http://www.fuelcellknowledge.

org/

Mean Free Path

The mean free path of molecules in a gas is a vital parameter for determining which
physical phenomena are most important in determining its behaviour. The mean free
path is the average distance molecules travel between collisions with other molecules
in the gas. As a general guideline, macroscopic/bulk/continuum analysis cannot be
applied to situations in which the length scale (e.g. diameter of pore, thickness of
boundary layer) is of the same order as (or lower than) the mean free path.

The mean free path can be estimated by considering the number density (number of
molecules of gas per unit volume) and the volume ‘swept out’ by a molecule per unit
time as it travels. Consider the following situation:

All molecules in a single component gas are frozen in space except for one molecule,
_
which is travelling through the gas at mean speed C . When two molecules collide,
the distance between their centres is d. Hence, this molecule will collide with any
molecule (the centre of) which lies within a cylinder of diameter 2d swept out as it
_
moves. In time dt the molecule moves a distance C dt, sweeping out a volume
π ⋅ d 2 ⋅ C ⋅ dt . If the number density of molecules in the gas is n, then the number of
collision per unit time per unit volume – the ‘collision frequency’ z – will be:

z = π ⋅ n ⋅ d 2 ⋅C (1)

Now, consider each molecule to be moving (with all the others stationary) in turn, the
total molecular collision rate will be:

π ⋅ n2 ⋅ d 2 ⋅ C
Z= (2)
2

Where the factor of ½ avoids counting each collision twice.

If this analysis is repeated assuming that all molecules are simultaneously moving and
_
have a Maxwellian distribution of velocities (mean speed C ) then the same
expression for Z is obtained, but with a factor of 2 . Hence (derivation to follow):

_
2 ⋅π ⋅ n2 ⋅ d 2 ⋅ C
Z= (3)
2

To determine the distance travelled between collisions we must divide the mean
molecular velocity by the collision frequency. Note however that we must do this for
the collision frequency of one molecule (divide Z by n and multiply by 2 to give z),
giving the mean free path as:

1
 1
λ= (4)
2 ⋅π ⋅ n ⋅ d 2

This analysis can be extended to mixtures of two gases (assuming both to have a
Maxwellian distribution of velocities) as follows:

The collision rate per unit volume between like molecules is given by (3) and the
collision rate between unlike molecules given by:

_
Z = 2 ⋅ π ⋅ n A ⋅ nB ⋅ d AB ⋅ C AB
2
(5)

Where,

d AB =
(d A + d B ) (6)
2

The difficulty in the case of unlike molecules is that the mean speeds of the different
gases are not the same, and hence the speed does not cancel from the calculation of
mean free path. Expressions for mean molecular speed can be derived from the
Maxwell-Boltzman distribution, in terms of temperature and molecular weight:

_
C=
(8kT ) (7)
π ⋅m

For mixtures of gases, m must be replaced by mAB where:

m AmB
m AB = (8)
(m A + m B )
However as mAB reduces to m/2 when mA=mB, the factor of 8 in the mean speed
expression is reduced to 4, thus:

_
C AB =
(4kT ) (9)
π ⋅ m AB

The collision rate of gas A molecules is given by ZAA+ZAB, and hence the collision
rate of a single gas A molecule given by:

2 × Z AA Z AB
zA = + (10)
nA nA

(ZAB does not need to be multiplied by 2 as the expression is for unlike molecules and
hence the number of collisions was not divided by 2). Substituting expressions for
ZAA, ZAB and mean molecular speed in (10) gives:

2
 8kT 4kT
z A = 2 ⋅π ⋅ nA ⋅ d A ⋅ + 2 ⋅ π ⋅ nB ⋅ d AB ⋅
2 2
(11)
π ⋅ mA π ⋅ m AB

Finally, dividing the mean speed of molecules of gas A by the collision rate of a
single molecule of gas A gives the mean free path of gas A in the mixture of gases A
and B:

1
mA
zA = (12)
1 1
2 ⋅π ⋅ nA ⋅ d A + ⋅π ⋅ nB ⋅ d AB
2 2

mA m AB

Note that this expression reduces to equation (4) when nB is zero, and when mA=mB
and dA=dB.

Calculations & Required Data

The expressions derived above are useful for understanding the origin and
dependencies of the mean free path, but are in fact of little help for calculating it, as
they involve the molecular diameter which is not a directly measurable quantity. The
Molecular diameter is typically determined from viscosity measurements. Hence, the
calculation of mean free path is most straightforward using the kinetic theory
expression for viscosity (derivation to follow):
_
ρCλ
µ= (13)
2

_
Where, ρ = n ⋅ m , λ is the mean free path (m), C the mean molecular speed (m/s),
µ is the viscosity (kg/ms) and rigorous Chapman-Enskog theory gives a coefficient
of 0.499 rather than 0.5. An experiment using this method is described at Georgia
Institute of Technology.

Whichever method is used molecular mass and number density values are required:

Molecular Mass data is widely available, see Materials section.

Number Density can be calculated using the Ideal Gas Law, see Thermodynamics
section.

Viscosity data is available for most SOFC gases at SOFC operation conditions, see
Materials section.
n N nm N a N p p
n= m a = = a =
V nm RT p RT kT

3
Where, n is the number density, nm the number of moles of gas present, V the volume
of gas present, Na Avogadros Number, p the gas pressure, R the Universal Gas
Constant, k Boltzmans Constant and T the gas temperature (K).

0.9
(N2) (O2) (H2) (H2O) (CO) (CO2) (CH4)
0.8

0.7
Mean Free Path (microns)

0.6

0.5

0.4

0.3

0.2

0.1

0
0 200 400 600 800 1000 1200 1400 1600
Temperature (K)

Figure 1. Mean Free Path Variation of SOFC Gases

References:

[1] J.B. Young, Molecular Thermodynamics Course Notes, Cambridge University

Engineering Department, 1999-2000.
[2] B. Todd, J.B. Young, Thermodynamic & Transport Properties for SOFC
Modelling, Journal of Power Sources 110 (2002) pp.186-200.

You might also want to visit HyperPhysics (http://hyperphysics.phy-

astr.gsu.edu/hbase/hframe.html)– gives a good general description of Mean Free Path
and associated theories.

 Ben Todd 2003

Last Updated: November 28, 2003

This document is from the Fuel Cell Knowledge website, which presents information on the research and development of fuel
cells and renewable energy systems. The focus is on Solid Oxide Fuel Cells (SOFCs) with particular emphasis on
thermodynamics, materials & gas properties and SOFC computational modelling. Material presented has arisen from my PhD at
Cambridge University Engineering Department.