Controlling Polymer Rheological Properties Using Long-Chain Branching

(1)

National Science Foundation WHERE DISCOVERIES BEGINNational Science Foundation WHERE DISCOVERIES BEGIN

Controlling Polymer Rheological Properties Using Long-Chain Branching

PI: Ronald Larson

Univ. of Mich., Dept of Chem. Eng., Macromolecular Science and Engineering Program

Possible co-PI: Michael Solomon

Univ. of Mich., Dept of Chem. Eng., Macromolecular Science and Engineering Program

Possible co-PI: Jimmy Mays

Univ. of Tennessee, Dept of Chemistry

(2)

National Science Foundation WHERE DISCOVERIES BEGIN

Industrial Relevance

“The flow behavior (‘rheology’) [of polymers] is enormously sensitive to LCB [long chain branching] concentrations far too low to be detectable by spectroscopic (NMR, IR) or chromatographic (SEC) techniques. Thus polyethylene

manufacturers are often faced with ‘processability’ issues that depend directly upon polymer properties that are not explainable with spectroscopic or chromatographic

characterization data. Rheological characterization becomes the method of last resort, but when the rheological data are in hand, we often still wonder what molecular structures gave rise to those results.”

Janzen and Colby, J. Molecular Structure, 1999

Innovation through Partnerships 2

(3)

Rheology, Processing and Long- Chain Branching

< 1 LCB’s per million carbons significantly affects

rheology!

branched thread-like micelles

branched polymers

(4)

Project Goals

• Develop industrially useful tools for inferring long-chain branching levels from rheology

• Develop optimization strategies for improving processing and product properties through

control of long-chain branching

• Provide software tools and training as needed for industrial applications

(5)

Objectives & Research Methods

• Measure rheology of commercial polymers

• Combine this with conventional characterization by SEC, light scattering, and knowledge of

reaction kinetics

• Use “Hierarchical model”, a computational tool, to determine a long-chain branching profile of commercial polymers.

• Determine how changes in the long-chain branching profile could alter rheological properties in desirable ways.

(6)

Hierarchical Model

Larson et al., (2001, 2006, 2011)

comb H star

linear

• A complex commercial branched polymer is represented by an ensemble of up to 10,000 chains.

• This ensemble represents the range of

branching structures and the molecular weight distribution of the commercial polymer.

•The ensemble is generated from a combination of GPC characterization,

knowledge of reaction kinetics, and rheology.

•The ensemble is fed into the “Hierarchical Code,” and a prediction of the linear rheology (G’ and G”) emerges.

Das, Inkson, Read, Kelmanson, J.

Rheol. (2006)

Relaxation of each molecule is tracked in the time domain, as it relaxes from the tips of the branches, inwards towards the backbone. At long times, branches act as drag centers, slowing down motion of the branch or backbone to which they are attached. The

contributions of all molecules in the ensemble to the rheology are combined, and converted to the

frequency domain to predict G’ and G’’.

(7)

Example 1: Characterization of Anionically Synthesized “H” Polymer

Synthesized by Rahman and Mays

Linear Mw/Mn

=1.01

Star Mw/Mn

=1.03

H

Mw/Mn

=1.07

(8)

Chemically Likely Structures

(9)

Identification of Structures

0 5 10 15 20 25 30 35

n (a.u.)

t_R (min)

17 18 19 20 21 22

T ( oC)

Mw 38.5k

54.8k

73.6k

Unfractionated Star

0 5 10 15 20 25 30

n (a.u.)

t_R (min)

T ( oC)

14 16 18 20 22 24 26 28 30

Fractionated H 32

71k 95k 129k

114k

Star (Semi- H):

H:

TGIC from

Hyojoon Lee and Taihyun Chang

Using TGIC: Temperature Gradient Interaction Chromatography

(10)

Identification of Structures

0 5 10 15 20 25 30 35

n (a.u.)

t_R (min)

17 18 19 20 21 22

T ( oC)

Mw 38.5k

54.8k

73.6k

Unfractionated Star

0 5 10 15 20 25 30

n (a.u.)

t_R (min)

T ( oC)

14 16 18 20 22 24 26 28 30

Fractionated H 32

71k 95k 129k

114k

Star

(Semi- H):

H:

TGIC from

Hyojoon Lee and Taihyun Chang

(11)

11

Comparisons of theoretical predictions and experimental measurements

Star (semi-H) H

blend

from Chen, Rahman, Mays, Lee, Chang, Larson

Xue Chen

(12)

12

Example 2: Blends of Linear Exact 3128 and Branched PL1880 Polyolefins

X.Chen, C. Costeux, R. Larson. J. of Rheology 54(6) 1185-1206, 2010

(13)

13

Rheology of Blends of Linear Exact 3128 and Branched PL1880 Polyolefins

T=150C

Increasing LCB

0.3 LCB’s per million carbons!

(14)

Start

Generate random number R: U(0,1)

R>pp

Propagation Termination

Save molecule

Add monomer Add macromonomer Generate random

number R: U(0,1)

R>lp

NO YES

monomer addition

addition of unsaturated chain

generation of dead structured chain

-hydride elimination

Reaction kinetics of LCB PE using single-site catalyst

Algorithm for Monte Carlo simulation of LCB PE using

single-site catalyst

Monte Carlo probabilities

Costeux et al., Macromolecules (2002)

propagation probability

monomer selection probability

Generating an Ensemble of Chains for a Commercial Single-Site Metallocenes

P_x,n+ M ¾ ® ¾ P^kp _x+1,n

P_x,n+ D_y,m⁼ ¾ ^kLCB¾ ¾ P® _x+y,n+m+1

P_x,n+CTA¾ ® ^kCTA¾ ¾ D_x,n⁺ + P_1,0

P_x,n ¾ ® ¾ D^kb _x,n⁼ + P_1,0

pp= R_p+ R_LCB R_p+ R_LCB+ R_T

lp= R_p R_p+ R_LCB

(15)

15

A Priori Predictions of Commercial Branched Polymer Rheology

(16)

Outcomes/Deliverables

• Measurements of rheological properties of commercial polymers

• Measurement of SEC curves for select commercial polymers

• Computer software and training for predicting rheological properties

• Assessment of impact of changing

• branching structure on rheology

(17)

Impact

• Improved ability to design and control polymer processing properties

• Ability to infer likely branching characteristics from rheology

• Develop methods of extracting “hidden”

features of molecular structure through rheology of samples blended with simpler linear polymers

(18)

Project Duration and Proposed Budget

• 1-4 years, depending on polymer to be

tackled, number of samples to be studied, availability of industrial data, such as GPC

data, and the solvents/conditions required for characterization

• Budget: $75,000/year