77 Standard electrode potentials 

743

(a) The electrochemical series value of K for the reaction is obtained by doing the calculations
described in Topic 76. It should always be remembered that
In Topic 76 it is shown that a cell reaction is spontaneous even for reactions that are thermodynamically favourable there
under standard conditions (and therefore in the sense K > 1, as may be kinetic factors that result in very slow rates of reaction.
<
explained in that Topic) if Ecell > 0 (corresponding to ΔrG<< 0).
We have seen that, for two redox couples Ox1/Red1 and Ox2/Red2,
the cell potential is given by eqn 77.1; for their standard values Brief illustration 77.1 The electrochemical series
Red L , Ox L || Red R , Ox R <
Ecell = ER< − EL< (77.5a) Because E < (Zn2+,Zn) = −0.76 V < E < (Cu 2+,Cu) = +0.34 V, zinc
has a thermodynamic tendency to reduce Cu 2+ ions in aque-
ous solution. More qualitatively: to determine whether zinc
It follows that the cell reaction (Ox + ν e− → Red)R − (Ox + ν e− → can displace magnesium from aqueous solutions, we note that
Red)L, which is zinc lies above magnesium in the electrochemical series, so
zinc cannot reduce magnesium ions in aqueous solution.
Red L + Ox R → Ox L + Red R (77.5b)
Self-test 77.3 Can zinc displace hydrogen from acid? Can
is spontaneous as written (in the sense K > 1) if E > E . < < copper?
R L
Answer: Yes; no
Because in the cell reaction RedL reduces OxR, we can conclude
that
Red L has a thermodynamic tendency to reduce Ox R
if EL< < ER<. (b) The determination of activity coefficients
More briefly: low reduces high.
Once the standard potential of an electrode in a cell is known,
Table 77.2 shows a part of the electrochemical series, the
perhaps by calculation from tables of standard electrode poten-
metallic elements (and hydrogen) arranged in the order of their
tials or by direct measurement, it can be used to determine
reducing power as measured by their standard potentials in
mean activity coefficients. The procedure involves measuring
aqueous solution. A metal low in the series (with a lower stand-
the cell potential for a known molality, b, of electroactive spe-
ard potential) can reduce the ions of metals with higher stand-
cies, using the Nernst equation to find the value of the activ-
ard potentials. This conclusion is qualitative. The quantitative
ity, a, that corresponds to the observed cell potential, and then
Table 77.2 The electrochemical series of the metals* determining the activity coefficient from an expression of the
form γ = a/(b/b<). For instance, if we write the Nernst equation
Least strongly reducing for the Harned cell in the form
Gold
2RT
Platinum g (b) − E < (Ag , AgCl, Cl − ) = − ln γ ± (77.6)
F
Silver
Mercury then all we need do is measure g(b) = Ecell + (2RT/F) ln b (with b
Copper interpreted as b/b<) for a given value of b.
(Hydrogen)
Lead Brief illustration 77.2 The activity coefficient
Tin
Nickel
From the data in Example 77.1 we know that E < (Ag, AgCl,
Cl −) = +0.2232 V and that when b = 9.138 mmol kg −1, E cell =
Iron
0.468 60 V, corresponding to g(b) = 0.2273 V. Therefore, with
Zinc
RT/F = 25.693 mV,
Chromium
0.2273V − 0.2232V
Aluminium ln γ ± = − = −0.0798…
2 × (25.693 ×10−3 V)
Magnesium
Sodium That is, γ ± = 0.9233.
Calcium
Self-test 77.4 Use data from Example 77.1 to determine γ ±
Potassium
when b = 25.63 mmol kg−1.
Most strongly reducing
Answer: 0.8778
* The complete series can be inferred from Table 77.1 in the Resource section.

Atkins09819.indb 743 9/11/2013 8:38:16 AM

 744 15 Chemical equilibria

The determination of
(c) The temperature coefficient of the standard cell potential,
<
equilibrium constants dEcell /dT , gives the standard entropy of the cell reaction. This con-
clusion follows from the thermodynamic relation (∂G/∂T)p = −S
The principal use for standard potentials is to calculate the (Topic 66) and eqn 76.3 of Topic 76 (Δ r G < = −␯FEcell <
), which
standard potential of a cell formed from any two electrodes. To combine to give
do so, we subtract the standard potential of the left-hand elec-
trode from the standard potential of the right-hand electrode, <
dEcell Δ S<
<
= ER< − EL< . Because ΔG < = −␯FEcell< = r Temperature coefficient of the cell potential (77.7)
Ecell , it then follows that if dT ␯F
<
the result gives Ecell > 0, then the corresponding cell reaction
<
has K > 1. To calculate the numerical value of K we use eqn 76.5 The derivative is complete because Ecell , like ΔrG<, is inde-
of Topic 76 (ln K = ␯FEcell<
/ RT ). pendent of the pressure. Hence we have an electrochemi-
cal technique for obtaining standard reaction entropies and
through them the entropies of ions in solution.
Brief illustration 77.3 Equilibrium constants Finally, we can combine the results obtained so far and
A disproportionation is a reaction in which a species is use them to obtain the standard reaction enthalpy by writing
both oxidized and reduced. To study the disproportiona- ΔG = ΔH − TΔS in the form ΔH = ΔG + TΔS:
tion 2 Cu + (aq) → Cu(s) + Cu 2+ (aq) we combine the following
electrodes:
⎛ < dE < ⎞
Δ r H < = −␯F ⎜ Ecell − T cell ⎟ Standard reaction enthalpy (77.8)
Right-hand electrode : Cu(s) | Cu + (aq ) ⎝ dT ⎠
Cu + (aq) + e − → Cu(aq) ER< = + 0.52 V
Left-hand electrode : Pt(s) | Cu 2+ (aq), Cu + (aq) This expression provides a non-calorimetric method for
Cu (aq) + e → Cu (s)
2+ − +
EL< = +0.16 V measuring ΔrH< and, through the convention ΔfH<(H+,
aq) = 0, the standard enthalpies of formation of ions in solu-
where the standard potentials are measured at 298 K. The tion (Topic 57).
standard cell potential is therefore

<
Ecell = + 0.52 V − 0.16 V = +0.36 V Example 77.3 Using the temperature coefficient
of the cell potential
We can now calculate the equilibrium constant of the cell
reaction. Because ν = 1 and RT/F = 25.693 V, The standard potential of the cell Pt(s)|H2(g)|HBr(aq)|AgBr(s)|
Ag(s) was measured over a range of temperatures, and the data
0.36V
ln K = = 14.0… were fitted to the following polynomial:
25.693 × 10−3 V
<
Ecell /V = 0.07131 − 4.99 × 10−4 (T /K − 298)
Hence, K = 1.2 × 106.
− 3.45 × 10−6 (T /K − 298)2
Self-test 77.5 Calculate the solubility constant (the equilibrium
constant for the reaction Hg 2Cl 2 (s)  Hg 22+ (aq ) + 2 Cl − (aq )) at Evaluate the standard reaction Gibbs energy, enthalpy, and
298.15. entropy at 298 K of the reaction AgBr(s) + 12 H2 (g ) → Ag(s) +
Answer: 2.6 × 10 −18 HBr(aq).
Method The standard Gibbs energy of reaction is obtained
by using ΔrG < = −νFE < after evaluating E < at 298 K, and by
using 1 V C = 1 J. The standard entropy of reaction is obtained
(d)The determination of by using eqn 77.7, which involves differentiating the polyno-
thermodynamic functions mial with respect to T and then setting T = 298 K. The reaction
enthalpy is obtained by combining the values of the standard
The standard cell potential is related to the standard reaction
Gibbs energy and entropy.
Gibbs energy through eqn 76.3 of Topic 76 (Δ r G < = −␯FEcell<
).
<
Therefore, by measuring Ecell or by constructing its value from Answer At T = 298 K, Ecell
<
= +0.07131 V, so
standard electrode potentials we can obtain this important Δ rG < = −␯F Ecell
<
= −(1) × (9.6485 ×104 C mol −1 ) × (+0.07131V)
thermodynamic quantity. Its value can then be used to calcu-
= −6.880 ×103 VC mol −1 = −6.880 kJmol −1
late the Gibbs energy of formation of ions by using the conven-
tion explained in Topic 65.

www.ebook3000.com

Atkins09819.indb 744 9/11/2013 8:38:30 AM

 77 Standard electrode potentials  745

The temperature coefficient of the standard cell potential is It then follows that

< Δ r H < = Δ rG < + T Δ r S <

dEcell
= − 4.99 ×10 − 4 V K −1 − 2(3.45 ×10 − 6 )(T /K − 298) V K −1 = −6.880 kJmol −1 + (298K) × (−0.0481kJK −1 mol −1 )
dT
= −21.2 kJmol −1
At T = 298 K this expression evaluates to
One difficulty with this procedure lies in the accurate meas-
< urement of small temperature coefficients of the cell potential.
dEcell
= − 4.99 × 10 − 4 V K −1 Nevertheless, it is another example of the striking ability of
dT
thermodynamics to relate the apparently unrelated, in this
So, from eqn 77.7, the reaction entropy is case to relate electrical measurements to thermal properties.

Δ r S < =1 × (9.6485 ×104 C mol −1 ) × (−4.99 ×10−4 V K −1 ) Self-test 77.6 Predict the standard potential of the Harned cell
at 303 K from tables of thermodynamic data.
= − 48.1JK mol
−1 −1
Answer: +0.2222 V

Checklist of concepts
☐ 1. Each electrode is considered to make a characteristic
contribution to the overall cell potential, the difference
of the two contributions being the cell potential.
☐ 2. Tables of standard electrode potentials are compiled on
the basis that the standard hydrogen electrode has zero
potential at all temperatures.
☐ 3. In the electrochemical series, the metallic elements
(and hydrogen) are arranged in the order of their reduc-
ing power as measured by their standard potentials in
aqueous solution.
☐ 4. A metal low in the electrochemical series (with a lower
standard potential) can reduce the ions of metals with
higher standard potentials.
☐ 5. The temperature coefficient of the standard cell poten-
<
tial, dEcell / dT , gives the standard entropy of the cell
reaction.
☐ 6. Standard potentials can be used to determine activity
coefficients, equilibrium constants, and thermody-
namic functions.

Checklist of equations

Property Equation Comment Equation number

Cell potential Ecell = ER − EL Definition 77.1

Standard hydrogen electrode E<(H+,H2) = 0 Convention, at all temperatures 77.2

Combinations of standard electrode potentials 77.4

νaE<(a) = νbE<(b) + νcE<(c)

Temperature coefficient of cell potential < /dT = Δ S < /␯F

dEcell 77.7
r

Standard reaction enthalpy Δ r H < = −␯F (Ecell

< − TdE < /dT )
cell 77.8

Atkins09819.indb 745 9/11/2013 8:38:36 AM

 746 15 Chemical equilibria

Focus 15 on Chemical equilibria

Topic 73 Chemical transformations

Discussion questions
73.1 Explain how the mixing of reactants and products affects the position of 73.3 Suggest how the thermodynamic equilibrium constant may respond
chemical equilibrium. differently to changes in pressure and temperature from the equilibrium
constant expressed in terms of partial pressures.
73.2 Explain how a reaction that is not spontaneous may be driven forward by
coupling to a spontaneous reaction.

Exercises
73.1(a) Write the expressions for the equilibrium constants of the following 73.4(a) The equilibrium pressure of O2 over solid silver and silver oxide, Ag2O,
reactions in terms of (i) activities and (ii) where appropriate, the ratios p/p< at 298 K is 11.85 Pa. Calculate the standard Gibbs energy of formation of
and the products γ b: Ag2O(s) at 298 K.
(a) CO(g) + Cl2(g) ⇌ COCl(g) + Cl(g) 73.4(b) The equilibrium pressure of H2 over solid uranium and uranium
(b) 2 SO2(g) + O2(g) ⇌ 2 SO3(g) hydride, UH3, at 500 K is 139 Pa. Calculate the standard Gibbs energy of
(c) Fe(s) + PbSO4(aq) ⇌ FeSO4(aq) + Pb(s) formation of UH3(s) at 500 K.
(d) Hg2Cl2(s) + H2(g) ⇌ 2 HCl(aq) + 2 Hg(l)
73.5(a) For CaF2(s) ⇌ Ca2+(aq) + 2 F−(aq), K = 3.9 × 10−11 at 25 °C and the
(e) 2 CuCl(aq) ⇌ Cu(s) + CuCl2(aq)
standard Gibbs energy of formation of CaF2(s) is −1167 kJ mol−1. Calculate the
73.1(b) Write the expressions for the equilibrium constants of the following
standard Gibbs energy of formation of CaF2(aq).
reactions in terms of (i) activities and (ii) where appropriate, the ratios p/p<
73.5(b) For PbI2(s) ⇌ Pb2+(aq) + 2 I−(aq), K = 1.4 × 10−8 at 25 °C and the
and the products γ b:
standard Gibbs energy of formation of PbI2(s) is −173.64 kJ mol−1. Calculate
(a) H2(g) + Br2(g) ⇌ 2 HBr(g)
the standard Gibbs energy of formation of PbI2(aq).
(b) 2 O3(g) ⇌ 3 O2(g)
(c) 2 H2(g) + O2(g) ⇌ 2 H2O(l) 73.6(a) In the gas-phase reaction 2 A + B ⇌ 3 C + 2 D, it was found that when
(d) H2(g) + O2(g) ⇌ H2O2(aq) 1.00 mol A, 2.00 mol B, and 1.00 mol D were mixed and allowed to come
(e) H2(g) + I2(g) ⇌ 2 HI(aq) to equilibrium at 25 °C, the resulting mixture contained 0.90 mol C at a
total pressure of 1.00 bar. Calculate (a) the mole fractions of each species at
73.2(a) Identify the stoichiometric numbers in the reaction Hg2Cl2(s) + H2(g) →
equilibrium, (b) Kx, (c) K, and (d) ΔrG<.
2 HCl(aq) + 2 Hg(l).
73.6(b) In the gas-phase reaction A + B ⇌ C + 2 D, it was found that when
73.2(b) Identify the stoichiometric numbers in the reaction CH4(g) +
2.00 mol A, 1.00 mol B, and 3.00 mol D were mixed and allowed to come
2 O2(g) → CO2(g) + 2 H2O(l).
to equilibrium at 25 °C, the resulting mixture contained 0.79 mol C at a
73.3(a) The standard reaction Gibbs energy of the isomerization of borneol total pressure of 1.00 bar. Calculate (a) the mole fractions of each species at
(C10H17OH) to isoborneol in the gas phase at 503 K is +9.4 kJ mol−1. Calculate equilibrium, (b) Kx, (c) K, and (d) ΔrG<.
the reaction Gibbs energy in a mixture consisting of 0.15 mol of borneol
73.7(a) The hydrolysis of ATP is written as ATP 4− (aq) + H2O(l) → ADP3− (aq)
molecules and 0.30 mol of isoborneol molecules when the total pressure is 600
+ HPO24− (aq) + H3O+ (aq ). For this reaction the standard reaction Gibbs
Torr. Under these conditions, is the isomerization of borneol spontaneous?
energy is +10 kJ mol−1 at 298 K. What is the biological standard state value?
73.3(b) The standard reaction Gibbs energy of the isomerization of cis-2-
73.7(b) The overall reaction for the glycolysis reaction is C 6 H12O6 (aq) +
butene to trans-2-butene in the gas phase at 298 K is −2.9 kJ mol−1. Calculate
2 NAD+ (aq) + 2 ADP3− (aq) + 2 HPO24− (aq) + 2 H2O(l) → 2 CH3COCO2− (aq)+
the reaction Gibbs energy in a mixture consisting of 0.25 mol of cis-2-butene
2 NADH(aq) + 2 ATP 4− (aq) + 2 H3O+ (aq). For this reaction, the standard
molecules and 0.95 mol of trans-2-butene molecules when the total pressure
reaction Gibbs energy is −80.6 kJ mol−1 at 298 K. What is the biological
is 600 Torr. Under these conditions, is the isomerization of cis-2-butene
standard state value?
spontaneous?

Problems
73.1 The equilibrium constant for the reaction, I2(s) + Br2(g) ⇌ 2 IBr(g) is 0.164 of iodine is negligible. (c) In fact, solid iodine has a measurable vapour
at 25 °C. (a) Calculate ΔrG< for this reaction. (b) Bromine gas is introduced pressure at 25 °C. In this case, how would the calculation have to be modified?
into a container with excess solid iodine. The pressure and temperature are
73.2 The standard Gibbs energy of formation of NH3(g) is −16.5 kJ mol−1 at
held at 0.164 atm and 25 °C. Find the partial pressure of IBr(g) at equilibrium.
298 K. What is the reaction Gibbs energy when the partial pressures of the
Assume that all the bromine is in the liquid form and that the vapour pressure

www.ebook3000.com

Atkins09819.indb 746 9/11/2013 8:38:40 AM

 Exercises and problems  747

N2, H2, and NH3 (treated as perfect gases) are 3.0 bar, 1.0 bar, and 4.0 bar, Δ␯gas
⎛ T ⎞ p<
respectively? What is the spontaneous direction of the reaction in this case? K = Kc ×⎜ < ⎟ T< =
⎝T ⎠ Rc <
73.3 The degree of dissociation, α, is defined as the fraction of reactant that
has decomposed; if the initial amount of reactant is n and the amount at where Δνgas = c + d − (a + b). Evaluate T<.
equilibrium is neq, then α = (n − neq)/n. The standard Gibbs energy of reaction
for the decomposition H2O(g) → H2 (g)+ 12 O2 (g ) is +118.08 kJ mol−1 at 73.10 The protein myoglobin (Mb) stores O2 in muscle and the protein
2300 K. What is the degree of dissociation of H2O at 2300 K and 1.00 bar? haemoglobin (Hb) transports O2 in blood; haemoglobin is composed of four
Hints: The equilibrium constant is obtained from the standard Gibbs energy of myoglobin-like molecules. Here we explore the chemical equilibria associated
reaction by using eqn 73.13, so the task is to relate the degree of dissociation, with binding of O2 in these proteins. (a) First, consider the equilibrium
α, to K and then to find its numerical value. Proceed by expressing the between Mb and O2:
equilibrium compositions in terms of α. For example, if an amount n H2O is
[MbO2 ]
present initially, then an amount αn H2O reacts to reach equilibrium and an Mb(aq) + O2 (g )  MbO2 (aq ) K=
[Mb]p
amount(1 − α)n H2O(g) is present at equilibrium. Then, solve for α in terms of
K. Because the standard Gibbs energy of reaction is large and positive, we can
where p is the numerical value of the partial pressure (in Torr) of O2 gas.
anticipate that K will be small and hence that α  1, which opens the way to
Show that the fractional saturation, s, the fraction of Mb molecules that are
making approximations to obtain its numerical value.
oxygenated, is
73.4 Calculate the equilibrium constant of the reaction CO(g) + H2(g) ⇌
H2CO(g) given that for the production of liquid formaldehyde ΔrG< = Kp
s=
+28.95 kJ mol−1 at 298 K and that the vapour pressure of formaldehyde is 1500 1+ Kp
Torr at that temperature.
and plot the dependence of s on p for K = 5 Torr. (b) Now consider the
73.5 A sealed container was filled with 0.300 mol H2(g), 0.400 mol I2(g), and equilibria between Hb and O2:
0.200 mol HI(g) at 870 K and total pressure 1.00 bar. Calculate the amounts
of the components in the mixture at equilibrium given that K = 870 for the
[HbO2 ]
reaction H2(g) + I2(g) ⇌ 2 HI(g). Hb(aq) + O2 (g )  HbO2 (aq) K1 =
[Hb] p
73.6‡ In a study of Cl2O(g) by photoelectron ionization (R.P. Thorn, et al. [Hb(O2 )2 ]
J. Phys. Chem. 100, 14178 (1996)), the authors report ΔfH<(Cl2O) = HbO2 (aq) + O2 (g )  Hb(O2 )2 (aq ) K2 =
[HbO2 ]p
+77.2 kJ mol−1. They combined this measurement with literature data on [Hb(O2 )3 ]
the reaction Cl2O(g) + H2O(g) → 2 HOCl(g), for which K = 8.2 × 10−2 and Hb(O2 )2 (aq) + O2 (g )  Hb(O2 )3 (aq ) K 3 =
[Hb(O2 )2 ]p
ΔrS< = +16.38 J K−1 mol−1, and with readily available thermodynamic data
[Hb(O2 )4 ]
on water vapour to report a value for ΔfH<(HOCl). Calculate that value. All Hb(O2 )3 (aq) + O2 (g )  Hb(O2 )4 (aq ) K 4 =
[Hb(O2 )3 ] p
quantities refer to 298 K.
73.7‡ The 1980s saw reports of ΔfH<(SiH2) ranging from 243 to 289 kJ mol−1. Show that
For example, the lower value was cited in the review article by R. Walsh (Acc.
[O2 ]bound AK1 p
Chem. Res. 14, 246 (1981)); Walsh later leant towards the upper end of the s= =
range (H.M. Frey, et al. J. Chem. Soc., Chem. Commun. 1189 (1986)). The 4[Hb]total 4B
higher value was reported by S.-K. Shin and J.L. Beauchamp (J. Phys. Chem.
90, 1507 (1986)). If the standard enthalpy of formation is uncertain by this with
amount, by what factor is the equilibrium constant for the formation of SiH2
A = 1 + 2 K 2 p + 3K 2 K 3 p 2 + 4 K 2 K 3 K 4 p 3
from its elements uncertain at (a) 298 K, (b) 700 K?
B = 1+ K1 p + K1K 2 p2 + K1K 2 K 3 p3 + K1K 2 K 3 K 4 p3
73.8 Express the equilibrium constant of a gas-phase reaction A + 3 B ⇌ 2 C in
terms of the equilibrium value of the extent of reaction, ξ, given that initially and plot the dependence of s on p (in torr) for K1 = 0.01, K2 = 0.02, K3 = 0.04,
A and B were present in stoichiometric proportions. Find an expression for and K4 = 0.08. Hints: To develop an expression for s, proceed as follows: (i)
ξ as a function of the total pressure, p, of the reaction mixture and sketch a express [Hb(O2)2] in terms of [HbO2] by using K2, then express [HbO2] in
graph of the expression obtained. terms of [Hb] by using K1, and likewise for all the other concentrations of
73.9 The equilibrium constant K calculated from thermodynamic data refers Hb(O2)3 and Hb(O2)4. (ii) Show that
to activities. For gas-phase reactions, that means partial pressures (and
explicitly, pJ/p<). However, in practical applications we might wish to discuss [O2 ]bound = [HbO2 ]+ 2[Hb(O2 )2 ]+ 3[Hb(O2 )3 ]+ 4[Hb(O2 )4 ]
gas-phase reactions in terms of molar concentrations. The equilibrium = AK1 p[Hb]
constant is then denoted Kc, and for the equilibrium a A(g) + b B(g) ⇌
[Hb]total = B[Hb]
c C(g) + d D(g), we write
(iii) Use the fact that each Hb molecule has four sites at which O2 can attach.
[C]c [D]d
Kc = (c) The binding of O2 to haemoglobin is an example of cooperative binding, in
[A]a [B]b
which the binding of a ligand (in this case O2) to a biopolymer (in this case
Hb) becomes more favourable thermodynamically (that is, the equilibrium
with, as usual, the molar concentration [J] interpreted as [J]/c< with
constant increases) as the number of bound ligands increases up to the
c< = 1 mol dm−3. Show that
maximum number of binding sites. Which features of the plot from part (b)
can be ascribed to cooperative binding of O2 to Hb?
‡ These problems were supplied by Charles Trapp and Carmen Giunta.

Atkins09819.indb 747 9/11/2013 8:38:47 AM

 748 15 Chemical equilibria

73.11 The curves you were asked to plot in Problem 73.10 may also be 73.14 Under biochemical standard conditions, aerobic respiration produces
modelled mathematically by the equation approximately 38 molecules of ATP per molecule of glucose that is completely
oxidized. (a) What is the percentage efficiency of aerobic respiration under
⎛ s ⎞ biochemical standard conditions? (b) The following conditions are more
log ⎜ = ␯ log p − ␯H log K likely to be observed in a living cell: pCO2 =5.3 × 10−2 atm, pO 2 = 0.132 atm,
⎝ 1− s ⎟⎠ H
[glucose] = 5.6 × 10−2 mol dm−3, [ATP] = [ADP] = [Pi] = 1.0 × 10−4 mol dm−3,
pH = 7.4, T = 310 K. Assuming that activities can be replaced by the numerical
where s is the saturation, p is the partial pressure of O2, K is a constant (not
values of molar concentrations, calculate the efficiency of aerobic respiration
the binding constant for one ligand), and vH is the Hill coefficient, which
under these physiological conditions. (c) A typical diesel engine operates
varies from 1, for no cooperativity, to N, for all-or-none binding of N ligands
between Tc = 873 K and Th = 1923 K with an efficiency that is approximately
(N = 4 in Hb). The Hill coefficient for myoglobin is 1, and for haemoglobin it
75 per cent of the theoretical limit of (1 − Tc/Th) (see Topic 61). Compare the
is 2.8. (a) Determine the constant K for both Mb and Hb from the graph of
efficiency of a typical diesel engine with that of aerobic respiration under
fractional saturation (at s = 0.5) and then calculate the fractional saturation
typical physiological conditions (see part b). Why is biological
of Mb and Hb for the following values of p/kPa: 1.0, 1.5, 2.5, 4.0, 8.0. (b)
energy conversion more or less efficient than energy conversion in a
Calculate the value of s at the same p values assuming vH has the theoretical
diesel engine?
maximum value of 4.
73.15 In anaerobic bacteria, the source of carbon may be a molecule other
73.12 Here we investigate the molecular basis for the observation that the
than glucose and the final electron acceptor is some molecule other than
hydrolysis of ATP is exergonic at pH = 7.0 and 310 K. (a) It is thought that
O2. Could a bacterium evolve to use the ethanol/nitrate pair instead of the
the exergonicity of ATP hydrolysis is due in part to the fact that the standard
glucose/O2 pair as a source of metabolic energy?
entropies of hydrolysis of polyphosphates are positive. Why would an
increase in entropy accompany the hydrolysis of a triphosphate group into 73.16‡ Nitric acid hydrates have received much attention as possible catalysts
a diphosphate and a phosphate group? (b) Under identical conditions, the for heterogeneous reactions which bring about the Antarctic ozone hole.
Gibbs energies of hydrolysis of H4ATP and MgATP2−, a complex between the Worsnop, et al. investigated the thermodynamic stability of these hydrates
Mg2+ ion and ATP4−, are less negative than the Gibbs energy of hydrolysis under conditions typical of the polar winter stratosphere (D.R. Worsnop, et al.
of ATP4−. This observation has been used to support the hypothesis that Science 259, 71 (1993)). Standard reaction Gibbs energies can be computed
electrostatic repulsion between adjacent phosphate groups is a factor that for the following reactions at 190 K from their data:
controls the exergonicity of ATP hydrolysis. Provide a rationale for the
(i) H2O(g) → H2O(s) ΔrG< = −23.6 kJ mol−1
hypothesis and discuss how the experimental evidence supports it. Do these
electrostatic effects contribute to the ΔrH or ΔrS terms that determine the (ii) H2O(g) + HNO3(g) → HNO3.H2O(s) ΔrG< = −57.2 kJ mol−1
exergonicity of the reaction? Hint: In the MgATP2−complex, the Mg2+ ion and
(iii) 2 H2O(g) + HNO3(g) → HNO3.2H2O(s) ΔrG< = −85.6 kJ mol−1
ATP4− anion form two bonds: one that involves a negatively charged oxygen
belonging to the terminal phosphate group of ATP4− and another that involves (iv) 3 H2O(g) + HNO3(g) → HNO3.3H2O(s) ΔrG< = −112.8 kJ mol−1
a negatively charged oxygen belonging to the phosphate group adjacent to the
terminal phosphate group of ATP4−. Which solid is thermodynamically most stable at 190 K if pH2 O = 1.3 ×10−7
and pHNO3 = 4.1×10−10 ? Hint: Try computing ΔrG for each reaction under
73.13 To get a sense of the effect of cellular conditions on the ability of ATP
the prevailing conditions; if more than one solid forms spontaneously,
to drive biochemical processes, compare the standard Gibbs energy of
examine ΔrG for the conversion of one solid to another.
hydrolysis of ATP to ADP with the reaction Gibbs energy in an environment
at 37 °C in which pH = 7.0 and the ATP, ADP, and Pi− concentrations are all
1.0 μmol dm−3.

Topic 74 The statistical description of equilibrium

Discussion question
74.1 Use concepts of statistical thermodynamics to describe the molecular
features that determine the magnitudes of equilibrium constants and their
variation with temperature.

Exercises
74.1(a) Calculate the value of K for the reaction I2(g) ⇌ 2 I(g) at 1000 K from
the following data for I2: ␯ = 214.36 cm −1 , B = 0.0373 cm −1 , De = 1.5422 eV. The
ground state of the I atoms is 2P3/2, implying fourfold degeneracy.
74.1(b) Calculate the value of K at 298 K for the gas-phase isotopic exchange
reaction 2 79Br81Br ⇌ 79Br79Br + 81Br81Br. The Br2 molecule has a non-degenerate
ground state, with no other electronic states nearby. Base the calculation on the
wavenumber of the vibration of 79Br81Br, which is 323.33 cm−1.

www.ebook3000.com

Atkins09819.indb 748 9/11/2013 8:38:54 AM


Problems
74.1 Calculate and plot as a function of temperature, in the
range 300 K to 1000 K, the equilibrium constant for the
reaction CD4(g) + HCl(g) ⇌ CHD3(g) + DCl(g) using the
following data (numbers in parentheses are degeneracies):
␯ (CHD3 )/cm −1 = 2993(1), 2142(1), 1003(3), 1291(2), 1036(2);
␯ (CD4 )/cm −1 = 2109(1), 1092(2), 2259(3), 996(3); ␯ (HCl)/cm −1 = 2991;
 HCl)/ cm −1 = 10.59; B(
␯ (DCl)/cm −1 = 2145; B(  DCl)/cm −1 = 5.445;
 CHD )/cm −1 = 3.28; B (CHD )/cm −1 = 2.63; B (CD )/cm −1 = 2.63.
A(
3 3 4
Topic 75 The response of equilibria to the conditions
Discussion questions
75.1 Account for Le Chatelier’s principle in terms of thermodynamic
quantities. Can you think of a reason why the principle might fail?
Exercises
75.1(a) Calculate the percentage change in Kx for the reaction
H2CO(g) ⇌ CO(g) + H2(g) when the total pressure is increased from 1.0 bar to
3.0 bar at constant temperature.
75.1(b) Calculate the percentage change in Kx for the reaction
CH3OH(g) + NOCl(g) ⇌ HCl(g) + CH3NO2(g) when the total pressure is
increased from 1.0 bar to 4.0 bar at constant temperature.
75.2(a) The standard reaction enthalpy of Zn(s) + H2O(g) → ZnO(s) + H2(g)
is approximately constant at +224 kJ mol−1 from 920 K up to 1280 K. The
standard reaction Gibbs energy is +33 kJ mol−1 at 1280 K. Estimate the
temperature at which the equilibrium constant becomes greater than 1.
75.2(b) The standard enthalpy of a certain reaction is approximately constant
at +125 kJ mol−1 from 800 K up to 1500 K. The standard reaction Gibbs energy
is +22 kJ mol−1 at 1120 K. Estimate the temperature at which the equilibrium
constant becomes greater than 1.
75.3(a) The equilibrium constant of the reaction 2 C3H6(g) ⇌ C2H4(g) +
C4H8(g) is found to fit the expression ln K = A + B/T + C/T2 between 300 K
and 600 K, with A = −1.04, B = −1088 K, and C = 1.51 × 105 K2. Calculate the
standard reaction enthalpy and standard reaction entropy at 450 K.
75.3(b) The equilibrium constant of a reaction is found to fit the expression
ln K = A + B/T + C/T3 between 400 K and 600 K with A = −2.01, B = −1170 K,
Problems
75.1 The dissociation vapour pressure of NH4Cl at 427 °C is 608 kPa but at 459 °C
it has risen to 1115 kPa. Calculate (a) the equilibrium constant, (b) the standard
reaction Gibbs energy, (c) the standard enthalpy, (d) the standard entropy of
dissociation, all at 427 °C. Assume that the vapour behaves as a perfect gas and
that ΔH< and ΔS< are independent of temperature in the range given.
75.2 Consider the dissociation of methane, CH4(g), into the elements H2(g)
and C(s, graphite). (a) Given that ΔfH<(CH4, g) = −74.85 kJ mol−1 and
Atkins09819.indb 749
Exercises and problems  749
74.2 The exchange of deuterium between acid and water is an important
type of equilibrium, and we can examine it using spectroscopic data on
the molecules. Calculate the equilibrium constant at (a) 298 K and (b)
800 K for the gas-phase exchange reaction H2O + DCl ⇌ HDO + HCl
from the following data: ␯ (H2O)/cm −1 = 3656.7, 1594.8, 3755.8;
␯ (HDO)/ cm −1 = 2726.7, 1402.2, 3707.5;
A (H2O)/cm −1 = 27.88; B (H2O)/ cm −1 = 14.51; C (H2O)/cm −1 = 9.29;
A (HDO)/cm −1 = 23.38; B (HDO)/cm −1 = 9.102; C (HDO)/cm −1 = 6.417;
B( HCl)/cm −1 = 10.59; B(
 DCl)/cm −1 = 5.449; ␯ (HCl)/cm −1 = 2991;
␯ (DCl)/cm −1 = 2145.
75.2 State the limits to the generality of the van ’t Hoff equation, written as in
eqn 75.5.
and C = 2.2 × 107 K3. Calculate the standard reaction enthalpy and standard
reaction entropy at 500 K.
75.4(a) What is the standard enthalpy of a reaction for which the equilibrium
constant is (a) doubled, (b) halved when the temperature is increased by 10 K
at 298 K?
75.4(b) What is the standard enthalpy of a reaction for which the equilibrium
constant is (a) doubled, (b) halved when the temperature is increased by 15 K
at 310 K?
75.5(a) Estimate the temperature at which CaCO3(calcite) decomposes.
75.5(b) Estimate the temperature at which CuSO4⋅5H2O undergoes
dehydration.
75.6(a) From information in the Resource section, calculate the standard
Gibbs energy and the equilibrium constant at (a) 298 K and (b) 400 K for the
reaction PbO(s) + CO(g) ⇌ Pb(s) + CO2(g). Assume that the reaction enthalpy
is independent of temperature.
75.6(b) From information in the Resource section, calculate the standard Gibbs
energy and the equilibrium constant at (a) 25 °C and (b) 50 °C for the reaction
CH4(g) + 3 Cl2(g) ⇌ CHCl3(l) + 3 HCl(g). Assume that the reaction enthalpy is
independent of temperature.
that ΔfS<(CH4, g) = −80.67 J K−1 mol−1 at 298 K, calculate the value of the
equilibrium constant at 298 K. (b) Assuming that ΔfH< is independent of
temperature, calculate K at 50 °C.
75.3 The equilibrium pressure of H2 over U(s) and UH3(s) between 450 K
and 715 K fits the expression ln(p/Pa) = A + B/T + C ln(T/K), with A = 69.32,
B = −1.464 × 104 K, and C = −5.65. Find an expression for the standard enthalpy
of formation of UH3(s) and from it calculate ΔrCp<.
9/11/2013 8:39:04 AM
 750 15 Chemical equilibria

75.4 The degree of dissociation, αe = (ninitial − nequilibrium)/ninitial, of CO2(g) into 75.8‡ Suppose that an iron catalyst at a particular manufacturing plant
CO(g) and O2(g) at high temperatures was found to vary with temperature as produces ammonia in the most cost-effective manner at 450 °C when the
follows: pressure is such that ΔrG for the reaction 12 N2 (g) + 23 H2 (g) → NH3 (g ) is
equal to −500 J mol−1. (a) What pressure is needed? (b) Now suppose that
T/K 1395 1443 1498 a new catalyst is developed that is most cost-effective at 400 °C when the
αe/10−4 1.44 2.50 4.71 pressure gives the same value of ΔrG. What pressure is needed when the new
catalyst is used? What are the advantages of the new catalyst? Assume that
(i) all gases are perfect gases or that (ii) all gases are van der Waals gases.
Assuming ΔrH< to be constant over this temperature range, calculate ΔrH<
Isotherms of ΔrG(T, p) in the pressure range 100 atm ≤ p ≤ 400 atm are needed
and, at each of the above three temperatures, K, ΔrG<, and ΔrS<. Make any
to derive the answer. (c) Do the isotherms you plotted confirm Le Chatelier’s
justifiable approximations.
principle concerning the response of equilibrium to changes in temperature
75.5 The standard reaction enthalpy for the decomposition of CaCl2⋅NH3(s) and pressure?
into CaCl2(s) and NH3(g) is nearly constant at +78 kJ mol−1 between 350 K
75.9 Find an expression for the standard reaction Gibbs energy at a
and 470 K. The equilibrium pressure of NH3 in the presence of CaCl2⋅NH3 is
temperature T ′ in terms of its value at another temperature T and the
1.71 kPa at 400 K. Find an expression for the temperature dependence of ΔrG<
coefficients a, b, and c in the expression for the molar heat capacity listed
in the same range.
in Table 56.1. Evaluate the standard Gibbs energy of formation of H2O(l) at
75.6 Acetic acid was evaporated in a container of volume 21.45 cm3 at 437 K 372 K from its value at 298 K.
and at an external pressure of 200 kPa. The container was then sealed. The
75.10‡ The dimerization of ClO in the Antarctic winter stratosphere is believed
mass of acid present in the sealed container was 0.0519 g. The experiment
to play an important part in that region’s severe seasonal depletion of ozone.
was repeated with the same container but at 471 K, and it was found that
The following equilibrium constants are based on measurements by Cox
0.0380 g of acetic acid was present. Calculate the equilibrium constant for the
and Hayman (R. A. Cox and G. D. Hayman, Nature 332, 796 (1988)) on the
dimerization of the acid in the vapour and the enthalpy of vaporization.
reaction 2 ClO(g) → (ClO)2(g).
75.7 The dissociation of I2 can be monitored by measuring the total pressure,
and three sets of results are as follows: T/K 233 248 258 268 273
K 4.13 × 108 5.00 × 107 1.45 × 107 5.37 × 106 3.20 × 106
T/K 973 1073 1173
100p/atm 6.244 7.500 9.181 T/K 280 288 295 303
104 nI 2.4709 2.4555 2.4366 K 9.62 × 105 4.28 × 105 1.67 × 105 7.02 × 104

where nI is the amount of I atoms per mole of I2 molecules in the mixture,
which occupied 342.68 cm3. Calculate the equilibrium constants of
the dissociation and the standard enthalpy of dissociation at the mean
temperature.
(a) Derive the values of ΔrH< and ΔrS< for this reaction. (b) Compute the
standard enthalpy of formation and the standard molar entropy of (ClO)2
given ΔfH<(ClO) = +101.8 kJ mol−1 and Sm < (ClO) = 226.6 J K −1 mol −1 .

Topic 76 Electrochemical cells

Discussion questions
76.1 Describe the various types of galvanic and electrolytic cells. 76.2 Explain why salt bridges are used in electrochemical cell measurements.

Exercise
76.1(a) Write the cell reaction and electrode half-reactions of each of the 76.1(b) Write the cell reaction and electrode half-reactions of each of the
following cells: following cells:
(a) Zn(s)|ZnSO4(aq)‖AgNO3(aq)|Ag(s) (a) Pt(s)|K3[Fe(CN)6](aq),K4[Fe(CN)6](aq)‖Mn2+(aq),H+(aq)|MnO2(s)|Pt(s)
(b) Cd(s)|CdCl2(aq)‖HNO3(aq)|H2(g)|Pt(s) (b) Cu(s)|Cu2+(aq)‖Mn2+(aq),H+(aq)|MnO2(s)|Pt(s)
(c) Pt(s)|K3[Fe(CN)6](aq),K4[Fe(CN)6](aq)‖CrCl3(aq)|Cr(s) (c) Pt(s)|Cl2(g)|HCl(aq)‖HBr(aq)|Br2(l)|Pt(s)
(d) Pt(s)|Fe3+(aq),Fe2+(aq)‖Sn4+(aq),Sn2+(aq)|Pt(s) (d) Fe(s)|Fe2+(aq)‖Mn2+(aq),H+(aq)|MnO2(s)|Pt(s)

Problems
76.1 A fuel cell develops an electric potential from the chemical reaction of a cell fuelled by (a) hydrogen and oxygen, (b) the complete oxidation of
between reagents supplied from an outside source. What is the potential benzene at 1.0 bar and 298 K?

www.ebook3000.com

Atkins09819.indb 750 9/11/2013 8:39:07 AM

 Exercises and problems  751

76.2 A fuel cell is constructed in which both electrodes make use of the makes use of the partial oxidation of methane to carbon monoxide and water.
oxidation of methane. The left-hand electrode makes use of the complete (a) Which electrode is the cathode? (b) What is the cell potential at 25 °C
oxidation of methane to carbon dioxide and water; the right-hand electrode when all gases are at 1 bar?

Topic 77 Standard electrode potentials

Discussion questions
77.1 Discuss how the electrochemical series can be used to determine if a 77.3 Describe at least one non-calorimetric experimental method for
redox reaction is spontaneous under standard conditions. determining a standard reaction enthalpy.
77.2 Describe a method for the determination of the standard potential of a
redox couple.

Exercises
77.1(a) Calculate the standard potential of the Ce4+/Ce couple from the values (c) Pt(s)|Cl2(g)|HCl(aq)‖HBr(aq)|Br2(l)|Pt(s)
for the Ce3+/Ce and Ce4+/Ce3+ couples.
(d) Fe(s)|Fe2+(aq)‖Mn2+(aq),H+(aq)|MnO2(s)|Pt(s)
77.1(b) Calculate the standard potential of the Au3+/Au+ couple from the
values for the Au3+/Au and Au+/Au couples. 77.6(a) Devise cells in which the following are the reactions and calculate the
standard cell potential in each case:
77.2(a) Can mercury produce zinc metal from aqueous zinc sulfate under
standard conditions? (a) Fe(s) + PbSO4(aq) → FeSO4(aq) + Pb(s)
77.2(b) Can chlorine gas oxidize water to oxygen gas under standard
(b) Hg2Cl2(s) + H2(g) → 2 HCl(aq) + 2 Hg(l)
conditions in basic solution?
(c) 2 H2(g) + O2(g) → 2 H2O(l)
77.3(a) The potential of the cell Ag(s)|AgI(s)|AgI(aq)|Ag(s) is +0.9509 V at
25 °C. Calculate (a) the solubility of AgI and (b) the equilibrium constant for 77.6(b) Devise cells in which the following are the reactions and calculate the
the dissolution of AgI(s). standard cell potential in each case:
77.3(b) The potential of the cell Bi(s)|Bi2S3(s)|Bi2S3(aq)|Bi(s) is −0.96 V at
(a) H2(g) + O2(g) → H2O2(aq)
25 °C. Calculate (a) the solubility of Bi2S3 and (b) the equilibrium constant for
the dissolution of Bi2S3(s). (b) H2(g) + I2(g) → 2 HI(aq)
77.4(a) The standard potential of the cell Pt(s)|H2(g)|HCl(aq)|Hg2Cl2(s)|Hg(l) (c) 2 CuCl(aq) → Cu(s) + CuCl2(aq)
was found to be +0.2699 V at 293 K and +0.2669 V at 303 K. Evaluate the
77.7(a) Consider the cell Ag|AgBr(s)|KBr(aq, 0.050 mol kg−1)‖Cd(NO3)2(aq,
standard reaction Gibbs energy, enthalpy, and entropy at 298 K of the reaction
0.010 mol kg−1)|Cd. (a) Write the cell reaction. (b) Write the Nernst equation
Hg2Cl2(s) + H2(g) → 2 Hg(l) + 2 HCl(aq).
for the cell. (c) Use the Debye–Hückel limiting law and the Nernst equation to
77.4(b) The standard potential of the cell Pt(s)|H2(g)|HBr(aq)|AgBr(s)|Ag(s)
estimate the cell potential at 25 °C.
was found to be +0.07372 V at 293 K and +0.06873 V at 303 K. Evaluate the
77.7(b) Consider the cell Pt|H2(g,p<)|HCl(aq, 0.010 mol kg−1)|AgCl(s)|Ag.
standard reaction Gibbs energy, enthalpy, and entropy at 298 K of the reaction
(a) Write the cell reaction. (b) Write the Nernst equation for the cell. (c) Use
AgBr(s) + ½ H2(g) → Ag(s) + HBr(aq).
the Debye–Hückel limiting law and the Nernst equation to estimate the cell
77.5(a) Calculate the standard cell potential and the value of ΔrG< at 25 °C for potential at 25 °C.
the reactions in the following cells:
77.8(a) Calculate the equilibrium constants of the following reactions at 25 °C
(a) Zn(s)|ZnSO4(aq)||AgNO3(aq)|Ag(s) from standard potential data:
(b) Cd(s)|CdCl2(aq)||HNO3(aq)|H2(g)|Pt(s) (a) Sn(s) + Sn4+(aq) ⇌ 2 Sn2+(aq)
(c) Pt(s)|K3[Fe(CN)6](aq),K4[Fe(CN)6](aq)||CrCl3(aq)|Cr(s) (b) Fe(s) + Hg(NO3)2(aq) ⇌ Hg(l) + Fe(NO3)2(aq)
(d) Pt(s)|Fe3+(aq),Fe2+(aq)||Sn4+(aq),Sn2+(aq)|Pt(s) 77.8(b) Calculate the equilibrium constants of the following reactions at 25 °C
from standard potential data:
77.5(b) Calculate the standard cell potential and the value of ΔrG< at 25 °C for
the reactions in the following cells: (a) Cd(s) + CuSO4(aq) ⇌ Cu(s) + CdSO4(aq)
(a) Pt(s)|K3[Fe(CN)6](aq),K4[Fe(CN)6](aq)||Mn2+(aq),H+(aq)|MnO2(s)|Pt(s) (b) 3 Au2+(aq) ⇌ Au(s) + 2 Au3+(aq)
(b) Cu(s)|Cu2+(aq)‖Mn2+(aq),H+(aq)|MnO 2(s)|Pt(s)

Problems
77.1 Given that ΔrG< = −212.7 kJ mol−1 for the reaction in the Daniell cell at coefficients in the compartments, (c) the reaction quotient, (d) the standard
25 °C, and b(CuSO4) = 1.0 × 10−3 mol kg−1 and b(ZnSO4) = 3.0 × 10−3 mol kg−1, cell potential, and (e) the cell potential. (Take γ+ = γ− = γ± in the respective
calculate (a) the ionic strengths of the solutions, (b) the mean ionic activity compartments.)

Atkins09819.indb 751 9/11/2013 8:39:08 AM

 752 15 Chemical equilibria

77.2 Although the hydrogen electrode may be simple conceptually, it is 77.8‡ The table below summarizes the potential observed for the cell
cumbersome to use and several substitutes have been devised. One of these Pd(s)|H2(g,1 bar)|BH(aq,b), B(aq,b)|AgCl(s)|Ag(s). Each measurement
alternatives is the quinhydrone electrode (quinhydrone, Q⋅QH2, is a complex is made at equimolar concentrations of 2-aminopyridinium chloride
of quinone, C6H4O2 = Q, and hydroquinone, C6H4O2H2 = QH2). The electrode (BH) and 2-aminopyridine (B). The data are for 25 °C and it is found that
half-reaction is Q(aq) + 2 H+(aq) + e− → QH2(aq), E< = +0.6994 V. If the cell E< = 0.22251 V. Use the data to determine pKa for the acid at 25 °C and the
Hg|Hg2Cl2(s)|HCl(aq)|Q⋅QH2|Au is prepared, and the measured cell potential mean activity coefficient (γ±) of BH as a function of molality (b) and ionic
is +0.190 V, what is the pH of the HCl solution? Assume that the Debye– strength (I). Use the Davies equation for the mean activity coefficient (see
Hückel limiting law is applicable. Problem 77.6) with A = 0.5091. Draw a graph of the mean activity coefficient
with b = 0.04 mol kg−1 and 0 ≤ I ≤ 0.1.
77.3 Consider the cell, Zn(s)|ZnCl2(0.0050 mol kg−1)|Hg2Cl2(s)|Hg(l), for
which the cell reaction is Hg2Cl2(s) + Zn(s) → 2 Hg(l) + 2 Cl−(aq) + Zn2+(aq).
Given that E<(Zn2+,Zn) = −0.7628 V, E<(Hg2Cl2,Hg) = +0.2676 V, and that b/(mol kg−1) 0.01 0.02 0.03 0.04 0.05
the cell potential is +1.2272 V, (a) write the Nernst equation for the cell. E<(25 °C)/V 0.74452 0.72853 0.71928 0.71314 0.70809
Determine (b) the standard cell potential; (c) ΔrG, ΔrG<, and K for the cell
reaction; (d) the mean activity and activity coefficient of ZnCl2 from the b/(mol kg−1) 0.06 0.07 0.08 0.09 0.10
measured cell potential; and (e) the mean activity coefficient of ZnCl2 from
the Debye–Hückel limiting law. (f) Given that (∂E/∂T)p = −4.52 × 10−4 V K−1, E<(25 °C)/V 0.70380 0.70059 0.69790 0.69571 0.69338
calculate ΔrS and ΔrH.
Hint: Use mathematical software or a spreadsheet.
77.4 The potential of the cell Pt(s)|H2(g, p<)|HCl(aq,b)|Hg2Cl2(s)|Hg(l) has
been measured with high precision (G. J. Hills and D. J. G. Ives, J. Chem. Soc., 77.9 Show that if the ionic strength of a solution of the sparingly soluble salt
311 (1951)), with the following results at 25 °C: MX and the freely soluble salt NX is dominated by the concentration C of the
latter, and if it is valid to use the Debye–Hückel limiting law, the solubility S′
b/(mmol kg−1) 1.6077 3.0769 5.0403 7.6938 10.9474 in the mixed solution is given by
E/V 0.60080 0.56825 0.54366 0.52267 0.50532
1/2
K s e 4.606 AC
S′ =
Determine the standard potential of the cell and the mean activity coefficient C
of HCl at these molalities. (Make a least-squares fit of the data to the best
straight line.) when Ks is small (in a sense to be specified).

77.5 Careful measurements of the potential of the cell Pt(s)|H2(g, p<)|NaOH 77.10 If the mitochondrial electric potential between matrix and the
(aq, 0.0100 mol kg−1), NaCl(aq, 0.01125 mol kg−1)|AgCl(s)|Ag(s) have been intermembrane space were 70 mV, as is common for other membranes, how
reported (C.P. Bezboruah, et al. J. Chem. Soc. Faraday Trans. I 69, 949 (1973)). much ATP could be synthesized from the transport of 4 mol H+, assuming the
Among the data is the following information: pH difference remains the same?
77.11 The standard potentials of proteins are not commonly measured
θ/°C 20.0 25.0 30.0
by the methods described in this Topic because proteins often lose their
E/V 1.04774 1.04864 1.04942 native structure and function when they react on the surfaces of electrodes.
In an alternative method, the oxidized protein is allowed to react with an
Calculate pKw at these temperatures and the standard enthalpy and entropy of appropriate electron donor in solution. The standard potential of the protein
the autoprotolysis of water at 25.0 °C. is then determined from the Nernst equation, the equilibrium concentrations
of all species in solution, and the known standard potential of the electron
77.6 Measurements of the potentials of cells of the type Ag(s)|AgX(s)MX(b1)| donor. We shall illustrate this method with the protein cytochrome c. The one-
MxHg|MX(b2)|AgX(s)|Ag(s), where MxHg denotes an amalgam and the electron reaction between cytochrome c, cyt, and 2,6-dichloroindophenol,
electrolyte is an alkali metal halide dissolved in ethylene glycol, have been D, can be followed spectrophotometrically because each of the four species
reported (U. Sen, J. Chem. Soc. Faraday Trans. I 69, 2006 (1973)), and in solution has a distinct colour, or absorption spectrum. We write the
some values for LiCl are given below. Estimate the activity coefficient at reaction as cytox + Dred ⇌ cytred + Dox , where the subscripts ‘ox’ and ‘red’ refer
the concentration marked * and then use this value to calculate activity to oxidized and reduced states, respectively. (a) Consider Ecyt < and E < to
D
coefficients from the measured cell potential at the other concentrations. Base be the standard potentials of cytochrome c and D, respectively. Show that, at
your answer on the Davies equation (eqn 72.29, log γ± = −AI1/2/(1 + BI1/2) + CI) equilibrium (‘eq’), a plot of ln([Dox]eq/[Dred]eq) versus ln([cytox]eq/[cytred]eq) is
for the mean ionic activity, with A = 1.461, B = 1.70, k = 0.20, and I = b/b<. For linear with slope of 1 and y-intercept F (Ecyt< − E < )/RT , where equilibrium
D
b2 = 0.09141 mol kg−1: activities are replaced by the numerical values of equilibrium molar
concentrations. (b) The following data were obtained for the reaction between
b1/(mol kg−1) 0.0555 0.09141* 0.1652 0.2171 1.040 1.350 oxidized cytochrome c and reduced D in a pH = 6.5 buffer at 298 K. The ratios
E/V −0.0220 0.0000 0.0263 0.0379 0.1156 0.1336 [Dox]eq/[Dred]eq and [cytox]eq/[cytred]eq were adjusted by titrating a solution
containing oxidized cytochrome c and reduced D with a solution of sodium
77.7 The standard potential of the AgCl/Ag,Cl− couple has been measured ascorbate, which is a strong reductant. From the data and the standard
very carefully over a range of temperature (R.G. Bates and V.E. Bowers, potential of D of 0.237 V, determine the standard potential of cytochrome c at
J. Res. Nat. Bur. Stand. 53, 283 (1954)) and the results were found to fit the pH = 6.5 and 298 K.
expression

[Dox]eq/[Dred]eq 0.00279 0.00843 0.0257 0.0497

E < /V = 0.23659 − 4.8564 ×10−4 (θ /°C) − 3.4205 ×10−6 (θ /°C)2
[cytox]eq/[cytred]eq 0.0106 0.0230 0.0894 0.197
+ 5.869 ×10−9 (θ /°C)3
[Dox]eq/[Dred]eq 0.0748 0.238 0.534
Calculate the standard Gibbs energy and enthalpy of formation of Cl−(aq) and
its entropy at 298 K. [cytox]eq/[cytred]eq 0.335 0.809 1.39

www.ebook3000.com

Atkins09819.indb 752 9/11/2013 8:39:13 AM


Integrated activities
F15.1 Nitrogen dioxide, a paramagnetic compound, is in equilibrium with its
dimer, dinitrogen tetroxide, a diamagnetic compound. Derive an expression
in terms of the equilibrium constant, K, for the dimerization to show how
the molar magnetic susceptibility varies with the pressure of the sample
(Topic 39). Suggest how the susceptibility might be expected to vary as the
temperature is changed at constant pressure.
F15.2 Use mathematical software, a spreadsheet, or the Living graphs (labelled
LG) on the website of this book for the following exercises:
(a) Plot xA and xB against the pressure p for several values of the equilibrium
constant K for a reaction of the form A(g) ⇌ 2 B(g).
Atkins09819.indb 753
Exercises and problems  753
(b) The equilibrium constant of a reaction is found to fit the expression
ln K = a + b/(T/K) + c/(T/K)3 over a range of temperatures. (a) Write
expressions for ΔrH< and ΔrS<. (b) Plot ln K against T between 400 K and
600 K for a = −2.0, b = −1.0 × 103, and c = 2.0 × 107.
(c) Plot the variation of cell potential with the value of the reaction quotient
for the cell reaction for different values of the temperature. Does the cell
potential become more or less sensitive to composition as the temperature
increases?
(d) For the cell described in Example 77.1, plot a family of curves of E against
b/b< for several values of the temperature T.
9/11/2013 8:39:13 AM
this page left intentionally blank

www.ebook3000.com
 FOCUS 16 ON Molecular motion

Topic 78 Topic 79 Topic 80 Topic 81

The kinetic Transport Motion Diffusion

theory properties in
of gases of gases liquids

Focus 14 Focus 18
Physical Reaction
equilibria dynamics

That so many chemical reactions take place in gas and liquid phases is due in part to the mobility of
molecules and ions in fluid media. To understand the rates of these reactions, which are treated as
an aspect of Reaction dynamics, we need to know how fast molecules migrate and encounter one
another.
A perfect gas is in principle very simple, and the model of a gas in which the molecules are in
ceaseless chaotic motion provides a way to predict many of its properties (Topic 78). In fact, the
kinetic theory of gases described in this Topic is a remarkable example of model-building in science,
for from the flimsiest of assumptions precise predictions can be made about the pressure and the
speeds of the molecules. Moreover, when molecules fly through space, they carry properties with
them. By noting how far and how fast molecules move, we can build quantitative expressions for a
number of these ‘transport properties’, including the thermal conductivity and viscosity of a perfect
gas (Topic 79).
The motion of ions and molecules in liquids is quite different: instead of flying freely through
space, a molecule or ion must migrate through a dense but mobile medium. One way to study this
motion is to apply an electric field and to monitor the rate at which ions migrate under its influ-
ence (Topic 80). All solute particles, however, migrate by jostling their way past the surrounding sol-
vent molecules, and their motion can be understood in terms of a ‘thermodynamic force’ that relies
on the concept of the chemical potential introduced in Physical equilibria; furthermore we see that
their average motion can be expressed in terms of two very powerful and general laws of diffusion
(Topic 81). In fact, by treating the motion as a random walk, we see that there are analogies between
the migration of molecules in liquids and in gases.

What is the impact of this material?

The migration of molecules and ions through matter is crucial to the occurrence of chemical reac-
tions, but it also plays a central role in biology. In Impact 16.1 we examine how the transport proper-
ties of molecules are used in genomics and proteomics to establish the composition of nucleic acids
and proteins. In Impact 16.2 we see how the experimental detection of the migration of ions across
biological membranes leads to insights about the function of biological cells.

Atkins09819.indb 755 9/11/2013 8:39:16 AM

 To read more about the impact of this material, scan the QR code or go to http://
bcs.whfreeman.com/webpub/chemistry/qmc2e/impact/qchem_impact16.
html.

www.ebook3000.com

Atkins09819.indb 756 9/11/2013 8:39:16 AM

 TOPIC 78

The kinetic theory of gases

Contents ➤ What do you need to know already?

78.1 The kinetic model 757 You need to be aware of Newton’s second law of motion,
(a) Pressure and molecular speeds 758 that the acceleration of a body is proportional to the force
Brief illustration 78.1:The number of molecules acting on it (Foundations, Topic 2).
with a range of speeds 760
(b) Mean values 761
Example 78.1: Calculating the mean speed
of molecules in a gas 762 In the kinetic theory of gases (which is sometimes called the
(c) The collision frequency 763 kinetic-molecular theory, KMT) it is assumed that the only
Brief illustration 78.2: Molecular collisions 763 contribution to the energy of the gas is from the kinetic ener-
(d) The mean free path 764 gies of the molecules. The kinetic model is one of the most
Brief illustration 78.3: The mean free path 764 remarkable—and arguably most beautiful—models in physical
78.2 Collisions with walls and surfaces 764 chemistry, for from a set of very slender assumptions, powerful
(a) The collision flux 764 quantitative conclusions can be reached.
Brief illustration 78.4: The collision flux 764
(b) Effusion 765
Example 78.2: Calculating the vapour pressure
from a mass loss 765 78.1 The kinetic model
Checklist of concepts 766
Checklist of equations 766 The kinetic model of a pure gas (of identical molecules) is based
on four assumptions:
1. The gas consists of molecules of mass m in ceaseless
random motion obeying the laws of classical
mechanics.
➤ Why do you need to know this material? 2. The size of the molecules is negligible, in the sense that
There are ‘scientific’ and practical reasons for the their diameters are much smaller than the average
importance of this material. The ‘scientific’ reason is that the distance travelled between collisions.
discussion shows how to extract quantitative information 3. Because the forces between molecules are assumed to
from a qualitative model. The practical importance of the have a short range, the molecules interact only through
material is that it is used in the discussion of the properties brief collisions.
of gases (Topic 36), in the discussion of catalysis (Topic 97), 4. The collisions are elastic.
and in the discussion of reaction rates in gases (Topic 87).
An elastic collision is a collision in which the total trans-
➤ What is the key idea? lational kinetic energy of the molecules is conserved. The
According to the kinetic model, a gas consists of molecules theory is applicable to a sample consisting of a large num-
of negligible size in ceaseless random motion and obeying ber of molecules in which it is safe to consider the average
the laws of classical mechanics in their collisions. behaviour of the molecules and to ignore fluctuations from
the mean.

Atkins09819.indb 757 9/11/2013 8:39:17 AM

 758 16 Molecular motion

|vxΔt|
(a) Pressure and molecular speeds
From the very economical assumptions of the kinetic model, Area, A
we show in the following Justification that the pressure, p, and
Won’t
volume, V, of the gas are related by
Will
pV = 13 nM vrms
2
Perfect gas Pressure (78.1)

where M = mNA, the molar mass of the molecules of mass m, Volume = |vxΔt|A
n is the amount of molecules in the sample, and vrms is the x
square root of the mean of the squares of the speeds, v, of the
molecules: Figure 78.2 A molecule will reach the wall on the right
within an interval Δt if it is within a distance vx Δt of the wall
vrms = 〈v 2 〉1/2 Definition Root mean square speed (78.2) and travelling to the right.

1
nN A Avx Δt / V .
The total momentum change in that interval
Justification 78.1The pressure of a gas according
2
is the product of this number and the change 2mvx:
to the kinetic model
nN A Avx Δt
Consider the arrangement in Fig. 78.1. When a particle of mass Momentum change = × 2mv x
2V
m that is travelling with a component of velocity vx parallel to
M
the x-axis collides with the wall on the right and is reflected, nmN A Avx2 Δt nMAvx2 Δt
= =
its linear momentum changes from mvx before the collision V V
to −mvx after the collision (when it is travelling in the oppo- Next, to find the force, we calculate the rate of change of
site direction). The x-component of momentum therefore momentum, which is this change of momentum divided by
changes by 2mvx on each collision (the y- and z-components the interval Δt during which it occurs:
are unchanged). Many molecules collide with the wall in an
interval Δt, and the total change of momentum is the product nMAvx2
Rate of change of momentum =
of the change in momentum of each molecule multiplied by the V
number of molecules that reach the wall during the interval. This rate of change of momentum is equal to the force (by
Because a molecule with velocity component vx can travel Newton’s second law of motion). It follows that the pressure,
a distance vx Δt along the x-axis in an interval Δt, all the mol- the force divided by the area, is
ecules within a distance vx Δt of the wall will strike it if they are
travelling towards it (Fig. 78.2). It follows that if the wall has nM vx2
Pressure =
area A, then all the particles in a volume A × vx Δt will reach the V
wall (if they are travelling towards it). The number density of
Not all the molecules travel with the same velocity, so the
particles is nNA/V, where n is the total amount of molecules in
detected pressure, p, is the average (denoted 〈…〉) of the quan-
the container of volume V and NA is Avogadro’s constant, so the
tity just calculated:
number of molecules in the volume Avx Δt is (nNA/V) × Avx Δt.
At any instant, half the particles are moving to the right nM 〈 vx2 〉
and half are moving to the left. Therefore, the average num- p=
V
ber of collisions with the wall during the interval Δt is
This expression already resembles the perfect gas equation of
state.
Before mvx
collision
To write an expression for the pressure in terms of the root
mean square speed, we begin by writing the speed of a single
molecule, v, as v 2 = vx2 + v 2y + vz2 . Because the root mean square
–mvx speed, v rms, is defined as v rms = 〈v2〉1/2 (as in eqn 78.2), it follows
After that
collision
x vrms
2
= 〈 v 2 〉 = 〈 vx2 〉 + 〈 v 2y 〉 + 〈 vz2 〉

Figure 78.1 The pressure of a gas arises from the impact However, because the molecules are moving randomly, all
of its molecules on the walls. In an elastic collision of a three averages are the same. It follows that vrms 2
= 3〈 v 2 〉.
molecule with a wall perpendicular to the x-axis, the Equation 78.1 follows immediately by substituting 〈 vx 〉 = 13 vrms
2 2
x-component of velocity is reversed but the y- and into p = nM 〈 vx 〉 / V .
2
z-components are unchanged.

www.ebook3000.com

Atkins09819.indb 758 9/11/2013 8:39:29 AM

 78 The kinetic theory of gases 759

Equation 78.1 is one of the key results of the kinetic model. ∞

⎛ 2πkT ⎞
1/2

1= Kx ∫ e − mvx /2 kT dvx = K x ⎜
2

⎝ m ⎟⎠
We see that, if the root mean square speed of the molecules
−∞
depends only on the temperature, then at constant temperature
Therefore, K x = (m/2πkT)1/2 and at this stage we can write
pV = constant
1/2
⎛ m ⎞
f (vx ) = ⎜ e − mvx /2 kT
2

which is the content of Boyle’s law. Moreover, for eqn 78.1 to be ⎝ 2πkT ⎟⎠
the equation of state of a perfect gas, its right-hand side must be
equal to nRT. It follows that the root mean square speed of the The probability that a molecule has a velocity in the range vx to
vx + dvx, vy to vy + dvy, vz to vz + dvz is therefore
molecules in a gas at a temperature T must be
f (vx ) f (v y ) f (vz )dv x dv y dvz
1/2
⎛ 3RT ⎞ 3/2
vrms = ⎜ Perfect gas Root mean square speed (78.3) ⎛ m ⎞
⎝ M ⎟⎠ e − mvx /2 kT e − mv y /2 kT e − mvz /2 kT dvx dv y dvz
2
=⎜
2 2

⎝ 2πkT ⎟⎠
3/2
It is also possible to confirm that the right-hand side of eqn ⎛ m ⎞
=⎜ e − mv dv x dv y dv z
2
/2 kT
78.1 is equal to nRT by appealing to the Boltzmann distribu- ⎝ 2πkT ⎟⎠
tion. To do so, we proceed in two steps. First, in the following
Justification we show that the fraction of molecules that have a where v 2 = vx2 + v 2y + vz2 .
speed in the range v to v + dv is f(v)dv, where To evaluate the probability that the molecules have a speed
in the range v to v + dv regardless of direction we think of the
three velocity components as defining three coordinates in
3/2
⎛ M ⎞ Perfect Maxwell– ‘velocity space’, with the same properties as ordinary space
f (v) = 4 π ⎜ v 2 e − M v /2 RT
2
Boltzmann (78.4)
⎝ 2πRT ⎟⎠
gas except that the coordinates are labelled (vx,v y,v z) instead of
distribution
(x,y,z). Just as the volume element in ordinary space is dxdydz,
so the volume element in velocity space is dvxdvydvz. The sum
Justification 78.2 of all the volume elements in ordinary space that lie at a dis-
The Maxwell–Boltzmann distribution
tance r from the centre is the volume of a spherical shell of
of speeds radius r and thickness dr. That volume is the product of its
The Boltzmann distribution implies that the fraction of surface area, 4πr2, and its thickness, dr, and is therefore 4πr2dr.
molecules with velocity components vx, vy, and vz is propor- Similarly, the analogous volume in velocity space is the volume
tional to an exponential function of their kinetic energy: of a shell of radius v and thickness dv, namely 4πv 2dv (Fig.
f(v) = Ke −ε/kT, where K is a constant of proportionality. The 78.3). Now, because f(vx)f(vy)f(vz), the term in blue in the last
kinetic energy is equation, depends only on v2 , and has the same value every-
where in a shell of radius v, the total probability of the mol-
ε = 12 mvx2 + 12 mv 2y + 12 mvz2 ecules possessing a speed in the range v to v + dv is the product

Therefore, we can use the relation a x+y+z = a xayaz to write

vz
Surface Thickness, dv
− (mv 2x +mv 2y +mvz2 )/2 kT − mv 2y /2 kT
f (v) = Ke = Ke − mv 2x /2 kT
e e − mvz2 /2kkT
area, 4πv2

The distribution factorizes into three terms, and we can write v

f(v) = f(vx) f(vy) f(vz) and K = K x KyK z, with

f (vx ) = K x e − mvx /2 kT
2

and likewise for the other two axes. vx

vy
To determine the constant K x , we note that a molecule
must have a velocity component somewhere in the range
−∞ < vx < ∞, so
Figure 78.3 To evaluate the probability that a molecule
∞
has a speed in the range v to v + dv, we evaluate the
∫−∞
f (v x )dvx = 1
total probability that the molecule will have a speed
that is anywhere on the surface of a sphere of radius
Substitution of the expression for f(vx) and use of Integral G.1 v = (vx2 + v 2y + vz2 )1/2 by summing the probabilities that it is in
in the Resource section then gives a volume element dvxdvydvz at a distance v from the origin.

Atkins09819.indb 759 9/11/2013 8:39:40 AM

 760 16 Molecular motion

of the term in blue and the volume of the shell of radius v and
thickness dv. If this probability is written f(v)dv, it follows that

Distribution function, f (v)

3/2
⎛ m ⎞
f (v)dv = 4 πv 2dv ⎜ e − mv
2
/2 kT
⎝ 2πkT ⎟⎠

and f(v) itself, after minor rearrangement, is

3/2
⎛ m ⎞
f (v) = 4 π ⎜ v 2e − mv
2
/2 kT
⎝ 2πkT ⎟⎠ v1 v2

Because m/k = M/R, this expression is eqn 78.4. Speed, v

Figure 78.5 To calculate the probability that a molecule will

have a speed in the range v1 to v2, we integrate the distribution
The function f(v) is called the Maxwell–Boltzmann distri- between those two limits; the integral is equal to the area of
bution of speeds. Then we use this distribution to calculate the the curve between the limits, as shown shaded here.
average value of v2.
The important features of the Maxwell–Boltzmann distribu- molecules can be expected to have high speeds at high
tion are as follows (and are shown pictorially in Fig. 78.4): temperatures than at low temperatures.
r
A factor v2 (the term before the e) multiplies the
r
Equation 78.4 includes a decaying exponential function
exponential. This factor goes to zero as v goes to zero, so
(more specifically, a Gaussian function). Its presence
the fraction of molecules with very low speeds will also
implies that the fraction of molecules with very high
be very small whatever their mass.
speeds will be very small because e − x becomes very
2

r
The remaining factors (the term in parentheses in eqn

Physical interpretation

small when x is large.

78.4 and the 4π) simply ensure that, when we sum the
r
The factor M/2RT multiplying v2 in the exponent is
fractions over the entire range of speeds from zero to
large when the molar mass, M, is large, so the
infinity, we get 1.
exponential factor goes most rapidly towards zero
when M is large. That is, heavy molecules are The Maxwell–Boltzmann distribution can be used to evalu-
unlikely to be found with very high speeds. ate the fraction of molecules in the range from v1 to v2 by evalu-
r
The opposite is true when the temperature, T, is high: ating the integral
then the factor M/2RT in the exponent is small, so the
v2

∫
exponential factor falls towards zero relatively slowly as
F (v1 , v2 ) = f (v)dv (78.5)
v increases. In other words, a greater fraction of the v1

This integral is the area under the graph of f as a function of v

Low temperature
or and in general has to be evaluated numerically by using math-
Distribution function, f(v)

high molecular mass ematical software (Fig. 78.5).

Intermediate temperature
or
molecular mass Brief illustration 78.1 The number of molecules with
High temperature
or a range of speeds
low molecular mass
To evaluate the number of N2 molecules that have speeds in
the range 200 to 220 m s −1 at 400 K we need to evaluate the
integral in eqn 78.5 with M = 28.02 g mol−1. First, we write
Speed, v
M 2.802 ×10−2 kg mol −1
Figure 78.4 The distribution of molecular speeds with = = 4.21…×10−6 m −2 s2
2RT 2 × (8.3145 JK −1 mol −1 ) × (400 K )
temperature and molar mass. Note that the most probable
speed (corresponding to the peak of the distribution) increases and denote it a 2 . Then (by using mathematical software to
with temperature and with decreasing molar mass, and evaluate the integral),
simultaneously the distribution becomes broader.

www.ebook3000.com

Atkins09819.indb 760 9/11/2013 8:39:45 AM

 78 The kinetic theory of gases 761

T. We can conclude that the root mean square speed of the

1.95 × 10 −8 m−3 s3
  7.32 × 105 m3 s−3
  molecules of a gas is proportional to the square root of the
3/2
⎛ a2 ⎞ 220 m s −1 temperature and inversely proportional to the square root
F (200 ms , 220 ms ) = 4 π ⎜ ⎟
−1 −1
⎝ π⎠ 200 m s −1 ∫
ve
2 − a2 v2
dv
of the molar mass. That is, the higher the temperature, the
= 0.0143 higher the root mean square speed of the molecules, and,
at a given temperature, heavy molecules travel more slowly
That is, 1.43 per cent of the molecules lie in the specified range. than light molecules. Sound waves are pressure waves, and
for them to propagate the molecules of the gas must move to
Self-test 78.1 Calculate the fraction of CO2 molecules that
form regions of high and low pressure. Therefore, we should
have speeds in the range 400 to 500 m s−1 at 298 K.
expect the root mean square speeds of molecules to be com-
Answer: 0.199
parable to the speed of sound in air (340 m s−1). The root mean
square speed of N2 molecules, for instance, is found from eqn
78.3 to be 515 m s−1 at 298 K.

Detector

The chemist’s toolkit 78.1 Mean values

Source
Selector The mean value (also called the expectation value) 〈x〉 of a
property (such as the speed v) is calculated by first multiplying
each discrete value x i that the property can have (such as v1,
v2 , …, vN) by the probability pi that x i occurs and then sum-
Figure 78.6 A velocity selector. The molecules are produced
ming these products over all possible N values:
in the source (which may be an oven with a small hole in one
N
wall), and travel in a beam towards the rotating discs. Only
if the speed of a molecule is such as to carry it through the 〈x〉 = ∑p x
i =1
i i

sequence of slits will it reach the detector. Thus, the number

of slow molecules can be counted by rotating the discs slowly, When N is very large and the xi values are so closely spaced that
and the number of fast molecules counted by rotating the discs the property can be regarded as varying continuously (as is the
rapidly. case with molecular speeds), it is useful to express the probabil-
ity that it can have a value between x and x + dx as

The Maxwell–Boltzmann distribution has been verified
experimentally. For example, molecular speeds can be meas-
ured directly with a velocity selector (Fig. 78.6). The spinning
cylinder has channels that permit the passage of only those
molecules moving through them at the appropriate speed, and
the number of molecules can be determined by collecting them
Probability of finding a value between x and x + dx = f (x )dx
where the function f(x) is the probability density. It follows that
the probability that the property has a value between x = a and
x = b is
b

at a detector. Probability of finding a value between a and b = ∫a

f (x )dx

The integral is equivalent to the area beneath the graph of the

(b) Mean values function f(x) between the limits a and b. If g(x) is a function of
the property (for instance, a power of x, such as x 2; in the case
Once we have the Maxwell–Boltzmann distribution, we can
of speeds, v2), then its mean value in the range is
calculate the mean value of any power of the speed by evalu-
b
ating the appropriate integral (see The chemist’s toolkit 78.1).
Thus, to evaluate the average value of vn we calculate
〈 g (x )〉 = ∫
a
g (x ) f (x )dx

∞
The mean value if x spans an infinite but positive range (like
∫ (78.6)
〈vn 〉 = vn f (v)dv molecular speeds) is
0
∞

In particular, integration with n = 2 results in eqn 78.3 for the

〈 g ( x )〉 = ∫
0
g (x ) f (x )dx

mean square speed (vrms 2

) of the molecules at a temperature

Atkins09819.indb 761 9/11/2013 8:39:53 AM

 762 16 Molecular motion

Example 78.1 Calculating the mean speed of molecules

vmp = (2RT/M)1/2
in a gas
vmean = (8RT/πM)1/2

f(v)/4π(M/2πRT)3/2
Calculate the mean speed, vmean, of N2 molecules in air at 25 °C. vrms = (3RT/M)1/2
Method The mean speed, v mean, is obtained by evaluating the
integral

∞
vmean = ∫
0
v f (v)dv

with f(v) given in eqn 78.4. Either use mathematical software v/(2RT/M)1/2
1 (4/π)1/2 (3/2)1/2
or use Integral G.4 given in the Resource section.
Answer The integral required is Figuer 78.7 A summary of the conclusions that can be
deduced from the Maxwell–Boltzmann distribution for
⎛ M ⎞
3/2 ∞ molecules of molar mass M at a temperature T: vmp is the most
vmean = 4 π ⎜ ∫ v 3e − mv dv
2
/2 kT
⎝ 2πRT ⎟⎠ 0
probable speed, vmean is the mean speed, and vrms is the root
3/2 1/2 1/2 mean square speed.
⎛ M ⎞ 1 ⎛ 2RT ⎞ ⎛ 8RT ⎞
= 4π ⎜ × ⎜ =⎜
⎝ 2πRT ⎟⎠ 2 ⎝ M ⎟⎠ ⎝ πM ⎟⎠ The mean relative speed, vrel, the mean speed with which one
molecule approaches another of the same kind, can also be cal-
Substitution of the data then gives culated from the distribution:
1/2 Perfect gas, Mean relative
⎛ 8 × (8.3145 JK −1 mol −1 ) × (298 K ) ⎞ vrel = 21/2 vmean identical molecules speed (78.9a)
vmean = ⎜ = 475 ms −1
⎝ π × (28.02 ×10 kg mol ) ⎟⎠
−3 −1

This result is much harder to derive, but the diagram in Fig.

We have used 1 J = 1 kg m2 s −2. 78.8 should help to show that it is plausible. For the relative
Self-test 78.2 Evaluate the root mean square speed of N2 mol-
mean speed of two dissimilar molecules of masses mA and mB,
ecules at 25 °C by integration. Use mathematical software or 1/2 Mean
Integral G.5 of the Resource Section. ⎛ 8RT ⎞ MA MB Perfect
vrel = ⎜ μ= relative
⎝ πμ ⎟⎠
(78.9b)
Answer: v rms = (3RT/M)1/2 = 515 m s −1 MA + MB gas
speed

v 21/2v v

As shown in Example 78.1, we can use the Maxwell–

Boltzmann distribution to evaluate the mean speed, vmean, of v
the molecules in a gas: v

1/2 v v v v
⎛ 8RT ⎞
vmean = ⎜ (78.7)
⎝ πM ⎟⎠
Perfect gas Mean speed

We can identify the most probable speed, vmp, from the loca-
tion of the peak of the distribution:
1/2
⎛ 2RT ⎞
vmp = ⎜
⎝ M ⎟⎠
Perfect gas
The location of the peak of the distribution is found by differ-
entiating f(v) with respect to v and looking for the value of v at
which the derivative is zero (other than at v = 0 and v = ∞; see
Problem 78. 5). Figure 78.7 summarizes these results.
Most probable speed
0 21/2v 2v
Figure 78.8 A simplified version of the argument to show
that the mean relative speed of molecules in a gas is related to
the mean speed. When the molecules are moving in the same
(78.8)
direction, the relative speed is zero; it is 2v when the molecules
are approaching each other. A typical mean direction of
approach is from the side, and the speed of approach is then
21/2v. The last direction of approach is the most characteristic,
so the speed of approach can be expected to be about 21/2v.
This value is confirmed by more detailed calculation.

www.ebook3000.com

Atkins09819.indb 762 9/11/2013 8:40:00 AM

 78 The kinetic theory of gases 763

Equation 78.9b turns into eqn 78.9a when the molecules are of the pressure of the gas is obtained by using the perfect gas
identical (that is when MA = MB = M so μ = M/2). equation and R = kNA to write
N nN A nN A p
N =  = = =
(c) The collision frequency V V nRT / p kT

Although the kinetic-molecular theory assumes that the mol-

ecules are point-like, we can count a ‘hit’ whenever the cen-
tres of two molecules come within a distance d of each other,
where d, the collision diameter, is of the order of the actual
diameters of the molecules (for impenetrable hard spheres d is
the diameter). As we show in the following Justification, we can
use the kinetic model to deduce that the collision frequency,
z, the number of collisions made by one molecule divided by
the time interval during which the collisions are counted, when
there are N molecules in a volume V, is
z =σvrel N Perfect gas Collision frequency
with N = N/V, the number density, and vrel given by eqn
78.9.The area σ = πd2 is called the collision cross-section of the
molecules. Some typical collision cross-sections are given in
Table 78.1. In terms of the pressure (as is also shown in the fol-
lowing Justification),
σvrel p
z= Perfect gas Collision frequency
kT
Table 78.1* Collision cross-sections
σ/nm2
Miss vrelΔt
d
d Hit
Area, σ
Figure 78.9 In an interval Δt, a molecule of diameter d
sweeps out a tube of radius d and length vrelΔt. As it does so
it encounters other molecules with centres that lie within
the tube, and each such encounter counts as one collision.
(78.10a) In reality, the tube is not straight, but changes direction at
each collision. Nevertheless, the volume swept out is the
same, and this straightened version of the tube can be used
as a basis of the calculation.
Equation 78.10a shows that, at constant volume, the collision
frequency increases with increasing temperature. Equation
78.10b shows that, at constant temperature, the collision fre-
(78.10b) quency is proportional to the pressure. Such a proportion-
ality is plausible because the greater the pressure, the greater
the number density of molecules in the sample, and the rate at
which they encounter one another is greater even though their
average speed remains the same.

C6H6 0.88
CO2 0.52
He 0.21
N2 0.43
* More values are given in the Resource section.
Justification 78.3 The collision frequency
Consider the positions of all the molecules except one to be
frozen. Then note what happens as one mobile molecule
travels through the gas with a mean relative speed v rel for a
time Δt. In doing so it sweeps out a ‘collision tube’ of cross-
sectional area σ = πd 2 and length v rel Δt and therefore of vol-
ume σv rel Δt (Fig. 78.9). The number of stationary molecules
with centres inside the collision tube is given by the volume
of the tube multiplied by the number density N = N/V, and is
Nσv rel Δt. The number of hits scored in the interval Δt is equal
to this number, so the number of collisions divided by the time
interval is Nσv rel, which is eqn 78.10a. The expression in terms
Brief illustration 78.2 Molecular collisions
For an N2 molecule in a sample at 1.00 atm (101 kPa) and 25 °C,
from Example 78.1 we know that v mean = 475 m s−1, so from eqn
78.9a, v rel = 21/2 × 475 m s −1. Therefore, from eqn 78.10b, and
taking σ = 0.43 nm 2 (corresponding to 0.43 × 10 −18 m 2) from
Table 78.1,
(0.43 × 10−18 m2 ) × (21/2 × 475 ms −1 ) × (1.01 × 105 Pa )
z=
(1.381 × 10−23 JK −1 ) × (298 K )
= 7.1 ×109 s −1
so a given molecule collides about 7 × 109 times each second.
We are beginning to appreciate the timescale of events in
gases.
Self-test 78.3 Evaluate the collision frequency between H 2
molecules in a gas under the same conditions.
Answer: 1.7 × 1010 s −1

Atkins09819.indb 763 9/11/2013 8:40:05 AM

 764 16 Molecular motion

(d) The mean free path 78.2Collisions with walls and

Once we have the collision frequency, we can calculate the surfaces
mean free path, λ (lambda), the average distance a molecule
travels between collisions. If a molecule collides with a fre- The key result for accounting for transport in the gas phase is
quency z, it spends a time 1/z in free flight between collisions, the rate at which molecules strike an area (which may be an
and therefore travels a distance (1/z)vrel. It follows that the imaginary area embedded in the gas, or part of a real wall). The
mean free path is collision flux, ZW is the number of collisions with the area in a
given time interval divided by the area and the duration of the
vrel interval.
λ= Perfect gas Mean free path (78.11)
z

Substitution of the expression for z in eqn 78.10b gives

(a) The collision flux
The collision frequency, the number of hits per second, is
kT obtained by multiplication of the collision flux by the area of
λ= Perfect gas Mean free path (78.12)
σp interest. We show in the following Justification that the collision
flux is
Doubling the pressure reduces the mean free path by half.
p
ZW = Perfect gas Collision flux (78.13)
(2πmkT )1/2
Brief illustration 78.3 The mean free path
In Brief illustration 78.2 we noted that v rel = 21/2 × 475 m s −1 for
N2 molecules at 25 °C, and z = 7.1 × 109 s −1 when the pressure is
Brief illustration 78.4 The collision flux
1.00 atm. Under these circumstances, the mean free path of N2
molecules is The collision f lux of O 2 molecules, with m = M/N A and
M = 32.00 g mol−1, at 25 °C and at 1.00 bar is
2 × 475 ms
1/2 −1
λ= = 9.5 × 10−8 m
7.1 × 109 s −1
kgm−1 s−2

or 95 nm, about 103 molecular diameters. 1.00 × 10 Pa 5

ZW =
{2π × (32.00 × 10−3 kg mol −1 / 6.022 × 1023 mol −1 )
Self-test 78.4 Evaluate the mean free path of benzene mol-
×(1.381 × 10−23 JK −1 ) × (298 K )}1/2
ecules at 25 °C in a sample where the pressure is 0.10 atm.
Answer: 460 nm = 2.70 × 1027 m −2 s −1

This flux corresponds to 2.70 × 1023 cm−2 s −1.

Although the temperature appears in eqn 78.12, in a sam-
ple of constant volume, the pressure is proportional to T, so
T/p remains constant when the temperature is increased.
Therefore, the mean free path is independent of the temper-
ature in a sample of gas in a container of fixed volume: the
distance between collisions is determined by the number of
molecules present in the given volume, not by the speed at
which they travel.
In summary, a typical gas (N2 or O2) at 1 atm and 25 °C
can be thought of as a collection of molecules travelling with
a mean speed of about 500 m s−1. Each molecule makes a colli-
sion within about 1 ns, and between collisions it travels about
103 molecular diameters. The kinetic model of gases is valid
and the gas behaves nearly perfectly if the diameter of the
Self-test 78.5 Evaluate the collision flux of H2 molecules under
the same conditions.
Answer: 1.07 × 1028 m−2 s −1
Justification 78.4 The collision flux
Consider a wall of area A perpendicular to the x-axis (as in
Fig. 78.2). If a molecule has v x > 0 (that is, it is travelling in
the direction of positive x), then it will strike the wall within
an interval Δt if it lies within a distance v x Δt of the wall.
Therefore, all molecules in the volume Avx Δt, and with posi-
tive x-component of velocities, will strike the wall in the
interval Δt. The total number of collisions in this interval is
molecules is much smaller than the mean free path (d  λ), therefore the volume Avx Δt multiplied by the number density,
for then the molecules spend most of their time far from one N, of molecules. However, to take account of the presence of a
another. range of velocities in the sample, we must sum the result over

www.ebook3000.com

Atkins09819.indb 764 9/11/2013 8:40:09 AM

 78 The kinetic theory of gases 765

all the positive values of vx weighted by the probability distri- particularly of substances with very low vapour pressures. In
bution of velocities given in Justification 78.2: this technique, a sample of the substance is enclosed in a cav-
∞
ity with a small hole and its mass is monitored as a function of
Number of collisions = NAΔt ∫ 0
vx f (vx )dvx time. The value of the vapour pressure, p, is then obtained by
applying eqn 78.14.
The collision flux is the number of collisions divided by A and
Δt, so
∞
ZW = N ∫0
vx f (vx )dvx Example 78.2 Calculating the vapour pressure from
a mass loss
Then, using the velocity distribution f(vx) found in Justification
78.2 and Integral G.2 from the Resource section, we find Caesium (m.p. 29 °C, b.p. 686 °C) was introduced into a con-
tainer and heated to 500 K. When a hole of diameter 0.50 mm
∞ 1/2 ∞ 1/2 was opened in the container for 100 s, a mass loss of 385 mg
⎛ m ⎞ ⎛ kT ⎞
∫ vx f (vx )dvx = ⎜ ∫ vx e − mvx /2 kT dvx = ⎜
2
was measured. Calculate the vapour pressure of liquid cae-
0 ⎝ 2πkT ⎟⎠ 0 ⎝ 2πm ⎟⎠
sium at 500 K.
Therefore, Method The pressure of vapour is constant inside the con-
tainer despite the effusion of atoms because the hot liquid
1/2
⎛ kT ⎞ metal replenishes the vapour. The rate of effusion is therefore
ZW = N ⎜
⎝ 2πm ⎟⎠ constant, and given by eqn 78.14. To express the rate in terms
of mass, multiply the number of atoms that escape by the mass
Substitution of N = p/kT then gives eqn 78.13. of each atom.
Answer The mass loss Δm in an interval Δt is related to the
collision flux by Δm = Z WA0mΔt, where A0 is the area of the
(b) Effusion hole and m is the mass of one atom. It follows from eqn 78.14
that
The essential empirical observations on effusion are summa-
pA0mΔt pA m1/2 Δt
rized by Graham’s law of effusion, which states that the rate Δm = = 0 1/2
(2πmkT )1/2
(2πkT )
of effusion is inversely proportional to the square root of the
molar mass. The basis of this result is that, as remarked above, and therefore that
the mean speed of molecules is inversely proportional to M1/2,
1/2 1/2
so the rate at which they strike the area of the hole is also ⎛ 2πkT ⎞ Δm ⎛ 2πRT ⎞ Δm
p=⎜ =
inversely proportional to M1/2. However, by using the expres- ⎝ m ⎟⎠ A0 Δt ⎜⎝ M ⎟⎠ A0 Δt
sion for the rate of collisions, we can obtain a more detailed
expression for the rate of effusion and hence use effusion data Substitution of the data and M = 132.9 g mol−1 gives
more effectively.
1/2
When a gas at a pressure p and temperature T is separated ⎛ 2π × (8.3145 JK −1 ) × (500 K ) ⎞
from a vacuum by a small hole, the rate of escape of its mole- p=⎜ ⎟⎠
⎝ 132.9 × 10 kg mol
−3 −1
cules is equal to the rate at which they strike the area of the hole,
385 ×10−6 kg
which is the product of the area and collision flux. Therefore, ×
for a hole of area A0 π × (0.25 × 10−3 m)2 × (100 s)
Pa 
= 8.7 × 10 kg m −1s −2
3
pA0 Perfect Effu-
Rate of effusion = Z W A0 = (78.14)
(2πmkT )1/2 gas sion
or 8.7 kPa.
Because M = mNA, this rate is inversely proportional to M1/2, in Self-test 78.6 How long would it take 1.0 g of Cs atoms to
accord with Graham’s law. effuse out of the oven under the same conditions?
Equation 78.14 is the basis of the Knudsen method for the Answer: 260 s
determination of the vapour pressures of liquids and solids,

Atkins09819.indb 765 9/11/2013 8:40:16 AM

 766 16 Molecular motion

Checklist of concepts
☐ 1. In the kinetic theory (or kinetic-molecular theory,
KMT) of gases it is assumed that the only contribution
to the energy of the gas is from the kinetic energies of
the molecules.
☐ 2. An elastic collision is a collision in which the total
translational kinetic energy of the molecules is
conserved.
☐ 3. The Maxwell–Boltzmann distribution of speeds gives
the probability that a molecule will have a given speed
(see below).
☐ 4. The root mean square speed of the molecules of a gas is
proportional to the square root of the temperature and
inversely proportional to the square root of the molar
mass.
☐ 5. The collision frequency is the number of collisions
made by one molecule divided by the time interval dur-
ing which the collisions are counted.
☐ 6. The mean free path is the average distance a molecule
travels between collisions.
☐ 7. The collision flux is the number of collisions with the
area in a given time interval divided by the area and the
duration of the interval.
☐ 8. Graham’s law of effusion states that the rate of effusion
is inversely proportional to the square root of the molar
mass.
☐ 9. The Knudsen method is a technique for the determi-
nation of the vapour pressures of liquids and solids by
monitoring mass loss by effusion.

Checklist of equations
Property Equation Comment Equation number

Pressure pV = 13 nM vrms
2 KMT* 78.1

Root mean square speed vrms = 〈v2〉1/2 Definition 78.2

vrms = (3RT/M)1/2 KMT 78.3

f (v) = 4 π ( M / 2πRT )
3/2 2 − M v 2 /2 RT
Maxwell–Boltzmann distribution v e KMT 78.4

Mean speed vmean = (8RT/πM)1/2 KMT 78.7

Most probable speed vmp = (2RT/M)1/2 KMT 78.8

Mean relative speed vrel = (8RT/πμ)1/2 μ = MAMB/(MA + MB), KMT 78.9b

Collision frequency z = σvrelN KMT 78.10a

Mean free path λ = vrel/z KMT 78.11

Collision flux ZW = p/(2πmkT)1/2 KMT 78.13

Rate of effusion Rate of effusion = ZWA0 KMT; rate in terms of number, not amount 78.14

* KMT denotes a result derived on the basis of kinetic-molecular theory of gases.

www.ebook3000.com

Atkins09819.indb 766 9/11/2013 8:40:18 AM

 TOPIC 79
Transport properties of gases
Contents
79.1 The phenomenological equations
Brief illustration 79.1: Energy flux
79.2 The transport parameters
The diffusion coefficient
(a)
Brief illustration 79.2: The diffusion coefficient
(b) Thermal conductivity
Brief illustration 79.3: The thermal conductivity
(c) Viscosity
Brief illustration 79.4: The viscosity
Checklist of concepts
Checklist of equations
➤ Why do you need to know this material?
The transport of properties by gas molecules plays an
important role in the atmosphere. It also extends the
approach of kinetic theory, showing how to extract
quantitative expressions from simple models.
➤ What is the key idea?
A molecule carries properties through space for about the
distance of its mean free path.
➤ What do you need to know already?
This Topic builds on and extends the kinetic theory of gases
(Topic 78) and you need to be familiar with the expressions
from that Topic for the mean speed of molecules and with
the significance of the mean free path and its pressure
dependence. It also uses the expression for the collision
flux.
Transport properties are commonly expressed in terms of a
number of ‘phenomenological’ equations, equations that are
empirical summaries of experimental observations. These
Atkins09819.indb 767
phenomenological equations apply to all kinds of properties
and media. In the following sections, we introduce the equa-
tions for the general case and then show how to calculate the
767
parameters that appear in them.
769
769
769
The phenomenological
770 79.1
770 equations
771
771 By a ‘phenomenological equation’, a term encountered com-
772 monly in the study of fluids, we mean an equation that summa-
772 rizes empirical observations on phenomena without, initially at
773
least, being based on an understanding of the molecular pro-
cesses responsible for the property.
The net rate of transport 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
the interval. If matter is flowing (as in diffusion), we speak of a
matter flux of so many molecules per square metre per second;
if the property is the energy of thermal motion (as in thermal
conduction), then we speak of the energy flux and express it in
joules per square metre per second, and so on.
Experimental observations on transport properties show
that the flux of a property is usually proportional to the first
derivative of some other related property. For example, the flux
of matter diffusing parallel to the z-axis of a container is found
to be proportional to the first derivative of the concentration:
dN
J (matter ) ∝ (79.1)
dz
where N is the number density of particles with units num-
ber per metre cubed (m−3). The SI units of J are number per
square metre per second (m−2 s−1). The proportionality of the
flux of matter to the concentration gradient is sometimes called
Fick’s first law of diffusion: the law implies that if the concen-
tration varies steeply with position, then diffusion will be fast.1
1 Fick’s ‘second law’ is treated in Topic 81.
9/11/2013 8:40:19 AM
 768 16 Molecular motion

Table 79.1* Transport properties of gases at 1 atm

κ/(mW K−1 m−1) η/μP†

N 273 K 273 K 293 K
dN z
<0 Ar 16.3 210 223
dz
CO2 14.5 136 147
He 144.2 187 196
N2 24.0 166 176
* More values are given in the Resource section.
J>0 † 1 μP = 10 −7 kg m−1 s −1.

Figure 79.1 The flux of particles down a concentration

gradient. Fick’s first law states that the flux of matter (the
number of particles passing through an imaginary window in a
given time interval divided by the area of the window and the
length of the interval) is proportional to the density gradient at
that point.
There is no net flux if the concentration is uniform (dN/dz = 0).
Similarly, the rate of thermal conduction (the flux of the energy
associated with thermal motion) is found to be proportional to
the temperature gradient:
dT
J (energy of thermal motion) ∝
dz
The SI units of this flux are joules per square metre per second
(J m−2 s−1).
A positive value of J signifies a flux towards positive z; a
negative value of J signifies a flux towards negative z. Because
matter flows down a concentration gradient, from high con-
centration to low concentration, J is positive if dN/dz is nega-
tive (Fig. 79.1). Therefore, the coefficient of proportionality in
eqn 79.1 must be negative, and we write it −D, with D a positive
constant:
To see the connection between the flux of momentum and
the viscosity, consider a fluid in a state of Newtonian flow,
which can be imagined as occurring by a series of layers mov-
ing past one another (Fig. 79.2). The layer next to the wall of
the vessel is stationary, and the velocity of successive layers
varies linearly with distance, z, from the wall. Molecules cease-
lessly move between the layers and bring with them the x-com-
ponent of linear momentum they possessed in their original
layer. A layer is retarded by molecules arriving from a more
slowly moving layer because they have a low momentum in the
x-direction. A layer is accelerated by molecules arriving from a
more rapidly moving layer. We interpret the net retarding effect
as the fluid's viscosity.
(79.2)
Because the retarding effect depends on the transfer of the
x-component of linear momentum into the layer of interest,
the viscosity depends on the flux of this x-component in the
z-direction. The flux of the x-component of momentum is pro-
portional to dvx/dz because there is no net flux when all the lay-
ers move at the same velocity. We can therefore write
J (x -component
dv (79.5)
of momentum) = −η x Definition Viscosity
dz

dN
J (matter ) = − D Definition Diffusion coefficient (79.3)
dz Brings Brings
Wall

x low high
The constant D is called the diffusion coefficient; its SI units are momentum momentum
metre squared per second (m2 s−1). Energy of thermal motion
(‘heat’) migrates down a temperature gradient, and the same
reasoning leads to

J (energy of thermal Thermal z

dT Definition (79.4)
motion) = −κ conductivity
dz Figure 79.2 The viscosity of a fluid arises from the transport
of linear momentum. In this illustration the fluid is undergoing
where κ (kappa) is the coefficient of thermal conductivity. The laminar (Newtonian) flow, and particles bring their initial
SI units of κ are joules per kelvin per metre per second, or watts momentum when they enter a new layer. If they arrive with
per kelvin per metre (J K−1 m−1 s−1, or W K−1 m−1). Some experi- high x-component of momentum they accelerate the layer; if
mental values are given in Table 79.1. with low x-component of momentum they retard the layer.

www.ebook3000.com

Atkins09819.indb 768 9/11/2013 8:40:25 AM

 79 Transport properties of gases  769

The constant of proportionality, η (eta), is the coefficient of
viscosity (or simply ‘the viscosity’). Its SI units are kilogram
per metre per second (kg m−1 s−1, which is equivalent to Pa s).
Viscosities are often reported in the non-SI unit poise (P),
where 1 P = 10−1 kg m−1 s−1. There is a variety of methods of
determining viscosity, including monitoring the rate of flow of
a fluid through a narrow tube. Some experimental values are
given in Table 79.1.
Justification 79.1 The diffusion coefficient
Consider the arrangement depicted in Fig. 79.3. On average,
the molecules passing through the area A at z = 0 have trav-
elled about one mean free path λ since their last collision.
Therefore, the number density where they originated is N(z)
evaluated at z = −λ. This number density is approximately
⎛ dN ⎞
N ( − λ ) = N (0) − λ ⎜
⎝ dz ⎟⎠ 0

Brief illustration 79.1 Energy flux where we have used a Taylor expansion of the form
f(x) = f(0) + (df/dx)0x + … truncated after the second term (see
Suppose that there is a temperature difference of 10 K between Mathematical background 1). Similarly, the number density at
two metal plates that are separated by 1.0 cm. The temperature an equal distance on the other side of the area is
gradient is
⎛ dN ⎞
dT 10 K N (λ ) = N (0) + λ ⎜
=− = −1.0 ×103 K m −1 ⎝ dz ⎟⎠ 0
dz 1.0 ×10−2 m

Therefore, the energy flux in air (Table 79.1) is The average number of impacts on the imaginary window of
area A0 during an interval Δt is Z WA0Δt, where Z W is the col-
J(energy of thermal motion) = −(0.0241JK −1 m −1 s −1 ) lision flux (from Justification 78.4 and eqn 78.7 for the mean
× (−1.0 ×103 K m −1 ) = + 24 Jm −2 s −1 speed, Z W = N (kT /2πm)1/2 = 14 Nvmean ). Therefore, the f lux
from left to right, J(L → R), arising from the supply of mole-
As a result, in 1.0 h (3600 s) the transfer of energy through an cules on the left, is
area of the opposite walls of 1.0 cm2 is
  ZW
 
Transfer = (24Jm s )× (1.0 ×10 m ) ×(3600 s) = 8.6J
−2 −1 −4 2 1
N ( − λ )v A0 Δt 1
J (L → R ) = 4 mean
= 4 N (− λ )vmean
A0 Δt
Self-test 79.1 The thermal conductivity of glass is 0.92 J K−1
m−1 s−1. What is the rate of energy transfer (in watts, 1 W = 1 J s−1) There is also a flux of molecules from right to left. On average,
through a window pane of thickness 0.50 cm and area 1.0 m2 the molecules making the journey have originated from z = +λ
when the room is at 22 °C and the exterior is at 0 °C? where the number density is N(λ). Therefore,
Answer: 4.0 kW
J(L ← R ) = 14 N (λ )vmean

The transport parameters

Number density, N

79.2 N(–λ)
N(0)
N(+λ)
Here we derive expressions for the diffusion characteristics z Area, A
(specifically, the diffusion coefficient, the thermal conduc-
tivity, and the viscosity) of a perfect gas on the basis of the
kinetic-molecular theory.

(a) The diffusion coefficient +λ

0
In the following Justification we show that the flux of molecules –λ
through a gas is in accord with Fick’s first law, and specifically
Figure 79.3 The calculation of the rate of diffusion of a gas
that
considers the net flux of molecules through a plane of area
⎛ dN ⎞ A as a result of arrivals from on average a distance λ away in
J z = − 12 vmean λ ⎜ (79.6)
⎝ dz ⎟⎠ 0 each direction, where λ is the mean free path.

Atkins09819.indb 769 9/11/2013 8:40:34 AM

 770 16 Molecular motion
The net flux from left to right is
J z = J (L → R ) − J (L ← R )
= 14 vmean {N (− λ ) − N (λ )}
⎧⎪ ⎡ ⎛ dN ⎞ ⎤ ⎡ ⎛ dN ⎞ ⎤ ⎫⎪
= 14 vmean ⎨ ⎢N (0) − λ ⎜ ⎟ ⎥ − ⎢ N (0 ) + λ ⎜
⎝ dz ⎠ 0 ⎦ ⎣ ⎝ dz ⎟⎠ 0 ⎦ ⎪
⎩⎪ ⎣
That is,
⎛ dN ⎞
J z = − 12 vmean λ ⎜
⎝ dz ⎠⎟ 0
which is eqn 79.6.
At this stage it looks as though we can pick out a value of
the diffusion coefficient by comparing eqns 79.6 and 79.3, so
obtaining D = 12 λvmean . It must be remembered, however, that
the calculation is quite crude, and is little more than an assess-
ment of the order of magnitude of D. One aspect that has not
been taken into account is illustrated in Fig. 79.4, which shows
that although a molecule may have begun its journey very close
to the window, it could have a long flight before it gets there.
Because the path is long, the molecule is likely to collide before
reaching the window, so it ought to be added to the graveyard
of other molecules that have collided. To take this effect into
account involves a lot of work, but the end result is the appear-
ance of a factor of 23 representing the lower flux. The modifica-
tion results in
D = 13 λvmean Diffusion coefficient
Brief illustration 79.2 The diffusion coefficient
In Brief illustration 78.3 of Topic 78 it is established that the
mean free path of N2 molecules in a gas at 1.0 bar is 95 nm; in
Example 78.1 of the same Topic it is calculated that the mean
speed of N2 molecules at 25 °C is 475 m s −1. Therefore, the dif-
fusion coefficient for N2 molecules under these conditions is
D = 13 × (9.5 ×10−8 m) × 475ms −1 = 1.5 × 10−5 m2 s −1
The experimental value (for N2 in O2) is 2.0 × 10−5 m2 s −1.
Self-test 79.2 Evaluate D for H2 under the same conditions.
Answer: 9.0 × 10 −5 m2 s −1
There are three points to note about eqn 79.7:
r
The mean free path, λ, decreases as the pressure is
increased (eqn 78.12 of Topic 78, λ = kT/σp), so D
decreases with increasing pressure and, as a result, the
gas molecules diffuse more slowly.
www.ebook3000.com
Atkins09819.indb 770
Short flight
(survives)
⎥⎬
⎭ Long flight
(collides in flight)
Figure 79.4 One approximation ignored in the simple
treatment is that some particles might make a long flight to
the plane even though they are only a short perpendicular
distance away, and therefore they have a higher chance of
colliding during their journey.
r
The mean speed, vmean, increases with the
temperature (eqn 78.7 of Topic 78, vmean=(8kT/πm)1/2),
so D also increases with temperature. As a result,
Physical interpretation
molecules in a hot sample diffuse more quickly than
those in a cool sample (for a given concentration
gradient).
r
Because the mean free path increases when the
collision cross-section of the molecules decreases
(eqn 78.12 of Topic 78, λ = kT/σp), the diffusion
coefficient is greater for small molecules than for large
molecules.
(b) Thermal conductivity
According to the equipartition theorem (Foundations, Topic 2),
(79.7) each molecule carries an average energy ε = νkT, where ν is a
number of the order of 1. For atoms, ␯ = 23 . When one mol-
ecule passes through the imaginary window, it transports that
average energy. We suppose that the number density is uniform
but that the temperature is not. Molecules arrive from the left
after travelling a mean free path from their last collision in a
hotter region, and therefore with a higher energy. Molecules
also arrive from the right after travelling a mean free path from
a cooler region. When this model is developed quantitatively,
as in the following Justification, it turns out that the energy flux
is proportional to the temperature gradient and that the ther-
mal conductivity is
κ = 13 ␯vmean λ N k Thermal conductivity (79.8a)
If we identify N = nNA/V = [J]NA, where [J] is the molar concen-
tration of the carrier particles J and NA is Avogadro’s constant,
and identify νkNA as the molar constant-volume heat capacity
of a perfect gas (which follows from CV,m = NA(∂ε/∂T)V), this
expression becomes
κ = 13 vmean λ[J]CV , m Thermal conductivity (79.8b)
9/11/2013 8:40:41 AM
 79 Transport properties of gases  771

Yet another form is found by recognizing that N = p/kT and the energy is transported is determined by the size of the con-
using the expression for D in eqn 79.7, for then tainer and not by the other molecules present. The flux is still
proportional to the number of carriers, but the length of the
␯pD journey no longer depends on λ, so κ ∝ [J], which implies that
κ= Thermal conductivity (79.8c)
T κ ∝ p.

Justification 79.2 Thermal conduction

Brief illustration 79.3 The thermal conductivity
The two opposing energy fluxes are
In Brief illustration 79.2 we calculated D = 1.5 × 10 −5 m 2 s −1
ZW

 ZW

 for N2 molecules at 25 °C. To use eqn 79.8c note that for N2
J (L → R ) = 4 Nvmean ε (− λ )
1
J (L ← R ) = 4 Nvmean ε (λ )
1
molecules ␯ = 25 (there are three translational modes and two
rotational; whereas the rotational modes are active at room
and the net flux is temperature, the vibrational mode is not). Therefore, at 1.0 bar,
J z = J (L → R ) − J (L ← R ) J m−3

= vmeanN {ε (− λ ) − ε(λ )}
1
4 × (1.0 ×105 Pa ) ×(1.5 × 10−5 m2 s −1 )
5
κ= 2
⎧⎪ ⎡ ⎛ dε ⎞ ⎤ ⎡ ⎛ dε ⎞ ⎤ ⎫⎪ 298 K
= 14 vmeanN ⎨ ⎢ε (0) − λ ⎜ ⎟ ⎥ − ⎢ε (0) + λ ⎜ ⎟ ⎥ ⎬
⎝ dz ⎠ 0 ⎦ ⎣ ⎝ dz ⎠ 0 ⎦ ⎪ = 1.3 ×10−2 JK −1 m −1 s −1
⎩⎪ ⎣ ⎭

That is, or 13 mW K−1 m−1.

Self-test 79.3 Estimate the thermal conductivity of argon gas
1 ⎛ dε ⎞ 1 ⎛ dT ⎞
J z = − vmean λN ⎜ ⎟ = − ␯vmean λN k ⎜ at 25 °C and 1.0 bar.
2 ⎝ ⎠0
d z 2 ⎝ dz ⎟⎠ 0
Answer: 7.6 mW K−1 m−1

The energy flux is proportional to the temperature gradient, as

we wanted to show. As before, we multiply by 23 to take long
flight paths into account, and after comparison of this equa-
tion with eqn 79.4 arrive at eqn 79.8.
To interpret eqn 79.8, we note that:
r
Because λ is inversely proportional to the pressure (eqn
78.12 of Topic 78, λ = kT/σp), and hence inversely
Physical interpretation
(c) Viscosity
Arguments very similar to those employed above, but applied
to linear momentum (see the following Justification), show that
the flux of momentum is proportional to the velocity gradient
in the fluid and that the viscosity of a perfect gas is given by
η = 13 vmean λ mN

proportional to the molar concentration of the gas, Viscosity (79.9a)

and N is proportional to the pressure (N = p/kT), the
thermal conductivity, which is proportional to the Two alternative forms of this expression (after using mNA = M)
product λp, is independent of the pressure. are
r
The thermal conductivity is greater for gases with a η = MD[J] Viscosity (79.9b)
high heat capacity (eqn 79.8b) because a given
temperature gradient then corresponds to a greater pMD
energy gradient. η= Viscosity (79.9c)
RT
The physical reason for the pressure independence of the ther- where [J] is the molar concentration of the gas molecules and
mal conductivity is that it can be expected to be large when M is their molar mass.
many molecules are available to transport the energy, but the
presence of so many molecules limits their mean free path Justification 79.3
and they cannot carry the energy over a great distance. These Viscosity
two effects balance. The thermal conductivity is indeed found Molecules travelling from the right in Fig. 79.5 (from a fast
experimentally to be independent of the pressure, except when layer to a slower one) transport a momentum mvx(λ) to their
the pressure is very low, when κ ∝ p. At low pressures λ exceeds new layer at z = 0; those travelling from the left transport
the dimensions of the apparatus, and the distance over which mvx(−λ) to it. If it is assumed that the density is uniform, the

Atkins09819.indb 771 9/11/2013 8:40:49 AM

 772 16 Molecular motion

We can interpret eqn 79.9a as follows:

x Fast
layer
r
Because λ ∝ 1/p (eqn 78.12 of Topic 78, λ = kT/σp))
and [J] ∝ p, it follows that η ∝ λN is independent of

interpretation
Slow
layer p. That is, the viscosity is independent of the

Physical
pressure.
r
Because v mean ∝ T1/2 (eqn 78.7 of Topic 78,
v mean = (8kT/πm)1/2), η ∝ T1/2. That is, the viscosity
z of a gas increases with temperature.
λ
0 The physical reason for the pressure independence of the vis-
–λ cosity is the same as for the thermal conductivity: more mol-
ecules are available to transport the momentum, but they
Figure 79.5 The calculation of the viscosity of a gas
examines the net x-component of momentum brought to a
carry it less far on account of the decrease in mean free path.
plane from faster and slower layers on average a mean free The increase of viscosity with temperature is explained when
path away in each direction. we remember that at high temperatures the molecules travel
more quickly, so the flux of momentum is greater. By contrast,
collision flux is 14 Nvmean . Those arriving from the right on as discussed in Topic 80, the viscosity of a liquid decreases with
average carry a momentum increase in temperature because intermolecular interactions
must be overcome.
⎛ dv ⎞
mvx (λ ) = mvx (0) + mλ ⎜ x ⎟
⎝ dz ⎠ 0
Brief illustration 79.4 The viscosity
Those arriving from the left bring a momentum
We have already calculated D = 1.5 × 10 −5 m2 s −1 for N2 at 25 °C.
⎛ dv ⎞ Because M = 28.02 g mol −1, for the gas at 1.0 bar, eqn 79.9c
mvx (− λ ) = mv x (0) − mλ ⎜ x ⎟
⎝ dz ⎠ 0 gives

The net flux of x-momentum in the z-direction is therefore ⎛ J m−3 ⎞

⎜ 1.0 ×105 Pa ⎟ ×(28.02 × 10−3 kg mol −1 )×(1.5 × 10−5 m2 s −1 )

⎜ ⎟
⎪⎧ ⎡ ⎛ dv ⎞ ⎤ η=
⎝ ⎠
J z = 14 vmeanN ⎨ ⎢mvx (0) − mλ ⎜ x ⎟ ⎥
⎝ dz ⎠ 0 ⎦ (8.3145 JK −1 mol −1 ) ×(298 K)
⎪⎩ ⎣
= 1.7 × 10−5 kg m −1 s −1
⎡ ⎛ dv ⎞ ⎤ ⎫⎪ ⎛ dv x ⎞
− ⎢mv x (0) + mλ ⎜ x ⎟ ⎥ ⎬ = − 12 vmean λmN ⎜⎝ dz ⎟⎠
⎣ ⎝ dz ⎠ 0 ⎦ ⎪ or 17 μPa s. The experimental value is 18 μPa s.
⎭ 0

Self-test 79.4 Evaluate the viscosity of benzene vapour at

The flux is proportional to the velocity gradient, as we wished
to show. Comparison of this expression with eqn 79.5, and 0.10 bar and 25 °C.
Answer: 1.4 × 10 −5 kg m−1 s −1
multiplication by 23 in the normal way, leads to eqn 79.9.

Checklist of concepts
☐ 1. The flux of a property is the quantity of that property
passing through a given area in a given time interval
divided by the area and the duration of the interval.
☐ 2. Fick’s first law of diffusion states that the flux of matter
is proportional to the concentration gradient.
☐ 3. In a state of Newtonian flow a series of layers of the
fluid are considered to move past one another.
☐ 4. The diffusion coefficient of a perfect gas decreases with
increasing pressure and increases with temperature; it
is greater for small molecules than for large molecules.

www.ebook3000.com

Atkins09819.indb 772 9/11/2013 8:40:54 AM

 79 Transport properties of gases  773

☐ 5. The thermal conductivity of a perfect gas is independ- ☐ 6. The viscosity of a perfect gas is independent of the pres-
ent of the pressure and is large for gases with a high heat sure and increases with temperature.
capacity.

Checklist of equations
Property Equation Comment Equation number

Matter flux J = −DdN/dz Fick’s first law 79.3

Energy of thermal motion flux J = −κdT/dz ‘Heat flux’ 79.4

x-component of momentum flux J = −ηdvx/dz Viscosity 79.5

1
Diffusion coefficient D = λvmean KMT* 79.7
3
1
Thermal conductivity κ = ␯vmeanNk KMT 79.8a
3
1
κ = vmean λ[J]CV , m 79.8b
3
κ = νpD/T 79.8c
1
Coefficient of viscosity η = vmean λ mN KMT 79.9a
3
η = MD[J] 79.9b

η = pMD/RT 79.9c

* KMT denotes that the expressions are derived on the basis of the kinetic theory of gases, a model of a perfect gas.

Atkins09819.indb 773 9/11/2013 8:40:57 AM

 TOPIC 80

Motion in liquids
In this Topic we consider two aspects of motion in liquids.
Contents First, we deal with pure liquids, and examine how the mobili-
ties of their molecules, as measured by their viscosity, vary with
80.1 Pure liquids 774 temperature. Then we consider the motion of solutes. A par-
Brief illustration 80.1: Liquid viscosity 775 ticularly simple and to some extent controllable type of motion
80.2 Electrolyte solutions 775 through a liquid is that of an ion, and we see that the informa-
(a) Conductivity 775 tion that motion provides can be used to infer the behaviour of
Example 80.1: Determining the limiting uncharged species too.
molar conductivity 776
(b) The mobilities of ions 777
Brief illustration 80.2: Ion mobility 777
(c)

(d)
Mobility and conductivity
The Einstein relations
Brief illustration 80.3: Ionic conductivity
Brief illustration 80.4: Mobility and diffusion
Checklist of concepts
Checklist of equations
➤ Why do you need to know this material?
Liquids are central to chemical reactions, and it is important
to know how the mobility of their molecules and solutes in
them varies with the conditions. Ionic motion is a way of
exploring this motion as forces to move them can be applied
electrically. From electrical measurements the properties of
diffusing neutral molecules may also be inferred.
➤ What is the key idea?
Ions reach a terminal velocity when the electrical force on
them is balanced by the drag due to the viscosity of the
solvent, which decreases with increasing temperature.
➤ What do you need to know already?
The discussion of viscosity starts with the definition of
viscosity coefficient introduced in Topic 79. One derivation
uses the same argument about flux as was used in Topic
79. The final section quotes the relation between the drift
speed and a generalized force acting on a solute particle,
which is derived in Topic 81.
778
779
80.1 Pure liquids
779 The motion of molecules in liquids can be studied experimen-
780 tally by a variety of methods. Relaxation time measurements in
780
NMR and EPR (Topics 49 and 50) can be interpreted in terms
780
of the mobilities of the molecules, and have been used to show
that big molecules in viscous fluids typically rotate in a series of
small (about 5°) steps, whereas small molecules in non-viscous
fluids typically jump through about 1 radian (57°) in each step.
Another important technique is inelastic neutron scattering,
in which the energy neutrons collect or discard as they pass
through a sample is interpreted in terms of the motion of its
particles. The same technique is used to examine the internal
dynamics of macromolecules.
More mundane than these experiments are viscosity meas-
urements. The coefficient of viscosity, η (eta), is introduced
in Topic 79 as a phenomenological coefficient, the constant of
proportionality between the flux of linear momentum and the
velocity gradient in a fluid:
dv x
J z (x -component of momentum) = −η Viscosity (80.1)
dz
(This is eqn 79.5 of Topic 79.) Some values are given in Table
80.1. The SI units of viscosity are kilograms per metre per sec-
ond (kg m−1 s−1), but they may also be reported in the equiva-
lent units of pascal seconds (Pa s). The non-SI unit poise (P)
and centipoise (cP) are still widely encountered: 1 P = 10−1 Pa s
and so 1 cP = 1 mPa s.

www.ebook3000.com

Atkins09819.indb 774 9/11/2013 8:40:58 AM

 80 Motion in liquids  775

Table 80.1* Viscosities of liquids at 298 K of the viscosity. Thus, the temperature dependence of viscosity
at constant volume, when the density is constant, is much less
η/(10−3 kg m−1 s−1)
than that at constant pressure. The intermolecular interactions
Benzene 0.601 between the molecules of the liquid govern the magnitude
Mercury 1.55 of Ea, but the problem of calculating it is immensely difficult
Pentane 0.224 and still largely unsolved. At low temperatures, the viscosity of
Water† 0.891 water decreases as the pressure is increased. This behaviour is
consistent with the need to rupture hydrogen bonds for migra-
*More values are given in the Resource section.
† The viscosity of water corresponds to 0.891 cP. tion to occur.

Unlike in a gas, for a molecule to move in a liquid it must

Brief illustration 80.1 Liquid viscosity
acquire at least a minimum energy (an ‘activation energy’ in the
language of Topic 87) to escape from its neighbours. The prob- The viscosity of water at 25 °C and 50 °C is 0.890 mPa s and
ability that a molecule has at least an energy Ea is proportional 0.547 mPa s, respectively. It follows from eqn 80.2 that the
to e − E /RT , so the mobility of the molecules in the liquid should
a
activation energy for molecular migration is the solution of
follow this type of temperature dependence. Because the coeffi-
η(T2 ) (Ea /R)(1/T2 −1/T1 )
cient of viscosity is inversely proportional to the mobility of the =e
η(T1 )
particles, we should expect that
which is
η = η0 e E /RT a
Temperature dependence of viscosity (liquid) (80.2)

Ea =
( ) (
R ln {η(T2 )/η(T1 )} 8.3145 JK −1 mol −1 ln 0.547/0.8900
=
)
where η0 is a constant independent of temperature. Note the 1/T2 −1/T1 1/(323 K) − 1/(298 K)
positive sign of the exponent, because the viscosity is inversely = 1.56 × 104 Jmol −1
proportional to the mobility. This expression implies that the
viscosity should decrease sharply with increasing temperature. or 15.6 kJ mol−1. That is comparable to the strength of a hydro-
Such a variation is found experimentally, at least over reason- gen bond (Topic 35).
ably small temperature ranges (Fig. 80.1). The activation energy
Self-test 80.1 The corresponding values of the viscosity of
typical of viscosity is comparable to the mean potential energy
benzene are 0.604 mPa s and 0.436 mPa s. Evaluate the activa-
of intermolecular interactions.
tion energy for viscosity.
One problem with the interpretation of viscosity measure-
Answer: 10.4 kJ mol−1
ments is that the change in density of the liquid as it is heated
makes a pronounced contribution to the temperature variation

Electrolyte solutions
2
80.2
Viscosity, η/(10–3 kg m–1 s–1)

1.6
Further insight into the nature of molecular motion can be
1.2
obtained by studying the net transport of charged species
through solution, for ions can be dragged through the solvent
0.8
by the application of a potential difference between two elec-
trodes immersed in the sample. By studying the transport of
0.4
charge through electrolyte solutions it is possible to build up
a picture of the events that occur in them and, in some cases,
0 to extrapolate the conclusions to species that have zero charge,
0 20 40 60 80 100 that is, to neutral molecules.
Temperature, θ/°C

Figure 80.1 The experimental temperature dependence of (a) Conductivity

the viscosity of water. As the temperature is increased, more
molecules are able to escape from the potential wells provided
by their neighbours, and so the liquid becomes more fluid. A
plot of ln η against 1/T is a straight line (over a small range) with
positive slope.
The fundamental measurement used to study the motion of
ions is that of the electrical resistance, R, of the solution. The
conductance, G, of a solution is the inverse of its resistance R:
G = 1/R. As resistance is expressed in ohms, Ω, the conductance

Atkins09819.indb 775 9/11/2013 8:41:02 AM

 776 16 Molecular motion

of a sample is expressed in Ω−1. The reciprocal ohm used to be where ν+ and ν− are the numbers of cations and anions per for-
called the mho, but its SI designation is now the siemens, S, and mula unit of electrolyte. For example, ν+ = ν− = 1 for HCl, NaCl,
1 S = 1 Ω−1 = 1 C V−1 s−1. It is found that the conductance of a and CuSO4, but ν+ = 1, ν− = 2 for MgCl2.
sample decreases with its length l and increases with its cross-
sectional area A. We therefore write
Example 80.1 Determining the limiting molar
A conductivity
G =κ Definition of κ Conductance (80.3)
l The conductivity of KCl(aq) at 25 °C is 14.668 mS m−1 when
c = 0.001 00 mol dm−3 and 71.740 mS m−1 when c = 0.005 00 mol
where κ (kappa) is the electrical conductance. With the con- dm−3. Determine the values of the limiting molar conductivity
ductance in siemens and the dimensions in metres, it follows Λ°m and the Kohlrausch constant K.
that the SI units of κ are siemens per metre (S m−1).
Method Use eqn 80.4 to determine the molar conductivities at
The conductivity of a solution depends on the number of
the two concentrations, then use the Kohlrausch law, eqn 80.5,
ions present, and it is normal to introduce the molar conduc-
in the form
tivity, Λm, which is defined as

κ
(
Λm (c2 ) − Λm (c1 ) = K c11/2 − c12/2 )
Λm = Definition Molar conductivity (80.4)
c to determine K. Then find Λ°m from the law in the form

where c is the molar concentration of the added electrolyte. Λ°m = Λm +K c1/2

The SI unit of molar conductivity is siemens metre-squared per
mole (S m2 mol−1), and typical values are about 10 mS m2 mol−1 With more data available, a better procedure is to perform a
(where 1 mS = 10−3 S). linear regression.
The values of the molar conductivity as calculated by eqn
Answer It follows that the molar conductivity of KCl when
80.4 are found to vary with the concentration. One reason for
c = 0.001 00 mol dm−3 is
this variation is that the number of ions in the solution might
not be proportional to the nominal concentration of the elec- 14.688 mSm −1
trolyte. For instance, the concentration of ions in a solution of Λm = = 14.688 mSm2 mol −1
0.00100 ×103 mol m −3
a weak electrolyte depends on the concentration of the solute
in a complicated way, and doubling the concentration of the Similarly, when c = 0.005 00 mol dm−3 the molar conductivity
solute added does not double the number of ions. Secondly, is 14.348 mS m2 mol−1. It then follows that
because ions interact strongly with one another, the conductiv-
Λm (c2 ) − Λm (c1 ) (14.348 −14.688) mSm2 mol −1
ity of a solution is not exactly proportional to the number of K= =
c1 − c2
1/2 1/2
(0.0010001/2 − 0.005001/2 ) (mol dm −3 )1/2
ions present.
In an extensive series of measurements during the nine- = 8.698 mSm2 mol −1/(mol dm −3 )1/2
teenth century, Friedrich Kohlrausch established what is now (It is best to keep this awkward but convenient array of units as
known as Kohlrausch’s law, that at low concentrations the they are rather than converting them to the equivalent 10 −3/2 S
molar conductivities of strong electrolytes vary linearly with m7/2 mol−3/2.) Now we find the limiting value from the data for
the square root of the concentration: c = 0.001 00 mol dm−3:

Λm = Λ°m − K c1/2 Kohlrausch’s law (80.5) mS m2 mol −1

The Kohlrausch constant K depends on the identity of the sol-
ute and the solvent. Kohlrausch also established that Λ°m, the
limiting molar conductivity, the molar conductivity in the
limit of zero concentration, is the sum of contributions from its
individual ions. If the limiting molar conductivity of the cati-
ons is denoted λ+ and that of the anions λ−, then his law of the
independent migration of ions states that
Law of the
Limiting
Λ°m = ␯ + λ+ + ␯ − λ− law independent
migration of ions
Λ°m = 14.688 mSm2 mol −1 + 8.698
(mol dm −3 )1/2
× (0.00100 mol dm −3 )1/2 = 14.963 mS m2 mol −1
Self-test 80.2 The conductivity of KClO4(aq) at 25 °C is 13.780
mS m−1 when c = 0.001 00 mol dm−3 and 67.045 mS m−1 when
c = 0.005 00 mol dm−3. Determine the values of the limiting
molar conductivity Λ°m and the Kohlrausch constant K for this
system.
Answer: K = 9.491 mS m2 mol −1 /(mol dm −3 )1/2 ,
(80.6) Λ°m = 14.08 mS m2 mol −1

www.ebook3000.com

Atkins09819.indb 776 9/11/2013 8:41:10 AM

 80 Motion in liquids  777

Kohlrausch’s law is used to provide an extrapolation pro- zeE

cedure for determining the limiting molar conductivity of an s= Drift speed (80.10)
f
electrolyte solution, and hence to assemble values of the limit-
ing conductivities of individual ions. As the following section It follows that the drift speed of an ion is proportional to the
explains, those values are used to explore the mobilities of ions strength of the applied field. We write
in solution and to construct models of their properties. s = uE Definition of u Mobility (80.11)

where u is called the mobility of the ion (Table 80.2).

(b) The mobilities of ions Comparison of the last two equations shows that
To interpret conductivity measurements we need to know why
ze ze
ions move at different rates, why they have different molar con- u= = Mobility (80.12)
f 6πηa
ductivities, and why the molar conductivities of strong electro-
lytes decrease with the square root of the molar concentration.
The central idea in this section is that although the motion of
an ion remains largely random, the presence of an electric field Table 80.2* Ionic mobilities in water at 298 K
biases its motion, and the ion undergoes net migration through u/(10−8 m2 s−1 V−1) u/(10−8 m2 s−1 V−1)
the solution.
When the potential difference between two planar elec- H+ 36.23 OH− 20.64

trodes a distance l apart is Δφ, the ions in the solution between Na+ 5.19 Cl− 7.91
them experience a uniform electric field of magnitude (see K+ 7.62 Br− 8.09
Foundations, Topic 2) Zn2+ 5.47 SO2−
4 8.29

Δφ * More values are given in the Resource section.

E= (80.7)
l
In such a field, an ion of charge ze experiences a force of
Brief illustration 80.2 Ion mobility
magnitude
For an order of magnitude estimate we can take z = 1 and a as
zeΔφ the radius of an ion such as Cs + (which might be typical of a
F = zeE = Electric force (80.8)
l smaller ion plus its hydration sphere), which is 170 pm. For the
viscosity, we use η = 1.0 cP (1.0 mPa s, Table 80.1). Then
where here and throughout this section we disregard the sign of
J V -1
the charge number and so avoid notational complications.

A cation responds to the application of the field by accelerat- 1.6 ×10 −19
C
u= = 5.0 ×10−8 m2 V −1 s −1
ing towards the negative electrode and an anion responds by ⎛ ⎞
accelerating towards the positive electrode. However, this accel- 6π × ⎜ 1.0 × 10−3 Pa
 s⎟ × (170 × 10 m)
−12
⎜⎝ ⎟
J m-3 ⎠
eration is short-lived. As the ion moves through the solvent it
experiences a frictional retarding force, Ffric, proportional to its
speed. For a spherical particle of radius a travelling at a speed This value means that when there is a potential difference of
s, this force is given by Stokes’s law, which was derived by con- 1 V across a solution of length 1 cm (so E = 100 V m−1), the drift
sidering the hydrodynamics of the passage of a sphere through speed is typically about 5 μm s −1. That speed might seem slow,
but not when expressed on a molecular scale, for it corresponds
a continuous fluid:
to an ion passing about 104 solvent molecules per second.
Ffric = fs f = 6πηa Stokes’s law (80.9) 4 ion in water at 25 °C is
Self-test 80.3 The mobility of an SO2−
8.29 × 10 −8 m2 V−1 s −1. What is its effective radius? Use the vis-
where η is the viscosity. In writing eqn 80.9, we have assumed cosity given above.
that it applies on a molecular scale, and independent evidence Answer: 205 pm
from magnetic resonance suggests that it often gives at least the
right order of magnitude.
The two forces act in opposite directions, and the ions
quickly reach a terminal speed, the drift speed, when the accel- Because the drift speed governs the rate at which charged
erating force is balanced by the viscous drag. The net force is species are transported, we might expect the conductivity
zero when fs = zeE, or to decrease with increasing solution viscosity and ion size.

Atkins09819.indb 777 9/11/2013 8:41:16 AM

 778 16 Molecular motion

Experiments confirm these predictions for bulky ions (such +

as R4N+ and RCO2− ) but not for small ions. For example, the
mobilities of the alkali metal ions in water increase from Li+ to
Cs+ (Table 80.2) even though the ionic radii increase. The para-
dox is resolved when we realize that the radius a in the Stokes
formula is the hydrodynamic radius (or ‘Stokes radius’) of the
ion, its effective radius in the solution taking into account all
the H2O molecules it carries in its hydration shell. Small ions
give rise to stronger electric fields than large ones (the electric
field at the surface of a sphere of radius r is proportional to ze/r2
and it follows that the smaller the radius, the stronger the field),
so small ions are more extensively solvated than big ions. Thus,
an ion of small ionic radius may have a large hydrodynamic
radius because it drags many solvent molecules through the
solution as it migrates. The hydrating H2O molecules are often
very labile, however, and NMR and isotope studies have shown
that the exchange between the coordination sphere of the ion
and the bulk solvent is very rapid for ions of low charge but may
be slow for ions of high charge (Fig. 80.2).
The proton, although it is very small, has a very high mobility
(Table 80.2)! Proton and 17O-NMR show that the times char-
acteristic of protons hopping from one molecule to the next
are about 1.5 ps, which is comparable to the time that inelas-
tic neutron scattering shows it takes a water molecule to re-
orientate through about 1 radian (1 to 2 ps). According to the
Grotthuss mechanism, there is an effective motion of a proton
that involves the rearrangement of bonds in a group of water
molecules (Fig. 80.3). However, the actual mechanism is still
highly contentious. The mobility of NH +4 in liquid ammonia
is also anomalous and presumably occurs by an analogous
mechanism.
(c) Mobility and conductivity
Ionic mobilities provide a link between measurable and theo-
retical quantities. As a first step we establish in the following
1 μs 1s 1h
+1
Na Li
H
+
Figure 80.3 The Grotthuss mechanism of conduction by
protons in water. There is a virtual transfer of a proton from one
end of the chain to the other as the bonds and hydrogen bonds
adjust.
Justification the relation between an ion’s mobility and its molar
conductivity:
λ = zuF Ion conductivity (80.13)
where F is Faraday’s constant (F = NAe).
Justification 80.1
The relation between ionic mobility
and molar conductivity
To keep the calculation simple, we ignore signs in the follow-
ing, and concentrate on the magnitudes of quantities.
Consider a solution of a fully dissociated strong electrolyte
at a molar concentration c. Let each formula unit give rise to
ν+ cations of charge z +e and ν− anions of charge z −e. The molar
concentration of each type of ion is therefore νc (with ν = ν+ or
ν−), and the number density of each type is νcNA. The number
of ions of one kind that pass through an imaginary window of
area A during an interval Δt is equal to the number within the
distance sΔt (Fig. 80.4), and therefore to the number in the vol-
ume sΔtA. (The same sort of argument is used in Topic 79 in
the discussion of the transport properties of gases.) The num-
ber of ions of that kind in this volume is equal to sΔtAνcNA.
The flux through the window (the number of this type of ion
1y
Anions
Area, A
Cations

ions
Zn, Mn
Fe, Mg
Hg, Ba
Cr, Ca

+2
Cu

Cd

Co

Ru
Ni

Pt
V

ions

s–Δt
La, Sc

+3 s+Δt
Gd

Ga

Rh
Ru
Fe

Cr
Al
In
Y

Ir

ions

–10 –9 –6 –4 –2 0 2 4 6 8 10 Figure 80.4 In the calculation of the current, all the cations
log(t1/2/s) within a distance s+Δt (that is, those in the volume s+AΔt) will
pass through the area A. The anions in the corresponding
Figure 80.2 The half-lives of water molecules in the hydration volume on the other side of the window will also contribute
spheres of ions. to the current similarly.

www.ebook3000.com

Atkins09819.indb 778 9/11/2013 8:41:20 AM

 80 Motion in liquids  779

passing through the window divided by the area of the win-

But 1 V−1 s −1 C = 1 S (see the remark preceding eqn 80.3), so
dow and the duration of the interval) is therefore
λ ≈ 5 mS m 2 mol−1, and, from eqn 80.14b, about twice that
sΔtA␯cN A value for Λ°m , in accord with experiment. The experimental
J (ions) = = s␯cN A
ΔtA value for KCl, for instance, is 15 mS m2 mol−1.
Each ion carries a charge ze, so the flux of charge is Self-test 80.4 Estimate the ionic conductivity of an SO2−
4 ion
in water at 25 °C from its mobility (Table 80.2).
J (charge) = zs␯ceN A = zs␯cF
Answer: 16 mS m 2 mol−1
Because s = uE, the flux is
J (charge) = zu␯cFE

The current, I, through the window due to the ions we are con- (d) The Einstein relations
sidering is the charge flux times the area: An important relation between the drift speed s and a force F of
I = JA = zu␯cFE A any kind acting on a particle is derived in Topic 81:

Because the electric field is the potential gradient (eqn 80.7, DF

s= Drift speed (80.15)
E = Δφ/l), we can write RT

zu␯cFAΔφ where D is the diffusion coefficient for the species and F is the
I= thermodynamic force (a molar quantity). We have seen that
l
an ion in solution has a drift speed s = uE in the presence of an
Current and potential difference are related by Ohm’s law, electric field of strength E and experiences a thermodynamic
Δφ = IR, so it follows that force of magnitude NAzeE. Therefore, substituting these known
Δφ κAΔφ
values into eqn 80.15 and using NAe = F gives uE = DFzE/RT
I= = GΔφ = and hence, on cancelling the E, we obtain the Einstein relation:
R l
zDF
where we have used eqn 80.3. Comparison of the last two u= Einstein relation (80.16)
RT
expressions gives κ = zuνcF. Division by the molar concentra-
tion of ions, νc, then results in eqn 80.13. The Einstein relation provides a link between the molar con-
ductivity of an electrolyte and the diffusion coefficients of its
ions. First, by using eqns 80.13 and 80.16 we write

Equation 80.13 applies to the cations and to the anions. z 2 DF 2

λ = zuF = (80.17)
Therefore, for the solution itself in the limit of zero concentra- RT
tion (when there are no ionic interactions),
for each type of ion. Then, from Λ°m = ␯ + λ+ + ␯ − λ− , the limiting
Λ°m = (z + u+ ␯ + + z − u− ␯ − )F (80.14a) molar conductivity is

For a symmetrical z:z electrolyte (for example, CuSO4 with z = 2 F2

Λ°m = (␯ + z +2 D+ + ␯ − z −2 D− ) Nernst–Einstein equation (80.18)
and ν+ = ν− = 1), this equation simplifies to RT

Λ°m = z (u+ + u− )F (80.14b) which is the Nernst–Einstein equation. An application of this

Brief illustration 80.3
Ionic conductivity
In Brief illustration 80.2, we estimated the typical ionic mobil-
ity as 5.0 × 10−8 m2 V−1 s−1; so, with z = 1 for both the cation and
anion, we can estimate that a typical limiting molar conduc-
tivity should be about
λ = (5.0 × 10−8 m2 V −1 s −1 ) × (9.648 × 104 C mol −1 )
= 4.8 ×10−3 m2 V −1 s −1 C mol −1
equation is to the determination of ionic diffusion coefficients
from conductivity measurements; another is to the prediction
of conductivities using models of ionic diffusion.
Equations 80.12 (u = ez/f) and 80.16 (u = zDF/RT in the form
u = zDe/kT) relate the mobility of an ion to the frictional force
and to the diffusion coefficient, respectively. We can combine
the two expressions and cancel the ze and obtain the Stokes–
Einstein equation:
kT
D= Stokes–Einstein equation (80.19a)
f

Atkins09819.indb 779 9/11/2013 8:41:30 AM

 780 16 Molecular motion

If the frictional force is described by Stokes’s law, then we also Brief illustration 80.4 Mobility and diffusion
obtain a relation between the diffusion coefficient and the vis-
4 is 8.29 ×10 m V s .
From Table 80.2, the mobility of SO2− −8 2 −1 −1
cosity of the medium:
It follows from eqn 80.16 in the form D = uRT/zF that the diffu-
kT sion coefficient for the ion in water at 25 °C is
D= Stokes–Einstein equation (80.19b)
6πηa (8.29 ×10−8 m2 V −1 s −1 ) × (8.3145 JK −1 mol −1 ) × (298 K )
D=
⎛ ⎞
An important feature of eqn 80.19 is that it makes no reference 2 × ⎜ 9.649 ×104 C mol −1
⎟

to the charge of the diffusing species. Therefore, the equation ⎜⎝ J V -1
⎟⎠
also applies in the limit of vanishingly small charge; that is, it = 1.06 × 10−9 m2 s −1
also applies to neutral molecules. This feature is taken further
in Topic 81 where the general features of diffusion are dis- Self-test 80.5 Repeat the calculation for the NH +4 ion.
cussed. It must not be forgotten, however, that both equations Answer: 1.96 × 10 −9 m2 s −1
depend on the assumption that the viscous drag is proportional
to the speed.

Checklist of concepts
☐ 1. The viscosity of a liquid decreases with increasing ☐ 5. An ion reaches a drift speed when the acceleration due
temperature. to the electrical force is balanced by the viscous drag.
☐ 2. The conductance of a solution is the inverse of its ☐ 6. The hydrodynamic radius of an ion may be greater
resistance. than its geometrical radius due to solvation.
☐ 3. Kohlrausch’s law states that at low concentrations the ☐ 7. The high mobility of a proton in water is explained by
molar conductivities of strong electrolytes vary linearly the Grotthuss mechanism.
with the square root of the concentration.
☐ 4. The law of the independent migration of ions states that
the molar conductivity in the limit of zero concentra-
tion is the sum of contributions from its individual ions.

Checklist of equations
Property Equation Comment Equation number

Viscosity η = η0 e Ea /RT Narrow temperature range 80.2

Conductivity κ = Gl/A Definition, G = 1/R 80.3

Molar conductivity Λm = κ/c Definition 80.4

Kohlrausch’s law Λm = Λ°m −K c1/2 Empirical observation 80.5

Law of independent migration of ions Λ°m = ␯ + λ+ + ␯ − λ− Limiting law 80.6

Stokes’s law Ffric = fs, f = 6πηa Hydrodynamic radius a 80.9

Drift speed s = uE Defines u 80.11

Ion mobility u = ze/6πηa Assumes Stokes’s law 80.12

Conductivity and mobility λ = zuF 80.13

Molar conductivity and mobility Λ°m = (z + u+ ␯ + + z − u− ␯ − )F Limiting law 80.14

www.ebook3000.com

Atkins09819.indb 780 9/11/2013 8:41:35 AM

 80 Motion in liquids  781

Property Equation Comment Equation number

Drift speed s = DF/RT 80.15

Einstein relation u = zDF/RT 80.16

Nernst–Einstein relation Λ°m = (␯ + z +2 D+ + ␯ − z −2 D− )(F 2 /RT ) 80.18

Stokes–Einstein relation D = kT/f 80.19

Atkins09819.indb 781 9/11/2013 8:41:36 AM

 TOPIC 81

Diffusion

Contents Topic 80. It uses the concept of chemical potential (Topic

81.1 The thermodynamic view
Brief illustration 81.1: The thermodynamic force
Brief illustration 81.2: The thermodynamic force
and the drift speed
81.2 The diffusion equation
(a) Simple diffusion
Brief illustration 81.3: The diffusion equation
(b) Diffusion with convection
Brief illustration 81.4: Convection
(c) Solutions of the diffusion equation
Example 81.1: Calculating the average
displacement
81.3 The statistical view
Brief illustration 81.5: Random walk
Checklist of concepts
Checklist of equations
➤ Why do you need to know this material?
Diffusion is a hugely important process both in the
atmosphere and in solution, and it is important to be able
to predict the spread of one material through another
when discussing reactions in solution and the spread of
substances into the environment.
➤ What is the key idea?
Particles tend to spread and achieve a uniform distribution.
➤ What do you need to know already?
This Topic draws on arguments relating to flux that are
treated in Topic 79, particularly the way to calculate
the flux of particles through a window of given area.
This Topic goes into more detail about the diffusion
coefficient, which was introduced in Topic 79 and used in
69) to discuss the direction of spontaneous change. One
782
of the mathematical manipulations draws on Stirling’s
783
approximation, which is introduced in Topic 51.
784
784
That solutes in gases, liquids, and solids have a tendency to
784
spread can be discussed from three points of view. One view-
785
point is from the Second Law of thermodynamics and the ten-
785
dency for entropy to increase or, if the temperature and pressure
785
are constant, for the Gibbs energy to decrease. When this law is
786
applied to solutes it appears that there is a force acting to disperse
the solute. That force is illusory, but it provides an interesting and
786
useful approach to discussing diffusion. The second approach
787
is to set up a differential equation for the change in concentra-
788
tion in a region by considering the flux of material through its
789
boundaries. The resulting ‘diffusion equation’ can then be solved
789
(in principle, at least) for various configurations of the system,
such as the shape of a reaction vessel. The third approach is
more mechanistic, and is to imagine diffusion as taking place in
a series of random small steps: it enriches the thermodynamic
view by providing a model of what is taking place when matter
diffuses. We explore all three viewpoints in this Topic.
81.1 The thermodynamic view
In Topic 64 it is established that, at constant temperature and
pressure, the maximum non-expansion work that can be done
when a substance moves from a location where its Gibbs energy
is G to a location where its Gibbs energy is G + dG is dw = dG.
Since the chemical potential is the (partial) molar Gibbs energy
(Topic 69), the maximum non-expansion work per mole is
dw = dμ. In a system in which the chemical potential depends
on the position x,
⎛ ∂μ ⎞
d w = dμ = ⎜ ⎟ d x
⎝ ∂ x ⎠ T, p

www.ebook3000.com

Atkins09819.indb 782 9/11/2013 8:41:37 AM

 81 Diffusion  783

It is also established in Foundations, Topic 2 that in general In Topic 79 it is established that Fick’s first law of diffusion,
work can always be expressed in terms of an opposing force which we write here in the form
(which here we write F), and that
dN
dw = −F dx J (number ) = − D Fick’s first law (81.3)
dx
By comparing these two expressions, we see that the slope of
the chemical potential can be interpreted as an effective force can be deduced from the kinetic model of gases. Here we gen-
per mole of molecules. We write this thermodynamic force as eralize that result. We show that it can be deduced more gener-
ally and that it applies to the diffusion of species in condensed
⎛ ∂μ ⎞
F = −⎜ ⎟ Thermodynamic force (81.1) phases too. To do so, we suppose that the flux of diffusing par-
⎝ ∂x ⎠ T , p
ticles is motion in response to a thermodynamic force arising
There is not necessarily a real force pushing the particles down from a concentration gradient. The diffusing particles reach
the slope of the chemical potential. As we shall see, the force a steady ‘drift speed’, s, when the thermodynamic force, F, is
may represent the spontaneous tendency of the molecules to matched by the drag due to the viscosity of the medium. This
disperse as a consequence of the Second Law and the hunt for drift speed is proportional to the thermodynamic force, and we
maximum entropy. write s ∝ F. However, the particle flux, J, is proportional to the
In a solution in which the activity of the solute is a, the drift speed, and the thermodynamic force is proportional to the
chemical potential is μ = μ< + RT ln a (see Table 72.1 of Topic concentration gradient, dc/dx. The chain of proportionalities
72). If the solution is not uniform the activity depends on the (J ∝ s, s ∝ F, and F ∝ dc/dx) implies that J ∝ dc/dx, which is the
position and we can write content of Fick’s law.
If we divide both sides of eqn 81.3 by Avogadro’s constant,
⎛ ∂ ln a ⎞ thereby converting numbers into amounts (numbers of moles),
F = − RT ⎜ (81.2a)
⎝ ∂x ⎟⎠ T , p noting that N/NA = (N/V)/NA = (nNA/V)/NA = n/V = c, the molar
concentration, then Fick’s law becomes
If the solution is ideal, a may be replaced by c/c<, and then
dc
⎛ ∂ ln(c / c < ) ⎞ RT ⎛ ∂c ⎞ J (amount) = − D (81.4)
F = − RT ⎜ ⎟⎠ = − c ⎜⎝ ∂x ⎟⎠ (81.2b)
dx
⎝ ∂x T,p T,p
In this expression, D is the diffusion coefficient and dc/dx is the
where we have also used the relation d ln y/dx = (1/y)(dy/dx). slope of the molar concentration. The flux is related to the drift
speed by
Brief illustration 81.1 The thermodynamic force J (amount) = sc (81.5)
Suppose a linear concentration gradient is set up across a
container at 25 °C, with points separated by 1.0 cm differing This relation follows from the argument used in Topic 79. Thus,
in concentration by 0.10 mol dm−3 around a mean value of all particles within a distance sΔt, and therefore in a volume
1.0 mol dm−3. According to eqn 81.2b, the solute experiences a sΔtA, can pass through a window of area A in an interval Δt.
thermodynamic force of magnitude Hence, the amount of substance that can pass through the win-
dow in that interval is sΔtAc. The particle flux is this quantity
(8.3145 JK −1 mol −1 ) × (298 K ) 0.10 mol dm −3 divided by the area A and the time interval Δt, and is therefore
F= ×
1.0 mol dm −3 1.0 ×10−2 m simply sc.

N
By combining the last two equations and using eqn 81.2
= 2.5 × 104 Jm −1 mol −1
dc DcF DF
or 25 kN mol−1. Note that the thermodynamic force is a molar sc = − D = or s = (81.6)
dx RT RT
quantity.
Self-test 81.1 Suppose that the concentration of a solute Therefore, once we know the effective force and the diffusion
decreases exponentially as c(x ) = c0e − x/λ . Derive an expres- coefficient, D, we can calculate the drift speed of the particles
sion for the thermodynamic force. (and vice versa), whatever the origin of the force. This equa-
Answer: F = RT/λ tion is used in Topic 80 where the force is applied electrically
to an  ion.

Atkins09819.indb 783 9/11/2013 8:41:45 AM

 784 16 Molecular motion

Brief illustration 81.2 The thermodynamic force Justification 81.1 The diffusion equation
and the drift speed
Consider a thin slab of cross-sectional area A that extends
Laser measurements showed that a molecule has a drift from x to x + λ (Fig. 81.1). Let the concentration at x be c at
speed of 1.0 μm s −1 in water at 25 °C, with diffusion coefficient the time t. The rate at which the amount (in moles) of particles
5.0 × 10 −9 m 2 s −1. The corresponding thermodynamic force enter the slab is JA, so the rate of increase in molar concentra-
from eqn 81.6 in the form F = sRT/D is tion inside the slab (which has volume Aλ) on account of the
flux from the left is
(1.0 ×10−6 ms −1 ) × (8.3145 JK −1 mol −1 ) × (298 K )
F= ∂c JA J
(5.0 ×10−9 m2 s −1 ) = =
∂t Aλ λ

N
= 5.0 × 105 Jm −1 mol −1 Volume, Aλ J(x+λ)
Area, A
or about 500 kN mol−1. J(x)

Self-test 81.2 What is the drift speed of a sucrose molecule in

water at 25 °C if the thermodynamic force is 250 kN mol−1? Use
D = 5.4 × 10−10 m2 s −1.
Answer: 5.3 μm s −1
x+λ
x

Figure 81.1 The net flux in a region is the difference

between the flux entering from the region of high
81.2 The diffusion equation concentration (on the left) and the flux leaving to the region
of low concentration (on the right).
We now turn to the discussion of time-dependent diffusion
processes, where we are interested in the spreading of inho- There is also an outflow through the right-hand window. The
mogeneities with time. One example is the temperature of a flux through that window is J′, and the rate of change of con-
metal bar that has been heated at one end: if the source of heat centration that results is
is removed, then the bar gradually settles down into a state of ∂c J′
uniform temperature. When the source of heat is maintained =−
∂t λ
and the bar can radiate, it settles down into a steady state of
non-uniform temperature. Another example (and one more The net rate of change of concentration is therefore
relevant to chemistry) is the concentration distribution in a sol- ∂c J − J ′
vent to which a solute is added. We shall focus on the descrip- =
∂t λ
tion of the diffusion of particles, but similar arguments apply to
Each flux is proportional to the concentration gradient at the
the diffusion of physical properties, such as temperature. Our
respective window. So, by using Fick’s first law, we can write
aim is to obtain an equation for the rate of change of the con-
centration of particles in an inhomogeneous region. ∂c ∂c ′
J − J ′ = −D +D
∂x ∂x

(a) Simple diffusion The concentration at the right-hand window is related to that
on the left by
The central equation of this section is the diffusion equation,
also called ‘Fick’s second law of diffusion’, which relates the rate ⎛ ∂c ⎞
c′ = c + ⎜ ⎟ λ
of change of concentration at a point to the spatial variation of ⎝ ∂x ⎠
the concentration at that point:
which implies that

∂c ∂2 c ∂c ∂ ⎧ ⎛ ∂c ⎞ ⎫ ∂ 2c
=D 2 Diffusion equation (81.7) J − J ′ = −D + D ⎨c + ⎜ ⎟ λ ⎬ = Dλ 2
∂t ∂x ∂x ∂x ⎩ ⎝ ∂x ⎠ ⎭ ∂x

We show in the following Justification that the diffusion equa- When this relation is substituted into the expression for the
tion follows from Fick’s first law of diffusion. rate of change of concentration in the slab, we get eqn 81.7.

www.ebook3000.com

Atkins09819.indb 784 9/11/2013 8:41:51 AM

 81 Diffusion  785

(b) Diffusion with convection

Spreads
The transport of particles arising from the motion of a stream-
Concentration, c

Negative
ing fluid is called convection. If for the moment we ignore dif-
curvature fusion, then the flux of particles through an area A in an interval
Δt when the fluid is flowing at a velocity v can be calculated in
Fills the way we have used several times elsewhere (such as in Topic
79, by counting the particles within a distance vΔt), and is
Positive
curvature cAvΔt
J conv = = cv Convective flux (81.8)
Position, x AΔt

Figure 81.2 Nature abhors a wrinkle. The diffusion equation This J is called the convective flux. The rate of change of con-
tells us that peaks in a distribution (regions of negative centration in a slab of thickness l and area A is, by the same
curvature) spread and troughs (regions of positive curvature) argument as before and assuming that the velocity does not
fill in. depend on the position,

∂c J conv − J conv
′ cv ⎧ ⎛ ∂c ⎞ ⎫ v
= = − ⎨c + ⎜ ⎟ λ ⎬
The diffusion equation shows that the rate of change of con- ∂t λ λ ⎩ ⎝ ∂x ⎠ ⎭ λ
Convection (81.9)
centration is proportional to the curvature (more precisely, to ⎛ ∂c ⎞
the second derivative) of the concentration with respect to dis- = −⎜ ⎟ v
⎝ ∂x ⎠
tance. If the concentration changes sharply from point to point
(if the distribution is highly wrinkled), then the concentration When both diffusion and convection occur, the total change of
changes rapidly with time. Where the curvature is positive concentration in a region is the sum of the two effects, and the
(a dip, Fig. 81.2), the change in concentration is positive; the generalized diffusion equation is
dip tends to fill. Where the curvature is negative (a heap), the
change in concentration is negative; the heap tends to spread. ∂c ∂2 c ∂c
= D 2 −v Generalized diffusion equation (81.10)
If the curvature is zero, then the concentration is constant in ∂t ∂x ∂x
time. If the concentration decreases linearly with distance, then
the concentration at any point is constant because the inflow of A further refinement, which is important in chemistry, is the
particles is exactly balanced by the outflow. possibility that the concentrations of particles may change as
The diffusion equation can be regarded as a mathematical a result of reaction. When reactions are included in eqn 81.10
formulation of the intuitive notion that there is a natural ten- (Topic 88) we get a powerful differential equation for discuss-
dency for the wrinkles in a distribution to disappear. More suc- ing the properties of reacting, diffusing, convecting systems
cinctly: Nature abhors a wrinkle. and which is the basis of reactor design in chemical industry
and of the utilization of resources in living cells.

Brief illustration 81.3 The diffusion equation
If a concentration falls linearly across a small region of space,
in the sense that c = c0 – ax, then ∂ 2c/∂x 2 = 0 and consequently
∂c/∂t = 0. The concentration in the region is constant because
the inward flow through one window is matched by the out-
ward f low through the other window. If the concentration
varies as c = c0 − 12 ax 2 , then ∂2 c/∂x 2 = −a and consequently
∂c/∂t = −Da. Now the concentration decreases, because there
is a greater outward flow than inward flow.
Self-test 81.3 What is the change in concentration when
the concentration falls exponentially across a region? Take
c = c0 e − x /λ .
Answer: ∂c/∂t = (D/λ 2)c
Brief illustration 81.4 Convection
Here we continue the discussion of the systems treated in
Brief illustration 81.3 and suppose that there is a convective
flow v. If the concentration falls linearly across a small region
of space, in the sense that c = c 0 − ax, then ∂c/∂x = −a and the
change in concentration in the region is ∂c/∂t = av. There is
now an increase in the region because the inward convective
flow outweighs the outward flow, and there is no diffusion. If
a = 0.010 mol dm−3 m−1 and v = +1.0 mm s −1,
∂c
= (0.010 mol dm −3 m −1 ) × (1.0 ×10−3 ms −1 )
∂t
= 1.0 ×10−5 mol dm −3 s −1

Atkins09819.indb 785 9/11/2013 8:41:56 AM

 786 16 Molecular motion

Table 81.1* Diffusion coefficients at 298 K

and the concentration increases at the rate of 10 μmol dm−3 s−1.
D/(10−9 m2 s−1)
Self-test 81.4 What rate of flow is needed to replenish the con-
centration when the concentration varies exponentially as H+ in water 9.31
c = c0e−x/λ across the region? I2 in hexane 4.05
Answer: v = D/λ
Na+ in water 1.33
Sucrose in water 0.522
* More values are given in the Resource section.
(c) Solutions of the diffusion equation
The diffusion equation is a second-order differential equation various times, and it is clear that the concentration spreads and
with respect to space and a first-order differential equation tends to uniformity.
with respect to time. Therefore, we must specify two boundary Another useful result is for a localized concentration of sol-
conditions for the spatial dependence and a single initial condi- ute in a three-dimensional solvent (a sugar lump suspended in
tion for the time dependence (see Mathematical background 2). a large flask of water). The concentration of diffused solute is
As an illustration, consider a solvent in which the solute is spherically symmetrical, and at a radius r is
initially coated on one surface of the container (for example,
a layer of sugar on the bottom of a deep beaker of water). The n0
c(r , t ) = e − r /4 Dt
2
Three-dimensional diffusion (81.12)
single initial condition is that at t = 0 all N0 particles are con- 8(πDt )3/2
centrated on the yz-plane (of area A) at x = 0. The two bound-
ary conditions are derived from the requirements (1) that the Other chemically (and physically) interesting arrangements,
concentration must everywhere be finite and (2) that the total such as transport of substances across biological mem-
amount (number of moles) of particles present is n0 (with branes, can be treated. In many cases the solutions are more
n0 = N0/NA) at all times. These requirements imply that the flux cumbersome.
of particles is zero at the top and bottom surfaces of the system. The solutions of the diffusion equation are useful for experi-
Under these conditions it is found that mental determinations of diffusion coefficients (Table 81.1).
In the capillary technique, a capillary tube, open at one end
n0 and containing a solution, is immersed in a well-stirred larger
c( x , t ) = e − x /4 Dt
2
One-dimensional diffusion (81.11)
A(πDt )1/2 quantity of solvent, and the change of concentration in the
tube is monitored. The solute diffuses from the open end of the
as may be verified by direct substitution (Problem 81.5). Figure capillary at a rate that can be calculated by solving the diffu-
81.3 shows the shape of the concentration distribution at sion equation with the appropriate boundary conditions, so D
may be determined. In the diaphragm technique, the diffu-
2.5
sion occurs through the capillary pores of a sintered glass dia-
0.05 phragm separating the well-stirred solution and solvent. The
2
concentrations are monitored and then related to the solutions
Concentration, c/(n0/A)

0.1 of the diffusion equation corresponding to this arrangement.

1.5
Diffusion coefficients may also be measured by a number of
techniques, including NMR spectroscopy.
1
0.3 The solutions of the diffusion equation can be used to predict
the concentration of particles (or the value of some other physi-
1.0
0.5
cal quantity, such as the temperature in a non-uniform system)
at any location. We can also use them to calculate the average
0 displacement of the particles in a given time.
0 0.5 1 1.5 2
Distance,x Example 81.1 Calculating the average displacement
Figure 81.3 The concentration profiles above a plane from
Calculate the average displacement of particles in a time t in
which a solute is diffusing. The curves are plots of eqn 81.11 and
a one-dimensional system if they have a diffusion constant D.
are labelled with different values of Dt. The units of Dt and x
are arbitrary, but are related so that Dt/x2 is dimensionless. For Method We need to use the results of probability theory
example, if x is in metres, Dt would be in metres squared; so, for summarized in Mathematical background 7. In this case, we
D = 10−9 m2 s−1, Dt = 0.1 m2 corresponds to t = 108 s.

www.ebook3000.com

Atkins09819.indb 786 9/11/2013 8:41:58 AM

 81 Diffusion  787

calculate the probability that a particle will be found at a cer- 0 1m

tain distance from the origin, and then calculate the average
–2 1 cm
by weighting each distance by that probability. You will need
1 mm
to use eqn 81.11 as well as Integral G.2 listed in the Resource

log(〈x 2〉1/2/m)
–4
section.
Answer The number of particles in a slab of thickness dx and –6 1 μm
area A at x, where the molar concentration is c, is cANAdx. The
probability that any of the N0 = n0NA particles is in the slab is –8
therefore cANAdx/N0. If the particle is in the slab, it has trav- 1 nm
1 μs 1 ms 1s 1h 1d 1y
elled a distance x from the origin. Therefore, the average dis- –10
placement of all the particles is the sum of each x weighted by –10 –6 –2 0 2 6
the probability of its occurrence: log(t/s)


Integral G.2
 Figure 81.4 The root mean square distance covered by
∞ ∞ 1/2
c(x , t )AN A 1 ⎛ Dt ⎞ particles with D = 5 × 10−10 m2 s−1. Note the great slowness of
〈x〉 = ∫
0
x
N
 0
dx =
( πDt )1/2 ∫0
xe − x 2 /4 Dt
dx = 2 ⎜ ⎟
⎝ π⎠ diffusion.
n0 N A

The average displacement varies as the square root of the

lapsed time. 81.3 The statistical view
Self-test 81.5 Derive an expression for the root mean square
distance travelled by diffusing particles in a time t in a one- An intuitive picture of diffusion is of the particles moving in a
dimensional system. You will need Integral G.3 listed in the series of small steps and gradually migrating from their origi-
Resource section. nal positions. We explore this idea by using a model in which
Answer: 〈x 2〉1/2 = (2Dt)1/2 the particles can jump through a distance λ in a time τ. The
total distance travelled by a particle in a time t is therefore tλ/τ.
However, the particle will not necessarily be found at that dis-
tance from the origin. The direction of each step may be dif-
As shown in Example 81.1, the average displacement of a dif- ferent, and the net distance travelled must take the changing
fusing particle in a time t in a one-dimensional system is directions into account.
If we simplify the discussion by allowing the particles to
1/2
⎛ Dt ⎞ travel only along a straight line (the x-axis), and for each step
〈x〉 = 2⎜ ⎟ One dimension Mean displacement (81.13) (to the left or the right) to be through the same distance λ, then
⎝ π⎠
we obtain the one-dimensional random walk. We show in the
and the root mean square displacement in the same time is following Justification that the probability of a particle being at
(from Self-test 81.5) a distance x from the origin after a time t is

1/2
〈 x 2 〉1/2 = (2Dt )1/2 One Root mean square ⎛ 2τ ⎞
P (x , t ) = ⎜ ⎟ e − x τ /2t λ
2 2
dimension displacement (81.14) One dimension Probability (81.15)
⎝ πt ⎠

The latter is a valuable measure of the spread of particles when

they can diffuse in both directions from the origin (for then
〈x〉 = 0 at all times). The root mean square displacement of par- Justification 81.2 The one-dimensional random walk
ticles with a typical diffusion coefficient (D = 5 × 10−10 m2 s−1) is
Consider a one-dimensional random walk of N steps in which
illustrated in Fig. 81.4, which shows how long it takes for diffu-
each step is through a distance λ to the left or right. The num-
sion to increase the net distance travelled on average to about ber of ways of performing a walk with NR steps to the right and
1 cm in an unstirred solution. The graph shows that diffusion NL to the left is given by the binomial coefficient (Mathematical
is a very slow process (which is why solutions are stirred, to background 7)
encourage mixing by convection). The diffusion of phero-
mones in still air is also very slow, and greatly accelerated by N! N!
W= =
convection. N L ! N R ! (N − N R )! N R !

Atkins09819.indb 787 9/11/2013 8:42:04 AM

 788 16 Molecular motion

λ2
The total possible number of paths is 2N as each of the N steps D= Einstein–Smoluchowski equation (81.16)
may be in either of two directions. The probability of the 2τ
net distance walked being nλ with n = N R − N L = 2N R − N is
therefore
Brief illustration 81.5 Random walk
number of paths with N R steps to the right
P (nλ ) = Suppose an SO2− ion jumps through its own diameter each
total number of paths 4

W N! time it makes a move in an aqueous solution. Then, because

= N = D = 1.1 × 10 −9 m 2 s −1 (Table 81.1) and a = 250 pm (as deduced
2 (N − N R )! N R !2N
from mobility measurements, Topic 80), it follows from λ = 2a
This expression can be developed by making use of Stirling’s that
approximation (Topic 51) in the form
(2a)2 2a2 2 × (250 × 10−12 pm)2
τ= = = = 1.1 × 10−10 s
ln x ! ≈ ln(2π) + (x +
1/2
2 )ln x − x
1 2D D 1.1 × 10−9 m2 s −1

and the parameter or τ = 110 ps. Because τ is the time for one jump, the ion makes
about 1 × 1010 jumps per second.
NR 1
μ= − 1 Self-test 81.6 Suppose the activation energy for diffusion is
N 2
40 kJ mol−1. What would be the jump time when the tempera-
which is small because almost exactly half the steps are to the ture was increased from 25 °C (as above) to 30 °C?
right. The smallness of μ allows us to use the expansion Answer: 84 ps

⎛1 ⎞
ln ⎜ ± μ ⎟ = − ln 2 ± 2 μ − 2 μ 2 +
⎝2 ⎠ For a one-dimensional random walk, eqn 81.13 in combi-
μ2
and retain terms through the order in the overall expres- nation with eqn 81.16 implies that the distance walked after
sion for ln P(nλ). The final result, after quite a lot of algebra N = t/τ steps of length λ is
(see Problem 81.10), is
1/2
⎛ 2N ⎞ One Mean distance
2N +1 e −2 N μ 2e −2 N μ 〈 x〉 = λ ⎜
2 2
(81.17a)
⎝ π ⎟⎠
dimension from origin
P (nλ ) = N =
2 (2πN ) 1/2
(2πN )1/2
The corresponding expressions for the mean radius reached in
At this point we recognize that
a random walk in two dimensions (like a molecule migrating
(2N R − N )2 (N R − N L )2 n2 on a flat surface) and in three dimensions (in a solution; see
N μ2 = = =
4N 4N 4N Problem 81.13) are

The net distance from the origin is x = nλ and the number of 〈r 〉 = λN 1/2 Two dimensions Mean distance from origin (81.17b)
steps taken in a time t is N = t/τ, so Nμ 2 = τx 2/4tλ 2. Substitution
of these quantities into the expression for P gives eqn 81.15. 1/2
⎛ 8N ⎞ Three Mean distance
〈r 〉 = λ ⎜
⎝ π ⎟⎠ dimensions from origin (81.17c)

The differences of detail between eqns 81.11 (for one-dimen-
sional diffusion) and 81.15 arise from the fact that in the pre-
sent calculation the particles can migrate in either direction
from the origin. Moreover, they can be found only at discrete
points separated by λ instead of being anywhere on a continu-
ous line. The fact that the two expressions are so similar sug-
gests that diffusion can indeed be interpreted as the outcome of
a large number of steps in random directions.
We can now relate the coefficient D to the step length λ and
the rate at which the jumps occur. Thus, by comparing the two
exponents in eqn 81.11 and eqn 81.15 we can immediately
write down the Einstein–Smoluchowski equation:
Note that the average distance increases as the dimensionality
increases, as the particles have greater freedom to escape than
in one dimension.
The Einstein–Smoluchowski equation is the central connec-
tion between the microscopic details of particle motion and
the macroscopic parameters relating to diffusion (for exam-
ple, the diffusion coefficient and, through the Stokes–Einstein
relation, eqn 80.19b of Topic 80, D = kT/6πηa, the viscosity). It
also brings us back full circle to the properties of the perfect gas
treated in Topic 79. For if we interpret λ/τ as vmean, the mean
speed of the molecules, and interpret λ as a mean free path,
then we can recognize in the Einstein–Smoluchowski equation

www.ebook3000.com

Atkins09819.indb 788 9/11/2013 8:42:14 AM

 81 Diffusion  789

essentially the same expression as we obtained from the kinetic diffusion of a perfect gas is a random walk with an average step
model of gases (eqn 79.7 of Topic 79, D = 13 λvmean ). That is, the size equal to the mean free path.

Checklist of concepts
☐ 1. The thermodynamic force represents the spontaneous
tendency of the molecules to disperse as a consequence
of the Second Law and the hunt for maximum entropy.
☐ 2. The diffusion equation (Fick’s second law; see below)
can be regarded as a mathematical formulation of the
notion that there is a natural tendency for concentra-
tion to become uniform.
☐ 3. Convection is the transport of particles arising from
the motion of a streaming fluid.
☐ 4. An intuitive picture of diffusion is of the particles mov-
ing in a series of small steps and gradually migrating
from their original positions.

Checklist of equations
Property Equation Comment Equation number

Thermodynamic force F = −(∂μ/∂x)T,p Definition 81.1

Fick’s first law J(amount) = −Ddc/dx 81.4

Diffusive flux J = sc 81.5

Drift speed s = DF/RT 81.6

Diffusion equation ∂c/∂t = D∂2c/∂x2 One dimension 81.7

Convective flux J = cv 81.8

Generalized diffusion equation ∂c/∂t = D∂2c/∂x2 − v∂c/∂x One dimension 81.10

Mean displacement 〈x〉 = 2(Dt/π)1/2 One-dimensional diffusion 81.13

Root mean square displacement 〈x2〉1/2 = (2Dt)1/2 One-dimensional diffusion 81.14

P (x , t ) = (2τ / πt )1/2 e − x τ /2t λ

2 2
Probability of displacement One-dimensional random walk 81.15

Einstein–Smoluchowski equation D = λ2/2τ One-dimensional random walk 81.16

Atkins09819.indb 789 9/11/2013 8:42:16 AM

 790 16 Molecular motion
Focus 16 on Molecular motion
Topic 78 The kinetic theory of gases
Discussion questions
78.1 Specify and analyse critically the assumptions that underlie the kinetic
model of gases.
Exercises
78.1(a) Determine the ratios of (a) the mean speeds, (b) the mean translational
kinetic energies of H2 molecules and Hg atoms at 20 °C.
78.1(b) Determine the ratios of (a) the mean speeds, (b) the mean kinetic
energies of He atoms and Hg atoms at 25 °C.
78.2(a) Calculate the root mean square speeds of H2 and O2 molecules at
20 °C.
78.2(b) Calculate the root mean square speeds of CO2 molecules and He atoms
at 20 °C.
78.3(a) Use the Maxwell distribution of speeds to estimate the fraction of N2
molecules at 400 K that have speeds in the range 200 to 210 m s−1.
78.3(b) Use the Maxwell distribution of speeds to estimate the fraction of CO2
molecules at 400 K that have speeds in the range 400 to 405 m s−1.
78.4(a) Calculate the most probable speed, the mean speed, and the mean
relative speed of CO2 molecules in air at 20 °C.
78.4(b) Calculate the most probable speed, the mean speed, and the mean
relative speed of H2 molecules in air at 20 °C.
78.5(a) Assume that air consists of N2 molecules with a collision diameter
of 395 pm. Calculate (a) the mean speed of the molecules, (b) the mean free
path, (c) the collision frequency in air at 1.0 atm and 25 °C.
78.5(b) The best laboratory vacuum pump can generate a vacuum of about
1 nTorr. At 25 °C and assuming that air consists of N2 molecules with a
collision diameter of 395 pm, calculate (a) the mean speed of the molecules,
(b) the mean free path, (c) the collision frequency in the gas.
78.6(a) At what pressure does the mean free path of argon at 20 °C become
comparable to the diameter of a 100 cm3 vessel that contains it? Take
σ = 0.36 nm2.
78.6(b) At what pressure does the mean free path of argon at 20 °C become
comparable to 10 times the diameters of the atoms themselves?
78.7(a) At an altitude of 20 km the temperature is 217 K and the pressure
0.050 atm. What is the mean free path of N2 molecules? (σ = 0.43 nm2).
78.7(b) At an altitude of 15 km the temperature is 217 K and the pressure
12.1 kPa. What is the mean free path of N2 molecules? (σ = 0.43 nm2).
78.8(a) How many collisions does a single Ar atom make in 1.0 s when the
temperature is 25 °C and the pressure is (a) 10 atm, (b) 1.0 atm, (c) 1.0 μatm?
78.8(b) How many collisions per second does an N2 molecule make at an
altitude of 15 km? (See Exercise 78.7b for data.)
www.ebook3000.com
Atkins09819.indb 790
78.2 Provide molecular interpretations for the dependencies of the mean free
path on the temperature, pressure, and size of gas molecules.
78.9(a) A solid surface with dimensions 5.0 mm × 4.0 mm is exposed to argon
gas at 25 Pa and 300 K. How many collisions do the Ar atoms make with this
surface in 100 s?
78.9(b) A solid surface with dimensions 2.0 cm × 10.0 cm is exposed to helium
gas at 120 Pa and 1200 K. How many collisions do the He atoms make with
this surface in 1.0 s?
78.10(a) If 125 cm3 of hydrogen gas effuses through a small hole in 135
seconds, how long will it take the same volume of oxygen gas to effuse under
the same temperature and pressure?
78.10(b) If 175 cm3 of carbon dioxide effuses through a small hole in 255
seconds, how long will it take the same volume of sulfur dioxide to effuse
under the same temperature and pressure?
78.11(a) An effusion cell has a circular hole of diameter 1.50 mm. If the molar
mass of the solid in the cell is 300 g mol−1 and its vapour pressure is 0.735 Pa at
500 K, by how much will the mass of the solid decrease in a period of 1.00 h?
78.11(b) An effusion cell has a circular hole of diameter 1.00 mm. If the molar
mass of the solid in the cell is 250 g mol−1 and its vapour pressure is 0.324 Pa
at 425 K, by how much will the mass of the solid decrease in a week (use
1 week = 7 × 24 h)?
78.12(a) A manometer was connected to a bulb containing a gaseous sample
under slight pressure. The gas was allowed to escape through a small pinhole,
and the time for the manometer reading to drop from 74 cm to 20 cm
was 152 s. When the experiment was repeated using nitrogen (for which
M = 28.02 g mol−1) the same fall took place in 45 s. Calculate the molar mass of
the sample.
78.12(b) A manometer was connected to a bulb containing nitrogen under
slight pressure. The gas was allowed to escape through a small pinhole, and
the time for the manometer reading to drop from 75.1 cm to 32.5 cm was
22.5 s. When the experiment was repeated using a fluorocarbon gas, the same
fall took place in 135.0 s. Calculate the molar mass of the fluorocarbon.
78.13(a) A space vehicle of internal volume 3.0 m3 is struck by a meteor and a
hole of radius 0.10 mm is formed. If the oxygen pressure within the vehicle is
initially 80 kPa and its temperature 298 K, how long will the pressure take to
fall to 70 kPa?
78.13(b) A container of internal volume 22.0 m3 was punctured, and a hole of
radius 0.050 mm was formed. If the nitrogen pressure within the vehicle is
initially 122 kPa and its temperature 293 K, how long will the pressure take to
fall to 105 kPa?
9/11/2013 8:42:17 AM

Problems
78.1 The speed of molecules can also be measured with a rotating slotted-disc
apparatus, which consists of five coaxial 5.0 cm diameter disks separated by
1.0 cm, the slots in their rims being displaced by 2.0° between neighbours.
The relative intensities, I, of the detected beam of Kr atoms for two different
temperatures and at a series of rotation rates were as follows:
ν/Hz 20 40 80 100 120
I (40 K ) 0.846 0.513 0.069 0.015 0.002
I (100 K) 0.592 0.485 0.217 0.119 0.057
Find the distributions of molecular velocities, f(vx), at these temperatures, and
check that they conform to the theoretical prediction for a one-dimensional
system.
78.2 A Knudsen cell was used to determine the vapour pressure of germanium
at 1000 °C. During an interval of 7200 s the mass loss through a hole of radius
0.50 mm amounted to 43 μg. What is the vapour pressure of germanium at
1000 °C? Assume the gas to be monatomic.
78.3 The pressure of a Knudsen cell of volume V in which a vapour is
confined (with no condensed phase to replenish the vapour phase) decays
exponentially with a time constant τ = (2πM/RT)1/2(V/A) (see Problem
78.12 for the derivation of a related expression). How long would it take
the pressure of barium vapour in a cell with V = 100 cm3 and A = 0.10 mm2 at
1300 °C to fall to 1/10 of its initial value?
78.4 The vapour pressure of zinc in the range 250 °C to 419 °C can be
estimated from the expression log(p/Torr) = a − b/T with a = 9.200 and
b = 6947 K. Calculate and plot the beam flux (in Zn atoms per second)
emerging from a hole of radius 0.20 mm as the temperature of the oven
containing solid zinc is raised from 250 °C to 400 °C.
78.5 Start from the Maxwell–Boltzmann distribution and derive an expression
for the most probable speed of a gas of molecules at a temperature T. Go on
to demonstrate the validity of the equipartition conclusion that the average
translational kinetic energy of molecules free to move in three dimensions
is  23 kT.
78.6 In Topic 58 it is established that the heat capacity of a collection of
molecules is proportional to the variance of their energy (the mean square
deviation of the energy from its mean value). Use the Maxwell–Boltzmann
distribution of speeds to calculate the translational contribution to the heat
capacity of a gas by this approach.
78.7 Consider molecules that are confined to move in a plane (a two-
dimensional gas). Calculate the distribution of speeds and determine the
mean speed of the molecules at a temperature T.
78.8 A specially constructed velocity selector accepts a beam of molecules
from an oven at a temperature T but blocks the passage of molecules with a
speed greater than the mean. What is the mean speed of the emerging beam,
relative to the initial value, treated as a one-dimensional problem?
78.9 What, according to the Maxwell–Boltzmann distribution, is the
proportion of gas molecules having (a) more than, (b) less than the root mean
square speed? (c) What are the proportions having speeds greater and smaller
than the mean speed?
78.10 Calculate the fractions of molecules in a gas that have a speed in a range
Δv at the speed nvmp relative to those in the same range at vmp itself? This
calculation can be used to estimate the fraction of very energetic molecules
(which is important for reactions). Evaluate the ratio for n = 3 and n = 4.
Atkins09819.indb 791
Exercises and problems  791
78.11 Derive an expression for 〈vn〉1/n from the Maxwell–Boltzmann
distribution of speeds. You will need Integrals G.7 and G.8 listed in the
Resource section.
78.12 Derive an expression that shows how the pressure of a gas inside an
effusion oven (a heated chamber with a small hole in one wall) varies with
time if the oven is not replenished as the gas escapes. Then show that t1/2,
the time required for the pressure to decrease to half its initial value, is
independent of the initial pressure. Hint: Begin by setting up a differential
equation relating dp/dt to p = NkT/V, and then integrating it.
78.13 Calculate the escape velocity (the minimum initial velocity that will
take an object to infinity) from the surface of a planet of radius R. What is the
value for (a) the Earth, R = 6.37 × 106 m, g = 9.81 m s−2; (b) Mars, R = 3.38 × 106
m, mMars/mEarth = 0.108. At what temperatures do H2, He, and O2 molecules
have mean speeds equal to their escape speeds? What proportion of the
molecules have enough speed to escape when the temperature is (a) 240 K,
(b) 1500 K? Calculations of this kind are very important in considering the
composition of planetary atmospheres.
78.14 The kinetic model of gases is valid when the size of the particles is
negligible compared with their mean free path. It may seem absurd, therefore,
to expect the kinetic theory and, as a consequence, the perfect gas law, to be
applicable to the dense matter of stellar interiors. In the Sun, for instance, the
density is 1.50 times that of liquid water at its centre and comparable to that
of water about half way to its surface. However, we have to realize that the
state of matter is that of a plasma, in which the electrons have been stripped
from the atoms of hydrogen and helium that make up the bulk of the matter
of stars. As a result, the particles making up the plasma have diameters
comparable to those of nuclei, or about 10 fm. Therefore, a mean free path
of only 0.1 pm satisfies the criterion for the validity of the kinetic model and
the perfect gas law. We can therefore use pV = nRT as the equation of state
for the stellar interior. (a) Calculate the pressure half way to the centre of
the Sun, assuming that the interior consists of ionized hydrogen atoms, the
temperature is 3.6 MK, and the mass density is 1.20 g cm−3 (slightly higher
than the density of water). (b) Combine the result from part (a) with the
expression for the pressure from the kinetic model to show that the pressure
of the plasma is related to its kinetic energy density, ρk = Ek/V, the kinetic
energy of the molecules in a region divided by the volume of the region, by
p = 23 ρ k. (c) What is the kinetic energy density half way to the centre of the
Sun? Compare your result with the (translational) kinetic energy density of
the Earth's atmosphere on a warm day (25 °C), 1.5 × 105 J m−3 (corresponding
to 0.15 J cm−3). (d) A star eventually depletes some of the hydrogen in its
core, which contracts and results in higher temperatures. The increased
temperature results in an increase in the rates of nuclear reactions, some of
which result in the formation of heavier nuclei, such as carbon. The outer part
of the star expands and cools to produce a red giant. Assume that halfway
to the centre a red giant has a temperature of 3500 K, is composed primarily
of fully ionized carbon atoms and electrons, and has a mass density of
1200 kg m−3. What is the pressure at this point? (e) If the red giant in part
(d) consisted of neutral carbon atoms, what would be the pressure at the same
point under the same conditions?
78.15 The principal components of the atmosphere of the Earth are diatomic
molecules, which can rotate as well as translate. Given that the translational
kinetic energy density of the atmosphere is 0.15 J cm−3, what is the total kinetic
energy density, including rotation?
9/11/2013 8:42:19 AM
 792 16 Molecular motion

Topic 79 Transport properties of gases

Discussion questions
79.1 Explain how Fick’s first law arises from the concentration gradient of gas 79.3 What might be the effect of molecular interactions on the transport
molecules. properties of a gas?
79.2 Provide molecular interpretations for the dependencies of the diffusion
coefficient and the viscosity on the temperature, pressure, and size of gas
molecules.

Exercises
79.1(a) Calculate the thermal conductivity of argon (CV,m = 12.5 J K−1 mol−1,
σ = 0.36 nm2) at 298 K.
79.1(b) Calculate the thermal conductivity of nitrogen (CV,m = 20.8 J K−1 mol−1,
σ = 0.43 nm2) at 298 K.
79.2(a) Calculate the diffusion constant of argon at 20 °C and (a) 1.00 Pa,
(b) 100 kPa, (c) 10.0 MPa. If a pressure gradient of 1.0 bar m−1 is established
in a pipe, what is the flow of gas due to diffusion?
79.2(b) Calculate the diffusion constant of nitrogen at 20 °C and (a) 100.0 Pa,
(b) 100 kPa, (c) 20.0 MPa. If a pressure gradient of 1.20 bar m−1 is established
in a pipe, what is the flow of gas due to diffusion?
79.3(a) Calculate the flux of energy arising from a temperature gradient of
10.5 K m−1 in a sample of argon in which the mean temperature is 280 K.
79.3(b) Calculate the flux of energy arising from a temperature gradient of
8.5 K m−1 in a sample of hydrogen in which the mean temperature is 290 K.
79.4(a) Use the experimental value of the thermal conductivity of neon (Table
79.1) to estimate the collision cross-section of Ne atoms at 273 K.
79.4(b) Use the experimental value of the thermal conductivity of nitrogen
(Table 79.1) to estimate the collision cross-section of N2 molecules at 298 K.
79.5(a) In a double-glazed window, the panes of glass are separated by 1.0 cm.
What is the rate of transfer of heat by conduction from the warm room
(28 °C) to the cold exterior (−15 °C) through a window of area 1.0 m2? What
power of heater is required to make good the loss of heat?
79.5(b) Two sheets of copper of area 2.00 m2 are separated by 5.00 cm. What is
the rate of transfer of heat by conduction from the warm sheet (70 °C) to the
cold sheet (0 °C)? What is the rate of loss of heat?
79.6(a) Use the experimental value of the coefficient of viscosity for neon
(Table 79.1) to estimate the collision cross-section of Ne atoms at 273 K.
79.6(b) Use the experimental value of the coefficient of viscosity for nitrogen
(Table 79.1) to estimate the collision cross-section of the molecules at 273 K.
79.7(a) Calculate the viscosity of air at (a) 273 K, (b) 298 K, (c) 1000 K. Take
σ ≈ 0.40 nm2. (The experimental values are 173 μP at 273 K, 182 μP at 20 °C,
and 394 μP at 600 °C.)
79.7(b) Calculate the viscosity of benzene vapour at (a) 273 K, (b) 298 K,
(c) 1000 K. Take σ ≈ 0.88 nm2.

Problems
79.1‡ Fenghour, et al. (J. Phys. Chem. Ref. Data 24, 1649 (1995)) have
compiled an extensive table of viscosity coefficients for ammonia in the
liquid and vapour phases. Deduce the effective molecular diameter of NH3
based on each of the following vapour-phase viscosity coefficients: (a)
η = 9.08 × 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 10.0 bar.
79.2 Calculate the ratio of the thermal conductivities of gaseous hydrogen
at 300 K to gaseous hydrogen at 10 K. Be circumspect, and think about the
modes of motion that are thermally active at the two temperatures.
79.3 Interstellar space is quite a different medium than the gaseous
environments we commonly encounter on Earth. For instance, a typical
density of the medium is about 1 atom cm−3 and that atom is typically H;
the effective temperature due to stellar background radiation is about 10 kK.
Estimate the diffusion coefficient and thermal conductivity of H under these
conditions. Comment: Energy is in fact transferred much more effectively by
radiation.

Topic 80 Motion in liquids

Discussion questions
80.1 Discuss the difference between the hydrodynamic radius of an ion and
its ionic radius and explain why a small ion can have a large hydrodynamic
radius.
80.2 Discuss the mechanism of proton conduction in water. How could the
model be tested?
80.3 Why is a proton less mobile in liquid ammonia than in water?

‡ These problems were supplied by Charles Trapp and Carmen Giunta.

www.ebook3000.com

Atkins09819.indb 792 9/11/2013 8:42:19 AM

 Exercises and problems  793

Exercises
80.1(a) The viscosity of water at 20 °C is 1.002 cP and 0.7975 cP at 30 °C. What
is the energy of activation for the transport process?
80.1(b) The viscosity of mercury at 20 °C is 1.554 cP and 1.450 cP at 40 °C.
What is the energy of activation for the transport process?
80.2(a) The mobility of a chloride ion in aqueous solution at 25 °C is
7.91 × 10−8 m2 s−1 V−1. Calculate the molar ionic conductivity.
80.2(b) The mobility of an acetate ion in aqueous solution at 25 °C is
4.24 × 10−8 m2 s−1 V−1. Calculate the molar ionic conductivity.
80.3(a) The mobility of a Rb+ ion in aqueous solution is 7.92 × 10−8 m2 s−1 V−1 at
25 °C. The potential difference between two electrodes placed in the solution
is 25.0 V. If the electrodes are 7.00 mm apart, what is the drift speed of the Rb+
ion?
80.3(b) The mobility of a Li+ ion in aqueous solution is 4.01 × 10−8 m2 s−1 V−1 at
25 °C. The potential difference between two electrodes placed in the solution
is 24.0 V. If the electrodes are 5.0 mm apart, what is the drift speed of the ion?
80.4(a) The limiting molar conductivities of NaI, NaNO3, and AgNO3 are
12.69 mS m2 mol−1, 12.16 mS m2 mol−1, and 13.34 mS m2 mol−1, respectively (all
at 25 °C). What is the limiting molar conductivity of AgI at this temperature?
80.4(b) The limiting molar conductivities of KF, KCH3CO2, and Mg(CH3CO2)2
are 12.89 mS m2 mol−1, 11.44 mS m2 mol−1, and 18.78 mS m2 mol−1, respectively
(all at 25 °C). What is the limiting molar conductivity of MgF2 at this
temperature?
80.5(a) At 25 °C the molar ionic conductivities of Li+, Na+, and K+ are
3.87 mS m2 mol−1, 5.01 mS m2 mol−1, and 7.35 mS m2 mol−1, respectively. What
are their mobilities?
80.5(b) At 25 °C the molar ionic conductivities of F−, Cl−, and Br− are
5.54 mS m2 mol−1, 7.635 mS m2 mol−1, and 7.81 mS m2 mol−1, respectively. What
are their mobilities?
80.6(a) Estimate the effective radius of a sucrose molecule in water at 25 °C
given that its diffusion coefficient is 5.2 × 10−10 m2 s−1 and that the viscosity of
water is 1.00 cP.
80.6(b) Estimate the effective radius of a glycine molecule in water at 25 °C
given that its diffusion coefficient is 1.055 × 10−9 m2 s−1 and that the viscosity of
water is 1.00 cP.

Problems
80.1 The viscosity of benzene varies with temperature as shown in the Verify that the molar conductivity follows the Kohlrausch law and find the
following table. Use the data to infer the activation energy for viscosity (the limiting molar conductivity. Determine the coefficient K. Use the value of K
parameter Ea in eqn 80.2). (which should depend only on the nature, not the identity, of the ions) and
the information that λ(Na+) = 5.01 mS m2 mol−1 and λ(I−) = 7.68 mS m2 mol−1
to predict (a) the molar conductivity, (b) the conductivity, (c) the resistance it
θ/°C 10 20 30 40 50 60 70 would show in the cell of 0.010 mol dm−3 NaI(aq) at 25 °C.
η/cP 0.758 0.652 0.564 0.503 0.442 0.392 0.358
80.6 What are the drift speeds of Li+, Na+, and K+ in water when a potential
80.2 An empirical expression that reproduces the viscosity of water in the difference of 100 V is applied across a 5.00 cm conductivity cell? How
range 20–100 °C is long would it take an ion to move from one electrode to the other? In
conductivity measurements it is normal to use alternating current: what are
η 1.3272(20 −θ / °C) − 0.001053(20 −θ /°C)2 the displacements of the ions in (a) centimetres, (b) solvent diameters (about
log =
η 20 θ / °C +105 300 pm) during a half-cycle of 2.0 kHz applied potential difference?

where η20 is the viscosity at 20 °C. Explore (by using mathematical software) 80.7‡ Bakale, et al. (J. Phys. Chem. 100, 12477 (1996)) measured the mobility
− ions in a variety of nonpolar solvents. In cyclohexane at
of singly charged C 60
the possibility of fitting an exponential curve to this expression and hence −
identify an activation energy for the viscosity. This approach is taken further 22 °C, the mobility is 1.1 cm2 V−1 s−1. Estimate the effective radius of the C 60
in Problem 80.10. ion. The viscosity of the solvent is 0.93 × 10−3 kg m−1 s−1. Suggest a reason why
there is a substantial difference between this number and the van der Waals
80.3 The conductivity of aqueous ammonium chloride at a series of radius of neutral C60.
concentrations is listed in the following table. Deduce the molar conductivity
and determine the parameters that occur in Kohlrausch’s law. 80.8 Estimate the diffusion coefficients and the effective hydrodynamic radii
of the alkali metal cations in water from their mobilities at 25 °C. Estimate the
c/(mol dm−3) 1.334 1.432 1.529 1.672 1.725 approximate number of water molecules that are dragged along by the cations.
κ/(mS cm−1) 131 139 147 156 164 Ionic radii are given Table 38.2.

80.4 Conductivities are often measured by comparing the resistance of a cell 80.9 Nuclear magnetic resonance can be used to determine the mobility
filled with the sample to its resistance when filled with some standard solution, of molecules in liquids. A set of measurements on methane in carbon
such as aqueous potassium chloride. The conductivity of water is 76 mS m−1 tetrachloride showed that its diffusion coefficient is 2.05 × 10−9 m2 s−1 at 0 °C
at 25 °C and the conductivity of 0.100 mol dm−3 KCl(aq) is 1.1639 S m−1. A and 2.89 × 10−9 m2 s−1 at 25 °C. Deduce what information you can about the
cell had a resistance of 33.21 Ω when filled with 0.100 mol dm−3 KCl(aq) and mobility of methane in carbon tetrachloride.
300.0 Ω when filled with 0.100 mol dm−3 CH3COOH(aq). What is the molar 80.10 In Topic 85 it is shown that a general expression for the activation
conductivity of acetic acid at that concentration and temperature? energy of a chemical reaction is Ea = RT2(d ln k/dT). Confirm that the same
80.5 The resistances of a series of aqueous NaCl solutions, formed by successive expression may be used to extract the activation energy from eqn 80.2
dilution of a sample, were measured in a cell with cell constant (the constant C for the viscosity and then apply the expression to deduce the temperature
in the relation κ = C/R) equal to 0.2063 cm−1. The following values were found: dependence of the activation energy when the viscosity of water is given
by the empirical expression in Problem 80.2. Plot this activation energy
c/(mol dm−3) 0.00050 0.0010 0.0050 0.010 0.020 0.050 as a function of temperature. Suggest an explanation of the temperature
dependence of Ea.
R/Ω 3314 1669 342.1 174.1 89.08 37.14

Atkins09819.indb 793 9/11/2013 8:42:21 AM