550

Lecture #16 of 26
 551

Liquid-Junction Potentials
Chapter 2
 552

Q: What’s in this set of lectures?

A: B&F Chapter 2 main concepts:

● “Section 2.1”: Salt; Activity; Underpotential deposition

● Section 2.3: Transference numbers; Liquid-junction

potentials

● Sections 2.2 & 2.4: Donnan potentials; Membrane potentials;

pH meter; Ion-selective electrodes
Now on to two general liquid junctions that we care about (the most)… 553

an SCE an ISE (for nitrate ions)

… liquid junctions: 554

when two ionic solutions are separated across an interface that

prevents bulk mixing of the ions, but has ionic permeability, a potential
(drop) develops called the liquid junction potential.

same salt; one ion in common;

everything else
different conc. same conc.
Bard & Faulkner, 2nd Ed., Wiley, 2001, Figure 2.3.2
… example “1”: ● starting at the side with larger ion concentration 555
● the ion the with larger mobility will impart its
charge to the opposite side of the junction

… conceptually, let’s think about a condition in the

limit where tH+ → 1 (say tH+ ≈ 0.9)…
… as H+ diffuses down its concentration gradient,
an electrostatic force is exerted on Cl– to pull it
along (at a larger flux) while at the same time
slowing down transport of H+
… this happens until ti–effective = 0.5 for both H+
and Cl–, and at which time the system has attained
steady-state mass transport and has generated a
maximum liquid-junction potential.

… FYI, in semiconductor physics this same process

results in a Dember potential… and the transport
same salt; process is called ambipolar diffusion
different conc.
Bard & Faulkner, 2nd Ed., Wiley, 2001, Figure 2.3.2
Recall that transport number, ti, (or transference number) is based on … 556

units: cm2/(s V)
(Stokes' law)

… and that the ionic conductivity, κ or σ, is defined as…

Siemens

units: S/cm or 1/(Ω cm)

… so ti is the fraction of the solution conductivity attributable to ion "i"

The Kohlrausch law (empirical) and Physicist
Debye–Hückel–Onsager equation
(theoretical) predict that the molar
conductivity is proportional to the
square root of the salt concentration
Friedrich Wilhelm Georg Kohlrausch
Λα (C)1/2 (1840–1910)
from Wiki
Recall that transport number, ti, (or transference number) is based on … 557

units: cm2/(s V)
(Stokes' law)

… and that the ionic conductivity, κ or σ, is defined as…

Siemens

units: S/cm or 1/(Ω cm)

… so ti is the fraction of the solution conductivity attributable to ion "i"

The Kohlrausch law (empirical) and P-Chemist & Physicist
Debye–Hückel–Onsager equation
(theoretical) predict that the molar
conductivity is proportional to the
square root of the salt concentration
Lars Onsager
Λ α (C)1/2 from Wiki (1903–1976)
 558
"equivalent"
molar (ionic)
conductivity

/
/
/
/
/
/
/
/
=F
/
/
/
/ … and recall the
/ Einstein–
/ Smoluchowski
/ equation to
/
/ calculate Di,
𝒛𝒊 𝑭𝑫𝒊
𝝁𝒊 =
𝑹𝑻
… example “2”: ● compare dissimilar ions (cations or anions) 559
● the ion with the larger mobility will impart its
charge to the opposite side of the junction

… the sign of the liquid-junction potential is

obvious for Types 1 and 2 (but not Type 3)
based on the mobilities of the individual ions…

… and so when in doubt, think logically about

the sign of the potential to verify answers…
Cl–
… and yes, Cl– will migrate/drift based on the
electric potential formed by cation transport

one ion in common;

same conc.
Bard & Faulkner, 2nd Ed., Wiley, 2001, Figure 2.3.2
560
we use ti values, which are based on kinetic transport to determine 561
the liquid-junction potential (for derivations, see B&F, pp. 70 – 72)…

Type 1 Type 2 Type 3

same salt; same cation or anion; no common ions,

different concentrations different counter ion; and/or one common
same concentration ion; different concs
we use ti values, which are based on kinetic transport to determine 562
the liquid-junction potential (for derivations, see B&F, pp. 70 – 72)…
(α)
Type 1 … use the
(β) activity of the
entire salt

Type 2 … use the conductivity due to

all ions, even the common one
(with a few assumptions, pg. 72)
… sign depends on the charge of the dissimilar ion:
(+) when cations are dissimilar, and (–) when anions are dissimilar

Type 3

the Henderson Eq. (with a few assumptions, pg. 72)

… as written, these equations calculate Ej at β vs α

example: B&F Problem 2.14d 563

Calculate Ej for NaNO3 (0.10 M) | NaOH (0.10M)

α|β
1. What Type? Type 2

2. Polarity?
564
example: B&F Problem 2.14d 565

Calculate Ej for NaNO3 (0.10 M) | NaOH (0.10M)

α|β
1. What Type? Type 2

2. Polarity? Polarity should be clear… compare mobilities… µOH– is larger…

… so NaNO3 side will be (–)… and so β vs α will be (+)

3. Calculate it:
… (–) due to
- anions moving

as predicted, a (+) LJ
µNO−3 + µNa+ potential correlates
𝐸𝑗 = −0.05916 log with the compartment
µOH− + µNa+ in the denominator, β,
vs α
7.404 + 5.913
𝐸𝑗 = −0.05916 log = 0.0176 V = +17.6 mV
20.5 + 5.913
… a rather large Ej (at β vs α)!
 566

… so, why do trained

electrochemists prefer to
use saturated KCl (or KNO3)
as the salt to fill reference
electrodes?

… similar mobilities and

thus, similar ti’s and thus,…

… vanishingly small LJ
potentials!
Last point: clarifying sign conventions in B&F so that this is crystal clear… 567

Type 3:

Type “2”:

where Λ (the equivalent conductivity) is defined as follows:

… and where Ceq is the concentration of positive or negative charges

associated with a particular salt in solution, and so rearranging 2.3.40…

, where

… and by comparing this to the form of our Type 2 equation, one sees…
Last point: clarifying sign conventions in B&F so that this is crystal clear… 568

Type 3:

Type “2”:

Type 2: Note: α and β are switched…

… they are off by a factor of (-1)… let’s look at that pre-factor…

… for a specific example:
0.1M HCl (α) | 0.1M KCl (β)
Last point: clarifying sign conventions in B&F so that this is crystal clear… 569

Type 3:

… so this means that the Lewis–Sargent

relation should have a (–) in front of it when
net cations diffuse based on our convention…
so switch the sign, as ∓… or use our equation
… they are off by a factor of (-1)… let’s look at that pre-factor…
… for a specific example:
0.1M HCl (α) | 0.1M KCl (β)
Last point: clarifying sign conventions in B&F so that this is crystal clear… 570

Type 3:

Type “2”:

… since we know that the β side will be (+) in the previous case,
this really means the Lewis–Sargent relation should have a (–) sign
in front of it when net cations diffuse…

… if we’re sticking with our convention that the potential is the β

(product/reduced phase) versus the α (reactant/oxidized phase)…
 571

… anyway…
Donnan potential: A special liquid-junction potential due to fixed charges 572
… here are two systems in which Donnan potentials play a prominent role:
an ionomer film a cell

semipermeable
membrane
Nafion membrane impermeable to
charged macromolecules
http://www.futuremorf.com/
http://www.nafion.mysite.com/ http://www.williamsclass.com/
… consider this model which applies to both scenarios… 573

a film of poly(styrene sulfonate)

CNaCl

R–
Na+
m m s s
cNa+ cCl- cNa+ cCl-
Because differences in electrochemical potential (𝜇ҧ𝑖𝑜 ) – think free energy 574
– drive net mass transport (of unstirred solutions), mobile Na+ and Cl–
partition between the membrane and the solution in compliance with
their electrochemical potentials:
𝜇𝑖𝑜,𝑚 + 𝑅𝑇 ln γ𝑚
𝑖 + 𝑅𝑇 ln 𝑐𝑖
𝑚
+ 𝑧𝑖 𝐹𝜙 𝑚 = 𝜇 𝑜,𝑠 + 𝑅𝑇 ln γ 𝑠 + 𝑅𝑇 ln 𝑐 𝑠 + 𝑧 𝐹𝜙 𝑠
𝑖 𝑖 𝑖 𝑖

m s
(for ion "i"… its electrochemical potential in the membrane… is the same as in
solution… this is the definition of something that has equilibrated!)
Because differences in electrochemical potential (𝜇ҧ𝑖𝑜 ) – think free energy 575
– drive net mass transport (of unstirred solutions), mobile Na+ and Cl–
partition between the membrane and the solution in compliance with
their electrochemical potentials:
𝜇𝑖𝑜,𝑚 + 𝑅𝑇 ln γ𝑚
𝑖 + 𝑅𝑇 ln 𝑐𝑖
𝑚
+ 𝑧𝑖 𝐹𝜙 𝑚 = 𝜇 𝑜,𝑠 + 𝑅𝑇 ln γ 𝑠 + 𝑅𝑇 ln 𝑐 𝑠 + 𝑧 𝐹𝜙 𝑠
𝑖 𝑖 𝑖 𝑖

… Assuming that standard state chemical potentials (𝜇𝑖𝑜 ) are the same inside and
outside of the membrane, we can easily solve for the ("Galvani" / inner) electric
potential difference, ϕm – ϕs
… which is exactly what was required to calculate liquid-junction potentials!
𝑠 𝑠
𝑚 𝑠
𝑅𝑇 γ𝑖 𝑐𝑖
𝜙 −𝜙 = ln 𝑚 𝑚 = 𝐸Donnan
𝑧𝑖 𝐹 γ𝑖 𝑐𝑖

… so we can express EDonnan, an equilibrium electric potential difference, in terms

of any ion that has access to both the membrane and the solution:
𝑠 𝑠
𝑅𝑇 𝑎Na + 𝑅𝑇 𝑎Cl−
𝐸Donnan = ln 𝑚 = ln 𝑚
1 𝐹 𝑎Na+ −1 𝐹 𝑎Cl−
 576
𝑠 𝑠
𝑅𝑇 𝑎Na + 𝑅𝑇 𝑎Cl −
𝐸Donnan = ln 𝑚 =− ln 𝑚
𝐹 𝑎Na+ 𝐹 𝑎Cl−

Aside #1: Recall Type 1 case of LJ potential… but now with t– = 0…

(α)
(β)

𝑅𝑇 𝑎1 α
𝐸j = ln = 𝐸Donnan Na+
𝐹 𝑎1 β
… with β being the membrane

Aside #2: This is what B&F writes for this (Donnan) potential…
… Check!

Eqn. 2.4.2
 577
𝑠 𝑠
𝑅𝑇 𝑎Na + 𝑅𝑇 𝑎Cl −
𝐸Donnan = ln 𝑚 =− ln 𝑚
𝐹 𝑎Na+ 𝐹 𝑎Cl−

Anyway… now divide both sides by RT/F and invert the argument of the
“ln()” on the right to eliminate the negative sign, and we have...

… or…
s s m m
𝑎Na+ 𝑎Cl− = 𝑎Na+ 𝑎Cl −
… recall the scenario we are analyzing… 578
… with R– representing the fixed charges…
a film of poly(styrene sulfonate)

CNaCl

R–
Na+
m m s s
cNa+ cCl- cNa+ cCl-
 579
s s m m
𝑎Na 𝑎
+ Cl − = 𝑎 𝑎
Na+ Cl−

… if these are dilute electrolytes, we can neglect activity coefficients…

s s m m
𝑐Na+ 𝑐Cl− = 𝑐Na+ 𝑐Cl−

now, there is an additional constraint: the bulk of the solution and the bulk
of the membrane must be electrically neutral:

s s m m m
𝑐Na+ = 𝑐Cl − 𝑐Na+ = 𝑐Cl − + 𝑐R−

… an equation quadratic in cmCl– is obtained as follows…

… an equation quadratic in cmCl– is obtained as follows… 580

s s m m
𝑐Na+ 𝑐Cl− = 𝑐Na+ 𝑐Cl−
s s because in solution, cNa+ = cCl– for
𝑐Na + = 𝑐Cl − goodness sakes!

s 2 m 2 m m
𝑐Cl− = 𝑐Cl− + 𝑐Cl − 𝑐R−
… an equation quadratic in cmCl– is obtained as follows… 581

s s m m
𝑐Na+ 𝑐Cl− = 𝑐Na+ 𝑐Cl−
s s m m m
𝑐Na + = 𝑐Cl− 𝑐Na+ = 𝑐Cl− + 𝑐R−

s 2 m 2 m
𝑐Cl− = 𝑐Cl− + 𝑐Rm− 𝑐Cl −

m 2 m s 2
0= 𝑐Cl− + 𝑐Rm− 𝑐Cl − − 𝑐Cl−
… use the quadratic formula to solve for cmCl– and one gets…

𝑚 2 2
−𝑐R𝑚− + 𝑐R− +4 𝑠
𝑐Cl− 𝑐R𝑚− 𝑠
𝑐Cl−
2
𝑚
𝑐Cl− = = 1+4 𝑅 −1
2 2 𝑐Cl−
 582

𝑚 2 2
−𝑐R𝑚− + 𝑐R− +4 𝑠
𝑐Cl− 𝑐R𝑚− 𝑠
𝑐Cl−
2
𝑚
𝑐Cl− = = 1+4 𝑚 −1
2 2 𝑐R−

𝑠
if 𝑐Cl− ≪ 𝑐R𝑚− (which is the typical case of interest), then…
𝑠 2 𝑠 2
𝑐Cl− 𝑐Cl−
1+4 ≈1+2
𝑐R𝑚− 𝑐R𝑚−
(Taylor/Maclaurin series expansion to the first 3 (or 4) terms)
 583

𝑚 2 2
−𝑐R𝑚− + 𝑐R− +4 𝑠
𝑐Cl− 𝑐R𝑚− 𝑠
𝑐Cl−
2
𝑚
𝑐Cl− = = 1+4 𝑚 −1
2 2 𝑐R−

𝑠
if 𝑐Cl− ≪ 𝑐R𝑚− (which is the typical case of interest), then…
𝑠 2 𝑠 2
𝑐Cl− 𝑐Cl−
1+4 ≈1+2
𝑐R𝑚− 𝑐R𝑚−
𝑠 2
𝑚 𝑐R𝑚− 𝑐Cl−
𝑐Cl− = 1+2 𝑚 −1 =
2 𝑐R−

… fixed charge sites are responsible for the electrostatic exclusion of mobile
“like” charges (co-ions) from a membrane, cell, etc. This is Donnan Exclusion.
… the larger is CR–m, the smaller is CCl–m
… so how excluded is excluded? 584

𝑠 𝑚
… is 𝑐Cl − ≪ 𝑐R− a reasonable assumption? What is CR–m?

… well, for Nafion 117, the sulfonate concentration is 1.13 M…

… for CR61 AZL from Ionics, the sulfonate concentration is 1.6 M…

so, as an example, if CCl–s = 0.1 M…

0.1 2
= = 0.01 M
1.0
… an order of magnitude lower
… but what if CCl–s is also large (e.g. 1 M)? than CCl–s… rather excluded!
… No more Donnan exclusion!
Source: Torben Smith Sørensen, Surface Chemistry and Electrochemistry
of Membranes, CRC Press, 1999 ISBN 0824719220, 9780824719227