J . Phys. Chem. 1993,97, 12798-12804
12798
-
Heterogeneous Reactions of HOCl + HCI C12 + H2O and CION02 + HCI Ice Surfaces at Polar Stratospheric Conditions
-
CI2 + HN03 on
Liang T. Chu, Ming-Taun Leu,' and Leon F. Keyser Earth and Space Sciences Division, Jet Propulsion Laboratory, California Institute of Technology, Pasadena, California 91 109 Received: July 30, 1993; In Final Form: September 13, 1993"
+
+
+
+
Heterogeneous reactions of HOCl HC1- C12 H2O (1) and ClONO2 HC1- Cl2 HNO3 (2) on ice surfaces at a temperature of 188 K have been investigated in a flow reactor interfaced with a differentially pumped quadrupole mass spectrometer. Partial pressures for HOCl and ClONO2 in the range of 6.5 X 10-8-2.0 X 10-6 Torr, which mimic conditions in the polar stratosphere, have been used. Uptake of HCl on ice surfaces using partial pressures of HCl in the range of 1.8 X 10-7-8.0 X 10" Torr has been measured prior to contact with HOCl or ClON02. Pseudo-first-order decays of HOCl and ClONO2 over HC1-coated ice surfaces have been observed in all experiments under the conditions of P H O C l < PHCIand PCIONO2 < PHCI used. Both the decay rates of HOCl and CION02 and the growth rates of C12 have been used to obtain reaction probabilities: ye (1) = 0.34 f 0.20 (1 a) and yg (2) = 0.27 f 0.19 (1 a) if we assume that the area of ice surfaces is equal to the geometric area of the flow-tube reactor. The results are in good agreement with previous measurements. By considering the morphology of ice films, we obtain true reaction probabilities yt( 1) = 0.13 f 0.08 and yt(2) = 0.10 f 0.08 using a previously published model of surface reaction and pore diffusion. In addition, the true reaction probability, yt (3), for the C 1 0 N 0 2 H20 HOCl HNO3 (3) reaction has been measured to be greater than 0.03 on ice surfaces. Reaction mechanisms for these heterogeneous reactions 1-3 are discussed, including a possible two-step mechanism for reaction 2 in terms of reaction 3 and reaction 1.
+
I. Introduction It is now well established that heterogeneous reactions on the surfaces of polar stratospheric clouds (PSCs) are of vital importance in converting inactive chlorine to active forms which subsequently deplete ozone through catalytic cycles,such as C10C10 and B M 1 0 mechanisms,in the polar ~tratosphere.l-~ These heterogeneous reactions include
+ HCl ClONO, + HCl ClONO, + H 2 0 HOCl
--
-
+H20 C1, + HNO, HOCl + HNO, C1,
(1)
(2) (3)
Recent model calculations4shave suggested that reaction 1 could be as important as reactions 2 and 3, particularly in the early Antarctic winter if ClONOz concentrations are lower than HOCl concentrations. Nevertheless, reaction 2 has been thought to comprise twoelementary steps: the first step is reaction 3 followed by reaction laG8 Reaction 1 on ice surfaces has been the subject of recent laboratory investigations. Abbatt and Molina7used an electronimpact ionization mass spectrometer (EIMS) interfaced to a flow reactor. These experiments were carried out at ice-film temperatures in the range 195-202 K and HCl partial pressures in the range 10-6-10-s Torr. Both temperatures and HCl partial pressures are significantly greater than those in the polar stratosphere (Le., T = 180-188 K for type I1 PSCs and P H C=~ 10-8-10-7 Torr). They reported a reaction probability yg (1) = 0.16 at 202 K and 0.24 at 195 K, respectively. A small negative temperature dependence was suggested. Hanson and Ravishankara8 used a chemical ionization mass spectrometer (CIMS) for detection in a flow tube reactor and measured the first-order loss of HOCl on ice in the presence of HCI partial pressures of 10-7-10-6 Torr. They reported ye (1) for reaction 1 to be >0.3 at 191 K. In these studies the surface area of the ice film was Abstract published in Advance ACS Abstracts, October 15, 1993.
0022-365419312097-12798%04.00/0
-
+
assumed to be equal to the geometric area of the flow reactor for the determination of yg (1); the internal surface area was not considered. There are several laboratory studies of reactions 2 and 3. Some early investigationsg-'' are believed to suffer from surface deactivation by HN03 because higher reactant concentrations of C1ONO2were used. Nevertheless,these studieswere instrumental in demonstrating that heterogeneous reactions on PSC surfaces are of importance in chlorine activation. Recently Hanson and Ravishankara8Jz used CIMS with HCl partial pressures of lO-'-lo-6 Torr. They reported a value of ye = 0.3 (+0.7,-0.1) at 200 K for reactions 2 and 3 on a surface consisting of NAT and ice. Another study was reported by Abbatt and Molina.6 They used an EIMS apparatus and HCl partial pressures of l 0-6-lO-5 Torr. They obtained ye(2) > 0.2 at 202 K for reaction 2 on a similar surface. Again, in these studies the geometric area was used to calculate the reaction probability. In this work we report a new measurement of the reaction probability for reaction 1 at 188 K using partial pressures of reactants (1W-10-6 Torr) similar to thosein the polar atmosphere. Furthermore, we have reinvestigated reactions 2 and 3 using our new EIMS apparatus taking advantage of higher sensitivity detection. In the following sections we will briefly describe the experimental procedures used in the reaction probability determination and present our experimental results. We then discuss the effect of ice-film morphology on our results and compare them with previous measurements. Finally, a possible two-step mechanism for reaction 2 in terms of reactions 3 and 1, and reaction mechanisms for these heterogeneous reactions on ice surfaces are briefly discussed. 11. Experimental Section
The reaction probability measurement was performed in a flow reactor coupled to a differentially pumped quadrupole mass spectrometer. The details of the apparatus have been discussed in our previous publi~ations,'~J~ and we will only briefly describe it in this article. 0 1993 American Chemical Society
Heterogeneous Reactions
Flow Reactor. The flow reactor was constructed of borosilicate glass, and its dimensions are 1.76 cm inside diameter and 33.0 cm in length. The geometric area of the flow reactor is about 182.5 cm2. The temperature of the reactor was regulated by a refrigerated methanol circulator (Haake, Model FK2) and measured by a pair of thermocouples located in the middle and at the downstream end. During the experiment the temperature was maintained at 188 K. The accuracy of this measurement is about k0.5K. The pressure inside the reactor was monitored by a high-precision pressure meter (MKS Instruments, Model 390 HS, 10 Torr full scale), which was located about 2 cm from the flow reactor at the downstream end. Preparation of Ice Films. The ice film was prepared as follows: Heliumcarrier gas was bubbled through a water reservoir which was kept in a constant-temperature circulator, normally at 293 K, and the helium gas saturated with the water vapor was admitted to the inlet of the sliding Pyrex injector. During the period of deposition the sliding injector was slowly pulled out at a constant speed, and a uniform ice film was deposited on the inner surface of the reactor, which was held at a temperature of 188 K. After the ice film was prepared, the injector was kept at the upstream end in order to prevent warming of the substrate. The amount of ice substrate deposited was calculated from the water vapor pressure, the mass flow rate of the helium-water mixture (which was measured by a Hastings mass flow meter), and the deposition time. Some of the substrates were transferred to a U-tube a t 77 K and weighed on an analytical balance. The results from these two methods are in good agreement. Average film thickness was calculated by using the measured geometric area and weight of the deposit with a value of 0.63 g/cm3 for the bulk density of vapor-deposited water ice;15 the results ranged from 3.7 to 34.1 pm. In addition, morphology of ice films was investigated under similar experimental conditions using an environmental scanning electron microscope.l6 The results suggest that ice films comprise layers of micron-sized granules. This observation is in agreement with surface area measurements obtained by using BET analysis of gas-adsorption isotherms.15-17 This information on the icefilm structure will be used to determine the true reaction probability in a later section. HCI Mixtures and Uptake Measurements. HCl/He mixtures were prepared by mixing Matheson semiconductor-purity HCl (99.995%) and Matheson-purity helium (99.9999%) in a glass manifold which was previously evacuated to 10-6 Torr. Flow rates of the mixtures were monitored by using a Hastings mass flowmeter. At first the He-HCl mixture was admitted to the flow reactor through an inlet located at the downstream end; this bypassed the ice film and allowed the vacuum lines to be conditioned with HCl. At the start of a typical uptake measurement, the flow was redirected through another inlet at the upstream end of the ice film. At saturation, the HCl uptake capacity on the ice surface based on the geometric area of the flow reactor was found to be 8 X 1014-1.1 X 1016 molecules/cm2 depending on the partial pressure of HCl used. These results are reasonably consistent with our earlier observations.14 HOC1 Preparation and Calibration. The HOCl solution was prepared by mixing NaOCl with MgS04.7H20.7J9 Some modifications in the synthesis of HOCl were necessary in order to achieve a stable yield. We dissolved 40 g of MgSO4.7H20 in 75 mL of distilled water. Then we added the MgSOd solution to 75 mL of 6% NaOCl solution drop by drop in the dark. The solution was slowly stirred during the reaction in which a white precipitate of Mg(OH)2 was formed. After completing the synthesis, we separated the Mg(OH)2 precipitate from the solution by decantation. A slightly yellowish clear HOCl (OC1-) solution was obtained. Since HOCl is in equilibrium with OC1- and H+ in solution, it is very important to control the pH of the prepared solution in order to ensure that the equilibrium is shifted toward
The Journal of Physical Chemistry, Vol. 97,No. 49, 1993 12799 the HOCl side. The pH of the solution should be close to neutral. By adding a small amount of dilute H2SO4 to the HOCl solution, we were able to shift the equilibrium toward HOCl and increase the yield of the synthesis. The HOCl was further purified by vacuum distillation in order to decrease impurities such as HCl, Cl2, and C120 which were checked by our mass spectrometer and were found to be smaller than 10%. Helium gas was bubbled through the HOCl solution which was maintained at 273 K. A small amount of water vapor (less than 1% of the ice-film mass) from the HOCl solution was also admitted into the reactor. The water vapor was needed in order to prevent HOCl from decomposing into C120 by the reaction 2HOC1- (2120 + H2O during transport to the flow reactor. The concentration of gas-phase HOCl was calibrated by its production from CION02 in reaction 3. For this we assume (1) a stoichiometric ratio of unity for HOCl formed to CION02 lost and (2) no significant adsorption of HOCl on the ice surface. In these experiments thicker ice films wereused in order to minimize deactivation by product H N 0 3 . CION02 Synthesis and Calibration. ClONO2 was synthesized by mixing a small amount of C120 in N2O5 at 195 K; the mixture was allowed to warm to 248 K to produce ClONOI. This procedureI1J3was repeated several times in order to prepare a sufficient amount of ClONO2. The ClONO2 thus prepared was further purified by vacuum distillation at 195 and 178 K. The Cl2 impurity, which could interfere with our reaction probability measurements, was found to be small. Vapor pressures of ClONO2 in the reservoir were measured by a high precision MKS pressure meter (MKS Instruments, Model 390 HS, 1 Torr full scale). Procedures for Reaction 1. The reaction probability of HOCl on an HC1-covered ice film was determined as follows. First, an ice film about 20 pm thick was prepared on the inner wall of the flow reactor. Second, the film surface was treated with HCl at pressures 9 X 10-7-8 X 10-6 Torr until its saturation uptake capacity was attained (see above). Then without turning off the HCl flow, HOCl at pressures between 1 X 10-7-2 X 10-6 Torr was admitted to the reactor. The loss rate of HOCl and the growth rate of Clz were measured as a function of injector distance, z. The parent peaks of HOCl at m/e = 52 and of Cl2 a t m/e = 70 were used. Knowing the gas flow velocity, u, in the reactor, the reaction time was calculated by using t = z/v. The firstorder rate constant, k,, was calculated from the slope of a linear least-squares fit to the experimental data (see next section). The gas-phase diffusion correction fork, was made by using a standard procedure,20 and the corrected rate, kg, was determined. The diffusion coefficient of HOCl in helium was estimated to be 193 Torr cm2 s-1 at 188 K.21 Based on the geometric area of the flow-tube reactor the reaction probability, -ye, was then calculated by using the following equztion 7,= 2rokg/G + ‘ok,) (4) where ro is the radius of the flow reactor (0.88 cm) and is the average molecular velocity for HOCl. In addition, corrections17J8 were made for the interaction of surface reaction and porediffusion to obtain the true reaction probability, yt (see the following section). Proceduresfor Reaction 2. The reaction probability for reaction 2 was measured in a manner similar to that used for reaction 1. Helium was bubbled through the CION02 reservoir which was maintained a t 155 K or lower, depending on the ClONOz concentrations required in the experiment. The low-temperature bath was prepared by mixing ethanol and liquid nitrogen. The helium gas saturated with ClONO2 was introduced into the flow reactor. The C10N02loss rate a t m/e = 46 and the Clz growth rate a t m/e = 70 were measured as a function of injector distance. The procedures used to obtain reaction probabilities are identical to those used for reaction 1. Thediffusion coefficient of CION02
Chu et al.
12800 The Journal of Physical Chemistry, Vol. 97, No. 49, 1993
TABLE I: Reaction Probability for the Reaction of HOCl+ HCI
-
CIS + H20 on Ice Surfaces
HOCl decay H (Torr) ~ 1.1 x 1v7 1.2 x 1v7 1.4 x 10-7 1.5 x 1v7 1.5 x 1v7 1.6 x 10-7 2.1 x 1v7 2.2 x 10-7 2.6 x 10-7 2.6 x 10-7 2.8 x 1v7 2.9 x 10-7 3.2 x 10-7 3.3 x 1v7 3.3 x 1v7 3.9 x 10-7 4.6 x 1v7 4.8 x 10-7 5.0 x 10-7 5.2 x 1v7 5.5 x 10-7 5.5 x 10-7 5.8 x 1v7 5.9 x 1v7 6.3 x 10-7 7.0 x 1v7 7.1 x 10-7 7.1 x 10-7 8.1 x 10-7 8.8 X lV7 9.9 x 1v7 1.0 x 10-6 1.1 x 10-6 1.3 X 10-6 1.3 X 10-6 1.3 X 10-6 1.7 X 10-6 1.7 X 10-6 1.8 X 10-6 1.8 X 10-6 2.0 x 10-6
PHCI(Torr) 2.82 X 10-6 1.11 x 10-6 2.27 X 10-6 2.84 X 10-6 1.62 X 10-6 9.38 X l t 7 1.11 x 10-6 1.13 X 10-6 2.84 X 10-6 2.27 X 10-6 1.62 X 10-6 3.62 X 10-6 2.27 X 10-6 1.12 x 10-6 9.31 X lV7 2.84 X 10-6 1.63 X 10-6 2.26 X 10-6 1.14 X 10-6 4.65 X 10-6 9.22 X le7 2.83 X 10-6 4.23 X 10-6 2.27 X 1od 1.62 X 10-6 4.65 X 10-6 1.13 X 10-6 1.69 X 10-6 2.10 x 10-6 9.24 X l V 7 1.65 X 10-6 4.65 X 10-6 4.65 X 10-6 1.35 X 10-6 1.38 X 10-6 3.46 X 1od 2.10 x 10-6 4.31 X 10-6 1.43 X 10-6 7.97 x 10-6 1.76 X lod
P
thickness (bm) 19.5 19.6 19.4 19.5 19.2 19.9 19.6 19.6 19.5 19.4 19.2 19.2 19.4 19.6 19.9 19.5 19.2 19.4 19.6 19.1 19.9 19.5 19.9 19.4 19.2 19.1 19.6 19.8 19.9 19.9 19.2 19.1 19.1 20.1 18.5 19.9 18.4 19.2 19.4 18.1 18.7
kp (1/s) 1.18 x 103 1.24 x 103 1-26x 103 1.36 x 103 1.42 x 103 1.03 x 103 1.07 x 103 1.09 x 103 1.32 x 103 1.27 x 103 1.21 x 103 1.19 x 103 1.19 x 103 1.05 x 103 1.22 x 103 1.28 x 103 1.04 x 103 1.61 x 103 1.45 x 103 1.36 x 103 1.44 x 103 1.51 x 103 1.75 x 103 1.04 x 103 1.57 x 103 1.44 x 103 1.70 x 103 1.40 x 103 1.63 x 103 1.30 x 103 1.39 x 103 9.95 x 102 1.40 x 103 1.80 x 103 1.50 x 103 1.44 x 103 1.31 x 103 1.33 x 103 1.57 x 103 1.75 x 103 1.34 x 103
C12 growth
k p (1 /SI 3.83 x 103 3.24 x 103 3.48 x 103 4-52 x 103 4.97 x 103 2.16 x 103 2.31 x 103 2.39 x 103 4.14 x 103 3.58 x 103 3.08 x 103 3.04 x 103 3.01 x 103 2.24 x 103 3.16 x 103 3.76 x 103 2.20 x 103 9.39 x 103 5.33 x 103 4.63 x 103 5.41 x 103 7.13 x 103 1.86 x 104 2.22 x 103 7.91 x 103 5.70 x 103 1.27 X 10' 4.95 x 103 1.02 x 10' 3.87 x 103 4.74 x 103 2.05 x 103 5.13 x 103 2.86 X 10' 6.50 x 103 5.45 x 103 4.09 x 103 4.19 x 103 8.61 x 103 2.11 x 104 4.41 x 103
:k (l/s) 3-83x 103 1.86 x 103 4.99 x 103 6.77 x 103 2.24 x 103
k t (1 /SI 1.18 x 103 9.17 X 102 1.29 x 103 1-37x 103 9.98 X 102 9.11 X 102 1.22 x 103 1.03 x 103 1.57 x 103 1.02 x 103 9.27 X 102 1.82 x 103 1.15 x 103 8.25 X 102 1.16 x 103 1.26 x 103 1.71 x 103 8.29 X 102 1.15 x 103 1-30x 103 1.39 x 103 1.41 x 103 9.17 X lo2 1.43 x 103 1.36 x 103 1.53 x 103 1.20 x 103 1.49 x 103 1.31 x 103 1.37 x 103 1.38 x 103 1.31 X lo3
YC
0.25 0.19 0.20 0.25 0.27 0.13 0.14 0.14 0.23 0.21 0.18 0.18 0.18 0.13 0.18 0.21 0.13 0.46 0.29 0.26 0.29 0.37 0.74 0.13 0.40 0.31 0.58 0.27 0.49 0.22 0.26 0.12 0.28 0.95 0.34 0.30 0.23 0.24 0.43 0.80 0.25
9.53 x 1.33 x 9.41 X 1.37 x 1-49x 1.34 x 1.52 x
YP
0.25 0.13 0.31 0.40 0.15 0.13 0.25 0.16 0.93 0.16 0.13 1.o 0.22 0.1 1 0.22 0.30 1.o 0.11 0.21 0.33 0.40 0.47 0.13 0.48 0.37 0.75 0.24 0.55 0.32 0.39 0.39 0.34
1.86 X 10, 3.85 x 103 2.42 X 10, 2.34 X 10' 2.40 x 103 1.92 x 103 >3 x 104 3.30 x 103 1.52 x 103 3-38 x 103 4.70 x 103 >3 x 104 1.54 x 103 3.26 X 10, 5.39 x 103 6.80 x 103 8.30 x 103 1.94 x 103 8.53 x 103 6.12 x 103 1.63 X 10' 3.70 x 103 1.04 X 104 5.24 x 103 6.52 x 103 6.62 x 103 5.62 x 103 2.07 x 5.67 x 2.02 x 7.00 x 1.13 X 6.05 x 1.42 X
102 103 102 103 103 103 103
103 103 103 103 104 103 10'
0.14 0.35 0.14 0.41 0.58 0.36 0.69
0 Obtained from q 5. b After correction for gas-phase axial and radial diffusion, see ref 20. C Obtained from eq 4 which is based on the geometric area of the flow reactor. d Obtained from eq 6.
HOC1
in helium was estimated to be 158 Torr cmz s-l at 188 K.13921 In addition, the reaction probability for reaction 3 was also measured on ice surfaces in the absence of an HC1 coating. 111. Results
+ HCI+
CI,
+ H,O
3
4
lo4
For an irreversiblepseudo-first-orderreaction under plug-flow conditions, the following equation holds for the reactant
In [S(z)] = -k,(z/u) + In [S(O)] whereS is the signal for either HOCl or ClONO2,O is the reference injector position, u is the average flow velocity, and z is the injector position. The corresponding equation for the product signal, assuming rapid desorption and unit stoichiometry, is given by
io3 102
r
10' r 0
ln[S(-) - S(z)] = ln[S(-)] - k,(z/u)
(6) where S(-) is the C12 signal when the reaction has reached completion. The left-hand sides of eqs 5 and 6 were plotted vs the reaction time, z/u, for reactant decay and product growth, respectively. Observed reaction rate constants, k,,were obtained from linear least-squares fits to these data. HOCl+ HCI c 1 ~ H20. The reaction probability for this reaction was measured by observingthe decay of HOCl, monitored at m / e = 52 as a function of the injection position. In a separate run using the same ice film immediately after the HOCl decay measurement, the growth of C12 at m / e = 70 was also monitored as a function of injector position. The thickness of the ice film
-
+
P I"
-1
0
1
2
5
6
T i m (mr) Figure 1. The HOCl dccay and the C12 growth in the reaction of HOCl + HC1- Cl2 + H20 on ice surface. The experimental conditions art Ptow 0.398 Torr, v 1720 cm/s, PHXI= 8.1 X lp7Torr, = 2.1 X 1V Torr, h = 19.9 pm, and T = 188 K.
was 19.4*0.4pm,andthetemperaturewas 1 8 8 A O S K . Typical data are shown in Figure 1 The results from these experiments using PHW~ in the range 1.3 X 1e7-1.8 X 1o-S Torr and PHCIin the range 9 X 1P7-8 X 1o-S Torr is summarized in Table I and I
Heterogeneous Reactions
The Journal of Physical Chemistry, Vol. 97, No. 49, 1993 12801 HOC1
+ HCI -.CI, + H,O (a)
-
0.75
HOC1
0.75
0'50
0.25
0.00'
'
1
6
I
I
2
3
4
I
1
0.75 -
0.251
o,o
',
,
,
1
2
3
4
1{
Q
2
d
0.00 5
(C)
0.75
0.00
t
1
L 1 2 3 5 4
A
A A A
1.5
1.00-
4
(d)
-
0.00
5
+ H C I d CI, + H,O
0.50
-
0.25
-
0.00
3.0
I
4.5
6.0
7.5
I
I
I
9.0
(e)
.
. . . I
I
l'o
0.75
0.50 1
0.00
1.5
P,,, (10.' Torr)
3.0
4.5
6.0
7.5
9.0
PHCl(lo-' Torr)
Figure 2. Summary of the data for the reaction of HOCl + HC1- C12 + H20 on ice surfaces. The data shown in this figure is the average value of the data obtained both from the HOCl decay rates and from the Cl2 growth rates. Partial HOCl pressures are (a) 1.3 X lC7, (b) 2.9 X lo", (c) 5.6 X lt7, (d) 7.3 X lt7(e) , 1.2 X l@, and (f) 1.8 X l@ Torr. See Table I for detailed experimental conditions.
-
CIONO, + HCI CI, + HNO, also shown in Figure 2. The measured reaction probabilities presented in Table I and Figure 2 have been corrected for external gas-phase diffusion.2'J The average reaction probability is 0.30 f 0.18 from the decay rate of HOCl and 0.38 f 0.24 from the CI, growth rate of C12. The overall average value of 41 experiments is ys ( I ) = 0.34f 0.20.The uncertainty represents one standard deviation (la). The results are approximately independent of P~wland PHC~ (see next section for detailed discussion). In addition, the C 4 yield based on HOCl reacted has been measured to be 0.80 f 0.20. Several experiments were also carried out in the absence of an HC1-coating on ice surfaces. The first-order rate constant of the HOCl decay is much smaller than that reported in the previous paragraph. Furthermore, no reaction products were observed in these experiments. ClONO2 HCl C12 + HNO3. The reaction probability for reaction 2was determined in the same manner as that for reaction 1. Typical data are shown in Figure 3. The decay of C10N02 -1 0 1 2 3 4 5 6 was monitored by its major fragment peak m / e = 46 while the Tim. (mol reaction product, Clz, was measured by its parent peak m / e = Figure 3. The ClON02 decay and the Clz growth in the reaction of 70. The concentration of HCl used in the investigation is always ClONOz + HCl- C12 + HNO3 on a surface consisting of ice and NAT. greater than that of ClON02, thus a first-order rate constant can The experimental conditions are P-1 = 0.400Torr, u = 1730 cm/s, be calculated from the observed decay. We have varied P C I O N ~ PCIONO~1.4 X le7Torr, P H C ~ 5.1 X le' Torr, h = 12.5 pm, and T = 188 K. in the range 6.5 X 10-8-9.7 X le7Torr and PHCIfrom 1.6 X 1e7-2.3 X 1od Torr. The ice film thickness was about 10.4 f 1.3 pm, and the temperature was 188 f 0.5 K. These data are We obtain an average yo = 0.26 f 0.19 from the decay rate of summarized in Table I1 and Figure 4. Some of data indicate that ClONO2 and yr = 0.28 f 0.18 from the growth rate of C12. The yl (2)may be greater at higher HCl partial pressures. However, overall average reaction probability of 24 experiments is y8 (2) the overall measured reaction probability yo(2)seems to be nearly = 0.27f 0.19. The uncertainty representsone standard deviation independent of P C I O N and ~ PHC~ after consideringthe uncertainty (lu). In addition, the C12 yield based on the CION02 reacted of the measurements (see next section for detailed discussion). has been measured to be 1.2 f 0.2.
t
+
-
12802 The Journal of Physical Chemistry, Vol. 97, No. 49, 1993
TABLE Ik Reaction hobability for the Reaction of CION02 + HCI
-
Chu et al.
Clz + HN03 on NAT/Ice Surfaces
ClONO2 decay
PCION (Torr) ~ 6.5 X 10-8 6.5 X 10-8 6.6 X 10-8 6.7 X 10-8 6.7 X 10-8 6.9 X 10-8 1.3 x 10-7 1.4 x 10-7 1.4 x 10-7 1.4 x 1v7 1.4 x 10-7 1.5 x 10-7 1.5 x 10-7 1.5 x 1v7 1.5 x 10-7 3.1 x 10-7 3.1 x 1v7 3.2 x 10-7 3.2 x 1v7 3.2 x 10-7 9.6 x 10-7 9.6 x 10-7 9.7 x 10-7 9.7 x 10-7
P H C(Torr) ~ 9.95 x 1v7 7.54 x 1v7 6.67 X lV7 4.36 X lV7 3.29 X lV7 2.82 X lV7 8.12 X l t 7 1.24 X 10-6 1.64 X lk7 2.68 X lV7 5.05 X 1.78 X 1.78 X 3.44 x 3.73 x 4.59 x 1.02 x 6.47 X 1.34 X 8.36 X 9.06 X 1.23 X 1.84 X 2.33 X
thickness (pm) 9.8 9.5 9.9 10.0 10.6 9.9 12.3 10.1 9.6 9.5 12.5 9.8 11.4 12.4 9.7 9.3 9.2 9.4 9.3 9.5 9.7 9.4 9.9 9.3
lk7
lV7 lV7 10-7 10-7 1v7 10-6 lV7 10-6 lW7 lt7 10-6 10-6 10-6
kIE(l/s) 1.48 x 103 6.84 X 102 1.51 x 103 1.10 x 103 9.29 X 102 9.51 X 102 1.03 x 103 1.40 x 103 7.47 x 102 8.23 X 102 8.41 X lo2 8.26 X 102 9.13 X lo2 7.36 X 102 8.75 X lo2 1.00 x 103 1.06 x 103 1.23 x 103 9.44 x 102 8.07 X 102 1.18 x 103 1.11 x 103 1.11 x 103 9.09 X lo2
C12 growth
krb (l/s) 1.45 X 104 1.13 x 103 1.57 X 104 3.09 x 103 2.01 x 103 2.13 x 103 2.66 x 103 9-61 x 103 1.33 x 103 1.61 x 103 1.64 x 103 1.62 x 103 1.95 x 103 1.29 x 103 1.84 x 103 2.50 x 103 3.00 x 103 4.76 x 103 2.21 x 103 1.57 x 103 4.04 x 103 3.29 x 103 3.48 x 103 2.04 x 103
TIC
0.77 0.094 0.81 0.24 0.16 0.17 0.21 0.59 0.1 1 0.13 0.13 0.13 0.16 0.11 0.15 0.20 0.23 0.34 0.18 0.13 0.30 0.25 0.26 0.16
ksd (l/s) 1.40 x 103 1.16 x 103 1.35 x 103 8.15 X lo2 1-44x 103 9.04 x 102 1.37 x 103 1.54 x 103
krb (l/s) 7.57 x 103 3.43 x 103 5.88 x 103 1.52 x 103 8.62 x 103 1.85 X 103 5.90 x 103 1.96 X 104
0.44 0.22 0.36 0.11 0.48 0.13 0.36 0.84
7.79 x 102
1.42 x 103
0.10
7.57 x 9.37 x 1.34 x 1.19 x 9.54 x 1.24 x 8.98 X 9.10 X 1.23 x 9.09 X
1.36 x 2.05 x 7.16 x 3.90 x 2.19 x 4.68 x 1.86 x 1.92 x 4.65 x 4.86 x
0.096 0.14 0.42 0.25 0.15 0.29 0.13 0.13 0.29 0.30
102 102 103 103 102 103 102 102 103 lo2
103 103 103 103 103 103 103 103 103 103
YIC
Obtained from q 5 . After correction for gas-phase axial and radial diffusion, see ref 20. Obtained from q 4 which is based on the geometric area of the flow reactor. Obtained from q 6.
+ HCI-
CIONO,
0.75 '.O0
0.50
0*25
3
1 1
0.00 I
0
P
e L
+ HCI
+
CI,
+ HNO,
: 1
7 I
k
CIONO,
+ HNO,
CI,
0
I
0
I
0
I
I
1
1
i
3
6
9
12
15
0
0
0.50
..
0.25
*
1
0.00 0
5
10
15
20
25
C
! t a
lb' 0.75
-
0.50
-
I
a
0.50 0.75 a
0.25 -
0.25
8.8 I
0.00
I
I
I
1I
0.00
'
0
PHa
Torr)
-
8
.
I
I
I
1
5
10
15
20
P,
8I 1
25
(IO-' Torr)
Figure 4. Summary of the data for the reaction of ClONOz + HCl C12 + HNO3 on NAT/ice surfaces. The data shown in this figure is the average value of the data obtained both from the CION02 decay rates and from the C12 growth rates. Partial ClONO2 pressures are (a) 6.6 X 10-8, (b) 1.4 X lV7, (c) 3.1 X 10-7, and (d) 9.7 X l V 7 Torr. See Table I1 for detailed experimental conditions.
-
+
-
CION02 + H20 HOCl HNO,. We have also remeasured the reaction probability for the reaction of ClONO2 + H20 HOCl + HNO3 on ice surfaces. As noted in the Introduction section, the ice is readily contaminated by the reaction product, HN03,which forms a thin layer of hydrated nitric acid (possibly nitric acid trihydrate, NAT) on the surface. Therefore a very
small reactant concentration of ClONO2 about 1.4 X lo-' Torr was used in these experiments. The results are summarized in Table 111. In every experiment we first measure the ClON02 decay and then measure the growth of the HOCl signal. In general, the data obtained from the HOCl growth is equal to or smaller than that obtained from the ClONO2 decay because of
Heterogeneous Reactions
TABLE 111. Reaction Probability of CION02 + HzO PCIONO~ (Torr) 1.45 X l t 7 1.44 X 1.46 X 1.21 x 1.43 X 1.45 X 1.42 X 1.21 x 1.42 X 1.42 X
lCF7 lCF7 10-7 le7 lt7 lt7 10-7 lt7
thickness (pm) 3.7 4.2 8.6 9.0 9.2 10.6 18.9 24.8 25.7 34.1
Ksa (1 /SI 4.33 x 102 6.78 X 102 7.75 x 102 5.43 x 102 6.50 X lo2 7.03 X lo2 7.35 x 102 8.49 x 102 8.11 X lo2 6.93 X lo2
The Journal of Physical Chemistry, Vol. 97, No. 49, 1993 12803
-
HOCl+ HNO3 on NAT/Ice Surfaces CION02 decay k,b '5.74 x 1.10 x 1.39 x 7.82 X 1.03 x 1.18 x 1.26 x 1.64 x 1.50 x 1.14 x
ksb (l/s) 4.48 X 102 8.91 X lo2 7.25 X lo2 1.34 x 103 1.56 x 103 1.22 x 103 1.58 x 103 1.69 x 103
k,d(l/s) 3.64 X lo2 6.14 X lo2 5.28 X lo2 7.80 X lo2 8.74 X lo2 7.54 x 102 8.83 X lo2 9.16 X lo2
YaC
102 103 103 lo2 103 103 103 103 103 103
HOCl growth
0.05 0.09 0.13 0.07 0.09 0.10 0.10 0.13 0.12 0.10
Yt
0.03 0.06 0.05 0.08 0.10 0.08 0.10 0.10
#Obtained from eq 5 . b After correction for gas-phase axial and radial diffusion, see ref 20. Obtained from eq 4 which is based on the geometric area of the flow reactor. Obtained from eq 6. HOC1
surface deactivation by product HN03. If the geometric area of the flow-tube reactor is used to obtain the reaction probability, ye (3) ranges from 0.03 on thinner ice films to 0.13 on thicker ice films.
IV. Discussion Correction for the Morphology of Ice Films. A model taking surfacereaction and pore diffusioninto account has been described recently.17J8Based on observationsusing environmentalscanning electron microscopy,16 H2O ice films can be described in terms of a model consistingof hexagonallyclose-packed(HCP) spherical granules stacked in layers. In this case the true reaction probability, yt, is related to the value, ys,by '
where 7 is the effectiveness factor, NL = 2 + 9 loglo h (pm) is the number of granule layers, and h is the film thickness. Using this model, we obtain the following true reaction probabilities: yt (1) = 0.13 f 0.08 for reaction 1 and yt (2) = 0.10 f 0.08 for reaction 2. The model includes corrections for external surface roughness and for internal porosity. Roughness is defined as the ratio of total external surface area to the area of the underlying substrate. For the layer model, the surface roughness has a value of about 2. The overall correction to ye for reactions 1 and 2 is about a factor of 2.6; thus, most of the correction is due to the surface roughness. This is consistent with the fact that at high yt (>0.1) gas-phase diffusion of HOCl and CION02 into the porous film cannot compete with reaction at the external surface, and most of the internal film surface does not participate in the rea~ti0n.l~ For Y~ > 0.1, surface roughness still should be considered, even though the mean free path of the reacting gas is greater than the dimensions of the r o ~ g h n e s s . ~Under ~ * ~ flow-reactor ~ conditions, thevelocity distribution directed at the surface is almost random; moreover, after the first collision for yt < 1, the gas molecules leave the surface with completely random velocities and can undergo secondary collisions with the surface. In effect, after the first collision the reacting gas becomes partially trapped within the surface irregularities. Thus, even for long mean free paths almost the entire external surface can be sampled. A similar treatment used to obtain true reaction probabilities for reactions 1 and 2 can be applied to the data (Table 111) for reaction 3. An average value yt (3) = 0.03 was found. This value should be considered as a lower limit because of surface deactivation by product HN03. Comparison with Previous Measurements. In Figure 5 we compareour data of ys (1) with previousmeasurementsofreaction 1 reported by Abbatt and Molina6and Hanson and Ravishankara.8 These data are not corrected for the internal surface area of ice films. Our measurement is in good agreement with these recent data. There is a possibility that a slight negative temperature dependence for reaction 1 may exist. However, within uncer-
+ HCI
-
CI,
+ H,O
I
I
I
k
p
a
e
n v
n
f at
0 0
1 180
This Work anson and Aavishankara
I
0 Abbatt and Molina 190
200
210
Temperatum (K)
220
-
Figure 5. Comparison of the present reaction probability data with
previous measurements for the reaction of HOCl + HCl Clz + H2O on ice surfaces: ( 0 )this work, (A)Hanson and Ravishankara! (0)and Abbatt and Molina.6 See text for details. tainties of these measurements ye (1) should be considered to be nearly independent of temperature. Similarly, our present data on reaction 2, ys (2), is compared with the previous measurements which are shown in Figure 6. Our present measurement is in excellentagreement with the recent data reported by Abbatt and Molina7 and Hanson and Ravishankara.8 Again, ye (2) should be considered to be nearly independent of temperature within uncertainties of these measurements. For reaction 3, surface deactivation by product HN03, which may form hydrates of nitric acid, reduces yt (3). Therefore, this value should be considered as a lower limit. This measurement is consistent with the recent data, ys (3) = 0.2-1.0, reported by Hanson and R a ~ i s h a n k a r a . ~ J ~ - ~ ~ Reaction Mechanism. It has been suggested that reaction 2 may proceed through reaction 3 and then reaction 1.H This is an intriguing question which needs to be resolved. From our measurements reported in the previous section, ye (3) could be smaller than ye (l), and thus, it could serve as a rate limiting step. Furthermore, in our experiments ys (2) is also greater than ye (3). Therefore, although the two-step mechanism is possible, a direct reaction between CION02 and HCl could be more important than the two-step mechanism. Abbatt and Molina7 also reached the same conclusion in their study of these reactions on a NAT surface. Adsorption isotherms have been used to discuss these heterogeneous reactions on ice surfaces2s28 and may be applicable to our experimental results. For reaction 1, since the adsorption equilibrium constant, b ~ mfor ~ HOCl , is smaller than that for
Chu et al.
12804 The Journal of Physical Chemistry, Vol. 97, No. 49, 1993
+ HCI
CIONO, 10 0
-
CI,
+ HNO,
,
I
0
on ice surfaces has been suggested in our previous article14and elsewhere.2+3l This suggestionisvery intriguingbecausemobility of HCI in the liquid is much greater than that in solid ice by several orders of magnitude. Therefore, HCI molecules would be readily available for reaction with HOCl. However, there is a question about the dissociation of HOCl forming HO- and C1+ (or possibly H+ and OCl-) in a quasi-liquid layer. In addition, the existence of this quasi-liquid layer under polar stratospheric conditions has been que~tioned.~~ Further investigationon singlecrystal ice as proposed by Kroes and ClaryZ7is certainly required.
This Work
Hanson and Ravishankara
0 Abban and Molina n
0
Conclusion
0
1 0 180
I
I
I
190
200
210
Temperature (K)
220
-
Figure 6. Comparison of the present reaction probability data with
+
previous measurements for the reaction of ClON02 + HCI C12 HNO3 on NAT/ice surfaces. ( 0 ) this work, (A) Hanson and Ravishankara>JZ and (0)Abbatt and M01ina.~ See text for details.
HCl, bHCl, and P~ml< PHC~, thus the bHmlPHml term can be neglected as compared with the ~ H C ~ Pterm. H C ~Also, HCl molecules on ice surfacesdissociateinto H+ and Cl-.14,29-31Based on the Langmuir-Hinshelwood mechanism, an equation for the relationship between yg and PHCI is given as follows
In summary, the present investigationhas been performed under experimentalconditions which mimic the polar environment, such as, temperature, ice substrate, and reactant concentrations. Therefore, the results are directly applicable to atmospheric modeling. The time constant in converting inactive chlorine to active forms on the basis of our reaction probability measurements and estimated surface area for type I1 PSCs is about a few hours in the wintertime polar stratosphere.32
Acknowledgment. The research described in this article was performed at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the National Aeronautics and Space Administration. References and Notes (1) Scientific Assessment of Ozone Depletion: 1991; World Meteorological Organization Global Ozone Research and Monitoring Project, Report No. 25. (2) Kolb,C. E.; Worsnop,D. R.; Zahniaer,M.S.;Davidovits,P.;Hanson,
D.R.;Ravishankara,A.R.;Keyser,L.F.;Leu,M.-T.;Williams,L.R.;Molina, M. J.; Tolbert, M. A. In Current Problems in Atmospheric Chemistry, Barker,
where yo is the reaction probability on an ice surface at its saturation capacity for HCl. This equation predicts that yg increases linearly with P H C ~at~ /low * HCl pressures and becomes nearly independent of P H C +at~pressures ~ where the surface is saturated with HCl. Our results in Figure 2 (parts a and b) show l predicted by eq 8. that ys increases with PHCIat low P ~ m as However, the results in Figure 2c-f at higher PHWIshow that within experimental uncertainties yg is nearly independent of PHCI; this suggests that during these experiments the ice surfaces are nearly saturated with HCl over the pressure range used in this study. Thus, the Langmuir-Hinshelwood mechanism does not fully account for our results on ice surfaces. It is worthwhile to note that the experimental technique used in this work is not particularly sensitive to yggreater than 0.2. Also, the formation of HCl hexahydrate or melting takes place on ice surfaces at PHCI greater than 1 X le5Torr at 188 K based on our previous measurements.I4 In order to validate this mechanism, more accurate measurements for reaction 1, perhaps using a different experimental approach over a wide range of reactant concentrations, are needed. Similarly, we reach the same conclusion for reaction 2. The results for this reaction are shown in Figure 4a-d. Molina29 has suggested an ionic mechanism for these heterogeneous reactions on ice surfaces; for example, reaction 1 can be depicted as HCl
+ HOC1
-
ioe H+Cl-
HO-Cl’
-
C12(g)
+ H20(s)
(9) In this mechanism a quasi-liquid layer could be formed after the solvation of HC1 on ice surfaces even at stratospheric conditions: T = 180-200 K and PHCI = lO-’ Torr. The dissociation of HCl
J. R., Ed.; World Scientific Publishing Company: In press. (3) Solomon, S. Nature 1990, 347, 347. (4) Prather, M. J. Nature 1992, 355, 534. (5) Crutzen, P. J.; Mlller, R.; Briihl,Ch.;Peter,Th. Geophys. Res. Lett. 1992, 19, 1113. (6) Abbatt, J. P. D.; Molina, M. J. J . Phys. Chem. 1992, 96, 7674. (7) Abbatt, J. P. D.; Molina, M. J. Geophys. Res. Lett. 1992, 19, 461. (8) Hanson, D. R.; Ravishankara, A. R. J . Phys. Chem. 1992,96,2682. (9) Molina, M. J.; Tso,T. L.; Molina, L. T.; Wang, F. C. Y.Science 1987, 238, 1253. (10) Tolbert, M. A.; Rossi, M. J.; Malhotra, R.; Golden, D. M. Science 1987, 238, 1258. (11) Leu, M.-T. Geophys. Res. Lett. 1988, 15, 17. (12) Hanson, D. R.; Ravishankara, A. R. J . Geophys. Res. 1991,96,5081. (13) Leu, M.-T.; Moore, S. B.;Keyser, L. F. J . Phys. Chem. 1991, 95, 7763. (14) Chu, L. T.;Leu, M.-T.; Keyser, L. F. J . Phys. Chem. 1993,97,7779. (15) Keyser, L. F.; Leu, M.-T. J . Colloid Interface Sci. 1993,155, 137. (16) Keyser, L. F.; Leu, M.-T. Microscopy Res. Technique 1993,25,434. (17) Keyser, L. F.; Moore, S. B.; Leu, M.-T. J. Phys. Chem. 1991, 95, 5496. (18) Keyser, L. F.; Leu, M.-T.; Moore, S. B. J . Phys. Chem. 1993, 97, 2800. (19) D’Ans, J.; Freund, H. E. Z . Elekrfochem. 1957,61, 10. 1978,81, 1. (20) Brown, R. L. J . Res. Natl. Bur. Stand. (U.S.) (21) Marrero, T. R.; Mason, E. A. J . Phys. Chem. Re/. Data 1972,1,3. (22) Takaishi, T. J . Phys. Chem. 1957, 61, 1450. (23) McKee, C. S.; Roberts, M. W. Trans. Faraday Soc. 1%7,63,1418. (24) Hanson, D. R.; Ravishankara, A. R.J . Phys. Chem. 1993,97,2802. (25) Elliot, S.; Turco, R. P.; Toon, 0. B.: Hamill, P. J . Atmos. Chem. 1991, 13, 211. (26) Mozurkewich, M. Geophys. Res. Lett. 1993, 20, 355. (27) Kroes, G.-J.; Clary, D. C. J . Phys. Chem. 1992,96, 7079. (28) Tabazadeh, A.; Turco, R. P. J. Geophys. Res. 1993,98, 12727. (29) Molina, M. J. In CHEMRAWN VII Chemistry of the Atmosphere: The Impact of Global Change;Calvert,J. G., Ed.; Blackwell Sci. Publ.: Oxford, In press. (30) Hanson, D. R.; Mauersberger, K. Geophys.Res. Letf.1988,15,1507. (31) Hanson, D. R.; Mauersberger, K. J . Phys. Chem. 1990, 94, 4700. (32) Turco,R. P.; Toon, 0.B.; Hamill, P. J. Geophys. Res. 1989, 94, 16493.
