P•

MM M M

PM• MM

M

M M

MM

M

M

M

M

M

M

I•

IM•

M•

M

M

M

M

M

Microemulsion Polymerization

Eric W. KalerDepartment of Chemical Engineering

University of DelawareNewark, DE 19716

Jen O’DonnellKevin Hermanson

Carlos Co (U. Cincinnati)Renko de Vries (Wageningen U.)

Characteristics and Structures

Emulsion:•Two Phase•Energy Needed to Form•Opaque•Monomer Drops > 1 µm

Microemulsion: •One Phase•Spontaneous•Transparent/ •Swollen Micelles ( D < 20 nm)

Why Study Microemulsion Polymerization?

Polymerization in a confined environment may lead to unique polymer morphologies, e.g. tacticityand knotting.

“Knotted” polymer chain in solution

Produces nanosized (~15 nm) latex particles smaller than those obtainable by emulsion polymerization

“Paint” the walls of a microporous material

Dry

“Seeds” for emulsion polymerization

Extremely high MW( ~20 million daltons)

are readily made

Outline

• Problems and Model Mixtures• How to Make Microemulsions• Microstructures• Polymerization – Kinetics and Model• Structure Evolution• Multiple Additions• The End!

Ternary Phase Diagrams at 60ºC

Added degree of freedomfrom mixing surfactants is used to tune one phase oil-in-water microemulsions

Gibbs Phase Prism at 60ºC

Mixing ratio sets surfactant

chemical potential

Free Radical Thermal Polymerization

kp 60°C (L mol-1 s-1)

3401015 1140 995

Monomers

CH2

CH3

OO

CH2CH2

CH2CH2

CH2CH3

n-C6MA

CH2

CH3

OO

CH2CH2

CH2CH3

n-C4MA

CH2

CH3

OOC

CH3

CH3CH3

t-C4MA

CH CH2

Styrene

Water Solubility60°C (mM) ~ 0.4 3.4 4.3 4.6

Tg (°C) - 5 20 128 106

Reaction at 60°C

A Basic Recipe

• Surfactant mixture to tune phase behavior

Cationic surfactantsDTAB

DDAB

• Monomer with low water solubility - hexylmethacrylate (0.4 mM)

• Polymer with low Tg - polyhexylmethacrylate (-5°C)

• Radical initiator with simple dissociation kinetics - V50 (not persulfates)

P•

MM M M

PM•

MM

M

M M

MM

M

M

M

M

M

M

I•

IM•M•

310~Particles#Micelle#

Initiation by IM•

Propagation

Initiationby M•

ChainTransfer

A Simple View Of Microemulsion Polymerization

M

M

M

M

M

5 nm

V – 50 Polymerizations

1. Rapid polymerization

2. 100% conversions

3. Rate profile parabolic

4. Average maximumrate at about 39%conversion

Microemulsion Polymerization – John Morgan

Modeling the V-50 Rate Curves

• Fundamental rate equation for addition polymerization:

• Microemulsion conversion form:

][ •R

cRktc

p ][ •=∂∂

−

c = monomer concentration at polymerization locus= concentration of propagating radicals

Mo = monomer concentration at polymerization locus (M)

N* = propagating radical concentration in whole microemulsion (M)

o

*p )()(

MfctNk

tf=

∂∂

Modeling the V-50 Rate Curves (cont.)

• Assumption: Monomer concentration within polymer particles given by:

c = co (1 – f )co = initial concentration of monomer, M

• Entry rate is constant; all radicals remain active

N*(t) = ρot

ρo = rate of radical entry, M s-1

Modeling the V-50 Rate Curves (cont.)

• Rate Equation:

o

oop

Mck

A

fAttf

ρ=

−=∂∂ )1(

)1()()(

o

p

o

*p ft

Mck

MfctNk

tf oo −==∂∂ ρ

Or

Parameter:

Conversion: )21exp(1 2Atf −−=

Modeling Implications

2/1

5.0

:Rate Maximum of Time

39.01 :Rate Maximumat Conversion

:Maximum Rate

−

−

=

=−=

=′

At

ef

eAf

• Dependence on Initiator Concentration:

Assume ρo = 2kd [I]

then A goes as [I]

Measured Rate Constants

– Propagation Rate Constant• kp = 995 M-1s-1 (Pulsed Laser Polymerization)

– Initiator Decomposition Rate Constant• kd = 3 x 10-5 M-1s-1 (Literature)

– Initial Monomer Concentration in Droplet• C0 = 1.0 M (SANS)

– Initial Monomer Concentration in microemulsion• M0 = 0.257 (Formulation)

N.B. No fitted parameters!

Comparison with Experiment

Quantitative Agreement Through to Full Conversion

Kinetic Modeling: Monomer Concentration at Locus of Polymerization

Carlos Co, Renko de Vries

How Does the Microstructure Evolve?Case I

Flory-Huggins for bothlatex particles and micelles.

All monomer is taken upby latex particles at ~ 5%

conversion.

Case IIFlory-Huggins for latexparticles and curvature

energy for micelles.Monomer partitions

between latex and micelles.

Case IIINo swelling of polymerlatex. Polymerizationoccurs in shell of latexparticles with monomerconcentration equal to

that in the micelles.

Mon

omer

C

once

ntra

tion

Conversion ConversionConversion

Monomer concentration is approximately linear with conversion.Can differentiate only using Small Angle Neutron Scattering (SANS).

Small Angle Neutron Scattering (SANS)

Length1~q⎥⎦

⎤⎢⎣⎡θ

λπ

=2

sin4q

I(q) = n P(q) S(q)

P(q) ⇒ Single particle properties(size, shape, composition)

S(q) ⇒Relative positionsof particles due to

interactions

FormFactor

StructureFactor

⇒

⇒

2D Detector

Neutron Source θSample

λ q

What Can SANS Tell Us?I(

q)

MicelleScattering

ParticleScattering

ObservedScattering

q q q

How does the SANS spectra change as the microstructureevolves from micelles to a mixture of polymer particles and micelles?

Reactor for Online Scattering During

Polymerization

Online SANS / Kinetics Experiments

Gradual shifts in SANS spectra indicate that monomer partitionsbetween the micelles and the polymer particles. Case I is incorrect.

q (Å-1)

Inte

nsity

(cm

-1)

1

10

100

1000

0.01

0.10

0%10%38%65%85%96%100%

IncreasingParticleSize andNumberDensity

DecreasingMicelle Size

C6MA (DTAB/DDAB)

Connection of particle size to MWD

• Basic Idea– Growing chain of L segments was initiated at an

earlier time t1

– At t1, calculate Δt for one propagation event– Number of chains initiated during Δt is N(L)

– Assume single chain particles• Final result is analytical• See Morgan and Kaler, Macromolecules, 1998

Predicted MW and Size Distributions

SANS Model for Online SANS Spectra

Discretize Model-Predicted Particle Size Distribution

Effective HS InteractionsForm Factors

Polymer Micelle

Three AdjustableParameters

Rmaj

RminREHS

Calculate model intensities using Vrij’s analytical equation

Online SANS Modeling Results

q (Å-1)

Inte

nsity

(cm

-1)

1

10

100

1000

0.01

0.02

0.04

0.06

0.08

0.10

0.20

C6MA (DTAB/DDAB)

0% 65% 100%

Minor Radius (Å) 29 23 20

Aspect Ratio 2.1 2.3 2.2

Minor/HSRadius Ratio 1.6 1.7 1.6

Micelle Dimensions(Fitted Parameters)

0% 65% 100%

Avg Radius (Å) 160 130

Stdev (Å) 50 70

Particle Size Distribution(Model Predictions)

To within the accuracy of the predicted particle size distributions, the polymer particles are not swollen by monomer.

0%

100%

65%

SANS Model Fitting Results

Particle size distribution model is consistent with SANS

Micelle size decreases steadilywith increasing conversion

Validation of SANS Swelling Experiments(n-C4MA)

q (Å-1)

Inte

nsity

(cm-1

)

1

10

100

1000

0.00 0.05 0.10 0.15

OnlineSwelling4%

33% OnlineSwelling

75% OnlineSwelling

cryo-TEM by Stefan Burauer (Universitaet zu Koeln)

Molecular Weight Distribution

By Pat Cotts, Dupont

1.422.41001.226.137

1.918.64

Mw/MnMw (106)Conv. (%)

GPC/MALLS/RI by Patricia Cotts (DuPont)

Molecular Weight Distributionsdw

/d(lo

gM

)

Molecular Weight (106)

n-C6MA

43%

79%

95%

n-C4MA

33%

60%

76%

94%

Polystyrene Exceeds Chain Transfer Limit

P • + M

PM •

P + M •

kp

ktr

kp/ktr ratio sets characteristic

limiting molecular weight

Styrene free-radical polymerizations (60°C) limited to ~2·106 daltons.

MW~15 ·106 polystyrene is consistently prepared by microemulsion polymerization

Styrene

dw/d

(log

M)

Molecular Weight (106)

27%

53%

70%

91%

Summary: A Simple Model for Microemulsion Polymerization

Rate Equation

( )

knownisMρck

A

f1Attf

o

oop=

−=∂∂

⎟⎟⎠

⎞⎜⎜⎝

⎛−−=

2tA1exp1f:Conversion

2

39.021exp1f:ratemaxatConversion =⎟⎠⎞

⎜⎝⎛−−=

Predictions for Particle Size and Molecular Weight Distributions

Kinetic Predictions

Molecular Weight (106)

dw/d

(log

M)

V50

Con

cent

ratio

n

How to Increase Polymer Loading?

…Sequential Addition Polymerization

[ ]( )⎟⎟⎠

⎞⎜⎜⎝

⎛+−−= tNtIk

Mck

exp1f o*2

do

op

N* = ρο t + N*o

[ ]( )od NtIkM

fcktf *

o

op 2)1(

+−

=∂∂

Kinetics

Initial Step 1

M

M

M

M

Refill

M

•Radicals present from prior monomer additions

1 2

3 4

5

Additional Steps

Step 2

Determine Cmon from Single Addition KineticsC6MA/DTAB/DDAB (5% Total Surfactant)

Conversion

0.0 0.2 0.4 0.6 0.8 1.0

Mon

omer

Con

cent

ratio

nC

mon

(M)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

wt%mon=0.475

wt%mon =1.90

Monomer Partitioning MapSingle Addition Polymerization

Single Addition Kinetics (Scaled*)

Increase Total Monomer

Concentration

*0.95% scaled 1.5X, 1.43% scaled 2X, 1.90% scaled 2.5X

Conversion (f)

0.0 0.2 0.4 0.6 0.8 1.0

d(C

onve

rsio

n)/d

(tim

e) (s

-1)

0.000

0.002

0.004

0.006

0.008

0.010

0.012

wt%mon=0.475

wt%mon=0.95

wt%mon=1.90

wt%mon=1.43

Determine Cmon for Multiple Addition Polymerization

Conversion (f)

0.0 0.2 0.4 0.6 0.8 1.0

Mon

omer

Con

cent

ratio

nC

mon

(M)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

Monomer Partitioning MapSingle Addition Polymerization

0.475% Monomer0.475% Polymer

0.475% Monomer0.95% Polymer

0.475% Monomer1.43% Polymer

New Co

Conversion (f)

0.0 0.2 0.4 0.6 0.8 1.0

Mon

omer

Con

cent

ratio

nC

mon

(M)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

Monomer Partitioning MapMultiple Addition Polymerization

Addition 4

Addition 1

Addition 1 Co=0.66MAddition 2 Co=0.56MAddition 3 Co=0.47MAddition 4 Co=0.37M

Cmon=Co(1-f)

Conversion (f)

0.0 0.2 0.4 0.6 0.8 1.0

d(C

onve

rsio

n)/d

(tim

e) (s

-1)

0.000

0.002

0.004

0.006

0.008 Addition 1

Addition 2

Addition 3

Addition 4

Multiple Addition Kinetics

Measured vs. Predicted Kinetics(offset*)

Conversion (f)

0.0 0.2 0.4 0.6 0.8 1.0

d(C

onve

rsio

n)/d

(tim

e) (s

-1)

0.000

0.001

0.002

0.003

0.004

0.005Co decreases lowering reaction rate

Measured Kinetics

[ ]( )o*

do

op NtIk2M

)f1(cktf

+−

=∂∂Multiple Addition Model

The predicted data matched the measured data when N*o=0

* Addition 2 offset 0.001, Addition 3 offset 0.002, Addition 4 offset 0.003

Addition 1Addition 2Addition 3Addition 4

2530

3540

4550

5560

65

0

20

40

60

80

100

423222

121

DiameterAddition

Inte

nsity

(I/I m

ax*1

00)

Generation Number0 10 20 30 40 50

Dia

met

er (n

m)

0

20

40

60

80

100

120

140

QLS Measured Particle SizeMaximum Predicted Particle Size

Particle Size DistributionContin AnalysisParticle Size

Multiple Addition Particle Size Measurements

0 8.84.6 16.212.6 19.4

% Polymer

Predicted particle size if no new particles are formed

After 43 additions latex contains 17% polymer and 4% surfactant

Experimentally Measured Particle Size

Summary

• Microemulsion polymerization produces small monodisperse particles

• Initiator charge plays no role (with pH control…)• Reaction rate and MWD can be modeled with minimal

assumptions and no free parameters• Microstructures “meter” monomer and control

polymerization• Commercially interesting concentrations can be

produced by sequential (or continuous) polymerization