FOCUS 16

Molecules in motion

This Focus is concerned with understanding how matter and resistance of electrolyte solutions and to analyse it in terms of
other physical properties (such as energy and momentum) the response of the ions to an applied electric field.
are transported from one place to another in both gases and 16B.1 Experimental results; 16B.2 The mobilities of ions
­liquids.

16C Diffusion
16A Transport properties of The diffusion of solutes and various physical properties is an
a perfect gas important process in liquids. It can be discussed by introduc-
ing the concept of a general ‘thermodynamic force’ that can
The transport of matter and physical properties can be be regarded as being responsible for the motion of molecules.
­described by a set of closely related empirical equations. For This apparent force can be used to construct the important
a gas, it is possible to understand the form of these equations ‘diffusion equation’, which describes how solutes spread out in
by building a model based on the kinetic theory of gases dis- space with increasing time. An alternative model of diffusion
cussed in Topic 1B. With this approach, the rate of diffusion, as a random walk gives further insight.
the rate of thermal conduction, viscosity, and effusion can all 16C.1 The thermodynamic view; 16C.2 The diffusion equation;
be related to quantities arising from the kinetic theory. 16C.3 The statistical view
16A.1 The phenomenological equations; 16A.2 The transport
parameters

Web resource What is an application

of this material?
16B Motion in liquids
A great deal of chemistry, chemical engineering, and b
­ iology
Molecular motion in liquids is different from that in gases on depends on the ability of molecules and ions to migrate
account of the presence of significant intermolecular interac- through media of various kinds. Impact 25 explains how con-
tions and the much higher density typical of a liquid. One way ductivity measurements are used to analyse the motion of ions
to monitor motion in such systems is to explore the electrical through biological membranes.
TOPIC 16A Transport properties of
a perfect gas

The parameter σ is the collision cross-section of a molecule,

➤ Why do you need to know this material? a measure of the target area it presents in a collision. For the
derivation and physical interpretation of this expression,
Many physical processes take place by the transfer of
see Topic 1B. Another important result from kinetic theory is
a property from one region to another, and gas-phase
the mean speed of molecules of molar mass M at a tempera-
chemical reactions depend on the rate at which mol-
ture T, which is given by eqn 1B.9:
ecules collide. The material presented here also includes
general aspects of transport in fluid systems of any kind, 1/2
 8 RT  Mean speed
and which are applicable to reactions taking place in vmean =  
 πM  [kinetic theory] (16A.1b)
solution.

➤ What is the key idea?

A molecule carries properties through space in steps of The phenomenological
16A.1
about the distance of its mean free path.
equations
➤ What do you need to know already?
A ‘phenomenological equation’ is an equation that summa-
This Topic builds on and extends the kinetic theory of
rizes empirical observations on phenomena without, initially
gases (Topic 1B). You need to be familiar with the concepts
at least, being based on an understanding of the molecular
of the mean speed of molecules and the mean free path
processes responsible for the property. Such equations are
and their dependence on pressure and temperature.
­encountered commonly in the study of fluids.
The rate of migration of a property is measured by its flux,
J, the quantity of that property passing through a given area in
a given time interval divided by the area and the duration of
A transport property is a process by which matter or an at- the interval. If matter is flowing (as in diffusion), the matter
tribute of matter, such as momentum, is carried through flux is reported as so many molecules per square metre per
a medium from one location to another. The rate of trans- second (number or amount m−2 s −1). If the property migrating
port is commonly expressed in terms of an equation that is is energy (as in thermal conduction), then the energy flux is
an empirical summary of experimental observations. These expressed in joules per square metre per second (J m−2 s −1), and
equations apply to all kinds of properties and media, and so on. The total quantity of a property transferred through
can be adapted to the discussion of transport properties of a given area A in a given time interval Δt is | J |AΔt. The flux
gases. In such cases, the kinetic theory of gases provides sim- J may be positive or negative: the significance of its sign is
ple expressions that show how the rates of transport of these ­discussed below.
properties depend on the pressure and the temperature. The Experimental observations on transport properties show
most important concept from the kinetic theory developed that the flux of a property is usually proportional to the first
in Topic 1B, and used throughout this Topic, is the mean derivative of a related quantity. For example, the flux of mat-
free path, λ , the average distance a molecule travels between ter diffusing parallel to the z-axis of a container is found to
collisions. According to eqn 1B.14, at a temperature T and a be proportional to the gradient of the concentration along the
pressure p same direction:

kT Mean free path dN

λ=  (16A.1a) J (matter) ∝  Fick’s first law of diffusion (16A.2)
σp [kinetic theory] dz
 16A Transport properties of a perfect gas 691

where N is the number density of particles, with units number Table 16A.1 Transport properties of gases at 1 atm*
per metre cubed (m−3). The proportionality of the flux of mat-
κ/(mW K−1 m−1) η/μP‡
ter to the concentration gradient is sometimes called Fick’s
273 K 273 K 293 K
first law of diffusion: the law implies that diffusion is faster
when the concentration varies steeply with position than Ar 16.3 210 223
when the concentration is nearly uniform. There is no net flux CO2 14.5 136 147
if the concentration is uniform (dN /dz = 0). Similarly, the rate He 144.2 187 196
of thermal conduction (the flux of the energy associated with N2 24.0 166 176
thermal motion) is found to be proportional to the tempera-
* More values are given in the Resource section.
ture gradient: ‡
1 μP = 10 −7 kg m−1 s −1.

dT
J (energy of thermal motion) ∝  Flux of energy (16A.3)
dz
Brief illustration 16A.1
A positive value of J signifies a flux towards positive z; a
negative value of J signifies a flux towards negative z. Because Suppose that two metal plates are placed perpendicular to the
matter flows down a concentration gradient, from high con- z-axis, with the first at z = 0 and the second at z = +1.0 cm, and
centration to low concentration, J is positive if dN /dz is nega- that the temperature of the second plate is 10 K higher than
tive (Fig. 16A.1). Therefore, the coefficient of proportionality that of the first plate. The temperature gradient is
in eqn 16A.2 must be negative, and it is written as −D:
dT 10K
= = + 1.0 ×103 K m −1
d z 1.0 × 10−2 m
dN Fick’s first law in terms of
J (matter) = − D  (16A.4)
dz the diffusion coefficient
If the plates are separated by air, for which κ = 24.1 mW K−1 m−1,
the energy flux is
The constant D is the called the diffusion coefficient; its SI
units are metre squared per second (m2 s−1). Energy migrates dT
down a temperature gradient, and the same reasoning leads to J = −κ = − (24.1mWK −1 m −1 ) × (1.0 ×103 K m −1 )
dz
= − 24 Wm −2
dT
J (energy of thermal motion) = −κ 
dz The flux is negative because energy is transferred from the
 Flux of energy in terms of the (16A.5) hotter plate at z = +1.0 cm to the cooler plate at z = 0, which is
coefficient of thermal conductivity
a flow in the −z direction. The energy transferred through an
area of 1.0 cm2 between the two plates in 1 h (3600 s) is
where κ (kappa) is the coefficient of thermal conductiv-
ity. The units of κ are joules per kelvin per metre per second Transfer = | J |AΔt = (24 W m−2) × (1.0 × 10−4 m2) × (3600 s) = 8.6 J
(J K−1 m−1 s−1) or, because 1 J s−1 = 1 W, watts per kelvin per
metre (W K−1 m−1). Some experimental values are given in
Table 16A.1. Viscosity arises from the flux of linear momentum. To
see the connection, consider a fluid in a state of Newtonian
flow, in which a series of layers move past one another and,
in this case, in the x-direction (Fig. 16A.2). The layer next
to the wall of the vessel is stationary, and the velocity of
N successive layers varies linearly with distance, z, from the
dN wall. Molecules ceaselessly move between the layers and
<0
dz bring with them the x-component of linear momentum they
z possessed in their original layer. A layer is retarded by mol-
ecules arriving from a more slowly moving layer because
such molecules have a lower momentum in the x-direction.
A layer is accelerated by molecules arriving from a more
J>0
rapidly moving layer. The net retarding effect is interpreted
Figure 16A.1 The flux of particles down a concentration gradient. as the viscosity of the fluid.
Fick’s first law states that the flux of matter is proportional to the Because the retarding effect depends on the transfer of the
concentration gradient at that point. x-component of linear momentum into the layer of interest,
692 16 Molecules in motion

How is that done? 16A.1 Deriving an expression for the

collision flux
Wall Brings Brings
x low high Consider a wall of area A perpendicular to the x-axis
momentum momentum
(Fig. 16A.3). In the following calculation, note that for a per-
fect gas the equation of state pV = nRT can be used to relate
the number density, N, to the pressure by N = N/V = nNA/V
= nNAp/nRT = p/kT. In the final equality R = NAk was used,
where k is Boltzmann’s constant.
z
|vxΔt|
Figure 16A.2 The viscosity of a fluid arises from the transport
of linear momentum. In this illustration the fluid is undergoing
Area, A
Newtonian (laminar) flow in the x-direction, and particles bring
their initial momentum when they enter a new layer. Won’t
Will

the viscosity depends on the flux of this x-component in the Volume = |vxΔt|A
z-direction. The flux of the x-component of momentum is x
­proportional to dvx/dz, where vx is the velocity in the x-direc-
Figure 16A.3 A molecule will reach the wall on the right
tion; the flux can therefore be written
within an interval Δt if it is within a distance vxΔt of the wall
and travelling to the right.
dvx
J ( x -component of momentum) = − η 
dz
Step 1 Identify the number of molecules that will strike an area
Momentum flux in terms of
 (16A.6)
the coefficient of viscosity If a molecule has vx > 0 (that is, it is travelling in the direction
of positive x), then it will strike the wall within an interval Δt
The constant of proportionality, η, is the coefficient of viscos- if it lies within a distance vxΔt of the wall. Therefore, all mol-
ity (or simply ‘the viscosity’). Its units are kilograms per metre ecules in the volume AvxΔt, and with a positive x-component
per second (kg m−1 s−1). Viscosities are often reported in the of velocity, will strike the wall in the interval Δt. The total
non-SI unit poise (P), with 1 P = 10−1 kg m−1 s−1. Some experi- number of collisions in this interval is therefore NAvxΔt,
mental values are given in Table 16A.1. where N is the number density of molecules.
Although it is not strictly a transport property, closely
Step 2 Take into account the range of velocities
­related to diffusion is effusion, the escape of matter through
The velocity vx has a range of values described by the prob-
a small hole. The essential empirical observations on effusion
ability distribution f (vx ) given in eqn 1B.3:
are summarized by Graham’s law of effusion, which states
that the rate of effusion is inversely proportional to the square 1/2
 m  − mvx2 /2 kT 
root of the molar mass, M. f (v x ) =  e
 2πkT 

with f (vx )dvx the probability of finding a molecule with

a component of velocity between vx and vx + dvx . The total
16A.2 The transport parameters number of collisions is found by summing NAvx Δt over all
positive values of vx (because only molecules with a positive
The kinetic theory of gases (Topic 1B) can be used to derive component of velocity are moving towards the area of inter-
expressions for the diffusion characteristics of a perfect est) with each value of vx being weighted by the probability of
gas. All the expressions depend on knowing the collision it occurring:
flux, ZW, which is the rate at which molecules strike a re- ∞
gion in the gas (this region may be an imaginary window, a Number of collisions = NA∆t ∫ vx f (vx )d vx
0
part of a wall, or a hole in a wall). Specifically, the collision
flux is the number of collisions divided by the area of the The collision flux is the number of collisions divided by A
region and the duration of the time interval. Its depend- and Δt, so
ence on pressure and temperature can be derived from the ∞

kinetic theory. ZW = N ∫0
vx f (vx )d vx
 16A Transport properties of a perfect gas 693

Step 3 Evaluate the integral kg m−1 s−2


Because 1.00 ×105 Pa
ZW = 1/2
Integral G.2  32.00 ×10−3 kg mol −1 −23 
 2π × 23 −1 × (1.381 ×10 JK −1 ) × (298K)
∞  m 
1/2
∞  kT 
1/2
 6.022 × 10 mol 
∫ ∫0 vx e x dvx =  2πm 
− mv 2 /2 kT
vx f (vx )d vx = 
0  2πkT  = 2.70 ×1027 m −2 s −1

it follows that This flux corresponds to 0.45 mol cm−2 s −1.

N = p/kT
1/2 1/2
 kT  p  kT 
ZW = N  =
 2πm  kT  2πm  (a) The diffusion coefficient
The first application of the result in eqn 16A.7b is to use it to
and therefore
find an expression for the net flux of molecules arising from a
p (16A.7a) concentration gradient.
ZW = 
(2πmkT )1/2 Collision flux in terms of pressure
[perfect gas]
How is that done? 16A.2 Deriving an equation for the net
Step 4 Develop an alternative expression in terms of the mean
flux of matter
speed
The mean speed is given by eqn 16A.1b as Consider the arrangement depicted in Fig. 16A.4. The mol-
ecules passing through the area A at z = 0 have travelled
R = NAk M = NAm
an average of about one mean free path, λ, since their last
collision.
1/2 1/2
 8 RT   8kT 
vmean =  =
 πM   πm 
Number density, N

N(–λ)
It follows that N(0)
N(+λ)
1/2
z Area, A
 kT 
 2πm = 14 vmean

and therefore Z W = N (kT /2πm)1/2 can be expressed as

 (16A.7b)
Z W = 14 Nvmean +λ
Collision flux
[perfect gas] 0
–λ

According to eqn 16A.7a, the collision flux increases with Figure 16A.4 The calculation of the rate of diffusion of a gas
pressure simply because increasing the pressure increases considers the net flux of molecules through a plane of area A
the number density and hence the number of collisions. The as a result of molecules arriving from an average distance λ
flux decreases with increasing mass of the molecules because away in each direction.
heavy molecules move more slowly than light molecules.
Caution, however, is needed with the interpretation of the role Step 1 Set up expressions for the flux in each direction
of temperature: it is wrong to conclude that because T 1/2 ap- If the number density at z is N (z), then the number density at
pears in the denominator that the collision flux decreases with z = λ can be estimated by using a Taylor expansion of the form
increasing temperature. If the system has constant volume, the f(x) = f(0) + (df/dx)0x + …, truncated after the second term (see
pressure increases with temperature (p ∝ T), so the collision The chemist’s toolkit 12 in Topic 5B):
flux is in fact proportional to T/T 1/2 = T 1/2, and increases with
temperature (because the molecules are moving faster).  dN 
N (+ λ ) = N (0) + λ 
 d z  0
Brief illustration 16A.2 Similarly, the number density at z = −λ is

The collision flux of O2 molecules, with m = M/NA and M =  dN 

N (− λ ) = N (0) − λ 
32.00 g mol−1, at 25 °C and 1.00 bar is  d z  0
694 16 Molecules in motion

The average number of impacts on the imaginary region

of area A during an interval Δt is ZWAΔt, where ZW is the
collision flux. Therefore, the matter flux from left to right,
J(L→R), arising from the supply of molecules on the left, is
this number of collisions divided by the time interval and Short flight
(survives)
the area:
Long flight
Z W = Nvmean
1 (collides in flight)
4

Z W A∆t
J (L → R) = = Z W = 14 N (− λ )vmean Figure 16A.5 One issue ignored in the simple treatment is that
A∆t
some molecules might make a long flight to the plane even
though they are only a short perpendicular distance away from
The number density is that at z = −λ , where the molecules
it. A molecule taking the longer flight has a higher chance of
originate before striking the area. There is also a flux of mol-
colliding during its journey.
ecules from right to left. The molecules making this journey
have originated from z = +λ where the number density is
N (λ). Therefore,
Brief illustration 16A.3
J(L ← R) = 14 N (λ )vmean
In Brief illustration 1B.3 of Topic 1B it is established that the
Step 2 Evaluate the net flux mean free path of N2 molecules in a gas at 1.0 atm and 25 °C is
91 nm; in Example 1B.1 of the same Topic it is calculated that
The net flux from left to right is
under the same conditions the mean speed of N2 molecules is
J z = J (L → R) − J (L ← R) 475 m s −1. Therefore, the diffusion coefficient for N2 molecules
under these conditions is
= 14 vmean {N (− λ ) − N (λ )}
D= 1
3 × (91 × 10−9 m) × (475 m s −1) = 1.4 × 10−5 m2 s −1
   dN     dN   
= 14 vmean   N (0) − λ    −  N (0) + λ  dz    The experimental value is 1.5 × 10−5 m2 s −1.
   dz 0  0 

That is,

 (16A.8) There are three points to note about eqn 16A.9:

 dN 
J z = − 12 vmean λ 
 dz  0 Net flux • The mean free path, λ , decreases as the pressure is
increased (eqn 16A.1a), so D decreases with increas-
ing pressure and, as a result, the gas molecules
This equation shows that the flux is proportional to the gra-
­diffuse more slowly.
dient of the concentration, in agreement with the empirical

Physical interpretation
observation expressed by Fick’s law, eqn 16A.2. • The mean speed, vmean , increases with the tempera-
At this stage it looks as though a value of the diffusion coef- ture (eqn 16A.1b), so D also increases with tempera-
ficient can be picked out by comparing eqns 16A.8 and 16A.4, ture. As a result, molecules in a hot gas diffuse more
so obtaining D = 12 λ vmean. It must be remembered, however, quickly than those when the gas is cool (for a given
that the calculation is quite crude, and is little more than an concentration gradient).
assessment of the order of magnitude of D. One aspect that has • Because the mean free path increases when the
not been taken into account is illustrated in Fig. 16A.5, which ­collision cross-section σ of the molecules decreases,
shows that although a molecule may have begun its journey the diffusion coefficient is greater for small mol-
very close to the window, it could have a long flight before ecules than for large molecules.
it gets there. Because the path is long, the molecule is likely
to collide before reaching the window, so it ought not to be
(b) Thermal conductivity
counted as passing through the window. Taking this effect
into account results in the appearance of a factor of 32 repre- According to the equipartition theorem (The chemist’s toolkit 7
senting the lower flux. The modification results in in Topic 2A), each molecule carries an average energy ε = νkT,
where ν depends on the number of quadratic contributions to
the energy of the molecule and is a number of the order of 1.
D = 13 λvmean Diffusion coefficient (16A.9)
For atoms, which have three translational degrees of freedom,
 16A Transport properties of a perfect gas 695

ν = 32 . When one molecule passes through the imaginary win- The number density can be written in terms of the molar
dow, it transports that average energy. An argument similar to concentration [J] of the carrier particles J: N = nNA/V =
that used for diffusion can be used to discuss the transport of [J]NA. Next, the quantity νkNA can be identified as the molar
energy through this window. constant-volume heat capacity of the gas: the molar energy
is NAνkT, so it follows from CV,m = (∂Um/∂T)V that CV,m = NAνk.
With these substitutions, eqn 16A.10a becomes
How is that done? 16A.3 Deriving an expression for the
thermal conductivity κ = 13 vmean λ [ J ]CV, m  Thermal conductivity (16A.10b)
Assume that the number density is uniform but that the
Yet another form is found by starting with eqn 16A.10a, rec-
­temperature, and hence the average energy of the molecules,
is not. Molecules arrive from the left after travelling a mean ognizing that N = p/kT and then using the expression for D in
free path from their last collision in a hotter region, and eqn 16A.9:
therefore arrive with a higher energy. Molecules also arrive N = p/kT D = 31 λvmean
from the right after travelling a mean free path from a cooler
p ν pD
region, and hence arrive with lower energy. κ = 13 ν kvmean λ N = 13 ν kvmean λ =
kT T
Step 1 Write expressions for the forward, reverse, and net  Thermal conductivity (16A.10c)
energy flux
If the average energy of the molecules at z is ε ( z ), the two
opposing energy fluxes are Brief illustration 16A.4

ZW ZW In Brief illustration 16A.3 the value D = 1.4 × 10−5 m2 s −1 was

      calculated for N2 molecules at 25 °C and 1.0 atm. The thermal
J (L → R) = 14 Nvmean ε (− λ ) J (L ← R) = 14 Nvmean ε (λ )
conductivity can be calculated by using eqn 16A.10c and not-
ing that for N2 molecules ν = 52 (three translational modes and
and the net flux is two rotational modes; the vibrational mode is inactive for this
‘stiff’ molecule).
J z = J (L → R) − J (L ← R)
−3
Jm

= 14 vmean N {ε (− λ ) − ε (λ )} × (1.01×10 Pa) × (1.4 ×10−5 m 2 s −1 )
5 5
κ= 2
298K
= 1.2 ×10−2 JK −1 m −1 s −1
ε(z) = ε(0) + z(dε/dz)0, z = ±λ

   dε     dε    or 12 mW K−1 m−1. The experimental value is 26 mW K−1 m−1.

= 14 vmean N  ε (0) − λ    − ε (0) + λ    
 dz   dz  0  
  0  

 dε  To interpret eqn 16A.10, note that:

= − 12 vmean λ N  
 dz  0
• The mean free path is λ is inversely proportional to
the pressure, but the number density N is propor-
Step 2 Express the energy gradient as a temperature gradient tional to the pressure (N = p/kT). It follows that the
Because ε = νkT, a gradient of energy can be expressed as a product λN, which appears eqn 16A.10a, is inde- Physical interpretation
gradient of temperature: d ε /d z = ν k(dT /d z); therefore pendent of pressure, and so therefore is the thermal
conductivity.
 dε   dT  • The thermal conductivity is greater for gases with
J z = − 12 vmean λ N   = − 12 ν kvmean λ N 
 dz  0  d z  0 a high heat capacity (eqn 16A.10b) because a given
temperature gradient then corresponds to a steeper
This equation shows that the energy flux is proportional to the energy gradient.
temperature gradient, which is the desired result. As before,
the constant is multiplied by 32 to take long flight paths into
account, and comparison of this equation with eqn 16A.5 The physical reason for the pressure independence of the ther-
shows that mal conductivity is that the conductivity can be expected to be
large when many molecules are available to transport the en-
 (16A.10a) ergy, but the presence of so many molecules limits their mean
κ = 13 ν kvmean λ N
Thermal conductivity free path and they cannot carry the energy over a great distance.
696 16 Molecules in motion

The net flux of x-momentum in the z-direction is therefore

x Fast
layer    dv     dv   
J z = 14 vmean N  mvx (0) − mλ  x   − mvx (0) + mλ  x   
   d z 0   dz  0 

Slow
layer  dv 
= − 12 vmean λmN  x 
 dz  0

Step 2 Identify the coefficient of viscosity

z The flux is proportional to the velocity gradient, in line with
λ the phenomenological equation. Comparison of this expres-
0 sion with eqn 16A.6, and multiplication by 32 in the normal
–λ way, leads to

Figure 16A.6 The calculation of the viscosity of a gas examines  (16A.11a)

the net x-component of momentum brought to a plane from η = 13 vmean λmN
Viscosity
faster and slower layers a mean free path away in each direction.
Two alternative forms of this expression (after using mNA = M
and eqn 16A.9, D = 13 λvmean) are
These two effects cancel. The thermal conductivity is indeed
found experimentally to be independent of the pressure, except η = MD [ J ] (16A.11b)
when the pressure is very low, and then κ ∝ p. At very low pres- pMD
sures it is possible for λ to exceed the dimensions of the appa- η=  (16A.11c)
RT
ratus, and the distance over which the energy is transported is
where [J] is the molar concentration of the gas molecules J and
then determined by the size of the container and not by colli-
M is their molar mass.
sions with the other molecules present. The flux is still propor-
tional to the number of carriers, but the length of the journey no
longer depends on λ , so κ ∝ [J], which implies that κ ∝ p.
Brief illustration 16A.5

(c) Viscosity From Brief illustration 16A.3 the value of D for N2 molecules at
25 °C and 1.0 atm is 1.4 × 10−5 m2 s −1. Because M = 28.02 g mol−1,
If the momentum in the x-direction depends on z as mvx(z), eqn 16A.11c gives
then molecules moving from the right in Fig. 16A.6 (from a
fast layer to a slower one) transport a momentum mvx(λ) to J m−3
their new layer at z = 0; those travelling from the left transport 
(1.01×10 Pa) × (28.02 ×10−3 kg mol −1 ) × (1.4 ×10−5 m 2 s −1 )
5

mvx(−λ) to it. This picture can be used to build an expression η=

(8.3145JK −1 mol −1 ) × (298K)
for the coefficient of viscosity.
= 1.6 ×10 −5 kg m −1 s −1

How is that done? 16A.4 Deriving an expression for or 160 μP. The experimental value is 177 μP.
the viscosity
The strategy is the same as in the previous derivations, but
The physical interpretation of eqn 16A.11a is as follows:
now the property being transported is the linear momentum
of the layers. • As has already been noted for thermal conductivity,
Physical interpretation

the product λN is independent of p. Therefore, like

Step 1 Set up expressions for the flux of momentum in each
the thermal conductivity, the viscosity is independ-
direction and the net flux
ent of the pressure.
Molecules arriving from the right bring a momentum
• Because vmean ∝ T 1/2, η ∝ T 1/2 at constant volume (and
 dv  η ∝ T 3/2 at constant pressure). That is, the viscosity of
mvx (λ ) = mvx (0) + mλ  x 
 dz  0 a gas increases with temperature.

Those arriving from the left bring a momentum The physical reason for the pressure-independence of the
­v iscosity is the same as for the thermal conductivity: more
 dv  molecules are available to transport the momentum, but they
mvx (− λ ) = mvx (0) − mλ  x 
 dz  0 carry it less far on account of the decrease in mean free path.
 16A Transport properties of a perfect gas 697

The increase of viscosity with temperature is explained when Example 16A.1 Calculating the vapour pressure from
it is recalled that at high temperatures the molecules travel
a mass loss
more quickly, so the flux of momentum is greater. In contrast,
as discussed in Topic 16B, the viscosity of a liquid decreases Caesium (m.p. 29 °C, b.p. 686 °C) was introduced into a con-
with increase in temperature because intermolecular interac- tainer and heated to 500 °C; the container is pierced by a hole
tions must be overcome. of diameter 1.0 mm. It is found that in 1.00 h (3600 s) the mass
of the container decreased by 84.4 mg. Calculate the vapour
pressure of liquid caesium at 500 °C.
(d) Effusion
Collect your thoughts The pressure of vapour is constant
Because the mean speed of molecules is inversely propor- inside the container despite the effusion of atoms because the
tional to M1/2, the rate at which they strike the area of the hole hot liquid metal replenishes the vapour. The rate of effusion
through which they are effusing is also inversely proportional is therefore constant, and given by eqn 16A.12. To express
to M1/2, in accord with Graham’s law. However, by using the the rate in terms of mass, multiply the number of atoms that
expression for the collision flux, a more detailed expression for escape by the mass of each atom.
the rate of effusion can be obtained and used to interpret effu-
The solution The mass loss Δm in an interval Δt is equal to
sion data in a more sophisticated way.
the number of molecules that strike the area of the hole in this
When a gas at a pressure p and temperature T is separated
interval multiplied by the mass of each molecule:
from a vacuum by a small hole, the rate of escape of its mol-
ecules is equal to the rate at which they strike the area of the eqn 16A.12
hole, which is the product of the collision flux and the area of 
pA0 N A
the hole, A0: ∆m = rate of effusion × ∆ t × m = ×∆ t × m
(2πMRT )1/2
m = M/NA, k = R/NA
where A0 is the area of the hole and M is the molar mass.
pA0 pA0 N A Rearrangement of this equation gives an expression for p:
Rate of effusion = Z W A0 = = 
(2πmkT )1/2 (2πMRT )1/2 1/2
 2πRT  ∆m
 p=
Rate of effusion (16A.12)  M  A0 ∆ t
Substitution of the data and M = 132.9 g mol−1 gives
This rate is inversely proportional to M1/2, in accord with
Graham’s law. Do not conclude, however, that because the ex- 1/2
 2π × (8.3145JK −1 mol −1 ) × (773K) 
pression includes a factor of T −1/2 the rate of effusion decreases p=  ×
 0.1329kg mol −1
as the temperature increases. Because p ∝ T, the rate is in fact
proportional to T 1/2 and increases with temperature. 84.4 ×10−3 kg
= 1.6 ×10 4 Pa or 16kPa
Equation 16A.12 is the basis of the Knudsen method for the π × (0.50 ×10 −3 m)2 × (3600s)
  
determination of the vapour pressures of liquids and solids, A0
particularly of substances with very low vapour pressures and
which cannot be measured directly. Thus, if the vapour pres- Self-test 16A.1 How long would it take for 200 mg of Cs atoms
sure of a sample is p, and it is enclosed in a cavity with a small to effuse out of the oven under the same conditions?
hole, then the rate of loss of mass from the container is propor-
Answer: 8.5 × 103 s or 2.4 h

tional to p.

Checklist of concepts
☐ 1. Flux is the quantity of a property passing through a ☐ 3. Fick’s first law of diffusion states that the flux of matter
given area in a given time interval divided by the area is proportional to the concentration gradient.
and the duration of the interval. ☐ 4. Thermal conduction is the migration of energy down a
☐ 2. Diffusion is the migration of matter down a concentra- temperature gradient; the flux of energy is proportional
tion gradient. to the temperature gradient.
698 16 Molecules in motion

☐ 5. Viscosity is the migration of linear momentum down ☐ 7. Graham’s law of effusion states that the rate of effusion
a velocity gradient; the flux of momentum is propor- is inversely proportional to the square root of the molar
tional to the velocity gradient. mass.
☐ 6. Effusion is the emergence of a gas from a container
through a small hole.

Checklist of equations
Property Equation Comment Equation number

Fick’s first law of diffusion J = −DdN /dz 16A.4

Flux of energy of thermal motion J = −κ dT /d z 16A.5

Flux of momentum along x J = − ηdvx /d z 16A.6

Diffusion coefficient of a perfect gas D = 13 λ vmean KMT* 16A.9

Coefficient of thermal conductivity of a perfect gas κ = 13 vmean λ [ J ]CV, m KMT and equipartition 16A.10b

Coefficient of viscosity of a perfect gas η = 13 vmean λmN KMT 16A.11a

1/2
Rate of effusion Rate ∝ 1/M Graham’s law 16A.12

* KMT indicates that the equation is based on the kinetic theory of gases.
TOPIC 16B Motion in liquids

(a) Liquid viscosity

➤ Why do you need to know this material?
Many chemical reactions take place in liquids, so for a full
The coefficient of viscosity, η (eta), is introduced in Topic 16A
understanding of them it is important to know how solute
as the constant of proportionality in the relation between the
molecules and ions move through such environments.
flux of linear momentum and the velocity gradient in a fluid:
➤ What is the key idea? dvx
Ions reach a terminal velocity when the electrical force on J z ( x -component of momentum) = − η  (16B.1)
dz
them is balanced by the drag due to the viscosity of the
The units of viscosity are kilograms per metre per second
solvent.
(kg m−1 s −1), but they may also be reported in the equivalent
➤ What do you need to know already? units of pascal seconds (Pa s). The non-SI units poise (P) and
centipoise (cP) are still widely encountered: 1 P = 10 −1 Pa s
The discussion of viscosity starts with the definition of the
and so 1 cP = 1 mPa s. Table 16B.1 lists some values of η for
coefficient of viscosity introduced in Topic 16A. Some cal-
liquids.
culations make use of the information about electrostatics
Unlike in a gas, for a molecule to move in a liquid it must
set out in The chemist’s toolkit 29.
acquire at least a minimum energy (an ‘activation energy’, Ea,
in the language of Topic 17D) to escape from its neighbours.
From the Boltzmann distribution it follows that the probabil-
ity that a molecule has at least an energy Ea is proportional to
The motion of ions and molecules is an important aspect of e − E /RT , so the mobility of the molecules in the liquid should
a

the properties of liquids and of the reactions taking place in follow this type of temperature dependence. As the tempera-
them. It can be studied experimentally by a variety of meth- ture increases, the molecules become more mobile and so the
ods. For example, bulk measurements of viscosity and its viscosity is reduced; the expected temperature dependence of
temperature dependence can be used to build models of the the viscosity, which decreases as the mobility of the molecules
motion. At a more detailed level, relaxation time measure- increases, is therefore of the form
ments in NMR (Topic 12C) and EPR can be used to show how
molecules rotate. For instance, these observations show that η = η0e E /RT 
a
Temperature dependence of viscosity (liquid) (16B.2)
large molecules in viscous fluids typically rotate in a series of
small (about 5°) steps, whereas small molecules in non-viscous (Note the positive sign of the exponent.) The activation ­energy
fluids typically jump through about 1 radian (57°) in each typical of viscosity is comparable to the mean potential en-
step. Another important technique is inelastic neutron scat- ergy of intermolecular interactions. Equation 16B.2 implies
tering, in which the energy neutrons collect or discard as they that the viscosity should decrease sharply with increasing
pass through a sample is interpreted in terms of the motion of temperature. Such a variation is found experimentally, at
its molecules.

Table 16B.1 Viscosities of liquids at 298 K*

η/(10−3 kg m−1 s−1)

16B.1 Experimental results Benzene 0.601

There are two ‘classical’ methods of investigating molecular Mercury 1.55

motion in liquids. One is the measurement of viscosity and its Pentane 0.224
temperature dependence. Another involves inferring details Water‡ 0.891
about molecular motion by dragging ions through a solvent * More values are given in the Resource section.
under the influence of an electric field. ‡
Note that 1 cP = 10 −3 kg m−1 s −1; the viscosity of water corresponds to 0.891 cP.
700 16 Molecules in motion

2
(b) Electrolyte solutions
Viscosity, η/(10–3 kg m–1 s–1)

1.6 When a potential difference is applied between two electrodes

immersed in a solution containing ions there is a flow of cur-
1.2 rent due to the migration of ions through the solution. The key
electrical property of the solution is its resistance, R, which
0.8 is expressed in ohms, Ω (1 Ω = 1 C −1 V s). It is often more con-
venient to work in terms of the conductance, G, which is the
0.4 inverse of the resistance: G = 1/R, and therefore expressed in
Ω−1. The reciprocal ohm used to be called the mho, but its SI
0
0 20 40 60 80 100
designation is now the siemens, S, and 1 S = 1 Ω−1 = 1 C V −1 s−1.
Temperature, θ/°C Electric current is expressed in amperes, A, with 1 A = 1 C s−1,
so a more physically revealing relation is 1 S = 1 A V −1.
Figure 16B.1 The temperature dependence of the viscosity of
The conductance of a sample is proportional to its cross-
water. As the temperature is increased, more molecules are able
to escape from the potential wells provided by their neighbours, sectional area, A, and inversely proportional to its length, l.
and so the liquid becomes less viscous. Therefore
A Conductivity
G =κ  (16B.3)
l [definition]

least over reasonably small temperature ranges (Fig. 16B.1). where the constant of proportionality κ (kappa) is the electri-
Intermolecular interactions govern the magnitude of Ea, but cal conductivity of the sample. With the conductance in sie-
the calculation of its value from an assumed form of the po- mens and the dimensions in metres, it follows that the units
tential energy due to the interactions between the molecules is of κ are siemens per metre (S m−1). The conductivity is a prop-
an immensely difficult and largely unsolved problem. erty of the material, whereas the conductance depends both
on the material and its dimensions. The conductivity depends
on the concentration of charge carriers in the sample, which
Brief illustration 16B.1
suggests that it is sensible to introduce the molar conductivity,
The viscosity of water at 25 °C and 50 °C is 0.890 mPa s and Λm, ­defined as
0.547 mPa s, respectively. It follows from eqn 16B.2 that the κ Molar conductivity
activation energy for molecular migration is the solution of Λm =  (16B.4)
c [definition]
η(T2 ) ( Ea /R )(1/T2 −1/T1 )
=e where c is the molar concentration of the added electrolyte.
η(T1 )
The units of molar conductivity are siemens metre-squared per
which, after taking logarithms of both sides, is mole (S m2 mol−1), and typical values are about 10 mS m2 mol−1
R ln{η(T2 )/η(T1 )}
(where 1 mS = 10−3 S).
Ea = Experimentally, it is found that the value of the molar con-
1/ T2 −1/ T1
ductivity varies with the concentration. One reason for this
(8.3145JK −1 mol −1 )ln{(0.547mPa s)/(0.890mPa s)}
= variation is that the number of ions in the solution might not
1/(323K) −1/(298K)
be proportional to the nominal concentration of the electro-
= 1.56 ×104 Jmol −1 lyte. For instance, the concentration of ions in a solution of a
or 15.6 kJ mol−1. This value is comparable to the strength of a weak electrolyte depends on the degree of dissociation, which
hydrogen bond. is a complicated function of the total amount of solute: dou-
bling the total concentration of the solute does not double the
number of ions. Secondly, even for fully dissociated strong
One problem with the interpretation of viscosity measure- electrolytes, because ions interact strongly with one another,
ments is that the change in density of the liquid as it is heated the conductivity of a solution is not exactly proportional to the
makes a pronounced contribution to the temperature vari- number of ions present.
ation of the viscosity. Thus, the temperature dependence of In an extensive series of measurements during the nineteenth
viscosity at constant volume, when the density is constant, is century, Friedrich Kohlrausch established the Kohlrausch law,
much less than that at constant pressure. At low temperatures, that at low concentrations the molar conductivities of strong
the viscosity of water decreases as the pressure is increased. electrolytes depend on the square root of the concentration:
This behaviour is consistent with the need to rupture hydro-
gen bonds for migration to occur. Λm = Λ°m −K c1/2 Kohlrausch law (16B.5)
 16B Motion in liquids 701

where Λ°m is the limiting molar conductivity, the molar con- Self-test 16B.1 The conductivity of KClO4(aq) at 25 °C is
ductivity in the limit of zero concentration when the ions 13.780 mS m−1 when c = 1.000 mmol dm−3 and 67.045 mS m−1
are so far apart that they move independently of each other. when c = 5.000 mmol dm−3. Determine the values of the limit-
Kohlrausch also established that this limiting molar conduc- ing molar conductivity Λ°m and the Kohlrausch constant K for
tivity is the sum of contributions from the individual ions this system.
present. If the limiting molar conductivity of the cations is Answer: K = 9.491mSm 2 mol −1 /(moldm −3 )1/2 , Λ°m = 14.080mSm 2 mol −1
denoted λ+ and that of the anions λ−, then his law of the inde-
pendent migration of ions states that

Law of the independent

Λ°m = ν+λ+ + ν−λ− migration of ions (16B.6)
16B.2 The mobilities of ions
where ν+ and ν− are the numbers of cations and anions pro-
vided by each formula unit of electrolyte. For example, for The reason why different ions have different molar conduc-
HCl, NaCl, and CuSO4, ν+ = ν− = 1, but for MgCl2 ν+ = 1, ν− = 2. tivities in solution can be understood by analysing the motion
of an ion subject to an electric field and at the same time sur-
Example 16B.1 rounded by a viscous medium.
Determining the limiting molar
conductivity
The conductivity of KCl(aq) at 25 °C is 14.688 mS m−1 when c =
(a) The drift speed
1.0000 mmol dm−3, and 71.740 mS m−1 when c = 5.0000 mmol An ion in a vacuum is accelerated by an electric field, but in a
dm−3. Determine the values of the limiting molar conductivity viscous liquid the motion of the ion is impeded by the need for
Λ°m and the Kohlrausch constant K . it to push its way through the tightly packed solvent molecules.
Collect your thoughts You need to use eqn 16B.4 to determine The latter effect is called viscous drag. As the ion accelerates
the molar conductivities at the two concentrations. Then, by under the influence of the field, the viscous drag increases and
using eqn 16B.5, you can express the difference between these the ion quickly reaches a steady terminal speed, called the
two values as Λm(c2) − Λm(c1) = K (c11/2 − c21/2). From this relation drift speed, which can be found by balancing the two forces.
you can determine K and then go on to find Λ°m by using one
of the values of the molar conductivity in eqn 16B.5, rear- How is that done? 16B.1 Deriving an expression for
ranged into Λ°m = Λm + K c1/2 .
the drift speed
The solution The molar conductivity of KCl(aq) when
The starting point for this calculation is the result from
c = 1.0000 mmol dm−3 (which is the same as 1.0000 mol m−3) is
electrostatics that when the potential difference between two
14.688mSm −1 planar electrodes a distance l apart is Δϕ, the ions in the solu-
Λm = = 14.688mSm 2 mol −1 tion between them experience a uniform electric field of mag-
1.0000molm −3
nitude E = Δϕ/l. Here, and throughout this section, the sign
Similarly, when c = 5.0000 mol dm−3 its molar conductivity is of the charge number is disregarded so as to avoid notational
14.348 mS m2 mol−1. It then follows that complications.
Λm (c2 ) − Λm (c1 ) (14.348 −14.688)mSm 2 mol −1 Step 1 Find the force on the ion due to the field
K= =
1/2
c1 − c21/2
(0.001 00001/2 − 0.005 00001/2 ) (mol dm −3 )1/2 In an electric field E , an ion of charge ze experiences a force of
2 −1
= 8.698mSm mol /(mol dm ) −3 1/2
magnitude zeE (see The chemist’s toolkit 29). Therefore,
ze∆φ
(For reasons that will become clear immediately below, it is Felectric =
l
best to keep this awkward but convenient array of units rather
than converting them to the equivalent 10−3/2 S m7/2 mol−3/2.) Step 2 Find the force on the ion due to viscous drag
The limiting molar conductivity is then found by using the As the ion moves through the solvent it experiences a fric-
data for c = 1.0000 mmol dm−3: tional retarding force proportional to its speed. For a spherical
mSm 2 mol −1 particle of radius a travelling at a speed s, this force is given by
Λ°m = 14.688mSm 2 mol −1 + 8.698 Stokes’ law, which Stokes derived by considering the hydro­
(mol dm −3 )1/2
dynamics of the passage of a sphere through a continuous fluid:
× (1.0000 ×10−3 moldm −3 )1/2 = 14.963mSm 2 mol −1
Fviscous = fs f = 6πηa Stokes’ law (16B.7)
Comment. Although the value of K has been given to four
significant figures in conformity with the data, that degree of where η is the coefficient of viscosity. In this calculation it
precision is probably over-optimistic in practice. is assumed that Stokes’ law applies on a molecular scale;
702 16 Molecules in motion

e­ xperimental evidence suggests that it often gives at least the The chemist’s toolkit 29
right order of magnitude for the viscous force. Electrostatics

Step 3 Find the drift speed by balancing the two forces A charge Q 1 (units: coulomb, C) gives rise to a Coulomb
potential ϕ (units: volt, V). The potential energy (units: joule,
The two forces act in opposite directions and the ions quickly J, with 1 J = 1 V C) of a second charge Q in that potential is
reach a terminal speed, the drift speed, s, when they are in bal-
ance. This balance occurs when fs = zeE , and therefore EP = _Q φ

 zeE (16B.8a) In one dimension, the electric field strength (units: volt per
s= metre, V m−1), E , is the negative of the gradient of the electric
f Drift speed
potential ϕ:
Equation 16B.8a shows that the drift speed is proportional dφ
to the electric field strength. The constant of proportionality E= _  Electric field strength
dx
is called the mobility of the ion, u:
In three dimensions the electric field is a vector, and
Mobility
s = uE (16B.8b)
[definition] E = _∇φ
With the electric field strength in volts per metre (V m−1) The electric field between two plane parallel plates separated
and the drift speed in metres per second (m s −1) the slightly by a distance l, and between which there is a potential differ-
awkward units of u are metres-squared per volt per second ence Δϕ, is uniform and given by
(m2 V −1 s −1; note that m2 V −1 s −1 × V m−1 = m s −1); a selection of
values is given in Table 16B.2. Comparison of the last two ∆φ
E= _
equations shows that l
A charge Q experiences a force proportional to the electric
f = 6πηa
field strength at its location:
ze ze
u= =
f 6πηa  (16B.9) Felectric = QE

where the Stokes’ law value for the frictional coefficient f has A potential gives rise to a force only if it varies with distance.
been used.

Brief illustration 16B.2

confirm these predictions for bulky ions (such as R4N+ and
An order-of-magnitude estimate of the mobility can be found RCO2−) but not for small ions. For example, the mobilities
using eqn 16B.9 with z = 1 and a = 130 pm, which is typical of of the alkali metal ions in water increase from Li+ to Rb+
the radius of a hydrated ion; the viscosity of water at 25 °C is (Table 16B.2) even though the ionic radii increase. The para-
0.9 cP, or 0.9 mPa s. Then dox is resolved when it is realized that the radius a in the Stokes
formula is the hydrodynamic radius (or ‘Stokes ­radius’) of the
J V −1 ion, its effective radius in the solution taking into account all

1.6 ×10 C −19 the H2O molecules it carries in its hydration shell. Small ions
u= = 7.3 ×10−8 m 2 V −1 s −1
6π × (0.9 ×10−3 Pa −12
 s) × (130 ×10 m)
J m−3

This value means that when there is a potential difference of Table 16B.2 Ionic mobilities in water at 298 K*
1.0 V across a solution of length 1.0 cm (so E = 100 V m−1), the
u/(10−8 m2 V −1 s−1) u/(10−8 m2 V −1 s−1)
drift speed is 7.3 μm s −1. That speed might seem slow, but not
when expressed on a molecular scale, because it corresponds H+ 36.23 OH− 20.64
to an ion passing about 104 solvent molecules per second. Li+
4.01 F− 5.70
+ −
Na 5.19 Cl 7.91
K+ 7.62 Br− 8.09
+ 2−
Rb 7.92 SO 8.29
Equation 16B.9 implies that the mobility of an ion decreases 4

with increasing solution viscosity and ion size. Experiments * More values are given in the Resource section.
 16B Motion in liquids 703

+ Step 1 Calculate the number of ions passing through an imagi-

nary window
By referring to Fig. 16B.3 you can see that the number of ions
of one kind, moving at speed s, that pass through an imagi-
+ nary window of area A during an interval Δt is equal to the
number within the distance sΔt, and therefore to the number
in the volume sΔtA. It follows that the number of ions passing
through the window in that period is NsΔtA = sΔtAνcNA.
Figure 16B.2 A highly schematic diagram showing the effective
motion of a proton in water.
Area, A

give rise to stronger electric fields than large ones (the electric
field at the surface of a sphere of radius r is proportional to z/r2,
so the smaller the radius the stronger the field). Consequently,
small ions are more extensively solvated than big ions and a
small ion may have a large hydrodynamic radius because it sΔt
drags many solvent molecules through the solution as it mi-
grates. The hydrating H2O molecules are often very labile, Figure 16B.3 In the calculation of the current, all the ions,
however, and NMR and isotope studies have shown that the moving at speed s, within a distance sΔt (i.e. those in the
exchange between the coordination sphere of the ion and the volume sAΔt) will pass through the area A.
bulk solvent is very rapid for ions of low charge but slow for
ions of high charge. Step 2 Calculate the charge passing through the window, and
The proton, although it is very small, has a very high mo- hence the current
bility (Table 16B.2). Proton and 17O-NMR show that the char- Each ion carries a charge ze, so the charge passing through
acteristic lifetime of protons hopping from one molecule to the window is zesΔtAνcNA which, by introducing Faraday’s
the next is about 1.5 ps, which is comparable to the time that constant F = eNA, can be written zsΔtAνcF. The electri-
inelastic neutron scattering shows it takes a water molecule cal current, I, is the rate of passage of charge, which in
to turn through about 1 radian (1 to 2 ps). According to the this case is the charge divided by the time interval Δt: so
Grotthuss mechanism, there is an effective motion of a proton I = zsΔtAνcF/Δt = zsAνcF.
that involves the rearrangement of bonds in a group of water
Step 3 Set up expressions for the conductance, the conductivity,
molecules (Fig. 16B.2). However, the actual mechanism is still and the molar conductivity
highly contentious. The mobility of protons in liquid ammo-
The conductance is given by G = I/Δϕ, where Δϕ is the poten-
nia is also anomalous and presumably occurs by an analogous
tial difference across the solution. It follows that
mechanism.
I zsAν cF
G= =
∆φ ∆φ
(b) Mobility and conductivity
The limiting molar conductivity of an ion is a measurable The conductivity is
quantity, and the mobility of an ion can, in principle, be cal-
Gl zsAν cFl zsν cFl
culated on the basis of a model of its motion through the sol- κ= = =
A ∆φ A ∆φ
vent. It should be possible to find a relation between these two
quantities. The limiting molar ionic conductivity is

κ zsν cFl zsFl

How is that done? 16B.2 λ= = =
Establishing the relation between ν c ν c∆φ ∆φ
ionic mobility and limiting molar conductivity
Step 4 Introduce the ionic mobility
To keep things notationally simple, ignore the signs of quanti-
At this point you can identify Δϕ/l as the electric field strength
ties in what follows and focus on their magnitudes. Consider a
E , and s/E as the mobility, u:
solution of an electrolyte at a molar concentration c. Let each
formula unit give rise to ν+ cations of charge z + e and ν− anions ∆φ
=E s = uE
of charge z − e. The molar concentration of each type of ion is l

therefore νc (with ν = ν+ or ν−), and the number density of each zsFl zsF Ion molar conductivity
λ= =
type is N = νcNA. ∆φ E = zuF  in terms of mobility
(16B.10)
704 16 Molecules in motion

Equation 16B.10 applies to the cations and to the anions. For Table 16B.3 Diffusion coefficients at 298 K, D/(10−9 m2 s−1)*
an electrolyte where there are ν+ and ν− cations and anions,
Molecules in liquids Ions in water
respectively, from each formula unit, it follows from eqn 16B.6
that Λ°m = ν+λ+ + ν−λ−. Therefore I2 in hexane 4.05 K+ 1.96 Br− 2.08
+ −
in benzene 2.13 H 9.31 Cl 2.03
 (16B.11a)
Λ°m = (z +u+ν+ + z −u−ν−)F Glycine in water 1.055 Na+ 1.33 I− 2.05
Limiting molar conductivity
−
in terms of mobilities H2O in water 2.26 OH 5.03
Sucrose in water 0.5216
For a symmetrical z:z electrolyte (e.g. CuSO4 with z + = z − = 2),
this equation simplifies to * More values are given in the Resource section.

Λ°m = z(u+ + u−)F (16B.11b)

the solute molecules are very mobile. Don’t be misled by the
presence of temperature in the denominator into thinking that
the ion mobility decreases with increasing temperature: the
Brief illustration 16B.3 diffusion coefficient increases more rapidly with temperature
than T itself, so u increases with increasing temperature.
In Brief illustration 16B.2 the mobility of a typical ion is esti-
mated as 7.3 × 10−8 m2 V −1 s −1. For z = 1, this value can be used
to estimate a typical limiting molar conductivity of the ion as
Brief illustration 16B.4
λ = 1 × (7.3 × 10−8 m2 V −1 s −1) × (9.649 × 104 C mol−1)
From Table 16B.2, the mobility of SO42− is 8.29 ×10−8 m2 V −1 s −1.
= 7.0 × 10−3 m2 V−1 s −1 C mol−1
It follows from eqn 16B.13 in the form D = uRT/zF that the
Because 1 V −1 C s −1 = 1 S, the value can be expressed as diffusion coefficient for the ion in water at 25 °C is
7.0 mS m 2 mol−1. The experimental value for K+ (aq) is
(8.29 ×10−8 m 2 V −1 s −1 ) × (8.3145JK −1 mol −1 ) × (298K)
7.4 mS m2 mol−1. D= −1
2 × (9.649 ×104 Cmol
 )
J V −1
−9 2 −1
= 1.06 ×10 m s
(c) The Einstein relations
The relation between drift speed and the electric field strength
in eqn 16B.8a (s = zeE /f ) is a special case of a more general re- The Einstein relation can be developed to provide a link be-
lation derived in Topic 16C (eqn 16C.5): tween the limiting molar conductivity of an electrolyte and the
diffusion coefficients of its ions. First, by using eqns 16B.10 and
DF Drift speed in terms of 16B.13 the limiting molar conductivity of an ion can be written
s=  (16B.12)
RT diffusion coefficient
u = zDF/RT

where F is the force (per mole of ions) driving the ions through z 2 DF 2
the viscous medium and D is the diffusion coefficient for the λ = zuF =  (16B.14)
RT
species (Table 16B.3). For an ion in solution the drift speed is
s = uE (eqn 16B.8b), and the force on each ion in an electric Then, by using Λ°m = ν+λ+ + ν−λ− (eqn 16B.6), the limiting molar
field of strength E is ezE . It follows that the force per mole of conductivity is
ions is NAezE which, by using NAe = F, can be written zFE .
Substitution of these expressions for s and F into eqn 16B.12 F2 Nernst–Einstein
Λ°m = (ν + z +2 D+ + ν − z −2 D− )  (16B.15)
gives, on cancelling the E , the Einstein relation: RT equation

zDF which is the Nernst–Einstein equation. One application of

u=  Einstein relation (16B.13)
RT
this equation is to the determination of ionic diffusion coeffi-
The mobility is high when the diffusion coefficient (which is cients from conductivity measurements; another is to the pre-
inversely proportional to the viscosity) is high, indicating that diction of conductivities based on models of ionic diffusion.
 16B Motion in liquids 705

Checklist of concepts
☐ 1. The viscosity of a liquid decreases with increasing tem- ☐ 4. An ion reaches a drift speed when the acceleration due
perature. to the electrical force is balanced by the viscous drag.
☐ 2. Kohlrausch’s law states that, at low concentrations, the ☐ 5. The hydrodynamic radius of an ion may be greater
molar conductivities of strong electrolytes vary as the than its ionic radius.
square root of the concentration. ☐ 6. The high mobility of a proton in water is explained by
☐ 3. The law of the independent migration of ions states the Grotthuss mechanism.
that the molar conductivity, in the limit of zero concen- ☐ 7. The mobility of an ion can be related to its limiting
tration, is the sum of contributions from its individual molar conductivity and, via the Einstein relation, to its
ions. diffusion coefficient.

Checklist of equations
Property Equation Comment Equation number

Viscosity of a liquid η = η0 e Ea / RT
Over a narrow temperature range 16B.2

Conductivity κ = Gl/A, G = 1/R Definition 16B.3

Molar conductivity Λm = κ/c Definition 16B.4

Kohlrausch’s law Λm = Λ°m − K c1/2 Empirical observation 16B.5

Law of independent migration of ions Λ°m = ν+λ+ + ν−λ− Limiting law 16B.6

Stokes’ law Fviscous = fs, f = 6πηa 16B.7

Drift speed s = uE Defines u 16B.8b

Ion mobility u = ze /6πηa Assumes Stokes’ law 16B.9

Conductivity and mobility λ = zuF 16B.10

Molar conductivity and mobility Λ°m = (z+u+ν+ + z−u−ν−)F 16B.11a

Drift speed s = DF /RT F is a general (molar) force 16B.12

Einstein relation u = zDF /RT 16B.13

Nernst–Einstein equation Λ°m = (ν z D+ + ν z D− )( F /RT )

2
+ +
2
− −
2
16B.15
FOCUS 16C Diffusion

16C.1 The thermodynamic view

➤ Why do you need to know this material?
At constant temperature and pressure, the maximum non-
The diffusion of chemical species through space deter-
expansion work that can be done by a spontaneous process
mines the rates of many chemical reactions in chemical
is equal to the change in the Gibbs energy (Topic 3D). In this
reactors, living cells, and the atmosphere.
case the spontaneous process is the spreading of a solute, and
➤ What is the key idea? the work it could achieve per mole of solute molecules can be
identified with the change in the chemical potential of the sol-
Molecules and ions tend to spread into a uniform distribu-
ute: dwm = dμ. The difference in chemical potential between
tion.
the locations x + dx and x is
➤ What do you need to know already?  ∂µ 
dµ = µ( x + dx ) − µ( x ) =  dx
This Topic draws on arguments relating to the calculation  ∂ x  T , p
of flux (Topic 16A) and the notion of drift speed intro-
so the molar work associated with migration through dx is
duced in Topic 16B. It also uses the concept of chemical
potential to discuss the direction of spontaneous change  ∂µ 
dw m =  dx
(Topic 5A). The final section uses a statistical argument  ∂ x  T , p
like that used to discuss the properties of a random coil
The work done in moving a distance dx against an opposing
in Topic 14D.
force F (in this context, a molar quantity) is dwm = −F dx. By
comparing the two expressions for dwm it is seen that the slope
of the chemical potential with respect to position can be inter-
preted as an effective force per mole of molecules. This ther-
That solutes in gases, liquids, and solids have a tendency to modynamic force is written as
spread can be discussed from three points of view. The ther-
 ∂µ  Thermodynamic force
modynamic viewpoint makes use of the Second Law of ther- F = −  (16C.2)
 ∂ x  T , p [definition]
modynamics and the tendency for entropy to increase and, if
the temperature and pressure are constant, for the Gibbs en- There is not a real force pushing the molecules down the slope
ergy to decrease. The second approach is to set up a differential of the chemical potential: the apparent force represents the spon-
equation for the change in concentration in a region by con- taneous tendency of the molecules to disperse as a consequence
sidering the flux of material through its boundaries. The third of the Second Law and the tendency towards greater entropy.
approach is based on a model in which diffusion is imagined In a solution in which the activity of the solute is a, the
as taking place in a series of random small steps. chemical potential is μ = μ + RT ln a. The thermodynamic
⦵

Several derivations in this Topic use Fick’s first law of dif- force can therefore be written in terms of the gradient of the
fusion, which is discussed in Topic 16A and repeated here for logarithm of the activity:
convenience: ⦵
(∂μ/∂x)T,p = ∂(μ + RT ln a)/∂x
dN Fick’s first law
J (number) = − D  (16C.1a)  ∂µ   ∂ln a 
dx [number]
F = − = − RT   (16C.3a)
 ∂x  T , p  ∂x  T , p
where N is the number density and D is the diffusion coef-
ficient. In a number of cases it is more convenient to discuss If the solution is ideal, a may be replaced by c/c , where c is the ⦵

the flux in terms of the amount of molecules and the molar molar concentration and c is its standard value (1 mol dm−3):
⦵

concentration, c. Division by Avogadro’s constant turns eqn

16C.1a into d ln y/dx = (1/y)(dy/dx)

dc RT  ∂c 
F = − RT  ∂ln(c / c ) 
−−
○
Fick’s first law =−
J (amount) = − D  (16C.1b)  (16C.3b)
dx [amount]  ∂x  T ,p
c  ∂x  T , p
 16C Diffusion 707

Example 16C.1 Self-test 16C.1 Suppose that the concentration of a solute

Calculating the thermodynamic force decreases exponentially to the right as c( x ) = c0e − x /l . Derive an
Suppose that the concentration of a solute varies linearly expression for the thermodynamic force at any position.
along x according to c = c0 + αx, where c0 is the concentra-
Answer: F = RT /l

tion at x = 0. Find an expression for the thermodynamic force

at position x, and evaluate this force at x = 0 and x = 1.0 cm The thermodynamic force acts in many respects like a real
for the case where c0 = 1.0 mol dm−3 and α = 10 mol dm−3 m−1. physical force. In particular it is responsible for accelerating
Take T = 298 K. solute molecules until the viscous drag they experience bal-
Collect your thoughts You will need to use eqn 16C.3b to find ances the apparent driving force and they settle down to a
the force, so begin by evaluating ∂c / ∂ x . steady ‘drift speed’ through the medium. By considering the
balance of apparent driving force and the retarding viscous
The solution The gradient of the molar concentration is force it is possible to derive Fick’s first law of diffusion and re-
 ∂c   ∂(c0 + α x )  late the diffusion coefficient to the properties of the medium.
 ∂ x  =  ∂x  = α
T ,p T ,p

Then, from eqn 16C.3b, the thermodynamic force is How is that done? 16C.1 Deriving Fick’s first law of diffusion

(∂c/∂x)T,p = α; c = c0 + α x
The flux due to a concentration gradient is the amount (in
RT  ∂c  α RT moles) of molecules passing through an area A in an interval
F = − c   = − c +αx Δt divided by the area and the interval. This flux can be related
 ∂x  T , p 0
to the drift speed, s, by using an approach like that used in
At x = 0, Topic 16A for diffusion in a gas.
(10moldm −3 m −1 )× (8.3145JK −1 mol −1 )× (298K)
F =− Step 1 Find an expression for the flux due to molecules moving
1.0moldm −3 at the drift speed
N
 In a time interval Δt all the particles within a distance sΔt can
= − 2.5 ×10 Jm −1 mol −1
4
pass through the window, which means that all of the parti-
or −25 kN mol−1. A similar calculation at x = 1.0 cm = 1.0 × cles in a volume sΔtA pass through the window (there is no
10−2 m gives the force as −23 kN mol−1. reverse flux because, unlike in a gas, in this model all the sol-
ute molecules are moving down the concentration gradient).
Comment. The negative sign indicates that the force is Hence, the amount (in moles) of solute molecules that can
towards the left (towards negative x), because the concentra- pass through the window is sΔtAc. The flux J is this number
tion increases towards the right (as c0 + α x ) and there is there- divided by the area A and by the time interval Δt:
fore a tendency for the solute to migrate to the left under the
J (amount) = sc
influence of that apparent force. The magnitude of the thermo-
dynamic force decreases as x increases because the gradient of Step 2 Find an expression for the drift speed
ln(c/c0), which is α /(c0 + α x ), becomes smaller on going to the
The apparent driving force (per mole) acting on the solute
right (Fig. 16C.1).
molecules is F = −(RT/c)dc/dx, eqn 16C.3b. The molecules
also experience a viscous drag which is assumed to be propor-
Concentration, c
tional to the speed: expressed as a molar quantity this force is
written NAfs, where f is a constant, the ‘frictional constant’,
Constant depending on the medium. When these two forces are in
positive
balance the molecules will be moving at the drift speed. It
Property

gradient

therefore follows that, with R/NA = k,

RT dc RT dc kT dc
Thermodynamic force, F N A fs = − hence s = − =−
c dx N A fc dx fc dx
The negative sign arises because the molecules are moving
opposite to the direction of increasing concentration.
Position, x
Step 3 Combine the two expressions
Figure 16C.1 The thermodynamic force is proportional to
−∂ In c / ∂ x = − (1/ c )∂c / ∂ x . The force thus drives the molecules Now substitute the expression for the drift speed into that for
from a region with higher concentration to one with lower the flux to give
concentration, and becomes smaller in magnitude as the kT dc kT dc
J (amount) = sc = − c=−
concentration increases. fc dx f dx
708 16 Molecules in motion

This expression has the same form as Fick’s first law, eqn
16C.1b, that the flux is proportional to the concentration gra- 16C.2 The diffusion equation
dient. In addition, the diffusion constant D can be identified
as kT/f, which is the Stokes–Einstein relation: Diffusion results in the modification of the distribution of
concentration of the solute (or of a physical property) as inho-
 kT (16C.4a)
D= mogeneities disappear. The discussion is expressed in terms of
f Stokes–Einstein relation
the diffusion of molecules, but similar arguments apply to the
diffusion of other entities, such as ions, and of various physical
The constant f that appears in the Stokes–Einstein rela- properties, such as temperature.
tion can be inferred from the hydrodynamic result known as
Stokes’ law for the viscous drag (this law is used in Topic 16B).
According to this law the magnitude of the viscous force is
(a) Simple diffusion
6πηas for a spherical particle of radius a. It therefore follows The diffusion equation, one of the most important equations
that f = 6πηa, and substituting this into eqn 16C.4a gives the for discussing the properties of fluids, is an equation that ex-
Stokes–Einstein equation presses the rate of change of concentration of a species in terms
of the inhomogeneity of its concentration. It is also called
kT ‘Fick’s second law of diffusion’, but that name is now rarely
D=
6πηa  Stokes–Einstein equation (16C.4b)
used. The diffusion equation can be derived on the basis of
Fick’s first law.
This equation is an explicit relation between the diffusion co-
efficient and the viscosity for a species of hydrodynamic radius How is that done? 16C.2 Deriving the diffusion equation
a and confirms that D is inversely proportional to η.
The drift speed can be related to the diffusion constant and The diffusion equation is developed by considering the net
the thermodynamic force by equating two expressions for the flux of particles entering a thin slab of cross-sectional area
flux, J = sc and J = −D dc/dx, to obtain s = −(D/c) dc/dx. The A that extends from x to x + l (Fig. 16C.2) and therefore has
concentration gradient can be expressed in terms of the ther- volume Al.
modynamic force, that is, eqn 16C.3b rearranged in the form
dc /dx = −cF /RT to give Volume, Al Flux JR
Area, A
DF Flux JL
s=  (16C.5)
RT
This relation provides a way to estimate the thermodynamic
force from measurements of the drift speed and the diffusion
coefficient.
x+l
x

Brief illustration 16C.1 Figure 16C.2 The net flux in a thin slab is the difference
between the flux entering from the region of high
Laser measurements show that a particular molecule has a concentration (on the left) and the flux leaving to the
drift speed of 1.0 μm s −1 in water at 25 °C, at which ­temperature region of low concentration (on the right).
the diffusion coefficient is 5.0 × 10−9 m2 s −1. The corresponding
thermodynamic force calculated using eqn 16C.5, rearranged Step 1 Find an expression for the net rate of change of the
into the form F = sRT/D, is concentration in the slab due to particles entering from each
side
(1.0 ×10−6 ms −1 ) × (8.3145JK −1 mol −1 ) × (298K) If the flux of molecules from the left is JL , then the rate at
F=
5.0 ×10−9 m 2 s −1 which the particles enter the slab is JL A. The rate of increase

N of the molar concentration inside the slab due to the flux in
= 5.0 ×105 Jm −1 mol −1 from the left is
 ∂c  JL A JL
or about 500 kN mol−1. This thermodynamic force is many  ∂t  = Al = l
L
times that of gravity, which explains why solutions do not
sediment. Molecules also flow out of the right face of the slab. If this flux
is JR, then by a similar argument
 16C Diffusion 709

 ∂c  JR
 ∂t  = − l Spreads
R

Note the minus sign: this flux reduces the concentration. The

Concentration, c
net rate of change of concentration in the slab is Negative
curvature
∂ c  ∂ c   ∂c  J −J
= + = L R
∂t  ∂t  L  ∂t  R l Fills

Step 2 Relate the fluxes to the concentration gradients Positive

curvature
From Fick’s first law (eqn 16C.1b), each flux can be expressed
in terms of the diffusion coefficient and the concentration Position, x
gradient at each face: Figure 16C.3 The diffusion equation implies that, over time,
peaks in a distribution (regions of negative curvature) spread and
J = −D(∂c/∂x) troughs (regions of positive curvature) fill in.
 ∂c   ∂c    ∂ c   ∂c  
J L − J R = − D   + D   = D   −   
 ∂x  L  ∂x  R  ∂ x  R  ∂ x  L  • Where the curvature is positive (a dip, Fig. 16C.3),
the change in concentration with time is positive; the

Physical interpretation
where (∂c / ∂ x )L is the concentration gradient at the left face of
dip tends to fill.
the slab, and similarly (∂c / ∂ x )R is that at the right face. The
concentration gradients at the two faces can be expressed in • Where the curvature is negative (a heap), the change
terms of the gradient (the first derivative of the concentration) in concentration with time is negative; the heap
at the centre of the slab, (∂c / ∂ x )0 , and the first derivative of tends to spread.
that gradient (which is the second derivative of the concentra- • If the curvature is zero, then the concentration is
tion), ∂2c /∂x 2. The distances of the faces from the centre are 12 l constant in time.
in each direction, so it follows that
The diffusion equation can be regarded as a mathemati-
 ∂ c   ∂c    ∂c  l ∂ c   ∂c 
2
l ∂c  ∂c
2 2 cal formulation of the intuitive notion that there is a natural
 ∂ x  −  ∂ x  =  ∂ x  + 2 ∂ x 2  −  ∂ x  − 2 ∂ x 2  = l ∂ x 2 ­tendency for the wrinkles in a distribution to disappear.
R L  0   0 

This expression is then substituted into that for the difference

Brief illustration 16C.2
of the fluxes to give
If a concentration across a small region of space varies linearly
(∂c/∂x)R − (∂c/∂x)L = l(∂2c/∂x 2)
as c = c0—α x then it follows that ∂ 2c/∂ x2 = 0 and so from eqn
∂ 2c 16C.6 ∂c/∂t = 0. The concentration does not vary with time
J L − J R = Dl because the flow into one face of a slab is exactly matched by the
∂x 2
flow out from the opposite face (Fig. 16C.4a). If the concentra-
Step 3 Combine the expression for the net flux with that for the tion varies as c = c0 − 12 β x2 then ∂ 2c/∂ x2 = −β and consequently
time dependence of concentration ∂c/∂t = –Dβ. The concentration decreases with time because the
Substitute the last expression into the equation for the rate of flow out of the slab is greater than the flow into it (Fig. 16C.4b).
change of concentration, cancel the l, and obtain the diffusion
equation:
Concentration

 ∂c ∂ 2c (16C.6)
=D 2
∂t ∂x Diffusion equation
Flux

The diffusion equation shows that the rate of change of

(a) (b)
c­oncentration in a region is proportional to the curvature
(more precisely, to the second derivative) of the concentra- Position
tion with respect to distance in that region. If the concentra- Figure 16C.4 The two instances treated in Brief illustration
tion changes sharply from point to point (if the distribution is 16C.2: (a) linear concentration gradient, (b) parabolic
highly wrinkled), then the concentration changes rapidly with concentration gradient.
time. Specifically:
710 16 Molecules in motion

(b) Diffusion with convection Brief illustration 16C.3

Convection is the bulk motion of regions of a fluid. This If it is assumed that the concentration across a small region of
­process contrasts with diffusion in which molecules move space varies linearly as c = c0 − α x, then ∂c/∂x = −α . If there
­individually through the fluid. The flux due to convection can is convective flow at velocity v, it follows from eqn 16C.8 that
be analysed in a similar way to diffusion. ∂c/∂t = α v. The concentration in the slab increases because
the convective flow into the left face outweighs the flow out
from the right face; with this linear concentration depend-
How is that done? 16C.3 Evaluating the change in ence, there is no diffusion. If α = 0.010 mol dm−3 m−1 and v =
concentration due to convection +1.0 mm s −1,
As in previous calculations, imagine the flux of molecules ∂c
= (0.010moldm −3 m −1 ) × (1.0 ×10−3 ms −1 )
through an area A in an interval Δt but now due to convective ∂t
flow in which the fluid moves at a speed v. = 1.0 ×10−5 moldm −3 s −1
Step 1 Express the rate of change of concentration in terms of The concentration increases at the rate of 10 μmol dm−3 s −1.
the net flux
As in the derivation of the diffusion equation, When diffusion and convection occur together, the total
∂c J L,conv − J R ,conv rate of change of concentration in a region is the sum of the
= two effects, which is described by the generalized diffusion
∂t l
equation:
where J L,conv and J R ,conv are, respectively, the fluxes from the
left into and on the right out of the slab, but here due to ∂c ∂ 2c ∂c
­convection. = D 2 −v  Generalized diffusion equation (16C.9)
∂t ∂x ∂x
Step 2 Evaluate the net flux A further refinement, which is important in chemistry, is the
In an interval Δt all the particles within a distance vΔt and possibility that the concentrations of molecules may change as
therefore in the volume AvΔt pass through a face of the slab. a result of reaction. When reactions are included in eqn 16C.9
If the molar concentration at the relevant face is c, then the (in Topic 18B) a differential equation is obtained that can be
amount passing through that face is cAvΔt. The ‘convective used to discuss the properties of reacting, diffusing, convect-
flux’ is this amount divided by the area of the face and the ing systems. This equation is the basis for modelling reactors
time interval: in the chemical industry and the utilization of resources in
cAv∆t living cells.
J conv = = cv Convective flux (16C.7)
A∆t
The concentrations on the left (c L) and right (c R) faces of the (c) Solutions of the diffusion equation
slab are related to the concentration at its centre, c0, by
The diffusion equation (eqn 16C.6) is a second-order
 ∂c   ∂c  ­d ifferential equation with respect to space and a first-order
cR = c0 + l  
1
cL = c0 − l  
1
 ∂x 
2 2
 ∂x  ­d ifferential equation with respect to time. To find solutions
where the first derivatives are evaluated at the centre of the it is necessary to know two boundary conditions for the spa-
slab. It follows from eqn 16C.7 that tial dependence and a single initial condition for the time
­dependence.
J = cv As an illustration, consider an arrangement in which there
 is a layer of a solute (such as sugar) at the bottom of a tall
 ∂c     ∂c  
J L,conv − J R ,conv = (cL − cR )v = c0 − 12 l    v − c0 + 12 l    v beaker of water (which for simplicity may be taken to be in-
  ∂ x     ∂x  
finitely tall), with base area A; x is the distance measured up
 ∂c  from the base. At t = 0 it is assumed that all N0 particles are
= −   lv
 ∂x  concentrated on the yz-plane at x = 0: this is the initial condi-
tion. The two boundary conditions are derived from the re-
Step 3 Evaluate the net rate of change of concentration
quirements that the concentration must everywhere be ­finite
Now substitute this result into the expression for the rate of and the total amount of particles present is n0 (with n0 = N0/NA)
change of concentration derived in Step 1, and obtain at all times. With these conditions,
 ∂c  ∂c  (16C.8)
= −  v n0 2
One-dimensional
∂t  ∂x  Convection c( x , t ) = e − x /4 Dt  (16C.10)
A(πDt )1/2 diffusion
 16C Diffusion 711

2.5
How is that done? 16C.4 Evaluating the average
0.05
2
displacement in a one-dimensional system
Concentration, c/(n0/A)

0.1
Increasing time To calculate the average value of the displacement x, denoted
1.5 〈x〉, you need to use eqn 16C.10 to find an expression for the
probability density of the particles P(x), defined such that
0.3
1 P(x)dx is the probability of finding a particle between x and
x + dx. Then
1.0
0.5 ∞
〈 x 〉 = ∫ xP( x )d x
0

0
0 0.5 1 1.5 2 Step 1 Set up an expression for the probability density
Distance,x/x0
The number of particles in a slab of thickness dx at x and at
Figure 16C.5 The concentration profiles above a plane from time t is the volume of the slab, Adx, multiplied by the molar
which a solute is diffusing into pure solvent. The curves are concentration at that location (and time) and Avogadro’s con-
labelled with the corresponding value of Dt, and x0 = {4 Dt}1/2. stant: N(x,t) = c(x,t)NA Adx. The molar concentration at the
position and time of interest is given by eqn 16C.10. The total
number of particles is NAn0, where n0 is the total amount, so
as may be verified by direct substitution. Figure 16C.5 shows
the probability of finding a molecule in the slab is
the shape of the concentration distribution at various times
and illustrates how the concentration spreads. 2 /4 Dt
c( x , t ) = n0
e− x
Another useful result is for the diffusion in three dimen- A( πDt ) 1/2

sions arising from an initially localized concentration of solute N Ac( x ,t )Adx A n0 2

(a sugar lump suspended in an infinitely large flask of water). P( x )dx = = e − x /4 Dt dx

N An0 n0 A(πDt )1/2
The concentration of diffused solute is spherically symmetri- 1 2

cal, and at a radius r is = e − x /4 Dt dx

(πDt )1/2

n0 2
Three-dimensional Step 2 Evaluate the integral
c (r , t ) = e − r /4 Dt  (16C.11)
8(πDt )3/2 diffusion
The integration required is

Other chemically (and physically) interesting arrangements, Integral G.2

such as transport of substances across biological membranes    1/2
∞ 1 1 ∞  Dt 
∫ (πDt )1/2 ∫0
2
− x 2 /4 Dt
can be treated, but the mathematical forms of the solutions are x e− x /4 Dt
dx = xe d x = 2  π 
0 (πDt )1/2
more cumbersome.
The diffusion equation is useful for the experimental deter-
mination of diffusion coefficients. In the capillary technique, That is, the average displacement of a diffusing particle in a
a capillary tube, open at one end and containing a solution, time t in a one-dimensional system is
is immersed in a well-stirred larger quantity of solvent, and 1/2
the change of concentration in the tube is monitored. The  Dt  (16C.12)
〈x 〉 = 2  
 π Mean displacement
­solute diffuses from the open end of the capillary at a rate [one dimension]
that can be calculated by solving the diffusion equation with
the appropriate boundary and initial conditions, so D may A similar calculation shows that the root-mean-square dis-
be ­determined. In the diaphragm technique, the diffusion placement in the same time is
occurs through the capillary pores of a sintered glass dia-
Root-mean-square displacement
phragm ­separating the well-stirred solution and solvent. The 〈x 2 〉1/2 = (2Dt)1/2 (16C.13a)
[one dimension]
concentrations are monitored and then related to the diffusion
equation that has been solved for this arrangement. Diffusion This result is a useful measure of the spread of particles when
coefficients may also be measured by a number of other tech- they can diffuse in both directions from the origin, because
niques, including NMR spectroscopy (Topic 12C). in that case 〈x〉 = 0 at all times. The time-dependence of the
The diffusion equation can be used to predict the concen- root-mean-square displacement for particles with a typical
tration of particles (or the value of some other physical quan- diffusion coefficient in a liquid (D = 5 × 10−10 m2 s−1) is illus-
tity, such as the temperature in a non-uniform system) at any trated in Fig. 16C.6. The graph shows that diffusion is a very
location. It can also be used to calculate the average displace- slow process (which is why solutions are stirred, to encourage
ment of the particles in a given time. mixing by convection).
712 16 Molecules in motion

0 1m Step 1 Set up an expression for the probability of achieving a

given final displacement
–2 1 cm Many sequences of individual steps can arrive at a given final
1 mm displacement. The number of these sequences is equal to the
log(〈x 2〉1/2/m)

–4 number of ways of choosing NR steps to the right and NL =

N − NR steps to the left:
–6 1 μm
N! N!
W= =
–8 N L ! N R ! ( N − N R )! N R !
1 nm
1 μs 1 ms 1s 1h 1d 1y At each step, the molecule can step to the left or right, so the
–10
total number of possible sequences of steps is 2N. The prob-
–10 –6 –2 0 2 6
log(t/s) ability of achieving a final displacement nd, P(nd), is therefore

Figure 16C.6 The root-mean-square distance covered by particles W N!

P(nd ) = =
with D = 5 × 10−10 m2 s−1. Note the great slowness of diffusion. 2 N ( N − N R )! N R !2 N

Step 2 Simplify this expression by using Stirling’s approxima-

tion
In three dimensions the root-mean-square displacement is
given by a similar expression: This expression can be simplified by taking its logarithm to
give
Root-mean-square displacement
〈r2〉1/2 = (6Dt)1/2 [three dimensions]
(16C.13b) ln P = ln N ! − {ln( N − N R )! + ln N R ! + ln2 N }
and then using Stirling’s approximation in the form

ln x! ≈ ln(2π)1/2 + ( x + 12 ) ln x − x
16C.3 The statistical view
to obtain
An intuitive picture of diffusion is of the particles moving 1
in a series of small steps and gradually migrating from their ln P = − ln(2π )1/2 2 N + ( N + 12 )ln
1 − N R /N
original positions. This picture suggests a model in which 1 − N R /N 1
each particle jumps through a distance d after a time τ. The + N R ln − 2 ln N R
N R /N
total distance travelled by a particle in time t is therefore td/τ.
For reasons that will become clear shortly, it is convenient to
However, it is most unlikely that a particle will end up at this
introduce the new variable μ:
distance from the origin because each jump might be in a dif-
ferent direction. NR
µ= − 1

The discussion is simplified by allowing the particles to N 2

travel only along a straight line (the x-axis) with each step a from which definition it follows that 1− N R /N = 12 − µ and
jump through distance d to the left or to the right. This model N R /N = 12 + µ . With these substitutions the expression for ln P
is called the one-dimensional random walk. The probabil- can be written in terms of N and μ alone:
ity that the walk will end up at a specific distance from the ln P = − ln(2π )1/2 2 N − ( N + 12 )ln( 12 − µ ) + N ( µ + 12 )ln( 12 − µ )
origin can be calculated by considering the statistics of the
− N ( µ + 12 )ln( µ + 12 ) − 12 ln N ( µ + 12 )
process.1
Step 3 Expand the logarithms
How is that done? 16C.5 Evaluating the probability
Because the probability of taking a step to the right or to the
distribution for a one-dimensional random walk left is the same, it is expected that after many steps the total
Imagine that a molecule has made N steps in total, NR of number to the right will be very close to half the number of
which are to the right and NL to the left. The displacement steps. That is NR /N ≈ 12 . It follows that µ <<1, so you can use
from the origin is therefore (NR − NL)d, which is written nd the series expansion
with n = NR − NL.
ln( 12 ± µ ) = − ln2 ± 2 µ − 2 µ 2 +

and retain terms up to that in μ2. After rather of lot of algebra

1
The calculation is essentially the same as in the discussion of the ran- (see A deeper look 11 on the website, where the details are
dom coil structures of denatured polymers (Topic 14D). given) you will obtain
 16C Diffusion 713

ln P = − ln(2πN )1/2 2 N + ln2 N +1 − 2( N − 1)µ 2 By comparing the two exponents in eqn 16C.10 and eqn
16C.14 it is possible to relate the diffusion coefficient D to the
At this point take antilogarithms of both sides and use N >> 1: step length d and the time between jumps, τ. The result is the
2 2 2
Einstein–Smoluchowski equation:
2 N +1 e −2( N −1) µ 2e −2( N −1) µ 2e −2 N µ
P= 1/2 = 1/2 ≈
N
2 (2πN ) (2πN ) (2πN )1/2 d2
D=  Einstein–Smoluchowski equation (16C.15)
2τ
Step 4 Recast the expression for the probability in terms of the
time for each step Brief illustration 16C.4
The exponent Nμ2 can be rewritten
Suppose that in aqueous solution an SO42− ion jumps through
μ = NR/N−1/2 n = NR − NL its own diameter of 500 pm each time it makes a move, then
because D = 1.1 × 10−9 m2 s −1 (as deduced from mobility meas-
(2 N R − N )2 ( N R − N L )2 n 2
Nµ2 = = = urements, Topic 16B), it follows from eqn 16C.15 that
4N 4N 4N
d2 (500 ×10−12 m)2
The final distance from the origin, x, is equal to nd, and the τ= = = 1.1×10−10 s
2D 2 × (1.1×10−9 m 2 s −1 )
number of steps taken in a time t is N = t/τ. It follows that
N µ 2 = n 2 /4 N = τ x 2 /4td 2 . Substitution of these expressions for or τ = 110 ps. Because τ is the time for one jump, the ion makes
N and N µ 2 into the expression for P gives about 1 × 1010 jumps per second.

  2τ 
1/2
(16C.14)
2
P( x ,t ) =   e − x τ /2td
2
The Einstein–Smoluchowski equation makes the connec-
 πt  One-dimensional random walk
tion between the microscopic details of particle motion and the
macroscopic parameters relating to diffusion. It also brings the
The differences of detail between eqns 16C.10 (for one-di- discussion back full circle to the properties of the perfect gas
mensional diffusion) and 16C.14 arises from the fact that in treated in Topic 16A. If d/τ as is interpreted as vmean, the mean
the present calculation the particles can migrate in either di- speed of the molecules, and d is interpreted as a mean free path
rection from the origin. Moreover, they can be found only at λ , then the Einstein–Smoluchowski equation becomes D =
discrete points separated by d instead of being anywhere on 2 d(d/τ) = 2 λ vmean, which is essentially the same expression as
1 1

a continuous line. The fact that the two expressions are so obtained from the kinetic model of gases (eqn 16A.9 of Topic
similar suggests that diffusion can indeed be interpreted as the 16A, D = 13 λ vmean). That is, the diffusion of a perfect gas is a ran-
outcome of a large number of steps in random directions. dom walk with an average step size equal to the mean free path.

Checklist of concepts
☐ 1. A thermodynamic force is an apparent force that that there is a natural tendency for concentration to
mirrors the spontaneous tendency of the molecules to become uniform.
disperse as a consequence of the Second Law and the ☐ 4. Convection is the bulk motion of regions of a fluid.
tendency towards greater entropy.
☐ 5. A model of diffusion is of the particles moving in a
☐ 2. The drift speed is achieved when the viscous retarding series of small steps, a random walk, and gradually
force matches the thermodynamic force. migrating from their original positions.
☐ 3. The diffusion equation (Fick’s second law) can be
regarded as a mathematical formulation of the notion
714 16 Molecules in motion

Checklist of equations
Property Equation Comment Equation number

Fick’s first law J (amount) = − Ddc /dx 16C.1b

Thermodynamic force F = − (∂ µ / ∂x )T , p Definition 16C.2

Stokes–Einstein relation D = kT/f fs is the frictional drag 16C.4a

Drift speed s = DF /RT 16C.5

Diffusion equation ∂c / ∂t = D ∂2c / ∂x 2 One dimension 16C.6

2 2
Generalized diffusion equation ∂c /∂t = D ∂ c /∂ x − v ∂c /∂ x One dimension 16C.9

Mean displacement x = 2(Dt /π )1/2 One-dimensional diffusion 16C.12

Root-mean-square displacement 〈x2〉1/2 = (2Dt)1/2 One-dimensional diffusion 16C.13a

2 1/2 1/2
〈r 〉 = (6Dt) Three-dimensional diffusion 16C.13b
2 2
Probability of displacement P( x , t ) = (2τ /πt )1/2 e − x τ /2td One-dimensional random walk 16C.14

Einstein–Smoluchowski equation D = d /2τ 2

One-dimensional random walk 16C.15
 Exercises and problems 715

FOCUS 16 Molecules in motion

TOPIC 16A Transport properties of a perfect gas

Discussion questions
D16A.1 Explain how Fick’s first law arises from considerations of the flux of D16A.2 Provide molecular interpretations for the dependencies of the
molecules due to a concentration gradient in a perfect gas. diffusion constant and the viscosity on the temperature, pressure, and size of
gas molecules.

Exercises
E16A.1(a) A solid surface with dimensions 2.5 mm × 3.0 mm is exposed to E16A.7(b) Calculate the viscosity of benzene vapour at (i) 273 K, (ii) 298 K,
argon gas at 90 Pa and 500 K. How many collisions do the Ar atoms make with (iii) 1000 K. Take σ as 0.88 nm2.
this surface in 15 s?
E16A.8(a) Use the experimental value of the coefficient of viscosity for neon
E16A.1(b) A solid surface with dimensions 3.5 cm × 4.0 cm is exposed to
(Table 16A.1) to estimate the collision cross-section of Ne atoms at 273 K.
helium gas at 111 Pa and 1500 K. How many collisions do the He atoms make
E16A.8(b) Use the experimental value of the coefficient of viscosity for nitrogen
with this surface in 10 s?
(Table 16A.1) to estimate the collision cross-section of the molecules at 273 K.
E16A.2(a) Calculate the diffusion constant of argon at 20 °C and (i) 1.00 Pa,
E16A.9(a) An effusion cell has a circular hole of diameter 2.50 mm. If the molar
(ii) 100 kPa, (iii) 10.0 MPa; take σ = 0.36 nm2. If a pressure gradient of
mass of the solid in the cell is 260 g mol−1 and its vapour pressure is 0.835 Pa at
1.0 bar m−1 is established in a pipe, what is the flow of gas due to diffusion at
400 K, by how much will the mass of the solid decrease in a period of 2.00 h?
each pressure?
E16A.9(b) An effusion cell has a circular hole of diameter 3.00 mm. If the molar
E16A.2(b) Calculate the diffusion constant of nitrogen at 20 °C and (i) 100.0 Pa,
mass of the solid in the cell is 300 g mol−1 and its vapour pressure is 0.224 Pa at
(ii) 100 kPa, (iii) 20.0 MPa; take σ = 0.43 nm2. If a pressure gradient of
450 K, by how much will the mass of the solid decrease in a period of 24.00 h?
1.20 bar m−1 is established in a pipe, what is the flow of gas due to diffusion
−1
at each pressure? E16A.10(a) A solid compound of molar mass 100 g mol was introduced into a
−1 −1 container and heated to 400 °C. When a hole of diameter 0.50 mm was opened
E16A.3(a) Calculate the thermal conductivity of argon (CV,m = 12.5 J K mol ,
in the container for 400 s, a mass loss of 285 mg was measured. Calculate the
σ = 0.36 nm2) at 298 K.
−1 −1 vapour pressure of the compound at 400 °C.
E16A.3(b) Calculate the thermal conductivity of nitrogen (CV,m = 20.8 J K mol , −1
E16A.10(b) A solid compound of molar mass 200 g mol was introduced into a
σ = 0.43 nm2) at 298 K.
container and heated to 300 °C. When a hole of diameter 0.50 mm was opened
E16A.4(a) Use the experimental value of the thermal conductivity of neon in the container for 500 s, a mass loss of 277 mg was measured. Calculate the
(Table 16A.1) to estimate the collision cross-section of Ne atoms at 273 K. vapour pressure of the compound at 300 °C.
E16A.4(b) Use the experimental value of the thermal conductivity of
E16A.11(a) A manometer was connected to a bulb containing an unknown
nitrogen (Table 16A.1) to estimate the collision cross-section of N2 molecules
gas under slight pressure. The gas was allowed to escape through a small
at 298 K.
pinhole, and the time for the manometer reading to drop from 75 cm to
E16A.5(a) Calculate the flux of energy arising from a temperature gradient of 50 cm was 52 s. When the experiment was repeated using nitrogen (for
10.5 K m−1 in a sample of argon in which the mean temperature is 280 K. The which M = 28.02 g mol−1) the same fall took place in 42 s. Calculate the molar
necessary data needed to calculate the thermal conductivity are in Exercise mass of the unknown gas. Hint: The pressure changes and, as a consequence,
E16A.3(a). so does the rate of effusion; note, however, that the change is the same in
E16A.5(b) Calculate the flux of energy arising from a temperature gradient both cases.
of 8.5 K m−1 in a sample of N2 in which the mean temperature is 290 K. The E16A.11(b) A manometer was connected to a bulb containing nitrogen
necessary data needed to calculate the thermal conductivity are in Exercise under slight pressure. The gas was allowed to escape through a small
E16A.3(b). pinhole, and the time for the manometer reading to drop from 65.1 cm to
42.1 cm was 18.5 s. When the experiment was repeated using a fluorocarbon
E16A.6(a) In a double-glazed window, the panes of glass are separated
gas, the same fall took place in 82.3 s. Calculate the molar mass of the
by 1.0 cm and the space is filled with a gas with thermal conductivity
fluorocarbon.
24 mW K−1 m−1. What is the rate of transfer of heat by conduction from the
3
warm room (28 °C) to the cold exterior (−15 °C) through a window of area E16A.12(a) A space vehicle of internal volume 3.0 m is struck by a meteor and
1.0 m2? You may assume that one pane of glass is at the same temperature as a hole of radius 0.10 mm is formed. If the oxygen pressure within the vehicle
the inside and the other as the outside. What power of heater is required to is initially 80 kPa and its temperature 298 K, how long will the pressure take to
make good the loss of heat? fall to 70 kPa assuming that the temperature is held constant?
2 3
E16A.6(b) Two sheets of copper of area 2.00 m are separated by 5.00 cm in E16A.12(b) A container of internal volume 22.0 m was punctured, and a hole
N2(g). What is the rate of transfer of heat by conduction from the warm sheet of radius 0.050 mm was formed. If the nitrogen pressure within the container
(70 °C) to the cold sheet (0 °C)? Refer to the Resource section for any necessary is initially 122 kPa and its temperature 293 K, how long will the pressure take
data. to fall to 105 kPa assuming that the temperature is held constant?
E16A.7(a) Calculate the viscosity of air at (i) 273 K, (ii) 298 K, (iii) 1000 K. Take
σ as 0.40 nm2 and M as 29.0 g mol−1.
716 16 Molecules in motion

Problems
P16A.1‡ A. Fenghour et al. (J. Phys. Chem. Ref. Data 24, 1649 (1995)) compiled P16A.4 A Knudsen cell was used to determine the vapour pressure of
an extensive table of viscosity coefficients for ammonia in the liquid and germanium at 1000 °C. During an interval of 7200 s the mass loss through a
vapour phases. Deduce the effective molecular diameter of NH3 based on hole of radius 0.50 mm amounted to 43 μg. What is the vapour pressure of
each of the following vapour-phase viscosity coefficients: (a) η = 9.08 × germanium at 1000 °C? Assume the gas to be monatomic.
10−6 kg m−1 s−1 at 270 K and 1.00 bar; (b) η = 1.749 × 10−5 kg m−1 s−1 at 490 K and
P16A.5 An atomic beam is designed to function with (a) cadmium,
10.0 bar.
(b) mercury. The source is an oven maintained at 380 K, and the vapour
P16A.2 Calculate the ratio of the thermal conductivities of gaseous hydrogen escapes though a small slit of dimensions 10 mm by 1.0 × 10−2 mm. The
at 300 K to gaseous hydrogen at 10 K. Hint: Think about the modes of motion vapour pressure of cadmium is 0.13 Pa and that of mercury is 12 Pa at this
that are thermally active at the two temperatures. temperature. What is the number of atoms per second in the beams?
P16A.3 Interstellar space is quite different from the gaseous environments we P16A.6 Derive an expression that shows how the pressure of a gas inside
commonly encounter on Earth. For instance, a typical density of the medium an effusion oven (a heated chamber with a small hole in one wall) varies
is about 1 atom cm−3 and that atom is typically H; the effective temperature with time if the oven is not replenished as the gas escapes. Then show that
due to stellar background radiation is about 10 kK. Estimate the diffusion t1/2, the time required for the pressure to decrease to half its initial value, is
coefficient and thermal conductivity of H under these conditions. Compare independent of the initial pressure. Hint: Start from the expression for the rate
your answers with the values for gases under typical terrestrial conditions. of effusion and rewrite it as a differential equation relating dp/dt to p; recall
Comment: Energy is in fact transferred much more effectively by radiation. that pV = NkT can be used to relate the pressure to the number density.

TOPIC 16B Motion in liquids

Discussion questions
D16B.1 Discuss the difference between the hydrodynamic radius of an ion and D16A.2 Discuss the mechanism of proton conduction in water. Might the same
its ionic radius. Explain how it is possible for the hydrodynamic radius of an mechanism also occur in ice?
ion to decrease down a group even though the ionic radius increases.

Exercises
E16B.1(a) The viscosity of water at 20 °C is 1.002 cP and 0.7975 cP at 30 °C. E16B.4(a) The mobility of a chloride ion in aqueous solution at 25 °C is 7.91 ×
What is the energy of activation associated with viscosity? 10−8 m2 s−1 V −1. Calculate the molar ionic conductivity.
E16B.1(b) The viscosity of mercury at 20 °C is 1.554 cP and 1.450 cP at 40 °C. E16B.4(b) The mobility of an ethanoate (acetate) ion in aqueous solution at
What is the energy of activation associated with viscosity? 25 °C is 4.24 × 10−8 m2 s−1 V −1. Calculate the molar ionic conductivity.
+ −8 2 −1 −1
E16B.2(a) The limiting molar conductivities of NaI, NaNO3, and AgNO3 are E16B.5(a) The mobility of a Rb ion in aqueous solution is 7.92 × 10 m s V
2 −1 2 −1 2 −1
12.69 mS m mol , 12.16 mS m mol , and 13.34 mS m mol , respectively (all at 25 °C. The potential difference between two electrodes, separated by
at 25 °C). What is the limiting molar conductivity of AgI at this temperature? 7.00 mm and placed in the solution, is 25.0 V. What is the drift speed of the
Hint: Each limiting molar conductivity can be expressed as a sum of two ionic Rb+ ion?
+ −8 2 −1 −1
conductivities. E16B.5(b) The mobility of a Li ion in aqueous solution is 4.01 × 10 m s V
E16B.2(b) The limiting molar conductivities of KF, KCH3CO2, and Mg(CH3CO2)2 at 25 °C. The potential difference between two electrodes separated by
are 12.89 mS m2 mol−1, 11.44 mS m2 mol−1, and 18.78 mS m2 mol−1, respectively (all 5.00 mm and placed in the solution, is 24.0 V. What is the drift speed of
at 25 °C). What is the limiting molar conductivity of MgF2 at this temperature? the ion?
+ + + −
E16B.3(a) At 25 °C the molar ionic conductivities of Li , Na , and K are E16B.6(a) The mobility of a NO3 ion in aqueous solution at 25 °C is 7.40 ×
3.87 mS m2 mol−1, 5.01 mS m2 mol−1, and 7.35 mS m2 mol−1, respectively. What 10−8 m2 s−1 V −1. Calculate its diffusion coefficient in water at 25 °C.
−
are their mobilities? E16B.6(b) The mobility of a CH3CO2 ion in aqueous solution at 25 °C is 4.24 ×
− − −
E16B.3(b) At 25 °C the molar ionic conductivities of F , Cl , and Br are 10−8 m2 s−1 V −1. Calculate its diffusion coefficient in water at 25 °C.
2 −1 2 −1 2 −1
5.54 mS m mol , 7.635 mS m mol , and 7.81 mS m mol , respectively. What
are their mobilities?

Problems
P16B.1 The viscosity of benzene varies with temperature as shown in the P16B.2 An empirical expression that reproduces the viscosity of water in the
following table. Use the data to infer the activation energy associated with range 20–100 °C is
viscosity.
η 1.3272(20 − θ / °C) − 0.001053(20 − θ / °C)2
θ/°C 10 20 30 40 50 60 70 log =
η20 θ / °C + 105
η/cP 0.758 0.652 0.564 0.503 0.442 0.392 0.358
where η20 is the viscosity at 20 °C. Explore (by using mathematical
software) the possibility of fitting these data to an expression of the form
‡
These problems were provided by Charles Trapp and Carmen Giunta. η = const × e Ea /RT and hence identifying an activation energy for the viscosity.
 Exercises and problems 717

P16B.3 The conductivity of aqueous ammonium chloride at a series difference. Calculate the displacement of each of the ions in (i) centimetres,
of concentrations is listed in the following table. Calculate the molar and (ii) solvent diameters (take as 300 pm), during a half cycle of an applied
conductivity at each concentration, and use the resulting data to determine potential at a frequency of 2.0 kHz. Hint: The drift speed will vary over the
the parameters that occur in the Kohlrausch law. half cycle, so you will need to integrate s over time to find the displacement.
P16B.7‡ G. Bakale et al. (J. Phys. Chem., 12477 (1996)) measured the mobility
c/(mol dm−3) 1.334 1.432 1.529 1.672 1.725
of singly charged C60− ions in a variety of nonpolar solvents. In cyclohexane at
κ/(mS cm−1) 131 139 147 156 164 22 °C (viscosity is 0.93 × 10−3 kg m−1 s−1), the mobility is 1.1 × 10−4 cm2 V −1 s−1.
Estimate the effective radius of the C60− ion. Suggest a reason why there is a
P16B.4 Conductivities are often measured by comparing the resistance of a substantial difference between this number and the van der Waals radius of
cell filled with the sample to its resistance when filled with some standard neutral C60.
solution, such as aqueous potassium chloride. The conductivity of water
P16B.8 Estimate the diffusion coefficients and the effective hydrodynamic
is 76 mS m−1 at 25 °C and the conductivity of 0.100 mol dm−3 KCl(aq) is
1.1639 S m−1. A cell had a resistance of 33.21 Ω when filled with 0.100 mol dm−3 radii of the alkali metal cations in water from their mobilities at 25 °C (refer
KCl(aq) and 300.0 Ω when filled with 0.100 mol dm−3 CH3COOH(aq). What to the Resource section for values of the mobilities). Estimate the approximate
is the molar conductivity of ethanoic (acetic) acid at that concentration and number of water molecules that are dragged along by the cations. Ionic radii
temperature? are given Table 15C.2.
P16B.9‡ (a) A dilute solution of a weak electrolyte AB, which dissociates to
P16B.5 A cell was used to measure the resistance R of a series of solutions.
The cell has been calibrated against a standard solution and as a result the A+ + B−, is prepared with an initial concentration cAB. Suppose that a fraction α of
conductivity of the solution is given by κ = C/R, where C = 0.2063 cm−1. The AB dissociates. Assuming that activities can be approximated by concentrations,
following values of R were found: show that the equilibrium constant K for dissociation may be written
α 2cAB
K= −○−
c/(mol dm )−3
0.00050 0.0010 0.0050 0.010 0.020 0.050 (1 − α )c
R/Ω 3314 1669 342.1 174.1 89.08 37.14 (b) The conductivity of the solution described in (a) is measured as κ, and
the molar conductivity is then calculated as Λm = κ/cAB. However, because
(a) Verify that the molar conductivity follows the Kohlrausch law, find the the degree of dissociation, and hence the concentration of the ions, varies
limiting molar conductivity and the coefficient K. (b) Consider a solution strongly with the initial concentration cAB, values of Λm calculated in this way
of 0.010 mol dm−3 NaI(aq) at 25 °C placed in the cell. Assume that the also vary strongly with cAB. Given that κ can be expected to be proportional
same value of K applies to this solution as to that in (a), and that λ(Na+) to the concentration of the ions, explain why α = Λm/Λm,1, where Λm,1 is the
= 5.01 mS m2 mol−1 and λ(I−) = 7.68 mS m2 mol−1. Predict (i) the molar molar conductivity in the limit of complete dissociation of AB. (c) Substitute
conductivity, (ii) the conductivity, and (iii) the resistance of the solution this expression for α into the above expression for K. You now have two
in the cell. expressions for K: one in terms of α and one in terms of Λm/Λm,1. Equate these
+ + +
two expressions and hence show (by rearranging your expression) that
P16B.6 (a) Calculate the drift speeds of Li , Na , and K in water when a
potential difference of 100 V is applied across a 5.00 cm conductivity cell. 1 1 Λ (1 − α )
= + m2 2
Refer to the Resource section for values of ion mobilities. (b) Calculate Λm Λm,1 Λm,1α
how long it takes each ion to move from one electrode to the other. (c)
In conductivity measurements it is normal to use an alternating potential

TOPIC 16C Diffusion

Discussion questions
D16C.1 Describe the origin of the thermodynamic force. To what extent can it D16C.2 Account physically for the form of the diffusion equation.
be regarded as an actual force?

Exercises
E16C.1(a) The diffusion coefficient of glucose in water at 25 °C is 6.73 × of iodine molecules at 5.0 cm above the original layer after (i) 10 s, (ii) 24 h?
10−10 m2 s−1. Estimate the time required for a glucose molecule to undergo a Assume diffusion is the only transport process, take D = 4.05 × 10−9 m2 s−1, and
root-mean-square displacement of 5.0 mm. assume that the layer of hexane in infinitely deep.
−9 2 −1
E16C.1(b) The diffusion coefficient of H2O in water at 25 °C is 2.26 × 10 m s .
E16C.3(a) Suppose the concentration of a solute decays linearly along
Estimate the time required for an H2O molecule to undergo a root-mean-
the length of a container according to c( x ) = c0 − α c0 x , where c0 is the
square displacement of 1.0 cm.
concentration at x = 0. Calculate the thermodynamic force on the solute at
E16C.2(a) A layer of 20.0 g of sucrose is spread uniformly over a surface of area 25 °C and at x = 10 cm and 20 cm given that the concentration falls to 12 c0
5.0 cm2 and covered in water. What will be the molar concentration of sucrose when x = 10 cm. Hint: Start by finding the value of α.
molecules at 10 cm above the original layer after (i) 10 s, (ii) 24 h? Assume E16C.3(b) Suppose the concentration of a solute varies along the length of a
diffusion is the only transport process, take D = 5.216 × 10−9 m2 s−1, and assume container according to c( x ) = c0 − β c0 x 2, where c0 is the concentration at x = 0.
that the layer of water in infinitely deep. Calculate the thermodynamic force on the solute at 25 °C and x = 8 cm and
E16C.2(b) A layer of 10.0 g of iodine is spread uniformly over a surface of 16 cm given that the concentration falls to 12 c0 when x = 15 cm. Hint: Start by
area 10.0 cm2 and covered in hexane. What will be the molar concentration finding the value of β.
718 16 Molecules in motion

E16C.4(a) Suppose the concentration of a solute follows a Gaussian E16C.6(b) Estimate the effective radius of a glycine molecule in water at 25 °C
2
distribution, c( x ) = c0e −α x , where c0 is the concentration at x = 0, along the given that its diffusion coefficient is 1.055 × 10−9 m2 s−1 and that the viscosity of
length of a container. Calculate the thermodynamic force on the solute at water is 1.00 cP.
20 °C and at x = 5.0 cm given that the concentration falls to 12 c0 when
E16C.7(a) The diffusion coefficient for molecular iodine in benzene is 2.13
x = 5.0 cm.
× 10−9 m2 s−1. How long does a molecule take to jump through about one
E16C.4(b) For the same arrangement as in Exercise E16C.4(a) calculate the
molecular diameter (approximately the fundamental jump length for
thermodynamic force on the solute at 18 °C and x = 10.0 cm given that the
translational motion)?
concentration falls to 12 c0 when x = 10.0 cm. −9 2 −1
E16C.7(b) The diffusion coefficient for CCl4 in heptane is 3.17 × 10 m s .
E16C.5(a) The diffusion coefficient of CCl4 in heptane at 25 °C is 3.17 × How long does a molecule take to jump through about one molecular
10−9 m2 s−1. Estimate the time required for a CCl4 molecule to have a root mean diameter (approximately the fundamental jump length for translational
square displacement of 5.0 mm. motion)? The viscosity of heptane is 0.386 cP.
−9 2 −1
E16C.5(b) The diffusion coefficient of I2 in hexane at 25 °C is 4.05 × 10 m s .
E16C.8(a) What are the root-mean-square distances travelled (in one
Estimate the time required for an iodine molecule to have a root mean square
dimension) by an iodine molecule in benzene and by a sucrose molecule in
displacement of 1.0 cm.
water at 25 °C in 1.0 s? Refer to the Resource section for the necessary data.
E16C.6(a) Estimate the effective radius of a sucrose molecule in water at 25 °C E16C.8(b) About how long does it take for the molecules referred to in
given that its diffusion coefficient is 5.2 × 10−10 m2 s−1 and that the viscosity of Exercise 16C.8(a) to drift to a point (i) 1.0 mm, (ii) 1.0 cm from their starting
water is 1.00 cP. points?

Problems
P16C.1 A dilute solution of potassium permanganate in water at 25 °C was you will need to consider quite long times and short distances. Similarly,
prepared. The solution was in a horizontal tube of length 10 cm, and at first consider slow convection speeds. (c) A different way of plotting this function
there was a linear gradation of intensity of the purple solution from the is first to define x c = x 0 + vt ; xc is the position to which the solute would move
1/2
left (where the concentration was 0.100 mol dm−3) to the right (where the if the only process was convection. Now define z = ( x − x c )/(4 D ) and use
concentration was 0.050 mol dm−3). What is the magnitude and sign of the this quantity to rewrite the expression for c( x , t ) as
thermodynamic force acting on the solute (a) close to the left face of the
container, (b) in the middle, (c) close to the right face. Give the force per mole (4 πD )1/2 c( x , t ) 1 − z 2 /t
= 1/2 e
and force per molecule in each case. c0 t
P16C.2 A dilute solution of potassium permanganate in water at 25 °C was Now plot the right-hand side as a function of z for some representative values
prepared. The solution was in a horizontal tube of length 10 cm, and at first of t.
there was a Gaussian distribution of concentration around the centre of the 2 1/2
2 P16C.7 Calculate the relation between 〈x 〉 and 〈x4〉1/4 for diffusing particles at
tube at x = 0, c( x ) = c0e −α x , with c0 = 0.100 mol dm−3 and a = 0.10 cm−2. Evaluate
a time t if they have a diffusion constant D.
the thermodynamic force acting on the solute as a function of location, x, and
plot the result. Give the force per mole and force per molecule in each case. P16C.8 The diffusion equation is valid when many elementary steps are taken
What do you expect to be the consequence of the thermodynamic force? in the time interval of interest, but the random walk calculation makes it
possible to discuss distributions for short times as well as for long. Use the
P16C.3 Instead of a Gaussian ‘heap’ of solute, as in Problem P16C.2,
expression P(nd) = N!/(N − NR)!NR!2N to calculate the probability of being
suppose that there is a Gaussian dip, a distribution of the form
2 six paces from the origin (i.e. at x = 6d) after (a) four, (b) six, (c) twelve steps.
c( x ) = c0 (1− e −α x ). Repeat the calculation in Problem P16C.2 and describe
Hint: Recall that n = NR − NL and N = NR + NL.
its consequences. Comment on the behaviour at x = 0.
P16C.9 Use mathematical software to calculate P(nd) in a one-dimensional
P16C.4 A lump of sucrose of mass 10.0 g is suspended in the middle of a
random walk, and evaluate the probability of being at x = 6d for N = 6, 10,
spherical flask of water of radius 10 cm at 25 °C. What is the concentration
14, …, 60. Compare the numerical value with the analytical value in the limit
of sucrose at the wall of the flask after (a) 1.0 h, (b) 1.0 week. Take D = 5.22 ×
of a large number of steps. At what value of N is the discrepancy no more than
10−10 m2 s−1.
0.1 per cent? Hint: Recall that n = NR − NL and N = NR + NL.
P16C.5 Confirm that eqn 16C.10 is a solution of the diffusion equation, eqn
P16C.10 The diffusion coefficient of a particular kind of t-RNA molecule is
16C.6, and that it has the correct initial value.
D = 1.0 × 10−11 m2 s−1 in the medium of a cell interior. How long does it take
P16C.6 (a) Confirm that molecules produced in the cell nucleus to reach the walls of the cell at a
c0 2 distance 1.0 μm, corresponding to the radius of the cell?
c( x , t ) = e − ( x − x0 −vt ) /4 Dt
(4 πDt )1/2 P16C.11 Nuclear magnetic resonance can be used to determine the mobility
is a solution of the diffusion equation with convection (eqn 16C.19) with all of molecules in liquids. A set of measurements on methane in carbon
the solute concentrated at x = x0 at t = 0. (b) Using mathematical software or a tetrachloride showed that its diffusion coefficient is 2.05 × 10−9 m2 s−1 at 0 °C
spreadsheet, plot c( x , t )/c0 as a function of t and separately as a function of x and 2.89 × 10−9 m2 s−1 at 25 °C. Deduce what information you can about the
for some typical values of D and v. Recall that diffusion is a slow process so mobility of methane in carbon tetrachloride.
 Exercises and problems 719

FOCUS 16 Molecules in motion

Integrated activities
I16.1 In Topic 17D it is shown that a general expression for the activation (0 < x < ∞) and c(0,t) = cs (0 ≤ t ≤ ∞) where c0 and cs are constants, the
energy of a chemical reaction is Ea = RT 2(d ln kr/dT). Confirm that the same diffusion equation (eqn 16C.6) is solved by the following expression for c(x,t)
expression may be used to extract the activation energy from eqn 16B.2
c(x,t) = c0 + (cs − c0){1 − erf(ξ)}
for the viscosity and then apply the expression to deduce the temperature-
where ξ = x/(4Dt) . In this expression erf(ξ) is the error function and
1/2
dependence of the activation energy when the viscosity of water is given by
the empirical expression in Problem P16B.2. Plot this activation energy as a the concentration c(x,t) evolves by diffusion from the yz-plane of constant
function of temperature. Suggest an explanation of the temperature depend- concentration, such as might occur if a condensed phase is absorbing a
ence of Ea. species from a gas phase. (b) Draw graphs of concentration profiles at several
different times of your choice for the diffusion of oxygen into water at 298 K
I16.2‡ In this problem, you are invited to use mathematical software to
(when D = 2.10 × 10−9 m2 s−1) on a spatial scale comparable to passage of
examine a model for the transport of oxygen from air in the lungs to blood.
oxygen from lungs through alveoli into the blood. Use c0 = 0 and set cs equal to
(a) Show that, for the initial and boundary conditions c(x,t) = c(x,0) = c0,
the solubility of oxygen in water, 2.9 × 10−4 mol dm−3.
FOCUS 17
Chemical kinetics

This Focus introduces the principles of ‘chemical kinetics’, results of the analysis are relations, which can be explored ex-
the study of reaction rates. The rate of a chemical reaction perimentally, between the equilibrium constant of the overall
might depend on variables that can be controlled, such as the process and the rate constants of the forward and reverse reac-
pressure, the temperature, and the presence of a catalyst, and tions in the proposed mechanism.
it is possible to optimize the rate by the appropriate choice of 17C.1 First-order reactions approaching equilibrium;
conditions. 17C.2 Relaxation methods

17A The rates of chemical reactions 17D The Arrhenius equation

This Topic discusses the definition of reaction rate and out- The rate constants of most reactions increase with increasing
lines the techniques for its measurement. The results of such temperature. This Topic introduces the ‘Arrhenius equation’,
measurements show that reaction rates depend on the concen- which captures this temperature dependence by using only
tration of reactants (and sometimes products) and ‘rate con- two parameters that can be determined experimentally.
stants’ that are characteristic of the reaction. This dependence 17D.1 The temperature dependence of reaction rates;
can be expressed in terms of differential equations known as 17D.2 The interpretation of the Arrhenius parameters
‘rate laws’.
17A.1 Monitoring the progress of a reaction; 17A.2 The rates of reactions

17E Reaction mechanisms

17B Integrated rate laws The study of reaction rates also leads to an understanding of
the ‘mechanisms’ of reactions, their analysis into a sequence of
‘Integrated rate laws’ are the solutions of the differential equa- elementary steps. This Topic shows how to construct rate laws
tions that describe rate laws. They are used to predict the con- from a proposed mechanism. The elementary steps themselves
centrations of species at any time after the start of the reaction have simple rate laws which can be combined into an overall
and to provide procedures for measuring rate constants. This rate law by invoking the concept of the ‘rate-determining step’
Topic explores some simple yet useful integrated rate laws that of a reaction, by making the ‘steady-state approximation’, or
appear throughout the Focus. by supposing that a ‘pre-equilibrium’ exists.
17B.1 Zeroth-order reactions; 17B.2 First-order reactions; 17E.1 Elementary reactions; 17E.2 Consecutive elementary reactions;
17B.3 Second-order reactions 17E.3 The steady-state approximation; 17E.4 The rate-determining
step; 17E.5 Pre-equilibria; 17E.6 Kinetic and thermodynamic control
of reactions

17C Reactions approaching equilibrium

17F Examples of reaction mechanisms
In general, rate laws must take into account both the forward
and reverse reactions and describe the approach to equi- This Topic develops three examples of reaction mechanisms.
librium, when the forward and reverse rates are equal. The The first describes a special class of reactions in the gas phase