Heterogeneous chemistry of Cl2O and HOCl on frozen natural sea salt, recrystallized sea salt, KCl and NaCl solutions at 200 and 215 K

Z. Phys. Chem. 224 (2010) 1119–1150 . DOI 10.1524.zpch.2010.6144

by Oldenbourg Wissenschaftsverlag, München

©

Heterogeneous Chemistry of Cl O and HOCl on

2

Frozen Natural Sea Salt, Recrystallized Sea

Salt, KCl and NaCl Solutions at 200 and 215 K

By Pascal Pratte and Michel J. Rossi

*

Ecole Polytechnique Fédérale de Lausanne (EPFL), Laboratoire de Pollution Atmosphérique

et sol (LPAS), Station 6, CH-1015 Lausanne, Switzerland

th

Dedicated to Prof. Dr. Reinhard Zellner on the occasion of his 65 birthday

(

Received December 11, 2009; accepted March 21, 2010)

Heterogeneous Chemistry . Sea Salt . Cl O . HOCl

2

The HOCl heterogeneous reaction on frozen natural (NSS) and recrystallized (RSS) sea salt,

KCl and NaCl solutions was studied using a low pressure flow reactor in order to measure the

uptake coefficient γ and products of reaction. The HOCl sample used in these experiments

2

always contained up to 25% Cl O which was also studied separately as a pure gas in order to

understand the heterogeneous chemistry of both gases. By performing HOCl uptake on frozen

NSS solution at 200 K and a gas-phase residence time of (1.6±0.6) s we obtained a steady state

K3

uptake coefficient γHOCl on NSS = (2.5±0.ꢀ)!10 and γCl2O on NSS = (2.8±0.8)!10 . On frozen

K3

KCl solution at 200 K we obtain γHOCl on KCl = (2.8±1.3)!10 , identical to NSS, and γCl2O on

KCl = (4.6±0.8)!10 . The main product formed during the uptake on frozen NSS solution is

K4

2 2

Cl which is sustained for at least one hour. In contrast, only a transient Cl flow (pulse)

decreasing on the time scale of 100 s was observed on frozen KCl (NaCl) solution. 25±10 %

of the HOCl taken up on all chloride-containing frozen substrates at 200 K react to produce

1

K3

Cl

2

at high HOCl concentration (4.5!10 molecule cm ) and at a residence time of 1.6 s in

O. For smaller concentrations such as [HOCl] =

.ꢀ!10 molecule cm and.or a shorter residence time (0.13ꢀ±0.004s), HOCl uptake did

not generate Cl in contrast to Cl O. A single Br burst event was monitored when a Cl O or

HOCl.Cl O mixture is taken up on fresh frozen NSS solution during the first uptake at 200 K.

Further Cl O or HOCl.Cl O uptake on the same sample, even after annealing at 240 K does

not show an additional Br pulse. This Br release may be significant in the autocatalytic

comparison with twice that for Cl

3

2

1

0

K3

2

ozone destruction mechanism in the troposphere during polar sunrise. Some of the atmospheric

implications of the present results are highlighted with emphasis on the preequilibrium

2 2 2

Cl O(ads) + H O(ice) $ 2 HOCl(ads) between adsorbed HOCl and Cl O, with the latter being

the gateway to reactive uptake of HOCl at low temperatures.

*

Present address: Paul Scherrer Institut (PSI), Laboratorium für Atmosphärenchemie

(LAC), OFLA008, CH-5232 Villigen PSI, Switzerland; E-mail: michel.rossi@epfl.ch.

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

1

120

P. Pratte and M. J. Rossi

1. Introduction

Essentially complete ozone loss is frequently observed in the boundary layer of

the Arctic in the springtime [1]. Field measurements revealed that the ozone

concentration decreases by approximately a factor of ten in a few hours [1]

which seems to be correlated to a local maximum of [BrO] during sunrise [2–

4]. This event has been called polar sunrise chemistry and has been associated

with the presence of bromine containing species collected on filters. This sug-

gests that bromine-containing species such as Br, HBr and HOBr play a leading

role in the mechanisms of O destruction. Moreover, there is experimental evi-

3

dence from the field that chlorine atom concentrations are also elevated during

these ozone-depletion periods owing to the observed chlorination of naturally

occurring hydrocarbons that is consistent with the presence of an important ma-

rine halogen source [5]. Recent experimental results have confirmed the presence

of Br and Cl precursor concentrations on the tens to hundreds of ppt level such

as Br [6,ꢀ], BrCl [6], HOBr, Cl and HOCl [ꢀ]. Therefore, the search for an

2

explanation of the underpinning chemical mechanism has focussed on catalytic

cycles of Br which is well known from studies of global and polar stratospheric

ozone depletion and is displayed in reactions (1) to (3):

Br + hν/2Br

(1)

(2)

(3)

2

Br + O /BrO + O₂

3

2

BrO/Br + O₂

2

Owing to the extremely short time scale of ozone destruction in polar sunrise

chemistry the search for a mechanism has focussed on autocatalytic processes

that include a degenerate chain branching step and are similar to combustion or

explosion of hydrocarbon fuels, hence the alternative term “bromine explosion”

for the description of polar sunrise chemistry. An autocatalytic process is one

where one or more of the resulting reaction products are catalysts for the reaction

under study in contrast to a case where a catalyst is added externally. A number

of possible chemical reactions listed in reactions (4) to (ꢀ) including photochem-

istry involving aqueous sea salt (marine aerosol) [8,9] or frozen sea salt solution

[10] have been proposed in order to explain the extremely short time scale of

ozone loss, the chlorination of hydrocarbons and the postulated high halogen

free radical concentration in the polar region which can attain values of up to

5

•

K3

1

0 Cl cm [11,12].

K

HOBr + Cl /BrCl +OH

(4)

(5)

(6)

K

HOBr + Br /Br + OH

2

K

HOCl + Br /BrCl + OH

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

Heterogeneous Chemistry of Cl

2

O and HOCl

1121

•

BrCl + hν/Br + Cl

(ꢀ)

During polar sunrise, a significant part of Br , Cl and BrCl may be generated

2

by heterogeneous chemistry of HOCl and HOBr which are mainly formed from

the homogeneous gas phase reaction of OH and XO (X = Cl, Br) radicals, reac-

tion (8):

•

HO + XO /HOX + O

(8)

2

The formation of HOX (X = Cl, Br) compounds contributes therefore to-

wards a temporary reservoir of halogen compounds activated by their heteroge-

neous reactions on sea salt aerosol and.or frozen sea salt according to reactions

(4) and (5) [11–13].

The autocatalytic nature stems from the fact that one Br atom, once reacted

to HOBr following reaction (8), is able to volatilise marine bromide from sea

spray aerosols or from frozen sea salt following reaction (5) and thereby form

two potentially active Br atoms after photolysis of the precursor Br . It is the

2

catalyst and needs to be present in only minute concentrations. This amounts to

degenerate chain branching afforded by a heterogeneous reaction and continues

until the BrO concentration is sufficiently high such that the catalytic cycle (reac-

tions (1) to (3)) will take over. The rate determining step of the autocatalytic

mechanism is the formation of HOBr in reaction (8) and is superseded by reac-

tion (3) once the reaction has switched from autocatalytic to catalytic. The mech-

anistic investigations have mainly focussed on the role of Br rather than Cl

because of faster heterogeneous reaction rates of HOBr [14]. In addition, there

are potential complications with HOCl because it is unstable with respect to its

K

anhydride Cl O which is also reactive with marine halides X as shown below.

2

Sea salt deposited on the polar snow pack may contribute to the conversion

of halogen reservoir molecules to reactive species by heterogeneous processes

that deplete the ice.salt surface. This process competes with similar heterogene-

ous reactions that involve deliquescent sea salt aerosols. Only when the heteroge-

neous reactions (4) and (5) of HOBr on ice and snow are included does the

simulation of tropospheric ozone destruction agree with field measurements

through halogen activation [15]. However, only few data are available for HOCl

heterogeneous reactions on marine aerosol and sea ice in comparison to uptake

experiments of HOBr or HOCl performed on solid substrates [16–20].

It was suggested that HOCl could be an important temporary reservoir for

chlorine in the marine boundary layer because it reacts very slowly with ozone

II

K16

3

K1 K1

(

k < 4.0!10

cm molecule

s

) [21] and does not react with hydrocar-

bons. Despite the fact that HOCl has never been measured in the dark winter

season in the polar boundary layer, it may in principle be formed by radical

recombination (8) [22]. The absorption coefficient of HOCl is given in the litera-

ture [23,24] within the critical 300–380 nm wavelength region and leads to a

photolysis lifetime of less than one hour using the measured photodissociation

rate constant at the surface of the earth. This short lifetime suggests that HOCl

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

1

122

P. Pratte and M. J. Rossi

is an ineffective chlorine reservoir under solar radiation but in turn could be

effective in the arctic region during the winter season in the absence of actinic

radiation.

Huff and Abbatt performed uptake experiments of HOCl(g) in a laminar low

pressure flow reactor in which the walls were coated with a frozen NaCl solution

12

K3

at 233 K [25]. They used [HOCl] = 3!10 cm at a total pressure of He

between 1 and 2 Torr and concluded that HOCl is unreactive on frozen sea salt.

We took this work as a starting point of the present contribution towards the

knowledge of heterogeneous reactions of HOCl which has not been as thor-

oughly investigated as HOBr owing to the fact that the role of HOCl in polar

sunrise chemistry is less clear than for HOBr. In addition, the equilibrium be-

tween HOCl and Cl O, the latter of which is currently of no atmospheric interest,

2

and the anticipated reactivity of both components on liquid and frozen halide

solutions make this a chemical system that we wished to investigate further. The

observational evidence for HOCl is fairly weak for the troposphere, however, it

has been unambiguously identified in the middle and lower stratosphere in satel-

lite IR and FIR spectrometry [26]. Recently, the rate constant of reaction (8) has

come under debate in the interest of obtaining a satisfactory fit to HOCl profiles

obtained by satellite measurements [26].

We present here kinetic results of HOCl uptake mainly on frozen KCl and

natural sea salt (NSS) solutions at 200 and 215 K that are in the range of relevant

temperatures during the polar winter in the MBL. In addition, the products of

reaction released into the gas-phase were monitored in order to quantify the rate

at which gas-phase HOCl is converted to Cl , the precursor for chlorination of

2

naturally occurring hydrocarbons. We also focussed on differences between

HOCl uptake coefficients measured on frozen KCl, recrystallised sea salt (RSS)

and natural sea salt (NSS) solutions as well as products formed. In addition, we

tested the role of acidity for the formation of products both on NSS and KCl

frozen solution.

2. Experimental setup

Experiments were performed in a Teflon-coated low pressure flow reactor that

is suited for the measurement of heterogeneous process. A detailed description

of this technique has been given elsewhere [28]. Briefly, the Knudsen flow reac-

tor is mounted on a differentially pumped high vacuum chamber equipped with

a cryogenic and a turbomolecular pump. A given number of molecules are in-

jected into the reactor by a glass capillary inlet before effusing out of the escape

K1

orifice after a lifetime τ (1.k (s ) with k_e_s_cbeing the measured escape rate

esc

K1

constant .s ) to form a modulated (thermal) molecular beam by a rotating

chopper located in the upper chamber. The quadrupole mass spectrometer signal

is then processed by a digital lock-in amplifier in order to separate the contribu-

tion of the molecular beam from the background gas phase.

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

Heterogeneous Chemistry of Cl

2

O and HOCl

1123

Table 1. Escape rate constants kesc of HOCl, Cl

2

and Cl

2

O measured at 300 K and typical gas

concentrations [G] for the 4 and the 14 mm diameter escape aperture.

k

esc (±0.2 s^K¹) . [G]!10¹¹molecule cm^K³

Cl Cl

Escape aper-

ture (mm)

HOCl

2

O

4

0.6 . 4.5

0.5 .

6.3 .

0.4 . 2.ꢀ

1

4

ꢀ.3 . 0.3ꢀ

5.ꢀ . 0.19

In order to measure the uptake coefficient γ(HOCl,Cl O) on the frozen sub-

2

strates we quantitatively measured the HOCl and the Cl O molecules that have

2

been taken up as a function of time. To this end, steady state experiments were

performed in the low pressure flow reactor using concentrations ranging from

10

12

K3

K1

1

0

to 10 molecule cm . An initial flow rate F (molecule s ) is admitted

in

into the reactor and after a short transition time, the effusive flow rate F_o_u_tequals

the incoming flow rate F_i_nif no chemical loss is admitted, that is when the

sample compartment is isolated from the inflow. The residence time τ and the

gas concentration may be varied by changing either the flow or the size of the

escape orifice. When the sample compartment is open to the flow, part of the

gas is lost on the substrate resulting in a competition between the escape rate

(k_e_s_cN, with N being the total number of gas molecules in the reactor) and the

rate of loss (k_l_o_s_sN). At steady state we obtain the following balance:

dN

dt

=

F KNk KNk = 0

(9)

in

esc

loss

Rearrangement of equation (9) leads to equation (10) which relates the ratio

F_i_n.F_o_u_tto the escape rate constant k_e_s_cand results in the heterogeneous loss rate

constant k_l_o_s_s

:

F_i_n

^Fout

k_l_o_s_s= (

K1)k_e_s_c

(10)

Commonly, we used a normalised form of k_l_o_s_s, namely the uptake coefficient

γ that expresses the probability that a molecule colliding with a surface is cap-

tured. We defined γ simply by the first order loss rate constant k divided by

loss

K1

the frequency of collision ω in s . In addition, k_e_s_cwas measured by monitoring

the exponential decay rate of the MS signal by halting F_i_nin the absence of

substrate. The escape rates k_e_s_cwere measured for the different escape apertures,

namely for the 4 and 14 mm diameter orifices (Table 1). The concentration of a

gas G was calculated by using the initial flow rate F_i_nand the measured k_e_s_c

using the expression [G] = F .Vk with values summarized in Table 1. HOCl,

in

esc

Cl O, Cl and Br were monitored at m.e 52, 51, ꢀ0 and 160, respectively.

2

2.1 Sample preparation

Three types of salt samples were used, namely KCl (NaCl) (Fluka, purity >

99.5%), recrystallized sea salt RSS (Hawaiian Salt, Honolulu, PA AKAI INC.)

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

1

124

P. Pratte and M. J. Rossi

and NSS (sel de Guérande, DIEPAL-NSA) which is a natural non-processed sea

salt obtained by pan evaporation (open air). Solutions of KCl (NaCl), NSS and

RSS were made from the dissolution of salts in bidistilled water (35g.L) and

subsequent filtration in order to mimic the salt concentration of sea water [29].

The salt solutions were degassed in a vacuum line by performing two freeze-

pump-thaw cycles before pouring the salt solution, typically 5 ml, into the gold

coated copper cryogenic support that is described elsewhere [30]. Subsequently,

K1

the temperature of the salt solution was decreased at a rate of 0.2 K s down

to 200 or 215 K in order to produce frozen sea salt [2ꢀ]. All frozen solutions

were cooled down following the same procedure. In fact, different ice prepara-

tion methods led to different uptake kinetics in the past [31]. It is known that

the cooling rate affects the halide distribution in the bulk and on the surface of

ice [32]. If the freezing rate is slow bulk ice will concentrate most of the salt on

the substrate surface by segregation processes. However, in an atmospheric con-

text sea water freezes too quickly to allow segregation of halide from the bulk.

In some cases, KCl (NaCl) solutions were acidified using a commercial buffered

solution containing 0.068 M NaOH, 0.056 M citric acid and 0.044 M NaCl to

buffer the frozen salt solution at pH = 4 in order to study the effect of acidity

on the rate of Cl formation.

2

.2 HOCl and Cl O synthesis

2

Cl was admitted to a vessel containing HgO powder (Fluka, puriss > 99.0%)

2

held at ꢀꢀ K which was kept for 24 hours at 190 K to allow the following

reaction to occur:

HgO(s)+2Cl (g)/Cl O(g)+HgCl (s)

(11)

2

Cl O, absorbing at λ < ꢀ00 nm, was transferred to a trap containing approxi-

2

mately 2 ml of frozen bidistilled water at ꢀꢀ K that was subsequently warmed

in a water bath to ambient temperature. The cherry-colored Cl O liquid reacts

2

with liquid water to produce HOCl(l) upon Cl O hydrolysis according to reaction

2

(12).

Cl O(l) + H O(l)$2 HOCl(l)

(12)

2

The HOCl sample was kept at 190 K as a stock solution for further use.

2.3 Calibration procedure

Because HOCl is in equilibrium (12) with Cl O, increasing concentrations of

2

HOCl will shift the equilibrium towards Cl O under static conditions. In order

2

to limit the formation of Cl O, we pumped on the sample in order to lower the

2

total pressure. Even under those non-equilibrium conditions 25±10 % of Cl O

2

remained in the HOCl sample. At these experimental conditions the flow rate

measurement of HOCl into the flow reactor (F (HOCl)) could not be performed

in

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

Heterogeneous Chemistry of Cl

2

O and HOCl

1125

by observing the pressure drop in the calibrated volume because the pressure in

the vacuum line is under 50 mTorr which is too small for a reliable pressure

measurement using an absolute Baratron gauge (MKS Baratron 122AA-

0

0010AB). In order to calibrate the MS signal for HOCl, a titration was per-

formed on a vapor-deposited ice substrate. In contrast, the impurity Cl O has

2

been calibrated in the usual way by measuring its pressure drop in a calibrated

volume as described in the literature [31]. In order to enable the titration of

HOCl, Bulk (B) ice was prepared at 200 K in the cryogenic support device [30].

Subsequently, we deposited an amount of HCl on B ice in excess over HOCl

before isolating the sample by sealing off the sample compartment. The HOCl

source was held at 19ꢀ±1 K under differential pumping conditions by using a

roughing pump protected by a cold trap. A steady-state flow of HOCl was estab-

lished into the flow reactor and was allowed to react with the HCl.ice sample.

The gas phase HOCl and the Cl O impurity are titrated by excess adsorbed HCl

2

following reactions (13) and (14):

Cl O(g) + HCl(ads)/Cl (g) + HOCl(ads)

(13)

(14)

2

HOCl(g) + HCl(ads)/H O(ads) + Cl (g)

2

Cl O(g) + 2HCl(ads)/2Cl (g) + H O(ads)

2

The same titration procedure was performed for pure Cl O. This leads to the

2

conclusion that for each Cl O taken up, two Cl were released to the gas phase.

2

In addition, we believe that HOCl(ads) generated in reaction (13) is converted

to Cl (g) in reaction (14). By using these reactions for the titration of the HOCl.

2

Cl O mixture, the calibration constant for HOCl is given by equation (15).

2

(

Cl

2

)

(Cl

2

)

2 2

(Cl O) (Cl O)

(

HOCl)

K

cal

I

K2Kcal

I

K_c_a_l

=

(15)

(

HOCl)

I

where I⁽^X⁾and K_c_a⁽_l^X⁾are the measured MS signal and the calibration constant

for the molecule X (X = Cl , Cl O, HOCl), respectively. The absolute calibration

2

for the Cl mass signal was obtained from authentic samples of pure Cl .

2

3

. Results and discussion on Cl O and HOCl uptake

2

3

.1 Uptake kinetics of HOCl and pure Cl O on bulk ice at 200 K

2

The HOCl sample used in the present work inevitably contains up to 25% Cl O

2

following equation (12). It is likely that the Cl O “impurity” in HOCl does

2

not correspond to the equilibrium fraction because the mixture was not let to

equilibration under the used experimental conditions. Consequently, we need to

take into account the presence of Cl O by performing uptake experiments of pure

2

Cl O on all studied substrates as a reference in order to provide a comparison for

2

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

1

126

P. Pratte and M. J. Rossi

the uptake experiment of the HOCl.Cl O mixture. Before performing uptake

2

experiments on the frozen salt solution, we have studied the uptake of pure Cl O

2

and the HOCl.Cl O mixture on bulk (B) ice from frozen liquid water [31].

2

The uptake coefficient γ for the HOCl.Cl O mixture on bulk ice is

2

K4

γ_s_s(HOCl) = (5.5±0.4)!10 and γ (Cl O) = (5.3±0.6)!10 , respectively. For

uptake of pure Cl O, we obtain γ (Cl O) = (2.8±0.ꢀ)!10 , lower by a factor

ss

2

K4

2

ss

2

of two compared to the mixture. These results directly follow from displays in

Figures 9 and 10 (see Appendix). These uptake experiments show that γ_s_sfor

each component of the HOCl.Cl O mixture is similar while pure Cl O uptake

2

results in a γ_s_svalue smaller by a factor of two compared to the simultaneous

measurement of γ (Cl O) and γ (HOCl). It seems that Cl O uptake on pure B

ss

2

ss

2

ice is accelerated by HOCl, perhaps by hydrogen bonding owing to the presence

of adsorbed HOCl.

In addition, we note (Figure 9) that no product was formed for the HOCl.

Cl O mixture at 200 K. Conversely, in the case of pure Cl O uptake, slow HOCl

2

formation and Cl O decomposition to Cl occurred (Figure 10) as displayed in

2

equations (16) and (1ꢀ):

Cl O/Cl + ½ O

2

(heterogeneous decomposition)

(heterogeneous hydrolysis)

(16)

(1ꢀ)

2

H O + Cl O/2HOCl

2

This shows that HOCl and Cl O are not rapidly converted to Cl on pure B

2

ice at 200 K. These reference experiments performed on Bulk (B) ice enable the

comparison of the conversion efficiency of HOCl and Cl O to Cl on frozen salt

2

solutions performed under the same conditions. The chosen HOCl and Cl O

2

concentrations are high in order to enable observation of the possible formation

11

K3

of Cl , namely 4.5!10 and 1.5!10 molecule cm , respectively. In the

2

absence of any major product formation we conclude that the uptake of both

HOCl and Cl O is non-reactive and leads to adsorption of both gases on the

2

surface of ice. Without investigation of the condensed phase we are unable to

state on the possible inverse of reaction (12) upon HOCl adsorption.

3

.2 Pure Cl O uptake on frozen KCl and NSS solution at [Cl O] >

2

11

−3

1

0

molecule cm

Figure 1a (upper left panel) displays a Cl O uptake experiment on frozen KCl

2

11

K3

solution as a function of time using [Cl O] = 2.3!10 molecule cm corre-

sponding to F (Cl O) = 1.ꢀ!10 molecule s . This is similar to the concentra-

2

14

K1

in

2

tion of the Cl O impurity found in the used HOCl.Cl O mixture. When the

2

frozen KCl solution is exposed to Cl O at t = 80 s, a transient production of Cl

2

is observed that is accompanied by a continuous production of HOCl reaching a

steady state level at t = 400 s. In contrast, the Cl production is decreasing to a

2

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

Heterogeneous Chemistry of Cl

2

O and HOCl

1127

2

Fig. 1. a) and b) represent Cl O uptake performed on frozen KCl and NSS solution, respectively,

at 200 K and using the 4 mm diameter escape aperture. c) and d) represent HOCl uptakes

performed on frozen KCl and NSS solution, respectively. The dotted line (trace a) represents

Cl

2

, the hashed curve is Cl

2

O (trace b) and the solid curve is the HOCl (trace c) MS signal. In

1

4

14

K1

14

a) FCl2O = (1.ꢀ±0.4)!10 , in c) FHOCl = ꢀ.6!10 and FCl2O = 1.9!10 molecule s . In b)

1

4

14

F

Cl2O = (1.9±0.5)!10 and in d) FHOCl = (9.0±0.3)!10 and FCl2O = (2.1±0.5)!10 mole-

K1

cule s

.

low steady state level after the primary burst, presumably because of the accumu-

¯

lation of OH on the surface according to reactions (12), (18) and (19):

Cl O(g) + H O(ads)$2 HOCl(ads)

(12)

(18)

(19)

2

K

+

K

+

++

HOCl(ads)+Cl +M /Cl (g)+M +OH (M = Na ,K ,Ca ,…)

2

K

Cl O(g) + Cl /ClO + Cl (g)

2

Uptake experiments of pure Cl O were also performed on frozen NSS solu-

2

tion at 200 K for which a typical uptake is shown in Figure 1Ib for [Cl O] given

2

in Table 1. At t = 30 s, the plunger is lifted and the substrate exposed to Cl O

2

which leads to steady state formation of Cl accompanied by slow formation of

2

HOCl. This steady state formation of Cl displayed in Figure 1b is in contrast to

2

the corresponding uptake of Cl O performed on KCl frozen solution (Figure 1a).

2

Cl O uptake was also performed on frozen NaCl and RSS solution which show

2

identical behaviour in terms of kinetics and formation of products to frozen KCl

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

1

128

P. Pratte and M. J. Rossi

solution. This is in contrast to uptake of Cl O on NSS that is characterized by

2

sustained formation of Cl and slow formation of HOCl. The major difference

2

between NSS and RSS frozen solution is that NSS contains fatty acids, polyfunc-

tional partially oxidized organics or humic acid-like substances (HULIS) [33]

that may affect the formation of Cl . We believe that NSS is sufficiently inter-

2

nally buffered to be able to neutralize the resulting hydroxyl ion in reaction (18)

so as to support a sustainable formation of Cl . In contrast, the formation of Cl

2

comes to a halt in unbuffered frozen solutions such as KCl and RSS because the

halide displacement resulting in Cl is a general acid catalyzed reaction requiring

2

K

the presence of HOCl rather than ClO . Alternatively, NSS may contain organic

matter that is easily oxidized by either HOCl or Cl O thereby reducing the chlo-

2

rine to Cl₂.

3

.3 Mass balance for pure Cl O uptake

2

After each uptake experiment, the substrate is isolated from the reactive gas in

order to halt the reaction and measure the flow rate. The area of the MS signal

corresponding to the reactant lost or product formed was evaluated and leads to

the number of molecules taken up or produced during the reaction. As an exam-

ple, the dashed area in Figure 1a corresponds to the yield of Cl on the time

2

scale of the experiment. After the experiment, the surface was annealed at 240 K

in order to measure desorbed products at 240 K.

The mass balance from Figure 1a between Cl O uptake on frozen KCl solu-

2

1

6

tion and product formation may be expressed as follows: ꢀ.3!10 Cl O mole-

2

cules were taken up against production of 8.4!10¹⁶HOCl and

4

.9!10¹⁶Cl molecules produced (see Table 5). After uptake, the frozen solu-

2

15

tion was annealed at 240 K resulting in the desorption of 1.6!10 Cl O corre-

2

sponding to approximately 2% of the total Cl O initially taken up. However,

2

an important fraction of adsorbed HOCl is collected during thermal desorption

corresponding to 2!10¹⁶molecules. During thermal desorption, no Cl was

2

measured even if the plunger was lowered for 20 minutes in order to let the

frozen film react. Consequently, no delayed Cl production is observed in the

2

isolated sample compartment.

The mass balance expressed per chlorine atom is established including the

products of desorption listed in Table 5 for Cl O uptake according to equation

2

(

20) where the brackets correspond to molar yields:

!(Cl O_t_a_k_e_n_u_p) = HOCl_p_r_o_d_u_c_e_d+ HOCl_d_e_s_o_r_b_e_d+ 2!(Cl O_d_e_s_o_r_b_e_d) + (Cl )

2 produced

2

(20)

•

In equation (20), we assume that for each Cl produced, one Cl came from

2

•

the Cl O source while the second Cl was contributed as halide from the salt

2

substrate surface.

The chlorine mass balance displayed in Table 5 between Cl dosed and recov-

ered is excellent, namely 1.46!10¹^ꢀCl dosed vs. 1.56!10 Cl recovered. This

1ꢀ

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

Heterogeneous Chemistry of Cl

2

O and HOCl

1129

means that all chlorine deposited is recovered from the KCl frozen solution held

1

K3

at 200 K for initial [Cl O] > 1!10 molecule cm . The data taken from

2

16

Figure 1a and Table 5 reveal that 4.9!10 Cl were produced which means

2

that 4.9!10¹⁶surface (or near-surface) chloride ions were consumed. The ratio

K

16

1ꢀ

of the surface Cl to the total chlorine taken up is 4.9!10 .2!0.ꢀ3!10

.33±005 which means that 33% of the deposited Cl from Cl O react with Cl

at the substrate interface on frozen KCl solution at 200 K for [Cl O] > 10 mole-

=

K

0

2

11

2

K3

cule cm (see Table 5, next to last column).

The same mass balance check was performed for pure Cl O uptake at 200 K

2

on NSS frozen solution and again leads to excellent agreement between depos-

ited and recovered chlorine as displayed in Table 5 as well. As an example, data

1

on Cl O uptake on NSS frozen solution at 200 K for [Cl O] > 10 mole-

2

K3

cule cm , displayed in Figure 1b are summarized in Table 5. In Figure 1b,

16

1ꢀ

1

.41!10¹^ꢀCl O were taken up whereas 0.9ꢀ!10 HOCl and 1.96!10 Cl₂

2

15

16

were generated. Thermal desorption led to 3.ꢀ!10 Cl O and ꢀ!10 HOCl.

2

Using equation (20) we obtain excellent agreement between dosed and recovered

chlorine (2.82!10¹^ꢀvs. 2.83!10 molecule). The ratio of the unreacted to

1ꢀ

dosed Cl O is approximately 2.5%, and approximately ꢀ0±ꢀ % of the deposited

2

K

chlorine reacts with surface Cl . We note an important relative increase in sur-

face reactivity of Cl to Cl in going from frozen KCl to NSS solution at 200 K

for [Cl O] > 10 molecule cm

K

2

11

K3

.

2

The yield of Cl on frozen KCl and NSS solution was plotted in Figure 2 as

2

a function of Cl O taken up for several experiments. Figure 2 shows that the

2

yield of Cl per Cl O taken up is larger on frozen NSS than on KCl solution at

2

1ꢀ

2

0

00 K. As an example, 1!10 Cl have been produced on NSS compared to

2

.3!10¹^ꢀon KCl, NaCl and RSS frozen solution for 1!10 Cl O taken up.

1ꢀ

2

This difference in slope is consistent with the fast rate of formation of Cl and

2

the slow rate of formation of HOCl as displayed in Figure 1b. It results in a

slope of 1.11±0.30 displayed in Figure 2. In contrast, the experiment displayed

in Figure 1a shows a slow rate of formation of Cl and a fast rate of formation

2

of HOCl resulting in a slope of 0.40±0.11 in Figure 2, a factor of 2.8 lower than

NSS. Apparently, the rates of formation of Cl by surface halogen exchange with

2

chloride and formation of HOCl by hydrolysis are competing against each other

in Cl O uptake at 200 K. The predominant formation of Cl from pure Cl O on

2

frozen NSS may be either explained by oxidation of organic matter and concomi-

tant reduction of Cl O or by predominant reverse disproportionation (or halogen

2

exchange). Hydrolysis of Cl O resulting in HOCl is apparently not competitive

2

and therefore represents the minority channel on NSS. In contrast, on frozen KCl

salt solutions the competitive situation is reversed with the hydrolysis channel

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

1

130

P. Pratte and M. J. Rossi

Fig. 2. Correlation between the yield of Cl

Cl

2

Cl O] = 2.ꢀ!10 molecule cm . Each point corresponds to a separate experiment and the

2

and Cl

2

O taken up. Uptake experiment of pure

2

O on frozen KCl (▲), NSS (+), NaCl (◆) and recrystallized RSS (■) solution at 200 K for

1

K3

[

integration time is ꢀ00±50 s.

being dominant and the oxidation.reduction channel being outrun by hydrolysis,

perhaps owing to the absence of organic matter in the KCl salt solution.

3

.4 Uptake of the HOCl/Cl O mixture on frozen KCl and NSS solution

2

The uptake of the HOCl.Cl O mixture on frozen KCl and NSS solution has

2

been measured relative to pure Cl O uptake. Figure 1c shows a typical uptake

2

experiment of a HOCl.Cl O mixture on frozen KCl solution as a function of

2

11

K3

time at 200 K for [HOCl] > 10 molecule cm . At t = 160 s, the plunger is

lifted and a constant rate of formation of Cl is observed to 250 s, after which

2

the Cl rate of formation decreases. In contrast, the uptake of a HOCl.Cl O

2

mixture on frozen NSS solution reveals a continuous formation of Cl upon

2

lifting the plunger at t = 50 s (Figure 1d). The formation of Cl remains constant

2

during more than one hour (data not shown). Repetitive uptake experiments of

HOCl.Cl O mixtures on fresh and poisoned frozen KCl solutions are displayed

2

in Figures 11a and 11b, respectively. It appears that the Cl yield is smaller for

2

frozen KCl solution that has already been exposed to the HOCl.Cl O mixture

2

compared to the fresh substrate. Furthermore, the Cl O uptake (Figure 11a and

2

1

1b, trace (b)) saturates at 250 s for a poisoned substrate whereas it saturates

after 400 s on a fresh frozen KCl solution. Despite the saturation of the Cl O

uptake Cl formation is still on-going at F(Cl ) = 3!10 molecule s . The

2

13

K1

2

only possible source for the production of Cl is therefore HOCl uptake from the

2

14

K1

HOCl.Cl O mixture at a flow rate of 1!10 molecule s . On a relative basis,

2

for ten HOCl taken up, three are converted to Cl while seven stay adsorbed on

2

the substrate and will be recovered during thermal desorption at 240 K in mass

balance experiments.

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

Heterogeneous Chemistry of Cl

2

O and HOCl

1131

Fig. 3. Correlation between the yield of Cl

HOCl.Cl

salt (■) at 200 K for [HOCl] = 4.5!10 molecule cm . Each point corresponds to a separate

2

and Cl

2

O taken up. Uptake experiment of the

2

O mixture on frozen KCl solution (▲), NSS (+), NaCl (◆) and recrystallized sea

1

K3

experiment and the integration time is ꢀ00±50 s.

Role of acidity on frozen KCl solution

In order to find out how acidity could affect the yield of Cl formation, we have

2

performed uptakes of HOCl.Cl O mixtures on buffered frozen KCl solutions at

2

11

K3

2

00 K and pH = 4 for [HOCl] > 10 molecule cm as displayed in Figure 12.

The formation of Cl remains nearly constant during uptake of the HOCl.Cl O

2

mixture on the buffered frozen KCl solution akin to neutral frozen NSS solution

as shown in Figure 1d. In fact, the buffered solid allows the neutralisation of

K

OH produced in reaction (18). Buffering is obviously not necessary for sustain-

ing Cl formation on frozen NSS solutions, presumably owing to their natural

2

K

buffering capacity. As long as acidity is available to neutralize OH (reaction

(18)), it prevents the neutralization of HOCl that would accumulate non-reactive

hypochlorite salts on the frozen KCl surface and thereby poison the surface with

respect to further reaction of both HOCl and Cl O.

2

Mass balance for the uptake of the HOCl.Cl O mixture on frozen KCl and

2

NSS solution

The mass balance was also checked for the uptake of the HOCl.Cl O mixture

2

on frozen KCl and NSS solution using equation (21):

2

!(Cl O_t_a_k_e_n_u_p) + (HOCl_t_a_k_e_n_u_p) = (HOCl_d_e_s_o_r_b_e_d) + 2!(Cl O_d_e_s_o_r_b_e_d) + (Cl )

2 2 2 produced

(21)

Taken as an example from Table 5, total chlorine taken up (1.1ꢀ!10¹^ꢀmole-

cule) equals the recovered amount (1.14!10¹^ꢀmolecule) for a frozen KCl solu-

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

1

132

P. Pratte and M. J. Rossi

Fig. 4. The fractional yield of Cl

2

generated from HOCl reacting at 200 K plotted as a function

of the HOCl taken up on frozen KCl (▲), NaCl (◆) and on recrystallized SS (■) after correction

for Cl production owing to Cl O uptake following equation (22), (a). Similar experiments were

2

1

K3

performed on frozen NSS solution for [HOCl] = 4.5!10 molecule cm , (b). Each point

represents a new sample at different reaction times and at identical [HOCl]; the integration time

is ꢀ00±50 s.

K

tion. The ratio between the surface Cl estimated from the measured Cl and

2

the total chlorine deposited yields 0.50±0.06 which is larger by 50% (0.50.

0

.33) compared to the value measured for pure Cl O interacting with frozen KCl

2

solution, in agreement with Table 5.

The presence of HOCl apparently leads to an additional depletion of surface

K

Cl . Consequently, the HOCl taken up could contribute to the formation of 50%

more Cl with respect to uptake of Cl O alone if we assume that Cl O and HOCl

2

are both taken up independently. As will be pointed out below, this assumption

may not be valid in the present case. Results displayed in Tables 2 to 4 reveal

that the uptake kinetics of Cl O is somewhat inhibited by the presence of HOCl

2

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

Heterogeneous Chemistry of Cl

2

O and HOCl

1133

Table 2. Summary of γ values for fresh, that is non-poisoned, frozen salt solution at 200 K

using the 4 and 14 mm diameter escape aperture.

Fresh frozen salt solution

Formation of

KCl

NSS

Slow for pure Cl

2

O uptake

None for pure Cl

2

O uptake

14

HOCl

Formation of

Transient for HOCl uptake and for

pure Cl O uptake

Fast for HOCl and pure Cl

4

2

Cl

2

Kinetics (γ)

4(mm)

4

14

HOCl.Cl

2

O mix-

HOCl

(1.2±0.5)!10^K²(8.3±2.5)!10^K²(1.6±0.5)!10^K²(5.4±1.6)!10^K²

ture

(Cl

HOCl

Cl

(Cl

Cl

2 0

O)

(2.8±1.3)!10^K³(2.5±0.ꢀ)!10^K²(2.5±0.ꢀ)!10^K³(2.6±0.8)!10^K²

(

2

O)SS

(

2

O)

0

(3.0±1.1)!10^K²(1.8±0.6)!10^K¹(1.6±0.6)!10^K²(1.8±0.4)!10^K¹

(4.6±0.8)!10^K⁴(1.1±0.3)!10^K¹(1.8±0.9)!10^K²(1.6±0.6)!10^K¹

(4.5±1.4)!10^K²(2.8±1.1)!10^K¹(3.3±1.4)!10^K²(2.5±0.ꢀ)!10^K¹

(4.ꢀ±1.2)!10^K³(1.4±0.5)!10^K¹(2.8±0.8)!10^K³(2.3±0.6)!10^K¹

(

2

O)SS

Pure Cl

2

O

2 0

O)

(

2

O)SS

Table 3. Summary of γ values for a poisoned (contaminated) frozen salt solution at 200 K

using the 4 and 14 mm diameter escape aperture.

Contaminated frozen salt solu-

tion

KCl

NSS

Formation of

HOCl

Slow for pure Cl

2

O uptake

None for pure Cl

2

O uptake

14

Formation of

Transient for HOCl uptake and for pure Fast for HOCl and pure Cl

Cl O uptake

2

Cl

4(mm)

HOCl (Cl

2

Kinetics (γ)

HOCl.Cl

mixture

4

14

4

(5.2±2.3)!10^K²

(3.6±0.5)!10^K¹(8.2±2.2)!10^K³(1.1±0.6)!10^K¹

2

O

2

O)

0

HOCl (Cl

Cl O)

Cl

(Cl

Cl

2

O)SS

(3.ꢀ±1.3)!10^K³

(2.1±0.9)!10^K²

0

(5.1±1.4)!10^K²

(1.9±0.ꢀ)!10^K³

(1.ꢀ±0.ꢀ)!10^K²(1.3±0.ꢀ)!10^K³(2.0±0.8)!10^K²

(2.ꢀ±1.6)!10^K¹(2.8±1.0)!10^K²(2.3±0.4)!10^K¹

(4.8±1.9)!10^K²(1.2±1.2)!10^K²(1.0±0.3)!10^K¹

(2.8±0.9)!10^K¹(3.4±1.4)!10^K²(2.4±0.9)!10^K¹

(8.0±2.1)!10^K²(6.ꢀ±0.8)!10^K³(2.4±1.1)!10^K¹

(

2

0

(

2

O)SS

Pure Cl

2

O

2

O)

0

(

2

O)SS

which suggests a lower reactivity of HOCl on both frozen KCl as well as NSS

solution compared to Cl O.

2

For the frozen NSS solution we found that 20% of the produced Cl origi-

2

nates from HOCl (0.85.0.ꢀ0 = 1.21, see Table 5). This seemingly lower enhance-

ment of Cl production owing to the presence of HOCl for frozen NSS compared

2

to KCl solution stems from the fact that the uptake of pure Cl O on frozen NSS

2

is already so important that it leaves no room for a large additional effect due to

HOCl adsorption because of surface saturation of the uptake. In order to illustrate

the yield difference of Cl on frozen KCl and NSS solution upon uptake of the

2

HOCl.Cl O mixture, we have plotted the yield of Cl as a function of the Cl O

2

taken up in Figure 3. We conclude that the Cl yield is marginally larger on

2

frozen KCl, NaCl or RSS solution compared to frozen NSS solution. When

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

1

134

P. Pratte and M. J. Rossi

Table 4. Summary of γ values for fresh frozen salt solutions at 215 K using the 4 mm diameter

escape aperture.

Fresh Sample frozen solution

KCl

NSS

Formation of

Slow for pure Cl

take

2

O up- None for pure Cl

2

O up-

HOCl

take

Formation of

Transient for HOCl up- Fast for HOCl and pure

Cl

2

take and for pure Cl

uptake

2

O

2

Cl O uptake

K3

(5.2±1.8)!10^K³

Kinetics (γ) HOCl. HOCl (Cl

Cl O mixture

2

O)

0

(3.ꢀ±0.8)!10

2

K3

HOCl (Cl

Cl O)

Cl

(Cl

Cl

2

O)SS

(2.1±0.6)!10

(4.6±1.4)!10

K2

(

2

0

(3.0±0.9)!10

(3.0±1.1)!10

K2

(

2

O)SS

(1.5±0.3)!10

(2.1±0.ꢀ)!10

K2

Pure Cl

2

O

2

O)

0

(4.2±1.1)!10

(4.5±1.2)!10

K3

K2

(

2

O)SS

(1.4±0.4)!10

(3.1±0.9)!10

comparing Figures 2 and 3 the most striking feature is the differing slope for Cl₂

formation between uptake of pure Cl O (Figure 2) or a mixture of Cl O.HOCl

2

(

Figure 3). Table 5 reveals that for pure Cl O uptake the Cl yield increases by

2 2

a factor of four in going from frozen KCl to NSS solution, whereas the yield of

Cl O taken up only increases by a mere factor of two at a given initial concentra-

2

tion. In contrast, for the mixture Cl O.HOCl both the Cl as well as the quantity

2

of Cl O taken up increase by a factor of four which leaves the slope in Figure 3

2

practically unchanged (1.53 vs. 1.39) when comparing frozen KCl with NSS

solution. One should be aware that the enhancement factors from Table 5 refer

to individual experiments whereas the slope changes in Figures 2 and 3 result

from a least-squares fit to several experiments, thus corresponding to an average.

This is the principal reason for the difference between the slope ratio of 2.8

displayed in Figure 2 compared to a factor of four difference in Cl yield from

2

Table 5 in going from KCl to NSS, both for pure Cl O and HOCl.Cl O.

2

Cl formation from HOCl uptake

2

In order to quantify the amount of HOCl taken up which forms Cl on frozen

2

KCl solution, several HOCl.Cl O uptakes were performed. Under the as-

2

sumption that HOCl and Cl O interact independently on frozen solution (see

2

remark above), the slopes in Figures 2 and 3 were used in order to separate

the contribution of the generated Cl of both HOCl and Cl O. Using equations

2

(

22) and (23), it is possible to calculate the fraction of the HOCl taken up

which leads to the formation of Cl₂:

Cl₂₍_H_O_C_l_c_o_n_t_r_i_b_u_t_i_o_n₎= Cl₂₍_H_O_C_l_._C_l₂_O_m_i_x_t_u_r_e₎– Cl₂₍_c_a_l_c_u_l_a_t_e_d_,_p_u_r_e_C_l₂_O_u_p_t_a_k_e₎(22)

Cl₂₍_c_a_l_c_u_l_a_t_e_d_,_p_u_r_e_C_l₂_O_u_p_t_a_k_e₎= Cl O₍_H_O_C_l_._C_l₂_O_m_i_x_t_u_r_e₎!S₍_p_u_r_e_C_l₂_O_u_p_t_a_k_e₎

(23)

2

where Cl₂₍_H_O_C_l_._C_l₂_O_m_i_x_t_u_r_e₎is the Cl yield generated from simultaneous uptake

2

of both Cl O and HOCl, Cl

is the Cl yield generated

2

2 (calculated, pure Cl2O uptake)

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

Heterogeneous Chemistry of Cl

2

O and HOCl

1135

according to reactions (18) and (19) and calculated using the corresponding

slopes S₍_p_u_r_e_C_l₂_O_u_p_t_a_k_e₎from Figure 2. In order to determine the fraction of HOCl

that results in Cl , Cl

calculated from equation (22) is plotted as

2

2 (HOCl contribution)

a function of HOCl taken up as shown in Figure 4. The slopes in Figure 4a and

4

0

b are 0.30±0.10 on frozen KCl, NaCl and recrystallized sea salt (RSS), and

.23±0.13 on frozen NSS solution, respectively, at 200 K. Given the scatter in

the results displayed in Figure 4b, we conclude that between 10 and 40% of the

HOCl taken up is converted to Cl on both frozen KCl and NSS solution at

2

00 K, independent of the frozen substrate and after correction for the formation

of Cl from Cl O following equations (22) and (23).

2

In conclusion the data displayed in Figure 2 reveal that the Cl yield per

2

Cl O taken up is 2.8 times higher for NSS than for the remainder of the substrates

2

for pure Cl O, that is without interference from HOCl. Figure 3 reveals that for

2

the Cl O.HOCl mixture the Cl yield per Cl O taken up is significantly higher

2

on all Cl-containing substrates compared to pure Cl O. It is suggested that this

2

yield increase, expressed as an average slope of 1.46 in Figure 3 compared to

1

.11 in Figure 2 is due to the presence of HOCl in the reacting gas. Figure 4

finally displays a fairly weak Cl yield of 0.3±0.1 per HOCl taken up, regardless

2

of the substrate and after correction for the Cl yield caused by Cl O. We propose

2

to use the slopes as a parameter describing the relative propensity of Cl O and

2

HOCl to generate Cl from chloride-containing substrates at 200 K. It thus turns

2

out that Cl O generates roughly twice the Cl yield compared to HOCl taking

2

the ratio of slopes of 1.11 vs. 0.3 and normalizing the yield per Cl atom capble

of generating Cl . In addition, Table 5 reveals that a large fraction of HOCl may

2

be recovered through desorption at 240 K in contrast to Cl O that has all but

2

disappeared. In summary, the quantitative results show that the Cl yields per

2

HOCl taken up are significantly lower than for pure Cl O taken up on both

2

frozen KCl, NaCl, RSS and NSS solutions thus reflecting the lower reactivity of

HOCl compared to Cl O for uptake on frozen halide solutions. This remains

2

valid even when taking into account the redox properties of humic-like substan-

ces that may occur in NSS [33,34].

4

. Br single pulse event

2

Immediately after the exposure of the HOCl.Cl O mixture or pure Cl O on the

2

11

K3

frozen NSS solution for concentrations larger than 10 molecule cm , a single

1

5

Br pulse event of approximately (1.0±0.3)!10 molecules occurs during the

2

first moments of the experiment. Br pulse events are displayed in Figures 5a

2

and 5b for pure Cl O and HOCl.Cl O uptake on NSS frozen solution, respec-

2

tively. By performing repetitive uptake experiment on frozen NSS solution, we

do not detect additional Br bursts. We thought at first that only the surface

2

–

adsorbed Br were consumed in the first Cl O or HOCl uptake. However, no

2

additional Br bursts could be observed once the exposed sample had melted and

2

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

1

136

P. Pratte and M. J. Rossi

Fig. 5. Uptake of pure Cl

2

O (a) and HOCl.Cl

2

O mixture (b) on frozen NSS solution at 200 K

for an escape orifice of 4 mm diameter. The symbols a, b, c and d correspond to the MS signal

of Br (m.e 160), HOCl (m.e 52), Cl (m.e ꢀ0) and Cl O (m.e 51), respectively.

2

let above 2ꢀ3 K for 20 minutes before freezing the sample and performing a

second HOCl or Cl O uptake. This may be related to the slow rate of diffusional

2

mixing at temperatures in the vicinity of 2ꢀ3 K which prevents the effective

replacement of a Br-depleted solution from the interface by fresh bulk.

K

One in 600 of NaCl is Br in NSS solution resulting in approximately

K

20

K

3

0

.001 M of Br . This represents approximately 6!10 Br per dm of solution.

K1

In the present experiments, 5 ml of solution containing 35 gl of NSS was

poured into the sample holder and resulted in V = V_s_o_l(r_f_r_o_z_e_n.r ) =

frozen

sol

.5 cm , where V = 5 cm³is the volume of the liquid solution, r

3

=

4

1

sol

K3

.02 g cm and r_f_r_o_z_e_n= 0.92 g cm are the densities of the liquid and of the

frozen solution, respectively. This amounts to 3!10 Br cm and reveals

that 2!10 .2.ꢀ!10 = 0.08% of the total Br in the sample was available to

18

K

K3

15

18

K

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

Heterogeneous Chemistry of Cl

2

O and HOCl

1137

react and produce Br on frozen NSS solution at 200 K for both Cl O and HOCl.

2

Cl O uptake.

2

It was found in previous work [35] that HOBr taken up on frozen chloride

solution is completely converted to BrCl rather than transformed to Br . In con-

2

trast, HOCl taken up on frozen bromide solution mostly produces Br . In conclu-

2

K

sion, even if only 0.08% Br is activated in the atmosphere, it could have signifi-

cant implication on the composition of the marine boundary layer because only

catalytic amounts of Br are necessary to start polar sunrise chemistry.

2

5

. Influence of concentration, residence time and

temperature on the Cl yield

2

Switching from the 4 to 14 mm diameter escape aperture reduces both the resi-

dence time as well as the pressure in the reactor by a factor of 12 according to

Table 1. Under those conditions the chlorine mass balance for uptake of both

pure Cl O and HOCl.Cl O at T = 200 K is not satisfied anymore as may be

2

seen from Table 6 because 20 to 30% of the dosed chlorine is not recoverable

by thermal desorption at 240 K. The quantity of dissolved Cl O and.or HOCl is

2

probably too small for it to desorb at 240 K

Figure 6b displays the yield of Cl resulting from pure Cl O uptake on frozen

2

KCl solution at 200 K as a function of time as well as of Cl O taken up for a

2

single uptake experiment. A slope of 1.0 means that 50% of the chlorine taken

up as Cl O turns up as Cl . We note that the scale of the exposure time is not

2

necessary linear with the uptake of Cl O as it depends on the temporal depend-

2

ence of the uptake. The data in Figure 6b are to be compared with Figure ꢀb

corresponding to HOCl.Cl O uptake on frozen KCl solution at 200 K. Compari-

2

son of the slopes in Figures 6b and ꢀb at steady state (for t > 350 s) leads to the

conclusion that they are identical within experimental uncertainty. The same is

true for HOCl.Cl O and Cl O uptake on frozen NSS solution at 200 K whose

2

data are not shown. This together with a comparison of the cumulative yields of

Cl relative to Cl O and HOCl in pure Cl O and HOCl.Cl O uptakes at 200 K,

2

respectively, strongly suggests that the uptake of HOCl by itself does not lead

to Cl formation at the low partial pressures used under the assumption that Cl O

2

and HOCl uptake are independent of each other. Table 5 hints at the inhibition

of Cl O uptake in the presence of HOCl as do the data of Table 5, as mentioned

2

above, and suggest that the assumption of complete independence between HOCl

and Cl O may not hold. Consequently, HOCl is reactive only on frozen salt

2

11

K3

solutions for concentrations on the order of 10 HOCl cm corresponding to

0 ppb or more. This in turn suggests that the species taken up is in reality Cl O

1

2

whose formation from HOCl is favored at high [HOCl].

Uptake experiments of pure Cl O and HOCl.Cl O at 215 K obtained in the

mm diameter aperture reactor revealed the same results as for the 14 mm

2

4

diameter orifice reactor discussed above, namely that there is no additional for-

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

1

138

P. Pratte and M. J. Rossi

Fig. 6. (a) and (b): pure Cl

2

O uptake on frozen KCl solution at 200 K for [Cl

.9!10 molecule cm using an escape orifice of 14 mm diameter. (b) Cl yield as a function

of Cl O taken up for raw data displayed in (a). On graph (a), the dotted and solid curves show

the formation of Cl and HOCl, respectively. The hashed curve corresponds to uptake of Cl O.

2

O] =

1

0

K3

1

2

mation of Cl due to heterogeneous interaction of HOCl with frozen salt solutions

2

beyond the one generated by the reaction of Cl O. Moreover, the Cl O heteroge-

2

neous reaction is again significantly inhibited by the presence of HOCl. Compari-

son of the Cl yields with Cl O and HOCl uptake at 215 K.4 mm diameter

2

orifice reactor listed in Table 6 confirms both results, namely absence of Cl₂

formation from HOCl alone, and strong inhibition of Cl O uptake in the presence

2

of HOCl. Figures 13 and 14 present time-dependent cumulative yields of Cl₂

also as a function of Cl O uptake, and confirm the above conclusions. It therefore

2

seems that Cl O is more reactive than HOCl on frozen salt solutions at 200 and

2

15 K as far as Cl release is concerned. In addition, the slightly larger slopes

2

in going from Figures 13a and 14a indicate that the Cl yields per Cl O taken

2

up increase with temperature in the range 200 to 215 K.

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

Heterogeneous Chemistry of Cl

2

O and HOCl

1139

Fig. 7. (a) and (b). HOCl.Cl

2

O uptake on frozen KCl solution at 200 K for [HOCl] =

.ꢀ!10 molecule cm using an escape orifice of 14 mm diameter. (b) Cl yield as a function

of Cl O taken up and exposure time t for raw data displayed in (a). On graph (a), the dotted

and solid curves show the formation of Cl and the uptake of HOCl, respectively. The dashed

curve corresponds to uptake of Cl O.

1

0

K3

3

2

6

. Uptake coefficient of Cl O and HOCl

2

The uptake kinetics for fresh frozen KCl solutions at 200 K are displayed in

Table 2 and Figure 8 in terms of γ (HOCl) which increases by a factor of ten

ss

when [HOCl] decreases from 10¹¹to 10 molecule cm (see Table 1). Uptake

10

K3

coefficients were obtained by averaging over at least two uptake experiments.

K3

We also note that γ (Cl O) for the mixture HOCl.Cl O is less than 10 for

ss

2

11

K3

11

[

HOCl] > 10 molecule cm , but is larger than 0.1 for [HOCl] < 10 mole-

cule cm . This inhibition effect was briefly discussed above in relation to the

total uptakes and may occur owing to high surface concentrations of both HOCl

K3

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

1

140

P. Pratte and M. J. Rossi

2

Fig. 8. Synopsis of steady state uptake coefficients γss for HOCl and Cl O on frozen salt

solutions of KCl and NSS at different experimental conditions. The uptake coefficient measured

corresponds to the species first mentioned in the legend. The uncertainties in γss are estimated

as ±10 %.

and.or Cl O. This may lead to surface site saturation on the time scale of the

2

experiment and reveals complex kinetics at the gas-condensed phase interface.

11

K3

γ (Cl O) for pure Cl O uptake at [Cl O] < 10 molecule cm is larger

ss

2

than 0.05 on fresh and poisoned frozen KCl solution as displayed in Tables 2

and 3, respectively. In addition, γ (Cl O) for pure Cl O on frozen KCl solution

ss

2

10

12

K3

for [Cl O] between 10 to 10 molecule cm decreases by a factor of two for

2

the second uptake. In contrast, γ (Cl O) for pure Cl O on frozen NSS solution

ss

2

1

increases at least by a factor of two for the second uptake for [Cl O] > 10 mole-

cule cm (4 mm diameter orifice). However, γ (HOCl(Cl O)) on frozen NSS

solution did not change for concentrations ranging from 10 to 10 mole-

cule cm and for a fresh or a previously poisoned sample within the experimen-

2

K3

ss

2

10

12

K3

tal uncertainty as displayed in Tables 3 and 4.

The measured uptake coefficient of Cl O for the mixture HOCl.Cl O,

2

γ (Cl O(HOCl)), interacting with frozen NSS solution for [HOCl] ranging from

ss

0

2

10

12

K3

1

to 10 molecule cm was found to be ten times smaller compared to the

uptake of pure Cl O on the same substrate, that is in the absence of HOCl. This

2

may be explained either by the reverse of reaction (12) when HOCl is admitted

into the reactor and more Cl O is generated on the ice that may lead to surface

2

saturation, or by the known propensity of HOCl to poison ice surfaces that is in

contrast to the adsorption of Cl O on pure ice discussed in 3.1.

2

In summary, the kinetics of repetitive uptake for both Cl O and HOCl on a

2

frozen KCl or NSS salt solution will lead to a decrease of γ . In contrast, γ

ss

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

Heterogeneous Chemistry of Cl

2

O and HOCl

1141

1

increases when the sum of [HOCl] and [Cl O] decrease below 10 mole-

2

K3

cule cm . In addition, γ is generally larger on frozen NSS compared to KCl

ss

salt solution for a given HOCl and Cl O concentration at 200 or 215 K (see

2

Tables 3, 4 and 5 and Figure 8). Figure 8 conveys an impression on the relative

magnitude of γ_s_sof HOCl vs. Cl O and underlines the fact that Cl O uptake is

2

significantly faster under all conditions. Previous work [31] has investigated the

HOCl uptake on frozen KCl ice film of 0.1% by weight at 233 K which resulted

K2

in γ (HOCl) < 4.ꢀ!10 . In the present work, we have used a concentration of

3

ss

K1

K

5 g l of KCl which corresponds to approximately 0.2% by weight of Cl .

For a frozen KCl and NSS salt solution at 215 K we have obtained γ (HOCl) =

ss

K3

(

2.1±0.6)!10 and γ (Cl O) = (4.6±1.4)!10 , respectively. These values

ss 2

are consistent with the results presented by Huff and Abbatt at 233 K [35].

7. Atmospheric implications and conclusions

We conclude that significant differences in the yield of Cl owing to Cl O uptake

2

were observed depending on the nature of the frozen substrate. For this reason

KCl and NaCl may not in general be used as a model for NSS that shows special

properties in comparison with the other three chloride-containing substrates,

namely KCl, NaCl and RSS frozen solution. The reason for this special behav-

iour may lie either in the buffering capacity or the presence of oxidizable organic

substances, as alluded to above.

HOCl is thermodynamically unstable and is in equilibrium with Cl O accord-

2

ing to reaction (12) and its inverse. The standard heat of reaction of equilibrium

0

(

12), ΔH , has been measured by several authors, but we prefer the value of

r

0

Knauth et al. [23] owing to the quality of this study with ΔH = 2.ꢀ±0.2 kcal.

r

Mol leading to K²= 0.082 (dimensionless). The kinetic results of this work

98

including the comparison of the reactivity of pure Cl O and a (non-equilibrium)

2

mixture of HOCl.Cl O which we are able to prepare with routinely 25% Cl O

2

in the HOCl sample (see Experimental) suggest that the equilibrium (12) lies on

the side of adsorbed HOCl in the presence of ice or frozen aqueous salt solutions

according to equilibrium (24):

Cl O(ads) + H O(ice)#2 HOCl(ads)

(24)

2

The present results are consistent with a parallel reaction scheme in which

K

Cl O directly reacts with Cl to Cl , either through halogen exchange, reaction

2

(

19), or through a redox reaction on NSS as mentioned above. Unreacted Cl O

2

may form HOCl(ads) that is non-reactive by itself but that slowly may react on

the ice surface by way of equilibrium (24) and regenerate reactive Cl O. There-

2

fore, equilibrium (24) is essential in considering the heterogeneous halogen form-

ing potential of HOCl in the troposphere at suitably low temperatures. The possi-

ble reason for the exothermicity of equilibrium (24) in the forward direction,

which is in contrast to the gas phase, may lie in the ability of HOCl to form

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

1

142

P. Pratte and M. J. Rossi

Fig. 9. HOCl uptake on pure bulk (B) ice held at 200 K using the 4 mm diameter escape

aperture. Curves a (HOCl), b (Cl O) and c (Cl ) were monitored at m.e = 52, 51 and ꢀ0,

respectively. The HOCl and Cl flow rates correspond to (5.0±1.5)!10

1.1±0.5)!10 molecule s , respectively. The corresponding concentrations of HOCl and

2

1

4

2

O

and

1

4

K1

(

Cl

1

11

K3

2

O were 4.5!10 and 1.5!10 molecule cm , respectively.

strong hydrogen bonds with the ice surface, a property that Cl O lacks. Although

2

we do not observe Cl formation by HOCl itself at low residence times (14 mm

2

diameter orifice) or at higher temperature of 215 K in the present experiments

we cannot discount the possibility that Cl may form at atmospherically meaning-

2

ful time scales such as days if we allow the ice substrate to accumulate sufficient

HOCl such that reactive Cl O may be generated following equilibrium (24).

2

Formation of Cl is subsequently expected in view of the high reactivity of

2

Cl O(ads) observed in the present study. We therefore postulate that there is a

2

role for HOCl in regards to formation of Cl in marine boundary layer chemistry

2

at low temperatures although there is no experimental proof available from gas

phase measurements yet.

There have been a number of theoretical studies on the interaction of HOCl

with ice surfaces, but none with Cl O such that the quantitative aspects of equi-

2

librium (24) remain to be elucidated [36–38]. However, the present results offer

a glimpse on the qualitative behaviour of adsorbed HOCl.Cl O by using the Cl₂

2

formation as a probe. The present results are also broadly consistent with Huff

and Abbatt [25] who conducted the only other study with which the present work

is comparable. They noted that HOCl is nonreactive on ice substrates at 233 and

248 K and provided upper limits for γ_s_sthat are consistent with the present

values.

In an atmospheric context, we may estimate the yield of Cl from the uptake

2

of HOCl on marine ices using the present experimental results. From the work

of Huff et Abbatt [25], it is possible to estimate the flux of the dry deposition of

K1

HOCl using F = v C, where v is the vertical deposition velocity in cm s and

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

Heterogeneous Chemistry of Cl

2

O and HOCl

1143

Fig. 10. Cl

aperture. Curves a (HOCl), b (Cl

respectively. The Cl O impurity flow rate was (2.4±0.6)!10 molecule s

2

O uptake on pure bulk (B) ice held at 200 K using the 4mm diameter escape

2

O) and c (Cl ) were monitored at m.e = 52, 51 and ꢀ0,

2

1

4

K1

2

.

C is the HOCl concentration in the atmosphere in molecule cm^K³. Results from

field measurements revealed that the concentration of chloride compounds that

could be photolysed (HOCl or Cl ) is approximately 10 ppt [ꢀ] which corre-

2

8

K3

sponds to 2.95!10 HOCl cm in the spring time. The vertical velocity of the

dry deposition is calculated using v = 1.(R + R + R ), where R is the transport

a

b

c

a

of HOCl across the atmospheric boundary layer, R is the molecular diffusion

b

term and R is the contribution of the heterogeneous reaction expressed in

c

K1

2

s cm . R = (1.uk )(ln(z.z )) , where u is the wind velocity taken as 1000 cm

a

0

K1

K5

s

, k is a proportionality factor equal to 0.4, z = 10 m is the length of the

0

ice roughness and z is the height of an atmospheric layer taken to be 100 m [39].

2

K1

In addition, R = z .D(HOCl) where D(HOCl) = 0.105 cm s is the HOCl

b

0

diffusion coefficient in air. Finally, R = 4.γc where c is the mean molecular

c

K3

velocity of HOCl. Using as an estimate γ (HOCl) ≈ 2!10 on frozen NSS

solution at 200 K for [HOCl] >10 molecule cm (see Table 2), we obtain

v(HOCl) = 0.589 cm s which leads to F(HOCl) = (2.95!10 molecule cm

0.589 cm s ) = 1.ꢀ4!10 molecule cm

ss

11

K3

K1

8

K3

)

K1

8

K2 K1

(

s

. Under the assumption that 30%

of HOCl taken up on marine ice is converted to Cl at 200 K, we estimate that

2

8

K2 K1

ꢀ

K2 K1

F(Cl ) = (1.ꢀ4!10 molecule cm

s

) (0.30) = 5.22!10 molecule cm

s

2

at 200 K.

By considering the formation of Cl as an upper limit from the reaction of

2

HOCl on marine ice for half

a

day, we obtain (5.22!10^ꢀmole-

K2 K1

12

2

cule cm

s

)(43200 s) = 2.25!10 molecule of Cl produced per cm of

2

marine ice per day. If Cl is mixed homogeneously in the atmosphere within a

2

layer of 100 m, the Cl₂concentration could reach up to 2.25!10⁸mole-

K3

cule cm or 9 ppt. This order of magnitude estimate may confirm that halogen

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

1

144

P. Pratte and M. J. Rossi

Fig. 11. Repetitive HOCl uptake on frozen KCl solution at 200 K. The HOCl and Cl

2

O flow

1

4

14

K1

rate were 8!10 and 2!10 molecule s , respectively. The traces a, b and c are related to

2 2

HOCl (m.e 52), Cl O (m.e 51) and Cl (m.e ꢀ0) MS signal, respectively at 200 K, 4 mm

diameter orifice.

activation effectively could lead to an important destruction of hydrocarbons and

ozone in the marine boundary layer during arctic spring time.

The conclusions from the present work may be summarized as follows:

From the comparison of the interaction of pure Cl O on the one hand, and of

2

a non-equilibrium mixture of HOCl.Cl O on several types of chlorine-containing

2

substrates in the range 200 to 215 K we obtain the indication that the reactivity

of Cl O is significantly larger than that of HOCl in regards to the formation of

2

Cl . As an example, the Cl yield resulting from the interaction of Cl O on a

2

frozen salt solution at 200 K is twice that of HOCl.

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

Heterogeneous Chemistry of Cl

2

O and HOCl

1145

Fig. 12. HOCl uptake on frozen acidified KCl solution at 200 K. The HOCl (a) and Cl

2

O (b)

(c) production.

1

4

14

K1

flow rate were 9!10 and 2.4!10 molecule s , respectively, and led to Cl

2

The frozen KCl solution was buffered with 0.068 M NaOH, 0.056 M citric acid and 0.044 M

NaCl (pH = 4.0) at 200 K, 4 mm diameter orifice.

Excellent chlorine mass balances are obtained under certain conditions at

200 K, but not at low residence times and.or low pressures as well as at 215 K.

Uptake of both Cl O and HOCl on frozen NSS (natural sea salt) solution is

2

sustained for at least an hour in contrast to other chlorine-containing substrates

such as frozen KCl, NaCl and RSS (recrystallized sea salt) solution. The facile

reactive uptake of HOCl and Cl O on NSS may be due to the natural buffering

2

capacity or the presence of oxidizable organics in NSS that both are thought to

support the formation of Cl₂.

Cl O and.or HOCl uptake on NSS leads to a one-time Br release in a burst

2

mode that may be significant in relation to marine boundary layer chemistry in

the springtime arctic.

Owing to the thermodynamic instability of HOCl in the gas phase the fate.

formation of the corresponding thermodynamically stable anhydride, Cl O, must

2

also be considered in the presence of a substrate with which both species interact.

Considering the equilibrium between adsorbed HOCl and Cl O on ice it is likely

2

that the gateway to formation of Cl passes through the formation of adsorbed,

2

reactive Cl O by way of adsorption of gas phase HOCl on the surface of the

2

frozen aqueous halide.

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

1

146

P. Pratte and M. J. Rossi

Fig. 13. Uptake of pure Cl

2

O (a) and HOCl.Cl

HOCl] = 4.5!10 molecule cm (4 mm diameter escape orifice). The Cl

2

O (b) on frozen KCl solution at 200 K for

1

K3

[

2

yield is plotted

as a function of exposure time and Cl O taken up.

2

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

Heterogeneous Chemistry of Cl

2

O and HOCl

1147

Fig. 14. Uptake of pure Cl

2

O (a) and HOCl.Cl

HOCl] = 4.5!10 molecule cm (4 mm diameter escape orifice). The Cl

2

O (b) on frozen KCl solution at 215 K for

1

K3

[

2

yield is plotted

as a function of exposure time and Cl O taken up.

2

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

1

148

P. Pratte and M. J. Rossi

Appendix

2 2

Table 5. Summary of mass balance for uptake experiments of HOCl.Cl O and pure Cl O on

frozen KCl and Natural Sea Salt (NSS) solutions at 200 K at concentrations in excess of

1

K3

a

10

molecule cm (4 mm diameter escape orifice) . Integration time is ꢀ00±50 s.

Reactants Escape

orifice

Uptake

Products

Thermal de- Uptake Recovery Relative Ratio de-

sorption at

of chlo- of chlo-

Cl

yield

2

sorbed

vs.remain-

ing HOCl

(

mm)

240 K of chlo-

rine

rine spe-

cies

rine-contain- species

ing species

(

10¹^ꢀ(10¹^ꢀ(10¹^ꢀ(10¹^ꢀ(10¹^ꢀ(10¹^ꢀ

Cl O) HOCl) Cl HOCl) Cl O) HOCl)

(10¹^ꢀ

)

(10¹^ꢀ

)

2

)

2

Cl

2

O.

4

0.ꢀ3

0.2ꢀ

1.41

1.12

0

0.49 0.84 0.016 0.20

1.4ꢀ

1.1ꢀ

2.82

3.04

1.56

1.14

2.83

3.30

0.33

0.50

0.ꢀ0

0.85

-

KCl

200 K)

(

HOCl.

0.63 0.58

-

0.034 0.49

0.ꢀꢀ

-

KCl

200 K)

(

Cl O.

2

0

1.96 0.09ꢀ 0.03ꢀ 0.ꢀ0

NSS

200 K)

(

HOCl.

0.80 2.58

-

0.043 0.63

0.ꢀ9

NSS

200 K)

(

a

Each individual yield measurement has an uncertainty of 5% resulting in a 20% uncertainty in the mass bal-

ance.

Brought to you by | La Trobe University

Authenticated

Download Date | 3/19/19 12:48 AM

Article Doi

Heterogeneous chemistry of Cl2O and HOCl on frozen natural sea salt, recrystallized sea salt, KCl and NaCl solutions at 200 and 215 K

DOI: 10.1524/zpch.2010.6144

Source and publish data:

Authors:

Article abstract of DOI:10.1524/zpch.2010.6144

Full text of DOI:10.1524/zpch.2010.6144

Products guided by the article

R&D Labs maybe for 7791-21-1

Relevant to this article

Hot Product