
N ) 10 and τ ) 2.18 were used in the SCP calculation.
Figure 4. Plot of γ
for HOBr on aerosol-deposited NaCl particle films. The plot shows
that γ decreases slightly with RH when RH < 18% and then shows
a tendency toward higher γ
w
as a function of relative humidity (RH) at 250 K
Figure 5. Plot of the initial uptake coefficient as a function of the
-
6
NaCl film thickness at PHOBr ) 3.2 × 10 Torr, RH ) 1.6%, and 250
w
K. The solid line was a least-squares fit of the data to the SCP pore-
w
values at RH > 18%. The solid line is a
diffusion model (eq 7). γ and τ were determined from the least-squares
t
-
5
prediction from a kinetic scheme (see text) with a function of γ )
fit to be 6.3 × 10 and 2.18, respectively. The dotted, double-dot-
dashed, and dashed lines were calculated using the model at τ ) 1, 4,
and 6, respectively. This suggests that the tortuosity factor τ is not
very sensitive to the shape of the curve in the range of τ ) 1-4. For
illustration purposes, one standard deviation of the mean is included
on some data points.
a(1 - dlog(bPH O)) where a ) 0.01341, b ) 0.03078, and d ) 0.1458.
2
The error bars represent one standard deviation of the mean.
were fitted to [BrCl] ) [BrCl]∞(1-exp(-kst)), where [BrCl]∞ is
the concentration of BrCl at t ) ∞. Both ks and [BrCl]∞ were
-
1
fitting parameters. ks was determined to be 13.5 s . This is in
good agreement with ks determined from the HOBr loss.
Figure 4 shows the initial uptake coefficient γw as a function
of RH. The detailed experimental conditions are listed in Table
µm, the diffusion time is longer than the HOBr surface residence
time and HOBr molecules cannot effectively access all internal
surfaces. This appears as a plateau in Figure 5. On the basis of
Fb/Ft ) 0.51, the NaCl particle film was modeled by SCP. A
nonlinear least-squares fit of all the data, γw, to eq 7 is shown
as a solid line in Figure 5. η in eq 7 was computed from the
1
. Each datum except γt in the table was the average of two to
six determinations. Every determination of ks was one experi-
ment (e.g., Figure 3) conducted on a freshly prepared NaCl
surface. Errors listed in Table 1 are one standard deviation of
the mean of a number of determinations. Systematic errors in
RH and T were included. The true uptake coefficient, γt, was
calculated from γw using eq 7; the parameters used in the
calculation will be discussed below. The uptake coefficient
decreases slightly with RH at RH < 18% (Figure 4). When the
RH is higher than 18%, the initial uptake coefficient shows a
small increase with RH. This observation may be related to the
multilayer adsorption of H2O on NaCl at higher RH (see
Discussion).
4
2
following equations
-
1
η ) φ tanh φ
(9)
1
/2
φ ) (N - 1)[3F /2(F - F )](3τγ )
(10)
L
b
t
b
t
where τ is the tortuosity factor. The nonlinear least-squares fit
was calculated by allowing both γt and τ to vary according to
eqs 7, 9, and 10; and γw as a function of NL(h) was obtained.
A functional relationship between NL and h was determined on
the basis of SEM micrographs such as Figure 2 (bottom). The
result is shown in Figure 6. The solid line in Figure 6 is a least-
squares fit to the experimental data and the fitted result was
used in this model. The true uptake coefficient, γt, and τ were
The initial uptake coefficient γw was determined as a function
of the thickness, h, of the particle films. The initial uptake
coefficient of HOBr rapidly increases with the film thickness
when h is less 50 µm, and then γw increases gradually up to
-
5
∼
165 µm at 250 K (Figure 5). h is defined as m/AFb. This
determined from the nonlinear least-squares fit to be 6.3 × 10
suggests that the particle film is porous and has internal surface
areas. HOBr molecules can access internal surfaces by gaseous
diffusion from the exterior of the solid. However, at h > 100
and 2.18, respectively. Figure 5 also illustrates the effect of the
tortuosity factor on the profile of γw. The curves were calculated
at constant τ ()1-6). This analysis suggests that a set of curves

5684 J. Phys. Chem. B, Vol. 106, No. 22, 2002
Chu et al.
TABLE 2: Uptake of HOBr on NaBr Filmsa
V (m‚s-1)
k
(s-1
)
k
(s-1
)
γ
γ
b
RH%
T (K)
s
w
w
t
-
-
-
-
-
-
-
-
-
-
-
-
-
3
3
3
3
3
3
3
3
3
3
3
3
3
0
0
.53 ( 0.04
.72 ( 0.06
250.1 ( 0.3
250.3 ( 0.2
250.1 ( 0.2
250.3 ( 0.2
250.2 ( 0.2
250.4 ( 0.2
250.4 ( 0.2
250.0 ( 0.2
260.1 ( 0.5
260.1 ( 0.2
259.9 ( 0.7
260.1 ( 0.3
260.0 ( 0.3
10.1
9.5
10.0
10.0
8.8
9.3
8.4
237 ( 24
208 ( 26
223 ( 22
223 ( 53
177 ( 42
227 ( 32
172 ( 23
169 ( 18
167 ( 74
198 ( 61
153 ( 28
239 ( 26
160 ( 55
307 ( 36
265 ( 39
285 ( 29
288 ( 81
224 ( 65
292 ( 49
218 ( 37
221 ( 24
206 ( 110
245 ( 88
180 ( 35
309 ( 43
195 ( 80
0.029 ( 0.004
0.025 ( 0.004
0.027 ( 0.004
0.027 ( 0.008
0.021 ( 0.006
0.028 ( 0.005
0.020 ( 0.003
0.021 ( 0.004
0.019 ( 0.009
0.022 ( 0.008
0.017 ( 0.003
0.028 ( 0.004
0.018 ( 0.007
3.1 × 10
2.4 × 10
2.7 × 10
2.7 × 10
1.8 × 10
2.9 × 10
1.6 × 10
1.8 × 10
1.5 × 10
1.9 × 10
1.2 × 10
2.9 × 10
1.4 × 10
2
4
7
9
.7 ( 0.2
.3 ( 0.3
.4 ( 0.7
.7 ( 0.8
1
1
0
1.5 ( 1.1
2.2 ( 1.9
.11 ( 0.04
7.2
10.3
11.1
10.1
10.4
9.9
1
2
2
4
.2 ( 0.2
.1 ( 0.3
.8 ( 0.3
.1 ( 0.4
a
HOBr ) (1.0 ( 0.3) × 10 Torr, Ptotal ) 0.500 ( 0.002 Torr, thickness ) (15 ( 2) µm. b
-6
P
L
N ) 12 and τ ) 2.0 were used in the HCP model
calculation.
Figure 6. Plot of number of layers N
thickness h. The solid line was fitted to an empirical correlation N
alog(h + b), where a ) 10.5217 and b ) 0.044.
L
versus the NaCl particle film
Figure 7. Plot of the initial uptake coefficient as a function of RH for
HOBr on NaBr particle surfaces at 250 K. The solid line is a fit to the
model as described in the text. Errors are one standard deviation of
the average of a number of determinations.
L
)
with τ ) 1-4 can reasonably represent the experimental result,
and the overall shape of the curve is not very sensitive to τ in
the range of 1-4. Mochida et al. adopted τ ) 2 for NaCl powder
and sprayed film surfaces.18 Their adopted value is very close
to our best fit. The pore-diffusion model was adopted to correct
all γw listed in Table 1 to obtain true uptake coefficient γt using
τ ) 2.18.
2
along with the detailed experimental conditions. Errors are
one standard deviation of the average of a number of determina-
tions. Systematic errors in RH and T were included. γ values
measured at 260 K are slightly lower than that of 250 K. The
w
difference in γw between 250 and 260 K is small, but the
difference is beyond the error of the measurements. γw was
corrected for the surface porosity using the HCP pore-diffusion
model (eq 8). It is not feasible to vary spray-deposited NaBr
film thickness in a wide range to extract surface morphology
information in this experiment. SEM micrographs showed that
NaBr particle films were porous. We adopted τ ) 2 in the
calculation. As we have shown previously (Figure 5) and by
Uptake of HOBr on NaBr Films. A typical plot for HOBr
loss on an NaBr spray-deposited film is shown in Figure 1. A
product, Br2, was observed and is plotted on the right-hand axis
of the figure. However, no net Br2 signal increase was observed
when HOBr was flowed through the reactor without NaBr. The
pseudo-first-order wall-loss rate constant ks was determined to
4
1
-
1
Keyser et al., the error in γ caused by τ is estimated to be
less than 10%. γt was calculated from the mean value of γw
1
2
86 s . The diffusion-corrected first-order rate constant kw was
t
-
1
-2
32 s , and γw ) 2.2 × 10 . ks was also determined from the
-
1
directly using eqs 8-10.
formation of the Br2 signal to be 194 s . Both results are in
very good agreement.
Surface Deactivation. Figure 8 shows that γ of HOBr on
the NaBr surface decreases as the number of repeated experi-
ments increases. The initial uptake coefficient is 0.022, and the
subsequent measurements yield γ ) 0.019, 0.014, and 0.010. γ
decreases about 14% after the first uptake determination and
that is about 10 min in real time. This indicates that the uptake
coefficient of HOBr on NaBr decreases with the increase of
the surface exposure, and it is a clear indication of surface
deactivation. A similar pattern was found for NaCl surfaces:
The initial uptake coefficient of HOBr on NaBr film was
studied as a function of RH at 250 and 260 K. Figure 7 shows
a tendency that γw decreases slightly from 0.029 ( 0.004 to
0
.021 ( 0.004 as RH increases from 0.53 to 12.2%. Since the
decrease in γw is faintly larger than the uncertainty of the
measurement (a typical (σ ) 0.004, see Figure 7), the effect
of RH on γw is very small over this RH range. γw of HOBr on
NaBr is a factor of 10-15 higher than that on the NaCl film.
This experiment was conducted at a total pressure of 0.5 Torr.
This restricts the water pressure allowed in the reactor or the
range of RH can be used for the experiments. However, it
reduces the gas-phase diffusion correction on ks, and a higher-
accuracy γw value can be obtained under the fast-flow condition.
The results measured at 250 and 260 K are tabulated in Table
-
3
the initial uptake coefficient is 1.6 × 10 , and the subsequent
-
3
-4
determinations of γ are 1.1 × 10 and 5.1 × 10 . These
observations suggest that both active NaBr and NaCl surface
sites could be either saturated with HOBr or sites are deactivated
by the formation of NaOH, a product of the reaction (see
Discussion).

Kinetics of HOBr Uptake on NaBr and NaCl Surfaces
J. Phys. Chem. B, Vol. 106, No. 22, 2002 5685
The surface morphology of the NaX particle films was taken
into consideration. We found that the surface morphology of
particle films, and the shapes and sizes of the salt granules,
depended on the methods of preparation. NaCl granules had
irregular shapes (Figure 2), and the rounded edges of granules
suggest that the surface layer of the granules was wet. However,
when NaCl particle films were prepared from saturated NaCl-
CH3OH solutions, the granules showed a cubic crystal-like
shape. The films remained porous with internal surface areas.
The use of the pore-diffusion model to correct for internal
surface areas of the particle films is an approximation. Figure
2
shows granules of salt particles are not perfect spheres as
assumed in the model. The packing is not exactly SCP either.
However, the model did very well in reproducing the thickness
dependence on γw (Figure 6). This suggests that the treatment
is acceptable provided that one can validate the method as
illustrated in Figure 6.
Figure 8. HOBr uptake on a spray-deposited NaBr film is shown as
-
6
repeated measurements at 250.3 K, PHOBr ) 1.5 × 10 Torr, RH )
.75%, and total pressure of 0.50 Torr. The arrows indicate the reference
position z of each determination. The initial uptake coefficient γ is
.022, and subsequent γ are 0.019, 0.014, and 0.010, respectively. The
0
Water on NaCl and NaBr Surfaces. Water molecules are
known to be adsorbed on both NaCl and NaBr surfaces.4
Adsorption of water on NaCl surfaces under ambient conditions
0
w
4,45
0
result suggests that the NaBr surface becomes saturated as HOBr
exposure increases. The real time of the experiment is shown on the
top axis of the plot.
2
5-27,46-49
has been extensively studied.
Ewing and co-workers
found that adsorption of water on the NaCl(100) surface at low
IV. Discussion
humidity (RH < 30%) was submonolayer and started to form
“
islands” and multilayer coverage at higher humidity. At
Uptake. We have shown that the NaX surface sites can be
deactivated upon the reaction with HOBr. The surface deactiva-
tion is attributed to either HOBr adsorbing on an active surface
site or a product, NaOH, consuming a site. Figure 8 demon-
strates that γ decreases ∼14% after the NaBr surface was taken
up by HOBr in the γw determination. In the subsequent
determinations, γ was even lower. Mochida et al. observed a
multilayer coverage, there was three-dimensional growth on the
surface with a low concentration of ions from the substrate being
incorporated into the multilayer. At midcoverage, there was a
transition between two-dimensional and three-dimensional
growth.
The deliquescence point (DP) is 80 ( 4% RH at 253 K for
5
0
NaCl. The phase diagram of NaCl-H2O indicates that the
similar but stronger trend of decreasing uptake coefficient with
eutectic temperature of NaCl‚2H2O-ice-liquid is 252 K.5
1-53
At
exposure to HOBr on KBr spray-deposited surfaces.18 They also
low RH (RH < 22%) and 250-260 K, metastable liquid phase
should not be formed in the region of interest in this study, and
the bulk composition of the solid phase is mainly NaCl as
showed that the initial uptake coefficient of HOBr on NaCl and
KBr decreased as [HOBr] increased, for [HOBr] less than 2 ×
10
3
1
0 molecules/cm . However, the decrease in the initial uptake
5
2
1
0
suggested in the phase diagram. The eutectic temperature for
coefficient was insignificant when [HOBr] was >2 × 10
52
3
ice/NaBr‚5H2O is ∼245 K. Similarly, metastable liquid is not
expected to form in a phase region with RH < 12%, as we
assume that the phase diagram of NaBr/H2O is similar to that
of NaCl/H2O at 250 K. This suggests that the reactions of HOBr
with NaCl and NaBr surfaces resemble the gas-surface
reactions at low RH.
molecules/cm . This also suggests that surfaces can be saturated
with HOBr.
A typical measurement error in γw is ∼15% or higher (see
Tables 1 and 2). A possible source of error includes some degree
of irreproducibility of particle surfaces from experiment to
experiment, the surface deactivation, and the HOBr signal
fluctuation (random error). To minimize the surface deactivation
effect, the following procedures were implemented during this
study: (1) a shorter data acquisition time (∼3 min in the real
time to determine γw, cf. Figure 1); (2) freshly prepared NaX
surfaces; and (3) lower [HOBr] (∼10 Torr). The error analysis
suggests that the effect of the surface deactivation on γw is
reduced to a level that is comparable to the other random errors
of the measurement.
γw was not measured as a function of [HOBr] because of a
limitation in adjusting both [HOBr] and RH over a wide range
of values. However, on the basis of the work by Mochida et al.
and the results from this study, we expect that γw would vary
slightly under the range of [HOBr] employed in this study.
Figure 1 shows the HOBr uptake remains first order for ∼10
ms. However, the Br2 release curves off at ∼8 ms. A possible
explanation is that some active NaBr surface sites were adsorbed
by the product toward the end of the measurement and these
sites were deactivated. Thus, the formation of Br2 appears to
be lower and shows the tread of curve off at 8 ms. Nevertheless,
gas-phase HOBr molecules may have other loss pathways, as
will be discussed later. It remains the first-order loss of almost
Hemminger, Finlayson-Pitts, and co-workers showed that
when gaseous HNO3 molecules adsorbed on the NaCl(100)
5
4,55
surface, a metastable nitrate film was initially formed.
Water
5
5
vapor (<75% RH, below the DP of NaNO3, 74.3% ) was then
-6
deposited on the film, the ionic mobility of metastable NaNO3
increased, and deliquescence-like behavior occurred over the
NaCl surface. This resulted in a quasi-liquid layer in which the
NaNO3 ion mobility was substantially enhanced. The precipita-
tion of stable NaNO3 finally occurred below DP.
A similar process may occur for the HOBr uptake on NaCl
surfaces. After HOBr is adsorbed on NaCl, it reacts with NaCl
to form NaOH and BrCl. NaOH is adsorbed on NaCl surfaces,
but crystallization is thermodynamically unfavored. The ad-
sorbed water increases the ionic mobility of NaOH and transfers
NaOH into stable thermodynamic states and regenerates active
-
Cl sites. This is a driving force for the reaction to be
55
maintained. The DP of bulk NaOH is ∼7%. We do not expect
the precipitation of NaOH crystals; rather, NaOH is in a
deliquescence-like state on the surface under most experimental
conditions. The rate-limiting step is expected to be the reaction
-
between HOBr(ad) and Cl in which water may not play a direct
1
0 ms.
role. This qualitatively explains the weak dependence of γw on

5
686 J. Phys. Chem. B, Vol. 106, No. 22, 2002
Chu et al.
(17)
-
RH over the RH range of 1-18% (Figures 4 and 7). A
quantitative discussion will be presented below.
Mechanism. The experimental observations for the HOBr
reaction with the water-adsorbed NaCl solid salt are consistent
with the reaction occurring in the following steps: (1) water
vapor is in adsorption equilibrium with the NaCl surface; (2)
HOBr reacts with Cl to form BrCl while H2O is adsorbed on
the surface; (3) either product or HOBr consumes active NaCl
sites; (4) at a higher water coverage, ions from NaCl crystal
reactive Cl site:
SoP
HOBr
A
V
rate ) kK K
(1 - θ)
1
21
+ 2K K P
1 2 HOBr
where K1 ) k12/k-12 and K2 ) k13/k-13. The uptake coefficient
4 rate
ω[HOBr] A
-
V
γ )
can be written as
2
6,27,47
So(1 - θ)
4RT
ωL
are incorporated into three-dimensional water multilayer,
and HOBr can react with Cl over the adlayer. These processes
can be expressed by the following reactions:
γ ) kK K
)
1
2
-
1 + 2K K P
1 2 HOBr
So(1 - C ln cP
)
H O
2
4RT
ωL
kK K
(18)
1
2
1
+ 2K K P
1 2 HOBr
H O(g) + NaCl(s) T H O(ad)
(11)
(12)
(13)
2
2
where the Temkin isotherm was used to determine θ. c and C
are constants used in the Temkin isotherm and L is the Avogadro
constant. The Temkin isotherm describes many adsorption
systems in the coverage of 0.2 < θ < 0.8. Some constants in
eq 18 can be combined (PHOBr is approximately a constant in
HOBr(g) T HOBr(p)
HOBr(p) + S T HOBr(ad)
5
6
k
-
HOBr(ad) + Cl (s) T HO‚‚‚Br‚‚‚Cl(ad)
9
8
this study) and the equation γ ) a(1 - dlog(bPH O)) was used
2
-
to fit uptake coefficients of HOBr on NaCl and NaBr at constant
temperature and [HOBr] conditions; we assume that both
mechanisms are the same. The results are plotted as solid lines
in Figures 4 and 7. Overall, the fitted curves represent the
experimental results well. This is a good indication that the
proposed mechanism represents a correct reaction pathway at
low RH (<18%) region. The uptake coefficient of HOBr on
NaCl surfaces in Figure 4 appears to show a tendency of a slight
increase in γw at RH > 18%. Equation 18 ignored the reaction
BrCl(g) + OH (ad) (14)
k′
HOBr(p) + H O(ad)
9
8 HOBr(ad)
(14′)
2
k′′
-
-
HOBr(ad) + Cl /H O(s)
9
8 BrCl(g) + OH (ad) (14′′)
2
Reaction 11 represents the water adsorption equilibrium, and
the surface is ionic polycrystalline NaCl. NaCl(s) should be
taken as Na and Cl ions. HOBr is assumed to be trapped in
a precursor state (reaction 12) before it migrates over the surface
and reacts with Cl ions as on the ice surface. Cl (s) can be
from both NaCl surfaces and in H2O multilayer. The observed
HOBr loss rate can be expressed as
+
-
-
of HOBr with Cl over the water multilayer (reaction 14′′). If
we take this factor into consideration, the rate expression
includes a term proportional to θ‚f. Qualitatively, this predicts
that γ increases with PH O or RH at higher humidity. In addition,
the loss of HOBr to OH is completely ignored in eq 15, since
the amount of OH or NaOH formed is low initially in the time
period over which the initial uptake coefficient was determined.
However, at a region where water coverage is either monolayer
or multilayer (RH ∼ 30%), a deliquescence-like NaOH layer
may cover some parts of the NaCl surface. In basic solutions,
-
36
-
2
-
-
d[HOBr]/dt ) {k[HOBr(ad)](1-θ) + k′ [HOBr(p)]θ +
k′′ [HOBr(ad)] f θ}A/V (15)
where θ is the water surface coverage, f is a fraction of
multilayer islands on the water adlayer, and A/V is the surface-
to-volume ratio of the reactor. The first term in eq 15 describes
the reaction of HOBr with Cl on NaCl surface sites (reaction
1
5
7
a disproportion reaction can take place.
-
-
-
HOBr + OH f H O + OBr
(19)
(20)
4), which are not adsorbed by water; the second term is the
2
loss of HOBr on adsorbed H2O (reaction 14′); and the last term
is the reaction of HOBr with Cl over multilayer water islands
-
-
-
-
3OBr f 2Br + BrO
3
(reaction 14′′). At RH < 25%, the water coverage is in the range
2
6,46,49
of submonolayer to monolayer and f ≈ 0.
The uptake
This channel creates an additional HOBr loss and the observed
initial uptake coefficient could increase at higher RH. M o¨ ssigner
and Cox showed the uptake of HOI on NaOH is higher than
coefficient of HOBr on adsorbed H2O at 250 K is expected to
-
4
be ∼10 or less. This is because the initial uptake coefficients
of HOBr on the ice surface are 1.2 × 10 and 5.8 × 10 at
29 and 239 K, respectively.36 Thus, it is expected that
-
3
-4
58
0.1 at RH ) 11% and 278 K.
2
Equation 18 shows that the uptake coefficient is inversely
proportional to PHOBr. This explains the uptake coefficient
decreases with the increase in PHOBr as observed by Mochida
et al. The observations in Figure 8 can be explained:
Subsequent increase in HOBr exposure to the surface resulted
in the adsorption of OH (ad) and HOBr(ad) on the surface;
this reduces the amount of available reactive Cl sites, S, on
the surface. A surface deactivation was observed in the
experiment.
k′[HOBr(p)]θ < k[HOBr(ad)](1-θ), after taking into consid-
eration the fraction of adsorbed H2O on the surface and the
HOBr uptake coefficient on ice. This implies that the reaction
occurs predominately on those sites without water adsorption,
but those sites next to adsorbed water, defects, and steps are
favored. As a first-order approximation, we can rewrite eq 15
as (i.e., reaction 14 only)
1
8
-
-
rate ≈ k[HOBr(ad)](1 - θ) (A/V)
(16)
The adsorption of water on NaBr surfaces is similar to that
on NaCl surfaces. The shape of water isotherms on both surfaces
44,45
A rate expression can be derived using the steady-state ap-
proximation and total available NaCl surface site So ) S +
at P/P < 0.2 is nearly the same,
and thus predicted profiles
o
of the uptake coefficient based on eq 18 fit the results of HOBr
on both NaCl and NaBr very well. The ionic nature of the
reaction (cf. eq 14) is expected on the basis of reactions HOBr
[
HOBr(ad)] + [OH(ad)] ≈ S + 2[HOBr(ad)], where So ) total
NaCl sites - water adsorbed NaCl sites and S is the available

Kinetics of HOBr Uptake on NaBr and NaCl Surfaces
J. Phys. Chem. B, Vol. 106, No. 22, 2002 5687
with HCl and HBr;36,59 therefore, the mechanism outlined above
applies to both cases, as was demonstrated in Figures 4 and 7.
The conclusion of this discussion is the following: Under
conditions of low RH (RH < 18%), the reaction takes place
predominately on NaCl surface sites without water adsorption.
The water coverage, θ, does not dramatically change with RH,
et al.18 also reported that addition of water vapor (10 -10
-4
-3
Torr, RH < 1%) did not have an effect on γ. Our study was
conducted at 250 and 260 K and RH ) 0.5-12.2%; we found
γw has a tendency to be slightly lower at higher T. γw ()0.029
( 0.04 at RH ) 0.53%) is within the ranges of γ. Kirchner et
-
3
al. reported a γ value of 3.3 × 10 on 30% NaBr frozen
-
59
suggesting active surface sites are regenerated as OH (ad) is
solution at 230-240 K. We may assume the surface of frozen
converted to the deliquescence-like state on the surface. The
role of water or RH is that it increases the ion mobility of NaOH
over the surface and provides a lower energy pathway for NaOH
to form a stable deliquescence-like state. The weak dependence
of γw on RH also suggests that water may not play a direct role
in the rate-limiting step (reaction 14).
solutions is smooth; γt is used to compare with the result of
-
3
Kirchner et al. With γt ) 1.8 × 10 at our highest RH ) 12.2
and 250 K, it is numerically in good agreement with the result
of Kirchner et al. despite the differences in the nature of the
surface.
Atmospheric Implications. Sea-salt particles exist in the
marine boundary layer and in the lower stratosphere after
volcanic eruption. Salt particles most often occur as deliquescent
aerosols in the marine boundary layer. However, some of these
particles may be transported to a dry atmosphere where they
are found as solid crystalline salt particles. The purpose of
this study was to examine the heterogeneous activation of Br
and Cl species and the role of adsorbed water on NaX. The
critical question is whether heterogeneous activation of HOBr
via the reactions of HOBr + NaCl and HOBr + NaBr is
competitive with photolysis of HOBr to active Br (and Cl)
species.
Comparison with Previous Studies. The uptake of HOBr
on NaCl and KBr was measured in the Knudsen reactor at very
low humidity conditions <1% and on deliquescent NaCl
aerosols (75% RH) and in frozen and aqueous salt solu-
6
0
1
8,19,21,59
tions.
Although we may compare some of our results
with these studies, a direct comparison for γ as a function of
RH is not feasible.
Mochida et al. reported that the initial uptake coefficient of
-
4
HOBr on powder and grain NaCl was from 65 to 6.5 × 10 at
1
3
14
9
10
3
2
× 10 to 6 × 10 molecules/s (∼10 -10 molecules/cm )
18
at room temperature. This initial uptake coefficient was
corrected to be equivalent to the initial uptake coefficient
obtained from a spray-deposited thin NaCl particle film, which
was assumed to be no internal surface areas. Spray-deposited
NaCl substrates could not be used for measuring the uptake
coefficient of HOBr because the HOBr signal change before
and after the reaction was too small. This initial uptake
-3
The photolysis rate constant J for HOBr f OH + Br is ∼10
s- at ground level. The photochemical lifetime of HOBr is
approximately 1 h. The photochemical lifetimes of Br2 and BrCl
are 1 and 3 min, respectively, on the basis of photolysis rates
1
61
-
2
-3 -1 15,62
of Br2, J ) 2 × 10 , and BrCl, J ) 6 × 10 s .
The
importance of reactions 2 and 3 is that two halogen atoms are
released into the gas phase per HOBr consumed, thus increasing
the halogen atom concentration in the troposphere by converting
photochemical inactive sea salt and temporally reservoir HOBr.
Heterogeneous loss lifetime tτ of HOBr on NaX can be estimated
as
-
4
coefficient is estimated to be ∼10 under our experimental
conditions. At RH ) 1.5-1.9%, we obtained γw ) 1.5-2.0 ×
-
3
1
0
at 250 K (Figure 4), and the true uptake coefficient γt is
-
5
6
.3 × 10 . γt is estimated to be a factor of ∼10 lower than the
initial uptake coefficient reported by Mochida et al. The NaCl
substrates used by Mochida et al. were ground in a ball mill,
with average diameters between 10 and 100 µm or 350-500
µm. Similar to our treatment, the internal surface areas were
corrected using the pore-diffusion model developed by Keyser
t ) 4/(ωγ A )
(21)
τ
t c
2
3
where Ac is a condensed surface-to-air volume ratio (cm /cm ).
-
6
2
When we assume a sea-salt aerosol loading of 4 × 10 cm /
4
1,42
3
3
et al.
However, surface reactivity of ground salt samples
cm , corresponding to approximately 10 particles/cm with a
-
4
63
may be different from that of aerosol-deposited substrates in
terms of surface defects and steps. Our sample involved
recrytallization from liquid aerosols to solid particles on the
detachable glass tube surface at 150 °C. The recrystallization
may create internal surfaces (Figure 2), but the annealing at
high temperature may reduce the defects or steps on the granule
surface. This may explain why the uptake coefficient γt is lower
than the one reported by Mochida et al.
particle radius of 10 cm, lifetimes of ∼4.5 h and ∼9 days
are estimated for HOBr on the NaBr and NaCl aerosol surface
-
3
-5
using γt ) 2.5 × 10 and 5 × 10 , respectively. There are
other factors such as morphology of aerosols that affect the
lifetime calculation, but they are omitted in this estimation. The
lifetime calculation suggests that bromine activation via reaction
2 is probably competitive with photolysis of HOBr in the
troposphere, whereas reaction 3 is approximately 2 orders of
magnitude slower than photolysis under dry sea-salt aerosol-
loading conditions. However, with a higher γ value on an acidic
deliquescent NaCl aerosol, the heterogeneous activation of
HOBr over NaCl aerosols would be favored over photolysis in
Br and Cl activation. On the other hand, we have shown that
sea-salt surfaces can be deactivated slowly when it reacts with
HOBr; if all active sea-salt aerosol surfaces were consumed,
heterogeneous reactions 2 and 3 play little role in halogen
activation. A surface regeneration process would be required
for nondeliquescent aerosols to play a major role in the
atmosphere. The surface regeneration can be achieved by
adsorbed water as has been discussed previously or by continued
supplies of aerosols in the boundary layer. Future studies are
required to assess surface regeneration processes.
Abbatt and Waschewsky reported an upper limit to the uptake
coefficient of HOBr on unbuffered deliquescent NaCl aerosols
2
1
-
3
12
13
(
75% RH) to be 1.5 × 10 at 298 K and [HOBr] ) 10 -10
3 21
molecules/cm . In acidified aerosols, γ was 0.5. As their
experimental conditions were very different from those of the
current study in terms of both RH (liquid versus solid) and
[HOBr], a quantitative comparison cannot be made. However,
γ measured in unbuffered aerosols is comparable with γw and
it is higher than γt. Combined with the results of our study, this
suggests that the effect of RH (1-75%) on the initial uptake
coefficient is not very large. The formation of BrCl was
observed both on NaCl powder surfaces and in aqueous salt
solutions.18,19
There is no reported study on the uptake of HOBr on NaBr
surfaces. Kirchner et al. studied the HOBr uptake in frozen NaBr
solutions.59 The closest system to compare is the HOBr uptake
Conclusions
1
8
-2
on KBr spray-deposited surfaces. The γ was 2-6 × 10 at
We have for the first time determined the uptake of HOBr
on NaCl and NaBr particle films as a function of RH. The initial
[
HOBr] ) 1010 molecules/cm and room temperature. Mochida
3

5688 J. Phys. Chem. B, Vol. 106, No. 22, 2002
Chu et al.
uptake coefficient γw of HOBr on NaCl surfaces decreases
(19) Fickert, S.; Adams, J. W.; Crowley, J. N. J. Geophys. Res. 1999,
104, 23719.
-
3
-4
slightly from 2.0 ( 0.3 × 10 to 8.1 ( 1.3 × 10 as RH
increases from 1.5 to 18.2% at 250 K. γw increases sharply as
the NaCl film thickness increases from 1 to 50 µm and then
levels out gradually. The γw value of HOBr on NaBr showed a
weak decreasing tendency from 0.029 ( 0.004 to 0.021 ( 0.004
at 250 K as RH increases from 0.53 to 12.2%. The tendency is
small, but it is outside the error of the measurement. The granule
size distribution of the NaCl particle film was studied using
SEM, and the tortuosity factor, τ, was 2.18. The true uptake
coefficients γt were calculated using the pore-diffusion model.
(
(
20) Gebel, M.; Finlayson-Pitts, B. J. J. Phys. Chem. A 2001, 105, 5178.
21) Abbatt, J. P. D.; Waschewsky, G. C. G. J. Phys. Chem. A 1998,
102, 3719.
(22) Vogt, R.; Crutzen, P. J.; Sander, R. Nature 1996, 383, 327.
(23) Mozurkewich, M. J. Geophys. Res. 1995, 100, 14199.
(24) Beichert, P.; Finlayson-Pitts, B. J. J. Phys. Chem. 1996, 100, 15218.
(25) Peters, S. J.; Ewing, G. E. Langmuir 1997, 26, 6345.
(26) Foster, M. C.; Ewing, G. E. J. Chem. Phys. 1999, 26, 102.
(27) Foster, M. C.; Ewing, G. E. J. Chem. Phys. 2000, 112, 6817.
(
(
28) Hu, J. H.; Abbatt, J. P. D. J. Phys. Chem. A 1997, 101, 871.
29) Hallquist, M.; Stewart, D. J.; Baker, J.; Cox, R. A. J. Phys. Chem.
A 2000, 104, 3984.
30) Kane, S. M.; Caloz, F.; Leu, M.-T. J. Phys. Chem. A 2001, 105,
5465.
(31) Behnke, W.; Scheer, V.; Zetzsch, C. J. Aerosol Sci. 1993, 24, S115.
32) George, Ch.; Ponche, J. L.; Mirabel, Ph.; Behnke, W.; Scheer, V.;
Zetzsch, C. J. Phys. Chem. 1994, 98, 8780.
33) Seinfeld, J. H.; Pandis, S. N. Atmospheric Chemistry and Physics;
-
5
γt of HOBr on NaCl also showed a decrease from 6.6 × 10
(
-
5
to 2.6 × 10 as RH increases from 1.5 to 18.2% at 250 K. A
-3
-
3
weak decreasing tendency in γt from 3.1 × 10 to 1.6 × 10
(
at 250 K was observed for HOBr on NaBr as RH increases
from 0.53 to 12.2%. The mechanism for HOBr + NaCl f BrCl
(
+
NaOH and HOBr + NaBr f Br2 + NaOH involves a step
Wiley: New York, 1998; chapter 9.
(34) Chu, L. T.; Heron, J. W. Geophys. Res. Lett. 1995, 22, 3211.
-
for HOBr interaction with X on the surface. We have shown
that the reactions occur predominantly on a NaX surface where
no water is adsorbed at the condition of low RH. Adsorbed water
could enhance the ion mobility of NaOH on salt surfaces, and
the weak dependence of γw on RH suggests water may not play
a direct role in the rate-limiting step (reaction 14). The
heterogeneous reaction rate of HOBr + NaBr f Br2 + NaOH
is competitive with HOBr photolysis rate at the boundary layer.
(35) Chu, L.; Chu, L. T. J. Phys. Chem. A. 1999, 103, 691.
(36) Chu, L.; Chu, L. T. J. Phys. Chem. A 1999, 103, 8640.
(37) Dean, J. A. Lange’s Handbook of Chemistry, 15th ed.; Wiley: New
York, 1999; p 5.26 and 3.49.
(38) Chu, L.; Chu, L. T. Recent Res. DeVel. Geophys. 2000, 3, 141.
(
(
39) Brown, R. L. J. Phys. Natl. Bur. Stand (U. S.). 1978, 83, 1.
40) Perry’s Chemical Engineering Handbook; Perry, R. H., Green, D.
W., Eds.; McGraw-Hill: New York, 1997; p 5.50.
(41) Keyser, L. F.; Moore, S. B.; Leu, M.-T. J. Phys. Chem. 1991, 95,
5
2
496.
(42) Keyser, L. F.; Leu, M.-T.; Moore, S. B. J. Phys. Chem. 1993, 97,
Acknowledgment. The authors acknowledge discussions
with Professors Barbara Finlayson-Pitts, George Ewing, and Scot
Martin. We also thank Dr. W. Samsonoff for assistance in
obtaining SEM micrographs of salt samples at the Wadsworth
Center’s Electron Microscopic Core Facility and Dr. Oliver
Rattigan and two anonymous reviewers for helpful suggestions
and comments on the manuscript. This work was supported by
the Atmospheric Chemistry Program, National Science Founda-
tion.
800.
(43) Underwood, G. M.; Li, P.; Usher, C. R.; Grassian, V. H. J. Phys.
Chem. A 2000, 104, 819.
(44) Barraclough, P. B.; Hall, P. G. Surf. Sci. 1974, 46, 393.
(
45) Chikazawa, M.; Kaiho, M.; Kanazawa, T. Nippon Kagaku Kaishi
1
976, 3, 410.
(
(
46) Peters, S. J.; Ewing, G. E. J. Phys. Chem. B 1997, 101, 10880.
47) Engkvist, O.; Stone, A. J. J. Chem. Phys. 2000, 112, 6827.
(48) Luna, M.; Rieutord, F.; Melman, N. A.; Dai, Q.; Salmeron, M. J.
Phys. Chem. A 1998, 102, 6793.
49) Romakkaniemi, S.; H a¨ meri, K.; V a¨ kev a¨ , M.; Laaksonen, A. J. Phys.
Chem. A 2001, 105, 8183.
(
References and Notes
(
(
50) Cziczo, D. J.; Abbatt, J. P. D. J. Phys. Chem. A 2000, 104, 2038.
51) International Critical Table; Washburn, E. W., Ed.; McGraw-
(
1) Barrie, L. A.; Bottenheim, J. W.; Schnell, R. C.; Crutzen, P. J.;
Rasmussen, R. A. Nature 1988, 334, 138.
2) McConnell, J. C.; Henderson, G. S.; Barrie, L. A.; Bottenheim, J.;
Niki, H.; Langford, C. H.; Templeton, E. M. J. Nature 1992, 355, 150.
Hill: New York, 1926; Vol. 1, p 63; Vol. 2, p 327.
52) Koop, T.; Kapilashrami, A.; Molina, L. T.; Molina, M. J. J.
Geophys. Res. 2000, 105, 26393.
(
(
(
53) Martin, S. T. Chem. ReV. 2000, 100, 3409.
(
(
(
3) Fan, S. M.; Jacob, D. J. Nature 1992, 359, 522.
4) Tang, T.; McConnell, J. C. Geophys. Res. Lett. 1996, 23, 2633.
5) Impey, G. A.; Shepson, P. B.; Hastie, D. R.; Barrie, L. A.; Anlauf,
(54) Allen, H. C.; Laux, J. M.; Vogt, R.; Finlayson-Pitts, B. J.;
Hemminger, J. C. J. Phys. Chem. 1996, 100, 6371.
55) Laux, J. M.; Fister, T. F.; Finlayson-Pitts, B. J.; Hemminger, J. C.
(
K. G. J. Geophys. Res. 1997, 102, 16005.
6) Hebestreit, K.; Stutz, J.; Rosen, D.; Matveev, V.; Peleg, M.; Luria,
M.; Platt, U. Science 1999, 283, 55.
7) Finlayson-Pitts, B. J.; Livingston, F. E.; Berko, H. N. Nature 1990,
43, 622.
J. Phys. Chem. 1996, 100, 19891.
(
(56) Rudzinski, W.; Everett, D. H. Adsorption of Gases on Heteroge-
neous Surfaces; Academic Press: San Diego, CA, 1992; pp 51-53.
(57) Cotton, F. A.; Wilkinson, G. AdVanced Inorganic Chemistry, 4th
ed.; Wiley: New York, 1980; pp 558.
(58) M o¨ ssinger, J. C.; Cox. R. A. J. Phys. Chem. A 2001, 105, 5165.
(59) Kirchner, U.; Benter, Th.; Schindler, R. N. Ber. Bunsen-Ges. Phys.
Chem. 1997, 101, 975.
(
3
(
8) Behnke, W.; Kr u¨ ger, H.-U.; Scheer, V.; Zetzsch, C. J. Aerosol Sci.
1
991, 22, S609.
(
9) Vogt, R.; Finlayson-Pitts, B. J. J. Phys. Chem. 1994, 98, 3747.
(
10) Fenter, F. F.; Caloz, F.; Rossi, M. J. J. Phys. Chem. 1996, 100,
1
008.
(60) Shaw, G. E. J. Geophys. Res. 1991, 96, 22369.
(61) Rattigan, O. V.; Lary, D. J.; Jones, R. J.; Cox, R. A. J. Geophys.
Res. 1996, 101, 23021.
(
(
(
11) Davis, J. A.; Cox, R. A. J. Phys. Chem. A 1998, 102, 7631.
12) Ghosal, S.; Hemminger, J. C. J. Phys. Chem. A 1999, 103, 4777.
13) Gershenzon, M. Y.; Il’In, S.; Fedotov, N. G.; Gershenzon, Y. M.
(62) Impey, G. A.; Mihele, C. M.; Anlauf, K. G.; Barrie, L. A.; Hastie,
J. Atmos. Chem. 1999, 34, 119.
D. R.; Shepson, P. B. J. Atmos. Chem. 1999, 34, 21.
-
3
(14) Weis, D. D.; Ewing, G. E. J. Phys. Chem. A 1999, 103, 4865.
(15) Sander, R.; Crutzen, P. J. J. Geophys. Res. 1996, 101, 9121.
(16) Timonen, R. S.; Chu, L. T.; Leu, M.-T.; Keyser, L. F. J. Phys.
(63) Using marine aerosol concentration of ∼2000 cm with the size
of a micron (Kim, Y.; Sievering, H.; Boatman, J.; Wellman, D.; Pszenny,
-
-
A. J. Geophys. Res. 1995, 100, 23027) and the ratio of [Cl ]/[Br ] near
aerosol surfaces of ∼188 (Foster, K. L. et al. Science 2001, 291, 471), the
Chem. 1994, 98, 9509.
17) Caloz, F.; Fenter, F. F.; Rossi, M. J. J. Phys. Chem. 1996, 100,
494.
18) Mochida, M.; Akimoto, H.; van den Bergh, H.; Rossi, M. J. J. Phys.
Chem. A 1998, 102, 4819.
-
3
(
bromine aerosol density is approximately equivalent to 10 cm . This
-
6
2
3
7
corresponds to bromine sea-salt aerosol loading of 10 cm /cm or lower
6
2
3
6
(
and chlorine sea-salt aerosol loading >10 cm /cm . A value of 4 × 10
2 3
cm /cm was used to estimate the lifetime.