IUPAC Subcommittee on Gas Kinetic Data Evaluation

(http://www.iupac-kinetic.ch.cam.ac.uk)

Introduction to Heterogeneous Reactions

The term "heterogeneous" as applied to the atmosphere refers to multiphase chemistry

that occurs on or in condensed phases which are in contact with the gas phase. Examples of
condensed phases found in the atmosphere are liquid aerosols, such as sulphuric acid, sea
spray and cloud droplets, and solid atmospheric particulate such as soot, mineral dust, salts or
frozen water-containing particles such as cirrus clouds and some polar stratospheric clouds
(PSC's). Heterogeneous reactions taking place at the gas-solid or gas-liquid interface are
complex and are made up of several elementary physico-chemical processes such as gas phase
diffusion towards the surface of interest, the mass accommodation rate, the desorption rate
from the surface back to the gas phase, the chemical reaction rate at the interface, the diffusion
into the bulk of the condensed phase and possible chemical reactions within the bulk condensed
phase 1.

This compilation presents kinetic data for the uptake of atmospheric trace gas species
resulting from physical transfer or chemical reaction in/on a condensed phase. Rates are
expressed in terms of dimensionless uptake and accommodation coefficients which are defined
below. The primary measurement in most experimental techniques of heterogeneous kinetics is
the net rate of uptake of a trace gas species and in some cases the rate of formation of one or
several product species in a system of defined surface to volume ratio. The advantage of
expressing the kinetics in terms of uptake coefficients lies in the fact that the results are
transferable from one experiment to another, even if those experiments involve significant
differences in the surface-to-volume ratio. This would not be the case if kinetic results were
expressed in terms of rate coefficients which are however the primary experimental observables
in gas uptake experiments. The rate of uptake is subsequently analyzed and interpreted in terms
of the contributing rate processes outlined above which are now summarized in more detail.

• Mass accommodation coefficient. This parameter, α, is a measure of the ratio of the

rate at which the molecules cross the interface region between the condensed phase and the gas
phase normalized to the gas collision rate at the interface between the phases:

α = number of gas molecules entering the condensed phase in unit time (1)
number of gas molecules striking the interface in unit time

In this definition α is defined as a mass transport parameter which has been demonstrated
through measurement in the condensed phase of species transferred from the gas phase. The
parameter α has often been called the sticking coefficient when obtained in single collision
experiments using a molecular beam impinging on a solid surface. Because solid phase

-1-
IUPAC Subcommittee on Gas Kinetic Data Evaluation
(http://www.iupac-kinetic.ch.cam.ac.uk)

diffusion is often slow, thermal accommodation followed by physical or chemical adsorption

are the processes taking place. The use of the term sticking coefficient in relation to the rate of
uptake processes resulting from gas diffusion experiments is not recommended.
Unfortunately, results from molecular beam and conventional gas uptake experiments such as
the ones performed on droplet trains cannot be directly compared to date in order to be tested
for self-consistency, partly because the mass accommodation coefficient is thought to be
strongly dependent on the temperature and somewhat dependent on composition for sulfuric
acid solutions, particularly at high acid weight fractions 2.

The uptake coefficient, γ , is a dimensionless parameter defined in a similar way to α,

but reflecting the overall net rate of uptake including both physical and chemical processes in
the gas and condensed phases as well as at the interface; γ is thus defined in terms of the net
rate of disappearance of a species from the gas phase in the presence of a condensed phase:

γ = number of gas molecules removed by the condensed phase in unit time (2)
number of gas molecules striking the interface in unit time

Uptake into liquid droplets

The mass accommodation coefficient α determines the maximum flux J (molecule cm-2
s-1) of gas into a liquid, which is given by

J = ([M]g <c> α) / 4 (3)

where [M]g is the trace gas concentration (molecule cm-3) and <c> is the trace gas average
thermal velocity (cm s-1). The magnitude of α determines the maximum rate of mass transport
across the gas-condensed phase interface region and may be equal to the uptake coefficient γ
measured under certain limiting conditions, for instance when evaporation of the molecule back
into the gas phase may be excluded owing to fast chemical reaction or high solubility of M.

Determinations of α from measurements of uptake rates of trace gases onto bulk liquid
droplets yield values ranging from approximately 1 to ~10-2, sometimes even smaller. A
negative temperature dependence of α has often been observed and this has been related to the
thermodynamics of interfacial complex formation3. Davidovits and coworkers have formulated
a quantitative chemical kinetic model of mass accommodation according to the following
scheme 1,3:

Mg <===> Ms ➞ Ms* ➞ Ml (kads, kdes, ks, ksol) (4)

-2-
IUPAC Subcommittee on Gas Kinetic Data Evaluation
(http://www.iupac-kinetic.ch.cam.ac.uk)

where Mg, Ms, Ms* and Ml represent the molecule in the gas, adsorbed, activated (transition
state) and dissolved states, respectively. This scheme presents a coherent framework within
which mass accommodation rates may be understood in terms of the structure of the solute
molecule interacting with an aqueous solution interface using classical nucleation theory which
describes the incorporation of critical size clusters consisting of trace gas surrounded by solvent
molecules into the bulk liquid. The simple kinetic scheme presented in equation (4) is
consistent with the existence of a precursor (Ms) stabilized with respect to the gas phase (Mg).
A similar precursor-mediated adsorption mechanism has been proposed to explain for instance
the mechanism and the kinetics of D2O vapor uptake on D2O-ice at temperatures of 140 to
220K4. In general, all uptake rates for which a negative temperature dependence has been
experimentally observed are expected to follow this precursor-mediated uptake mechanism
given in equation (4).

The rate of uptake of trace gas onto a fluid is expressed by the overall uptake coefficient
γ , which includes contributions from different processes such as gas diffusion, mass
accommodation, reevaporation, bulk reaction and dissolution. The total rate may be calculated
according to a resistance model which is based on steady state solutions to the decoupled
differential equations describing each separate process by itself5. The formulation is based on
approximations to the exact solutions and the overall net uptake rate coefficient is expressed in
terms of the sum of resistances which are the inverse of the dimensionless uptake rate
coefficients, Γi, of the separate processes involved:

1/γ = 1/Γg + 1/α + 1/(Γsol + Γrxn) (5)

In the above expression 1/Γ g, 1/Γsol and 1/Γrxn are the resistances corresponding to gas
diffusion of the trace gas to the interface, the competing (non-reactive) dissolution following
Henry's law and chemical reaction in the condensed phase, respectively. In order to calculate
the flux of trace gas across the interface one simply replaces γ for α in expression (3). The
resistance model is a useful approximation which is usually sufficiently accurate. However, in
the case of an interfacial reaction of a molecule which is only sparingly soluble and which
slowly reacts in the condensed phase, chemical reaction could take place on the same time scale
as gas dissolution in the interface region so that the third term on the right hand-side in equation
(5) will not be accurate. A more complete treatment of interfacial reactions including the
formation of interfacial complexes has recently been described by Shi et al.6.

• Gas Diffusion. Gas diffusion of the trace gas towards the interface is often reduced
from its maximum value, the Boltzmann gas kinetic flux given by [M]g<c>/4 owing to the

-3-
IUPAC Subcommittee on Gas Kinetic Data Evaluation
(http://www.iupac-kinetic.ch.cam.ac.uk)

build-up of a concentration gradient from the interface to the bulk gas phase. The gas transport
resistance 1/Γg is approximated in expression (6):

1/Γg = <c>d/8Dg - 1/2 (6)

where d is a characteristic geometrical dimension such as the aerosol diameter in aerosol uptake
experiments or the diameter of the flow tube in bulk uptake studies using laminar coated-wall
flow tubes, and Dg is the molecular diffusion coefficient of the trace gas. The first term in
expression (6) is obtained from the steady state solution for gas phase diffusion to a sphere of
diameter d, appropriate for liquid droplets. The term -1/2 in equation (6) is important when gas
uptake distorts the equilibrium Boltzmann distribution in the proximity of the interface and
results in γ = 2 for α = 1 when all other resistances are negligible7. This negative resistance
term (-1/2) is an approximation of the fact that owing to uptake at the interface there is a
shortage of molecules with velocity vectors pointing away from the gettering surface. This
term may be important in cases where fast rates of uptake are studied in the laboratory7. A
quantitative approach based on Monte-Carlo trajectories of individual molecules interacting with
a condensed phase sample and taking into account the appropriate reactor geometry has been
used successfully to estimate the departure from a thermal distribution of velocities in a
heterogeneous reaction system8. In addition, 1/Γg is often underestimated when uptake rates
on aerosols are evaluated because the smallest size fraction contributes more heavily by virtue
of the dependence of 1/Γg on d and because of the larger surface-to-volume ratio of the small
size fraction 1,9.
• Solubility Limited Uptake. Many trace gases are only slightly soluble in the
condensed phase. In this case the net rate of uptake slows down markedly with time owing to
reevaporation of molecules which were initially taken up - the saturation effect. At long
exposure times equilibrium will be reached where the rate (flux) of uptake is balanced by the
rate (flux) of evaporation. The approach to saturation in the liquid may be described by Γ sol,
the solubility limited uptake coefficient:

Γsol = (4HRT/<c>π1/2)(Dl/t)1/2 (7)

where H is Henry's law constant expressed in M-1 atm-1 and Dl (cm2 s-1) is the diffusion
coefficient of the trace gas in the fluid at time t after initial exposure. At equilibrium the entire
condensed phase is saturated with the trace gas. Net uptake is zero, and the gas flux entering
the interface region (controlled by α) and the flux leaving it are equal. For non-equilibrium
conditions, Γsol accounts for the rate of evaporation which is zero at the beginning of the
uptake and becomes progressively larger with time.

-4-
IUPAC Subcommittee on Gas Kinetic Data Evaluation
(http://www.iupac-kinetic.ch.cam.ac.uk)

• Reaction Limited Uptake. When the dissolved gas irreversibly reacts in the condensed
phase a reactive uptake coefficient may be defined as follows:

Γrxn = (4HRT/<c>)(Dlkrxn)1/2 (8)

where krxn is the first-order rate constant for reaction in the condensed phase. Thus for steady
state uptake Γrxn is proportional to the product (kII[M]l)1/2 where kII is the bimolecular rate
constant for liquid phase reaction with [M]l being the liquid phase reactant concentration.
Equation (7) and (8) are analogous with respect of the equivalence of t in equation (7) with
1/krxn in equation (8). The interfacial reaction takes place in a surface layer whose thickness is
approximately equal to (Dl/krxn)1/2, the reacto-diffusive length.

Uptake on Solids
Uptake on solids is in principle similar to uptake on liquids except that the diffusion
coefficient for the condensed phase (Dl) is several orders of magnitude smaller than the
corresponding value for a liquid. Consequently, uptake on solids is prone to saturation
phenomena once the surface has been covered or has reacted, unless the interface is modified so
as to generate new adsorption/reaction sites. The uptake kinetics is described in terms of the
uptake coefficient γ, as given in equation (2).

The uptake coefficient is defined in terms of equation (9) or (10) depending on whether
the rate law for adsorption of the gas species Mg is known (equation (9)) or not (equation (10)):

γ = kup / ω with ω = As <c>/4V (9)

γ = 4VRup / As <c> [M]g (10)

where kup is the experimentally determined uptake rate constant (s-1), Rup the rate of uptake of
the reactant gas defined as d[M]g/dt (molecule cm-3 s-1), ω the gas-surface collision frequency
in s-1, [M]g the concentration of the molecular species in the gas phase, V and As the volume
of the reactor and the reactive sample surface, respectively. The rate law on solids, such as
soot, salt, ices or mineral dust, can be complex and vary from one sample to the next,
depending on experimental conditions such as mode of generation, internal surface, humidity,
aging, exposure history, presence of light , etc.

-5-
IUPAC Subcommittee on Gas Kinetic Data Evaluation
(http://www.iupac-kinetic.ch.cam.ac.uk)

Data in the Tables

Definitions

The uptake coefficient or reaction probability γ listed in the Tables is presented in

different forms depending on the experimental technique used:
• γ is an uptake coefficient calculated from the measured loss rate coefficient after
correction for gas phase diffusion, liquid phase saturation in relation to limiting Henry’s law
solubilities and condensed phase diffusion, and assumptions regarding the surface area of the
sample, using equation (9).
• γ ss describes an experimentally determined uptake coefficient obtained from
measurements at steady state when γ does not change with time on the time scale of the
particular experiment.
• γ0 is an experimentally determined uptake coefficient in the initial stages after exposure
of the gas to a fresh surface. This may be obtained through extrapolation of time dependent γ
values back to t=0 or it corresponds to an uptake coefficient resulting from a time-resolved, that
is pulsed dose experiment. If gas diffusion is not rate-limiting γ 0 is thought to closely
approximate the mass accommodation coefficient α.
• γ obs , γmax correspond to the observed value of the uptake coefficient under defined
experimental conditions of partial pressure or time and to its maximum value, respectively.
• γX corresponds to the reactive uptake coefficient of species X in cases where more than
one gas phase species has been exposed to the surface of interest.

Independent information on Henry's law solubility of the reactant (H) or its diffusion
coefficient in the condensed phase (Dl) under the relevant experimental conditions are required
to separate the physico-chemical constants in the condensed phase from the expressions
H(Dl/t)1/2 (equation (7)) or H(Dlkrxn)1/2 (equation (8)).

Condensed Phases

For mixtures of H2SO4/H2O, HNO3/H2O and saturated ternary solutions (STS)

containing HCl the composition is given in weight per cent (%H2SO4, %HNO3, %HCl), the
concentrations are given in mol l-1; rh means relative humidity in per cent based on the partial
pressure of H2O in the gas phase relative to that over condensed phase water, even when solid
substrates other than pure water ice are used. It is given by the following expression:

rh(%) = 100 p(H2O) / pice(T) (11)

-6-
IUPAC Subcommittee on Gas Kinetic Data Evaluation
(http://www.iupac-kinetic.ch.cam.ac.uk)

where pice(T) corresponds to the vapor pressure of ice at temperature T. In some cases
experimental results are listed without explicit knowledge of the nature and composition of the
substrate. NAT, NAD and SAT mean nitric acid trihydrate (HNO3•3H2O), nitric acid
dihydrate (HNO3•2H2O) and sulfuric acid tetrahydrate (H2SO4•4H2O), respectively.
Additional hydrates are listed by name rather than by their acronyms.

Surface area and Porosity

In order to obtain the true uptake coefficient on a per collision basis the sample
morphology or sample presentation has to be taken into account to calculate the collision
frequency, ω. Samples may be presented as single crystals, monodisperse grains, thin solid
films or polycrystalline powders. The geometric surface area of the bulk sample is usually used
to calculate ω for the reactant gas in flow reactor experiments. However if the sample is porous
the reactant may penetrate into the bulk of the sample and undergo many more collisions with
the internal surface of the sample compared to the calculated gas kinetic collision frequency
based merely on the external surface of the sample. An adaptation of pore diffusion theory
quantitatively expresses this enhancement in ω as a function of material and structural
parameters of the powder. The magnitude of the correction depends on the pore size and
geometry as well as on the magnitude of the uptake coefficient10. Full validation of this theory
has so far not been achieved with experimental data for atmospheric sustrates10,11.

• Surface Reactions on gas/liquid Interface

The bulk of the uptake studies on aqueous solutions so far undertaken fits into the
scheme that chemical reaction takes place in the bulk fluid condensed phase and that the
interfacial component is restricted to mass transfer across the gas-condensed phase interface.
However, there are distinct cases where the interpretation of uptake experiments involve surface
mediated chemical reaction akin to gas-solid reaction systems which exclusively favor such
surface-mediated reactivity which manifests itself whenever the affinity of one or both
components towards the condensed phase is weak12. Examples include interfacial reactions of
Cl213, Br213, SO214, ClONO215, ClNO216, CH2O17,18 and other aldehydes18,
gloyoxal19 as well as HONO in its reaction with HCl on ice20. In most cases the nature of the
surface complex preceding the chemical reaction is unknown, and in the cases of the aldehydes
a surface protonated species has been invoked.

-7-
IUPAC Subcommittee on Gas Kinetic Data Evaluation
(http://www.iupac-kinetic.ch.cam.ac.uk)

References

1 C. E. Kolb, D. R. Worsnop, M. S. Zahniser, P. Davidovits, L. F. Keyser, M.-T. Leu,

M. J. Molina, D. R. Hanson, A. R. Ravishankara, L. R. Williams and M. A. Tolbert,
"Laboratory Studies of Atmospheric Heterogeneous Chemistry", in Progress and
Problems in Atmospheric Chemistry (J. R. Barker, ed.), Adv. Ser. Phys. Chem., 3,
771, World Sci., River Edge, N. J., 1995.
2 G. M. Nathanson, P. Davidovits, D. R. Worsnop and C. E. Kolb, J. Phys. Chem .
100, 13007 (1996).
3 P. Davidovits, J.T.Jayne, S.X.Duan, D. R. Worsnop, M. S. Zahniser and C. E. Kolb,
J.Phys.Chem., 95, 6337 (1991).
4 L. Chaix, H. van den Bergh and M. J. Rossi, J. Phys. Chem. A 102, 10300 (1998).
5 P. V. Danckwerts, Trans. Faraday Soc. 47, 1014 (1951).
6 Q.Shi, P Davidovits, J.T.Jayne, D.R.Worsnop and C.E.Kolb, J.Phys.Chem.. A 103,
8812, (1999).
7 H. Motz and H. Wise, J. Chem. Phys. 32, 1893 (1960).
8 F. F. Fenter, F. Caloz and M. J. Rossi, Rev. Sci, Instrum. 68, 3180 (1997).
9 D. R. Hanson, A. R. Ravishankara and S. Solomon, J. Geophys. Res. 99D, 3615
(1994).
10 L. F. Keyser, M.-T. Leu and S. B. Moore, J. Phys. Chem. 97, 2800 (1993).
11 D. R. Hanson and A. R. Ravishankara, J. Phys. Chem. 97, 2802 (1993).
12 D. R. Hanson, J. Phys. Chem . B 101, 4998 (1997).
13 J.H.Hu, Q.Shi, P.Davidovits, D.R.Worsnop, M.S.Zahniser and C.E.Kolb,
J.Phys.Chem. 99, 8768, (1995).
14 D. R. Worsnop, M. S. Zahniser, C. E. Kolb, J. A. Gardner, L. R. Watson, J. M. Van
Doren, J. T. Jayne and P. Davidovits, J. Phys. Chem. 93, 1159 (1989).
15 D. R. Hanson and A. R. Ravishankara, J. Phys. Chem. 98, 5728 (1994).
16 Ch. George, W. Behnke, V. Scheer, C. Zetzsch, L. Magi, J. L. Ponche and Ph.
Mirabel, Geophys. Res. Lett. 22, 1505 (1995).
17 M. A. Tolbert, J. Pfaff, I. Jayaweera and M. J. Prather, J. Geophys. Res. 98, 2957
(1993).
18 J. T. Jayne, S. X. Duan, P. Davidovits, D. R. Worsnop, M. S. Zahniser and C. E.
Kolb, J. Phys. Chem. 96, 5452 (1992).
19 F. Schweitzer, L. Magi, Ph. Mirabel and Ch. George, J. Phys. Chem. A 102, 593
(1998).
20 F. F. Fenter and M. J. Rossi, J. Phys. Chem. 100, 13765 (1996).

-8-