15/10/2016

1

Cellulose-based nanocomposites

Kristofer Gamstedt

October 14, 2016

Presentation at CICY, Mérida, Mexico

Uppsala University

StockholmUppsala

15/10/2016

2

Need for renewable and biodegradable

and renewable materials

• Conventional plastics are petroleum-based

and do not degrade

Length scales of wood

15/10/2016

3

Infrastructure: Pulp and paper industry

Pros …

Inexpensive

Light

From renewable resource

Recyclable

Decent mechanical properties

Good for local industry

15/10/2016

4

…and cons

Sensitive to heat

Processing difficulties

Need for materials development

Sensitive to moisture

‘Poor’ mechanical properties

Nanocomposites

• Clay reinforced polymers

• Toyota Central R&D labs, 1980s

• Surprising property improvements

15/10/2016

5

Microfibrillated cellulose

Turbak et al. 1983

• Microfibrillation Cellulose fibrils with nm

lateral dimensions

• AFM image (1 µm2) of MFC:

TEMPO mediated oxidation

• Reduction in energy consumption in the

fibrillation process

• Saito et al., with and without TEMPO:

15/10/2016

6

Structure-property relations

Processing Mechanicalproperties

Nano- and microstructure

Mechanisms

Modelling

Nano- and microstructural tailoringfor optimal mechanical properties

Vision:

Empirical approach Modelling

Chemical synthesis

OHO

OO

O

OO

OO

O

H

WOOD FIBRIL

WOOD FIBRIL

Experimentally measured macroscopic properties

ε

σ

Composite microstructure

∑ ∫=

−

−+′+=N

k

z

z

kkjjjkljll

k

k

zHzKQNN1

m0

1

d ))(()( βε

∑ ∫=

−

−+′+=N

k

z

z

kkjjjkljll

k

k

zzHzKQMM1

m0

1

d ))(()( βε

Mathematical modelling

Predicted properties ε

σ

15/10/2016

7

Reinforcement

L/D =50-100L/D=1-2

Cellulose particles Cellulose fibrils

σ σ σ σ

Stress concentration Efficient load carrying

Increased strength and stiffness with long fibrils!

0

20

40

60

80

100

120

140

0 0.5 1 1.5 2 2.5

Strain [%]

Str

ess [M

Pa

]

Transverse direction

Longitudinal direction

Strength = 55 MPa

Strength = 125 MPa

Tensile behaviour of wood-fibre

reinforced composite

15/10/2016

8

Fibre orientation

Longitudinal (MD) Transverse (CD)

Protruding partiallypulled-out fibres

Cracks aroundoblique fibres

Efficient use ofload-carrying fibres

Inefficient useof reinforcement

Transverse (CD)

Cracks aroundoblique fibres

Inefficient useof reinforcement

Strength

Depends onmicrostructuralinhomogeneities

Concentrationsof transversebundles

Interfaces mayplay a rôle

15/10/2016

9

Failed specimens

Failure at different locations

at microstructural ‘defects’

Empirical approachMechanism-based

approach

Improved materials development !

Micromechanics as a tool for quantitative materials design !

Tailoring of material structures

15/10/2016

10

Pulp industry decline

• New products?

• New direction in property space?

• Market pull vs tech push

Property space

Carbon composite

Cellulosics

Ceramics

Hydrophobicity

Fracture toughness

Cost

Fatigue resistance

Themral expansionStrength

Stiffness

15/10/2016

11

Engineering challenges

• Upscaling from lab to industrial scale

• Dispersion (and orientation) of nanofibrils for

enhanced properties

• Energy consumption in processing + cycle

times

• Cost competitiveness

• Moisture issues in outdoor applications

Underlying scientific challenges

• Drying, consolidation, shaping of

nanocellulose components

• Moisture resistance: Chemical treatment

• Dispersion and orientation

• Rationalize materials development by phyiscal

understanding and modelling

15/10/2016

12

Selected topics

1. How stiff are cellulose nanofibrils really?

2. Moisture induced swelling

3. Nanoscale fracture processes

Those that did the work:

Cristian NeaguDevelopment engineer at Tetra Pak

Karin AlmgrenManager of B.Eng. programmes at KTH

Thomas JoffrePost-doc in biomaterials at Uppsala University

Gabriella JosefssonProduct specialist at spin-off Disruptive Materials

15/10/2016

13

Prediction of elastic properties of wood cellulose nanofibrils

Gabriella Josefsson, Kristofer Gamstedt

Ångström Laboratory, Uppsala University, Sweden

Bjørn Steinar Tanem, Yanjun Li, Per Erik Vullum

SINTEF Materials & Chemistry, Trondheim, Norway

Estimation of stiffness of nanofibrillated cellulose

E = 10 GPa

E = 20 GPa

E = 150 GPa

E = ? GPaNFC

15/10/2016

14

Size of NFCs

Cross-section

5 nm5 nm

NFC Ø 20 nm

Elementary fibril Ø 5 nm

Tensile tests not possible very complicated

Work strategy

High-resolution TEM and X-ray diffraction analysis: Crystal nanostructure of wood-based NFC

Use structure and theoretical estimates of crystal stiffness to predict Young’s modulus of NFC

Compare stiffness with 3-point bending in AFM FNFC

Span L

AFM tip

15/10/2016

15

Wood NFC from homogenized Norway spruce sulphite pulp (Borregaard)

S

Cellulose microfibril

Fibril

Hemicellulose

Lignin

Wood pulp fibre

S

S

P

3

2

1

Ø 20-30 nm

Ø 3-5 nm

NFC dimensions

15/10/2016

16

Self-consistent Mori-Tanaka model(e.g. Laws & McLaughlin, JPMS, 1977)

Input data:

• Stiffness of cellulosecrystallite, transverselyisotropic

• Stiffness of isotropicamorphous cellulose matrix

• Molecular dynamicssimulations for stiffness of the matrix and ellipsoid

• X-ray diffraction data for aspect ratio and degree ofcrystallinity

High-resolution TEM on (un)stained NFC

Crystals with same size as revealed on unstained NFC

5nm10nm

15/10/2016

17

X-ray diffraction results

10 15 20 25 30 35 40

2θ (degree)

Inte

ns

ity

Estimates based on Scherrer equation

Crystal width: D =0.9λ/Bcos(θ) at (200)

Crystal length: D =0.9λ/Bcos(θ) at (004)

Crystallinity: (I200 – IAm)/I200

Iβ(200)

Iβ(004)

An aspect ratio of 1.9 is obtained

A crystallinity of 69% is obtained

Crystals oriented along fibril direction

Crystallite aspect ratio

X-ray diffraction data from wood cell wall typically give an aspect ratio of ~ 5-8

M. Peura et. al. Trees 2008, 22, 49-61 (Norway spruce)

M. Peura et. al, Wood Sci Tecnol 2007, 41, 565-583 (Norway spruce)

Chemical/enzymatic pretreatment and homogenization reduce the crystal aspect ratio in MFC?

15/10/2016

18

Structure

1. Aspect ratio: 1.8 (from HR-TEM + image analysis/X-ray diffraction)

2. Crystalline volume fraction of crystals: 69% (from X-ray diffraction)

3. Orientation distribution function: Unidirectional (HR-TEM)

Elastic properties of constituents

1. Elastic properties of crystalline cellulose (Tashiro et al., Polymer 1991)

E_longitudinal = 167.79 GPaE_transverse = 34.86 GPaG (shear modulus) = 5.81 GPa

2. Elastic properties of amorphous cellulose (Eichhorn et al., Cellulose 2001)

E = 5.0 GPa

Input parameters

Effect of crystallite aspect ratio

15/10/2016

19

Comparison with literature

Cotton nanowhiskers: 57-105 GPaElastic modulus of cotton cellulose nanowhiskers determined from Raman spectroscopy: R. Rusli and S.J. Eichhorn, Appl. Phys. Lett 2008, 93(3),

Tunicate nanowhiskers: 143-150 GPaElastic modulus of tunicate cellulose nanowhiskers, determined from Raman spectroscopy: A. Sturcová, G. R. Davies, S. J. Eichhorn, Biomacromolecules 2005, 1055-1061Elastic Modulus of Single Cellulose Microfibrils from Tunicate Measured by AFMS. Iwamoto, W. Kai, A. Isogai, T. Iwata, Biomacromolecules 2009, 10, 2571-2576.

Bacterial cellulose: 78-114 GPaStiffness of bacterial cellulose filament (Raman spectroscopy + back calculation): Y-S Hsieh, H. Jano, M. Nogi, S. J. Eichhorn, Cellulose 2008, 15, 507-513Stiffness of bacterial cellulose filament from AFM force-deflection curves: G. Guhados, W. Wan, J. L. Hutter, Langmuir 2005, 21, 6642-6646

Lyocell fibrils: 93 GPaA method for testing the elastic modulus of single cellulose fibrils by AFM Q. Cheng and S. Wang; Composites Part A 2008, 393, 1838-1843

Wood NFC: 84 ±±±±23 GPaEffects of process and source on elastic modulus of single cellulose fibrils evaluated by AFMQ. Cheng, S. Wang, D. Harper; Composites Part A 2009, 40, 583-588

Mechanical testing of MFC by 3-point bending in AFM

Standard grating, MicroMarch

A drop of NFC suspension put on the substrate

15/10/2016

20

FMFC

Span L

AFM tip

Beam load-deflection relation

FL3

192δIE =

πD4

64I =

100 nm100 nm

Shape of MFC cross-section

TEM cross section of macrofibrils TEM cross section of MFC

Moment of inertia: Solid cylinder

5 nm

15/10/2016

21

Preliminary result on macrofibrils (Ø 100-150nm)

Young’s modulus: 50 GPa± 10 GPa

Predicted: 68 GPa(TEM + XRD + modelling)

AFM tapping mode

Remarks on MFC stiffness

• Microfibrillated cellulose has high potential as reinforcement in nanocomposites

• Stiffness of present MFC is lower than expected: 40-70 GPa

• Need for development of processing procedures to disintegrate MFC with retained high native stiffness

• Processabilty vs. performance should be addressed

15/10/2016

22

Moisture-induced dimensionally instability

Contribution from cellulose microfibrils

Moisture uptake in wood-

based materials

polysaccharides

Dimensional instability, degradation of properties, fungal attack etc.

15/10/2016

23

Wood-fibre composites

Slender fibres as reinforcements, not wood particles as fillers…

Uniform properties (in-plane isotropic) 3D shapes Polymer surface

2 cm

Fibre swelling

c

rr

∆=

εβRadial hygroexpansion

c

ll

∆=

εβAxial hygroexpansion

lr ββ >> (El >> Er)

K. Schulgasser, J Mech Phys Solids 35, 35 (1987)

15/10/2016

24

Hygoexpansion measurement of individual

fibres?

Dry Moist

Cumbersome, time consuming, …

Hygroexpansion from

consolidated fibre mats?

Relative contribution from fibres?

15/10/2016

25

Composite hygroexpansion

Out-of-plane hygroexpansion

Thickness directionIn-plane hygroexpansion

Radial direction

Random in-plane isotropic fiber orientation

Out-of-plane swelling is larger

Fibrous microstructure

Concentric cylinder assembly,

Self-consistent scheme

Z. Hashin, J. Appl. Mech., 50, 481 (1983).

15/10/2016

26

Laminate analogy

Laminate theory

)42(8

1

)46(8

1

)4233(8

1

66122211

LAM

66

66122211

LAM

12

66122211

LAM

22

LAM

11

QQQQQ

QQQQQ

QQQQQQ

+−+=

−++=

+++==

TLTTL

TLTLTTTLLAM

2

)(

EEE

EEEL

rν

ββνβββ

++

+++=

In-plane stiffness

In-plane hygroexpansion

15/10/2016

27

Swelling constraint

L

T

z

In-plane isotropic substrate

Thin unidirectional ply

r

Calculated out-of-plane

hygroexpansion

TLTTL

LTTL

L

LT

T

T'T

LAM

2

)(

EEE

EE

EEz

ν

ββννββ

++

−

−+=

Experimentally measured

Depends on E and β of consituents

( )∑=

−N

i

iziz

1

2

fT

LAMLAM

,

fT

),(min

ββββ

x

Back out fibre hygroexpansion:

15/10/2016

28

Materials

Fibres

• Bleached kraft softwood fibres, Imatra Mill, Finland

Untreated - reference

• Same fibres

Cross-linked BTCA (butyltetracarboxylic acid)

Matrix

• Polylactic acid

Composite

• 30 and 40 v% fibres

• Hot-press moulding

Fibres of polylactic acid

Hot-pressing of commingled fibre mats

Maintained fibre slenderness

Wood fibresConsolidated solid composite

HeatPressure

15/10/2016

29

Cell-wall cross-linking

Cross-linking

molecules

Restrained transverse swelling

Experiments

Out-of-plane swelling:

50%RH to FSP

Tensile testing

Ef from Neagu et al. J Comp. Mater. 40, 663 (2006)

Microscopy

15/10/2016

30

Input parameters

Em

(GPa)νm

βm

(ε/RH)

EfL BTCA

(GPa)νfL ref, νfT ref

EfL ref

(GPa)νfL BTCA, νfT BTCA

3.6 0. 35 10-4 34.4 0.3 37.7 0.3

Constituent properties (measured or assumed)

Composite swelling (Vf = 0.40)

Reference 8.5 %

Cross-linked 4.8 %

Back calculation

δ

RH

ββββfibre = ?

15/10/2016

31

Results

Type β (ε/RH)

Reference 0.17

Cross-linked 0.10

Estimated from paper properties 0.22I. Kajanto & K. Niskanen. Paper Physics, Fapet(1998)

Determined by FEM simulations 0.12-0.13K. Persson, PhD thesis LTH (2000)

Estimated from wood samples 0.21 L. Wallström et al. Holz Roh Werkst. 53, 87 (1995)

Effect of anisotropy ratio

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

0.18

0.2

1 3 5 7 9

Anisotropy ratio E L/E T

Tra

ns

ve

rse

co

eff

icie

nt

of

hy

gro

ex

pa

ns

ion

Reference

BTCA

L

T

T

L

β

β=

E

Emust be assumed

15/10/2016

32

Concluding remarks

• Sensitivity to moisture

• Inverse modelling to identify transverse fibre hygroexpansion

coefficient

• Cell-wall cross-linking reduce hygroexpansion coefficient

significantly

• Simple method that could be used in ranking fibre modification

• Almgren et al. Polym. Compos. 31, 762 (2010), Joffre et al.

(2014, 2015)

Characterization of nanoscale fracture processes

Fengzhen Sun (PhD student)Kristofer Gamstedt

Collaboration wtih Hu Li (graduate student), Klaus Leifer (professor)

Applied Materials Science, Uppsala University

Henrik Lindberg (professor emeritus)Luleå University of Technology 64

15/10/2016

33

Background & motivation

Why the study of nanosectioning?

65

Science:

Material failure mechanism at nanoscale (brittle, ductile failure)

Local deformation (strain localization), size effect

Potential approach for experimental mechanics

Engineering:

Micro/nanofabrication (sectioning exhibates many advantages over

the MEMS-based process)

Ultra precision manufacturing of plastic components, e.g., optics, electronics

Uncut thickness: 30~200 nm

Minimum feed:5 nm

Force resolution (PCB 209A12) : <1 mN (Fc)

Diamond knife: edge radius ~ 7 nm

CCD eyepiece camera (S3CMOS05000KPA)

Signal amplifiers (PCB 480E09)

Data acquisition device (Agilent U2352A)

Experiment introductionExperimental setup

66

15/10/2016

34

• Material: Polymethyl Methacrylate (PMMA)

• Sectioning speed of 1 mm/s, rake angle of 40 degrees

• Nanosectioning at the thickness of 60, 85, 110, 140, 170 and 200 nm

• The widths and lengths of the chips were measured using the CCD eyepiece

• The sectioned surfaces were observed using the atomic force microscope (AFM:

Multimode 8, Burker)

Experiment introductionCritical thickness

Uncut

thickness

(nm)

Uncut width

(mm)

Uncut

length (mm)

Chip width

(mm)

Chip length

(mm)

Chip

thickness

(nm)

60 0.61 0.545 0.605 0.408 81

85 0.505 0.52 0.793 0.44 102

110 0.53 0.625 0.518 0.438 155

140 0.47 0.465 0.482 0.406 160

170 0.515 0.54 0.523 0.443 204

200 0.47 0.57 0.468 0.47 244

67

(b)

crackcrack

FIGURE Macroscopic image of chips sectioned at (a) 60 nm and (b) 140 nm.

(a)

0 40 80 120 160 200 2400

10

20

30

40

50

60

70

80

90

Fc/w Group 1

Ft/w Group 1

Fc/w Group 2

Ft/w Group 2

Fo

rce

pe

r w

idth

(N

/m)

Uncut thickness (nm)

Mechanical resultSectioning force

FIGURE Sectioning force at different uncut thickness68

15/10/2016

35

[ ] 32144444 344444 21

4434421fracture

friction

c

ndeformatioplastic

yc RwVV

FwVtVF +−

⋅−+⋅=)cos(

sinsin)sec()()( 0

αφ

φβαβγτ

[ ]

−

−+

−−⋅+−+−=

−

−⋅

−−−

)(cos

)sin(sin

)cos(

cos

)cos(

sin)tan(cot

sin

1

)(cos

1

)cos()cos(

sinsin1

2

22

αφ

αφφ

αφ

φ

αβ

βαφφ

φαφαφαβ

φβ

Z

Q

Rwt

Q

wF

y

c +

= 0

γτ

[ ])cos()cos(/sinsin1correctionfriction αφαβφβ −−−=Q

Mechanical resultAnalysis using Atkins’ thoery

)tan(cotstrainshear αφφγ −+=

αβ += )arctan(frictionCoulomb ct FF

Q, optimum theFind φ

wIQR / of candidates calculate I, intercerpt expt. From =

γτ wSQ / of candidates calculate S, slope expt. From y =

correct Z theestablish to

S/I expt. with theS/I calculated theCompare

y and Rcorrect theoff read correct Z, theFrom τ

ZChoosing

0/ tRZ yτ=

69

The fracture energy (specific surface work for surface formation) during

nanosectioning of PMMA is around 6 J/m2 , shear yield stress 110 MPa.

Theoretical surface energy: take the bond dissociation energy as 400 kJ and the concentration of

molecular chains as 1 chain per 0.2 nm2, giving 5×××× 1018 molecular chains m-2. (I. Ward) Therefore

it requires about 1.5 J to form 1 m2 of new surface.

40 60 80 100 120 140 160 180 200 2200

10

20

30

40

50

60

70

Pro

port

ion

(%

)

Uncut thickness (nm)

Plasticity

Friction

Fracture

PMMA PS PC DGEBA/DETA0

10

20

30

Exp

eri

me

nta

l se

ctio

nin

g w

ork

, (J

/m2)

Mechanical resultFracture energy, shear yield stress

FIGURE Fracture energy for polymersFIGURE Proportion of energy dissipation

70

15/10/2016

36

500 nm

(a) 60 nm (b) 85 nm (c) 110 nm

(d) 140 nm (e) 170 nm (f) 200 nm

AFM MorphologyPeriodic structures

FIGURE Height profile of the sectioned surface

71

(a) 85 nm (b) 200 nm

FIGURE Young’s modulus of the sectioned surfaces

AFM MorphologyYoung’s modulus

72

15/10/2016

37

• 60 nm: the sectioned surface is flat and smooth

• 85 nm: short and feeble wave-like features appears

• 110 nm: long and periodic wave-like structures (perpendicular to the

sectioning direction, bandwidth: 54 nm)

• 140, 170 and 200 nm: the periodicity becomes more pronounced

Uncut thickness (nm) Chip thickness (nm) Average spacing (nm)

60 81 --

85 102 --

110 155 150

140 160 215

170 204 209

200 244 383

AFM Morphology

Table 1 Average spacing between two adjecent structures

Periodicity of band

73

J. Black. On microscopic plastic instabilities in metal

machining chips, Metallurgical Transactions,1972

M Jiang, L Dai. Formation mechanism of lamellar chips

during machining of bulk metallic glass, Acta

Materialia, 2009

AFM MorphologyFormation of shear band

74

15/10/2016

38

AFM Morphology

(a) (b)

(c) (d)

Formation of periodic structures

FIGURE Schematic of the formation of periodic structures

sV

sV

FIGURE Schematic of the periodic structures

The influence of sectioning speed

76

15/10/2016

39

77

• Material: Polymethyl Methacrylate (PMMA)

• Sectioning depth of 85 nm, rake angle of 40 degrees

• Nanosectioning at the speed of 0.25, 0.5, 1.0 ,3.0, 10.0 mm/s

• Measurement of the widths and lengths of the chips by CCD eyepiece

• Morphology of the sectioned surfaces by atomic force microscope (AFM:

Multimode 8, Burker)

Experiment introductionCritical speed

0.5 mm/s0.25 mm/s

1.0 mm/s 3.0 mm/s 10 mm/s

AFM Morphology

15/10/2016

40

Computation

• Constitutive model of PMMA

• Infinitely long-line heat (Carslaw and Jaeger)

79

n

iykTHV

kTmT

1

0

1

)exp(sinh

2)0(

∆−+−= −

βε

εσσ

&

&

−=

t

r

tc

Q

tt

Mαπαρ

θ4

exp)4(

2

1

4 heat sources for the locally planar region

80

( ) ( )

( ) ( ) ( ) ( )t

rpduu

upduu

up

dyprt

qdyp

t

tq

t

i

pp

l

yii

t

l

yiM

s

i

s

i

αχ

χπλαπλ

θ

4,exp

1,exp

1

1621

2

2

3

2

0

2

0

0

00

0

=−=−=Ω

+Ω

−=

∫∫

∫∫∞∞

==

ωω

ωωαπλ

θα

duYV

dWq toi

i

ttVW

Wt

iM

−−

−= ∫∫

−

= 4exp

1

2exp

4

24/)(

00

pl2

−−+

−−

−

−=

∫

∫∫

ts

ts

tV

s

t

tV

W

t

siM

du

tV

du

t

t

YVdW

q

α

α

ωω

ωα

ωω

ωω

απλθ

4/

0

2

0

2

24/

00

0

maxpl

2

2

4exp

4

4exp

11

2exp

4

ωω

ωωαπλ

θα

duYV

dWq toittVW

t

iM

−−

−= ∫∫

−

4exp

1

2exp

4

24/)(

00

pl2

( ) ( )

( ) ( ) ( ) ( )

)(4,

4

,exp1

,',exp1

,'

,'16

,'2

12

h

21

2

321

2

21

021

2

0

0

021

0

0

2

1

2

1

tt

rp

t

rp

duuu

ppduuu

pp

dypprt

qdypp

t

tq

t

i

t

i

p

p

p

p

l

yi

t

l

yiM

s

i

s

i

−==

−=−=Ω

+Ω

−=

∫∫

∫∫ ==

αα

χ

χπλαπλ

θ

4 pre-heat sources for the bulk material

ComputationKomanduri model

…

…

…

…

…

15/10/2016

41

81

Computation

Computational speed for the onset of periodic structures is ≥ 2.6 mm/s,

experimental result is ≥ 1.0 mm/s

Conclusions nanosectioning

• The fracture energy for different polymers are quite close to the theoretical

specific surface energy.

• Critical conditions for Periodic and long wave-like structures:

uncut thickness ≥ 85 nm for sectioning speed of 1 mm/s

sectioning speed ≥ 1.0 mm/s for sectioning thickness of 85 nm

• The periodic wave-like structures stem from the plastic instability caused by

adiabatic shearing. The computational result based on Komanduri’s model agrees

with the experimental result.

• Unaddressed question: What is the response of CNT reinforcment on

nanosectioning?

82

15/10/2016

42

Thank you for your attention!

83