The Structural Rheology of Long Chain

Branched Polymers:
Measuring and Modelling
Macromolecules in Flow

Tom McLeish
University of Durham

BASF, Basell, Dow, Ineos,

DSM, ICI, Lucite, Mitsubishi
TU Eindhoven, UCL, Reading
 The Team
•  Leeds Oliver Harlen, Daniel Read, David Groves, Alan Duckett, John
Embery, Chinmay Das, Harley Klein, Dietmar Auhl, Ahamadi Malidi,
Michael Kapnistos, Peter Hine, Peter Jimack, Nat Inkson, Rosen
Tenchev, Mark Walkley, David Hoyle, Suneel Kunanameni
•  Bradford Phil Coates, Tim Gough, Mike Martyn, Rob Spares
•  Durham Lian Hutchins, Nigel Clarke, Eduardo de Luca, Jonathan
Dodds, Solomon Kumani
•  Sheffield Tony Ryan, Ellen Heeley, Patrick Fairclough, Ron Young,
Christine Fernyhough, Sasha Mykhaylyk
•  Cambridge Malcolm Mackley, Karen Lee, Ashish Lele, Mark Collis,
David Hassell, Tim Lord, Lino Scelsi, Simon Butler
•  Oxford Paul Buckley, Junjie Wu, David de Foccatis, Huaxuang Li
•  UCL Helen Wilson, Mehmet Sahin, Tim Reis
•  Reading Alexei Likhtman, Jorge Ramirez, Bart Vorslaars
•  Eindhoven Han Meijer, Gerrit Peters, Rudi Steenbakkers
•  Nottingham Richard Graham
•  Queensland Tim Nicholson
 Polymer Processing in the 21st century?

Reaction Chemistry

Molecular shape

Melt Rheology

“Reversing the design arrow”

Erik Wassner, BASF “Good processing”
 What are the Questions?

•  Can we understand the physics?

•  Can we predict macroscopic flows?
•  Can we engineer molecules for PPP?
•  Can we do this at the bench-top scale?
•  Can we do this at the industrial scale?
Project Idea 1:
Industrial LCB and the “Buffer Zone”
THEORY MODEL MATERIALS INDUSTRIAL RESINS
Project Idea 2:REPTATE Select view
(data
representation)
Data
manipulation

Modules
Data files

Select theory
Data and
theory plots
Theory
controls

Saved
parameters

Theory
Fitting log parameters
window
Project Idea 3:
Lagrangian viscoelastic CFD tool: flowSolve
 Chain motions in the Tube Model

•  Linear polymers: •  Branched polymers: arm

reptation fluctuation
Update on Tube Model physics:

Reptation + Contour Length Fluctuation +

Constraint Release

Two parameters: G0, τe(T)

Detailed Chain Formulation
(GLAMM model)
Graham, Likhtman, Milner, TCBM, J. Rheol, 47, 1171-1200 (2003).
s

R(s)

Reptation +CLF flow CR

retraction
Polyisoprene samples
Pierre Chambon (Sheffield)
Poly(cis-1,4-isoprene)
Sample Mw/Mn Mw Z=Mw/
[-] [kg/mol]a Me
[-]b
PI 2k 1.07 2.4 0.5

PI 4k 1.05 5.2 1.1

PI 8k 1.04 9.2 1.9

PI 14k 1.04 13.6 3.0

PI 20k 1.04 23.5 4.8

PI 30k 1.04 33.6 6.9

PI 40K 1.04 46.2 9.1

PI 90K 1.03 94.4 19.4

PI 200K 1.04 223.2 46.4

a Determined by GPC (triple detector) in THF at

30°C.
b Me = 4.86 kg/mol, determined from linear theory,

(Likhtman and McLeish 2002).

3,4 units per chain: 5% - 7%

Linear shear rheology and predictions
D. Auhl et al., Macromolecules, 2009
µPP2 software tool: RepTate

Lines are predictions from linear theory Tref. = 25 °C

(Likhtman & McLeish 2002)
Model Parameters
from linear theory:
(Likhtman & McLeish 2002)

τe (25°C) = 0.003 s
Ge (25°C) = 0.569 MPa
log G‘, G‘‘ [Pa]

Me = 4.86 kg/mol
cv = 0.1

PI-200k PI-90k PI-30k PI-14k PI-4k

log ω [s-1]
Neutron Spin Echo

Wischnewski et al., PRL (2002)

- Jülich link

Without CLF

0 100 200 ns
Non-linear Rheology
(GLAMM + RoliePoly models)

Strain-hardening from linear melts when WiR>1

Rheology from Dietmar Auhl, PI melts Pierre Chambon

Polydispersity:-Bimodal Blends
(GLAMM + RoliePoly models)

Relaxation of fat tube stretch via motion

along thin tube (Daniel Read)

•  Effective stretch relaxation time:

•  Above argument works provided
•  It isn’t faster to relax via CR of thin tube along fat tube
•  Thin tube has time to equilibrate locally in fat tube
•  In either case, expect constant effective stretch relaxation time on
further dilution

Auhl et al. Phys. Rev. Lett. 103, 136001 (2009).

Measure stretch relaxation times using extensional rheology

•  Series of 480k / 33k polyisoprene blends, measured in extension

at range of rates (SER rheometer)

Qualitatively,
hardening at
lower rates at
greater dilution!
Fit data using a
multimode
“Rolie-Poly”
model – allows
extraction of
effective stretch
relaxation time.

Phys. Rev. Lett. 103, 136001 (2009).

Putting it all together: SANS in flow
Model World Real World
 The Recirculating SANS cell
(Bradford/Durham)

T. Gough, J. Bent, R. Richards, N. Clarke, E. de Luca, P. Coates,

250k monodisperse PS; Weo=13, Wes=1
J. Bent et al, Science, 301, 1691-1695 (2003).

D
E

B
A
250k monodisperse PS; Weo=13, Wes=1
J. Bent et al, Science, 301, 1691-1695 (2003).

J. Bent et al, Science, 301, 1691-1695 (2003).

250k monodisperse PS; Weo=13, Wes=1
J. Bent et al, Science, 301, 1691-1695 (2003).
250k monodisperse PS; Weo=13, Wes=1
J. Bent et al, Science, 301, 1691-1695 (2003).
250k monodisperse PS; Weo=13, Wes=1
J. Bent et al, Science, 301, 1691-1695 (2003).
250k monodisperse PS; Weo=13, Wes=1
J. Bent et al, Science, 301, 1691-1695 (2003).
250k monodisperse PS; Weo=13, Wes=1
J. Bent et al, Science, 301, 1691-1695 (2003).
250k monodisperse PS; Weo=13, Wes=1
J. Bent et al, Science, 301, 1691-1695 (2003).
Decay of Rg and Stress

SANS Birefringence

Different decay times/lengths

 13

SANS total flow 6

mapping
12 11
4 5

10 9

2 3

8
1
 Macromolecules, 43, 1539-1542 (2010)
SANS near corner flow
250K PS linear Soft Matter, 5, 4426-4432 (2009).

Eduardo de Luca, Kamakshi Jagannathan, Richard Graham, Harley Klein, Nigel Clarke
 Higher-Order Branched Polymers

Branched polymers relax

from outside in – e.g. H-polymer

First, the arms relax by star-like

breathing modes

Then, the backbone relaxes by

“reptation” – but with friction
concentrated at the ends of
the chain
Check BPW on H-polymers: transient rheology

BPW limit for q=2

Check BPW on H-polymers: SANS

Synthesis: J. Allgaier, Juelich

Ma=25k

Mb=57k (deuterated)
λ=2
FlowSANS: Combs in Linear melt
McLeish et al. Soft Matter, 2009, in press
Daniel Read and TCBM, PRL, 2007

Flow

Linear Melt
Linear rheology of arbitrarily branched polymers
Relaxation of a branched polymer

Occurs from the outside of the polymer

towards the inside
Linear rheology of arbitrarily branched polymers
Relaxation of a branched polymer

Occurs from the outside of the polymer

towards the inside
Linear rheology of arbitrarily branched polymers
Relaxation of a branched polymer

Sometimes side arms relax – relaxation

cannot proceed further until the main arm
“catches up”.

Side arms give extra “friction”

Linear rheology of arbitrarily branched polymers
Relaxation of a branched polymer

Sometimes side arms relax – relaxation

cannot proceed further until the main arm
“catches up”.

Side arms give extra “friction”

Linear rheology of arbitrarily branched polymers
Relaxation of a branched polymer

Eventually there is an effectively linear

section which relaxes via reptation, with
side-arms providing the friction.

c.f. H-polymer terminal relaxation

Let’s try to be useful: LCB melts as a
spectrum of pom-poms

  Linear relaxation spectrum

=> τbi, gi
  ‘decorate’ these modes
using nonlinear extensional
data => qi, τsi
 Multi-mode pompom - an example
107 107
Rates
0.001 s-1
Viscosity /Pa·s

0.003 s-1
Extension 0.010 s-1

Viscosity /Pa·s
106 106 Extension 0.030 s-1
0.100 s-1
0.300 s-1
1.000 s-1
3.000 s-1
105 105 10.000 s-1
30.000 s-1

104 Shear 104

Shear
103 10-3 10-2 10-1 100 101 102 103 10-1 100 101 102 103 104
Strain Rate /s-1 Time /s

2.5
Rates
10 s-1
2.0

Stress /105 Pa
5 s-1
2 s-1
1.5 1 s-1

1.0

0.5

0.0 10-1 100 101 102 103

Time /s
Data from Meissner (1972, 1975) and Münstedt and Laun (1979).
Looking at LDPE downstream from
a contraction with flowSolve….
… and with the Cambridge MPR:

Characteristic
‘Fangs’ are
observed
downstream
of contraction
But in MuPP2 it was a different story…

D. Hassell, M. Mackley

D. Hoyle, O. Harlen, TCBM

Linear rheology of arbitrarily branched polymers
Relaxation of a branched polymer

•  Suggests an algorithm which monitors amount of polymer

relaxation as a function of (logarithmic) time
•  Requires a time-integration along chain-contour variables
•  Chains communicate via “constraint release” (tube
dilation)

z(t)

http://sourceforge.net/projects/bob-rheology
Linear rheology of arbitrarily branched polymers
Two relaxation fronts and a widening tube

D.J. Read

z(t)

Inner front determines depth of coherent length

fluctuations: effective potential strength
 Linear rheology of arbitrarily branched polymers
Fix parameters by comparison with data from “monodisperse”
polymers

asymmetric stars H-polymers

Can put in (properly) the polydispersity (i.e. range of polymer sizes)

present in even these polymers.
In polymers, “monodisperse” means same size on a logarithmic scale!
 Linear rheology of arbitrarily branched polymers
Then we can make more predictions for “model” polymers

PB “Dendrimac”
(two-level Cayley tree)

Made by Lian Hutchings, Durham

Can put in (properly) the polydispersity (i.e. range of polymer sizes)

present in even these polymers.
In polymers, “monodisperse” means same size on a logarithmic
scale!
 Linear rheology of arbitrarily branched polymers
Then we can make more predictions for “model” polymers

PS Combs

Made by Pierre
Chambon +
Christine
Fernyhough
(Sheffield)
TGIC by
Taihyun Chang
(Pohang, South
Korea)
Statistical modelling of molecular architectures
Single-site metallocenes, CSTR

UP
Molecules are self-similar and
directional!

Cat
D
Allows:
•  analytical calculations to calculate MWD, DOWN
branching distribution,
(D.J. Read, TCBM, Macromolecules 2001)
•  Monte-Carlo generation of representative set
of molecular architectures
(various algorithms in literature, but easiest is
in Das et al, Journal of Rheology 2006).
 Linear rheology of arbitrarily branched polymers
Can then make predictions for truly polydisperse polymers

Set of metallocene
polyethylenes (Dow).
Branching probability and
molecular weight by
scattering.
Only two remaining
adjustable parameters –
matches entire dataset.
Das et al
J. Rheology, 2006, 50, 207

UP

Cat
D

DOWN
From chemistry to Processing: LDPE
Statistical modelling of molecular architectures (2)

Monte-carlo simulation of free-radical polymerisation;fundamental processes of:

• initiation Ri

Rp=kp[R][m]
•  chain propagation

•  branch formation

•  scission

•  termination by combination
(H. Tobita, J. Polym. Sci. B, 2001)
Assumptions:
•  “batch reaction” and “tubular reactor” are equivalent (i.e. plug flow approximation).
•  continuous initiation
•  Rp>> reaction timescale (“instantaneous chain propagation”)

4. Did the first segment initiate from a branch point?

2. Did it branch?

5. Did any termination combine?

0 3. Did the daughters branch? xf

1. Form a segment

6. Apply the branching algorithm recursively

Generates a 5-parameter representative set of molecular architectures

Termination

Combination

Branching

Scission

xf Conversion
Statistical modelling of molecular architectures (2)
6 well-characterised tubular LDPEs from Basell and Dow

Fitting MWD + g-factor gives indicative prediction

of linear rheology (all with same τe, Me)
Some small tweaks of reaction parameters then
used to fit linear rheology
Approximate non-linear rheology: mapping to multimode pom-pom
The “top-level” division for a Maxwell mode

Stress relaxation due to

constraint release on remaining
Stress relaxation due to polymer chains trapped within
escape of polymer confining tube at this timescale
chains from the confining
tube at this timescale

….then subdivide further into modes with different priority and stretch time
BUT NOTE constraint release physics is not embedded in the pom-pom model.
Approximate non-linear rheology: mapping to multimode pom-pom
Priorities
•  The maximum stretch ratio of a
1 1 chain strand in its tube
1 2 •  Obtained from a “force balance” at
3 branch points
2 5
•  Some chain ends might not
1 5 1 contribute – either too short, or
1 2 relaxing too fast (idea of “snipping”)
4 1 1 … we should use some “time-
dependent” priority.

1 2 2 1
1
1
Approximate non-linear rheology: mapping to multimode pom-pom
One of the tubular LDPEs (1840H).....
•  Transient viscosity during start-up extension / shear
•  Dashed curves – no snipping. Stress levels too high.
•  Dashed –dot – full snipping. Stress levels too low.
•  Full curves – renormalised snipping.
Approximate non-linear rheology: mapping to multimode pom-pom
Tubular LDPE family: analysing how LDPE works

Segments
relaxing at short,
medium and long
times
From reaction chemistry…….

Reaction Chemistry Statistical methods for

polymerisation reactions, e.g. Tobita
algorithm (Monte Carlo) for LDPE

Molecular shape Linear rheology: “BoB”

(predicts linear rheology of
arbitrary branch-on-branch
polymer distributions)
Non-linear rheology
Melt Rheology Approximate parameter
mapping from BoB onto a
multimode pom-pom ensemble

Linear rheology: looking in reasonable shape

Non-linear rheology: not bad, given many of the approximations
X-Slot rheometry for extensional steady-state
can give true q-value spectra

D. Hoyle, T. Lord and D. Auhl (2010)

But both X-slot rheometry (compared to
extensional) and controlled-strain filament
stretching suggest overshoots in LCB flows

D. Hoyle, T. Lord and D. Auhl (2010)

Rasmussen et al. (2005)

Multi-mode pom-pom with overshoot

D. Hoyle, T. Lord and D. Auhl (2010)

LDPE flows calculated in full 3D

O. Harlen, R. Tenchev, P. Jimack, D. Hoyle

From chemistry to Processing: LDPE flows

H. Klein, D. Hassell, O. Harlen, TCBM (2009)

 Conclusions
•  Model materials give insight into entanglement physics

•  Molecular structure has flow-field consequences

•  LCB tailors rates for melt strength

•  Chemistry→Rheology→Flow is possible

•  Teamwork is essential!

•  See more at www.mupp2.com

•  And REPTATE.com !!!