Enhancing the Sustainability of Epoxy Resins and theirFiber Composites

Daniel F. Schmidt*

Associate Professor of Plastics Engineering

University of Massachusetts Lowell

*Lead Research & Technology Associate

Department of Materials Research and Technology

Luxembourg Institute of Science & Technology

(since September 1, 2017)

Luxinnovation Greater Region Plastics WorkshopLuxembourg

December 4th, 2018

Introduction

Epoxy resins find many uses Adhesives and binders

Composites

Coatings

Encapsulants and potting compounds

Nearly all epoxies are petroleum derived

Health and safety can be a concern Acute toxicity (hardeners especially)

Chronic toxicity (hardeners, bisphenols, etc.)

Recycling and reuse are extremely challenging

Many opportunities to improve sustainability!

Selecting a Sustainable Epoxy Resin

Epoxidized Linseed Oil (ELO)

Clean, single-step synthesis Good availability (multiple large suppliers) Inexpensive (<€2/kg) Low viscosity (~1,000 cps) High functionality (f ~ 6, EEW ~ 170-180) Minimally toxic (FDA approved for food contact) Derived from a non-food crop Low reactivity (all secondary epoxies)

http://dawnofthenewage.com/wp-content/uploads/2013/01/linseed-oil-and-flax-seeds.jpg

O

O

O

O

O

O

O

O

O

CH3

O

O

CH3

CH3

O

O

H2O2

RCOOH

Forming a Network:Anhydride Cure

Found workable with a range of liquid anhydrides Catalysis required for curing to proceed High curing temperatures necessary Homogeneous, void-free material produced Highest hardness, modulus values realized

ELO +

Ex.: Nadic Methyl Anhydride(NMA)

O

O

O

CH3

Variouscatalysts

160-200°C2-24 hr.

For more, see: Ind. Eng. Chem. Res., 2017, 56 (10), pp 2658–2666

Ind. Eng. Chem. Res., 2017, 56 (10), pp 2673–2679

Anhydride-cured ELO:DMA & Tensile Properties

Methyltetrahydrophthalic anhydride (MTHPA) identified as optimal hardener; two catalysts studied: DBU = 1,8-Diazabicyclo(5.4.0)undec-7-ene – liquid, cures well but induces voiding during composite formation 2E4MI = 2-Ethyl-4-methylimidazole – requires pre-heating, similar cure levels to DBU but no void formation

Standard (9.6% oxirane oxygen) and high oxirane (10.4% oxirane oxygen) ELO used Control was Hexion RIM 145, a high performance anhydride-cured epoxy used in wind energy

DBU

cat

.

2E4M

I cat

.

2E4M

I cat

., high

ox.

Con

trol (

RIM

145

)

0

20

40

60

80

100

Tem

pera

ture

(°C

)

Main relaxation (from E")

DBU

cat

.

2E4M

I cat

.

2E4M

I cat

., high

ox.

Con

trol (

RIM

145

)

0

20

40

60

80

100

Tem

pera

ture

(°C

)

Main relaxation (from E")

Width of main relaxation

DBU

cat

.

2E4M

I cat

.

2E4M

I cat

., high

ox.

Con

trol (

RIM

145

)

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Tensile

Modulu

s (

GP

a)

DBU

cat

.

2E4M

I cat

.

2E4M

I cat

., high

ox.

Con

trol (

RIM

145

)

0

10

20

30

40

50

60

70

80

90

Bre

ak S

tress (M

Pa)

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Tensile

Modulu

s (

GP

a)

DBU

cat

.

2E4M

I cat

.

2E4M

I cat

., high

ox.

Con

trol (

RIM

145

)

0

10

20

30

40

50

60

70

80

90

Bre

ak S

tress (M

Pa)

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Tensile

Modulu

s (

GP

a)

0

1

2

3

4

5

6

7

Bre

ak S

train

(%)

Anhydride-cured ELO:Composite Production

Vacuum Assisted Resin Transfer Molding (VARTM) gives test coupons

Bioepoxies are ELO cured with MTHPA or NMA, catalyzed with DBU or 2E4MI

Conventional controls are Hexion RIM 135 (amine-cured) and RIM 145 (anhydride-cured), both used in wind energy

Unidirectional (UD) stitched E-glass (Saertex 955) provides reinforcement

Constituent component analysis: Fiber fraction = 52-57 vol%

Resin fraction = 42-46 vol%

Void fraction = 0.7-1.4 vol% (5.5 vol% for ELO-MTHPA-DBU)

Pump

Resin Trap

ResinMold

180°C

UD E-glass Composites:Flexural Properties

Resin dominates transverse properties

Axial modulus is fiber-dominated, while strength is more sensitive to interface

Excessive voiding compromises properties of ELO-MTHPA-DBU in particular

ELO

-NM

A-D

BU

ELO

-MTH

PA-D

BU

ELO

-NM

A-2

E4M

I

ELO

-MTH

PA-2

E4M

I

RIM

145

RIM

135

0

100

200

300

400

500

600

700

800

900

1000

1100

Fle

xu

ral S

tre

ng

th (

MP

a)

Axial

Transverse

ELO

-NM

A-D

BU

ELO

-MTH

PA-D

BU

ELO

-NM

A-2

E4M

I

ELO

-MTH

PA-2

E4M

I

RIM

145

RIM

135

0

10

20

30

40

Fle

xu

ral M

od

ulu

s (

GP

a)

Axial

Transverse

UD E-glass Composites:Flexural Properties

0.00 0.01 0.02 0.03 0.04 0.05

0

100

200

300

400

500

600

700

800

900

1000

1100 ELO-MTHPA-2E4MI

Str

ess (

MP

a)

Strain (mm/mm)

0

5

10

15

20

25

Work

of D

efo

rmation (

MJ/m

3)

Post-fracture

Pre-fracture

ELO-N

MA-D

BU

ELO-M

THPA-D

BU

ELO-N

MA-2

E4MI

ELO-M

THPA-2

E4MI

RIM

145

RIM

135

Axial

Transverse Post-fracture

Pre-fracture

0.00 0.01 0.02 0.03 0.04 0.05

0

100

200

300

400

500

600

700

800

900

1000

1100

Str

ess (

MP

a)

Strain (mm/mm)

RIM145

Conventional systems show catastrophic failure

Bioepoxies provide greater damage tolerance

UD E-glass Composites:Post-fracture Analysis

RIM145 ELO-MTHPA-2E4MI

Conventional control shows strong matrix adhesion, fiber breakage

Bioepoxy shows much more debonding, implying a weaker interface

Adding Reworkability

Montarnal et al achieve reworkability in epoxies via transesterification – Science 334 965 (2011) DGEBA / dicarboxylic acid / tricarboxylic acid

Modulus = 4 MPa

Failure stress = 9 MPa

Failure strain = 180%

DGEBA / glutaric anhydride Modulus = 1.8 GPa

Failure stress = 55 MPa

Tg ~ 80°C

Zinc acetylacetonate used as transesterification catalyst

Implication is that excess hydroxyls are needed

We successfully apply this approach to systems withoutsignificant quantities of excess hydroxyls

Assessing Reworkability:Catalyst Screening

Static load applied to RIM 145 specimens for 4 hours

In the absence of catalysts, <4% strain is observed

In the presence of catalysts, can see up to ~70% strain

No significant changes in hardness after testing

Assessing Reworkability:Stress Relaxation

Characteristic relaxation time (τ*) defined according to Brutman et al.(ACS Macro Lett., 2014, 3, 607)

G/G0 = 1/e τ*

Values of τ* follow Arrhenius relation in RIM 145 control

Ea ~ 95-170 kJ/mol

τ* ~ 80-500 s @ 270°C

Correlation between ln τ* and strain under static load

Using Reworkability:Mechanical Recycling

Mechanical grinding

Reprocessing via compression molding

25 mm

Regrind 0.2 mm Regrind Composite

25 mm

Regrind 2 mm

25 mm

Regrind 5 mm

25 mm

Pristine cast resin

Regrind 5 mm

Regrind 2 mm

Regrind 0.2 mm

Regrind Composite

Using Reworkability:Mechanical Recycling

Pris

tine

Cas

t Res

in

Reg

rind

5 m

m

Reg

rind

2 m

m

Reg

rind

0.2

mm

Reg

rind

Com

posi

te

Reg

rind

Com

posi

te +

2 m

m

0

1

2

3

4

5

6

7

8

9

E' a

t 2

3°C

(G

Pa

)Pr

istin

e C

ast R

esin

Reg

rind

5 m

m

Reg

rind

2 m

m

Reg

rind

0.2

mm

Reg

rind

Com

posi

te

Reg

rind

Com

posi

te +

2 m

m

0

1

2

3

4

5

6

7

8

9

E' a

t 2

3°C

(G

Pa

)

0

10

20

30

40

50

60

70

T (°C

)

Modulus and Tα are mostly retained regardless of particle size(similar results in bioepoxy systems)

Stress and strain at failure are highest with finest particle size

Pris

tine

Cas

t Res

in

Reg

rind

5 m

m

Reg

rind

2 m

m

Reg

rind

0.2

mm

0

10

20

30

40

50

60

70

Str

es

s a

t F

ailu

re (

MP

a)

Pris

tine

Cas

t Res

in

Reg

rind

5 m

m

Reg

rind

2 m

m

Reg

rind

0.2

mm

0

10

20

30

40

50

60

70

Str

es

s a

t F

ailu

re (

MP

a)

0

1

2

3

4

Stra

in a

t Fa

ilure

(%)

Using Reworkability:Chemical Recycling

Composite: RIM145 / E-glass

Solvent: 1-Dodecanol

Catalyst: n-butyltin tris(2-ethylhexanoate)

Heated for 12 hours

Cleaned in solvent, then water

Using Reworkability:Chemical Recycling

Reclaimed E-glass composite

Reclaimed E-glass fibers

Stiffness entirely retained

Strength reduced, likely due to changes in fiber sizing

Using Reworkability:Chemical Recycling

0

100

200

300

400

500

600

700

0

10000

20000

30000

40000

50000

60000

Fle

xura

l Str

en

gth

[M

Pa]

Fle

xura

l Mo

du

lus

[MP

a]

Flexural Modulus and Flexural Strength

Flexural Modulus

Flexural Strength

Bioepoxy / E glass

Bioepoxy /

carbon

Can improve properties with reclaimed fibers in bioepoxies(!)

Rate of chemical recycling is much faster as well

Fle

xu

ral

Mo

du

lus (

MP

a) F

lex

ura

l Stre

ng

th (M

Pa

)

Using Reworkability:Reuse via thermoforming

Flat sheet of RIM 145 epoxy resin preparedin the presence of transesterification catalyst

Sample placed in tooling, heated to rework temperature, pressure applied

Sample cooled in water to yield rigid, permanently deformed epoxy part!

Summary & Conclusions

Anhydride-cured structural thermosets successfully produced based solely on epoxidized linseed oil (ELO)

Bioepoxy composites provide good mechanical properties

Results are competitive with highly optimized controls

Bioepoxy composites are more damage tolerant than controls

Interfacial debonding, lower axial strength imply weak interface

Performance may be improved via optimization of fiber sizing

Transesterification catalysts enable recycling and reuse

Mechanical recycling demonstrated, particle size effect noted

Chemical recycling demonstrated, reclaimed fibers give high performance composites, especially effective with bioepoxies

Thermoforming demonstrated, promises reuse of existing parts

AcknowledgementsCollaborators and Group Members

Assoc. Prof. Emmanuelle Reynaud (UML Mechanical Engineering)

Chris Kuncho (PhD), Wenhao Liu (PhD), Johannes Möller (PhD),Julia Kammleiter (MS), Julia Stehle (MS), Dr. Akshay Kokil

Financial Support

National Science Foundation (Award #1230884)

Massachusetts Toxics Use Reduction Institute

Materials and Analytical Support

ACS Technical Products (Epoxidized linseed oil)

Huntsman Advanced Materials (MTHPA)

Hexion (RIM 135 & 145)

General Support

W. Liu, Dr. A. Kokil & the Reynaud-Schmidt Research Group

P. Casey, Dr. E. Ada & the UML Core Research Facilities

D. Rondeau, M. Shone, Dr. X. Chen, Prof. S. Johnston,UML Plastics Engineering and the UML Composites Lab

FOR MORE INFORMATION:ACS Symp. Ser., 2018, vol. 1310, ch. 18, pp 281-295

THANK YOU FOR YOUR ATTENTION!