Course M6 Lecture 6
28/1/2004
Course M6 Lecture 6
28/1/2004
Diffusion in polymers
Permeation and diffusion of
solvents into amorphous
polymers
Dr James Elliott
6.1 Introduction
z
z
z
In the previous lecture, we talked about the dynamics of
polymer chains in the melt (polymer only) and dilute
solution (polymer minority component)
But, when the polymer is the majority component, what
can we say about the motion of the solvent?
In particular, we are interested in the degree to which the
solvent permeates the polymer, how much is absorbed
and how fast it can move through the sample
Restrict attention to mainly amorphous structures
Rubbery polymers (extensively studied since Graham 1866)
Glassy polymers: more complicated
Semi-crystalline polymers: crystallites embedded in matrix
Course M6 Lecture 6
28/1/2004
6.2 Motivation for studying diffusion in polymers
z
Important for many design applications
Application
Penetrant
Design goal
packaging
additives
gas separation
analytical chemistry
monomer removal
polymer electrolytes
drug implants
biosensors
gas, moisture
plasticizers, dyes
gases
ions
unreacted monomer
ions
pharmaceuticals
biomolecules
high barrier
high barrier
selectivity
high selectivity
low barrier
ionic conductivity
controlled release
high selectivity
6.3 Macroscopic manifestations
z
Sorption isotherm
Diffusivity as a function of temperature
Solubility and diffusivity of various gases in natural rubber
Permeability of various gases in natural rubber
Permeability of various gases in different polymers
Course M6 Lecture 6
28/1/2004
6.4 Sorption isotherm for CO2 in PET at 25C
z
Typical curve shape for a glassy polymer
6.5 Solubility and diffusivity of gases in rubber
z
z
Diffusivity drops with penetrant size (Stokes)
Solubility increases with penetrant size (condensation)
Course M6 Lecture 6
28/1/2004
6.6 Permeability of gases in rubber
z
Minimum due to combined low solubility and diffusivity
6.7 Permeability of gases in various polymers
z
Similar trends for a wide range of polymers
Course M6 Lecture 6
28/1/2004
6.8.1 Revision of diffusion theory
z
Linear flux-force relationship
J = A
dX
dx
where A is some phenomenological coefficient and X is a
potential
When X is concentration, gives Ficks first law
J = D
dC
dx
where D is the diffusion coefficient
6.8.2 Revision of diffusion theory
z
Concentration of penetrant proportional to pressure on
polymer (Henrys law)
C = kp
where k is a solubility coefficient and p is the vapour
pressure of penetrant
Steady state permeability = flux/pressure gradient:
P = Dk = Diffusivity Solubility
z
z
In an ideal system, solubility and diffusivity are constant
When they are concentration dependent, the behaviour is
usually called anomalous
Course M6 Lecture 6
28/1/2004
6.9 Solution-diffusion model
z
Treat permeation as a three stage process
3
2
1. Adsorption onto polymer surface
2. Diffusion through bulk polymer
3. Desorption into external phase
6.10.1 Sorption
z
z
Rubbery polymers ideal behaviour
Glassy polymers anomalous behaviour via the dual
mode sorption mechanisms
Adsoprtion in microvoids (holes), ch
Ordinary dissolution, cd
ch = kp + ch bp/ (1 + bp)
cd
Henry
b is a hole affinity constant
ch
Langmuir
ch
p
Course M6 Lecture 6
28/1/2004
6.10.2 Sorption
z
Reason for the microvoids is that free volume is frozen
into structure at the glass transition
Specific
Volume
liquid
glass
Vexc
Tg
z
z
z
Temperature
Solubility given by enthalpy of dissolution Hs
Above Tg, there is endothermic chain separation
Below Tg, exothermic heat of absorption
6.11 Macroscopic diffusion
z
Ficks second law
dc
dJ
d 2c
=
= D 2
dt
dx
dx
z
Activated process Arrhenius behaviour:
D = D0 exp( Ed / RT )
z
z
z
z
Effect of glass transition is changes in slope
Ed(rubber) > Ed(glass)
Rubber: penetrant must create channels, activate chains
Glass: penetrant moves through holes, local activation
Course M6 Lecture 6
28/1/2004
6.12 Arrhenius plots for O2 and Ar in PVA
z
Shows changes in D around Tg
6.13 Free volume theory of diffusion
z
Molecules can only diffuse through free volume
Probability of finding a hole ~ exp(B / vf), where B
depends on the volume needed by the penetrant and vf is
the fractional free volume
Diffusion depends on penetrant mobility:
D = RTAf exp( B / vf )
Course M6 Lecture 6
28/1/2004
6.14 Microscopic picture of diffusion
z
Related to the autocorrelation function of penetrant
currents. If different particles are uncorrelated (low
concentration, short-ranged interactions):
D = 13 v(t )v(0) dt
z
For long times, equivalent to
D = 61t R(t ) - R(0)
z
z
(Green - Kubo)
(Einstein)
Assumes a random walk: no memory
Correlated motions give rise to anomalous diffusion
(restricted motion or conformational changes due to the
environment)
6.15 Time lag measurement of diffusion
z
z
z
Starting with a membrane of thickness l free of penetrant
Apply a penetrant pressure on one side, and measure
the flow Qt
Dci
l2
Q
=
t
In the steady state: t
6D
Qt / lc
transient
state
steady
state
l2
=
6D
Course M6 Lecture 6
28/1/2004
6.16 Pressure decay measurement of diffusion
z
z
Polymer sample is put in a closed, fixed volume
After evacuation, penetrant gas is applied at a certain
pressure. Pressure in the chamber then decreases due
to mass uptake Mt by the polymer, according to
Mt / M
1.0
Mt
4
=
M
Dt
l2
t1/2
6.17.1 Computer simulation of diffusion
z
z
z
z
Construct an atomistic scale model of the bulk material
including the gas molecules
Simulate the behaviour of the system for sufficient time to
observe diffusion
Analyse to obtain the mean square displacement (MSD)
of the gas molecules
The self diffusion coefficient (equivalent to the diffusivity
for low concentrations of solute) can be obtained from
curve of MSD versus time
Example: diffusion of nitrogen gas through
polydimethylsiloxane (PDMS)
10
Course M6 Lecture 6
28/1/2004
6.17.2 Building a model
z
Take some N2 and a PDMS chain
6.17.3 Building a model
z
Pack them (RIS) into a box with periodic boundaries to
an appropriate density
11
Course M6 Lecture 6
28/1/2004
6.17.4 Calculating mean-squared displacement
z
Evolve system in time, and observe penetrant motion
6.17.5 Calculating mean-squared displacement
Looking at motion of a single penetrant molecule
projected onto a plane, can see more clearly
25
20
15
y ()
10
5
0
-5
0
10
15
20
25
x ()
12
Course M6 Lecture 6
28/1/2004
6.17.6 Calculating mean-squared displacement
z
M ean-Squared-Displacement (2)
MSD plot confirms the jump diffusion mechanism
Average over many molecules and long times to obtain D
from Einstein formula
3500
3000
2500
2000
1500
1000
500
0
100
200
300
400
Time (ps)
Summary of M6 Part 1
z
z
z
z
In these last six lectures, we have discussed a number of
key issues in polymer physics
I will try to summarise the most important concept of each
lecture in one bullet point!!
Here goes
Although a freely jointed chain of uncorrelated segments
is sufficient to differentiate between polymers and simple
molecules, correct account of the couplings between the
torsional degrees of freedom is required to distinguish
between conformations of chemically dissimilar polymers
Polymer coils in the melt or at their -point adopt ideal
conformations, but in a good solvent they are expanded
13
Course M6 Lecture 6
28/1/2004
Summary of M6 Part 1
z
z
In general, different polymers are immiscible unless there
are strongly favourable enthalpic interactions
The phase diagram and morphologies of a binary diblock
copolymer are governed by balance between interfacial
forces and chain entropy, but allow mixing of chemically
dissimilar polymers on a microscopic scale
Diffusion of polymers is governed by their molecular
weight and concentration, with abrupt changes in scaling
behaviour at critical values of these
Permeability of polymers by solvents can be explained on
the basis of their diffusivity and solubility, and the
structure of the polymer matrix
Endings, and new beginnings
14
