LE-GALLIC 2022 Archivage

Understanding of the wrinkles formation during the
coating functionalization of lightweight fibrous material :

a multiscale approach
Florian Le Gallic
To cite this version:

Florian Le Gallic. Understanding of the wrinkles formation during the coating functionalization of
lightweight fibrous material : a multiscale approach. Chemical and Process Engineering. Université
Grenoble Alpes [2020-..], 2022. English. �NNT : 2022GRALI091�. �tel-04048638�
HAL Id: tel-04048638

https://theses.hal.science/tel-04048638v1
Submitted on 28 Mar 2023
HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
THÈSE
Pour obtenir le grade de
DOCTEUR DE L’UNIVERSITÉ GRENOBLE ALPES

École doctorale : I-MEP2 - Ingénierie - Matériaux, Mécanique, Environnement, Energétique, Procédés,
Production
Spécialité : MEP : Mécanique des fluides Energétique, Procédés
Unité de recherche : Laboratoire de Génie des Procédés pour la Bioraffinerie, les Matériaux Bio-sourcés
et l'Impression Fonctionnelle
Compréhension des phénomènes physiques relatifs à l'apparition de

plis générés lors de la fonctionnalisation de surface par enduction
des matériaux fibreux de faible grammage : une approche multi-
échelles.
Understanding of the wrinkles formation during the coating
functionalization of lightweight fibrous material: a multiscale
approach.
Présentée par :
Florian LE GALLIC
Direction de thèse :
Raphaël PASSAS Directeur de thèse
Ingénieur HDR, Université Grenoble Alpes
Céline MARTIN Co-directrice de thèse
Maitre de Conférences, Grenoble INP
Rapporteurs :
Dominique Knittel
Professeur des Universités, Université de Strasbourg
Paulo Jorge Tavares Ferreira
Maitre de Conférence, Universidade de Coimbra
Thèse soutenue publiquement le 5 décembre 2022, devant le jury composé de :
Dominique Knittel Rapporteur
Professeur des Universités, Université de Strasbourg
Paulo Jorge Tavares Ferreira Rapporteur
Maitre de Conférences, Universidade de Coimbra
Pascale Balland
Professeure des Universités, Université Savoie Mont-Blanc Examinatrice
Ana Paula Costa
Maitre de Conférences, Universidade da Beira Interior Examinatrice
Jean-Claude Roux
Professeur des Universités, Grenoble-INP Président - Examinateur
Raphaël Passas
Ingénieur HDR, Grenoble-INP - Agefpi Directeur de thèse
Céline Martin
Maitre de Conférences, Grenoble-INP Co-directrice de thèse
Invités :
iii
The necessaries of life for man in this climate may, accurately

enough, be distributed under several heads of Food, Shelter,
Clothing, and Fuel; for not till we have secured these are we
prepared to entertain the true problems of life with freedom
and a prospect of success.
Henry David Thoreau in Walden, 1854
iv
Remerciements
L’ensemble de ce qui est présenté dans cette thèse a été réalisé au sein du laboratoire LGP2 de
Grenoble, qui a su m’accueillir, me supporter, m’encourager, m’héberger, me fournir l’ensemble
des dispositifs numériques et expérimentaux et auprès duquel j’ai eu l’occasion de faire de nom-
breuses et belles rencontres ces dernières années. Merci à l’IMEP-2, Tec 21 et la Region Rhône-
Alpes pour le financement.
Je tiens tout d’abord à remercier mes 2 encadrants Céline et Raphaël. Merci pour les conseils,
la gentillesse, le temps, la bienveillance, l’humanité et la compassion (au sens de « la force secrète
de son étymologie » comme l’a si bien dit Kundera). Merci de m’avoir formé à la recherche par
la recherche, d’avoir soutenu ma future carrière sans chercher à gonfler les vôtres et merci d’avoir
prouvé que la recherche académique peut permettre de s’élever ensemble et non l’un par rapport
à l’autre. Soyez les moteurs du changement en montrant à la nouvelle génération de chercheurs
l’intérêt du management bienveillant ; et puisse ces derniers oublier leur h-index dénué de sens,
pour produire une science de qualité, non de quantité, sociétale et accessible.
Les travaux menés et présentés dans ce manuscrit n’auraient pas pu être réalisés sans l’aide
des supports scientifiques et techniques du laboratoire. Ainsi merci à Maxime, Cécile, Karine
et Bertine pour le concours apporté à ce projet. Je pense également aux services techniques
et informatiques ainsi qu’à Stéphane et Mikaël. Des collaborations internes et externes au
laboratoire ont été réalisé. Merci donc à Jérémie Viguié et Denis Curtil du LGP2 pour leur aide
exceptionnelle concernant la partie expérimentale autour du pilote et de la mise en place de la
D.I.C ainsi que la mécanique du flambement. Pour les collaborations externes, je remercie Pierre
Caulet (Ahlstrom-Munksjö), Nicolas Jacques (IRDL Brest), Pierre Vacher (USMB), Axel Le
Nozahic (SWM), Romuald Leon (SWM), François Bonnel (CMTC), Georges Kapelsky (SIMAP)
et Mamadou Toungara (ENETP, Bamako) pour le concours apporter à ce projet. Enfin merci
aux étudiants que j’ai encadrés, notamment Cédric Dumont, Julien Dufieux, Elsa Rozzi et
Romain Dubuis
Enfin, il est bien plus facile de travailler dans un lieu et une ambiance propice au travail
: il est important pour moi de mentionner le bon esprit, l’optimisme et la motivation des
doctorants et post-doctorats et que j’ai eu la chance de côtoyer ces dernières années. Les
v
vi
longues discussions, les sympathiques conversations tenues durant les pauses cafés/thés et les
points de cultures générales ou les échanges scientifiques font partis intégrants des très bons
moments passés en votre compagnie. Se reconnaîtront, entre autres, Maxime Teil, Arnaud,
Loreleï, Mathieu, Bastien, Gabriel, João, Lorette, Amélie, Fanny, Khawla, Clémentine, Julia,
Emilien, Laura et Julie. J’ai une pensée toute particulière pour Claire, qui a finalement reconnu
avoir été mon encadrante de stage et qui est aujourd’hui une formidable amie ; ainsi que pour
Marlène, ayant commencé sa thèse en même temps que moi (à quelques mois près...) : merci
pour ta présence rassurante et réconfortante (sans t’en rendre compte), sans oublier ton humour
et tes conseils.
Au moment où ces lignes sont écrites, j’ai déjà eu l’opportunité de soutenir la thèse. C’était
un moment que j’attendais avec beaucoup d’impatience (non sans stress) pour la richesse des
échanges scientifiques. J’ai été servi et suis aujourd’hui ravi du moment passé où j’ai pu consid-
érer de nombreuses perspectives pour la suite du projet. Merci encore aux rapporteurs, Paulo
Ferreira et Dominique Knittel pour leur travail de relecture du manuscrit, leurs rapports et nos
discussions. Et merci également à mes trois examinateurs, Ana Paula Costa, Pascale Balland et
Jean-Claude Roux, pour les échanges notamment sur la partie mécanique de la thèse, qui m’ont
beaucoup intéressé.
Pour terminer, même s’ils sont moins concernés par les travaux présentés ici, j’aimerais re-
mercier l’ensemble de mes amis, de mes proches et de ma famille pour leur soutien. Je pense
notamment à mes amis de Bretagne Clémentine, Noémie, David, Gwen et Agathe, ainsi qu’à
mes amis de l’ENSTIB et du Moulin Simon, Lucas et Franck. J’ai une pensée toute particulière
pour Margaux qui a toujours été là pour me réconforter, m’aider et je souhaite à tout le monde
de rencontrer une aussi merveilleuse personne au moins une fois dans sa vie. Enfin, je remercie
mes parents, Marie-Hélène et Sylvain, qui m’ont toujours soutenu et encouragé pour aller dans
la voie qui me plaisait le plus.
Abbreviations and Notations
Latin letters
Lowercase letters
a Web span (m)
amn Amplitude (m)
ay Described transition between the first Newtonian plateau and the power law zone
bw,tan WSD y-intercept (MPa)
b Web width (m)
c Constant (m5 )
dc−IR Distance between the coating station and the IR dryer inlet (m)
de Drop diameter (m)
dparticle Particles diameter (m)
ds Drop diameter measured a a distance de of the drop bottom (m)
dA Elementary surface (m2 )
dW Elementary work to increase a surface area of dA (N · m)
exp Exponential function
f Constant
g Gravitational acceleration (9.81 m · s−2 )
hjurin Liquid height in a single in capillary, determined by Jurin’s law (m)
hs Measuring height (mm)
h Liquid height in a single in capillary (m)
i Number of wrinkles (used by Hashimoto)
kc Buckling coefficient
kjp Transverse buckling wave number
kn Constant used by Puntel
kodw Flow consistency index (P a · snodw )
kp Wave number
ks Inverse of a characteristic diameter (m−1 )
kW r Shape parameter Weibull law
vii
viii
lc,jp Characteristic length (m)

lcm Wrinkling wavelength
lres Resonance length (m)
lx,jp longitudinal buckling wavelength,
m Number of wrinkles in longitudinal direction of a tensile test
m0 Initial mass (g)
mbinder Mass of starch used as a binder (g)
mc(%) Moisture content of paper (%)
mf Final mass (g)
mparticles Mass of uncooked starch particles in the coating color (g)
mwa,eq Mass at equilibrium (g)
mwa (t) Mass of water absorbed into porous media (g)
mwater Mass of water (g)
n Number of wrinkles in transverse direction of a tensile test
na Severity of water absorption
nc Flow behaviour in Cross’s law
ne Factor of severity of the Young’s modulus decay
nkp Number of wrinkles used by Puntel
nodw Flow behaviour in Ostwaeld-de-Waele’s law
rcap Capillary radius (m)
rparticle Particles radius (m)
rw Wenzel’s roughness factor
t1 Transition time between inertial and visco-inertial regime (s)
t2 Transition time between visco-inertial and purely viscous regime (s)
t95 Time value at inflection point of the DPM curve (s)
tc Contact time between paper and water on the coating pilot (s)
t Time (s)
tf Paper thickness (m)
tmax Time value at the highest transmission value in DPM (s)
twet Wetting time (s)
wmn Out of plane displacement (m)
xc Filler ratio in paper
Capital letters
A, B and D Constants
Acm Wrinkling amplitude used by Cerda (m)
Ae Coefficient depending upon the intermolecular vibration (P a · s)
Ae0 , ae and be Constants
An Constant
Ap Cross-sectional area of pores (m2 )
ix
As,eq Plateau value of the aborptiveness (gwater · m−2

paper
As Water absorptiveness of paper (gwater · m−2
paper
AW r Amplitude in Weibull law (µm2 )
Diso Web bending stiffness for isotropic materials (N · m)
DT Tortuosity fractal dimension
Dxx Web bending stiffness in x-direction (N · m)
Ea Activation energy (J · mol−1 )
ECD,max Maximum value of Young’s modulus following CD (GPa)
ECD,min Minimum value of Young’s modulus following CD (GPa)
ECD Young’s modulus in CD (GPa)
EM D,max Maximum value of Young’s modulus following MD (GPa)
EM D,min Minimum value of Young’s modulus following MD (GPa)
EM D Young’s modulus in MD (GPa)
Ew,tan WSD tangent modulus (MPa)
Etan Elastic tangent modulus (GPa)
Ex Young’s modulus in x-direction (GPa)
Ez Young’s modulus z-direction (GPa)
E Young’s modulus for isotropic materials (GPa)
Fdw Strength measured during a Du Noüy ring or Wihelmy plate test (N)
Gxz,Baum Shear modulus in the xz plane (GPa)
G Grammage (g · m−2 )
H Shape factor used in pendant drop method
J Proportionality factor
Kc Penetration or spreading constant (s−1 )
Kp Strain plateau in WSD test (%)
Ldw Length considered during a Du Noüy ring or Wihelmy plate test (m)
M Energy emittance (W · m−2 )
Nx Force in x-direction per unit length (N · m−1 )
Nz Force in z-direction per unit length (N · m−1 )
R 8.314J · mol−1 · K −1
Re Reynolds number
R2 Coefficient of determination
Sbs Bending stiffness (mN · m)
SDP M Maximum slope measured in DPM (% of transmission ·s−1 )
Sp Cross-sectional area of the porous media (m2 )
S Dimensionless coefficient used in Richardson-Zaki law
Sdr Developped interfacial areas ratio (%)
Tcr Critical web tension (N · m−1 )
Theating Heating temperature (°C)
Tmeasured Measured temperature (°C)
x
Tx Tension in x-direction (N · m−1 )

T Fluid temperature (K)
VRichardson−Zaki Particle speed calculated using Richardson-Zaki law (m · s−1 )
Vsheart hinning Particle speed calculated using Stokes lax in a shear-thinning fluid (m · s−1 )
VStokes Particle speed calculated using Stokes law (m · s−1 )
Vweb Web speed (m · s−1 )
WA Adhesion work (or Dupré’s Energy) (J · m−2 or N · m−1 )
Wlbt W function of Lambert
Greek letters
Lowercase letters
β Web shape factor
δαr Relative error (%)
δp Shape factor proposed by Benevante (m)
εc Porosity (%)
εs Sensor emissivity
εx Strain in x-direction (%)
εz Strain in z-direction (%)
η Dynamic (or apparent) viscosity (P a · s)
η∞ Viscosity of the second Newtonian plateau(at high shear rates) (P a · s)
η0 Viscosity of the first Newtonian plateau(at low shear rates) (P a · s)
γlg Liquid surface tension (J · m−2 or N · m−1 )
d
γlg Dispersive part of the liquid surface tension (J · m−2 or N · m−1 )
p
γlg Polar part of the liquid surface tension (J · m−2 or N · m−1 )
γsg Surface Free Energy (SFE) of a solid (J · m−2 or N · m−1 )
d
γsg Dispersive part of the Surface Free Energy (SFE) of a solid (J · m−2 or N · m−1 )
p
γsg Polar part Surface Free Energy (SFE) of a solid (J · m−2 or N · m−1 )
γsl Solid-liquid interfacial energy (J · m−2 or N · m−1 )
γ̇ Shear rate (s−1 )
κs Measuring angle (°)
λa Mass constant to model paper disc stacking (g)
λc Cut-off value (mm)
−1
λc Shear rate at which shear thinning behaviour appears (s−1 )
λ−1
s Absorptiveness value at middle height (in gwater · m−2
paper )
λW r Scale parameter in Weibull law
µ Friction coefficient
νiso Poisson’s ratio for isotropic materials
νxz,Baum Poisson ratio calculated with Baum’s method
νxz Poisson’s ratio for transverse strain in the z-direction
xi
νzx Poisson’s ratio for transverse strain in the x-direction

ϖ Constant
ϕ Volume fraction in particle (%)
ϕA Fraction of surface of a chemical compound A
ϕB Fraction of surface of a chemical compound B
ϕp Inclination angle of a capillary tube with respect to the horizontal plane (°)
ϕf Angle between the tensile test direction and MD (°)
ρ Mass density of suspending fluid (kg · m−3 )
ρf iber Volumetric mass density of fiber (g · cm−3 )
ρf iler Volumetric mass density of filler (g · cm−3 )
σ Stephan-Boltzmann constant (5.6697.10−8 W · m−2 · K −4 )
σe,iso Normalized stress for isotropic materials (Pa)
σe,ortho Normalized stress for orthotropic materials (Pa)
σx,cr Critical tensile stress in x-direction (Pa)
σx Tensile test in x-direction (Pa)
σz,cr Critical compressive stress in z-direction (Pa)
σz Stress in z-direction (Pa)
τ Characteristic transition time between inertial and viscous regime (s)
τ Shear stress (Pa)
τ1 Time constant (s)
τ2 Time constant (s)
τa Time constant (s)
τcr Critical shear stress (Pa)
τd Time constant (s)
τp Correction factor of tortuosity
θc,A Contact angle of a compound A (°)
θc,B Contact angle of a compound B (°)
θc,i Initial contact angle (°)
θc,eq Contact angle at equilibrium (°)
θc Contact angle (°)
θc∗ Corrected contact angle for a rough/heterogeneous surface (°)
θcr Critical misalignment angle (°)
θW r Position parameter in Weibull law
ξ1 Anisotropy factor of paper
ξ2 Constant
Capital letters
∆ρ Difference fluid and particle volumetric mass density (kg · m−3 )
∆T Work done by forces acting in the middle plane of the plate (N · m)
∆tf Increase in the paper thickness (m)
xii
∆U Strain energy of bending (N · m)

∆ Constant
Φbinder Cooked starch mass fraction (wt%)
Φparticles Mass fraction of uncooked starch grains in the coating gel (wt%)
Πp Perimeter of pores, respectively (m)
Σr2 Sum of the squared deviations
Θt,0 Initial tensile stiffness (kN · m−1 )
Θt Tensile stiffness (kN · m−1 )
Contents
1 General Introduction 1
2 Literature review 5
2.1 Paper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.1.1 Paper origin and composition . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.1.2 Paper making process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.1.3 Structure and surface of paper . . . . . . . . . . . . . . . . . . . . . . . . 14
2.2 Sizing and coating process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.2.1 Coating and sizing color . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.2.2 Sizing and coating process . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.2.3 Web handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.3 Liquid absorption by paper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.3.1 Paper wettability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.3.2 Water absorption in paper . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.3.3 Rheology of coating colors . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
2.4 Buckling and wrinkling of paper . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
2.4.1 Tension buckling and fluting . . . . . . . . . . . . . . . . . . . . . . . . . 47
2.4.2 Shear wrinkling of a web in a roll-to-roll process . . . . . . . . . . . . . . 52
3 Materials and methods 57

3.1 Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
3.1.1 Papers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
3.1.2 Starch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.1.3 Chemicals used for contact angle measurement . . . . . . . . . . . . . . . 59
3.2 Coating color . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
3.2.1 Preparation of the starch gel . . . . . . . . . . . . . . . . . . . . . . . . . 60
3.2.2 Devices and methods for characterising the coating color . . . . . . . . . . 60
3.3 Devices and methods for characterising the paper . . . . . . . . . . . . . . . . . . 65
3.3.1 Room conditions of tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
3.3.2 Composition and structure . . . . . . . . . . . . . . . . . . . . . . . . . . 65
xiii
xiv CONTENTS
3.3.3 Mechanical properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

3.3.4 Water absorption and wettability . . . . . . . . . . . . . . . . . . . . . . . 68
3.3.5 Coating process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.3.6 Wrinkles and buckling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
3.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
4 Absorption of the coating color by papers 79

4.1 Properties of the coating color . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
4.1.1 Preparation of the coating color . . . . . . . . . . . . . . . . . . . . . . . . 80
4.1.2 Rheology of the starch gel . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
4.1.3 Sedimentation of starch . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
4.2 Study of the wettability and water absorption by papers . . . . . . . . . . . . . . 91
4.2.1 Study of the papers wettability . . . . . . . . . . . . . . . . . . . . . . . . 92
4.2.2 Papers water absorption analysis . . . . . . . . . . . . . . . . . . . . . . . 100
4.3 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
5 The wrinkling phenomena of paper 111

5.1 Observation of the wrinkles phenomena . . . . . . . . . . . . . . . . . . . . . . . 112
5.1.1 In-situ observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
5.1.2 Ex-situ observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
5.1.3 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
5.2 Modelling of the wrinkling phenomenon . . . . . . . . . . . . . . . . . . . . . . . 126
5.2.1 Paper mechanical characteristics . . . . . . . . . . . . . . . . . . . . . . . 126
5.2.2 Is it possible to forecast wrinkles with models? . . . . . . . . . . . . . . . 140
5.3 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
6 Study of the coating process on the laboratory pilot 147

6.1 Control of the temperature during the process . . . . . . . . . . . . . . . . . . . . 148
6.1.1 Factors influencing a non contact IR temperature measurement . . . . . . 148
6.1.2 Inhomogeneous temperature and sensor position on the web width . . . . 152
6.1.3 Following the process temperature . . . . . . . . . . . . . . . . . . . . . . 155
6.2 Process optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
6.2.1 Process regulation: web speed, tension and misalignment angle . . . . . . 158
6.2.2 Characteristics of finished products . . . . . . . . . . . . . . . . . . . . . . 167
6.3 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
7 General conclusion 175
References 179
A Coating pilot diagram 205
B Paper disc stacking : complementary results 207

CONTENTS xv
C Generation of speckle for tensile test used in D.I.C 209
D Out of plane displacement of sample during tensile test 211
E Mechanical properties of papers 215
F S.E.M picture of CPRAW and CPBS paper 217
G WSD: complementary data 219
H Determination of the elliptical shape of the intersection between a plane and

a cone 221
I French extended abstract 225

xvi CONTENTS
Chapter 1
General Introduction
Context
Since its creation, paper as a material has constantly evolved to meet society’s needs. Currently,
research focus in the reduction of grammages with constant mechanical characteristics and the
functionalization of papers in order to obtain specific surface properties (roughness, oil and
moisture barrier, oxygen and water permeability, etc.) for the reduction of the environmental
foot printing of the paper industry. These characteristics can be obtained by coating paper
surface with suitable aqueous coating color. The quality of the coating obtained depends on
the triplet: paper/coating color/coating process. Once the coated paper obtained, important
properties must be controlled: water absorption, basic weight, surface and hydro-expansion. For
the coating color, the rheological properties, dry matter content and chemical nature must be
considered. As for the process, the main parameters are the quantity of coating color transferred,
the travel time from coating to drying, the drying conditions, the web tension and the rollers
alignment.
When coating very lightweight papers, wrinkles may appear at the exit of the coating area
(Figure 1.1) which, after drying and depending on the nature of the coating color, may be
irreversible and may cause non-conformities and downgrading of the products.
Here, a coupling between water absorption, which reduces the mechanical properties of paper,
and buckling phenomena, possibly due to shear, misalignment and tension, has to be studied.
Figure 1.1: Conformed (left) and non-conformed (right) paper.
1
2 CHAPTER 1. GENERAL INTRODUCTION
Firstly, water coming from the coating sauce weakens the mechanical fibre network properties.
Secondly, the loading imposed by the web tension (used to tension the paper during the
process) may be non-uniform across the web. This may be due to roll misalignment. This non-
uniform distribution of the load could generate secondary compressive stresses perpendicular to
the paper transport direction and lead to an out-of-plane deformation: this is buckling. Then
passing to a web span to another it becomes winkle (on the roller). If the buckled paper passed
over the roll it will create creases. Buckling could also just be" tension buckling" if the buckling
criterion is low.
A question then immediately comes to mind: "what is a wrinkle?" Wrinkling is a most
common and costly web defect encountered in the manufacturing and converting of lightweight
grades of paper, film and textiles. It is so common, that a lot of synonyms existed: baggy lane,
creases, troughs, soft/hard wrinkles . . . In this study, the following terms will be used as follows
[1]:
→ A web trough is an out-of-plane deformation of the web that occur between rollers in
web-handling system (= web buckled in web span)
→ A baggy web is a web with one edge stretched and the other floating
→ A web wrinkles is an out-of-plane deformation of the web that is transferred over rollers
from one web span to the next (= web buckled on roller).
→ A web creases is permanent deformation from wrinkling
Figure 1.2 represented the evolution from troughs to creases.
Figure 1.2: From troughs to creases on the RK coating pilot of the LGP2 laboratory.
This project is taking place within the MatBio and BioChip teams,of the LGP2 laboratory,
in Grenoble (France). It is also a continuation of a collaboration with INSA Lyon, supported by
3
the LGP2 (Roll and Fun project, 2018) as the part of the development of a paper-based thermal
insulator. The targeted objective of this former project was to produce paper-based thermal
insulation continuously. However, the paper wrinkling issue after coating and drying remains.
It should also be noted that this theme of dimensional variations linked to the hydro-expansion
of fibrous supports and out-of-plane deformation is a recurrent problem to manufacture fibrous
products.
Objectives of this PhD thesis

Thus, the main objective of this PhD thesis is to fully understand the mechanisms underlying the
wrinkle appearance during the roll to roll coating process. The final goal is to limit or totally
avoid wrinkles appearance when depositing a functional layer on the surface of lightweight
papers. This project is also based a multi-scale study of the physical phenomena involved and
their fundamental understating/approach. The understanding and technological resolution of
this challenge can subsequently lead to the development of new functionalities for Roll to Roll
technology. To achieve the project goal, three major objectives will have to be achieved, each
being a results chapter, listed below and illustrated Figure 1.3.
Figure 1.3: Visual abstract presenting the manuscript organisation.
→ Chapter 4 - knowledge of the paper behaviour when coated : This first results
chapter will focus on understanding the water absorption phenomena of GHDS 35 paper.
The paper wettability will be studied, taking into account its roughness. Then the water
absorption will be analysed using 3 techniques: Cobb, PDA and "Paper discs stacking".
This chapter will also deal with the coating sauce, focusing on the starch cooking, rheology
and the sedimentation phenomena that can occur with starch particles.
→ Chapter 5 - understanding of the wrinkle formation mechanics throughout a

multi-scale approach: The second chapter of results will strongly deal with the wrin-
kles mechanics. Through static and dynamic observation of the wrinkling/buckling phe-
4 CHAPTER 1. GENERAL INTRODUCTION
nomenon, observations will be made during tensile tests by using Digital Image Correlation
(D.I.C) and tests on a coating pilot. The wrinkles produced on the latter will be charac-
terised a posteriori on macroscopic and microscopic scales. Models from the literature will
be used to better understand the phenomena.
→ Chapter 6 - instrumentation of the coating pilot and development of a driving

assistance tool: The last chapter of results will present the instrumentation of the coating
pilot in the first place. The study of the temperature sensor and its monitoring during the
process will be the focus of the first part of this chapter. The second part will focus on
the realisation of a control map tool, in order to better master the process. The papers
coated in the laboratory will also be presented, as well as their characteristics.
Theses 3 objectives (and chapters) will be preceded by a literature review (Chapter 2) and a
presentation of the materials and methods used in this project (Chapter 3).
Scientific valorisation and presentation

The work presented in this manuscript was presented during 5 conferences through posters, oral
presentations and one conference article:
→ Le Gallic & al. Highlighting of the factors influencing a non-contact infrared temperature
measurement during a paper coating process, 25th TECNICELPA, Coimbra (Portugal),
March 2021 (oral presentation)
→ Le Gallic & al. Study of the rheological properties of a starch-based coating color for the
production of a multilayer paper insulation, 7th EPNOE International congress, Nantes
(France), October 2021 (poster presentation)
→ Le Gallic & al. Control of wrinkling phenomenon during surface functionalization process:
example of coating in paper industry, CODEGEPRA 2021, Grenoble (France), November
2021 (poster and oral presentation)
→ Le Gallic & al. Approaching Young’s modulus changes during aqueous coating, CIADICYP
2022, Girona (Spain), June 2022 (oral presentation)
→ Le Gallic & al. Wrinkles formation and origins: from theory of web handling to coating
pilot scale experimentation, 17th Fundamental Research Symposium, Robinson College,
Cambridge (United Kingdom), September 2022 (oral presentation and conference article)
Chapter 2
Literature review
Contents
2.1 Paper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.1.1 Paper origin and composition . . . . . . . . . . . . . . . . . . . . . . . . 6
2.1.1.1 Cellulose and hemicellulose . . . . . . . . . . . . . . . . . . . . 9
2.1.1.2 Lignin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.1.1.3 Interaction between lignin and cellulose . . . . . . . . . . . . . 10
2.1.1.4 The others components of wood . . . . . . . . . . . . . . . . . 10
2.1.2 Paper making process . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.1.2.1 A short history of paper . . . . . . . . . . . . . . . . . . . . . . 10
2.1.2.2 Pulp manufacturing . . . . . . . . . . . . . . . . . . . . . . . . 11
2.1.2.3 The manufacture and processing of the sheet of paper and
cardboard . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.1.3 Structure and surface of paper . . . . . . . . . . . . . . . . . . . . . . . 14
2.2 Sizing and coating process . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.2.1 Coating and sizing color . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.2.1.1 Pigments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.2.1.2 Binders and co-binders . . . . . . . . . . . . . . . . . . . . . . 18
2.2.1.3 Additives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.2.1.4 Formulation and preparation of coating color . . . . . . . . . . 22
2.2.1.5 Sedimentation . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.2.2 Sizing and coating process . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.2.2.1 Coating color application on papers . . . . . . . . . . . . . . . 24
2.2.2.2 Metering systems . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.2.2.3 Drying systems . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.2.3 Web handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.2.3.1 Unwind and Rewind . . . . . . . . . . . . . . . . . . . . . . . . 27
2.2.3.2 Spreading devices . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.2.3.3 Guiding equipment . . . . . . . . . . . . . . . . . . . . . . . . 29
5
6 CHAPTER 2. LITERATURE REVIEW
2.2.3.4 Tension control and sensing . . . . . . . . . . . . . . . . . . . . 30

2.3 Liquid absorption by paper . . . . . . . . . . . . . . . . . . . . . . . . 31
2.3.1 Paper wettability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.3.1.1 Surface tension . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.3.1.2 Contact angle on homogeneous surfaces . . . . . . . . . . . . . 32
2.3.1.3 Contact angle on heterogeneous surfaces . . . . . . . . . . . . 33
2.3.1.4 Techniques to determine liquid surface tension . . . . . . . . . 34
2.3.2 Water absorption in paper . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.3.2.1 The Lucas Washburn law . . . . . . . . . . . . . . . . . . . . . 35
2.3.2.2 Limits and modifications of the Lucas-Washburn law . . . . . 36
2.3.2.3 Measure the penetration of liquid into paper . . . . . . . . . . 39
2.3.3 Rheology of coating colors . . . . . . . . . . . . . . . . . . . . . . . . . . 41
2.3.3.1 Fundamental rheology tools for this project . . . . . . . . . . . 42
2.3.3.2 Rheology of starch . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.3.3.3 Techniques for measuring rheological properties [94] [83] . . . . 46
2.4 Buckling and wrinkling of paper . . . . . . . . . . . . . . . . . . . . . 47
2.4.1 Tension buckling and fluting . . . . . . . . . . . . . . . . . . . . . . . . 47
2.4.1.1 Tension buckling from the mechanistic point of view . . . . . . 47
2.4.1.2 Tension buckling from the paper maker point of view . . . . . 51
2.4.2 Shear wrinkling of a web in a roll-to-roll process . . . . . . . . . . . . . 52
This chapter will be dedicated to a literature review of the different topics discussed in
this thesis. First, the paper as a material will be reviewed: origin, manufacture, structure and
surface. The coating processes will then be discussed, firstly on the coating color (pigments,
binders, formulation and sedimentation) before focussing on the coating techniques and web
handling that can be found during the coating process. The third part will deal with the
phenomena of paper wettability and the absorption of liquids, including water by paper. The
rheology, of starch in particular, will be discussed. Finally, the fourth and last part of this
chapter will focus on the buckling phenomena that can occur on webs, such as paper, through
two phenomena: tension buckling and shear wrinkling.
2.1 Paper
2.1.1 Paper origin and composition
Paper is first and foremost fibres from trees or annual plants, growing around the world in
fields or forests. So to begin this chapter, let’s take a quick look at the origins of paper in
terms of raw materials. Forest, according to FAO (Food and Agriculture Organization) is a land
spanning more than 0.5 hectares with trees higher than 5 meters and a canopy cover of more
than 10 percent, or trees able to reach these thresholds in situ. It does not include land that is
predominantly under agricultural or urban land use. Forests cover 31% of the total land area
2.1. PAPER 7
of the world. The Forests of Former USSR, South America (mainly Brazil), and Asia-Oceania
(mainly China) alone accounts for more than half of the total world’s forest area.
Although, large scale tree plantation and natural expansion has significantly reduced the
net loss of forests in the previous decade but deforestation rate is still alarming. Deforestation
in Africa, Oceania and Latin America was the main cause for the net loss of forests in the
previous decade due to deforestation are mainly social and economic politics. In Africa the
main cause of deforestation is the dependence of its population on wood as a source of energy.
Moreover, in South-East Asia, some companies does not hesitate to chop natural forest for paper
production and to replace the destroyed forest hectares with arable lands with palm trees for
the production of palm oil, leading to the destruction of biodiversity and single-wood species
forest (case of Asian Pulp and Paper companies in Indonesia (APP) [2][3].
From the Neolithic period onwards, humans have shaped the forest, its boundaries and its
landscape to cope with the increasing settlement of the population. Since the industrial revolu-
tion, everything has accelerated and since 1950, wood is exploited more and more throughout
the world. Global industrial groups appeared, polarising the resource and making it a major
strategic issue [4].
Wood is an oriented composite and plant-based material. Its fields of use are determined by
its behaviour when subjected to physical or chemical treatment but also by its quality which
is linked to its macroscopic, microscopic structure and its chemical constituents: cellulose (40-
50%), hemicellulose (20-25%) and lignin (25-30%) [5]. Looking more specifically at its structure,
wood can be defined as a set of plant cells that settle into an organized and dynamic complex
with a certain rigidity. The cell wall of these plant cells is composed of three different parts
(Figure 2.1) [6][7]:
Figure 2.1: From living cell to wood fibre [8].
→ The middle lamina, which corresponds to the cement between the cells. Its thickness is
from 0.5 to 1.5 µm. It is essentially made up of pectins and lignin;
→ The primary wall, which corresponds to the outer layer of the cell. It has a thickness of
0.1 µm. It is essentially made up of cellulose micro-fibrils, hemicellulose and pectins;
→ The secondary wall, which is the inner wall of the plant cell. This wall is subdivided into
3 layers: the outer layer S1, the middle layer S2 and the inner layer S3. Its constitution is
similar to the primary wall but it contains more lignin;
The wood constituents are distributed differently within the middle lamina, the primary wall
and the secondary wall. This distribution of the polymers present in the cell varies according to
the plant studied (Figure 2.2).
Figure 2.2: Distribution of principal chemical constituents in softwood cell wall layers [9].
The primary wall represents 1% of the relative thickness of the cell wall. The outer layer
is between 10 and 20% of the thickness. The middle layer S2 is the major part of the cell wall
with 70 to 90% of the wall thickness. Finally the inner layer S3 is between 2 and 8% of the
thickness. In the different walls representing the wood, fibre are made of cellulose micro-fibrils.
This micro-fibrils are not oriented in the same direction in function of the layer considered and
the wood species. To determine this Micro-Fibrils Angle (MFA), the Page method could be used
[10]. In general, the primary wall has randomly MFA. The S1 layer MFA is between 50 and 70°.
For the S2 layer, MFA is from 10 to 30° and for S3 layer from 60 to 90°. Note that a 0° MFA
mean a vertical fibril. In the paper industry different types of wood are used:
→ Softwood (pine, fir, spruce, etc.), which contains long cellulosic fibres (2 to 4 mm), which
give papers good mechanical properties.
→ - Hardwood (birch, beech, aspen, hornbeam, eucalyptus, etc.), which contains shorter
fibres (0.6 to 1.5 mm), which provide opacity and printability.
2.1. PAPER 9
2.1.1.1 Cellulose and hemicellulose
Cellulose is the most abundant polymer on earth. It is estimated that 50 billion tons of cellulose
are produced annually, 80% of which comes from wood. It is also present in annual plants such
as cotton, hemp or flax as well as in algae, fungi and bacteria. Cellulose consists of the repetition
(homopolymer) of a molecular unit, cellobiose, with a size of 1.025 nm. Two cellobiose units are
linked by a β1-4 bond. The degree of polymerization (DP) of cellulose (i.e. number of units)
is on average 10000 in wood, which corresponds to a fiber of about 5 µm. Cellobiose itself
consists of 2 units of β-D-glucopyranose (Figure 2.3). Hemicellulose is an amorphous polymer
Figure 2.3: Cellobiose (between brackets) and 2 β-D-glucopyranose units (which composed
cellobiose unit) [11].
with a lower DP and molecular weight than cellulose (DP<200). Unlike cellulose, hemicellulose
is a heteropolymer with ramifications. The main types of hemicelluloses are xylans which are
5-carbon sugars and gluco-mannans which are 6-carbon sugars. Hemicelluloses are different
between hardwood and conifers.
2.1.1.2 Lignin
Lignin is the second largest constituent of wood by weight: 20 to 30%. It is a non-linear,

amorphous, water-insoluble polymer of phenolic nature, i.e. it is based on a phenol group in the
constituent units of lignins. Lignins are heteropolymers whose constituents are not always the
same: there are therefore an infinite number of lignins.
Figure 2.4: The 3 lignin precursors for plants (from left to right): p-hydroxyphenyl unit,
guaiacyl unit and syringyl unit [12].
Indeed, there is a great variability due to the botanical origin, the essence and the nature of
the cells. It should be noted that there is more lignin in conifers than in hardwoods. Lignins
are mainly composed of 3 types of constituent units: p-hydroxyphenyl units (H unit), guaiacyl
units (G unit) and syringyl units (S unit) (Figure 2.4). These units also allow differentiation
between coniferous, hardwood and herbaceous plants.
2.1.1.3 Interaction between lignin and cellulose
Wood is a composite. Cellulose is the wood’s framework and plays the role of axial reinforcement:
it is a structural element of wood. Lignin is the matrix of this material. It cements the fibres
together, fixes and waterproofs the cell. It also participates in the mechanical maintenance as
a glue between the cells. It is a protective barrier against microbial and fungal attacks, against
physical or chemical attacks: hydrophobic and antioxidant properties thanks to the phenol
group. Hemicelluloses are intended to be an agent of compatibility between lignin and cellulose
[6][13].
2.1.1.4 The others components of wood
In the other components, 2 families are present: extractables (organic) and ashes (mineral).
These are secondary metabolites: they do not contribute to the growth, and therefore to the
mechanical properties of the wood. However, they are the ones that bring color, odour, repellent,
antifungal and antioxidant properties to the wood. The main families of extractives are lipids,
phenolic compounds and carbohydrates [11][14].
2.1.2 Paper making process
A sheet of paper or cardboard is made from cellulose fibres. These fibres are mainly extracted
from wood or coming from recycled paper and cardboard. They are of various origins: industrial
(processing scraps, industrial and commercial packaging, unsold newspapers, office products...)
and household (old newspapers, household packaging...). Auxiliary materials may be added
during the manufacture of paper (mineral fillers, additives, etc.) to improve the characteristics
of the paper [15].
2.1.2.1 A short history of paper
The first signs written by cavemen were written on stone, bone, then on wood, metal and clay.
It is with the Egyptians that papyrus, extracted from a reed, appeared. It competed with
parchment, developed in Minor Asia, from washed and sanded calf and goat skins. In the year
105 A.D., paper was discovered in China. Taso-Lun, Minister of Agriculture, designed a paper
pulp made from old rags, bark and nets reduced to pulp [16]. In France, it was not until the
13th century that the first archival documents written on paper appeared, but it was not until
the 14th century that the first French paper mills were built.
On January 18, 1799, Louis-Nicolas Robert (1761-1828), a young inspector at the Essonnes
paper mill, obtained a patent for his invention: the first continuous paper machine, which made
it possible to produce more paper at lower prices. In 1803, Didot Saint Léger, Bryan Donkin and
the Fourdrinier brothers operated the first English paper machine in Frogmore (Kent). Around
1850, the first machine for making multi-ply cardboard appeared. In 1856, Edward C. Haley
filed the first patent in England for corrugated paper used in particular for making hats. The
first French corrugating machine is installed in 1888 in Limousin[3][16].
2.1. PAPER 11
Today, paper technologies are constantly evolving, contributing to the development of indus-
try, the dissemination of knowledge and the packaging of consumer products. It is also, in France
and in 2018, 5.7 billion turnover (paper and cardboard), know-how and many jobs. As for the
production of paper, cardboard and pulp, it has grown steadily over the last 10 years. However,
the relocation of some companies to Asia or South America has transformed the landscape of
the paper industry. Between 2007 and 2017, China has become the worlds leading producer of
paper and cardboard (115766 thousand tons in 2017 against 73500 in 2007), thus overtaking the
USA (72279 thousand tons in 2017 against 83664 in 2007).
2.1.2.2 Pulp manufacturing
The manufacture of pulp consists of separating the cellulose fibres from wood and/or from paper
and cardboard to be recycled. This separation can be obtained either by mechanical means or
by the action of chemical reagents. Mechanical or thermo-mechanical pulps are obtained by
destructuring of the wood combined with the action of water and temperature. It is carried
out using grinding wheels for logs and defibrators for shavings. The heating and the mechanical
friction action allow the wood to be broken down into fibres, this is called wood defibration.
The pulp obtained is very heterogeneous (heaps of fibres, logs, individual fibres, fine elements).
Additional grading and refining steps are necessary to produce a pulp with homogeneous prop-
erties. Steam treatments prior to the defibering operation are carried out for thermo-mechanical
pulps (TMP). This softens the lignin in the wood, facilitating the separation of the fibres. In the
case of chemical-thermo-mechanical pulps (CTMP), chemical agents are used at temperatures
above 100°C up to 180 °C to further improve fibre separation [17], like sodium hydroxide for
example.
Chemical pulps are obtained by cooking wood at high temperatures (between 130 and 180°C
depending on the process) and under pressure in digesters in the presence of chemicals, called
cooking liquor which can be acidic or basic. Lignin will be preferentially removed although
cellulose and hemicellulose will also be degraded. Fibers are separated from each other in a
blow tank with limited mechanical action (Figure 2.5) [18]. There are different types of process
and therefore different types of pulp obtained: sulfite/bisulfite cooking, organosolv cooking,
alkaline cooking (kraft process), free sulphur cooking (soda process).
Mechanical pulps are called high yield pulps compared to chemical pulps. In addition, they
offer good opacity and good calendering properties. However, their mechanical characteristics
are weaker than those of chemical pulps. The color of the pastes manufactured by the processes
mentioned above is different. Depending on their use, they require more or less pronounced
bleaching. For example, printing and writing papers require a pulp bleaching step.
2.1.2.3 The manufacture and processing of the sheet of paper and cardboard
Mills that do not make their own pulp receive baled pulp from suppliers. Once the fibers
cooked, the first step in making the sheet of paper is to suspend the fibres in water. The
purpose of disintegration is to disperse and individualized the fibers. This step is facilitated by
the hydrophilic nature of cellulose. These fibres are then refined to obtain the properties required
Figure 2.5: An overview of chemical and papermaking process: from logs to customers [19].
2.1. PAPER 13
Figure 2.6: Example of double disks refiner.
for sheet formation and for the final paper product. This operation consists of conveying the
pulp between two discs (one fixed, one rotating), lined with blades (Figure 2.6)[20]. The fibres
are hydrated (swelling and softening of the fibre) and fibrillated (partial tearing of the outer wall
of the fibre). Refining results in an increase of specific fiber surface area and helps the creation
inter-fibre bonds.
This treatment also leads to a cutting of the cellulosic fibres. Finally, the pulp is cleaned,
screened and diluted to reduce its concentration. Fillers and chemicals may be added before it
arrives on the paper machine.
Before becoming a sheet of paper or cardboard, the pulp travels a long way through the
paper machine. The paper machine will be used to continuously produce a material sheet, with
a given width, which can be up to 10 m, and corresponding to a well-defined specification (weight,
thickness, given physical and optical properties). A paper machine achieves impressive speeds:
from 1800 m.min−1 (newsprint) to 2200 m · min−1 (special machines). Some machine builders
are even looking to increase speeds up to 2500 m.min−1 and several pilot paper machines are
already running at speeds above 2000 m.min−1 for light grades (newspaper). The paper machine
consists of three main parts: the wet end, the press section and the dry end (Figure 2.7).
In the wet section, fibres are distributed by the headbox at low consistency (2 to 10 g.L−1 )
on one wire (Fourdrinier formation) or between two wires where the sheet is formed and then
drained, first by gravity and then by vacuum. In the press section, the sheet is compressed be-
tween two cylinders covered with absorbent felt. It is then mechanically dewatered as much
as possible. A film-size press (or coating) section could be placed between pre and post-
dryingFinally, it is in the dryer section where the remaining water is thermally removed (evap-
oration). The sheet is dried against internally steam-heated cast iron drums. Calenders could
precede winders section. The dryness of the sheet is 20 % after the forming area and about 50
% after the press section. On leaving the dryer, the sheet has a moisture content of between 4
and 9% (by mass). It is stored in roll form [21][22]. It should be noted that a roll at the exit of
the machine is more than 40 km of paper and about 2.5 tons, if it is not more.
Before this final rewinding step, which is the end of the paper machine, the sheet can undergo
a number of additional in-line or out-line treatments: coating, smoothing, calendering, etc [16].
Figure 2.7: Scheme of a paper machine with size press section and calenders [23]
In parallel to the papermaking process, on-line measurements are carried out on the paper
material, most often with the aim of regulation: weight (by beta radiation), paper moisture
content (infrared or microwave), paper thickness (by laser, image analysis), filler rate (by X-
rays), color, whiteness, gloss, opacity and fluorescence (by optical methods).
2.1.3 Structure and surface of paper
Paper products could be divided in 4 categories [24]:
→ Graphic papers: these are the supports of communication, literature, knowledge and in-
formation.
→ Wrapping and packaging paper-cardboard: they protect and preserve large-consumer

goods because paper makes their packaging flexible and strong and has a high degree
of machinability, whereas corrugated paper used in the manufacture of corrugated card-
board.
→ Hygiene papers: also known as "tissue" is the most common paper used in the manufacture
of toilet paper, paper towels, handkerchiefs, diapers...
→ Industrial and special papers: different types of application exist such as fiduciary uses
(identity papers, bank notes, etc.), specific graphic uses (transfer papers, etc.), industrial
uses (laminated paper, abrasive papers, etc.), and other specific uses (cigarette papers,
filters, etc.).
2.1. PAPER 15
Figure 2.8: Possible pore structures in paper, from left to right : surface pore, connected pore
and cavity.
First thing to look at on this different kind of paper it is their surface structure. It is
often characterized by its roughness. The roughness of typical uncoated paper and paperboard
surfaces, when evaluated by profilometry, were found to be within the range of 3 to 5 µm. Lower
values could be found for coated paper[25]. The most commonly used roughness parameter is the
average distance between the peak tops and valley bottoms of the surface Ra . Paper roughness is
determined by fibre morphology, forming and calendering. Paper roughness may be categorized
into three length scales [26]:
→ Optical roughness: at length scales smaller than 1 µm, it is related to surface properties
of individual fibres) and to internal cracks of other solid materials;
→ Micro-roughness: at 1-100 µm, it is related to shapes and positions of fibres and fines in
the network structure;
→ Macro-roughness: at 100 µm-1 mm,is the result of paper formation;
A typical paper surface is porous. At the microscopic scale, there are spaces between the
fibers in the paper web. It makes paper a network with fibres of different sizes that are me-
chanically and chemically linked together. The spaces between the fibres are the paper pores.
Paper could also contain mineral fillers retained in the pores. Paper has a random and irregular
structure, with pores of different sizes and shapes that could be interconnected. The paper
porous structure is often described by a model of capillaries with a certain radius distribution.
Pores start to appear due to stiffness of fibres: it prevents them from bending that they
would always touch their nearest neighbours in the thickness direction [26]. Different types of
pores can be distinguished in the paper [27] as shown Figure 2.8:
→ Pores connected to one side of the paper only;
→ Pores connected to both sides of the paper. They may also have branches and intercon-
nections;
→ Pores not connected to the overall pore network : they form cavities. There are small
cavities within the fibres called intra-fibre pores. The volume occupied by these pores is
often negligible compared to the volume of the inter-fibre pores;
The pore network is involved with the fibrous paper structure [26]. Therefore, paper can
be described as a two-phase material: a filling phase and a solid phase (Bristow 1986). Paper
can be characterized by its porosity, specific surface area, permeability and tortuosity. Paper
porosity is the fraction of void in the material in relation to its total volume. In the case of
paper containing fillers, it is given by equation 2.1.
1 − xc
!
G xc
ε=1− + (2.1)
tf ρf iber ρf iller
where G is the paper grammage, tf the paper thickness, xc the filler ratio in paper, ρf iber the
volumetric mass density of fiber and ρf iller the volumetric mass density of filler.
The pore structure could be changed by different factors. Paper is compressible meaning
that its pore structure changes when it is compressed, sometimes irreversibly. Paper is hygro-
expansive. Thus, pores shape and size change when the paper moisture content changes, due to
the fibres swelling and the breaking of the inter-fibre hydrogen bonds when paper becomes wet.
The paper structure depends on all the papermaking unit operations: drainage, wet pressing,
drying and calendering [26]. The spaces among fibers in a paper web will be greatly affected by
the degree of refining of fibers before forming the web. The reduced porosity of paper resulting
from refining also can be partly attributed to generation of cellulosic fines, which tend to fill in
the spaces among the fibers of paper [25].
The paper apparent density varies according to the web thickness. The surface pores are
larger than the pores inside the web. The paper density varies from 200 kg · m−3 for common
filter paper (ε = 0.87) to 1400 kg · m−3 for glazed greaseproof paper (ε = 0.1). The fines affect
the thickness density distribution in the web. Paper is anisotropic in the plane and mostly in the
thickness. The paper pores are larger in the web plane than in thickness. There is non-uniform
fibre orientation distributions and also flocculation affecting porosity [26].
Different characteristics of paper depend of its porosity and pores structure. The specific
surface area is the surface area of the pores per unit mass. It depends on the pore size distribu-
tion. Permeability is the capacity of a porous medium to let fluids or gaz flow through it in a
given direction. Tortuosity is the ratio of the actual distance travelled by a fluid flowing in one
direction through the paper to the apparent length of the paper in that direction. Tortuosity
accounts for pore shape and inter-pore connections (Clennell 1997). Porosity and specific sur-
face area are determining parameters of the paper permeability. Porosity, pore size distribution
and pore tortuosity also control the liquid transport in paper [28]. Note that the paper poros-
ity considers the occluded pores although these do not participate in the fluid flow (gases or
liquids). On one hand, porosity and specific surface depend exclusively on the dimensions and
number of pores. They do not take into account the possible anisotropy of the porous medium.
This information is obtained experimentally with mercury intrusion porosimetry [29]. Recent
X-ray microtomographic techniques have also provided a way to non-destructively image the
paper structure in three dimensions [30]. Tortuosity and permeability, on the other hand, are
directional quantities. They depend on the direction in which the measurement is made. For
papers, the permeability in the web plane is greater than the permeability in the web thickness
[27].
2.2. SIZING AND COATING PROCESS 17
2.2 Sizing and coating process

"Coated papers were manufactured in the USA as early as the first decade of the 19th century.
The first coated paper was manufactured principally for labels by hand brushing the coating
colors onto paper sheets which were then hung on racks to dry" wrote R. N. Thompson in [31].
Today things have changed a lot in terms of the coating colors used and their formulations, the
coating processes and their speed and the use of coated papers. This section aims to provide
readers with the key elements to understand the coating process through 3 points: coating color
and their components, coating processes and finally web handling techniques.
2.2.1 Coating and sizing color
Aqueous coating for paper or board applications generally contain three distinctive groups of
materials besides water: pigments, binders and additives. Surface sizing aims to increase or
reduce liquids penetration into papers. It allows the deposition of 1 to 4 g · m−2 of dry matter
on paper. Pigment coating aims to improve the optical properties of paper surface and its topog-
raphy to improve its printing quality. The most important coated paper grades are lightweight
coated where the basis weight of the coated layers (both sides) is less than g · m−2 .
The materials found in coating and sizing as well as the techniques used are presented in the
following paragraphs using the following references [32][33][34][35][36].
2.2.1.1 Pigments
Kaolin clay - Kaolin clay is by far the largest coating pigment used today. It gives papers an
excellent surface finish and good printability. Kaolin clay can be divided into different groups.
Airfloat clays are used in filler applications and pre-coat. Delaminated clays are produced
by applying mechanical shear force to a clay slurry. Clay stacks are separated into individual
platelets. This lamellar structure is used in paper coating for its good coverage quality. Calcined
clays are made from hydrous clay exposed to very high temperatures (around 1050°C) in a
calciner. The plate-like structure of the clay becomes a voluminous and highly porous structure.
It improves the whiteness of kaolin but highly increases porosity and inks/varnishes penetration
and then reduces gloss.
Calcium Carbonate - This pigment, with rhombohedral shape, can be produced from quar-
ries (Ground Calcium Carbonate (GCC)) or by precipitation (Precipitated Calcium Carbonate
(PCC)). GCC is from crushed chalk, limestone or marble obtain by dry or wet process. PCC
is produced by reacting calcium hydroxide with carbon dioxide, and produce pigment of high
brightness and purity. Depending upon reaction conditions, particles of different shape and crys-
talline structure are produced: aragonite or calcite crystals (either scalenohedral or prismatic
shape). PCC is particularly interesting for its high whiteness and its low degree of yellowing.
Titanium Dioxide - This compound is extracted from Ilmenite mines and treated by me-
chanical crushing followed by sulphuric acid treatment. Titanium dioxide is then precipitated
and calcined. Another treatment is possible with chlorine in a furnace. Two abrasive crystalline
forms exist for titanium oxide: anatase and rutile. Titanium oxide has very good dispersion
properties, high whiteness with an excellent opacifying power, high brightness and blue white
shade. Titanium dioxide is used as a complement to others pigments like calcined clays, for
premium grades or with unbleached kraft. However, this compound is less and less used.
Talc - This is a natural pigment that does not require any special chemical treatment. It is
extracted from a quarry and ground to the desired particle size. Talc is used in coating processes
for its low hardness and its ability to reduce the abrasiveness of the pigment layers. Talc also
allows to obtain a matt coating finish while preserving the gloss of inks and varnishes. Talc
slurries have low dry content: around 50 to 60%.
Plastic pigments - They are synthetic spherical pigments obtained by emulsion polymerisa-
tion of vinyl monomers (styrene). They have several advantages: lowest density of all pigments
(density close to 1); high gloss, especially when high temperature calendering; good optical
properties; high micro porosity; almost no abrasiveness; particles do not form films and remain
globally individualised during their use.
2.2.1.2 Binders and co-binders
Binders are indispensable, although expensive. The main role of binders and co-binders is to
ensure cohesion between pigments and adhesion with substrate. Their choice and dosage have
direct impact on the dispersion rheology. The products most frequently encountered come from
three sources: vegetable (starch or soya proteins), animal (milk casein) or synthetic (latex or
polyvinyl alcohols) [35]. In a coating formulation, it represents between 8 and 18% of the layer
(by mass) [37][38]. However, they also have certain disadvantages depending on the binders and
the chosen process: they could foam, reduce opacity, printability and gloss. Moreover, they may
be incompatible with some pigments.
Starch - As an adhesive for paper coating, starch was introduced around 1900 [32]. Starch
is an abundant renewable resources, particularly used in the paper industry, especially in size-
press sauce and also in this project. Thus, the understanding and knowledge of this material is
essential in the present context and requires a more detailed explanation. The general references
used to present starch are the following one [34] [37] [39], [40] [41], [42], [43], [32], [33]. Starch is
a high molecular weight polymer. It is a mixture of two different chemical compounds: amylose
and amylopectin (Figure 2.9). Both are formed from anhydroglucose unit (AGU) which has 3
hydroxyl groups that help the formation of inter- and intra-molecular hydrogen bonds. It gives
starch its hydrophilic character, just like cellulose. In the case of starch, the AGU are α-(1-4)
and α-(1-6) linked. Amylose is a linear polymer with AGU linked by α-(1-4)-D-glycoside bonds,
slightly branched by α-(1-6) linkages and a DP around 1000. Amylopectin is a highly branched
polymer consisting of relatively short branches of α-D-(1-4) glycopyranose that are interlinked
by α-D-(1-6) –glucosidic linkages approximately every 20 to 30 glucose units. It has a DP of
100,000.
Figure 2.9: Amylose (top) and amylopectin (bottom) structure [40].
Starch comes from various plants: corn and potato are the most commonly used, as well
as wheat, rice and tapioca. It is granules with dimensions between 2 and 100 µm and it is
commercially available as a white dry powder. These granules have a multi-scale structure:
granule (2-100 µm), into which growth rings are found (120-500 nm) composed of blocklets
(20-50 nm) made of amorphous and crystalline lamellae (9 nm) containing amylopectin and
amylose chains (0.1-1 nm). These granules are also insoluble in cold water and find few uses
in industry without hydrothermal or thermo-mechanical treatments. Indeed, the polymerised
structure of starch presents a large number of secondary bonds, hydrogen bonds between the
different chains, making it particularly difficult to access. These hydrothermal treatments are
commonly used in the paper industry, food industry or in the case of adhesives. Nonetheless in
presence of water excess (water-starch ratios > 1.5:1 )[44] heated to at least 60°C, starch grains
can absorb water and thus swell, gelatinize, solubilize and turn into a gel when cooled. These 3
stages called sorption, gelatinization and retrogradation are explained below:
→ Sorption: this is explained by the attachment of water molecules on starch hydroxyl groups,
through hydrogen bonds. Water firstly breaks weak bonds between hydroxyl groups and
continues with the strongest bonds This dilate the macromolecular network: the starch
grains swell. Starch grain are capable of absorbing water (0.48–0.56 g/g dry starch) due
to their semi-amorphous nature. Hydration (at 25 °C) increases the grains diameter of
approximately 10% to 25%.
→ Gelatinization: water absorption increases with temperature and starch granules collapse.
Amylose and amylopectin solubilize to form a colloidal solution and the granular structure
disappears. When the gelatinization temperature is reached, the starch solution viscosity
is at its highest because the grains are swollen by the water absorption. Then, viscosity
decreases to a minimum value corresponding to the bursting of all the grains.
A feature called the Maltese cross is a characteristic of starch grains as illustrated Figure
2.10. This cross divides the granule into quadrants and it is linked to systematical arrange-
ment of macromolecules grains. This cross disappears during gelatinisation process and is
therefore a good indicator to check if starch is fully cooked. The presence/absence of the
Figure 2.10: Maltese cross observed within potato starch granules viewed under
plane-polarized light (x120) [42].
Maltese could be checked with polarized light microscopy. Moreover, gelatinization tran-
sitions of starches may be measured using Differential Scanning Calorimetry (DSC). Due
to the heterogeneous grain structure, starch thermal transitions occur over a temperature
range rather than at a defined temperature. Gelatinization generally occurs over a 10 to
15 °C temperature range [41]. For a native normal starch (i.e. amylose content between
20 and 30%), gelatinization temperature is found between 55 and 80 °C.
→ Retrogradation: while cooling and in absence of any starch chemical or thermal modifi-
cation, amylose and amylopectin macromolecules can reassociate by recreating hydrogen
bonds initially destroyed by hydration. This is retrogradation which is the crystallization
(or recrystallization) of hydrated starch. Amylose and amylopectin retrograde on very
different timescales: amylose crystallization being very rapid (minutes to hours) and amy-
lopectin recrystallization being very slow (days to weeks). For starch pastes with amylose
and amylopectin, retrogradation involves phase separation (formation of polymer-rich and
polymer-deficient regions). Retrogradation may be monitored via a multitude of techniques
(DSC, Fourier-transform infrared spectroscopy and viscometry, etc)[41].
Retrogradation should be avoided as much as possible in the paper industry. The gel
produced cannot be pumped, transported and deposited on paper due to its high viscosity.
It also leads to a decrease in the binding capacity of starch. Retrogradation is favoured
by conditions of low agitation, low temperature and pH < 7.5. Chemical or thermal
modifications can prevent or delay this phenomenon. The advantages of using starch in
paper making are its relatively low price, the gave stiffness to the coated layer and its good
water retention properties in coating formulation.
The operation which leads to the heating of the starch in the presence of water is called
"cooking" and takes place in specially designed thermostatically controlled tanks, either in a
jet cooker or in a batch cooker. The first one allows a more regular cooking than the second
one. Following the starch aqueous suspension viscosity during a specified heating-cooling cycle,
in the presence of shear, over a defined analysis time period is very useful. This is called the
pasting profile. Heating or cooling rates/temperatures, length of hold and shear rate may all
be varied. It is possible to mention some measured characteristics defined below and illustrated
Figure 2.11:
Figure 2.11: Pasting profile of a wheat starch granule suspension over the course of a heating
and cooling cycle [44].
→ Pasting time: the point in time, to a given heating rate, where enough starch grains have
swelled to measured an initial rise in viscosity (above the original ungelatinized granule
slurry). This value is reported in regard to the pasting temperature (i.e minimum tem-
perature needed to cook starch under given conditions) [44].
→ Peak viscosity: it results from a rapid swelling of granules upon gelatinization. Paste
composition at the peak is composed of both highly swollen and ruptured granules in
a continuous phase of solubilized starch. Peak viscosity is indicative of starch thickening
power. Then the starch paste viscosity will decrease due to rupturing of the fragile, swollen
grains until a trough viscosity is reached. This value is obtain during a hold period at
peak temperature (usually 95 °C) [44].
→ Trough viscosity (hot paste viscosity): is the solubilized starch continuous phase viscosity
containing dispersed grains remnants. The "breakdown" is the peak viscosity minus trough
viscosity. It indicates the paste relative stability at high temperature and shear conditions
[44].
→ Setback: is the final viscosity minus trough viscosity. It reflects the extent of amylose
recrystallization occurring within a starch paste [44].
Beyond pasting properties, more complete analysis of starch dispersion and/or gel properties
may be conducted using uni-axial compression testing and rheometry [36] [45]
Latex - They are aqueous dispersions of insoluble high molecular weight polymers obtained
by emulsion. Latexes are dispersions of spherical copolymer of controlled sizes particles (like
plastic pigments) between 0.1 and 0.2 µm, in water with surfactants added for stability. Three
classes of latex are used as coating binders: styrene-butadiene (SBR), styrene-butyl acrylate and
vinyl acetate homo and copolymer. Latexes do not provide water retention. Soluble binders or
soluble co-binders are usually associated with latexes. SBR is the dominant synthetic binder
used in paper coating today. It was developed during World War II to replace natural rubber.
It is a tough and strong binder. Depending on the styrene butadiene ratio, different glass
transition temperatures are obtain resulting in different properties and behaviour during coating
[34].Other latexes are also used: PVAc (Polyvinyl acetate), acrylic latex and PVA (Polyvinyl
alcohol) [36][45].
Co-binders - The role of co-binders is to adjust the coating rheological properties and its
water retention. The co-binders are also used as optical brightener carriers while having a certain
binding power. The most commonly used co-binders are carboxymethyl cellulose (CMC) and
starch; however, soy protein and polyvinyl alcohol can also be found [36].
Water - It represents 35 to 40 % of coating color mass. Being the vector of all others elements,
it is a very important constituent of the coating color. Water should be evaporated during
drying to let the coating layer being fixed. But water could also lead to different issues during
the process: local decreasing of the mechanical properties, curling or wrinkles [38].
2.2.1.3 Additives
Coating additives are very common in paper coating processes today [34]. Dispersant are used
to make easier pigments dispersion. They could be organic (sodium polyacrylate, the most
used) or inorganic (sodium hexameta phosphate). Concerning the viscosity modifiers, they are
thickeners (caseinates, CMC, Hydroxyethyl cellulose (HEC), high viscosity starch) or thinners
like urea. The water-holding agents, like CMC and some polyacrylates, are used to prevent
water in the coating color from penetrating paper too quickly. The lubricants are very useful to
increase the production tools life and to improve calendering ability by reducing the abrasiveness
of the pigment layer. Calcium stearate is mainly used. Biocides or biostats, like inorganic boron
and organo-sulfur molecules, are unavoidable in coating processes to avoid fermentation and
bacteria attack. Concerning the pH adjusters, soda ash and ammonia are commonly found is
the industry. Repellents are useful to release paper and the most common are silicon compounds.
Insolubilizers are used to prevent certain coated products from re-solubilising when in contact
with water after coating (in offset printing, for example). This is particularly the case for starch
and casein. Glyoxal-based resins are commonly used. Optical brighteners are often used in
premium paper grade. They make the coated paper look very white while using stilbene or
benzophenone derivates. Dyes are also found to give specific shade to the coated paper. Finally
foam control agents are found in coating formulation.
2.2.1.4 Formulation and preparation of coating color
The workshop for preparing a coating color is called a "kitchen". Many of the preparation
systems are automated, making the operators’ task easier. The main components of a kitchen
for the preparation of coating color are: storage tanks for the different products, mixers mounted
on load cells and cookers. All of this is connected by transport and dosing systems (pipes and
pumps) [35].
Whether dry pigments or slurries are used, the order of preparation remains the same: the
pigments are first dispersed in water with the addition of dispersant to minimise caking. The
binder is then slowly added to avoid rheological shocks, while adapting the stirring speed to
avoid binder degradation. The pH is controlled all along the preparation to avoid dispersion
problems for example. Finally, the additives are added to the mixture, before it is transported
to the coater feed tank. Once the coating color is obtained, different properties are measured to
ensure optimal quality: dry matter content, viscosity, pH and water retention. It is sometimes
interesting to know some other properties of the coating color such as surface tension, fluid
dynamic, elongational viscosity and dynamic surface tension [36].
2.2.1.5 Sedimentation
After formulating a coating color, stability over time is required. A suspension is a heterogeneous
mixture of solid objects in a liquid phase. The instability of a suspension could be explain by
gravitational attraction which leads to sedimentation of the solid object in the liquid phase.
The sedimentation of an isolated spherical particle in an infinite Newtonian fluid is known
analytically. A particle speed can be described using the Stokes law (equation 2.2), when the
Reynolds number is small Re << 1 (equation 2.3).
d2particle · g · ∆ρ
VStokes = (2.2)
18 · η
and
ρ · VStokes · dparticle
Re = (2.3)
η
If the particles number increases sufficiently, it is no longer possible to describe the interactions
between all particles. Thus, for high volume fractions in particles (> 5%), Richardson-Zaki [46]
law is used and the sedimentation speed as shown equation 2.4 with S = 4.65 + 19.5d/D for a
Reynolds number below 0.2.
VRichardson−Zaki = VStokes (1 − ϕ)S (2.4)
In the case of a non-Newtonian fluid, the resolution of the sedimentation problem is more
complex. Thus, a simple approach is to introduce the characteristic shear rate created by a
particle with a given speed. Using Stokes velocity expression combined with Ostwald-de-Waele’s
law (described later in the document equation2.29), the particle velocity is described equation
2.5 [47].
2 · g · ∆ρ 1/n
n
n
Vshear−thinning = · (0.5rparticle ) n+1 (2.5)
9 · kodw
2.2.2 Sizing and coating process
Coating devices are generally referred as "coaters". They are located off or on machine. A
coating station has 3 functions: to deposit the coating color on the paper, to remove the coating
color excess on the paper and to spread the coating. However, new technologies allows to put
the exact amount directly on the paper like curtain coating. In this section, the coating delivery
systems will be described first, before focusing on the dosing system [48].
2.2.2.1 Coating color application on papers
The main systems used to deposit the coating color on paper are the following ones and are
presented Figure 2.12:
Roller applicator - It brings the excess coating color on the paper to be coated. It generally
soaks in a tank where the coating color is located. Different parameters can be adjusted: height
of the coating color in the tank, roller speed, and its distance from a counter roller. In order to
avoid evaporation or thickening of the coating color, the roller can be cooled or heated. Above
1250m · min−1 , problems with coating color splashing may occur. Note that dwell times (i.e
transportation time) are long in roll applicator system [49] [36].
Short dwell - This process corresponds to an injection of the coating color tangentially to
the coating support and very close to the coating dosing system (generally a blade, described
later 2.2.2.2). To dose the coating weight, the blade is mounted on a system of micrometric
screws over its entire width, allowing a localized adjustment of the deposited coating color. This
type of system has several advantages. Lower excess coating color is deposited with this system
and allows the blade or the dosing rod to work at lower pressures, improving runnability. The
contact time between paper and excess coating color is minimized. The paper is therefore less
weakened by the rewetting which limits breakage. The coating color penetration in the web is
also limited. However, at high speeds, vortexes appear in the coating color, before the blade,
leading to coating weight heterogeneity on the paper web. In addition, this system is more
sensitive to air bubbles and needs a deaeration device [36].
Curtain coating - Another system developed is the curtain coating. It is a relatively old
coating process, dating from the end of the 19th century. It was first used to coat candies with
a vertical slot. In the second part of the 20th century, the slide process was developed to apply
simultaneously up to ten photosensitive layers in a single curtain on photographic films and
papers. The principle is simple: a vertical plane flow is created with the coating color. The
latter is deposited by gravity on the horizontal coating support. The curtain must be stable.
Then, flow control is needed as well as viscosity, surface tension control and the absence of air
bubbles. This process produces layers with excellent regularity and allows the application of
several layers simultaneously. [36].
(a) Roller applicator[11] (b) Short dwell [11] (c) Curtain coating [50]
Figure 2.12: Three different way to deposit coating color on paper.
Figure 2.13: A size press system [50].
Sizing press - Surface sizing is a critical unit operation which has significant effects on coater
runnability and on the quality of the final coated web. Starch films is typically used to fill in
the surface voids in the web, reducing the pore sizes and therefore the fluids penetration rate.
This process is made with size press. During this sizing operation the objective is to flood a
nip entrance with sizing solution. The paper will absorb some of the solution and the excess
is removed in the nip. The overflow solution is collected in pans below the size press and is
recirculated to the nip. Three mains size press configurations are possible: vertical, inclined,
horizontal as illustrated Figure 2.13 [34][36].
2.2.2.2 Metering systems
Blades - This is the most developed system and the most installed on industrial coaters
(Figure 2.14a). The blade dosing system allows speeds between 1800 and 2500m · min−1 with
high mass concentration colors (60 to 65%) [51]. To regulate the coating color, the blade pressure
on web is adjusted by pneumatic or mechanical systems. Two types of blades are commonly
found: bevelled blade and bent-blade. Bevelled blade is made of special hardened steel, with
a thickness between 0.25 and 0.65 mm. It is mounted in such a way that only blade end is in
contact with the paper surface. These blades are beveled with a grinding angle between 15 and
55°. It allows to deposit between 5 and 13 g · m−2 of dry coating color on the web. Bent-blade
is mounted in S-shape by a double stop system so as to present one of its faces in flat support.
It is not bevelled but their tips are coated with ceramic or metal alloy layer to increase their
abrasion resistance. Bent-blade allows to deposit between 10 and 25 g · m−2 . The blade support
(a) Blade coating [11] (b) Rod coating [11] (c) Air knife [11]
Figure 2.14: Three different way to meter coating color on paper.
is generally a mechanical system [52][36].
Rod coater - The Mayer Coating Company of Rochester (NY) founded in 1905 by Charles
W Mayer was the company which developed this system. Rod coaters are more commonly used
for pre-coating. (Figure 2.14b) The rod coater rotation is opposite to the web direction [34] and
its speed is slow (10 to 30 revolutions per minute). The rod diameter is between 10 and 20 mm.
A numbering of the Mayer rods makes it possible to know the mass of coating color deposited.
This numbering is a function of the winding (this point will be presented later Table 3.5) With
this system, it is necessary to avoid too high web-rod contact pressures otherwise the coating
surface is quickly degraded. Rods can be smooth or threaded (by winding), it is also important
to adapt the coating color viscosity otherwise pigment particles can clog the rod. In the case
of Mayer rod, it is the wire diameter on the rod that controls the final weights. There is long
dwell time in these system. [36].
Air knife - There is no contact between the web and the dosing system (Figure 2.14c). The
layer thickness obtained is more regular. The coating amount is control by the air pressure, the
nozzle gap and the distance from web. However, mastering such a system can quickly become
tricky. Indeed, projections of drops occur around the air knife, generating clogging. Moreover,
there is also foam formation. This requires regular maintenance of the system. The coating
colors are less concentrated (and therefore less viscous) than in the case of metal blade coatings
[36].
2.2.2.3 Drying systems
Drying is a delicate operation that can lead to quality issues on the final product. Drying
consists in removing the water contained in the coating or any other solvent, and the water
that has penetrated in the paper. To dry paper, conduction (impingement drying), convection
(air float drying and air foil drying) and radiation (infrared drying) are used. Impingement
drying use a hood with a controlled hot air circulation and a tunnel with idler rolls that could
be heated. In the hood, air blows trough nozzles onto the web supported by rolls. Air speed
and the temperature control the drying. The temperature control are made at the hood inlet
to avoid danger linked to over-heating [52].
Air float and foil dryers are built in way that air supports the web. With a two sides drying,
the foil and float dryers are more effective and smaller compared to impingement dryer . This
is mainly found with high and large machine.
The infra-red (IR) dryers are mainly use to warm-up the web before the air dryers. This is
due to the property of IR to go deeper into coated material. It is a non contact drying system,
with a good heat "penetration" and a higher heat transfer. It takes less place than other system,
while improving quality and having a higher energy efficiency (45 to 55 %). IR dryers use either
gas or electricity to produced the IR with wavelength from 2 to 5 µm . The IR lamps are used
with reflectors to minimize energy loss. Nonetheless, it is only really worth it with water based
coating [52], [53].
2.2.3 Web handling
Coating is an element of what is called "web converting". Web converting is the fact of taking a
web and transformed it irremediably to confer it new property. Contrary to web converting, web
handling does not seek to modify web. Web handling seeks to preserve all the properties of web.
The problems encountered by web handling specialists are mostly geometric: for example a flat
web that becomes wrinkled and unusable. In this part, some aspects of web handling will be
discussed. The references used in this part are: [54][55], [56] [57][58] . For further information,
the reader is invited to consult David Roisum’s books on the subject: The mechanics of web
handling [54] and The mechanics of rollers [55].
2.2.3.1 Unwind and Rewind
Unwinder and rewinder have some common functions. They must support the paper reel during
the process. To do so roll spool, expanding shaft or expanding cores could be used. In these
devices mechanical or pneumatic systems perform the expansion of shafts. Unwinders control
the required web tension (must remain constant) with mechanical brake or motorized brakes
and can guide the web.
2.2.3.2 Spreading devices
Spreaders are required on most web processes to prevent or remove wrinkles, to widen a web
or to separate slit during winding. It exists a wild diversity of spreading system. One of the
most important things in web handling is the "Normal entry law" [59] "A web will seek to enter
a downstream roller at a right angle to the roller’s axis at the point of first contact. If the web
is not entering at right angles, it will move sideways at an ever decreasing rate". This law is
valid only if there is no slippage between web and rollers and describes web behaviour at guides,
winders and misaligned rollers. Keeping this law in mind, three different spreading mechanism
can be presented: the concaved spreader roller, the bowed spreader roller, the D-bar spreaders.
Other way to spread web exist: dual bowed roller spreader, pos-Z spreader, expander roller,
edge pull web stretchers, compliant cover rollers, grooved spreader roller,. . . but will not be
presented in this document.
(a) Concave spreader roller (b) Bowed spreader roller (c) D-bar spreader
Figure 2.15: Three different way to spread web on roll to roll system [60].
Concave spreader roller - This spreader is a roller whose diameter at the ends is slightly
greater than at the center (cf. Figure 2.15a. In its ideal configuration, the diameter profile of
the roller is cut as an arc of a circle. Notice that spreading effects can appear on any roller,
due to very small diameter changes. Concave roller has constant rotational speed (rpm) on
its width. However, it has a varying surface speed (m.min−1 ) depending of roller diameter at
every position. Thus, concave roller surface speeds is higher at the extremities than at the roller
center. This surface speed difference lead to ingoing web tension distribution that is shaped
similarly to the speed profile, leading to spreading.
Bowed spreader roller- The bowed spreader has a curved stationary axle upon which a
rotating sleeve is mounted in bearing sets (cf. Figure 2.15b). The web center travels exactly
in MD (Machine Direction), while edges travel perpendicular to the roller axis at the edges.
The MD velocity distribution is maximum at the machine centreline and decreases toward the
edges, which creates an outward rotation of the web. The CD (Cross Direction) velocity is zero
at the centreline and increases as it progresses to the edges, which causes an outward lateral
displacement of the web and thus spreading.
D-bar spreaders- It takes its name from the characteristic “D” cross-section of the bar as
illustrated Figure 2.15c. D-bar’s shape can be adjusted by intermediate jacks. With this adjust-
ment, local baggy or wrinkled spot can be spread. The problem, however, is that the bar soon
ends up looking like a snake after several adjustments, which affects spreading uniformity.
"Free” spreader - A “free” spreader uses Poisson’s contraction. As a web is stretched under
increasing tension, it necks-down in width. This decrease in width may translate into flatter
webs or better slit separation. However, increasing tensions will increase the propensity to web
breaks. Moreover, when tension is increased in some span, it becomes difficult to reduce tension
further downstream without causing wrinkle.
Flattening - Usually all that is need to flatten a web is to lightly wrap a large diameter roller.
Caution must be exercised because loss of traction also means a certain loss of web handling
control. However, if often is possible to maintain tension and speed match on a flattening roller
yet still allow wrinkle to slide outward. In general, wrinkles will tend to begin on the most
highly wrapped rollers in a system.
(a) Displacement guide (b) Steering guide
(c) Unwind guide (d) Rewing guide
Figure 2.16: The four guiding technologies [56] [56].
2.2.3.3 Guiding equipment
Lateral misalignment have a lot of different causes: material properties, roller misalignments,
unmaintained machines, environmental conditions and processing conditions. It leads to tele-
scoped rolls, badly wound rolls or web thickness variations. Moreover changes in roller drag and
air entrainment between the web and roller can also cause lateral misalignment. So, web guiding
help to improve the process runnability by deliberately changing rotation axis of the guide roller
and conduct the web to different positions.
Web guides are mainly electro-mechanical devices that automatically control the web po-
sition. There are 3 main components in web guiding system : sensors, controller and guide
mechanism. Sensors are used to spot the current web position. The type of sensor used is
usually based on the web type and process. The controller reads the sensor information and
determines the required correction for the guide mechanism. The controller then commands the
actuator to move the web material to the desired position. To guide a web, there are different
techniques presented below.
Web edge guiding is usually referred to guiding on a single web edge. This is the most
common type of web guiding. This method is suitable when the web width does not change
during and between runs. Center guiding is used when the web edges are uneven or if the web
width changes during or between runs. The center line of the web is guided using two sensors,
one on each web edge. Line guiding may be used on web with a printed line, using camera based
solution to guide the web. Contrast guiding is similar to line guiding. The web may be guided
based on a contrasting feature on the web. A printed contrast or a contrast due to a coating
process can be used as the reference for contrast guiding.
Depending on the location where the guiding is required, four different types of guiding
technologies are available: displacement guides, steering guides, unwind guides and rewind guides
as illustrated Figure 2.16. Displacement guides is the most common web guides used in industry.
It includes two rollers on the web guide mechanism that rotate. It enables web displacement
with minimal bending stresses on the web. This system is very simple and compact. Nonetheless
they are some situations where displacement guides cannot be used: after long span, locations
with web wrap constraints and installation with space constraints. To overcome this problem,
another steering guides are typically used. The system is made of one or two rollers that travel
over an arc. The steering guide bends the web in the entry span to create the lateral motion. The
steering depends on different factors: web speed, web stiffness, entry span length,... Moreover,
the installation of a steering guide requires a lot of attention to prevent web breakage.
Specific guiding system are used in unwind and rewind. Unwind guides are a type of terminal
web guides that align the web fed into the machine. Typically a carriage or a side lay carrying
the entire web roll is moved in CD. The carriage is typically mounted on high precision linear
bearings with linear actuators that can move the entire system (carriage + roll). An idle roller
is also used to stabilize the web to avoid vertical displacement during unwinding caused by the
roll diameter changes. Rewind guides are used at the end of a roll to roll system. The web
wound on the roller is guided to avoid web telescoping. Rewind guides are similar to unwind
guides. The rewind guide chases the web to align the different layers of the wound roll.
Other important parameters have also to be taken in count as roller diameter and roller
smoothness that play a critical part in web guiding. Moreover, a proper installation geometry
is needed focusing on three important parameters: entry spans, exit span and web wrap around
rollers.
2.2.3.4 Tension control and sensing
Different tension control equipments can be used. They must be as close as possible to the point
it affects and different web tensions are necessary in a process. The tension must be controlled
after unwinding, just before laminating or coating operation, in the drying tunnel and at the
rewind area to optimize the web tension control. [56].
Tension control To control and regulate tension, two possibilities exist: open loop or closed
loop control. Open loop tension control is selected when the cost of a tension control is a
consideration. An open loop tension control system has no actual tension feedback. It is based
only on a tension set point without feedback or influence from actual conditions and cannot
respond to tension disturbances. Tension is regulated using torque control in the drive system.
Closed loop tension control incorporated a force measuring sensor or a dancer roll system to
control tension. Tension is regulated using speed or torque control in the driving system [56].
2.3. LIQUID ABSORPTION BY PAPER 31
There are two options to measure tension. The first one is the load cells; as illustrated
Figure 2.17a, which is specifically designed for web tension measurement and transform it into
a proportional electrical signal. The majority of load cell designs use strain gauges as sensing
element and are mounted in pairs, supporting an idler roll. It forms a full Wheatstone bridge.
The strain gauges are attached to a mechanical structure that deflects when a tension force is
applied. The other option is the dancer roll as shown Figure 2.17b. It is a moving preloaded
idler arm. Its force and positioning is derived from web tension. Pivoting dancers are the most
common. Dancer roll system requires a position sensor for the dancer position control like optical
encoders. Friction should be minimise to avoid correction delay so low friction bearing can be
used and dancer mass should be as low as possible to minimize inertia effect. Dancer rolls are
also used in accumulators systems because of their storage capability.
(a) Load cell (b) Dancer roll system
Figure 2.17: Tension sensing system [58].
2.3 Liquid absorption by paper

Now that the coating processes have been presented, it is necessary to look at the physical
phenomena that can occur during these processes. The first that may come to mind is the
absorption of water that may be in the coating color. Thus, this section will look at the liquid
absorption of paper.
2.3.1 Paper wettability

When a coating color is applied to paper, it does not immediately penetrate the fibrous network.
There is a first contact because the coating color wets the paper. It is on this first point that
this sub-section will concentrate: the paper wettability.
2.3.1.1 Surface tension
The cohesive forces between the molecules in a liquid are responsible of the surface tension.
Molecules in a liquid are subject to equal attractive forces in all directions, whereas those on
the surface of the liquid have less interactions (Figure 2.18. This lack of bonding leads to liquid
surface tension phenomena. An energy model helps to obtain an equation describing surface
Figure 2.18: Surface tension appearing at the interface of gas (G) and liquid (L).
tension: "when increasing the free surface area by an elementary amount, molecules must be
transferred to the surface. The elementary work dW to increase the surface area of dA" is then
written:
dW = γlg · dA (2.6)
With γlg the liquid surface tension (J · m−2 or N · m−1 ).
2.3.1.2 Contact angle on homogeneous surfaces
A drop of liquid deposited on a perfectly smooth horizontal solid surface, chemically homoge-
neous, adopts at equilibrium a conformation depending on the interactions between the liquid,
the solid and the gas phases. At equilibrium the system minimises its energy and in the case
of partial wetting, the drop takes a spherical shape if the liquid volume is small. In this case,
Young’s equation (equation 2.7) expresses the contact angle formed at the intersection of the
three phases of the system:
γsg − γsl
cos(θc ) = (2.7)
γlg
with γsg the surface free energy (SFE) of the solid in equilibrium with the liquid vapour, γsl
the solid-liquid interfacial energy and γlg the surface tension of the liquid in equilibrium with
its vapour. The Young-Dupré’s equation could also be written as equation 2.8:
WA = γlg · [1 + cos(θc )] (2.8)
where WA = γsg +γlg −γsl and it corresponds to the adhesion work (sometimes called Dupré’s
energy) in J · m−2 or N · m−1 .
Different models are used to determine the SFE. The most known are the Fowkes method,
the Owens, Wendt, Rabel and Kaelble (OWRK) method and the Van Oss-Chaudhury-Good
method.
OWRK method- A model (equation 2.9), in which a dispersion term is combined with a
polar term, has been considered by researchers [61][62]. Using just two probe liquids, such as
di-iodomethane and water, different surfaces can be classified with respect to their dispersion
and polar nature. The strength of this approach is that by using water as a probe, the method
is very useful in applications where hydrogen bonding may be important [25].
q q
d · γd + 2 γp · γp
γlg · [1 + cos(θc )] = 2 γsg (2.9)
lg sg lg
The OWRK method is used more frequently in practice than the Fowkes method. Then ,the
OWRK will be used in this study.
Fowkes method - Fowkes assumed that the SFE of a solid (and surface tension of a liquid) is a
sum of independent components, associated with specific interactions considering the dispersion,
polar, hydrogen, induction and acid-base components, respectively. The Fowkes method follows
a similar way to the OWRK method. However, this method can only be used for fully dispersive
system [63].
Van Oss and Good method - The Van Oss and Good method goes further than the OWRK
method. The SFE is divided into a disperse part along with a Lewis acid part and a Lewis basic
part. According to the Lewis acid-base theory, polar interactions take place when an electron
acceptor (+) impinges on an electron donor (-) [63].
2.3.1.3 Contact angle on heterogeneous surfaces
Wettability and contact angle concepts all assume a smooth and uniform surface. Nonetheless
real case are not perfect: cellulosic surfaces are no exception to the rule. Roughness and chemical
defects are easily found. More precisely, it is the wetting hysteresis that reflects the presence
and action of surface heterogeneities on the wetting. This wetting hysteresis is defined as the
difference between a maximum contact angle reached during an increase in drop volume (advance
contact angle) and a minimum contact angle reached during a decrease in drop volume (receding
contact angle). Two models describe the behaviour of liquids on surfaces with roughness or
chemical heterogeneities but only apply when the drop size is large compared to the scale of the
surface roughness.
The Wenzel’s model - Wenzel in 1936 was the first researcher who focused on he roughness
contribution in wetting[25], [64], using a geometrical argument based on a roughness factor rw
as shown in equation 2.10.
Sdr
rw = 1 + (2.10)
100
where Sdr is the developed interfacial area ratio and corresponds to the area’s additional surface
contributed by the material texture as compared to the planar definition area. The Sdr of a
completely level surface is 0. Thus, the Young’s equation needs a correction as shown in equation
2.11 [65] [66].
cos(θc∗ ) = rw · cos(θc ) (2.11)
where θc∗ is the corrected contact angle for a rough and/or heterogeneous surface. Wenzel’s
equation normally only applies to systems with no hysteresis. This would normally exclude most
industrial surfaces in general and paper in particular [67]. Even in cases where contact angle
hysteresis is observed, average value between advancing and a receding angle can be taken, and
then applies the Wenzel equation to the result [25]. Wenzel equation is pretty accurate when
applied to hypothetical surfaces having regular roughness. Different experiments have been
made to verify Wenzel’s model [25].
The Cassie and Baxter model - Cassie and Baxter proposed in 1944 [68] the expression
of a drop contact angle on a chemically heterogeneous surface. In the case of a bicomponent
heterogeneous surface (compounds A and B), the expression of the resulting contact angle is
given by equation 2.12 :
cos(θc∗ ) = ϕA · cos(θc,A ) + ϕB · cos(θc,B ) (2.12)
where ϕA represents the fraction of surface of chemical compound A for which the contact
angle has a value θc,A on a surface consisting exclusively of this compound, and ϕB the fraction
of surface and the corresponding contact angle θc,B for compound B.
2.3.1.4 Techniques to determine liquid surface tension
Du Noüy ring A maximum force Fdw occurs when the lamella, produced when a ring (Figure
2.19a) moves through a liquid phase boundary, is aligned vertically to the ring plane. This
maximum correlates with the surface tension γlg according to equation 2.13.
Fdw
γlv = (2.13)
Ldw · cos(θc )
The wetted length Ldw of the ring is the sum of the inner and outer ring circumference. Platinum-
iridium is used as the material for the ring, as this alloy is optimally wettable due to its very
high surface free energy and therefore generally forms a contact angle θc ) = 0◦ with liquids. The
material is also chemically inert and easy to clean.
Wihelmy plate - When a vertically suspended plate is in contact with a liquid surface or
interfaceas shwon Figure 2.19b), a force Fdw acts on this plate as shown equation 2.13. The
difference here is that Ldw , the wetted length of the plate, is equal to its perimeter. To measure
the force Fdw , the plate is attached to a force sensor of a tensiometer. The material used for the
Wilhelmy plate is the same as for the Du Noüy ring. With the Wilhelmy method, a dynamic
contact angle could be measured by slowly immersing and then withdrawing the tested solid. It
also enables to measure SFE from different materials using liquids with known properties.
Jurin’s law - During the wetting of the internal walls of a capillary, the spreading of the
liquid on the walls imposes a curvature on the fluid which induces a depression and thus the
rise of the liquid in the capillary (capillary rise). It is this depression which is responsible for
the spontaneous imbibition of porous media. It is possible to calculate the equilibrium height
hjurin reached by the liquid when the hydrostatic forces (weight of the liquid) compensate the
capillary forces. This relation is the Jurin’s law, firstly described in 1718 [69] [70], and presented
equation 2.14.
2γlg · cos(θc )
hjurin = (2.14)
ρ · g · rcap
(a) Du Noüy ring (b) Wihelmy plate (c) Pendant drop method
Figure 2.19: Surface tension measuring systems.
where ρ the density of the liquid (kg·m−3 ), rcap (m) the radius of the tube, and g the acceleration
of gravity (9.81 m · s−2 ).
Pendant drop method - This method is based on the shape of drops at equilibrium as
illustrated Figure 2.19c. Small drops tend to be spherical because the surface tension effects
depend on the surface area of the drop (sphere = minimum surface area). As the hydrostatic
pressure is dependent on height, the curvature of the drop interface also changes in the vertical
direction. When equilibrium is reached, the liquid surface tension (γlg ) of a liquid can be
determined using equation 2.15.
∆ρ · g · d2e
γlv = (2.15)
H
where ∆ρ (kg.m−3 ) is the density difference between the liquid in the drop and the surrounding
phase, g = 9.81 m.s−2 ; de (m) is the drop equatorial diameter and H is a shape factor. H −1
is a function of ds /de , where ds , is the diameter measured at a distance de from the bottom
of the drop. The values of H −1 were determined empirically. This method cannot be used for
volatile liquids.
2.3.2 Water absorption in paper

Now that the coating color has wetted the paper, it can penetrate the fibre network. This
sub-section will therefore look at the absorption of liquids, including water, by the paper.
2.3.2.1 The Lucas Washburn law
Liquid transports in porous material are possible with the presence of a driving potential. These
driving potentials may be multiple at a same time [71]. Generally, penetration of a liquid into a
porous substrate takes place by capillary flow into pores. The classical liquid transport model is
based on Lucas (1918) and Washburn (1921) works [69] [72] [70]. Before focusing on the Lucas
- Washburn model (LW), it is important to present the prerequisites [25]:
→ the capillaries are cylindrical and have uniform diameter;
→ the walls are perfectly smooth;
→ the contact angle of the liquid with the solid surface is constant, even when the liquid
front is moving;
→ the liquid flow is laminar;
→ the liquid is Newtonian;
→ the capillary is connected to an unlimited tank of the pure liquid.
The kinetics of spontaneous ascent in a tube of radius rc less than the capillary length has
been studied by many authors such as Washburn [73], Quéré [74] and Fries [72]. They have
highlighted different regimes during a capillary ascent and defined transition times between
these different regimes. Fries [72] writes a momentum balance on the system composed of a
tube of radius rc and a reservoir containing a fluid with characteristics of density ρ, dynamic
viscosity η, liquid/gas surface tension γsl and static contact angle between the fluid and the solid
wall θc . This balance is a complex equation which is usually simplified, leading to equation 2.16
dh d2 h 8η dh 2γlg · cos(θc )
ρ· =− 2 ·h· + −ρ·g·h (2.16)
dt dt2 rc dt rc
Because it is rare that all of these assumptions are strictly true, the approach of Lucas
and Washburn can be regarded as an approximation in most practical cases. The LW model,
presented equation 2.17 links the height or the depth of liquid penetration into the materials
with different parameters.
0.5
γlg · rcap · cos(θc )

h= · t0.5 (2.17)
2η
where rcap is the capillary radius (m), η is the fluid viscosity (P a·s) and t is the time (s). Two
important points have to be highlighted. Firstly the penetration depth is proportional to the
square root of time. Second, the penetration is driven by the contact angle between the liquid
and the solid. Therefore, the contact angle between the liquid and the solid is a key criterion to
determine if capillary penetration is possible [75]. Experimental results indicate that this model
can be used to describe the average large-scale time evolution of liquid penetration also in more
complex capillary systems and in porous media such as paper [76].
2.3.2.2 Limits and modifications of the Lucas-Washburn law
The analytical solution of the LW equation and some experimental results showed a good con-
sistency but also some limitations in the following situations:
→ porous medium is simplified as a combination of many idealized capillary pores with the
shape of parallel cylinders;
→ the capillary imbibition model considers one-dimensional (1D) flow in straight tubes;
→ the interconnection of pores is not considered and their complex geometries of pore have
to be taken into account;
→ the static contact angle and tube radius are presumed known;
→ the tortuosity is neglected, and the actual flow channel is replaced with a straight-line fluid
motion;
→ the liquid velocity into the material channels will not necessarily be uniform; [25]. Roberts
et al. [77] revealed that some channels within a paper web may be entirely bypassed as
liquid chooses "easier" and unblocked channels;
→ the swelling of cellulose materials can constrict the open areas of pore cross-sections [71];
This is why other researchers set out to solve the equation under other hypotheses. A capillary
ascent can thus be successively described by: an inertial regime, a visco-inertial regime, a viscous
regime and a visco-gravitational regime. Quéré has studied the cas of inertial regime[74]. To
establish the solution associated with this regime, the terms related to viscous and gravitational
forces are neglected in the differential equation. The following solution expressing the fluid
height in the porous medium as a function of time is shown in equation 2.18.
s
2γlg · cos(θc )
h(t) = t (2.18)
ρ · rc
In this case, the ascent rate is constant. This law remains correct for a limited time interval
beyond which the viscous forces are no longer reasonably negligible. Fries [72] defines transition
times between regimes corresponding to the instants beyond which the deviation between the
solutions in each regime is greater than 3%. The transition time t1 between the inertial regime
and the visco-inertial regime is thus defined by equation 2.19.
t1 = 0.1856 · τ (2.19)
ρ · rc2
where τ = is commonly defined as the characteristic transition time between the inertial
8η
and the viscous regime. Concerning the visco-inertial regime, only the influence of gravity is
neglected. Bosanquet [73] obtains the following solution presented equation 2.20.
s
γlg · cos(θc )
h(t) = (t − τ (1 − exp (−τ · t))) (2.20)
2η
The transition time t2 between the visco-inertial regime and the purely viscous regime is defined
by Fries [48] by the following relation:
t2 = 16.921 · τ (2.21)
The viscous regime is the one described by Lucas and Whasburn. The solution given by this law
at short times is not representative of real physics. When t tends to zero, the impregnation rate
then tends towards infinity. It is essential for very short times to consider inertial effects as Quéré.
Finally, Fries studied the visco-gravitational regime. They defined a limit fluid height beyond
which it is essential to consider gravitational effects. This limiting fluid height corresponds to
10% of the equilibrium height given by Jurin’s law (equation 2.14). It enables the calculation of
the capillary imbibition process for a longer time than is the case in the classical LW equation
[72].
3 · t · sin2 (Φ )
ρ2 · g 2 · rcap
2γlv · cos(θc )
( " !!#)
p
h(t) = · 1 + Wlbt −exp −1 − (2.22)
ρ · g · rcap · sin(Φp ) 16η · γlv · cos(θc )
where sin(Φp ) is the is the inclination angle of the capillary tube with respect to a horizontal
plane and Wlbt is the Lambert function. At the end of this visco-gravitational regime, equilibrium
between capillary pressure and hydrostatic pressure is reached. The equilibrium height is given
by Jurin’s law. A discussed limitation of these models concerns the contact angle considered. In
all these laws, it is the static contact angle θc . This choice can be questioned especially in the
first moments of a capillary rise. It may be wise to couple these models with the laws expressing
a dynamic contact angle.
Other works more related to paper could be cited. Bristow [71] was able to measure water
penetration at considerably shorter contact times. The non-linearity between the obtained
sorption values and the square root of contact time was explained by the concept wetting delay:
there is a wetting time, twet , before the water starts to penetrate into the paper. After this
wetting delay, the water sorption was suggested to follow the square root relationship indicated
in the Washburn model. Bristow explained that swelling must be taken into account when water
transport phenomena in cellulosic material is discussed. He separated the total water sorption
into a pore sorption part describing the transport in the existing pores, and a fibre sorption part
depending on the water taken up by the fibre material. Hoyland [71] proposed a modified LW
model taking into account the thickness variations of paper :
s
γlv · rcap · cos(θc · (t − twet )
h(t) = − ϖ∆tf (2.23)
2η
where ∆tf is the increase in the thickness at time t and ϖ is a constant. Research in recent years
has been conducted to investigate the capillary-driven flow in porous systems by modifying and
extending the LW model.
Finally, more general works concerning the modelling of capillary-driven flow in porous
systems could be quickly presented. Adding various correction factors into the microstructure of
rocks, including tortuosity and pore shape, [78] modified the LW equation adding a pore shape
factor δp presented equation 2.24.
2 π · Ap
p
δp = (2.24)
Π2p
where Ap and Πp denote the cross-sectional area and the perimeter of pores, respectively. The
mass m of the absorbed water into porous media can also be expressed as follows:
s
γlv · rcap · δp · cos(θc )
m(t) = εc · Sp · ρ · (2.25)
2τp · η
where εc and Sp are porosity and the cross-sectional area of the porous media, respectively. τp a
correction factor of tortuosity. On the basis of tortuous streamline with fractal characteristics,
Cai et al. [78] modified and extended the classical LW equation. They demonstrated that the
more tortuous the streamline, the longer the time it takes to reach a given height. It can be
expressed with equation 2.26:
dh (2rcap )2DT −1 · γlv · cos(θp ) (2rcap )2DT · ρ · g

= − (2.26)
dt 8η · DT · h2DT −1 32η · DT · h2DT −2
where DT is the tortuosity fractal dimension.

Many models exist to approximate the absorption of liquid by porous materials. They
integrate more or less components of the fluid or the material, in order to approach a behaviour
as realistic as possible. However, some of the characteristics used are not always easy to acquire.
Thus, in this study, we will remain close to the classical LW theory, and will not complicate the
model.
2.3.2.3 Measure the penetration of liquid into paper
There are different methods for measuring the absorption/penetration of liquids in papers. Six
methods will be presented: some are regularly found in the literature, others are more innovative.
Cobb index - The measurement of water absorption by the Cobb method (according to ISO
532-5) consists of wetting one side of a paper with a given quantity of water for a given time
on a given surface (key parameter). The measurement obtained is given in gwater · m2paper . This
value depends on the height of the water column used and the water/paper contact time. The
test is only valid if there is no leakage during the test.
Bristow wheel - Developed by Bristow in 1967 [71], this method consists of attaching a strip
of paper to a wheel with a rotation speed between 0.01 cm · s−1 and 10 cm · s−1 . Above this
wheel, a small reservoir containing the test liquid is attached. At a given moment, the tank is
brought into contact with the paper. The liquid then flows onto the paper through a slit of 1
mm wide under the reservoir. This test therefore allows a paper of known speed to be wetted
with a known volume of liquid. It gives access to the depth of penetration of the liquid into the
paper for different application times.
Scanning absorptometer - This absorptometer is based on the Bristow’s method [79]. Dur-
ing a scanning absorptometer measurement, liquid is supplied from a scanning head moving
along a spiral path on paper as illustrated Figure 2.20. The speed of the head moving over the
paper surface is kept constant over a certain part of the paper sample. The head speed could
Figure 2.20: Scanning absorptometer apparatus [79]
Figure 2.21: Salminen apparatus [71].
be increased , enable the system to measure the liquid penetration at different contact times
(between 10 ms and 10 s). The total absorbed liquid is obtained in ml.m−2 .
Salminen apparatus [71] - This device was designed with the possibility to measure liquid
penetration at different pressure and temperature, with high contact time possibility. It consists
of a liquid applicator with a slit opening, a liquid flow measurement through the slit opening, a
backing roll and a web velocity control system (cf. Figure 2.21). The web velocity and the time
needed for the penetration of a given liquid volume into the paper are measured. The liquid
sorption per area unit and the contact time between liquid and paper are obtained with this
method.
DPM[75], [79] - This test enable to determine liquid penetration at very short time using
the evaluation of ultrasound transmitted through paper thickness during wetting. The sample
is fixed on a sample holder with double-sided tape. When the test starts, the sample holder is
immersed into the test liquid (water here). When the immersion is complete, a high frequency
(1 or 2 MHz), low energy ultrasound signal is transmitted through the sample in thickness
direction.
During all the test, the transmitted signals are recorded over time. There are two types
of sample holders: a flat sample holder where backside is completely covered with the tape; a
grooved sample holder where the sample is not in contact with the tape in the grooves. The
typical curves obtained by PDA are shown in Figure 2.22. The abscissa corresponds to the
measurement time. The ordinate indicates the transmission rate of the transmitted signal. By
convention, the maximum transmission point is set to 100%. The curve shows an increasing
transmission at the beginning of the measurement. This is followed by a decreasing one. During
the first phase, many air bubbles get stuck in the paper. They are gradually replaced by the
liquid in the tank. This wetting phase is proportional to the roughness of the paper. Thus, the
test starts with a lot of liquid-air-paper interaction to then give way to liquid-paper interactions
only.
Figure 2.22: Parameters for the evaluation of DPM measurement proposed by Waldner and
Hirn [75].
The change of behaviour can be materialized by the time tmax , corresponding to the highest
transmission and it is often interpreted as the wetting time. The decreasing phase appears when
the wetting is realized and the air in the paper is driven out little by little from the paper. The
intensity of the transmitted signal then decreases with time. The slope of the decreasing phase
is proportional to the penetration speed. The time t95 defined on the Figure 2.22 corresponds to
the inflection point of the curve and the index SDP M is the maximum slope (in % of transmission
·s−1 ). This method is very practical to determine the wetting time tmax however it does not
allow to quantify the exact quantity of absorbed liquid contrary to other methods (Cobb and
Bristow wheel for example).
Karppinen method - Based on diffuse light reflection from paper, Karppinen [80]studied
the penetration of different liquids: water, mineral oil and pure isopropanol into paper samples.
Their method could identify the individual wetting behaviour of different paper qualities and
different wetting agents. In practice, this method is well suited for dynamic testing of paper
wetting featuring up from 10 ms single droplets applied on paper.
2.3.3 Rheology of coating colors
Starches differ significantly in physicochemical, rheological, thermal and retrogradation proper-

ties. Potato starch exhibits higher swelling power, solubility, paste clarity and viscosity than
wheat, rice or corn starches. Morphological characteristics, such as shape and size of the starch
granules, are also different . Gelatinization temperatures and gelatinization enthalpies from dif-
ferent sources are also different depending on the origin. Potato starch shows a higher tendency
for retrogradation than cereal starches [81]. In another study, Sing et al [82] also showed that
even within the same starch botanical source (potato), the impact of cultivars could modify
Figure 2.23: Picture of the measurement device proposed by Karppinen [80] and its schematic
presentation.
the properties during cooking and use. Via DSC measurement, they shown that starches with
higher crystallinity have higher transition temperatures as well as gelatinization enthalpy. This
is linked to a higher degree of crystallinity, providing structural stability and makes the granules
more resistant to gelatinization. Moreover, starches with long branched chain length in amy-
lopectin molecules have higher gelatinization enthalpy, meaning that more energy is required to
gelatinize.
Rheology of paper coatings is a key factor influencing runnability, water retention and coating
holdout [33]. For example viscoelastic coating baths have an impact on the porosity of coating
layer and carboxymethyl cellulose slows water penetration in paper. In addition, in the previous
section dealing with liquid absorption in papers, the viscosity of the penetrating fluid is taken
into account in the physical models. In this subsection, the rheology fundamentals necessary
for this project will be presented as well as the techniques for measuring fluid viscosity. Starch
being used in this project, a given part fill also focused on the rheological behaviour.
2.3.3.1 Fundamental rheology tools for this project
Viscosity can be defined as the resistance of a fluid to flow and can be related to the internal
friction in the fluid, [83]. Dynamic viscosity η (P a · s) (sometimes named apparent viscosity)
can be defined (equation 2.27).
τ
η= (2.27)
γ̇
where τ is the shear stress (P a) and γ̇ is the shear rate (s−1 ). Different parameters could impact
fluid viscosity. For all liquids and pastes, η decreases when temperature increases. The thermal
agitation destroyed fluid order at short distance. The liquid viscosity variation with temperature
can be described approximately by the Eyring law as presented equation 2.28 below:
Ea

η = Ae · exp (2.28)
R·T
where Ae is a coefficient depending upon the intermolecular vibration (P a · s), Ea is the
activation energy (J · mol−1 ), R is equal to 8.314J · mol−1 · K −1 and T is the fluid temperature
(K) [84]. The chemical fluid nature and the concentration of solute are also influencing factors.
The fluid nature (mass fraction or volume fraction), aspect ratio of suspended particles impact
fluid behaviour. The viscosity of liquids increases slightly with pressure but this is often negligible
except in applications such as lubrication. The influence of time, combined with that of the shear
stress, can impact fluid viscosity: this is thixotropic and antithixotropic fluids.
The variation of viscosity with the shear stress allows the discrimination of fluids according
to their rheological behaviour. Thus, three categories of fluid are made: Newtonian fluids,
non-Newtonian fluids dependent or not of time and viscoelastic fluids.
Newtonian fluids - For a Newtonian fluid the viscosity η is constant regardless of the shear
rate γ̇.
Shear thinning or pseudoplastic fluids - These are non-Newtonian fluids. Their viscosity
decreases when shear rate increases. Two Newtonian plateaus can exist at low and high shear
rates but may be difficult to observe experimentally. Several empirical equations have been
proposed to describe the rheological behaviour of shear thinning fluids. The most used is the
Ostwald-de-Waele model, described equation 2.29, where the constant nodw is the flow behaviour
index (dimensionless) and kodw is the flow consistency index (P a.snodw ).
η = kodw · γ̇ nodw −1 (2.29)
In the case of shear thinning fluid, nodw ∈ ]0; 1[. If nodw = 1, the fluid is a Newtonian
fluid. The other widely used model is the Cross model. It considers three area : a high viscous
Newtonian plateau, a shear thinning transition and a low viscous Newtonian plateau as presented
equation 2.30.
η0 − η
= (λc · γ̇)nc (2.30)
η − η∞
where η0 and η∞ are the viscosities of the first (at low shear rates) and second (at high shear
rates) Newtonian plateaus. λc is a time homogeneous constant and λ−1
c is the shear rate at
which shear thinning behaviour appears. nc has the same dimensionless meaning that nodw in
the Ostwald-de-Waele model. When η∞ << η << η0 then this model is like a power law model
[85]. There is also the Carreau-Yasuda model presented equation 2.31 where nc and λc have
similar significance that in the Cross law.
η − η∞ nc −1
= [1 + (λc γ̇)ay ] ay (2.31)
η0 − η∞
where ay is a dimensionless parameter describing the transition between the first Newtonian
plateau and the power law zone [86]. If ay = 2, the model is called the Carreau model. The
Carreau Yasuada model is used to fit data from a fluid that has 2 Newtonians plateau and with
a power law region in between. The parameter ay adjusts the transition from the Newtonian
plateau into the power law region. In fact, the Carreau-Yasuda model is an extension of the
Carreau model. The advantage over the Carreau model is that it is continuous for all shear rates
(γ̇ ≥ 0).
It can be said that, under the flow conditions prevailing during the coating process, coating
colors exhibit a non-Newtonian viscoelastic behaviour [49]. These non-Newtonian behaviour of

paper coating color is due to the presence of particles in the coating formulation [33].
Shear thickening fluids - The apparent viscosity of these fluids increases with shear stress.
They can have an initial and final Newtonian domain. An equation of the Ostwald-de-Waele
type can represent this behaviour. The difference with shear thinning fluid is that nodw > 1 in
equation 2.29. They are sometimes called dilatant fluids because an increase in fluid volume can
occur at high stresses
Viscoplastic fluids or threshold fluids - They are characterized by a threshold shear stress
τc below which the flow is impossible. A first approach to this behaviour is the Bingham (1922).
However, it does not take into account the shear thinning behaviour that a fluid can exhibit
beyond the threshold flow. The Herschel-Bulkley equation (1926) can then be used. It models
properly paper coating flow behaviour at shear rates below 1000 s−1 [33].
Thixotropic fluids - Thixotropic fluids are fluids that are dependent on both shear stress (i.e.
non-Newtonian) and time. A thixotropic fluid, starting from a state of rest for a sufficiently long
period, sees its viscosity decreasing with time and its structure changing under the application
of a constant shear rate. The rheogram of a thixotropic fluid shows a hysteresis. The area of
the hysteresis loop is related to the shear history. Hysteresis loop consists of a complete flow
curve for a thixotropic materiel, obtained by measurements on a rotational viscometer, showing
for each shear rate value two values of shearing stress: one for increasing shear rate and one for
decreasing shear rate [32]. Note that thixotropy is quite common in paper coating formulations
[33].
Viscoelastic fluids - A viscoelastic fluid is a material that has both viscous and elastic prop-
erties. The quantities characterizing these properties are dependent on the time of solicitation.
2.3.3.2 Rheology of starch
Nuru et al [87] exhibited the shear thinning behaviour of a sago starch paste as well as the relation
of this behaviour with temperature. Again, they showed that shear rate, starch concentration
and temperature influence rheological properties. With their results, they proposed a single
relation to link starch gel dynamic viscosity η with these 3 parameters (cf. equation 2.32)
Ea

η = A · γ̇ · C · exp
B D
(2.32)
R·T
where A, B and D are constants. This relation is a mixed between power law model and
Eyring law. Concerning the rheological characteristics of starch paste, it depends on many
factors: chemical structure of starch, modification, starch concentration, pasting conditions
(temperature, shear rate and heating rate) and storage conditions (temperature and time) [88].
The viscosity profile of native and modified starch is given in Figure 2.24. It shows the shear
thinning behaviour of starch gel. Moreover, it is shown that the lower the starch concentration,
the lower the viscosity at a given shear rate. Such behaviour will therefore be expected in this
study. There is an overall viscosity reduction in the modified starch, but there is no change in
the shear thinning property.
Figure 2.24: Viscosity profile of native and modified starch as a function of shear rate [89].
Doublier, in 1981 [90] also worked on the impact of concentration on rheology properties.
With a study of wheat starch pastes flow behaviour trough different concentrations (0.3 to 8%),
it has been showed that flow curves indicated a non-Newtonian and shear thinning behaviour
whatever the concentration but at different scale. Shear-thinning happens due to the disentan-
glement of the polymer network. These curves have also a Newtonian limit viscosity between
0.3% and 1.5%. Above this concentration, the flow curves could be described by an Ostwald-
De Waele equation over a limited range of shear rates. Concerning the impact of temperature
Figure 2.25 shows that the higher the temperature, the lower the viscosity η.
In a recent publication, van der Sman et al [91] shown that much of the rheological behaviour
of concentrated polysaccharides, as starch, is governed by the ratio Tg /T . They showed that the
Carreau-Yasuda model well describes the flow curves. In this model presented earlier equation
2.31, inverse of critical shear rate represents a fundamental relaxation time, which is dependent
on both temperature and moisture content. Furthermore, for linear polymers it is found that
this relaxation time also depends on polymer molar weight. Nonetheless, other components
impact the rheological properties of the coating color and make the rheological behaviour not
totally predictable [92] [93].
The viscosity of starch paste determines the thickening power of starch for various applica-
Figure 2.25: Starch viscosity as a function of temperature [87].
tions. After cooling and storage, and depending on the starch botanical origins, the starch paste
do not act the same. Wheat and maize starch pastes (at 6%, w/w) can form gels with a defined
shape and exhibit a viscoelastic property. In the other hand, waxy maize, tapioca, and potato
starches remain as pastes or form rather weak gels at higher concentration [88].
2.3.3.3 Techniques for measuring rheological properties [94] [83]
There are different techniques for measuring rheological properties in different configurations.
In this project, only the parallel plate cone plane and capillary configurations will be presented
(Figure 2.26). Regardless of the basic geometry, the rheology equations assume non-slip of the
fluid at the wall. Techniques exist to detect slip or fracturing of the fluids during testing. In fact,
during rheology measurements, the onset of slip means that there is a loss of contact between the
sample and the tool. The adhesion condition is a prerequisite for the interpretation of rheometric
tests. Thus, the measurement becomes dependent on the study conditions. Within the sample,
the flow is no longer controlled and the shear state becomes unknown [95].
Parallel-plate viscosimeter A plate of known radius is rotated at angular speed parallel

to a fixed plate of the same radius. Unlike the cone-plane system, the velocity gradient is not
constant between the two plates and is even less suitable for low viscosity liquids. The stress
value is taken for the maximum radius value.
Cone-plane viscosimeter A cone with a known half-angle at the apex, is rotated under the
action of a constant torque (or at constant angular velocity). A fixed plane is in contact, without
friction, with the top of the cone; this plane is placed perpendicular to the axis of the cone. The
cone rotation speed is measured, or the cone torque. No correction is necessary provided that
measurements are not made at very high speed (turbulence). This method has the advantage of
allowing a uniform gradient and shear stress between the cone and plate, a wide measurement
range and the use of a small volume of fluid required. However the device is difficult to use with
low viscous or low surface tension liquids.
2.4. BUCKLING AND WRINKLING OF PAPER 47
(a) Parallel plate (b) Cone plate (c) Capillary rheometer
Figure 2.26: Three different rheometer techniques.
Capillary rheometer - To fully characterize the rheology of coating colors, the tested shear
rates are from 0.01 to 1 000 000 s−1 which is complicated to obtain. Nonetheless, capillary
rheometer can be a powerful tool for characterizing the high-shear behaviour of coating colors.
A correction should be applied with this method to correct inertial effect[33].
2.4 Buckling and wrinkling of paper
2.4.1 Tension buckling and fluting
2.4.1.1 Tension buckling from the mechanistic point of view
Anisotropic membranes with fibrous textures are omnipresent both in organisms and manufac-
turing like biofilms, skins, biological tissues, woven fabrics, paper and composite materials [96].
They are made of a deformable matrix and fibres with some privileged directions. Zhu et al.
[97] experimentally showed that an orthotropic film remains flat when stretched in the high-
stiffness direction, while wrinkles occur when it is stretched in the low-stiffness direction (Figure
2.27). Liu et al. [96] showed that the degree of orthotropy and the shear modulus significantly
affect wrinkling orientation and stability boundary for orthotropic, elastic films during uniaxial
stretching.
One of the first theory used is the tension-field theory, first introduced by Wagner [96]. It
assumes membrane have infinitely low bending stiffness.This theory was used to model operations
on the human skin performed by plastic surgeons for example [96]. However, more sophisticated
plate/shell theories that account for both stretching and bending contributions to elastic strain
energy are required to predict wrinkling phenomena. Thus the small-strain wrinkling of stiff
sheets was used. Friedl et al [98] [99] first uncovered that a stretched rectangular sheet induces
transverse compression somewhere which leads to wrinkles. A model of a simply supported plate
buckling under uniform membrane loads is proposed to obtain analytical information. Assuming
that lateral compression σz and global tension are proportional, we have: σx : σz = J · σx . The
a
analytical expression of critical longitudinal stress σx,cr as a function of the aspect ratio and
b
proportionality factor J ≤ 0 is proposed equation 2.33.
Figure 2.27: Deformation at different tensile strains during tensile test of highly orthotropic
films in low stiffness direction. Each line represents a different sample width (name W here).
Each column is a different strain rate. The red rectangles are configurations where out of plane
displacement appear [97]
 4
mn 2

n

π 2 · Et2f +2 + 4
 a ab

σx,cr = − (2.33)

12(1 − νiso
2 )  2 2 
m n
+

J
b a
where tf is the film thickness, E the Young’s modulus, νiso is Poisson’s ratio and m, n are
the half-buckled wave numbers along the longitudinal and transverse directions of the test,
respectively. Furthermore, to simply determine the critical condition, Friedl et al.[98] provided
a diagram (Figure 2.28) of the dimensionless buckling coefficient kc to calculate the critical
tensile stress based on equation 2.34
2
tf

σx,cr = kc · E (2.34)
b
Cerda and Mahadevan [100] studied the instability problem of webs/plates with two opposite
edges clamped and subjected to uniform tension (Figure 2.29).
They determined the transverse wrinkling wavelength lcm and amplitude Acm by minimizing
bending and stretching energies. They showed that wrinkling is fully linked to geometry. They
considered that the film is subjected to a longitudinal stretching strain ε. The condition of
transverse inextensibility is (Acm /lcm )2 ∼ µiso · ε .Thus, they minimized the energy of the
system and it leads to the wrinkling wavelength (equation 2.35) and amplitude (equation 2.36).
0.25
√ S

lcm =2 π a0.5 (2.35)
Tx
Figure 2.28: Critical buckling coefficient according to Friedl et al. versus the length/width
ratio [98].
Figure 2.29: Example of wrinkles formation, during an uni-axial test, on a polyethylene sample
of length a = 250 mm and width b = 100 mm. The thickness tf is around 0.1 mm. On the
picture, the strain is about 10 % [100] .
√ 0.5
2 ∆
Acm = · lcm (2.36)
π b
b 2πn
where ∆ = A2n · kn2 and kn = . Jacques and Potier-Ferry [101] [102] [103] further
8 b
carried out an analytical calculation from an energy standpoint to explore wavelength selection
and mode localization in a long tensile plate. Figure 2.30 illustrates the studied mechanism, in
the different stages of wrinkles formation.
The considered problem has no analytical solution, but the critical tensile stress σx,cr (equa-
tion 2.37) and transverse wavelength ljp (equation 2.38) can be determined.
Eπ 2
σx,cr = (2.37)
3(1 − µ2iso )J 2
2π q
ljp = = 2|J|lx,jp lc,jp (2.38)
kjp
where lx,jp is the longitudinal buckling wavelength, lc,jp is a characteristic length and kjp is
the transverse buckling wave number. Puntel et al [104] showed that the applied critical strain
Figure 2.30: Appearance of wrinkle during sheet conveying. Here TB is the critial buckling
load and TW is the critical wrinkling load [101].
depends on aspect ratios and verified the scaling law obtained by Cerda and Mahadevan [100],
providing the critical strain (equation 2.39).
π 2 t2f a2 n2kp
" ! !#
a2
εcr = 1 + 2 · (nkp + 2)2 1 + + n2kp (2.39)
48(1 − µ2iso )a2 b b2
where nkp ∈ N is related to the wave number kp , satisfying nkp = kp − 2.

In this study, we will use the kc criterion of Zhu et al [97]. Indeed, it is the theory taking into
account the orthotropy of paper. Thus, the critical tensile stress can be calculated. However,
when some papers deal with isotropic films [100] [99] [104] [98], it is necessary to use a formulation
that considers the anisotropy (or orthotropy) of the paper. Thus, based on Zhu et al.[97] work
on highly orthotropic films, equation 2.40 can be used.
Ex π 2
2
tf

σx,cr = kc · (2.40)
12 (1 − ξ1 νxz
2 ) b
where equation ξ1 (2.41) is the anisotropic coefficient and ξ2 (equation 2.42) a coefficient
depending on the Poisson’s ratios and ξ1 .
Ez
ξ1 = (2.41)
Ex
4 (1 − νxz νzx )
ξ2 = + νzx + νxz ξ1 (2.42)
1 + νxz + 1+νξ1
xz
Thus, kc can be isolated in equation 2.43.
3ξ1 − 1 + 2ξ12 νxz 2

( )
2
h i
kc = β + ξ1 β −2 n2 (n + 2) + n2 + (n + 2)2 (2.43)
ξ1 + 1 + 2ξ1 νxz
b
where β = . However, this relationship is only valid when the plate is clamped at its ends,
a
contrary to the assumption made by Good where the plate, i.e. the paper, is simply supported
and thus generates a different behaviour with respect to the Poisson’s effect. Moreover, Juner’s
work is also based on the Föppl-von Karman equation and must therefore be used with the
assumption of its complexity in mind [105]. In all these previous studies, webs were modelled
as plates subjected to uniform biaxial loads, which did not conform to the practical stress
distributions in the webs. Some works try to used stress function to approach case where
tension field are not uniformly distributed [106] [107] [108].
2.4.1.2 Tension buckling from the paper maker point of view
Wrinkles, web weaving, fluting and cockling are commonly seen in any web transport systems,
such as paper machine, winder, converting machine, and printing press (cf. Figure 2.31). Paper
is a viscoelastic, time-dependent material, greatly influenced by water, tension and heat. The
paper web tension - coating color - heat interaction have complex effects on the paper material.
Thus, linked to these four parameters, two effects have been highlighted concerning buckling
appearence: Poisson effect due to web tension and differential shrinkage[109].
(a) Cockling appearance [110] (b) Typical fluting pattern [111]
Figure 2.31: Highlighting of cockling and fluting appearence on paper.
In a coating process, paper is fixed by rollers at both ends. Cross Direction (CD) shrinkage is
then restricted. Web tends to buckle and out of plane displacement appears when the tension is
high enough. Moreover, when web tension is not uniform across the paper width it also leads to
buckling by secondary compression stress in CD as a result of the Machine Direction (MD) web
tension. It lead to the formation of troughs to minimize the stress in thickness direction. This
tension wrinkling occurs when thin materials are stretched under constraint. Tension wrinkles
disappear when the tension is released if the elastic limit was not reached. In coating process,
paper is subjected to variations due to its contact with water and heat during the drying. Both
of them decrease the elastic limit. However, the tension used in coating process is high enough to
exceed the elastic limit, especially when the paper is wet. If the tensile buckling is not corrected
before drying, it will stay on the final product. Paper may wrinkle by tension only but adding
heat creates a sharper and narrower fluting pattern than tension alone [112][113][110][114].
Another factor is the non-uniform CD moisture profile often seen at dry end of papermaking
process: areas where the moisture content is high have lower web tension. The elastic stiffness
in MD is usually 3 times larger than that in CD. In addition, the hygroexpansion coefficient
of paper in CD is usually 3 times larger than in MD. So, if fluting is driven by hygro-strains,
changing the direction of paper fiber increases the critical moisture change, needed to initiate
fluting, approximately by a factor of 9. Differential shrinkage between printed and non-printed
areas in printing process could be a reason of fluting. They postulated that image areas lose
less moisture during drying as the ink layer prevents moisture from evaporation. Coffin[115]
predicted the wavelength of fluting and showed that with this theory, tension is an important
factor determining the fluting wavelengths [116][117][111].
Paper may take up and lose moisture during the coating process causing important hygroex-
pansion or shrinkage in paper. These mechanisms tend to predict much larger wave length of
fluting than what were observed in printing. However, if paper is assumed to be non-uniform,
local buckling can occur [118]. One example of non-uniformity is local variations of fibre ori-
entation (its direction and degree). This structural variation could trigger local variations of
hygroexpansion, and thus local buckling . Cockling is, on the other hand, more random features
without any specific orientation direction. Fundamental mechanisms of fluting and cockling are
still somewhat controversial. Cockle can be created by producing local variations of drying “his-
tories”. If paper sheet has some wet and dry areas during drying, these areas develop residual
strains differently, and then buckle after drying is completed, causing cockles [119].
2.4.2 Shear wrinkling of a web in a roll-to-roll process
The study of wrinkles on a moving web is not new. The theory setting out the foundations for
studying this problem was proposed by Leonhard Euler, around 1750, who formulated a first
theory of beams by defining the elastica, a medium or neutral fiber that deforms in bending
without contracting or compressing. The study of plate bending returned during the 19th century
with the studies of Young, Laplace, Poisson, Navier, Cauchy, Fourrier, Germain, Lamé and
Clapeyron. In 1888, Love used the assumptions of Gustav Kirchhoff, themselves inspired by the
Euler-Bernoulli assumptions for beams, to found a theory of thin plates. The plate theory was
consolidated by Rayleigh (1877), von Kármán (1910), Timoshenko (1921), Reissner (1945) and
Uflyand (1948).
It is important to note that in the literature, the Timoshenko’s theory of elastic stability is
frequently found and cited as well as Föppl–von Kármán equations that described large deflec-
tions of thin flat plates. The physical validity of these last equations stay questionable [105]. It
is then possible to draw a parallel between a plate and a web. It is known in the web handling
industries that web shear wrinkles come from misaligned rollers. Therefore, many investigators
formulated the prediction model of wrinkles for isotropic webs with constant Young’s modulus
in MD and CD.
The plates buckling theory considers the webs as plates with low bending stiffness. Analyses
have been done using the energy methods or numerical methods such as finite element anal-
Parameters Variation Wrinkle risk Impact

Web span ↘ ***** ***** on critical misalignement an-
gle ↘
Web thickness ↘ **** * on critical misalignment angle ↗
**** on the critical web tension ↘
Friction ↗ *** *** on the critical web tension ↘
Web Poisson’s ratios ↘ * * on the critical web tension ↘
Web Young’s Modulus ↘ **** ** on the misalignment angle ↘
**** on the critical web tension ↘
Table 2.1: Prediction of web wrinkles proposed by Martz [126](high arrow: increase, low
arrow: decrease, *****: very important impact, *: low weak)
ysis [120] [120] [103] [121]. Shelton [59] studied the boundary conditions of webs transported
through misaligned rollers: webs were modelled as Euler beams, and the normal entrance law
was proposed to determine the boundary conditions. Then, the stress distribution in the web
could be obtained. Furthermore, based on Shelton’s studies, many other buckling analyses of
webs have been done and published. For example, a series of studies were carried out in the Web
Handling Research Center (Oklahoma State University, USA). In 1989, experimental investiga-
tions show that two types of wrinkling failures can exist in webs subjected to shear due to roller
misalignment [122]. Good et al. [123] presented a buckle prediction model for isotropic webs
with roller misalignment, which reliably forecasted the web behaviour. However, anisotropic
materials such as paper, have different Young’s modulus in MD and CD. Moreover, in the case
of paper-webs, the web air permeability could impact wrinkles formation. Accordingly, Good
and Beisel analysed the mechanism of buckling of anisotropic PET rolls. Hashimoto [124] pro-
posed a theoretical modelling of paper-web wrinkling generation due to misalignment of roller
and also considers web speed and web tension. It is completed with experimental verifications.
Beisel [125] studied the buckling phenomenon caused by the imperfect structure of rollers,
such as coin or concave profile of rollers. Moreover, Good and Beisel [125] used the biaxial
loaded plates buckling criterion to study the buckling of the web caused by the misalignment
of the roller. These theories have been used by Martz in his PhD thesis [126] where he also
proposed a table that sums up the impact of each process parameters (Table 2.1).
Kim Lin and Mote [107] [108] studied the critical buckling loads of webs with all edges simply
supported and subjected to non-uniform tension. Tang et al. [106] proposed a theoretical
calculating process for the elastic buckling of webs transported through a pair of misaligned
rollers, by making use of the plate buckling theory and determining the stress distribution of
webs by using the Timoshenko beam theory.
Theoretical models: Webs are composite membranes and plate structures that react to
compressive stresses until they reach the critical buckling stress [127] [123]. Wrinkles in web are
caused by the compression exceeding the critical buckling stress (when plates don’t shorten but
bend) and can be loaded in three directions. In this case, the focus is made on wrinkles, with
machine direction web tension, induced by lateral shearing forces in the web, and the wrinkles
that result are called “shear wrinkles” (Figure 2.32a). Timoshenko proposed a solution for the
buckling stresses for a rectangular plate [128]. It allows tension on two opposing sides and
compression on the other two opposing sides.
(a) Misaligned roller and wrinkle (b) Impact of the friction force
Figure 2.32: Buckling generation by shearing and impact of friction force .
Assuming that the plate is slightly buckling under the forces applied in its middle plane and
then to calculate the magnitudes that the forces must have in order to keep the plate in such a
slightly buckled shape. The solution uses the energy method. It states that when the work done
by forces acting in the middle plane of the plate ∆T becomes larger than the strain energy of
bending for any shape of out of plane deflection ∆U , the plate is unstable and buckling occurs
[122]. Thus, to find the critical buckling load, the work done by the forces is set equal to the
strain energy in bending as shown equation 2.44.
∆U = ∆T (2.44)
Isotropic materials (Timoshenko’s theory) [128]: Introducing forces Nx , Nz , the Pois-

son’s ratio for an isotropic material νiso , the web bending stiffness Diso and out of plane dis-
placements w, it leads to equation 2.45.
1
ZZ " 2 2 #
∂w ∂w

− Nx + Nz dxdz =
2 ∂x ∂z
 !2  !2 
Diso ∂2w ∂2w ∂2w ∂2w ∂2w
ZZ  
+ − 2 (1 − νiso ) −  dxdz (2.45)
2

 ∂x2 ∂z 2 ∂x2 ∂z 2 ∂x∂z 
Assuming that x-direction is MD and z-direction is CD, wmn (x, z) = w (equation 2.46) could be
described as the solution of linear partially differential equation and amn is the amplitude [24].
mπx nπz

wmn (x, y) = amn sin sin (2.46)
a b
It is assumed that the edges of the plates are simply supported. b is the web width (mm), a
the web span (mm). The number of half waves in the x-direction is m, and in z-direction in n.
Thus, combining the previous equation, dividing by the web thickness tf and introducing the
stress yield σx and σz , it leads to equation 2.47 and equation 2.48.
2 2 " 2 2 #2
mπ nπ mπ nπ

Nx + Nz = Diso + (2.47)
a b a b
2 " 2 # 2
a a

σx m + σz
2
n· = σe,iso m + n·
2
(2.48)
b b
where σe,iso is a normalized stress [128] presented equation 2.49.
π 2 Diso
σe,iso = (2.49)
a2 tf
Setting m = 1 because only one-half wave appeared in MD direction, equation 2.48 become
equation 2.50.
2 " 2 #2
a a

σx + σz n· = σe,iso 1+ n· (2.50)
b b
Equation 2.50 is the basic equation for failure criteria used to indicate if wrinkling appeared
on the web. Finally, after some calculation, for an isotropic plate, it leads to the critical normal
stress necessary to generate wrinkles σz,cr , where i is the number of half waves for a specific web
span that is wrinkled (equation 2.51).
 !2 
b2 i2 a2
σz,cr = 2 2 σe,iso 1 + 2 − σx  (2.51)
i a b
Orthotropic materials: In a similar way, by taking into account the anisotropy of paper,
with a thickness tf (mm), σz,cr is given by [15] according to equation 2.52.
" ! #
b2 a4 a2
σz,cr = 2 2 σe,ortho 1 + ξ1 i4 4 + ξ2 i2 2 − σx (2.52)
i a b b
where σe,ortho is the the normalized stress for orthotropic plates given equation 2.53. Dxx is the
web bending stiffness in x-direction given equation 2.54. σx is the stresses generated by the web
tension Tx , given equation 2.55.
π 2 Dxx
σe,ortho = (2.53)
a2 tf
Ex t3f
Dxx = (2.54)
12 (1 − νxz νzx )
Tx
σx = with Tx the web tension (2.55)
tf
Ex and Ez are the elastic modulus (MPa) in x-direction (loading direction and web transport
direction) and z-direction (transverse direction), νxz is the Poisson’s ratio for transverse strain
in the z-direction εz when stretched in the x-direction, where εx is the strain in x-direction.
Ez
Since the ratio is defined as ξ1 , the Poisson’s ratios in the same plane are related due to
Ex
the symmetry of the stiffness matrix as presented equation 2.56.
−εz
νxz = and νzx = ξ1 νxz (2.56)
εx
To define the interger i, equation 2.57 must be true.
" # " #
a4 a4
σe,ortho 1 − i (i + 1) ξ1 4 < σx < σe,ortho 1 − i2 (i − 1)2 ξ1 4
2 2
(2.57)
b b
It is possible to determine the maximum critical shear stress that the web must withstand τcr
(equation 2.58) and link it with the tram error θcr (°) (also called misalignment angle) (equation
2.59) according to equation 2.46, invoking Castigliano’s second theorem:
q
τcr = 2
σz,cr − σx σz,cr (2.58)
6τcr a2
θcr = (2.59)
Ex b2
Another point to consider is the friction between the paper and the surface of the roll.
Indeed, this friction will determine the web tension T at which wrinkles may or may not appear.
This critical tension value Tcr is inversely proportional to the value of the friction coefficient µ.
Thus, as presented by Hashimoto [124], a simple empirical model, leading to equation 2.60.
2t2f
s
Ex Ez
T > Tcr = (2.60)
µb 3 (1 − νxz νzx )
Chapter 3
Materials and methods
Contents
3.1 Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
3.1.1 Papers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
3.1.2 Starch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.1.3 Chemicals used for contact angle measurement . . . . . . . . . . . . . . 59
3.2 Coating color . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
3.2.1 Preparation of the starch gel . . . . . . . . . . . . . . . . . . . . . . . . 60
3.2.2 Devices and methods for characterising the coating color . . . . . . . . . 60
3.2.2.1 Determination of the starch grain sizes . . . . . . . . . . . . . 61
3.2.2.2 Sedimentation studies . . . . . . . . . . . . . . . . . . . . . . . 61
3.2.2.3 Differential Scanning Calorimetry (DSC) . . . . . . . . . . . . 61
3.2.2.4 Starch gel cooking control . . . . . . . . . . . . . . . . . . . . . 62
3.2.2.5 Rotational rheological measurement of the starch gel . . . . . . 62
3.2.2.6 Measure of the starch gel density . . . . . . . . . . . . . . . . . 63
3.2.2.7 Determination of starch gel surface tension . . . . . . . . . . . 63
3.3 Devices and methods for characterising the paper . . . . . . . . . . . 65
3.3.1 Room conditions of tests . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
3.3.2 Composition and structure . . . . . . . . . . . . . . . . . . . . . . . . . 65
3.3.3 Mechanical properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
3.3.3.1 Tensile test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
3.3.3.2 Kodak rigidity . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
3.3.3.3 Water Stretch Dynamics analyser (WSD) . . . . . . . . . . . . 67
3.3.3.4 Finch’s test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.3.4 Water absorption and wettability . . . . . . . . . . . . . . . . . . . . . . 68
3.3.4.1 Surface roughness . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.3.4.2 Static contact angle: Sessile Drop Method . . . . . . . . . . . 69
3.3.4.3 Dynamic Contact Angle (DCA) . . . . . . . . . . . . . . . . . 70
57
58 CHAPTER 3. MATERIALS AND METHODS
3.3.4.4 Cobb test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

3.3.4.5 Water absorption by paper discs stack . . . . . . . . . . . . . . 72
3.3.4.6 Dynamic Penetration Measurement (DPM) . . . . . . . . . . . 72
3.3.5 Coating process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.3.5.1 Coating pilot . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.3.5.2 Temperature measurement . . . . . . . . . . . . . . . . . . . . 75
3.3.5.3 Web tension measurement . . . . . . . . . . . . . . . . . . . . 76
3.3.6 Wrinkles and buckling . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
3.3.6.1 Digital Image Correlation (DIC) . . . . . . . . . . . . . . . . . 77
3.3.6.2 Friction coefficient . . . . . . . . . . . . . . . . . . . . . . . . . 77
3.3.6.3 X-Ray tomography . . . . . . . . . . . . . . . . . . . . . . . . 77
3.3.6.4 SEM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
3.3.6.5 Wrinkle profile determination . . . . . . . . . . . . . . . . . . . 78
3.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
This chapter will present the materials and techniques used in this project. The four papers
used will be presented as well as the chemicals used to perform contact angle tests. The tech-
niques for preparing and characterising the starch gel and the coating color will be presented.
Finally, all the equipment and techniques used to produce the coated paper and to characterise
it (with or without coating) will be detailed.
3.1 Materials
3.1.1 Papers
In this project, four different papers are tested. The common point between them is the low or
very low grammage they have. General characteristics of these papers are:
→ GHDS 35 and GHDS 40 are one-side coated glossy papers used for flexible packaging,
made with hardwood and softwood. The coated side is made of calcium carbonate, kaolin
and SBR Latex. GHDS 35 is a 35 g.m−2 paper and GHDS 40 is a 40 g.m−2 paper.
Compared to GHDS 40, GHDS 35 provides a 12.5% weight reduction enabling a lower
environment footprint impact. It also has good performance during lamination, printing
and converting. Main applications include pouches or sachets, for food and non-food
end-uses such as sugar, coffee, aromatic herbs or dehydrated products. These papers are
suitable for extrusion with PE or lamination with alufoil and PET. They can be printed
after lamination with flexo and helio printing. There lower basis weight also facilitates
sealing process according to the manufacturer.
→ CPRAW is a notebook paper, i.e. a paper for rolling cigarettes. This grade is made of
100% chemical unbleached softwood pulp. The pulp is highly refined and no fillers are
introduced into the paper. With its very low grammage of 14 g.m−2 , the CPRAW paper
has very little opacity. This paper has not been treated for wet strength and it is not
waterproof.
3.1. MATERIALS 59
→ CPBS is a cigarette paper used in cigarette making machines. It is composed of a mixture

of so-called textile pulps (linen pulp and hemp pulp), bleached eucalyptus pulp and calcium
carbonate as fillers. With its very low grammage of 23.5 g.m−2 , this paper is laid: the drip
roller on the flat table is marked with a striped pattern that leaves a watermark in the
paper, characteristic of cigarette papers. This paper has not been treated for wet strength
and it is not waterproof.
An important point was mentioned: the project was supposed to be done only with GHDS
35. However, it was necessary to order new GHDS 35 reel, but the manufacturer had stopped
production and offered GHDS 40. In addition, an opportunity emerged at the end of the project
to have one CPRAW reel and one CPBS reel with another manufacturer. This is why the project
studies were mainly carried out on GHDS 35.
3.1.2 Starch
The starch used is a STABILYS A025-thermally modified potatoes starch from Roquette (France).
It is a white and odourless powder. It has a pH of 6.2 in a 20 %wt water suspension. Its relative
density is 0.5 and its solubility in water is 350 g.L−1 . The median grain size is 18.7 µm accord
to manfufacturer (NFX 11-666 standard). The formulation used in this project is then made of
a starch gel (starch cooked in water), at different mass fraction, with uncooked starch particles,
at different mass fraction.
3.1.3 Chemicals used for contact angle measurement
Liquids used for contact measurements are: Hexadecane (CAS: 544-76-3) and 1-Bromonaphtalene
(CAS: 90-11-9) are from Fischer Scientific. Glycerol (CAS: 56-81-5) is from Sigma Aldrich.
Ethanol (>99.8%) (CAS: 64-17-5) is provided by Honeywell. Ethylene Glycol (CAS: 107-21-1)
is from Acros Organics. Home-made distilled water is used. The surface tensions, dispersive
and polar components of these test liquids are given Table 3.1.
Polar Dispersive
Surface tension
Tested liquid component component
(mJ.m−2 )
(mJ.m−2 ) (mJ.m−2 )
Distilled water 72.8 51 21.8
Glycerol 64 30 34
Hexadecane 27.5 0 27.5
Ethylen glycol 48 19 29
1-Bromonaphtalene 44.6 0 44.6
Table 3.1: Surface tension, polar and dispersive components of liquids used during contact
angle measurements [63].
3.2 Coating color

In this project, a difference will be made between the starch gel and the coating color. Starch gel
is starch cooked in water. The coating color consists of starch gel with uncooked starch grains
added.
3.2.1 Preparation of the starch gel
In order to prepare the coating color, a cooking robot (Moulinex Companion XL) is used. The
temperature is adjustable from 30 to 150 °C by steps of 5 °C. Several rotation speeds are available
but are given in arbitrary units. Figure 3.1 shows the blade used for mixing during cooking.
The blades are 50 mm high with a base diameter of 160 mm and a top diameter of 170 mm.
They are inclined at 45° and the percentage of empty space is about 35 %. It is also possible to
set the cooking time. The cooking parameters are chosen following the advice of the supplier:
30 min of cooking at 95 °C at the slowest robot speed at a mass concentration 15 wt% and 20
wt%
Figure 3.1: Blade used to cook and blend starch gel in cooking robot.
The cooked starch mass fraction Φbinder in the starch gel is calculated as presented in equation
3.1.
mbinder
Φbinder = (3.1)
mbinder + mwater
where mbinder is the mass of starch used as a binder and mwater is the mass of water used
for the coating color.
The mass fraction of uncooked starch grains in the coating gel Φparticles is calculated ac-
cording to equation 3.2 using mparticles as the mass of uncooked starch particles in the coating
color:
mparticles
Φparticles = (3.2)
mbinder + mwater + mparticles
3.2.2 Devices and methods for characterising the coating color
The starch and the starch gel are characterised following a wide range of techniques to have
different useful information for the project: rheological behaviour, water absorption and surface
properties.
3.2. COATING COLOR 61
3.2.2.1 Determination of the starch grain sizes
To confirm the supplier specifications, starch particles sizes are measured in water using a laser
granulometer (Cilas particle size analyser 1190) based on the Fraunhofer method . During the
test, ultrasonic system is on to maximise the dispersion of the particles and avoid aggregates.
The obscuration level is set around 15 points to obtain a good measure while slowly adding
starch particles. In addition, Scanning Electron Microscope (SEM) pictures of dry starch grains
are taken to have additional information about the size and shape of the grain. A FEI Quanta
250 (SEM) is used equipped with a field emission gun in ETD (Everhart–Thornley Detector) or
SDD (Silicon Drift Detector) mode.
3.2.2.2 Sedimentation studies
To study the behaviour of starch grains in water, a sedimentation study is performed. It will help
to know if during the coating process, starch grain could sediment and then lead to heterogeneity
when deposited on the paper. It consists in suspending a known quantity of uncooked starch
grains in water. Different phases will appear over time which are linked to particles sizes. The
test is carried out in a 1 L graduated test tube at 20.5 °C as shown Figure 3.2. After a time
Figure 3.2: Sedimentation test evolution at 2 different times : t and t + ∆t.
∆t, sedimentation occurred. With no evaporation, h0 is total height, h0 = h1 + h2 , where h1

is the height of the first sedimentation phase and h2 is the height of the second sedimentation
phase. During the study time, the heights of the various phases are measured by taking pictures
at different times, using a Sony Alpha 6000 with a 35-70 (f 3.5 -4.5) lens on a tripod.
3.2.2.3 Differential Scanning Calorimetry (DSC)
DSC measurements are performed to characterized the temperature at which starch grains ex-
ploded. This helps to ensure a good coating quality while being certain to be at the best cooking
temperature. This is done with a DSC Q100 (TA Instruments, NewCastle) filled with liquid
nitrogen cooling system. A suspension of starch in water (Φbinder = 20 wt%) is placed in a

hermetically sealed pan. Each sample is heated from 5 °C to 145 °C with a heating rate of 10
°C.min−1 . Measurements are done in duplicate to ensure repeatability. The sample weight was
between 15 and 25 mg.
3.2.2.4 Starch gel cooking control
To optimize the starch gel cooking, the cooking is followed by optical microscopy . The aim
of the method is to take a sample at different cooking times. A drop of this sample is placed
on a microscope slide as shown Figure in 3.3. The microscope used is a Zeiss microscope Axio
Imager M1 with AxioVision SE64 software.
Figure 3.3: Highlighting of the cooking evolution of starch particle by optical microscope
analysis.
Particular attention is paid to the quantity deposited. Indeed, a too important quantity
of starch grains deposited on the slide would lead to obtain several layers of starch grains and
would interfere with the observation. While the cooking progressed, less and less grain should
be observed. When no more grains are seen, the cooking is done. It is a complementary measure
to the DSC.
3.2.2.5 Rotational rheological measurement of the starch gel
To determine the rheological behaviour of the fluids used during coating, rheological measure-
ments are performed. The used device is an Anton Paar MCR 302 rheometer in a cone-plan
set up (CP50-1 measuring cone) (Figure 3.4) and with Anton Paar 1.25 software. The mea-
surements are done in a steady state regime and in a saturated moist environment to minimize
evaporation.
The starch gels were tested at different shear rates: from 0.1 to 10000 s−1 and at different
temperatures: from 20 to 50 °C. The gap between the measuring cone and the heating plate is
set to 102 µm. Two starch mass fractions Φbinder are tested: Φbinder = 15 %wt and 20 %wt.
An average viscosity value is measured by considering at least 200 points on the stabilisation
viscosity plateau. The results will be used to follow the viscosity evolution at different shear
rate and temperature. They will be compared with different models from literature.
3.2. COATING COLOR 63
Figure 3.4: Rotationnal rheological measurement with a cone-plan system.
3.2.2.6 Measure of the starch gel density
To forecast particles sedimentation in starch gel, density characterisation is realized. The ex-
periment will be performed in two stages: the volume of the density probe is determined with
a liquid with known density (calibration step), hereafter the density of unknown liquids can be
determined as illustrated in Figure 3.5 while recording the mass variation over time and using
Archimede’s principle. The device used is a Sigma 701 tensiometer from Biolin Scientific with a
12 mm diameter glass sphere. It is needed to heat the starch gel at 40°C to avoid formation of
a starch skin link to gelatinization.
Figure 3.5: Density measurement of a 20 %wt starch gel with a glass sphere.
3.2.2.7 Determination of starch gel surface tension
To study the wettability of paper by the starch gel, it is mandatory to know its surface tension.
To perform this measurement, the device used is a Sigma 700 tensiometer from Biolin Scientific
with Du Noüy ring [129] or Wilhelmy plate . Whatever the chosen probe, it should be cleaned
with ethanol and flamed prior to use (temperature > 1000 °C).
The De Noüy ring method uses the interaction of a platinum ring with the surface of the
tested liquid, starch gel here. The ring is submerged below the starch gel /air interface. After
immersion, the beaker with the starch gel gradually decreased and the ring pulls up the meniscus
of the liquid. During all the operation, the force acting on the support is recorded. The Wilhelmy
plate method follows the same principle. It utilizes the interaction between a platinum plate
Figure 3.6: Use of a platinum Wilhelmy plate (19.5 mm x 0,12 mm) to measure starch gel
surface tension with 5 round-trip cycles.
(the Wilhelmy plate) and the surface of the starch gel as shown Figure 3.6. In this method the
position of the probe relative to the surface is significant. As the surface is brought in contact
with the probe liquid, the force sensor will detect this by the change in force it experiences
(Figure 3.7).
Figure 3.7: Highlighting of the phenomena that occurs during a surface tension measure: (1)
the plate is at its "zero" position; (2) the plate is in contact with the liquid, the advancing
contact angle can be recorded; (3) the plate continues to go deeper in the liquid; (4) the plate
is going back to its "zero" position, the surface tension is measured.The receding contact angle
can be recorded [25].
The plate is then wetted to a set depth. When the plate is later returned to the zero depth
of immersion, the force is registered. In order to obtain significant value, 10 round trips in
the starch gel are performed. Moreover, the starch gel can quickly become very viscous and
gelatinize. To avoid this phenomenon, the beaker is set in a thermostatic vessel at 40 °C. The
gel has a starch mass fraction of 20 %wt. The test depth is set to 6 mm and the test speed is
20 m · min−1 .
3.3. DEVICES AND METHODS FOR CHARACTERISING THE PAPER 65
3.3 Devices and methods for characterising the paper

3.3.1 Room conditions of tests
Some of the tests shown below were done in conditioned room at 23.5 ±1 °C and 50 ±5HR%.
Moreover all the samples are stored at least 24 h in these conditions before any testing. Table
3.2 sum up which tests are done in conditioned room.
Done in conditioned
Tests
room
Thickness and grammage determination Yes
Mercury porosimetry No
Tensile test Yes
Kodak rigidity No
Water stretch dynamic analyser No
Finch’s test No
Surface roughness measurement No
Static contact angle measurement No
Dynamic contact angle measurement No
Cobb test Yes
Water absorption by paper discs stack No
Dynamic penetration measurement Yes
All experiments done in the coating pilot, including
No
temperature and web tension measurement
Digital image correlation during tensile test Yes
Measure of the friction coefficient Yes
X-ray tomography No
Scanning electron microscope No
Wrinkle profile determination No
Table 3.2: Tests carried out in conditioned or unconditioned rooms
3.3.2 Composition and structure

In order to determine the characteristics of the papers used, the following methods are used.
The grammages G of the samples are determined in accordance with ISO 536. The thicknesses
tf of the samples are measured following ISO 534. For thickness and grammage, 10 samples are
measured.
The pore size distribution and surface area of the papers are measured using mercury
porosimetry from Micromeritics AutoPore IV 9500 V1.09 (located at Institute of Environmental
Geosciences, Grenoble). The sample weight are 460 and 470 mg.
3.3.3 Mechanical properties

3.3.3.1 Tensile test
To carry out the test, an Instron 5965 testing system of 5 kN maximum force (Instron, USA)
is used, with an accuracy of 0.5% of the read value. Samples with dimensions of 15 mm width
and 100 mm length are tested with a constant rate of elongation set at 10 mm.min−1 . For each
paper, Cross Direction (CD) and Machine Direction (MD) are tested, 10 times each to have
representative average values and standard deviations. The geometry allows to be far enough
from the jaws and allows the Poisson effect.
The values obtain using tensile test are also used to determine the Poisson’s ratios according
to two different methods: an empirical model equation and one from mechanics for orthotropic
materials. The empirical model equation is used by Biancolini [130] and was first proposed
by Baum [131][132]. This empirical correlation bounds shear modulus and Poisson ratios as
presented equation 3.3 and equation 3.4.
Gxz,Baum = 0.387 Ex · Ez (3.3)

p
s
Ez
νxz,Baum = 0.293 (3.4)
Ex
Gxz,Baum is the shear modulus in the xz plane, νxz,Baum is the Poisson ratio, Ex is the
Young’s modulus in MD and Ez is the Young’s modulus in CD.
The other method comes from the constitutive equations of mechanics for an orthotropic ma-
terial and is proposed by Viguié [133].Thus, the experimentally measured Etan can be expressed
as presented equation 3.5.
1 1 −νxz 1 1

= · cos4 ϕf + 2 · + · sin2 ϕf · cos2 ϕf + · sin4 ϕf (3.5)
Etan Ex Ex Gxz Ez
In order to obtain Gxz and νxz , it is necessary to measure the tangent modulus of paper
sample Etan at a same strain. Etan is the slope, in the elastic domain, of the nstress versus
strain plot obtained during the tensile test. The fibres in the sample are not aligned with the
measurement direction: ϕf is the angle between the tested direction and MD. Once all the moduli
are measured, a fitting between the experimental values and the modelled values is performed
by the least squares method.
3.3.3.2 Kodak rigidity
The "Kodak rigidity" method enables the determination of the bending stiffness of a test piece
clamped at one end by measuring its vibrational resonant length. The vibrational resonant
length of a material, in one direction, when it is subjected to vibration of a given frequency
depends on its bending stiffness in that direction and its grammage. To determine the bending
stiffness of the sample, a resonance method is used following the ISO 5629 standard as shown
Figure in 3.8. The bending stiffness S (mN.m) is obtained as presented 3.6.
4
lres G

Sbs = 20 · · (3.6)
100 100
where lres is the resonance length (mm). However, this method leads to a high dispersion
coefficient for the lres values even if the precision the precision is around 0.1 mm.
Figure 3.8: The resonance method. Resonance corresponds to maximum deflection δ and w to
the out of plane displacement [134].
3.3.3.3 Water Stretch Dynamics analyser (WSD)
The device comes from Emtec Electronic GmbH (Leipzig, Germany). It consists of measuring
the elongation of the paper while it is in contact with water on one side only as explained in
Figure 3.9. The paper sample has a width of 60 mm for a length of 210 mm and the water
surface is about 55 x 70 mm2 . The duration of the test is 120 s and the strain measurement
starts as soon as the paper is in contact with water. This test is carried out at constant stress
(assuming negligible variations in paper cross-section). MD and CD are tested as coated and
uncoated face in contact with water.
Figure 3.9: WSD (Water Stretch Dynamics analyser) test principle.
For each combination, 4 different forces could be applied using an adjustable lever arm to
create the torque: 1, 1.5, 2 and 2.5 N. For each chosen force, a combination is tested 5 times.
3.3.3.4 Finch’s test
The Finch’s test is used to measure the tensile strength of paper after a 20 s immersion in water.
The sample is wrapped around a horizontal dry rod and blocked in the upper jaw.The wrapped
part is immersed in water and removed from it. Then the tensile test begin. It follows the
ISO-3781 standard. The test is performed with a 15 mm width paper strip. The strip is cut to
have a length between the dry horizontal rod and the upper jaw of 150 mm as presented Figure
3.10.
Figure 3.10: Finch’s test principle.
The immersed length of the strip should be between 25 and 50 mm. The test is performed
at a speed of 10 m · min−1 . The samples are tested 10 times in the MD and in the CD.
3.3.4 Water absorption and wettability
3.3.4.1 Surface roughness
The roughness is determined using the Alicona InfiniteFocus (Alicona, Alicona Imaging GmBH,
Raaba, Austria). It is an optical system which takes pictures of a sample at different heights and
then reconstructs a 3D view of the sample from the stack of taken images. It is used for dimen-
sional measurements, surface analysis and characterisation and allows measurement of textures
and profiles, according to ISO 25178 standard. In the present study, the measured value is the
Sdr (Developed Interfacial Area Ratio). It is expressed as the percentage of additional surface
area contributed by the texture, as compared to an ideal plane, the size of the measurement re-
gion. To obtain the Sdr value it is necessary to apply a cut-off filter to separate micro-roughness,
roughness, waviness and form as explained Figure 3.11.
Figure 3.11: Highlighting of the different scales that could be measured while focusing on
surface roughness.
The cut-off value is chosen from the RSM (mean spacing of profile irregularities of primary
profile). It follows the DIN EN ISO 4288, ASME B46.1 standards and Table 3.3 indicates the
link between the RSM and the cut-off valueλc .
Periodic profile RSM (mm) Cut-off filter λc (mm)

0.013 < RSM ≤ 0.04 0.08
0.04 < RSM ≤ 0.13 0.25
0.13 < RSM ≤ 0.4 0.8
0.4 < RSM ≤ 1.3 2.5
Table 3.3: Cut-off values as a function of periodic profile RSM
3.3.4.2 Static contact angle: Sessile Drop Method
To know the behaviour of the liquid drop on the papers studied in this thesis and to determine
their Surface Free Energy (SFE), the sessile drop technique is used. By placing a droplet of
liquid with a known surface energy, the contact angle over time is recording. While knowing
the surface energy of the prob liquid, the SFE is determined following the Owens, Wendt, Rabel
and Kaelble (OWRK) method [61] [62]. The use of several different probe liquids is required.
In the present case, hexadecane, 1-bromonaphtalene, glycerol, ethylene glycol and home-made
distilled water are used.
The tested paper is placed under the needle tip of a syringe which deposited 5 µL liquid
droplets on the sample as shown Figure 3.12. To minimize shrinkage, the paper sample is fixed
with tape at its two edges. The “static” contact angle of the droplets is measured over 1 minute
with an acquisition frequency of 20 frames per second. Due to the device, there is an initial time
of 3 to 10 s between probe liquid-paper contact and the start of the measurement. The optical
system used is the OCA 20 goniometer from Dataphysics Instruments GmBH, Filderstadt,
Germany. The device provides a ±0.2° reading accuracy. Then, the movie obtained is analysed
with SCA 20 OCA Control software (Dataphysics Instruments GmBH, Filderstadt, Germany)
to determine the contact angle of the drop on the paper over time. 5 tests are done for each
Figure 3.12: Evolution of a contact angle over time during a sessile drop method test.
liquids : it enables to drown a average curve for each liquids.
3.3.4.3 Dynamic Contact Angle (DCA)
This technique is used to measure the receding and advancing contact angle and to demonstrate
hysteresis-like behaviour of a material during its entry and exit into a liquid [25]. The DCA
method follows the same principle as the surface tension measurement of a liquid. The difference
lies in the object of the measurement. In the surface tension measurement, the properties of
a liquid are determined with a Wilhelmy plate or a de Noüy ring, with a known materials
(platinum). In the DCA technique, the liquid is known and the material characteristics (plate)
are unknown.
In the present case, the paper samples are cut to have a width of half an inch (12.7 mm). The
length of the sample must be greater than 15 mm in order to be sufficiently wetted. However,
for DCA test, it is necessary to have symmetry of the material and the same surface everywhere.
Therefore, to minimise the problems associated with asymmetry of the samples, two pieces of
paper are stuck together, with the same two sides facing outwards, using double-sided tape as
shown Figure 3.13.
Figure 3.13: Size and organisation of a typical paper sample for DCA measurements.
It must be kept in mind that the tape will influence the behaviour of the sandwich: it is
the perimeter P = 2 · width + 2 · thickness, of the sample that is considered in determining the
contact angles. So, to minimise the impact of the tape, the following condition is imposed on
the double side tape thickness tdst (equation 3.7).
2tdst ≤ 0.01P (3.7)
For this test, 5 up and down cycles are done per sample. The test liquids are water and/or
starch gel. The test speed is 5 m · min−1 and the immersion depth of 3 mm.
3.3.4.4 Cobb test
The Cobb test follows the ISO 535 standard. The principle of this test is to measure the amount
of water that a paper can absorbed during a given time, on a given surface. To do so, the
apparatus showed Figure 3.14 is used. A 150 mm square sample is set and blocked under a
metallic ring. This ring is locked to avoid water leaks.
Figure 3.14: Cobb test principle
The tests are done at different contact times in a conditioned room with conditioned paper
at 23.5 ±1 °C and 50 ±2 % HR. The Table 3.4 sum up the used contact time.
Water excess Drying start time

Trial time (s) Symbol
removal time (s) (s)
30 Cobb 30 20 ±1 30 ±1
60 Cobb 60 45 ±1 60 ±2
120 Cobb 120 105 ±1 120 ±2
300 Cobb 300 285 ±1 300 ±2
1800 Cobb 1800 1755 to 1815 1800 ±15
Table 3.4: Test duration for Cobb tests according to ISO 535 standard.
To determine the water absorptiveness of paper As (gwater · m−2

paper ), the equation 3.8 is used.
mwet − mdry
As = (3.8)
Spaper
where mwet is the wet mass, in grams of the tested sample, mdry is the dry mass, in grams,
of the tested sample and Spaper is the tested area (0.01 m2 ).
3.3.4.5 Water absorption by paper discs stack
Capillary rise or Washburn method can be done by using force tensiometer and a special type
of sample holder. The measurement set up is shown in Figure 3.15. The sample holder is a
cylinder with holes in its base, filled up with 15 mm diameter paper discs . It is immersed into
water and the amount of water absorbed into the paper bed is measured against time. One side
of the tested paper is coated, the other not: thus, every discs of one trial are set coated side face
up or face down.
Figure 3.15: Setting up the water absorption measurement by stacking paper discs on top of
each other and immersed it in water.
Different tests are carried out with different numbers of discs. Note that a paper disc has a
mass of 7 mg and the balance has a precision of 1 mg. Thus, with only one disc, the measurement
is already at the edge of the precision of the apparatus. The immersion depth is set to 0.50 mm,
the sampling interval to the minimum (around 10 Hz). The measure lasted 30 min maximum
and between 1 and 100 paper discs are used during one trial.
3.3.4.6 Dynamic Penetration Measurement (DPM)
This test enable to determine liquid penetration at very short time using the evaluation of
ultrasound transmitted trough paper thickness during wetting. The sample is fixed on a sample
holder with double-sided tape. When the test starts, the sample holder is immersed into the
test liquid (water here). When the immersion is complete, a high frequency (1 or 2 MHz), low
energy ultrasound signal is transmitted through the sample in thickness direction as explained
Figure 3.16).
During all the test, the transmitted signals are recorded over time. There are two types
of sample holders: a flat sample holder where backside is completely covered with the tape; a
grooved sample holder where the sample is not in contact with the tape in the grooves. Three
tests are carried out with the flat sample holder (2 with coated side in contact with water and
1 with uncoated side) and 3 others on the grooved sample holder (2 with coated side in contact
Figure 3.16: DPM measurement principle [75].
with water and 1 with uncoated side).
3.3.5 Coating process
3.3.5.1 Coating pilot
The coating laboratory pilot used is from RK Print, UK. This laboratory driver is versatile as it
allows different types of techniques to be used for coating flexible films such as plastics or paper.
In this study, the material used is paper and the coating technique chosen is the Mayer bar
coating. Figure 3.17 represents a simplified diagram of the pilot. Appendix A shows a complete
diagram with dimensions.
Figure 3.17: RK laboratory coating pilot.
The coating section consists of a tank for the coating color. This is where an applicator
roller dips to collect the coating color. The color is deposited on the paper by contact. In order
to ensure good contact between the paper and the applicator roller, a pressure roller is placed
upstream of the coating zone so that the paper is pressed properly against the applicator roller
(Figure 3.18.
Figure 3.18: The coating station with the different coating tools.
It is once the color has been applied to the paper that the Mayer bar comes into play. The
purpose of the Mayer bar is to metering the coating color onto the paper web, to remove excess
color from the paper. Rods are available in a wide variety of wire sizes to give a range of coating
weights. Table 3.5 sum up the rod characteristics.
Wire diameter Wet film thickness

Code
(µm) (µm)
.002 0.05 4
.003 0.08 6
.006 0.15 12
.012 0.30 24
.020 0.51 40
Table 3.5: Link between Mayer rod characteristics and the thickness of the wet film deposited.
Different widths of applicator rollers are available: 230, 290, 300 and 310 mm. Two pneumatic
vane motors drive either the applicator roller and the Mayer bar. The working pressure of the
motors is between 3.5 and 4 bars.
To dry after coating, an infra-red drying ramp is installed on the web path. This ramp
consists of 5 independent infra-red lamps, each with a power of 500 W. Therefore, the total
available power is 2.5 kW. Two hot air dryers are installed just after the infra-red dryer to
complete the drying process. These dryers can heat the air up to 200 °C and are temperature-
controlled (unlike the infra-red dryer).
The drive part consists of a variable speed DC main drive motor rated at 0.56 kW coupled
to a 10:1 ratio worm gearbox giving a final web speed from 5 to 50 m · min−1 . A toothed
belt transmission connects the core holder and a pneumatic clutch. The pneumatic clutch is
connected to the motor by an angle gear. The working pressure of the pneumatic clutch is
between 1.5 and 4 bars. The roller preceding the winding is a speed measuring roller. This
control has been installed during this project because it hasn’t been available previously.
On the unwinding section, the used paper reel should have an intern diameter of 76.2 mm
(3 inch) to be used on the coating pilot. The minimum web width is 220 mm and the maximum
width is 305 mm. The unwinding station is equipped with a manually adjustable brake for
controlling the web tension.
3.3.5.2 Temperature measurement
In order to carry out the various tests about temperature measurement, sensors from Micro-
Epsilon (CT-SF15-C1) were used [135]. The spectral wavelength range of these sensor is 8 to
14 µm. The measuring temperature range is -50 °C to 600 °C, for non-metallic materials. The
optical resolution of the sensors is 15:1 (ratio between sensor-object distance and spot size at
the focus point). Their accuracy is ±1 °C, with a given response time of 150 ms. Note that the
electric heating time is 10 min. The emissivity of the sensors can be adjusted between 0.1 and
1. Finally, the sensors can operate in working environments between -20 °C and 180 °C. The
emissivity of the paper is considered equal to 0.95.
Figure 3.19 is the test bench set up in order to carry out different tests. It consists of a flat
base on which a Peltier device is placed and used as a heating source. A sample of paper of 30
x 30 mm2 is fixed on the plate. On the base, a rail is fixed on which the sensor can move. The
rail can be tilted in a perpendicular plan to the base, allowing to change the angle between the
sensor and the paper sample.
Figure 3.19: Non-contact IR temperature measurement system.

The impact of the measuring height hs , the measuring angle κs and the sensor emissivity
εs are studied. The measuring height hs varied from 70 to 300 mm, the measuring angle κs
varied from 0 ° to 60 ° and the emissivity εs from 0.1 to 1.0. All the tests are made at different
temperatures: from ambient temperature 21.3 °C to 190 °C.
Once the IR temperature sensors characterised, the goal is to see where, on the paper width,
after an IR dryer on a coating pilot, it is the most relevant to fix the sensor. In order to do so
a thermal camera (FLIR - i7) is used to obtain the various images presented in this document.
With a resolution of 140 x 140 pixels, the camera works on a spectral wavelength range from
7.5 to 13 µm. The camera measurement accuracy is ±2 °C and can take measurements between
-20 °C and 250 °C. The images are taken at the output of an infra-red dryer fixed on a coating
pilot (Figure 3.20).
Figure 3.20: R pictures taking with Flir-i7 (Web speed: 5 m · min−1 .
The images are taken in such a way that the whole width of the paper (300 mm) is visible.
About the image processing, the Fiji software is used in order to obtain the different temperature
profiles for each picture. The web speed varies from 5 m.min−1 to 40 m.min−1 . All the 5 lights
of the IR dryer are switch on. The web tension is set to 20 N.m−1 . The temperature of the pilot
room varies from 15 to 35 °C and the relative humidity from 30% to 65%.
3.3.5.3 Web tension measurement
In order to measure the paper web tension, a web tension measuring roller is installed on the
coating pilot. The system comes from Erhardt & Leimer. First, a PD30 sensor roll is mounted on
the coating pilot. This measures the radial forces that the web exerts on the sensor roller in the
direction of measurement. The sensor roller consists of a roller in contact with 2 Wheatstone
bridge-type force sensors. These measuring elements are formed as double bending bars on
which the extension measuring strips are placed. Switched together, they form a measuring
bridge. A mechanical lock also protects the system from overloading. The second component of
the tape tension measurement is a DC62 tape tension regulator from the same supplier. This
system allows the signal measured by the PD 30 roller to be recovered and the web tension to
be accessed in real time via the DC62 display. This system would also allow the installation of
a web tension control via a brake fixed on the core holder of the master reel.
3.3.6 Wrinkles and buckling

3.3.6.1 Digital Image Correlation (DIC)
The wrinkle formation mechanism is reproduced on an Instron 5965 tensile testing machine to
recover displacement and deformation information thanks to 2D and 3D image correlation. Sony
Alpha 6000 with a 35-70 (f 3.5 - 4.5) lens is used for 2D image correlation and two Canon EOS
1200D DSLR (Digital Single-Lens Reflex) cameras with Canon EFS 24mm 2.8 STM lenses for
3D stereo-correlation (3.21.
Figure 3.21: DIC system installed on the tensile test (left) and on the coating pilot (right).
The software used for the tensile is 7D (developped by Pierre Vacher, from Université Savoie
Mont-Blanc). For the DIC study, speckled samples are used. The sample are tested in CD and
MD. In the 7D software, an area of interest is defined per set of photos. The algorithm for
correlating images is a combination of Harris and Pyramid algorithms. In addition, to carry out
the DIC tests, a random speckle is printed on the industrially coated side. These speckles are
needed to make difference between each studied point on the samples. The printing is carried
out with a Xerox laser printer at a resolution of 600 DPI. Printed papers are also characterised
and compared with non-printed papers.
3.3.6.2 Friction coefficient
The surfaces to be tested are placed together in plane contact and under apparent uniform
contact pressure. The forces needed to initiate sliding (the static friction force) and to slide the
surface relative to each other (the kinetic friction force) are recorded. This follows ISO-15359
standard.
3.3.6.3 X-Ray tomography
A sample of a wrinkled paper and coated with the starch gel is sent to the company Novitom
for X-ray tomography on a line at the ESRF (European Synchrotron Radiation Facility) in
Grenoble. The sample has a thickness of approximately 35 µm. The voxel resolution is 0.7 µm
on each direction.
3.3.6.4 SEM
SEM pictures of wrinkle and non-wrinkle paper are taken to have more information. A FEI
Quanta 250 tungsten SEM is used for the microscopic observation.Both secondary and backscat-
tered electrons are used to perform the images.
3.3.6.5 Wrinkle profile determination
The roughness is determined using the Alicona InfiniteFocus (Alicona, Alicona Imaging GmBH,
Raaba, Austria) (cf.3.3.4.1). For the wrinkle characterisation, a wrinkle paper is set under the
objective and then pictures are taken in order to obtain the profile of the paper as Figure 3.22.
Figure 3.22: Example of result that could be obtain with Alicona InfinteFocus device.
3.4 Conclusion
The different techniques used in this project were presented in this chapter. They will be used
in the following chapters presenting the results of this project. The techniques of rheology,
sedimentation, wettability and water absorption will be used first in the following chapter.
Then, techniques related to the mechanical behaviour of papers will be used. Finally, the
process aspects will be discussed, with the related techniques, in the last results chapter of this
document.
Chapter 4
Absorption of the coating color by papers
Contents
4.1 Properties of the coating color . . . . . . . . . . . . . . . . . . . . . . 80
4.1.1 Preparation of the coating color . . . . . . . . . . . . . . . . . . . . . . . 80
4.1.1.1 Control of starch cooking . . . . . . . . . . . . . . . . . . . . . 80
4.1.2 Rheology of the starch gel . . . . . . . . . . . . . . . . . . . . . . . . . . 83
4.1.2.1 Shear rate dependency of the starch gel viscosity . . . . . . . . 83
4.1.2.2 Temperature dependency of viscosity . . . . . . . . . . . . . . 85
4.1.3 Sedimentation of starch . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
4.2 Study of the wettability and water absorption by papers . . . . . . 91
4.2.1 Study of the papers wettability . . . . . . . . . . . . . . . . . . . . . . . 92
4.2.1.1 Contact angle betwen paper and liquids . . . . . . . . . . . . . 92
4.2.1.2 Surface free energy of paper . . . . . . . . . . . . . . . . . . . 97
4.2.1.3 Surface tension of starch gel . . . . . . . . . . . . . . . . . . . 99
4.2.2 Papers water absorption analysis . . . . . . . . . . . . . . . . . . . . . . 100
4.2.2.1 Paper pore sizes . . . . . . . . . . . . . . . . . . . . . . . . . . 100
4.2.2.2 First approach: Cobb method (ISO 532-5) . . . . . . . . . . . 101
4.2.2.3 Second approach: DPM . . . . . . . . . . . . . . . . . . . . . . 102
4.2.2.4 Third approach: Paper discs stack . . . . . . . . . . . . . . . . 105
4.3 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
In this first chapter of results, the preparation of the coating color and its characterisation
(as well as its compounds) will be presented. It is fundamental insofar as it plays an integral
part in optimising the coating process. In a second step, the phenomena of wettability and water
absorption by the paper will be discussed. These are important insofar as they are involved in
the coating process and can impact the mechanical properties of the paper. Please note that, in
this chapter, only the GHDS 35 paper have been tested.
79
80 CHAPTER 4. ABSORPTION OF THE COATING COLOR BY PAPERS
4.1 Properties of the coating color

Before looking at the penetration of liquids into the paper, it is important to know the fluids used.
This section will focus on the starch gel especially dealing with its cooking, the sedimentation
of starch particle in water and the rheology of the coating color under different conditions.
4.1.1 Preparation of the coating color
4.1.1.1 Control of starch cooking
In order to prepare the coating color used in this project, it is necessary to cook the starch in
water to make a starch gel. The cooking of the starch gel must be optimal in order to ensure
an explosion of all the starch grains in the water. For this purpose two techniques presented
in the materials and methods section are used. The first one is an optical microscopy of the
cooking of the grains. The second method is the use of DSC. Both methods aim to determine
the temperature at which the explosion of the starch grains is effective.
Thus, different images were taken by optical microscopy in order to follow the cooking of the
starch in water. The images presented Figure 4.1 were made with a concentrated starch gel at
15%wt at different temperatures.
When the suspension is cooked at 40.0 ±0.1 °C. The slide is covered with starch grains, and
on different planes: a superposition of grains is visible. The image 4.1b shows a suspension
being cooked at 54.5 ±0.1°C. Black areas surrounding the starch grains are visible but remain
very few. The image 4.1c was taken for a suspension at 64.0 ±0.1°C. The number of grains in
the picture is much lower than in the two previous pictures. Moreover, the image is blurred:
this is due to the heat of the fluid, leading to a movement of the particles. The image 4.1d is
the suspension being fired at 75.0 ±0.1°C. It has even less visible starch particles compared to
the previous image. Shapes in the background are still present. However, these are misshapen
and difficult to identify. The last image (Figure 4.1e) is the suspension at 86.0 ±0.1°C. It is very
blurred. It corresponds to a moving fluid without any particle shape. The suspension is cooked
and has become the desired coating color. Then it is possible to say that the cooking and the
total explosion of the grains takes place at a temperature between 75.0 ±0.1°C and 86.0 ±0.1°C.
In order to obtain more precise information about the temperature required to explode the
grains, another technique was used: DSC. The latter was also presented in the materials and
methods section. The measurements were carried out with a starch suspension in water at an
initial mass fraction of 20%wt. A triplicate was performed. The curves of the measurements are
shown in Figure 4.2.
The measured heat flow is plotted as a function of the sample heating temperature. The
reader will immediately notice the difference between the "starch 20%wt_1" curve and the
"starch 20%wt_2" and "starch 20%wt_3" curves. We spoke of the initial mass fraction: this
corresponds to the mass fraction in the container where the suspension was made. Although
care was taken to have a homogeneous mixture, the samples taken for measurement are of
about ten milligrams. It is therefore possible that during the sampling carried out for the
"starch 20%wt_1" measurement, the mass fraction of the sample was greater than the original
4.1. PROPERTIES OF THE COATING COLOR 81
(a) 40.0 ±0.1 °C (b) 54.5 ±0.1°C
(c) 64.0 ±0.1° (d) 75.0 ±0.1°C
(e) 86.0 ±0.1°C
Figure 4.1: Evolution of starch suspension at different temperature during cooking (Optical
microscopy).
Figure 4.2: Measured temperature by the IR sensor (°C) versus the heating temperature (°C)
and measuring height hs (mm) (standard deviation: ±1 °C). 3 trials were carried out.
mass fraction. However, this was beneficial. Indeed, the curves of the "starch 20%wt_2" and
"starch 20%wt_3" measurements seem to be smoothed: no clear peaks appear because the water
smoothed the measurement. Nevertheless, it is still possible to observe a slight variation of the
slope for a temperature close to 72°C.
The "starch 20%wt_1" measurement shows two very distinct peaks. The first is at 67.61°C.
The second is at 82.62°C. It is quite common to find a gelatinization temperature between 55 and
80°C in the literature [41]. Our values are therefore consistent on this point. It is still necessary
to understand the appearance of two peaks and not just one. In the work of Singh et al 2008 [82],
a second smaller peak is also observed for the sample named K.Sindhuri. The reasons for the
presence of two peaks could be related to the difference in gelatinization temperature between
amylose and amylopectin or to the presence of crystalline zones with long chains, thus requiring
more energy to be destroyed [82]. For the samples "starch 20%wt_2" and "starch 20%wt_3",
the peak values, although smoothed by water, are 73.92°C and 71.22°C respectively.
If we now compare the results of the optical and DSC measurements, some points can be
highlighted. First of all, a first phenomenon seems to appear at around 75°C as observed by
microscopy. The DSC measurements allow us to locate this first change more precisely at 70.91
±3.16 °C. A second phenomenon appears around 82°C.
Thus, in order to ensure optimal cooking of the starch gel, the said cooking will be carried
out at a temperature of 95°C, for 30 minutes in order to allow the temperature of the robot-
cooker to rise sufficiently. In addition, the cooking will be done under agitation. However, the
latter cannot be measured. Indeed, the food processor used is normally intended for domestic
use, so the control of the speed of rotation of the stirring blade has been simplified. In order to
obtain more information about the cooking of the starch gel, the investment in a RVA (Rapid
Visco Analyser) device would have been a plus. This would also have allowed a comparison with
many results in the literature.
4.1.2 Rheology of the starch gel

Rheology has a significant role in this project. Indeed, this science allows to have access to
different information concerning a fluid subjected to different constraints. The knowledge of the
viscosity of the starch gel is therefore necessary for 3 reasons:
→ To optimise the coating process by knowing the rheological behaviour of the fluid at
different shear rates and temperatures
→ To study the sedimentation of starch particles in the starch gel in the coating color storage
tank
→ To evaluate the penetration of water in the starch gel in order to quantify the evolution
of the mechanical properties of the paper during coating.
This rheological study will be divided into two parts. The first one will focus on the evolution
of the viscosity of the starch gel as a function of the shear rate. The second part will be dedicated
to the impact that temperature can have on the viscosity of starch gel.
4.1.2.1 Shear rate dependency of the starch gel viscosity
The objective of this section is to study the rheological behaviour of the starch gel of this project.
Through the literature review, it can be said that starch gels are shear thinning fluids, i.e. their
viscosity decreases as the shear rate increases. In addition, the starch mass fraction of the gels
also affects the rheological behaviour.
Figure 4.3 shows the evolution of the viscosity of starch gels at 15%wt (orange) and 20%wt
(black), at different temperatures: 20°C (triangle), 35°C (cross) and 50°C (square). It was
not possible to go beyond 50°C because the water in the starch gels evaporated despite the
precautions taken (moisture saturation of the measurement environment). The gels dried out
and became brittle starch films. Note that each mean value was calculated with at least 100
points. Standard deviations were determined but are not shown in Figure 4.3 because they are
small. The highest coefficient of variation is equal to 5.96% and is obtained for γ̇ = 1 s−1 and
η = 42.17 P a · s for a starch mass fraction of 15% at 35°C.
First of all, we note the unquestionable decrease in viscosity as a function of the shear rate,
whatever the temperature or the mass fraction. The impact of the mass fraction on the viscosity
should also be noted. Indeed, at the same temperature and shear rate, the 20%wt gel is more
viscous than the 15% gel. This is consistent with the information presented in the literature
review (see Figure 2.24). A higher stress is required to disentangle a more concentrated polymer
network.
Tests were carried out from 0.1 to 30,000 s−1 but when γ̇ > 10, 000s−1 starch gels are ejected
from the measuring plate due to the centrifugal force generated. At very low measurement rate
Figure 4.3: Evolution of starch gel viscosity versus shear rate, for different starch mass
fraction, at different temperature.
Code Temperature [°C] Φbinder [wt%] nodw [-] kodw (P a.sn ) R2

20°C-15% 20.00 15% 0.10 42 0.991
35°C-15% 35.00 15% 0.13 32 0.998
50°C-15% 50.00 15% 0.27 5 0.980
20°C-20% 20.00 20% 0.04 187 0.992
35°C-20% 35.00 20% 0.09 104 0.950
50°C-20% 50.00 20% 0.23 18 0.996
Table 4.1: Parameters obtained for the OdW model.Φbinder is the cooked starch mass fraction.
the rheometer is extremely sensitive to external disturbances (footsteps, closing doors) and these
appear on the rheograms. These events are not negligible and have a negative impact on the
stabilisation of the system. Moreover, with heating, the starch gel dries out before stabilisation
(Figure 4.4). For these reasons, no stabilisation plateau could be detected although stabilisation
seems to occur at high shear rates. Therefore the Cross and Yasuda models will not be used.
The Ostwald-de-Waele model will be kept.
In order to see whether the experimental results can be modelled by the Ostwald-de-Waele
model, the least squares method (LSM) is used. The problem variables are nodw and kodw ,
respectively the flow behaviour index and the flow consistency index. In order to determine
these two variables, the Generalized Reduced Gradient (GRG) algorithm is used. The results of
the modelling are presented in Table 4.1 based on 6 sets of experiments (3 temperatures and 2
mass fractions). The coefficients of determination R2 (a measure of the quality of the prediction
of a linear regression) are very good: all greater than or equal to 0.950, and for the majority
greater than 0.990. It is therefore possible to say that the experimental results can be modelled
by the Ostwald-de-Waele (OdW) model. For each of the 6 sets, the flow behaviour index nodw is
strictly between 0 and 1. This means that the shear thinning behaviour has been well modelled.
In addition, nodw increases when the temperature increases. Conversely, the kodw coefficient
Figure 4.4: Evolution of viscosity of a starch gel at 15%wt,at 50 °C and γ̇ = 0.1 s−1 . The
starch dried and now stabilization of the viscosity appears.
Shear rate Ea (J · mol−1 ) Ae (P a · s) R2

10 8968 39.16 0.78
100 8325 5.42 0.99
1000 8674 1.53 0.95
10000 16102 0.08 0.94
Table 4.2: Parameters obtained for the Eyring model.
decreases with increasing temperature.

It also appears that temperature plays a significant role on the viscosity of starch gels.
Indeed, at identical concentration and shear rate, the viscosity of the gels decreases when the
temperature of the latter increases. However, it will be noted that the experimental points for
the temperatures of 20 and 35°C are merged several times (always at fixed concentration). The
impact of temperature will be discussed in the next section.
4.1.2.2 Temperature dependency of viscosity
As mentioned earlier, the temperature of the starch gels tested seems to impact the viscosity of
the gels. This point was presented in the literature review and is not new. This kind of behaviour
is regularly modelled by Eyring type law and it will be applied to the present case. Figure 4.5
shows the evolution of the viscosity of a 20%wt starch gel as a function of temperature, at
different shear rates.
In the figure, it can be seen that for a given shear rate, the viscosity decreases with an
increase of the temperature. The Eyring law will be used to model the experimental behaviour
presented Figure 4.5. The LSM method combined with the GRG algorithm was also used as
before. The results are presented in Table 4.2.
Few remarks could be done. First of all, the R2 obtained for shear rate equal to 10 s− 1 is
0.78. This is low compared to the other values (0.94, 0.95 and 0.99). The reason for this low
Figure 4.5: Evolution of starch gel viscosity versus temperature at different shear rate for
20%wt starch gel.
value may be the long stabilisation time at γ̇ = 10 s−1 , allowing the gel to dry and thus leading
to a wider spread of viscosity values. The second point to address concerns is the values of the
activation energies, which are low except at γ̇ = 10000 s−1 . In fact, the activation energy of
water, which is a simple molecule easy to move, is 15.6 kJ · mol−1 . For γ̇ = 10000 s−1 , the
hypothesis made is that at high shear rates the starch and water in the gel separate and the
behaviour studied is similar to water. Dolz et al (2004)[136] showed the dependence of Ae and
Ea on shear rate for biopolymer emulsions. In the present case we do not have emulsions but
we try this model anyway (equation 4.1).
ae
Ae = Ae0 + (4.1)
(be + γ̇)2
where Ae0 , ae and be are constants. The obtained values are : Ae0 = 8.40 · 10−4 ; ae = 87.62;
be = 29.93 with R2 = 1. The maximum value of Ae could be related to η0 (zero shear viscosity).
The decay follows the shear-thinning behaviour of the fluid [137], the model is therefore correct.
To conclude this section, let us consider the use of these models and experimental results for
the process. It has been shown that starch gels are shear thinning fluids. Their behaviour is also
impacted by the gel mass fraction of starch and the temperature. In the context of a coating
process, the best compromise between all these parameters must be found. A fluid that is too
viscous will be difficult to apply to the paper. On the other hand, if its viscosity is too low, it
will run, splash and leave traces. Thus, the gel must not be cold. A coating at 20°C should
therefore be avoided. Moreover, if the fluid is too hot, the water contained in it will evaporate
and modify the mass fraction of the gel. It is therefore necessary to try to coat between 35°C
and 40°C. Rheological modifiers could have been used but it was not is the project philosophy
(using the less components and as much as possible bio-sourced). The coating speed will impact
the shear rate that the gel will undergo during coating, and consequently the viscosity. In this
case it is difficult to know the exact shear rate. However, the speeds are known at all times
(provided there is no slip between the sheet and the roller). The speed of the web is between 5
and 40 m · min−1 . The speed of the applicator roller is considered to be zero (this is not the
case in reality but the applicator roller has no speed control). The distance between the sheet
and the applicator roller during the application is difficult to measure. It is assumed that this
distance is 10−4 m (close to the air gaps found in industry: speed in industry 1800 m · min−1,
shear rate: 105 to 106 s−1 1 by fast calculation air gap between 30 and 300 µm. We therefore
choose 100 µm. Thus, knowing the air gap and the sheet speed, the shear rate can be estimated
at 1666 s−1 at 10 m · min−1 and 6666 s−1 at 40 m · min−1 . As first approximation, the shear
rate will be set at 3000 s−1 . At this shear rate, the viscosities of the gels are of the order of 0.03
P a · s for a 15%wt gel and 0.07 P a · s for a 20%wt gel. To favour the binding power of the gel
and limit the quantity of water in it, the 20%wt gel is chosen for the rest of the project (except
in specific cases).
The original project of this thesis was to produce a starch gel in which starch particles are
suspended to form the coating color. However, no study was carried out on starch gel with
uncooked starch particles in it. Indeed, the suspension of uncooked starch particles in the starch
gel is not easy to handle and aggregates may appear, causing a problem during the viscosity
measurement.
4.1.3 Sedimentation of starch
During the coating process, coating color may stay several minutes on the tank. If the agitation
is not enough efficient starch particles can sediment. It is therefore necessary to examine the
extent to which these particles sediment in the water first and then in the starch gel, all in order
to ensure homogeneity during the coating process. For this we will try to determine the size of
the starch particles experimentally. With these data, the sedimentation rate can be calculated
with models and compared with experimental results. Depending on the results, a decision will
be made to implement a system to prevent sedimentation.
To carry out this study, different information are needed. As presented in the literature
review, the Stokes equation refers to the size of the particles but also to the viscosity of the
suspending medium. Thus starch particles sizes and starch gel viscosity should be known. For
this purpose, a particle size analysis was carried out. SEM photos of the starch particles will
illustrate this (Figure 4.7). The other important point is the knowledge of the viscosity of the
starch gel. Let us first look at the particle size analysis in order to know the sizes of the particles
used. Figures 4.6b and 4.6a show the particle population, by number and volume respectively.
In Figure 4.6b and Figure 4.6a, we have taken care to isolate the small and large particles. For
this purpose, we have left the particles sedimented (this already provides an answer concerning
sedimentation).
Concerning Figure 4.6b showing the particle population (in numbers), there are some points
to discuss. First of all, the heavy particles curve shows 3 peaks: a first one very flattened
for a particle diameter close to 15.0 µm, a second one at 0.8 µm and a third one at 0.2 µm.
Furthermore, the shape of the curve between 1.6 µm and 3.0 µm could indicate the existence
(a) Volume (b) Number
Figure 4.6: Population of starch particles, expressed in volume or number, as a function of

their diameter. The cross symbols are the experimental values. The dotted lines are used to
facilitate reading.
of a fourth peak. The curve for the "Lightweight particles" also shows several peaks: the first is
easily identifiable at 2.6 µm, the second at 1.2 µm and the third at 0.2 µm. There also appears
to be a fourth peak at 6.0 µm, also flattened.
Figure 4.6a showing the population by volume, 3 peaks are identifiable on the curve of the
"Heavy particles": at 15.0 µm, at 4.0 µm and a smaller, almost flattened one at 1.2 µm. The
lightweight particle curve shows two peaks: the first near 4.0 µm and the second at 1.2 µm.
Thus, it is already possible to identify three different particle sizes. It is not unimportant to find
2 peaks at identical diameters on the curves of "Heavy particles" and "Lightweight particles".
This can be explained by the possibility of having small particles that have already sedimented
when the heavy particles were collected.
Thus, we see the appearance of many small particles, which had been obscured by the large
particles, and which were not identifiable in the volume population figure. Moreover, we note
that in figure 4.6b with the population by number, the peaks, whatever the curve, are found
at globally similar values: 0.2 µm, between 0.8 and 1.0 µm and a possible peak at 2.6 µm. In
order to identify all the families that may exist a mixed Gaussian model was used with Python.
This allowed to identify the particle size families presented in Table 4.3.
Here the standard deviation gives an idea of the precision of the parameters that are the
amplitude, the mean value and the standard deviation of the Gaussian model. It should be
noted that the standard deviation of the parameters can be quite large. However, what we are
particularly interested in here is the mean value obtained for each Gaussian. If we look at them
carefully, we can see that 4 families of particle size appear. The first is centred on 0.5 µm, the
second on 1.1 µm, the third on 4.2 µm and the last is centred on 13.7 µm. These values will be
used later to study the sedimentation of particles. The SEM picture Figure 4.7 illustrates the
particle size variation
Let’s now look at the sedimentation of particles. Figure 4.8a shows the evolution of the height
of the two experimentally observed sedimentation fronts as a function of time. The particle
"Family" Amplitude Average value (µm) Standard deviation (µm)

Heavy number 1 7.5 ±1.5 0.5 ±0.1 0.3 ±0.1
Heavy number 2 3.3 ±1.7 1.4 ±0.5 0.8 ±0.4
Heavy number 3 1.9 ±4.4 12.3 ±10.7 4.0 ±11.5
Heavy volume 1 0.4 ±0.4 1.3 ±0.4 0.5±0.4
Heavy volume 2 1.1 ±0.6 3.3 ±0.5 0.7 ±0.2
Heavy volume 3 96.6 ±1.8 15.2 ±0.1 3.9 ±0.1
Light number 1 2.7 ±1.3 0.5±0.1 0.2 ±0.1
Light number 2 6.8 ±1.4 1.1 ±0.1 0.2±0.1
Light number 3 2.9 ±1.9 4.4 ±1.2 1.5 ±1.1
Light volume 1 0.6±0.1 1.0 ±0.1 0.2 ±0.1
Light volume 2 64.6 ±5.5 4.9 ±1.1 1.4 ±10.3
Table 4.3: Particle familly. "Heavy" or "light" are linked to the phase study, "number" or
"volume" are linked to the analyse of population.
Figure 4.7: Starch particles of different sizes. Photo obtain by SEM.
density is 1500kg · m−3 . We can see that one front, called ’Heavy particles’, sedimented much
faster until 19000 s, then tended more slowly towards a plateau. The front called "Ligthweight
particles" settles much more slowly. Figure 4.8b shows the sedimentation rates as a function of
time.
It can be seen that for the "Heavy particles", the sedimentation speed is divided by 10
in 1h45 min (6330 s). Indeed, these "heavy" particles become compacted at the bottom of the
beaker and its height eventually tends towards a plateau. Consequently, the sedimentation speed
associated with this front tends towards 0 mm · s−1 . It is possible to compare the experimental
sedimentation values with the different sedimentation models proposed in the literature review
(equation 2.2 and equation 2.4). It is assumed that Brownian motion is negligible in this case.
By fixing the volume fraction at 14.28% (calculated from the mass fraction), the Richardson-
Zaki (R-Z) law [46] is applied to the 4 particle size diameters (0.5 µm, 1.1 µm, 4.2 µm and
13.7 µm) to determine sedimentation velocities (starch grains density is 1.5). The value of S
is fixed at 4.65 using LSM methode with GRG algorithm. Table 4.4 presents the results of the
(a) Height of the sedimentation front (b) Sedimentation rate
Figure 4.8: Evolution of the height of the sedimentation front and of the sedimentation rate
over time.
Diameter (µm) Stokes speed (mm.s−1 ) VRichardson−Zaki (mm.s−1 )

0.5 6.85 · 10−5 3.35 · 10−5
1.1 3.27 · 10−4 1.60 · 10−4
4.2 5.06 · 10−3 2.47 · 10−3
13.7 5.39 · 10−2 2.64 · 10−2
Table 4.4: Particle speed results from the 2 different models.
modelling obtained and also includes the velocity of the particles according to Stokes’ law.
According to the literature [47], it is expected that the velocity according to the R-Z model
is higher than that of the Stokes model. Indeed, the interactions between particles slow down
the falling motion. This point is verified in the models. Now it is possible to compare the model
and experimental values. The experimental maximum sedimentation velocities are between
10−3 mm · s−1 and 4.3 · 10−2 mm · s−1 . Only the particles with diameters of 4.2 µm and 13.7µm
are exactly in the experimental range. As for the particles with diameters of 0.5 µm and 1.1 µm,
their velocities modelled by the Richardson-Zaki law are outside the experimentally obtained
velocity range. This could be explained by a possible Brownian behaviour (which contradicts
our first hypothesis) of the smallest particles, which could remain suspended in the water, above
the sedimented layer. Consequently, it is possible to associate the 4.2 µm particles with the
sedimentation front of the "ligthweight particles" and the 13.7 µm particles with the front of the
"Heavy particles".
The final point of this section to be discussed is the practical use of this information. The
storage tank for the coating color is a trapezoidal shape as shown in Figure 4.9. The maximum
depth is therefore 35 mm. As the storage container is never full for ease of handling, the
maximum depth will be taken to be 30 mm. Considering the maximum sedimentation speed
measured experimentally equal to 4.3 · 10−2 mm · s−1 , it takes the associated sedimentation front
("heavy particles") 11 min 37s to sediment in the water and cover the 30 mm of the storage
tank. This is a calculation made for a quiet environment with no agitation or coating process
4.2. STUDY OF THE WETTABILITY AND WATER ABSORPTION BY PAPERS 91
Figure 4.9: Dimension of the coating tank used on the coating pilot.
taking place. This means that :
→ Sedimentation on the slopes of the tank occurs rapidly and will therefore not be negligible.
In addition, it has been found experimentally that once sedimented, the starch particles
form a thick paste, which is difficult to agitate.
→ The bubbling roller of the coating process is located in the centre of the tank. It creates
little movement in the coating color because it only "licks" the surface of the fluid. It
therefore does not allow sufficient agitation to avoid sedimentation of the starch particles
in the water.
However, in this case, the suspending medium is not water but starch. Therefore the viscosity
of the fluid is different. Furthermore, starch gel is a shear thinning fluid unlike water which is a
Newtonian fluid. So, it is necessary to take into account the characteristic shear gradient which
is 0.3 s−1 . The temperature of the starch gel is fixed at 35°C. The value of nodw is fixed at 0.09
and ko dw is fixed at 104 P a · snodw at 35°C for a mass fraction of 20%. The viscosity obtained
is therefore 836.51 P a · s. Thus the sedimentation speed obtained with equation 2.5 is equal to
6.05 · 10−9 mm · s−1 . This speed is 10,000 times lower than the sedimentation speeds in water.
This also means that a particle settling in starch gel would take several days to settle. This
does not take into account possible interactions between the suspending fluid and the settling
particles.
These results, although derived from modelling, suggest that no sedimentation problems
are to be expected when using a starch gel. However, the experiments carried out proved the
contrary. Indeed, sedimentation, although slower and less important than in water, appeared in
the starch gel. This can be due to three things: the dispersion of the starch particles was not
done correctly in the starch gel (which was also observed). Agglomerates formed and settled
faster than expected. The other possibility is the presence of physico-chemical interactions
neglected here. Finally, a third possibility is that the particles could have settled on the sloping
walls and then slid to the bottom of the coating color storage tank.
4.2 Study of the wettability and water absorption by papers

In the context of liquid absorption in porous media, the Lucas-Whasburn model is regularly
quoted and used as a basis for work. Equation 4.2 expresses the liquid intake (height or mass)
of a porous material as a function of time
0.5
γlg · rcap · cos(θc )

h= · t0.5 (4.2)
2η
. As a reminder, it integrates different characteristics of the material and the fluid:
→ The surface tension of the liquid γlg
→ The radius of the pore(s) of the material rcap
→ The contact angle between the fluid and the material θc
→ The viscosity of the fluid η
If the fluid studied is the starch gel studied previously, then only its viscosity as a function of
temperature is known. Finally, no information about the paper used is known. In this section,
the missing characteristics of the paper and the starch gel will be determined in order to see to
what extent the Lucas-Washburn model applies in this case. The modelling will be set alongside
experimental results, which will be used to confirm or refute the relevance of the models used.
The final objective is to obtain a relationship between the mass of liquid (water or starch gel)
absorbed by the paper as a function of time.
4.2.1 Study of the papers wettability

This sub-section is concerned with the wettability of the papers used and the surface phenomena
taking place between the paper and the water or starch gel. This will allow us to know the contact
angle θc existing between the paper and the water or the starch gel. Furthermore, by using other
liquids the SFE of the paper will be determined. The surface tension of the starch gel will also
be determined.
4.2.1.1 Contact angle betwen paper and liquids
The static contact angle between the paper used and different liquids was measured by the
sessile drop method. The aim of these measurements is to be able to trace the SFE of the
paper and also to know the behaviour with water. As a reminder, the different liquids used are
: Hexadecane (Hxdc), 1-Bromonaphthalene (1-Br-Napht), Glycerol (Gl-ol), Ethanol, Ethylene
Glycol (Et-G) and lab-made distilled water.
The results of the contact angle measurements are shown in Figure 4.10 for the coated side
of the GHDS 35 paper.
Between 10.00 and 60.00 s, the contact angles are all lower than 90°, whatever the test
liquid used. The material is therefore hydrophilic. A decrease in the contact angle over time
is observed which is more or less important. This can be explained by a spreading of the drop
or its absorption. The signals are also noisy due to the safety ventilation and lead to high
standard deviations values (represented by a light area around the main curve) especially for
water, glycerol and ethylene glycol. This can be explained by the different surface roughness of
the paper from one point to another and the chemical heterogeneity that the paper surface can
offer.
Figure 4.10: Evolution of the contact angle between different liquids and GHDS 35 paper
coated side.
The results for the uncoated side are shown in Figure 4.11.
The hexadecane curve is missing because on the uncoated side the product was wetting the
paper directly and spreading. Consequently, it was not possible to obtain a measurement of the
contact angle. On the other hand, it was found that the contact angle values for water are above
90°, meaning that the uncoated side of the paper is hydrophobic. This is relatively surprising.
However, the supplier told us that the paper was treated to resist the wet state. This could
explain the high contact angle value, as the paper fibres have been made hydrophobic (impact
of SBR latex or AKD treatment).
With this techniques and the studied paper, no information are available before 3 s (at
best), due to the measurement protocol. However, this is an area where it is necessary to have
results. In order to have access to this information, the behaviour at very short time (< 3 s)
was modelled according to models proposed by Stehr et al [138](cf. equation 4.3) and f Sarma
and Vest [139](cf. equation 4.4).
θc,i · θc,eq
θc = (4.3)
θc,eq

Kc ·t·
θc,i + (θc,eq − θc,i ) · exp θc,eq − θc,i
The physical meaning of Kc represents how fast the liquid spreads and penetrates into the porous
structure of wood. In the present case the unknowns are θc,i and Kc .

ks · γlg 
cos(θc,i )
!
− ·t
cos(θc ) = 1 − 1 −

· exp η  · cos(θc,eq )

(4.4)
cos(θc,eq )
Figure 4.11: Evolution of the contact angle between different liquids and GHDS 35 paper
uncoated side.
λc (µm) 10 50 80 250 400 500 800

Roughness (%) 0.66 4.2 4.39 4.45 4.44 4.44 4.43
Waviness (%) 0.21 0.07 0.02 0 0 0 0
Table 4.5: Sdr (Developed interfacial area ratio) obtained for different cut-off valuesλc , in the
case of roughness or waviness. The GHDS 35 coated face is studied here.
The model of Sarma has the advantage of taking into account two parameters of the test
liquid: its viscosity and its surface tension. However it is not perfect to predict behaviour at
short time contact [27].
A contact angle correction taking into account the roughness of the paper was also considered
with the Wenzel equation presented in the literature review. The Sdr’s were thus determined.
Table 4.5 presents the results. The Sdr measurements were performed for different cut-off filter
values λc , and at the scales of roughness and waviness. The scale to be taken into account is
the roughness. To choose the right cut-off value, the RSM has been measured and is equal to
39.81 µm. The Sdr is thus equal to 4.45 % and the rw =1.0445. The same method was applied
for the uncoated face, and in this case Sdr = 2.44 %. rw =1.0244 [66].
According to the Wenzel model, the higher the value of Sdr, the more complex the surface
under consideration is. In this case, one would expect the coated surface to have less roughness
than the uncoated surface. However, this is not the case. In order to understand this result, a
cross-sectional view of the GHDS 35 paper was taken with a SEM (see Figure 4.12). It allows
us to see that the coatedlayer has a certain undulation and that the pigments seem to create
asperities (at very small scales). This could explain the difference between Sdr value. However,
in needs more investigations.
Tables 4.6 and 4.7 present respectively the results of the modelling for the Stehr and Sarma
models, with and without the Wenzel correction. The models were performed for the coated
(a) x2000 (b) x8000 (c) x30000
Figure 4.12: View of the GHDS 35 paper section at different magnifications (SEM).
and uncoated sides of the GHDS 35 paper.

The results from the two models model the experimental data more or less well. First of
all, in the case of hexadecane coated side, the Wenzel correction leads to consider a complete
wetting of the paper because cos(θc∗ ) > 1. Moreover, for the uncoated side, the wetting of the
paper was experimentally visible and no data could be recovered. Thus, only the behaviour for
the coated side, without Wenzel correction, could be modelled. Considering the R2 and the sum
of the squared deviations Σr2 , the Stehr model best models the behaviour of hexadecane over
the experimental range.
In the case of water, the behaviour on the uncoated side could not be modelled. Regarding the
coated side, whatever the models (Stehr or Sarma) and the use or not of the Wenzel correction,
the R2 = 0.98. The discrimination criterion that can be used is the sum of the squared deviations
linked to the use of the LSM method. Thus the lowest sum of squared deviations is obtained
for the Sarma method, without Wenzel correction.
The behaviour of 1-bromonaphthalene is poorly modelled with or without contact angle
correction and whatever the model for the coated side. However, for the uncoated side, the
prediction quality of the models is much better. The Stehr model, without Wenzel correction,
will be considered.
Finally glycerol and ethylene glycol are correctly modelled. For the coated face, the values
obtained with the Stehr model without Wenzel correction will be used. For the uncoated side,
the quality of the prediction is lower. Thus, for glycerol, the Sarma model without Wenzel
correction will be kept for the uncoated side. For ethylene glycol, the Stehr model without
Wenzel correction will be kept.
From this study of the contact angle between the two sides of the GHDS 35 paper and
different liquids, two points are to be retained:
→ The models and predictions of the contact angles are not always good (especially for the
uncoated side of the paper); the SFE of each side will be calculated from the values at 6.35
s. A value at 0 s will also be calculated from the predicted initial contact angle values,
bearing in mind that the quality of the model predictions is not always perfect.
→ The contact angle between the water and the coated side will be taken as 74.52 ±5.8 ° at
Coated side Sdr Parameters Water Hxdc Gl-ol Et-G 1-Br-Napht

Yes No Kc 1.01·10−2 5.17·10−3 1.69·10−2 2.34·10−2 1.90·10−3
Yes No θc,i 78.14 11.70 88.72 57.56 26.11
Yes No Σr2 88.71 3.92 139.49 399.34 19.45
Yes No R2 0.98 0.87 0.99 0.94 0.72
Yes Yes Kc 1.08·10−2 n.a 1.79·10−2 2.66·10−2 3.31·10−3
Yes Yes θc,i 77.65 n.a 88.75 56.11 20.30
Yes Yes Σr2 98.42 n.a 150.92 461.84 34.32
Yes Yes R2 0.98 n.a 0.99 0.94 0.72
No No Kc 9.37·10−5 n.a 4.72·10−3 6.60·10−3 1.92·10−2
No No θc,i 97.22 n 77.42 70.05 34.61
No No Σr2 400.86 n.a 2500.03 301.13 160.33
No No R2 0.03 n.a 0.71 0.79 0.96
No Yes Kc 9.59·10−5 n.a 4.72·10−3 6.85·10−3 2.29·10−2
No Yes θc,i 97.40 n.a 77.42 69.54 32.59
No Yes Σr2 420.98 n.a 2678.37 318.15 185.05
No Yes R2 0.03 n.a 0.70 0.79 0.96
Table 4.6: Parameters obtained from the modelling using the Stehr model. The first column
informs about the GHDS 35 paper side tested (coated or uncoated). The second column
indicates if the Wenzel correction is used or not. θc,i is expressed in ° and Kc in s−1 .
Coated side Sdr Parameters Water Hxdc Gl-ol Et-G 1-Br-Napht

Yes No ks 8.99 ·10−4 5.21·10−3 1.57 2.68·10−2 5.50·10−3
Yes No θc,i 77.59 11.68 87.38 56.15 26.10
Yes No Σr2 85.16 4.06 174.70 466.98 19.67
Yes No R2 0.98 0.86 0.98 0.93 0.72
Yes Yes ks 8.99·10−4 n.a 1.57 2.67·10−2 5.44·10−3
Yes Yes θc,i 77.03 n.a 87.25 54.36 20.27
Yes Yes Σr2 94.15 n.a 191.47 552.93 35.00
Yes Yes R2 0.98 n 0.98 0.93 0.72
No No ks 6.18·10−5 n.a 5.16·10−1 2.82·10−2 8.68·10−3
No No θc,i 97.22 n.a 77.33 69.89 36.40
No No Σr2 400.86 n.a 2471.50 309.36 199.27
No No R2 0.03 n 0.71 0.78 0.94
No Yes ks 6.18·10−5 n 5.62·10−1 2.82·10−2 8.59·10−3
No Yes θc,i 97.40 n.a 77.14 69.37 34.36
No Yes Σr2 420.98 n.a 2647.63 327.22 244.28
No Yes R2 0.03 n.a 0.71 0.78 0.94
Table 4.7: Parameters obtained from the modelling using the Sarma model.The first column
informs about the GHDS 35 paper side tested (coated or uncoated). The second column
indicates if the Wenzel correction is used or not. θc,i is expressed n ° and ks in m−1
6.35 s and 77.59° at 0 s. For the uncoated face, the contact angle will be taken to be 98.22
±3.71 at 6.35 s. 6.35s is choose because there is at least one contact angle value for all
faces and liquids at that given time.
4.2.1.2 Surface free energy of paper
In order to calculate the SFEs for each side of the paper, the contact angles had to be known.
This is now the case. Thus, for each side of the paper, two SFE will be calculated. The first
will be calculated using the contact angle at 0 s predicted by the models. The second will be
calculated at 6.35 s using the experimental values. Figure 4.13 presents the results obtained
using the OWRK method.
(a) Coated side - 0 s (b) Coated side - 10 s
(c) Uncoated side - 0 s (d) Uncoated side - 10 s
Figure 4.13: Ω versus Π for the coated and uncoated side of GHDS 35, at different time.
v
u p
γlg · [1 + cos(θc )] u γlg
where Ω = q and Π = t d . Table 4.8 sum up the results. First of all, it is
2 γlgd γlg
possible to note the very variable R2 values on Figure 4.13, some regressions being very correct,
others very poor. This is quite common for paper, which is a very homogeneous material.
Furthermore, the impact of the CD or MD directions was not taken into account. This leads to
differences in the contact angle depending on the position of the observer during the measurement
Component Coated - 0s Coated - 6.35s Uncoated - 0s Uncoated - 6.35s

Polar mJ · m−2 7.73 8.41 0.16 1.37
Dispersive mJ · m−2 25 23.04 36.6 30.25
SFE mJ · m−2 32.73 31.45 36.76 31.62
Table 4.8: Surface Free Energy (SFE) results of GHDS 35 paper for the 2 faces, at two
different time.
Figure 4.14: FTIR spectrum of the GHDS 35 paper sides and bleached cellulose (for reference).
[37].The SFEs of the coated and uncoated sides are between 31.45 (6.35 s) and 36.76 mJ · m−2
(0 s). The SFE of cellulose is about 60 mJ · m−2 [140].
The values of the typical latex, calcium carbonate and kaolin coatings are between 30 and
35 mJ · m−2 [141] The values measured on the coated side are therefore those for coating. For
the uncoated side, values closer to those of cellulose were expected. Two hypotheses can be
formulated to explain these SFE values: a migration of the coating color from one side to the
other but it could be seen with SEM: it is not the case. The other possibility is that fibers
were treated, making them more hydrophobic, by increasing the dispersive component on their
surface. To see if this second hypothesis could be true, FTIR measurement were done on the
two sides of the GHDS 35 paper (Figure 4.14).
The orange and grey cruves are those of the GHDS 35 paper. It can be seen that they overlap
almost perfectly (small variations in intensity). This could suggest that, despite the coating
applied by the manufacturer, the two sides have more similarities than we think. Although
further investigations are required, this superposition of the spectra could indirectly explain the
similarity of the SFE of the two sides of the GHDS 35 paper trough a fiber treatment.
In any case, the SFE on the coated and uncoated sides are mostly dispersive and not very
polar. In the case of water, this could mean that a drop of water does not spread on the paper
faces. This was observed in particular for the uncoated side, and to a lesser extent for the coated
side.
4.2.1.3 Surface tension of starch gel
The starch gel could not be tested by the sessile drop method. Indeed, the system present in
the laboratory does not allow to heat the syringe in which the liquid to be tested is located.
At room temperature, after cooking, the starch gel has retrograded and is very viscous. It is
therefore very difficult or impossible to make a drop with this fluid without heating. So, the
results from another method are presented here as a solution to determine the contact angle
between the paper and the starch gel. The surface tension of a starch gel at 20% and 40°C was
measured and illustrated with Figure 4.15.
Figure 4.15: Evolution of the starch gel (20%wt) surface tension over time at 40°C. A, B, C
and D are different tests made in same conditions.
A great disparity can be seen in the results obtained over time and between the tests them-
selves. First of all, for test C, the increase in the surface tension value measured is explained
by the formation of a skin on the gel during the measurement. Although the gel was heated to
40°C, it was not possible to stir it (this would have disturbed the measurement). As a result,
the retrogradation of the starch took place on the surface (air colder than the gel), thus creating
a skin on the surface which disturbed the measurement. The same phenomenon is found, to a
lesser extent, in tests B and D. In fact, only the first test A does not have this phenomenon.
Indeed, from a temporal point of view, it was the first one, and the effects of the retrogradation
were still minimal. The surface tension value is not stable but over the duration of the test it
was calculated at γlg = 59 ±1.33 mN · m−1 .
In order to know the properties of the starch with the coated and uncoated sides of the
paper, the Dynamic Contact Angle (DCA) technique was used. In this case, only the first value
of advancing contact angle and receding contact angle are of interest. The measurement results
are presented Table 4.9/
There is a statistical difference between the advancing angle and receding angle for both
sides. During the first contact, the starch gel at 20% and 40°C does not wet the paper, whatever
Face ACA std RCA SD

Uncoated 106.37 0.87 48.10 3.78
Coated 94.89 5.44 36.37 2.94
Table 4.9: Advancing and receding contact angle between starch gel and GHDS 35 paper,
obtained with DCA. ACA is Advancing Contact Angle, RCA is Receding Contact Angle and
SD is the Standard Deviation.
Figure 4.16: Differential intrusion of mercury as a of the GHDS 35 paper pores diameter.
the side. When the sample is removed from the starch gel, wetting occurs.
4.2.2 Papers water absorption analysis

The contact angle between the water (or starch gel) and the paper faces is now known; so is the
surface tension of the starch gel. The missing data of the problem is the pore size of the paper.
In this sub-section, this will be measured, and subsequently, 3 experimental methods will be
presented to measure the water absorption by the paper.
4.2.2.1 Paper pore sizes
The pore size of the paper was measured by mercury porosimeter. The results of the two trials
carried out are shown in Figure 4.16.
It is possible to identify 3 main pore size families: 0.05 µm, 0.78 µm and 4.40 µm. In
addition, the porosity was measured and for "sample 1" it is 35.15% and for "sample 2" 31.58%.
This method does not allow to differentiate the porosity of the kaolin, calcium carbonate layer
and the fibrous structure. Furthermore, the pressure applied during the measurement (up to
200 MPa) can crush the paper and modify its original structure. These are points to keep in
Figure 4.17: Absorptiveness of paper over time obtained trough the Cobb method.Water is in
contact with the coated side of GHDS 35 paper.
mind for the rest of this document.
4.2.2.2 First approach: Cobb method (ISO 532-5)
The first method used to determine the amount of water that can be absorbed into a paper
is the Cobb method (according to ISO 535-5). The experimental results are shown in Figure
4.17, representing the absorptiveness of paper as a function of time and the modelled results.
It is first possible to notice a rapid increase in the absorptiveness of the paper before it tends
towards a plateau. This behaviour strongly resembles the index response of a first or second
order system in the aperiodic or critical aperiodic regime (damping factor greater than or equal
to 1). Thus, the two models tested on the experimental values shown in Figure 4.17 are a first
order model (equation 4.5) and a second order model (equation 4.6).
−t
 
As (t) = As,eq 1 − exp τ2  (4.5)

 
−t −t
  
1
As (t) = As,eq · 1 − · τ1 · exp τ1 − τ2 · exp τ2  (4.6)
  
τ1 − τ2
where As,eq is the plateau value of the aborptiveness , τ2 a time constant and τ1 is another
time constant.
In order to visualise the quality of the modelling a parity diagram is shown in Figure 4.18.
The modelling was performed with the LSM method and a solver using the GRG algorithm.
The variables of the model as well as the R2 are summarised in Table 4.10.
The first remarkable point is the similar value of the plateau K whatever the model. The
second remarkable point is the proximity of the time constant τ2 in the case of the two models.
Figure 4.18: Parity plot comparing the modelled and experimental values obtained with the
Cobb method. Water is in contact with the coated side of GHDS 35 paper.
Model As,eq (gwater · m2paper ) τ1 (s) τ2 (s) R2

First order 21.69 - 18.90 0.98
Second order 21.02 <0.01 17.58 0.97
Table 4.10: Parameters obtained after the modelling of the paper absorptiveness over time,
during Cobb test. Water is in contact with the coated side of GHDS 35 paper.
Finally the third remarkable point is the very low value of τ1 compared to τ2 . The problem
is that no Cobb value can be measured for a time less than 10 s, otherwise it will be very
inaccurate. Finally the Cobb values will be modelled by a first order model with As,eq = 21.69
(gwater · m2paper ) and τ2 = 18.90 s.
The Lucas - Washburn model is convenient for modelling absorption at short times or for
infinitely long media with an unlimited amount of liquid. However, in the case of the Cobb test,
as mentioned later, the lack of information below 10 s would make the quality of the LW model
poor. It is therefore necessary to have access to very short time information.
4.2.2.3 Second approach: DPM
The second method used is the DPM method. As a reminder, this method measures the evolution
of the intensity of an ultrasound signal, passing through a paper in water, over time. Figure 4.19
shows the results with the GHDS 35 paper, for the coated or uncoated side facing the water,
and for the flat or grooved surface. A zoom on the area of the first two seconds is proposed each
time to better visualise the behaviour at very short times.
For all curves, a rise in signal intensity is observed before the long decrease in signal intensity.
According to Waldner and Hirn [75], a signal decrease is caused by entrapped, resonating air
bubbles. However, paper where capillary penetration is not possible because they are heavily
sized (θc > 90) can still exhibit a decrease in signal transmission. Air bubbles cannot be the
(b) coated and uncoated - flat - zoom in

(a) coated and uncoated - flat 0-2s
(d) coated - flat and grooved - zoom

(c) coated - flat and grooved between 0-2 s
(f) coated and uncoated - grooved - zoom

(e) coated and uncoated - grooved between 0-2 s
(h) Uncoated - flat and grooved - zoom

(g) Uncoated - flat and grooved between 0-2 s
Figure 4.19: Signal intensity overs time curves obtained from DPM measurement.Coated and
uncoated refereed to the GHDS 35 paper side facing water. Grooved and flat refereed to the
sample holder surface.
Parameters Coated - flat Uncoated - flat Coated - grooved Uncoated - grooved

t95 (s) 1.00 0.76 1.44 1.12
slop (%/s) -4.87 -6.60 -4.44 -4.87
tmax (s) 0.16 0.16 0.04 0.28
Table 4.11: Characteristics parameters measured with DPM method. Coated and uncoated
refereed to the GHDS 35 paper side facing water. Grooved and flat refereed to the sample
holder surface.
reason for the decrease in this case, as there is no liquid entering the substrate pores. To verify
this point, it is sufficient to compare the results for a flat and grooved substrate. In the case
of the flat holder, air cannot escape. In the case of the grooved holder, air can escape thanks
to the grooves. If air was the reason for this decrease in signal strength, then the curves would
be different depending on the sample holder. However, this is not the case: in both cases, an
increase and then a decrease in signal strength is observed.
To quote Waldner and Hirn’s excellent explanation, "air bubbles cannot be the reason for a
signal decrease if capillary penetration is not possible, since there is no liquid front advancing
through the sample. However, as mentioned earlier, water can still enter the sample via vapor
and surface diffusion, as well as penetration within the paper fibers (which are porous). In that
way, water will diffuse into the fibers, filling voids within the fibers, causing them to swell. The
filling of the voids should lead to an increase in the speed of sound in the fibers. That would
lead consequently to an increased reflection coefficient for air-fiber boundaries. As the fibers are
still surrounded by air, this would explain the slowly decreasing ultrasound signal. [75].
This would confirm the hypothesis formulated earlier concerning the treatment of the fi-
bres/paper, explaining also the contact angle value between water and the uncoated side of the
paper. Furthermore, it is still possible to measure some parameters with these measurements
like the tmax , the t95 and the slope of the signal intensity decay. These results are summarised
in Table 4.11.
The standard deviations are significant and not negligible due to the heterogeneity of GHDS
35 paper. Therefore, when the tmax and t95 times are determined by reading the bound intensity
value, the latter is subject to statistical error. Consequently, the characteristic times are also
subject to statistical error. Therefore, it is difficult to tell the difference between the different
tmax . The same is true for the t95 and the slopes, although for this last parameter, the uncoated
flat samples seem to show a much faster decrease than the samples.
A statistical difference appears at a time greater than 15 s between the coated and uncoated
faces, independently of the geometry of the sample holder (flat or grooved). Indeed, when the
uncoated side is exposed to water (coated side stuck to the sample holder), we notice that, for
a given time greater than 15 s, the decrease in signal is less.
This method is very interesting for comparing the water behaviour of different papers. How-
ever, it does not give any information about the amount of water a paper can absorb per unit
of volume, surface or mass. A third method is therefore necessary to know the water absorption
at short time in a quantitative way.
Figure 4.20: GHDS 35 paper water absorption versus time. 50 stacks of paper are used, the
coated side facing the water.
√
Figure 4.21: GHDS 35 paper water absorption versus t. 50 stacks of paper are used, the
4.2.2.4 Third approach: Paper discs stack
The last method presented to measure the amount of water absorbed by a paper over time
is new. Figure 4.20 shows the average of the experimental results obtained for 50 stacks of
paper, coated face down. Further tests were carried out and are presented in Appendix B. The
other results in the appendix raise many questions, especially about the repeatability of the
measurement, which may be related to the construction of the paper column.
In this case, the standard deviation obtained is not negligible. It is also possible to notice
a semblance of linear behaviour between 0 and 2 s. Another way of representing this curve is
√
to plot the mass as a function of t according to the LW model. The evolution of the mass is
√
linear as a function of t and the slope is related to different characteristic of the fluid and the
paper (cf. Figure 4.21).
In Figure 4.21, a regression line has also been drawn. The coefficient of determination is 0.968.
However, it should be noted that the regression line greatly overestimates the experimental data
Name of the model Equation of the model

c · γlg · rcap · cos(θc ) 0.5 0.5

L-W mwa (t) = ·t
2η
−t
 
First order model mwa (t) = mwa,eq 1 − exp τa 

 
Power law model mwa (t) = λa tna
Table 4.12: The 3 different models used to modelled experiemntal data obtained with the
paper disc stacking method.
√
for t < 1.200 s0.5 . This non-linear relationship at the beginning of the absorption is also cited
in the literature [142]. Moreover, at the beginning of the absorption a linear relation between
the absorbed mass and time would be observed, as evoked by Quéré [74], for whom this would
be due to a variable contact angle [74]; this linear relation was noticed in Figure 4.20 and also
cited in the literature review. Experimental values have been modelled via three different models
presented in Table 4.12.
The modelling was carried out between 0 and 10 s. The results of the modelling are presented
in Table 4.13
L-W First order Power law
c (m5 ) = 7.40 10−12 mwa,eq = 0.029 λa = 0.0063
ρ (kg · m3 ) = 1000 τa = 4.900 na = 0.6376
γlg [0.001N · m−1 ] = 0.072 Σr2 = 0.00000438 Σr2 = 0.00007
θc = 75° R2 = 0.999 R2 = 0.988
η P a · s−1 = 0.001 - -
Σr2 = 1.64 10−4 - -
R2 = 0.993 - -
Table 4.13: The modelled parameters obtained from the 3 different models used. ρ, γlg , θc , η
are input values.
The models are generally very good if we look at the R2 . However, Figure 4.22 showing
the parity diagrams and comparing the experimental and modelled values provides additional
information. Indeed, it is possible to observe the global overestimation of the models for the low
experimental values (i.e. values at short time). However, this overestimation does not have the
same intensity and the first order model is the one that most closely matches the x = y curve.
This model gives two pieces of information: a characteristic time τ and a mass related to the
plateau K. τ corresponds to the time to reach 63% of the maximum value and 3 · τ corresponds
to the time necessary to reach 95% of the maximum value.
The same method and models were used for the uncoated side of the paper. Figure 4.23
shows the average of the experimental results obtained for 50 stacks of paper, uncoated side
down. Note the linear behaviour between 0 and 1 of the mass as a function of time.
√
Figure 4.24 is the same as Figure 4.23 but as a function of t. The linear behaviour for
t0.5 > 1s0.5 for the mass as a function of the square root of the time can be seen.
The experimental results were compared with the same 3 models as the recumbent face (LW,
paper discs stacking method for the GHDS 35 coated side down.
Figure 4.23: GHDS 35 paper water absorption versus time. 50 stacks of paper are used, the
uncoated side facing the water.
√
Figure 4.24: GHDS 35 paper water absorption versus t. 50 stacks of paper are used, the
paper discs stacking method for the uncoated face.
first order and power law) (cf Figure 4.25. The results are presented in Table 4.14.
L-W First order Power law

c (m5 ) = 7.40 10−12 mwa,eq = 0.014 λa = 0.0070
ρ (kg · m3 ) = 1000 τa = 1.717 na = 0.3526
γlv (0.001N · m−1 ) = 0.072 Σr2 = 0.00003718 Σr2 = 0.00007
θc = 75 R2 = 0.970 R2 = 0.962
η P a · s−1 = 0.001
Σr2 = 3.97 10-̂3
R2 = 0.926
Table 4.14: The modelled parameters obtained from the 3 different models used.ρ, γlg , θc , η
are input values
Overall the quality of the models is less good. The most convincing model remains the first
order model with a lower τ and K than before. It is not impossible that the absorption of water
by the uncoated face is slower in this direction: lower contact angle of the surface, creating a
brake on wettability, and the latex layer slowing down the progression in this direction.
The last point to be discussed concerns the linear part of the curve. The question is: up to
what time, the evolution of the mass can be considered linear as a function of time. For this
purpose, the experimental results for the face (coated and uncoated) were modelled as linear
function, for different durations between 0 and 10 s. For each range of duration, the R2 was
calculated. Figure 4.26a shows this for the coated side, and Figure 4.26b for the uncoated side.
However, this method, although very interesting, raises questions about the repeatability
of the measurements. A major point for improvement would be the construction of the paper
column. Short-cut phenomena and privileged passages could be at the origin of these important
standard deviations. Moreover, the very small size of the discs (15 mm) poses a problem for
handling. Finally, the cylinder-piston pair is not designed for this type of handling and is not
perfectly sealed.
4.3. CONCLUSION 109
(a) Coated side facing the water (b) Uncoated side facing the water
Figure 4.26: Evolution of the mass of water absorbed by the paper over time; and the
coefficient of determination for a linear model applied to the mass, as a function of time. The
GHDS 35 paper is used with the coated and uncoated sides facing the water.
4.3 Conclusion
In the study of the coating color it was shown that the cooking of the starch is an important
point in the process, as this impacts on the rheology of the sauce. It has been proved that the
starch gel has a shear thinning behaviour (Ostwald-de-Waele law) and that its behaviour also
depends on the temperature at which it is used (Eyring law), which is in line with the literature.
The sedimentation of starch particles has also been studied in water and quickly observed in
starch gel. The particles have a significant sedimentation speed in water when considering the
dimensions of the sauce storage tank on the pilot: it needs 12 minutes for the heaviest particles
to sediment in the storage tank. For the coating sauce, the models calculated a sedimentation
rate 10,000 times lower than in water. However, experimentally, sedimentation and aggregation
of the starch particles in the starch gel was observed. A recirculation of the coating sauce should
be installed to avoid this kind of sedimentation phenomenon.
In the a second time, the wettability and water absorption of GHDS 35 have been discussed.
The SFEs of the paper showed larger dispersive components than the polar components on both
sides. This possibly means that the fibres were treated with wet strength, thus explaining the
high contact angles between water and paper on all sides. The study of water absorption by
the paper was carried out using three methods: the COBB method, the DPM method, and
the paper discs stacking method. The Cobb method showed a first order water absorption
behaviour of the paper. The paper disc stacking method is a new method proposed to follow
water absorption in paper at short times. A first order model seems to be the one that best
models the observed behaviour. However, very large standard deviation of the experimental
data should be considered. Then comes the need to optimize the protocol and the measurement
apparatus. Finally, the DPM technique has shown a particular behaviour of the coated papers
looking at an AKD treated paper. This confirms the hypothesis put forward following the
determination of the SFE.
Chapter 5
The wrinkling phenomena of paper
Contents
5.1 Observation of the wrinkles phenomena . . . . . . . . . . . . . . . . . 112
5.1.1 In-situ observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
5.1.1.1 Tensile test and tensile buckling . . . . . . . . . . . . . . . . . 112
5.1.1.2 Pilot scale wrinkles formation . . . . . . . . . . . . . . . . . . 115
5.1.2 Ex-situ observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
5.1.2.1 Macro-scale observation . . . . . . . . . . . . . . . . . . . . . . 119
5.1.2.2 Micro-scale observation . . . . . . . . . . . . . . . . . . . . . . 123
5.1.3 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
5.2 Modelling of the wrinkling phenomenon . . . . . . . . . . . . . . . . . 126
5.2.1 Paper mechanical characteristics . . . . . . . . . . . . . . . . . . . . . . 126
5.2.1.1 Dry state . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
5.2.1.2 Wet state . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
5.2.2 Is it possible to forecast wrinkles with models? . . . . . . . . . . . . . . 140
5.2.2.1 Tensile test case . . . . . . . . . . . . . . . . . . . . . . . . . . 140
5.2.2.2 Pilot case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
5.3 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
Before being able to make the wrinkles disappear, it is necessary to be able to make them
appear and to understand the underlying mechanical phenomena. This chapter will therefore
focus on the appearance of wrinkles on a web paper. This has been described in the literature
review and two central equations have been isolated. The first one puts forward a misalignment
angle of a roll (equation 5.1). The second allows us to determine a critical strip tension above
which wrinkles can appear (equation 5.2).
6τcr a2
θcr = (5.1)
Ex b2
2t2f
s
Ex · Ez
Tcr = (5.2)
111
112 CHAPTER 5. THE WRINKLING PHENOMENA OF PAPER
Figure 5.1: 3D topographic view of speckled GHDS 35 paper obtained with the AIF set up.
Before using equation 5.1 and equation 5.2 it is necessary to observe the wrinkling phe-
nomenon and to collect experimental data. It will be performed trough a multi scale approach,
using in-situ and ex-situ experiments. The observations will be made during the pilot and tensile
tests: this is the first observation time (in-situ), where the formation of wrinkles will be observed.
The second observation phase will be carried out on wrinkles that have already formed (ex-situ)
on a macroscopic and microscopic scales. This will be the subject of the first part of this chap-
ter. The second part will allow the modelling of the wrinkle phenomenon and the comparison
of the modelling with the experimental values. In order to carry out, it is necessary to know the
paper properties. These will also be presented and discussed in the second part of this chapter,
separating the "dry" and "wet" properties of the papers.
5.1 Observation of the wrinkles phenomena

As explained, this first section will focus on the observation of wrinkles and the recovery of
experimental data to feed the different models in the literature.
5.1.1 In-situ observations
5.1.1.1 Tensile test and tensile buckling
The purpose of the tensile tests presented here is to see to what extent the wrinkling phenomenon
can be described on a microscopic scale and whether it can be reproduced outside of the coating
pilot. The measurements were carried out according following MD and CD, on GHDS 35 paper.
To follow the paper displacement over time with the D.I.C, it was necessary to speckle the paper
samples. A surface reconstruction of the speckle is shown in Figure 5.1, obtained with the AIF
(Alicona Infinite Focus). In Appendix C, an example of speckle pattern is presented.
It can be seen from the height scale that the printed ink has a height of up to 4 µm. The
GHDS 35 paper has a thickness of 33 ±1 µm. The printed ink is therefore up to 12% of the
original height of the paper. This is significant. In addition, this printing can cause mechanical
5.1. OBSERVATION OF THE WRINKLES PHENOMENA 113
Figure 5.2: Evolution of the stress versus time during a tensile test in MD. The associated
pictures correspond to the out of plane displacement of the sample at a given time (speed: 5
mm · min−1 ).
discontinuity and variation in local mechanical properties. However, this is the best technique
found for using D.I.C. on paper samples. This point should be kept in mind for the rest of this
chapter.
The results obtained during the tests are shown in Figure 5.2 for MD and Figure 5.3 for
CD. As a reminder, the dimensions of the samples are 15 mm wide by 100 mm long. The ratio
width
β= is therefore 0.15. Other results are presented in Appendix D.
length
Figure 5.2 shows the evolution of the tensile stress over time on a MD sample. The first few
seconds of the test show zero stress: the sample was relaxed. Once the specimen is under tension,
the stress evolves until the specimen breaks at 76 MPa. A linear zone can be distinguished,
comparable to the elastic zone of the material between 1.5 and 5 s. Beyond that, the behaviour
of the material can be considered as plastic. For 4 pairs of values (time; stress), it was possible
to reconstruct, with the 7D and Matlab software, the evolution of the out-of-plane displacements
w(x, z) over time. The more red an area is, the greater the out-of-plane deformation. At 5 s,
the out-of-plane deformation is very small, and may be related to measurement/reconstruction
artefacts. At 10 s, it starts to appear more clearly in the centre of the sample. At 15 s, a narrow
yellow band appears and the out-of-plane deformation is about 0.10 mm. At 20 s, just before
breakage, an orange and yellow band appears clearly on the reconstruction. The out-of-plane
deformation is close to 0.15 mm.
It is possible to describe the same phenomenon on CD specimens (cf. Figure 5.3). In order to
gain more information on the buckling phenomenon, the test presented Figure 5.3 was performed
at 2m · min−1 . The first few seconds of the test allow the sample to be stretched, which is why
no stresses appear. At 30 s of testing, a central area begins to appears. At 90 s, the central zone
Figure 5.3: Evolution of the stress versus time during a tensile test in CD. The associated
pictures correspond to the out of plane displacement of the sample at a given time (speed: 2
mm · min−1 ).
buckles more strongly and the out-of-plane strain is close to 0.10 mm. At 160 s, the central area
that was buckling showed 2 orange stripes. During the test, it was possible to distinguish not
one but two sharp edges. Finally, just before breakage, at 190 s, the two bands are separately
identifiable, with one having an out-of-plane deformation greater than 0.15 mm.
Thus, it is indeed possible to detect out-of-plane deformations, related to buckling, during
a tensile test. As buckling is an instability phenomenon, it is "all or nothing". Using Figures
5.2 and 5.3 it is possible to link out-of-plane deformation and tensile stress. The only point to
be determined is: "at what value of out-of-plane displacement is it possible to consider buckling
?" It is considered that there is buckling if the out-of-plane displacements w(x, z) respect the
inequality equation 5.3.
w(x, z) > 2 · tf (5.3)
In the present case, if the out-of-plane displacements are greater than 0.06 mm, buckling is
considered. Using the results presented in Figure 5.2 and Figure 5.3, as well as those presented
in Appendix D, it is possible to present Figure 5.4, linking out-of-plane test displacement, tensile
stress and number of strips. 3 clusters are clearly identifiable: "MD", "CD (1 strip)" and "CD (2
strips)".
Zhu et al [97] found the same phenomenon but much more pronounced on very orthotropic
polypropylene films. However, buckling does occur in the tensile tests, regardless of the direction
tested (CD or MD). In the present case, it is found that the stress to form a strip following MD
is very close to the stress at break. This suggests that in order to generate two strips, the stress
Figure 5.4: Tensile stress linked to out of plane displacement obtained during tensile tests of
speckle GHDS 35 paper.
(a) Troughs on coated dry side

observed, misalignment angle: (b) Wrinkles on coated dry (c) Wrinkles on wet coated
2.5° - Web tension: 150 side, misalignment angle: 2.5° side, misalignment angle: 2.5°
N · m−1 - Web tension: 260 N · m−1 - Web tension: 80 N · m−1
Figure 5.5: The evolution of the wrinkles formation with GHDS 35 paper, on the coating pilot
in function of the paper state, the misalignment angle and the web tension.
value at break would have to be exceeded. It can also be seen that the magnitude of the out-of-
plane displacement increases with the number of strips for the samples following CD. Moreover,
to have a second strip, an increase in tensile stress of 80% is required. It is now time to see if
this phenomenon appears in the same way on the coating pilot.
5.1.1.2 Pilot scale wrinkles formation
The appearance of wrinkles when using the coating pilot is not trivial. As pointed out in the
literature review and at the beginning of this chapter, the misalignment angle θcr and the critical
web tension Tcr are important process parameters. Then, by controlling these two parameters,
wrinkles were deliberately generated. Figure 5.5 shows the appearance of wrinkles for different
misalignment angle - web tension combinations on GHDS 35 paper.
Once the wrinkles and troughs have been generated, various points need to be highlighted.
First of all, one to many troughs (Figure 5.5a) are formed at the downstream roll with lengths
of approximately 30 to 40 mm. The lengths of the troughs vary until they become wrinkles. As
Figure 5.6: Experimental highlighting the appearance of wrinkles as a function of the pair:
web tension - angle of misalignment (GHDS 35 paper, uncoated side touching the rollers).
a matter of fact, the troughs gradually grow and reach lengths of approximately 150 mm under
certain conditions in favour of wrinkles formations. Finally, when the mechanical limit between
no-wrinkle/wrinkle is reached, the troughs have a length between 580 and 590 mm (diagonal on
the web span). The next step is the formation of wrinkle. Concerning the appearance of the
wrinkles (Figure 5.5b), they are never parallel to the machine direction of the paper but tilted.
Before the non-reversible wrinkle formation, a trough from the upstream roll to the downstream
roll is coming out. This trough starts at the upstream roll, widens to half of the web span before
tapering and converging to the downstream roll. As soon as the trough passes the downstream
roll, and is still there, it becomes a wrinkle and irreversibly marks the paper.
The last point to look at is the formation of wrinkles on wet paper (Figure 5.5c). Preliminary
tests were performed by wetting the paper just before the web span on which the study is focused.
It was found that for a web tension and misalignment angle for which no wrinkles were observed
on dry paper (despite the presence of troughs), it was not necessarily the case for wet paper.
Wrinkles starts to appear much more easily due to the presence of water.
From the observations made, as well as the measurements of the misalignment angle - web
tension pairs, Figure 5.6 and Figure 5.8 could be plotted for the GHDS 35 paper. Figure 5.6
was plotted for the uncoated side of the GHDS 35 against the roll. Figure 5.8 was plotted for
the coated side against the roll. These two figures represent the presence of wrinkles or troughs
for different misalignment angle - web tension combinations (M - T combinations).
First, region I (with the squares) corresponds to the θ - T combinations for which no wrinkles
are detected. The orange diamonds correspond to the θ - T combinations for which out-of-plane
deformations (troughs) are detected. No wrinkles or creases were detected. The same is true
for the orange triangles. Finally, the yellow circles correspond to θ - T combinations for which
wrinkles appear. It is therefore possible to isolate three zones from this graph: Region I where
the web is flat, Region II where the web has troughs and Region III where wrinkles appear on the
Figure 5.7: Web tension versus misalignment angle proposed by David Roisum [54]. 3 Regions
are highlighted: planar web, slack edge web and wrinkled web.
Figure 5.8: Experimental highlighting the appearance of wrinkles as a function of the pair:
web tension - angle of misalignment (GHDS 35 paper, coated side touching the rollers).
web. These 3 regions are also found in the literature. Region I is for a flat and unwrinkled web
which occurs when misalignment is sufficiently small. Region II is for a web with a "slack edge"
but no wrinkling which occurs when tension is sufficiently small (Figure 5.7). However, an out
of plane displacement appear. Region III is where the "shear wrinkles" crosses the downstream
roll [54].
The same observation can be made for Figure 5.8. However, no wrinkle could be detected
in this configuration.
Furthermore, the results between the coated and uncoated side of the paper are compared to
study the influence of the friction coefficient. It has been seeing that differences exist between
the areas of regions I and II of each face, especially when the angle of misalignment is greater
than 1.5°. Region III didn’t appear because no wrinkles could be formed on the uncoated face.
This could be explained by the difference in the friction coefficient of the 2 faces of the paper.
A friction test has been made on the two GHDS 35 paper sides in MD. It was found that there
(c) Oriented troughs and

(a) Troughs after the (b) Troughs after the Mayer futures wrinkles after IR
applicator roller rod drying
Figure 5.9: Troughs and wrinkles appearing on CPRAW paper during "coating" with water,
while using the coating pilot.
was more friction on the coating side. Other wrinkles are observed in tests with other papers.
Figure 5.9 shows the appearance of wrinkles on CPRAW paper. Figure 5.9a, is of the paper just
after the applicator roller. There is no coating color, only water in the coating tank. You can
see the beginnings of wrinkles appearing. In Figure 5.9b, the paper passes the Mayer rod of the
coating system and troughs are seen. Finally, Figure 5.9c is the paper after the IR dryer. The
roll visible in the picture was deliberately and roughly misaligned (the misalignment angle is
unknown). Before the roller, troughs are visible and are tilted. If one zooms on the roller, it is
possible to see wrinkles forming and becoming creases. The number of troughs is also significant:
the paper is still a little wet, which seems to be favourable to the appearance of troughs. The
roller is also hot (at least 80° C or more) as it is close to the dryers. Moreover, this paper, with
a lower grammage than GHDS 35, seems to be a perfect candidate for wrinkles to appear.
Wrinkles therefore appear more or less easily on the coating pilot depending on the web
tension and the angle of misalignment of the rollers with respect to each other. This phenomenon
is paper dependent: some papers seem to be affected more strongly than others : refining, fibre
lengths and fibre composition can be related to wrinkle appearance.
5.1.2 Ex-situ observations
The in-situ observations (i.e. during the formation of the wrinkles) provided a lot of information.
However, not everything could be characterised. It is therefore necessary to characterise the
wrinkles formed. The following observations were made on wrinkles formed on the coating pilot.
(a) Wrinkled CPRAW paper (b) Wrinkled CPBS paper (c) Wrinkled GHDS 35 paper
Figure 5.10: CPRAW, CPBS, GHDS 35 papers appearence after "coating" with water, using
the coating pilot.
5.1.2.1 Macro-scale observation
Firstly, macroscopic observations are made with the "naked eye" and with the help of the Alicona
Infinite Focus (AIF) on CPRAW, CPBS and GHDS 35 paper on Figure 5.10. Each picture of
the different papers corresponds to a representative sample of the paper after "coating" with
water (for CPRAW AND CPBS) and with starch gel (for GHDS 35). For each picture, an
overall view and a focus on a wrinkle is presented. For GHDS 35, a second focus is proposed
on a phenomenon frequently found and very visible on GHDS which is a pinch of the wrinkle.
On each paper, an asymmetric wrinkle is visible. In addition, the surfaces of the papers have
an "orange peel" effect, like small scale cockling. CPRAW has a "silk paper" or "crepe paper"
feel/touch. The wrinkle on CPRAW paper is crushed on one side and the asymmetry is therefore
more pronounced. GHDS 35 has a sticky feel/touch due to starch gel coating. As shown in Figure
5.10c, the wrinkle is oriented and very pronounced. In addition, there is the aforementioned
pinch. And this is repeated more or less regularly along the wrinkle: 23 ±2 pinches over 300 mm
(values taken from 3 wrinkles), about 0.08 pinch · mm−1 . AIF allows access to more information
about wrinkles, at an intermediate scale. Thanks to the surface reconstructions, it is possible
to see the elevation profiles of wrinkles as well as the possible periodicity of wrinkles. First of
all, an analysis has been made Figure 5.11 on the wrinkled GHDS 35.
The first analysis is carried out on a 1.39 mm depth analysis zone. Figures 5.11a and Figure
5.11b represent 5 elevation profiles of a wrinkle measured at different depths in 3D view and
in front view. The first thing that stands out is the asymmetry of the wrinkle. The wrinkle
height, whatever the chosen depth, increases drastically to reach its maximum value between
620 and 630 µm. The wrinkle height decreases slowly between 0.700 and 2.400 mm. The other
remarkable point is the strong overlap of the different elevation profiles. This suggests that
over the 1.39 mm considered, wrinkle does not change its orientation. This impression is also
supported by Figure 5.12, which is a wrinkle height map. It shows the variation in wrinkle
height as a function of wrinkle height and depth.
The rapid decrease of the wrinkle height for a width between 0.300 and 0.500 mm also poses
(a) Elevation profiles of 5 areas of a wrinkled in

3D (b) Overlay of the 5 elevation profiles
Figure 5.11: 3D and 2D reconstructions of a wrinkled GHDS 35 paper.
Figure 5.12: Wrinkle height map for a GHDS 35 paper (1.39 mm deep)
a problem with regard to the low number of data. Indeed, it is difficult for the AIF device
to reconstruct surface variations for which part of the surface is almost vertical. This leads to
a large standard deviation when calculating an average wrinkle representing the studied area.
This average wrinkle is used as a basis for modelling the wrinkle shape. In the literature,
a polynomial curve model of degree 3 is used [143]. However, the wrinkle asymmetry is not
negligible. Therefore, the Weibull model is proposed in this study to model the wrinkle height
as a function of its width. This model is normally used to model failure laws (failure time of an
electronic component or the failure time for car tyres). This law is presented in equation 5.4.
kW r −1 kW r
kW r lW r − θW r lW r − θW r

−
HW r (lW r , kW r , λW r , θW r , AW r ) = AW r · · ·e (5.4)
λW r λW r λW r
It normally incorporates 2 or 3 parameters: kW r > 0 the shape parameter; λW r > 0 the

scale parameter (directs the dispersion of the distribution) and θW r the position parameter
(determines the position of the origin). In this case, the Weibull distribution must integrate a
fourth parameter: the amplitude AW r (the Weibull distribution is a probability distribution,
Figure 5.13: Wrinkle height as a function of its width. Comparison between experimental and
modelled values obtained with the Weibull law on a wrinkled GHDS 35 paper. The
experimental values are the average of a wrinkle with a depth of 1.39 mm.
Parameters Model (3 parameters) Model (4 parameters)

kW r 2.7 1.6
λW r [µm] 905.8 638.2
θW r [µm] 0.00 300.00
AW r [µm2 ] 509723 517676
Σr2 [µm2 ] 4940828 635923
R2 0.92 0.99
Table 5.1: Parameters obtained from the 3 and 4 parameters Weibull models (Σr2 is the sum
of the squares of the deviations used is the LSM methode). The experimental values are the
average of a wrinkle with a depth of 1.39 mm. The paper is the GHDS 35.
the obtained values are generally between 0 and 1, which is not the case here). Note that for
the Weibull law lW r ≥ θW r and HW r (lW r , kW r , λW r , θW r , AW r ) = 0, ∀x < θW r . To determine
the model parameters the LSM method combined with the GRG algorithm is used. Figure 5.13
shows the shape of an average wrinkle as well as two modelled curves. The black diamonds are
the Weibull model with θW r = 0 (model with only 3 parameters) and the yellow triangles are
the Weibull model with θW r variable. The light orange band represents the standard deviation
of the experimental values.
With this model, it is possible to model the wrinkles more or less correctly. The results of
the modelling are presented in Table 5.1 .
The 3-parameter model gives R2 = 0.92 while the 4-parameter model gives R2 = 0.99. More-
over, λW r of the 4-parameters model is closer to the experimental data than the 3-parameters
model. The abrupt height decay is also better represented with the 4-parameters model. The
same analyses were performed on another wrinkle but at a greater depth (25 mm). The elevation
profiles are shown in Figure 5.14a and the level map is shown in Figure 5.14b.
The difference with the previous dataset is the lack of overlap between the different elevation
profiles of the wrinkle at different depths. Differences in width are also observed, suggesting the
pinches mentioned earlier. The abrupt variation in height is also observed but is less important.
(b) Wrinkle height map for a GHDS 35 paper (25

(a) Overlay of the 6 elevation profiles mm deep)
Figure 5.14: Elevation profiles of a 25 mm deep wrinkle (GHDS 35 paper) and its height map.
Figure 5.15: Wrinkle height as a function of its width. Comparison between experimental and
modelled values obtained with the Weibull law on a wrinkled GHDS 35 paper. The
experimental values are the average of a wrinkle with a depth of 25 mm.
The Weibull model with 3 or 4 parameters is also used. The results are shown in Figure 5.15
and the modelling parameters are summarised in Table 5.2 .
The R2 of the models are 0.96 and 0.98, which is very correct. The 4-parameters model
seems to model the wrinkle shape better for widths greater than 700 µm, notably thanks to
the value of λW r which is very close to the experimental value. However, for widths less than
700µm, the wrinkle shape is not perfectly represented by the 4-parameters model.
On the 2 data sets studied, the Weibull model allows to model the wrinkles very correctly.
It is however recommended to use the 4-parameters model. These first ex-situ observations have
already provided a lot of information on the wrinkles: asymmetric shape, widths, heights and
pinches. It is now necessary to decrease the scale of observation to complete the information
already acquired.
Parameters Model (3 parameters) Model (4 parameters)

kW r 3.44 2.23
λW r [µm] 1170.85 773.24
θW r [µm] 0.00 409.98
AW r [µm2 ] 542668.90 535711.73
Σr2 [µm2 ] 2415756.72 884593.99
R2 0.96 0.98
Table 5.2: Parameters obtained from the 3 and 4 parameters Weibull models (Σr2 is the sum
of the squares of the deviations used is the LSM methode).The experimental values are the
average of a wrinkle with a depth of 25 mm. The paper is the GHDS 35.
(a) GHDS 35 coated side (x200) (b) GHDS 35 uncoated side (x200)
Figure 5.16: The 2 sides of GHDS 35 paper observed using SEM (Low vaccum mode,
secondary electrons).
5.1.2.2 Micro-scale observation
Microscopic observations were carried out with 2 techniques: S.E.M and X-Ray tomography.
Information about the techniques is given in the chapter materials and methods. The S.E.M.
images of the both sides of GHDS 35 are shown in Figure 5.16 for reference.
It is easy to identify the two faces du to the fact that fibers are not visible on the coated
side. The wrinkled GHDS 35 is taken in a S.E.M. photo (Figure 5.17).The target area is a
pinch zone as previously shown in Figure 5.17. Thanks to the S.E.M. image, it is possible to see
the delamination of the pinched area. For further information, a wrinkled GHDS 35 chip was
subjected to X-Ray tomography. Figure 5.18 shows the results of the 3D reconstruction, with a
focus on a "damaged" area (Figure 5.18b).
Two cracks can be seen: one at the beginning of the wrinkle and the other at the top of the
wrinkle. Furthermore, on the "damaged" area, a delamination of the pigment coating layer is
seen. Images of two cross-sections of the 3D reconstruction are taken to gain more information
about this delamination. These slice planes are shown in Figure 5.19 for a non-damaged wrinkle
area and Figure 5.20 for a damage wrinkle area (with a focus on the damaged wrinkle).
Figure 5.17: Delamination phenomenon observed by SEM on a wrinkled and pinched area of
GHDS 35 paper.
(a) GHDS 35 paper sample reconstruction using (b) Focus on a damaged area of a GHDS 35
X-Ray tomography paper sample)
Figure 5.18: X-Ray tomography of a GHDS 35 paper sample.

Figure 5.19: Cross-section of a wrinkled GHDS 35 sample observed by X-ray tomography.
Figure 5.20: Damaged cross-section of a wrinkled GHDS 35 sample observed by X-ray

tomography. The observed area is delaminated.
For an undamaged wrinkle (cf. Figure 5.19) asymmetry observed previously is found. The
second section is an area of the "damaged" wrinkle (Figure 5.20). It can be seen that a non-
negligible area of the wrinkle is delaminated and that some of the fibres appear to be out of
contact on this section. The study of the causes of the delamination will not be presented in this
manuscript. The CPRAW and CPBS papers have not been studied by SEM and tomography.
However, the SEM images of these unwrinkled papers are presented in the Appendix F to allow
these two papers to be viewed at different scales.
It was possible to reuse Weibull model for the sectional plane of the undamaged wrinkle in
Figure 5.19. Figure 5.21 presents the result of the modelling with a 4-parameters model. Table
5.3 summarises the modelling parameters.
The R2 is 0.99. The modelling is very correct as soon as the width in abscissa is higher than
the value of θW r . θW r can be considered as the abscissa from which the wrinkle starts. The
kW r 2.64
λW r (µm) 740.75
θW r (µm) 549.32
AW r (µm2 ) 380396.88
Σr2 (µm2 ) 23471.59
R2 0.99
Table 5.3: Parameters obtained from the 4 parameters Weibull models (Σr2 is the sum of the
squares of the deviations used is the LSM methode).The experimental values (height as a
function of width) are obtained from Figure 5.19.
Figure 5.21: Modelisation of X-ray tomography wrinkle using 4 -parameters Weibull law.
Weibull model still allows the wrinkle shape to be modelled properly.
5.1.3 Conclusion
In this section we have seen the formation of buckling and wrinkles through different tools and
scales. The in-situ observation has allowed to acquire mechanical data which will be compared
with the models used in the following section. It has already been possible to observe a different
behaviour between the tensile tests and the pilot case. On the wrinkles formed with the pilot,
whatever the paper used, more or less oriented wrinkles are observed. The ex-situ observations,
at different scales, allowed to highlight the wrinkle asymmetry, as modelled by the Weibull
model, on 3 different data sets. The modelling had very good results. In addition, pinching
phenomena, also linked to a decohesion/delamination of the industrial pigment coated layer, are
observed for the GHDS 35 paper.
5.2 Modelling of the wrinkling phenomenon

In this section, the models presented in the literature review will be used and compared to the
experimental values. Before that, the results concerning the mechanical characterisation of the
papers in dry and wet conditions will be discussed.
5.2.1 Paper mechanical characteristics
5.2.1.1 Dry state
Table 5.4 below summarises the properties of the paper used, speckled or not. The difference
in grammage and thickness between the papers is clear. It is also important to note that the
printing of speckles on the papers has an impact on their weight and thickness: around 2 g · m−2
5.2. MODELLING OF THE WRINKLING PHENOMENON 127
Figure 5.22: Typical stress - strain curves of the 4 different papers in CD and MD.
and 2 µm with speckle. However, regarding the thickness of the speckled papers, it should be
noted that despite the different initial thickness, the two speckled papers have the same final
thickness (37 ±1 µm). As for the mechanical properties, the GHDS 35 and GHDS 40 have
equivalent Young’s MD and CD moduli. The CPRAW and CPBS have lower Young’s moduli
compared to the GHDS 35 and GHDS 40.
Name G (g · m−2 ) tf (µm) EM D (GPa) ECD (GPa)

GHDS 35 36.2 ±0.2 33 ±1 8.92 ±0.88 4.15 ±1.17
GHDS 40 40.6 ±0.2 35 ±1 9.20 ±0.17 3.93 ±0.35
GHDS 35 speckled 38.2 ±0.4 37 ±1 8.46 ±0.34 4.27 ±0.23
GHDS 40 speckled 42.4 ±0.2 37±1 9.88 ±0.33 4.08 ±0.36
CPRAW 14.1 ±0.2 24 ±1 7.21 ± 0.33 2.81 ±0.36
CPBS 23.7 ±0.4 37 ±1 4.47 ±0.74 2.06 ±0.10
Table 5.4: Mechanical characteristics of tested papers.
When it comes to the impact of the speckle, it seems that the speckle increases the MD
Young’s modulus for GHDS 40 reference and has no statistically significant impact on the
GHDS 35. Figure 5.22 represents the stress - strain curves of the 4 papers in CD and MD.
Table 5.5 gives access to the bending stiffness of GHDS 35 and GHDS 40 in both directions.
Although there is a difference between the bending stiffness following MD and CD for each
reference or speckled paper, it is not possible to make a statistically significant difference between
the papers themselves. The bending stiffness of CPRAW and CPBS papers were not measured
because these papers are too flexible. Using the Baum method, the calculated values of the
shear modulus Gxz,Baum are between 2.32 and 2.45 GPa (Table 5.6). As for the values of the
Poisson’s ratio νxz,Baum and νzx,Baum , the values are between 0.18 and 0.20 or 0.41 and 0.47.
Gxz,Baum and νxz,Baum can be compared with the results obtained with the method derived
from the constitutive equations of mechanics for orthotropic materials.
Name Bending stiffness MD (mN.m) Bending stiffness CD (mN.m)

GHDS 35 0.030 ±0.014 0.014 ±0.002
GHDS 40 0.034 ±0.013 0.016 ±0.002
GHDS 35 speckled 0.032 ±0.013 0.015 ±0.002
GHDS 40 speckled 0.035 ±0.014 0.017 ±0.002
Table 5.5: Bending stiffness in CD and MD of GHDS 35 and GHDS 40 papers with and
without speckle.
Name Gxz,Baum (MPa) νxz,Baum νzx,Baum

GHDS 35 2.35 0.19 0.42
GHDS 40 2.32 0.19 0.44
GHDS 35 speckled 2.32 0.20 0.41
GHDS 40 speckled 2.45 0.18 0.45
CPRAW 1.74 0.18 0.47
CPBS 1.17 0.20 0.43
Table 5.6: Poisson’s coefficient νxz,Baum , νzx,Baum and shear modulus Gxz,Baum of tested
papers calculated following Baum’s method.
Determination of Poisson’s ratio and shear modulus using constitutive equations of

mechanics for GHDS 35 and GHDS 40: To determine the Poisson’s ratio νxz and the
shear modulus Gxz,Baum , the tangent elastic modulus Etan is plotted as a function of the angle
Φ (Figure 5.23). Φ is the angle between the direction of the tensile test and MD. Thus, when
Φ = 0°, the samples are tested along MD and when Φ is equal to 90°, the samples are tested in
the CD. Once the curve plotted, the LSM is applied to compare experimental and model values
(cf. equation 3.5. In order to highlight the correspondence of the values, parity diagrams are
plotted Figure 5.23. In this case, y = 0.99x for GHDS 40 paper (R2 = 0.98) and y = 0.96x
for GHDS 35 paper (R2 = 0.99). Therefore, there is a good match between theoretical and
experimental values (cf. Table 5.7). Note that the values predicted by the Baum model for the
Poisson’s ratio match the experimental values perfectly. However, this is not the case for the
shear moduli where absolute relative errors of 26.79% and 16.24% are calculated.
As mechanical properties in the dry state are now known, let’s have a look at the properties
of papers when they are wet and contain a large amount of water. In this way it will be possible
to approach the conditions close to the coating process.
Name Gxz (MPa) Relative error on Gxz νxz Relative error on νxz
GHDS 35 3.21 -26.79% 0.19 0%
GHDS 40 2.77 -16.24% 0.19 0%
Table 5.7: Poisson’s coefficient νxz and shear modulus Gxz of tested papers calculated
following mechanical constitutive equation and error compare to Baum’s method.
(a) Etan versus Φ - GHDS 35 paper (b) Parity plot - GHDS 35 paper
(c) Etan versus Φ - GHDS 40 paper (d) Parity plot - GHDS 40 paper
Figure 5.23: Evolution of tangent elastic modulus Etan versus the angle bewtween MD and
test direction Φ and parity plot comparing experimental and modelled for GHDS 35 and
GHDS 40 papers.
5.2.1.2 Wet state
The mechanical properties for wet papers will be discussed through 3 aspects: the Young’s
moduli of papers with different water contents, the elongation of paper under constant stress
when one side is in contact with water (WSD) and the wet tensile strength (Finch test). This is
intended to approximate the evolution of the mechanical properties of the papers studied during
the coating process. It should be noted that in this subsection only GHDS 35 and GHDS 40
papers are tested.
Young’s moduli of papers at different water contents: The first point is the Young’s
moduli/ It is well known that paper and water do not mix well. Thus, it is easy to understand
that the water brought by the coating color during the coating process on the pilot will have an
unfavourable impact on the mechanical properties of the paper. The drying kinetics of a tensile
specimen after immersion and saturation in water had to be determined. Figure 5.24 shows
the evolution of the mass of a specimen after water saturation. Two specimens are blotted to
remove excess water before air drying.
Figure 5.24: Drying dynamic of tensile test sample (GHDS 35 paper) in conditionned room.
The first striking point is the decrease of the mass with time which seems exponential. It is
also noticeable that the signals are noisy. The balance used is a precision balance. This balance
has doors to isolate the weight measurement. In this case, to be sure to be as faithful as possible
to the drying phenomenon taking place in room where the tensile tests are carried out, the doors
of the balance were left open. This explains the noisy signal. It can also be seen that when the
excess water is removed with blotters, the mass value at t = 0 s is lower than when the excess
water is not removed. Blotters save drying time in one direction. For the rest of the study, the
values with blotters will be retained and modelled. The drying behaviour of the tensile specimen
Figure 5.25: Parity plot comparing modelled and experimental values from the drying of a
tensile test sample (GHDS 35 paper) in a conditioned room.
was modelled with a decreasing exponential law as presented equation 5.5.
t
m(t) = Kd · e τd + mf (5.5)
with Kd = m0 − mf in grams where m0 is the initial mass of the sample, τd is a time constant
(s) corresponding to the loss of 63% of the initial mass (3τd corresponds to the loss of 95% of the
initial mass and mf is the final mass of the sample). The parity plot comparing experimental
and model values is shown in Figure 5.25. The experimental data are globally well represented
(R2 = 0.97).
The modelled parameters are τd = 306.66 s, mf = 0.142 g and Kd = 0.088 g. The values
mf and Kd correspond to the experimental values. This model will therefore be retained for
the rest of the study. However, it should be kept in mind that this model is valid for the drying
of paper and not for the absorption of water: a hysteresis certainly exists but is not taken into
account in this study [144][26].
It is now possible to measure the Young’s moduli of different samples, following MD and
CD, at different amounts of water (thanks to equation 5.5 when dividing the calculate mass by
the sample surface). Thus, Figure 5.26 presents the evolution of the Young’s modulus for GHDS
35 as a function of the absorptiveness As , following MD and CD.
Whatever the direction (MD or CD), two plateaus can be seen on the two curves in Figure
5.26. Between these two plateaus, a more or less important decrease of the Young’s moduli is
visible. It should also be noted that the maximum values of the Young’s moduli are different
from those presented previously for the GHDS 35. The drying of the specimens is carried out
without stress contrary to what happens on a paper machine as well as on the coating pilot
where the paper is under tension. Without this pre-stress imposed on the paper, the Young’s
(a) MD (b) CD
Figure 5.26: Evolution of Young’s modulus in MD and CD versus absorptiveness As for GHDS
35 paper.
modulus value is lower than with pre-stress :
→ 8.92 GPa for EM D under constrained drying and 5.74 GPa under free drying;
→ 4.15 GPa for ECD under constrained drying and 2.20 GPa under free drying;
The shape of the curves is reminiscent of the behaviour found in rheology for shear thinning
fluids with two Newtonian plateaus. These fluids are modelled by Cross’s law (presented in the
literature review of this document). The form of this law will be used to model the evolution of
the Young’s modulus (CD or MD) as a function of absorptiveness. The following law equation
5.6 is therefore proposed for the Young’s modulus following MD (EM D ) :
EM D,min · (λs · As )ne + EM D,max

EM D = (5.6)
1 + (λs · As )ne
where EM D,min is the minimum value of Young’s modulus following MD (value of the high
absorptiveness plateau) in GPa and EM D,max is the maximum value of Young’s modulus fol-
lowing MD (value of the low absorptiveness plateau) in GPa. λ−1 s (in gwater · m−2
paper ) is the
EM D,min + EM D,max
absorptiveness value for which the Young’s modulus is . Finally ne is a
2
factor of severity of the Young’s modulus decay, it is dimensionless.
A similar reasoning is made for the Young’s modulus following CD ECD equation 5.7. The
meaning of the coefficients is the same as the equation 5.6 but following CD.
ECD,min · (λs As )ne + ECD,max

ECD = (5.7)
1 + (λs As )ne
Using the method of least squares and the non-linear GRG algorithm, this model was fitted to
the experimental data. The results of the modelling are presented in Table 5.8.
The models have a R2 > 0.98. There are nevertheless areas of under- and over-estimation.
−1
It is also worth noting the equality of the λs values between CD and MD (0.13 m2paper · gwater ).
Furthermore, the value of ne for ECD is larger than that for EM D , indicating a more severe
Parameters CD MD
λs 0.14 0.13
ne 38.43 16.66
Σr2 0.13 0.15
R2 0.986 0.995
Table 5.8: Results of modelling Young’s modulus versus absorptiveness using a Cross like
model for GHDS 35 paper.
Figure 5.27: Impact of the absorptiveness at half height λs value on the evolution of the CD
Young’s modulus ECD as a function of the GHDS 35 paper absorptiveness As , for ne = 38.43.
decay of ECD . To confirm the significance of λs and ne , a sensitivity study is carried out on
these two parameters. Thus, Figure 5.27 represents the evolution of ECD as a function of As ,
for different values of λs (ne being fixed at 38.43).
The black dots correspond to the experimental values. As the value of λs increases, the fall
in Young’s modulus occurs at low As . This validates the hypothesis of the significance of λs .
Similarly, Figure 5.28 corresponds to the study of the sensitivity of the parameter ne on the
model, λs being fixed at 0.13.
Here again, the black dots are the experimental values. When ne increases, the decrease of
ECD occurs for a larger value of As , but it is more abrupt. After verification, this value is found
to be λ−1
s . In any case, our hypothesis posing ne as a factor of severity of the growth of ECD is
validated.
The Cross like model is therefore presented to model the evolution of the Young’s moduli of a
paper as a function of the absorptivenes. This model differs from one proposed in the literature
[145]. It concerned the evolution of the tensile stiffness (kN · m−1 ). It is presented as equation
5.8.
Θt = Θt,0 · e−f ·mc(%) (5.8)
However, this model is used in an experiment where the paper sample under test was moist-
ened by water vapour, and no drying preceded the test. Furthermore, it does not consider an
upper plateau. Salmen also proposed a figure (cf. Figure 5.29) where the relative modulus of
paper is drawn as a function of moisture content [146]. If we just look at the curve shape, it can
Figure 5.28: Impact of the factor of severity ne value on the evolution of the CD Young’s
modulus ECD as a function of the GHDS 35 paper absorptiveness As , for λs = 0.14.
Figure 5.29: The relative modulus, where E0 is the modulus at 0 % moisture content, as a
function of moisture content. Solid line is the laminate model and the black dots the measured
value. The paper sheet used was isotropic [146].
be seen a similarity with our models. The question is still the values of λ−1
s . For λs = 7.69,
−1
we estimated the moisture content to 18% . When compared to Salmen results, the equivalent
s is associated to 10% of moisture content. This may suggest that our moisture content
to λ−1
value (18 %) is high, but the order of magnitude is not totally inconsistent.
s is large and one would expect
However, some points remain unresolved: the value of λ−1
Young’s moduli to fall for lower values of As . Furthermore, this model is based on the drying
of the paper and not on the absorption of water. As mentioned before, a hysteresis certainly
exists and it is not impossible that this leads to a more abrupt drop in Young’s modulus for
much lower values of As . Finally, the control of the drying can gain in precision and the tests
could give different values if the excess of water were not removed with the blotters.
Wet Stretch Dynamics analyses: This paragraph focuses on the results obtained using
the WSD technique presented earlier. In this case only GHDS 35 paper is tested using this
technique. Both sides of the paper are tested, following CD and MD, for 4 different forces.
Figure 5.30 shows the 5 experimental curves for the coated side of GHDS 35 against water, in
Figure 5.30: Example of experimental values and average value obtained during WSD test.
The sample here are in CD, tested 5 times with a force of 2 N. The average curve is also
represented. The GHDS 35 paper is used, with the coated face in contact with water.
CD (coloured markers), with a 2 N tensile strength of 2 N applied on the paper. The average
curve is also shown (black markers - 2N _ Average).
The high reproducibility of the measurements can already be noted. The standard deviation
is difficult to see, its maximum value is 0.14% (for an associated dispersion coefficient of 0.03%).
Thus, for the GHDS 35 having two different faces, two directions (CD and MD) and for 4
different test forces, the mean curves are plotted and modelled by a second order law according
to equation 5.9.
−t −t
  
1
εW (t) = Kp 1 − · τ1 e τ1 − τ2 e τ2  (5.9)
  
τ1 − τ2
where εW (t) is the deformation of the sample over time, Kw is the gain (%) and its value
corresponds to the maximum value of deformation. tau1 and tau2 are two time constants specific
to each system. The model curves are shown in Figure 5.31.
First of all, if the models between MD and CD are compared (Figure 5.31, a difference in
behaviour and maximum values can be observed. Indeed, the maximum elongation is much
higher in CD than in MD (identical behaviour to a tensile test). Furthermore, the behaviour
during the first 2 seconds is different. A horizontal slope is visible on the CD models and then
an inflection and growth of the slope value. This could indicate a second order behaviour for the
CD samples. It should also be noted that for times greater than 20 s for the CD models and for
times greater than 5 s for the MD models, the curves are always ordered in the same way: for
a given time, the greater the force, the greater the elongation. However, there is no regularity
of spacing between the curves. The other experimental curves can be found in Appendix G.
In addition, the parameters of the 16 models are presented in through histograms for a
(a) CD - coated side in contact with water (b) CD - uncoated side in contact with water
(c) MD - coated side in contact with water (d) MD - uncoated side in contact with water
Figure 5.31: Model curves obtained from WSD tests on GHDS 35 paper. The tests were
performed according to CD and MD (row), with the coated or uncoated side of the paper in
contact with water (column). The tests were carried out for 4 different strength: 1, 1.5, 2 and
2.5 N.
(a) tau1 (b) tau2
(c) Kp (d) R2
Figure 5.32: Parameters of the second order models obtained from the WSD curves, at
different forces, according to CD or MD, with the coated or uncoated side of the GHDS 35
paper in contact with water.
better reading Figure 5.32. First of all, for the tests following MD, the values of τ1 are equal to
0.00s. This means that the hypothesis made concerning the first order behaviour of the samples
following MD is validated. For the CD samples, 0.20 s ≤ τ1 ≤ 1.70 s. No logical order appears
in these values according to the force applied during the test. The second-order behaviour of
the tests following CD could be explained by a difference in water penetration between CD and
MD, which would result in an "ignition delay".
The values of τ2 are globally ordered: the greater the force applied during the test, the
greater the value of τ2 . The values of τ2 for the tests following CD are greater, at equivalent
force, than the values of τ2 following MD. Finally, the values of the coefficient of determination
R2 show that the models are, overall, much better for the CD samples than for the MD samples.
Using the models presented in Figure 5.31, 4 stress [MPa] versus strain [%] graphs could be
drawn. They were constructed in the following way: for each of the 4 tests, at a given time, 4
couples (strain, stress) are accessible. Figure 5.33 is one of these graphs (for CD_ uncoated).
For better readability only a selected number of times are shown but the data are available every
0.1 s. For short times, the points are aligned vertically. As time passes, they trace a straight
line with slope Ew,tan (WSD tangent modulus) and y-intercepts bw,tan . Ew,tan and bw,tan seem
Figure 5.33: Stress versus strain curves, at different time, from WSD. The sample are in CD,
with the GHDS 35 paper uncoated side facing water. Each markers represent a given time in
second.
Figure 5.34: Evolution of WSD tangent modulus Ew,tan over time. The values are for GHDS
35 paper in CD and MD direction, with the coated side or uncoated side facing water.
to tend towards a constant value.

Figure 5.34 is plotted adding the other 3 tests (MD_ coated, MD_ uncoated and CD_
coated). The slopes Ew,tan and orders at the origin bw,tan were determined for each curve and
plotted Figure 5.34 for Ew,tan and Figure 5.35 for bw,tan . For these two figures, the triangle
shape corresponds to the CD direction and the circles to the MD direction. Similarly, black is
related to the uncoated side facing the water and orange to the coated side facing the water.
Concerning the Figure 5.34 presenting the evolution of Ew,tan versus time, different points
are to be noticed. First of all, the values for the CD tests are lower than those for the MD
tests. The tested face has little or no impact. The values of Ew,tan at time 0.07 s are 8000.00
MPa (coated side) and 8150.00 MPa (uncoated side) for MD and 3925.00 MPa (coated side) and
3827.24 MPa (uncoated side) for CD. These values are very close to the values of the Young’s
moduli EM D and ECD of the tested paper (GHDS 35). Thus, the following hypothesis can be
Figure 5.35: Evolution of WSD vertical intercept bw,tan over time.The values are for GHDS 35
paper in CD and MD direction, with the coated side or uncoated side facing water.
made (cf. equations 5.10 and 5.11).
lim Ew,tan,M D = EM D (5.10)

t→0
lim Ew,tan,CD = ECD (5.11)

t→0
Thus the parameter Ew,tan could be a characteristic parameter of the evolution of the properties
of a paper during a WSD test, and of processes approaching this test, such as one-sided coating
for example.
With regard to Figure 5.35, which shows the evolution of bw,tan as a function of time, things
are different. The physical meaning of bw,tan is still difficult to approach. However, there is a
convergence of bw,tan at long times, the values are different between CD and MD. The behaviour
for times below 10 s is more difficult to explain. There is just one more difference between the
behaviour of the CD and MD samples. Moreover the curves "coated" or "uncoated" in CDs are
overlaid. As for the curves following MD, they are parallel for a time greater than 20 s, with a
difference of less than 1 MPa.
The WSD is not found much in the literature. However, the measurement principle is very
similar to what can happen on the coating pilot. The measurements carried out have shown a
difference in the behaviour of the different directions of the paper and a new parameter Ew,tan
has been developed to characterise the evolution of the mechanical properties of the paper as a
function of time (and consequently the quantity of water) during the WSD.
Finch test and wet tensile strength: The last paragraph of this sub-section will briefly
present the results of the Finch test for determining the wet tensile strength of a paper. Figure
5.36 illustrates the results obtained.
The GHDS 35 and GHDS 40 papers were tested according to MD and CD. The average
values according to MD are higher than those according to CD. However, this value is very
interesting for controlling the web tension in the coating process. Indeed, we now know that
for the GHDS 35, the process must be conducted in such a way as not to exceed a web tension
Figure 5.36: Wet tensile strength, obtained with the Finch’s test, for GHDS 35 and GHDS 40
in CD and MD.
of 85.27 ±15.25 N · m−1 with a wet paper, so ideally avoid exceeding 70 N · m−1 in the areas
where the paper still has a very high water content (typically between the coating zone and the
IR dryer).
5.2.2 Is it possible to forecast wrinkles with models?

The dry and wet properties of different papers are now known. It is therefore possible to model
the different behaviours that can occur during a tensile test or on the coating pilot. The results
of the models will be compared with the experimental values obtained in the first part of this
chapter.
5.2.2.1 Tensile test case
As described previously, the buckling coefficient kc (equation 2.43) is used in literature dealing
with plates stability under compressive or shear loading. Knowing the aspect ratio β, it is
possible to calculate kc and then to know the critical compression value σz,cr for which a plate
buckles. In the case of buckling due to constrained edges, a similar formulation can be obtained
for tensile loading σx,cr . Figure 5.37 presents the results obtained for the modelling of the
buckling coefficient kc versus plate aspect ratio β for GHDS 35 paper.
Based on the work of Zhu et al. [97] and the experimentally determined data, the kc criterion
was determined for different β and i (number of strip on the sample). The graph shows that for
a large aspect β > 1 ratio it is easier to buckle the plate than for an aspect ratio β < 1. Also,
for a given aspect ratio, kc increases with the number of wrinkles on the plate. This means that
it will take more stress to form more wrinkles. In this case, the web tension is increasing and
then the secondary compressive stress too.
For the tensile test of this study, the plate aspect ratio β = 0.15. The results are summarized
Table 5.9. First of all, it is possible to notice that it is easier to form wrinkles (or waves in this
case) when the sample is tested following CD than following MD. Moreover, the logical order of
things is respected: it takes more stress to form 2 waves than to form 1 wave.
(a) MD (b) CD
Figure 5.37: Buckling coefficient kc versus aspect ratio β for GHDS 35 for tensile test in CD
direction (left) and MD (right). i is the number of strip on the sample.
Direction i kc Critical tensile stress [MPa]

CD 1 190.75 5.78
CD 2 1332.68 40.36
MD 1 881.3 26.7
MD 2 6156.87 186.57
Table 5.9: Buckling coefficient and critical tension stress to have buckling determined with Zhu
model, for the GHDS 35 paper following CD and MD.
Figure 5.38 shows the experimental results and compares them with the results presented in
Table 5.9. The lines represent the critical tensile stress to go from i waves to i +1, depending
on the orientation of the fibres in the sample (MD or CD), on a dry paper.
Concerning the threshold values following CD, it can be seen that the values chosen experi-
mentally to go from 0 to 1 wave are very high compared to the models. The same observation
is made for the samples following MD. This suggests that the limit criterion for considering
experimental buckling of a specimen is possibly overestimated. As for the threshold value be-
tween 1 and 2 waves, following CD, it is close to the values considered experimentally. The
difference is on the out-of-plane displacement value considered. Concerning MD, the threshold
value between 1 and 2 waves is higher (186.57 MPa) than the value of stress at failure of the
samples. Physically, it does not exist and cannot be reached, which agrees with reality.
However, two points should be kept in mind: in the present study kc is calculated in the
case of a plate in tension where both ends are clamped, and the load is uniformly distributed.
However, these two assumptions cannot be applied in the context of the work carried out on the
coating pilot (simply supported plate and non-uniform web tension on the width). Nevertheless,
this criterion is in line with the reality of the observed facts. It would just be necessary to review
the conditions at the limits and to consider the use of a stress function to determine the evolution
of the web tension over the width. Finally, such a criterion would be ideal for an industrial
production. It would help to dimension roll-to-roll process by knowing the characteristics of the
web. It makes possible to choose the good span between rollers in order to avoid wrinkles while
Figure 5.38: Comparison between experimental data and model values of critical tensile
buckling limits for GHDS 35 paper of "dry" GHDS 35 paper.
having a sufficiently large web width to allow the production.
5.2.2.2 Pilot case
The experimental data are compared with the theoretical model for orthotropic materials pro-
posed by [124] (equation 5.1). Figure 5.39 compares them. In this case, the friction coefficient
µ between paper and roll is unknown. It is a delicate parameter to determine. Thus, this
factor was considered as a variable to match as well as possible with the experimental values
and to be able to determine the critical tension Tcr from which wrinkles can appear (equation
5.2). It was set at µ = 0.17 as it seems consistent with the observed phenomena: this value
enable to match the experimental boundary between troughs and wrinkles, especially at high
misalignment values.
However, the experimental misalignment angle values obtained in this project are high com-
pared to the values that can be found in the literature (Figure 5.40 et Figure 5.41). This point
is still under investigation.
Impact of the discretization of the wrinkles number i: During this study, it was found
that making the variable representing the number of wrinkles i continuous raise two questions.
The first of these issues is to ask what is the physical representation of a non-integer number
of wrinkles. The second issue is the impact of the discretization of the variable i on the model
linking misalignment angle and web tension. In order to answer these questions, two figures
have been drawn: one for a paper considered dry (Figure 5.42a), one for a paper saturated with
water (Figure 5.42b). They represent the evolution of the misalignment angle as a function
of the web tension (equation 5.1) as well as the evolution of the number of wrinkles i versus
the web tension. They are plotted twice: once considering i continuous, once considering i as a
integer. In the case of dry paper, the maximum number of wrinkles in the considered web tension
Figure 5.39: Comparison of experimental and model values for the evolution of misalignment
angle versus web tension for dry paper (µ= 0.17). The paper studied is the GHDS 35,
uncoated side touching the rollers.
Figure 5.40: Highlighting the web width (L here) (row) and web speed (columns) on the areas
of wrinkle appearance by Hashimoto [124].
Figure 5.41: Demonstration of the impact of the friction coefficient on the critical band voltage
value by Martz [126]. The critical band tension is inversely proportional to the friction
coefficient (all other things being equal).
range is 4 (integer i) and less than 4 (continuous i). In the case of water-saturated paper, the
maximum value is 6 (integer i) and is slightly lower than 6 in the continuous case. Comparing
the model curves linking misalignment angle and web tension, it appears that the misalignment
angle values are alternately overestimated and underestimated when i is continuous.
The model curves for discontinuous i also show a sharp drop in the permissible angle value.
These drops correspond to the appearance of a new wrinkle according to the variation of dis-
continuous i. Thus, these sudden variations can be interpreted as a zone of instability: they
are i wrinkles and the system tends towards a zone of unstable equilibrium before having i+1
wrinkle. This would also lead to the following interpretation: having a non-integer number of
wrinkles would be a transition phase between two different equilibria. It should also be noted
that as the number of wrinkles i increases, the range of web tension allowing i wrinkles increases.
It is also observed that the curves representing the number of continuous and discrete wrinkles
as a function of the web tension intersect at several points. When there is a break with a
horizontal plateau, the value of the associated web tension corresponds to a value for which the
boundary criterion (equation 2.57) is no longer verified for a discontinuous number of wrinkles.
When there is a break with a vertical zone, the boundary criterion is true again. Thus, to the
right of a cut with a vertical zone and to the left of a cut with a horizontal plateau, the boundary
equation is true. To the right of a cut with a horizontal plateau and to the left of a cut with a
vertical zone, the bounding equation i is not verified.
5.3. CONCLUSION 145
(a) Dry (b) Wet
Figure 5.42: Evolution of misalignment angle and number of wrinkles versus web tension for a
considered dry or wet GHDS 35 paper, with discrete and continuous i. The uncoated side is
touching the rollers.
5.3 Conclusion
In this chapter different points concerning the wrinkling mechanics have been put forward. The
buckling phenomenon has been highlighted in two different ways. The first one is to show
buckling in a tensile test. During these tests, 1 (MD) or 2 (CD) buckling waves appeared.
The values were compared with the values that could be obtained with the Zhu model [97].
The experimental criterion chosen over estimated the stress values for which buckling could
occur. However, the orders of magnitude are respected. The other way to generate buckling
and by extension wrinkles is to use the coating driver by choosing the values of misalignment
angle and web tension. Wrinkling phenomena have been observed on different papers and
at different scales. The wrinkle shapes could be modelled with a 4-parameter Weibull law.
Knowing the wrinkle shape and their location on the web width makes it possible to choose the
type of correction to be made (roll shape, orientation, etc.) and can be a measure of sensitivity
/ a criterion of discrimination to the wrinkle for different papers. At the microscopic scale,
delamination was observed, where pinching was observed at the macroscopic scale. In order to
model the wrinkling phenomenon on the pilot, laws proposed in the literature by Gehlbach or
Hashimoto were used [124][122]. In order to use these laws, it was first necessary to know the
dry and wet properties of the papers used. This study has enabled the development of a new
parameter called "WSD tangent modulus" and noted Ew,tan . This parameter makes it possible
to quantify the evolution of the mechanical properties of the paper over time, during a WSD test.
In addition, thanks to tensile tests on wet paper specimens, it was possible to model the evolution
of Young’s moduli, as a function of absorptiveness As , using a Cross-like model. Finally, the
comparison between the experimental values obtained on the pilot and the models showed that
the experimental values were very far from those found in the literature. The hypothesis made
to explain this difference is that in our study, the web speed was much lower than in Hashimoto’s
work. This impacts the air film that may exist between the roll and the paper. This affects the
friction coefficient µ between paper and roller and therefore the critical misalignment angle θcr .
However the model isolated properly the wrinkled and unwrinkled areas.
Chapter 6
Study of the coating process on the
laboratory pilot
Contents
6.1 Control of the temperature during the process . . . . . . . . . . . . . 148
6.1.1 Factors influencing a non contact IR temperature measurement . . . . . 148
6.1.2 Inhomogeneous temperature and sensor position on the web width . . . 152
6.1.3 Following the process temperature . . . . . . . . . . . . . . . . . . . . . 155
6.2 Process optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
6.2.1 Process regulation: web speed, tension and misalignment angle . . . . . 158
6.2.1.1 Creation of a control map to run the coating process . . . . . . 158
6.2.1.2 Use of D.I.C during the process . . . . . . . . . . . . . . . . . 165
6.2.2 Characteristics of finished products . . . . . . . . . . . . . . . . . . . . . 167
6.2.2.1 A wrinkle free coated paper . . . . . . . . . . . . . . . . . . . . 167
6.2.2.2 Extension to very lightweight paper grades . . . . . . . . . . . 169
6.3 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
In this last chapter results, the process aspects of this project will be discussed. First, the
temperature measurement during the process on the laboratory coating pilot will be presented.
A study of the factors impacting the non-contact temperature measurement of the paper web
by IR measurement will first be carried out; before looking at the temperature profile on the
web during the process. Finally, it will be shown to what extent temperature monitoring on the
coating pilot is useful in the context of laboratory coating. This section only focussed on the
GHDS 35 paper.
In a second step, an optimisation of the process will be presented. This will use the previous
chapters in order to create a control map that will allow the process to be carried out under the
best conditions while controlling the web speed, the web tension and the misalignment angle.
Following this optimisation, the papers obtained will be presented and a perspective will be
proposed for the use of CPRAW and CPBS papers.
147
148CHAPTER 6. STUDY OF THE COATING PROCESS ON THE LABORATORY PILOT
and measuring height hs (mm) (standard deviation: ±1 °C).
6.1 Control of the temperature during the process

6.1.1 Factors influencing a non contact IR temperature measurement
First of all, the impact on the temperature measurement of the measuring height hs (mm), the
measuring angle κs (°) as well as the emissivity set on the sensor εs have been studied. In
Figure 6.1, the temperature measured by the sensor Tmeasured is compared with the heating
temperature Theating , which the paper sample is supposed to reach, for different measuring
heights. It should also be noted that the measurements are made for a measurement angle κs
= 0° and an emissivity εs = 0.95 [147].
Thus, Figure 6.1 shows that for temperatures between 20 °C and 60 °C, no major difference is
found between the heating temperature and the measured temperature. However, from a heating
temperature greater than 60 °C, the measuring height starts to impact more and more the
temperature measured by the sensor. Indeed, a difference of 20 °C can be observed between the
heating temperature and the measured one for a measuring height hs = 300 mm. This continues
to increase up to 190 °C where a difference of 40 °C can be observed between the measured and
heating temperature for hs = 250mm. The origin of this variation seems to be twofold: the
further the sensor moves away from the measuring sample, the more the measuring surface grows
and thus if the heating of the sample is heterogeneous, the measured temperature (which is an
average over the measuring surface) will be affected. The other important point is the increasing
impact of the environment, whose temperature can change the measured temperature, as well
as a draught that will be more easily noticed if the sensor moves away from the sample.
Figure 6.2 shows the volution of the measured temepratud versus the heating temperature
at different measure angle κs which varies from 0 ° to 60 ° (hs = 150 mm and εs = 0.95). Small
6.1. CONTROL OF THE TEMPERATURE DURING THE PROCESS 149
Figure 6.2: Measured temperature by the IR (°C) sensor versus the heating temperature (°C)
and measuring angle κs (°) (standard deviation: ±1 °C).
Figure 6.3: Schematic views of the impact of the measurement angle on measuring area shape
(h = 150 mm).
differences are observed between the heating temperature and the measured temperature up to
60 °C. However, as for Figure 6.1, above 60 °C significant and not negligible differences between
Theating and Tmeasured are observed. Indeed, differences of 50 °C are observed for Theating = 190
°C with a measurement angle κs = 60 °. In addition, it is also possible to note that the curve
corresponding to the equality between Theating and Tmeasured is close, or merge with the curve
for a measurement angle of 0 °.
The origin of these variations seems to be the variation in size and shape of the measuring
area. Figure 6.3 may help to understand this phenomenon. Using the sensor manufacturer’s
data [3], it is possible to obtain the size and shape of the measuring area as a function of hs
(mm) and κs (°). Two measuring angles κs were tested in order to show the evolution of the
measured surface.
For κs = 0 ° (Figure 6.3 (a)), the measurement surface is a disc with a center of coordinates
Figure 6.4: Highlighting of the evolution of the surface of (a) the measurement area and (b)
the evolution of the ellipse-centre distance of the sample (OM distance).
(0.0), whose radius increases as the measuring height hs increases. However, when κs increases,
the measuring disc becomes a measuring ellipse whose center is no longer (0,0) but M (x, 0)
where x ∈ R∗+ and such as x = f (κs ) (Figure 6.3 (b)). This symbolizes the translation of the
center of the ellipse M along the horizontal axis x. The variation of the large radius of the ellipse
remains minimal for κs = 30 ° compared to the radius of the circle for a given hs . However, at
κs = 60 °, the elliptic shape is quite recognizable and the variation of the large radius of the
ellipse is very important.
It is also illustrated with Figure 6.4. In fact, with this figure, it is easy to show the impact of
hs and κs on the measuring area and the M -point shift. One will even notice that for hs = 300
mm, the measuring ellipse is outside the sample area therefore unheated areas could be seen by
the sensor. It is also necessary to consider the possible variations of out-of-plane heights which
could cause other problems.
The last point of attention is the emissivity set on the sensor εs . In Figure 6.5, the measured
temperature is again compared to the heating temperature but this time by varying the emissiv-
ity εs between 0.1 and 1.In this case hs = 150 mm and κs = 0 °. As a reminder, the emissivity
εs corresponds to the radiative flux of the thermal radiation emitted by a surface element at a
given temperature, compared to the reference value which is the flux emitted by a black body
at this same temperature.
As soon as the temperature exceeds 25°C, the emissivity impacts more and more the tem-
perature measurement when εs < 0.80. Indeed, when Theating = 190°C, a difference of 610°C
is noted between the heating and the measured temperature for εs = 0.1. This gap decreases
with the increase of εs until it reaches the limit represented by the heating temperature (curve
Measured temperature = Heating temperature on Figure 6.5).
In order to understand this phenomenon, it seems interesting to refer to Stefan-Boltzmann’s
law (equation 6.1 [148] considering the emissivity εs :
M = εs · σ · T 4 (6.1)
where Tmeasured is the absolute temperature (K), M is the energy emittance (W · m−2 ) and
σ the Stephan-Boltzmann constant (5.6697.10−8 W · m−2 · K −4 ). It is therefore realized that a
and sensor emissivity εs (standard deviation: ± 1 °C).
Emissivity
Theating (°C) 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.85 0.90 0.95 1.00
21.3 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.0 0.0 0.0 0.0 0.0
57.5 50.4 26.7 17.0 11.4 7.7 5.2 3.2 1.7 1.1 0.5 0.0 0.5
100.0 78.2 42.0 26.5 17.6 12.0 7.8 4.9 2.6 1.7 0.8 0.0 0.9
120.0 92.4 48.7 31.0 20.9 14.1 9.4 5.9 3.1 1.8 0.7 0.0 1.1
150.0 108.9 57.0 35.8 24.9 16.6 11.5 7.4 3.6 2.6 1.4 0.0 0.7
190.0 130.1 69.1 43.6 29.4 20.2 13.6 8.6 4.6 2.9 1.4 0.0 1.3
Table 6.1: Relative error (%) on the measured temperature compared to a reference
(εs = 0.95).
change in emissivity will impact the temperature read by the sensor and thus generate an error
[149]. The relative error δαr (%) was calculated as follows (equation 6.2) and reported in Table
6.1
Tmeasured,εs − Theating,εs =0.95

δαr = × 100 (6.2)
Theating,εs =0.95
Table 6.1 shows that the relative error on the measured temperature related to the emissivity
εs setting has a non-negligible impact. Indeed, relative errors of 5% or more are common.
Moreover, this is mostly the case for temperatures greater than or equal to 100 °C. For example,
when εs = 0.1, δαr = 78.2% for Theating = 100°C, which can have a considerable impact on the
process and the product.
6.1.2 Inhomogeneous temperature and sensor position on the web width
This part deals with the positioning of the temperature sensor on the paper width. Preliminary
tests had highlighted a temperature heterogeneity at the exit of the IR dryer. The hypothesis
is that the positioning of the sensor on the web width would impact the measurement and thus
mislead an operator in a paper-mill. Figure 6.6 represents the position of the isotherm curves
on a web coming out of an IR dryer, for web speed ranging from 5 m · min−1 to 10 m · min−1 .
As can be seen, the area under a given isotherm decreases as the web speed increases. In
fact, the dryer’s power is constant, it delivers the same energy in a given time. However, as the
speed increases, the time available for the paper to recover this energy and to dry is reduced.
Thus, at a higher speed, paper will heat less and will exit the dryer colder.
About the the shape of the temperature profiles, it can be seen that the different isotherms
are not straight vertical lines. This could be related to the difference in energy that the paper web
receives at its centre and at its edges. Figure 6.8 propose an explanation to this phenomenon.
By considering the problem symmetrical (OO’ axis symmetry), it is noted that the reception
areas 1 and 1’ are very brightly illuminated by the IR lamp: they receive energy from the
emission points O, A, A’, B and B’. As for the reception zones 3 and 3’ only receive energy from
the emission points B, B’, C and C’. Moreover, the emission zone of these last 2 points is only
50% useful. Please note that, as the IR dryer manufacturer does not give all the characteristics
of their equipment, Figure 6.8 is a schematic figure to illustrate the above hypothesis. However,
one side (0 mm) seems hotter than the other (300 mm). Asymmetric heating of the IR lamps
or an off-centred web could be the reason.
If the attention is now focused on the positioning of the sensor on the web, considering
the information given in Figure 6.6, it is legitimate to wonder where and how exactly to put
the sensor. Depending on the position on the web width and the temperature profiles, the
measurement would not be the same everywhere on the web width. Figure 6.9 shows an IR
image of the paper web after the IR dryer. Using processing software, grey levels (corresponding
to temperatures) were measured in 3 different surfaces to see the impact of the shape of the
measurement area.
It is now possible to see that in a disc-shaped measurement area (sensor perpendicular to the
web), an average temperature of 113 ±9 °C is measured, just like for the horizontal ellipse. For
the vertical ellipse, the average temperature measured is 128 ±15 °C. With the uncertainties,
it can be seen that no statistical difference is seen. However, when the temperature is read on
the sensor by a machine operator, for example, only the mean value will be visible. This will
result in an incomplete interpretation of the measurement. Thus, it will be more suitable to
work with a disc-shaped measurement area to avoid any lecture measurement error and to keep
a good runnability.
Thus, when using a temperature sensor, care must be taken to ensure that the emissivity
εs is correctly adjusted if possible. An error of 0.2 on the emissivity is acceptable to have an
accurate measure (with 2.6% error at 100 °C). The angle of measurement κs must be equal to
0 ±2 °, between the studied support and the sensor. The distance hs between the sensor and
the substrate depends on what must be measured but the further away you are, the larger the
(a) Speed: 5 m · min−1
(b) Speed: 10 m · min−1
(c) Speed: 20 m · min−1
Figure 6.6: Highlighting of the evolution of temperature isotherms on the paper web according
to different web speeds (standard deviation: ±2 °C). 300 mm correspond to the driver side of
the coating pilot.
Figure 6.7: Displacement of the 85°C isotherm at different web speeds (standard deviation: ±2
°C). 300 mm correspond to the driver side of the coating pilot.
Figure 6.8: Area illuminated by an IR lamp and overlapping of the illumination areas).
Figure 6.9: Different placement for IR sensors on a web after an IR dryer (web speed: 5
m · min−1 , hs and εs constant, κs = 60° for the ellipses).
measurement area will be and will consider temperature heterogeneity. In the present case, a
measurement angle κs = 0, a measurement height hs = 30mm and an emissivity of εs = 0.95
are kept and will enable to have good measures.
The study of the temperature profile at the exit of the dryer can be a good element of
Figure 6.10: Temperature monitoring screen installed next to the driver.
decision to choose the exact location on the web width of a temperature sensor with the couple
(height h, position on width). Moreover, it can also help to understand problems induced by
heterogeneous drying. In our case, the placement at 30 mm from the dryer outlet and in the
middle of the web would be relevant.
6.1.3 Following the process temperature
This sub-section will look at the practical use of the temperature sensors studied earlier. These
are installed on the coating pilot and are connected to a data acquisition card, which is connected
to a computer. A home made python code is used for data acquisition and communication
between the different elements (cf. Figure 6.10).
It can be seen immediately that the signal is very noisy. This phenomenon will be found on
the other figures resulting from the measurements of temperatures during coating on the pilot.
No explanation has been found to explain this phenomenon which was not present during the
laboratory measurements presented subsection 6.1.1. But, speed could be the reason of this
noisy signal. Figure 6.11 shows the temperature data recorded during a run on 22 March 2022.
The temperature T 0 is measured after the coating station, 200 mm before the entrance to the IR
dryer. The temperature T 1 is measured 300 mm after the exit of the IR dryer. The temperature
T 2 is the temperature of the paper web between the two hot air dryers. Finally the temperature
T 3 is measured 230 mm after the exit of the second hot air dryer.
Between 0 and 2000 s, the two hot air dryers are switched on, so that the set temperatures,
130 and 150 °C respectively for hot air dryers A and B, can be reached, according to an internal
temperature control loop. Once the plateaus are stabilised, the IR dryer is switched on. However,
for safety reasons, the IR dryer only works if the paper web has a minimum speed of 3 m·min−1 .
Therefore, the paper speed is set to 5 m · min−1 in order to allow the IR dryer to start. We
can now look at the events taking place between 2100 and 3900 s. It can be seen that the
temperature T 1 (red curve) drops 5 times in the considered time interval. This is due to the
contact of the paper with the applicator roller. The starch gel is then deposited on the paper.
The energy used to heat the paper before is used to dry the deposited (aqueous) starch gel. The
Figure 6.11: Trial conducted on 22-03-2022 on the GHDS 35 paper with a starch gel coating.
paper therefore comes out cooler. The different peaks can be explained by repeating the tests
with different parameters. Shortly before 4000 s, there is a clear drop in the temperatures T 1,
T 2 and T 3. This phenomenon is due to a breakage of the paper web. Following this problem, it
takes about 30 minutes to clean and replace the web. It is also necessary to wait for the hot air
dryers temperature to decrease to avoid any risk of burning. This time is reduced by allowing
the hot air to come on. However, this will make it more difficult for the web to pass through
the hot air dryers because of the air circulation. At 6000 s, the same process resumes: heating
the hot air dryers, heating the IR dryers and resuming the tests.
Figure 6.12 shows the measurements taken on 23 May 2022, on CPRAW paper and GHDS
35, during a water coatin. The paper is just re-wetted with the applicator roller and the excess
is removed with a Mayer bar .002 or .012 here.
Between 0 and 400s, the system worked with GHBS 35 paper. At 400s, the IR and hot air
drying systems are voluntarily shut down. This can be seen by the stepwise decrease of T 1: first
the dryer is stopped but the web is still moving to avoid any risk of overheating (first step);
then the web is also stopped (second step). In 400 and 1200 s the GHBS 35 is replaced by
CPRAW paper. A first heating phase is performed ("Heating n°1"). This is fast (250 s) because
the system was not completely cooled down. Then the first test starts with the Mayer bar .002.
We start slowly to see the behaviour of the paper by increasing the speed to 20 m · min−1 . Once
this speed has been reached, we go back up to 15, 10, 7.5 and 5 m · min−1 in order to be able
to take photos with the D.I.C. system. It is then decided to carry out the same tests with the
Mayer .012 bar (which means that the excess of water removed will be more than with the .002).
At 20 m · min−1 , the web resisted. At 15 m · min−1 , it breaks directly after the IR dryer. The
test is stopped.
Figure 6.12: Trial conducted on 23-05-2022 on CPRAW paper and GHDS 35 paper with a
"water coating".
Finally, Figure 6.13 represents the data acquired during a trial carried out on 24 May 2022.
"Trial n°1" was performed on the CPRAW with a .006 Mayer bar. Then the reel is changed to
CPBS paper. The latter is very difficult to use and broke several times as it seams very sensitive
to water. It was possible to use the CPBS paper only at 30 m · min−1 , with a Mayer bar .002.
Figure 6.13: Trial conducted on 24-05-2022 on CPRAW paper and CPBS paper with a "water
coating".
Thus, after observing the 3 previous figures, several conclusion can be drawn. First of all,
temperature sensors are very useful in order to better understand the phenomena taking place
during the process. They allow the monitoring, indirectly, of certain production events: change of
equipment, web breakage, coating, preheating of dryers. They also make it possible to determine,
thanks to these same indirect measurements, the time required for a given operation. The most
obvious example is web breaking. It takes up to 30 minutes to restart paper production on our
pilot. In this case, the temperature sensors really allow us to acquire information on production
and their importance is not negligible. Overheating can lead to deformation of the material, and
underheating will lead to inadequate moisture content of the material, leading to problems for
future use. In an industrial case, the means available and implemented are considerably more
important and the use of temperature sensors is one means among others to monitor production.
6.2 Process optimization
In this part, process optimization will be discussed. In a first step, the results presented in the
previous chapters will be used together to present a control map to drive the process in the best
way. To do so, the models described and developed previously will be used together. In a second
step, the coated papers will be presented and briefly studied.
6.2.1 Process regulation: web speed, tension and misalignment angle
On the coating pilot presented previously, different parameters can be controlled: temperature,
speed, web tension, misalignment angle, number of IR lamps on,... Within the framework of
this project, we have chosen to take into account the speed, web tension and misalignment angle
in order to build a process control document.
6.2.1.1 Creation of a control map to run the coating process
In order to run the process in the best possible way, the idea came up to create a document that
could be used by all people, from operators to engineers. The choice fell on a control map. The
"manufacturing" philosophy of the latter is presented in Figure 6.14.
The starting point is the web speed Vweb : this is between 5 and 40 m · min−1 on the pilot.
From this speed, the contact time tc between the water and the coating is defined. This contact
time is defined as a function of the distance that the paper travels between the coating station
and the IR dryer inlet. It is noted dc−IR and is equal to 0.70 m. Consequently, tc is defined by
the equation 6.3
dc−IR
tc = (6.3)
Vweb
The time tc is used in the following equation relating the absorptiveness As (t) and time. The
equation 6.4 is thus obtained.
6.2. PROCESS OPTIMIZATION 159
Figure 6.14: Construction diagram of the control maps as a function of the paper
absorptiveness or of the web speed.

dc−IR 
As (Vweb ) = Aeq 1 − exp τ2 · Vweb  (6.4)
 
Subsequently, the equation 5.6 and the equation 5.7 will allow the link between the Young’s
moduli EM D and ECD and the absorptiveness As (Vweb ). Note that in this scheme a correction
of the Poisson’s ratio νxz and νzx will be made thanks to Baum’s empiric models (equation 3.4).
The Poisson’s ratio and Young’s modulus will be inserted in the equations of the misalignment
angle (equation 5.1) and the critical band tension (equation 5.2). With this approach, it was
possible to generate control maps as a function of absorptiveness Figure 6.15 and Figure 6.16.
The difference between Figure 6.15 and Figure 6.16 lies in the variation of the β value, which
corresponds to the web aspect ratio. Similarly, the control maps Figure 6.17 and Figure 6.18
were generated but this time as a function of the web speed.
What can be seen in Figures 6.15, 6.16, 6.17, 6.18 is that the higher the coefficient of
friction, the lower the critical web tension. Moreover, this critical web tension decreases as the
absorptiveness increases (and therefore as the web speed decreases). Regarding the impact of
β, we note that the greater it is, the less tolerant the system is to de-salting, especially at high
absorptiveness (and therefore at low web speed).
The remaining question is: how to use these maps? We invite the reader to proceed as
follows using Figure 6.17b. Choose a web speed value and draw a vertical line on the map
starting from the chosen speed. Choose a web tension value and draw a horizontal line from
that web tension value. The intersection of the horizontal line and the vertical line will give
the maximum permissible misalignment angle value to avoid shear wrinkles. If the web tension
value is below the critical web tension value, it is even better: you have "the boots and the
umbrella". To give an example, if the chosen speed is 5 m · min−1 and the chosen web tension
is 200 N · m−1 , then the intersection of the two lines gives a critical misalignment angle value
of 1.20 ° (in Figure 6.17b). If this angle is too high, it is then possible to decrease the speed or
increase the web tension. This corresponds to what David Roisum [54] already proposed for the
web tension but it is the opposite of what he proposed for the speed. Indeed, for the speed, the
graph proposed by David Roisum was not in the framework of a coating process.
However, one point remains sensitive. Experimentally, creases could appear although no roll
misalignment could be measured or at very small misalignments, especially on rewetted papers.
On these wrinkles, one point was striking: it was parallel to the direction of paper travel and
not aligned like the classic shear wrinkle. Earlier in this paper, tension buckling phenomena
were discussed, but in the context of tensile testing. It would appear that tension buckling is
also a phenomenon present on the pilot but more in the context of rewetted papers. To justify
this, the evolution of the buckling coefficient (according to MD) kc as a function of β is again
presented in the dry state, but also for a paper saturated with water Figure 6.19.
In Figure 6.19 , it can be seen that for the same value of β in dry or wet conditions,
and for the same number of folds i, the buckling coefficient is lower in wet conditions than in
dry conditions. The buckling coefficient even collapses. By extension, the lower kc is (for a
given β and i), the lower the associated critical buckling stress will be to make the i troughs
(a) β = 0.2 and µ = 0.2
(b) β = 0.2 and µ = 0.9
Figure 6.15: Control map for GHDS 35, at different paper friction coefficient µ and for a web
aspect ratio β = 0.2.The green lines at the bottom of each subfigures corresponds to the value
of the critical web tension that should not be exceeded in order to avoid shear wrinkling.
(a) β = 0.5 and µ = 0.2
(b) β = 0.5 and µ = 0.9
Figure 6.16: Control map for GHDS 35, at different paper friction coefficient µ and for a web
aspect ratio β = 0.5. The green lines at the bottom of each subfigures corresponds to the value
of the critical web tension that should not be exceeded in order to avoid shear wrinkling.
(a) β = 0.2 and µ = 0.2
(b) β = 0.2 and µ = 0.9
Figure 6.17: Control map for GHDS 35 as a function of web speed, at different paper friction
coefficient µ and for a web aspect ratio β = 0.2. The purple lines at the bottom of each
subfigures corresponds to the value of the critical web tension that should not be exceeded in
order to avoid shear wrinkling.
(a) β = 0.5 and µ = 0.2
(b) β = 0.5 and µ = 0.9
Figure 6.18: Control map for GHDS 35 as a function of web speed, at different paper friction
coefficient µ and for a web aspect ratio β = 0.5. The purple lines at the bottom of each
subfigures corresponds to the value of the critical web tension that should not be exceeded in
order to avoid shear wrinkling.
(a) dry (b) wet
Figure 6.19: Buckling coefficient kc versus aspect ratio β for GHDS 35 in MD, in dry (left) or
fully wet (right) state.
appear. Thus, it is possible to understand that in the case of the pilot, when the paper is wet, a
combination of shear buckling and tension buckling effects can exist and lead to the appearance
of wrinkles although very little (if any) misalignment has been measured. It also highlight that
the Hashimoto model is incomplete : tension buckling without misalignment should be added
to the model. In addition, the control charts do not take into account the change in stress at
break between dry and wet paper, and the deformation phenomena (as observed on the WSD)
that may be present.
The readers must keep in mind that the presented control maps are valid for a given triptych
paper, coating color and coating device. Nevertheless, the control map is a very useful tool to
help the process control and the concept could be reused in the industrial context. It could
even be improved thanks to the experience of the operators on the machine. Indeed, operators
with 10, 15, 20 or 30 years of experience in paper mills are a very important resource, allowing
to advance the understanding of certain phenomena sometimes difficult to approach from a
theoretical point of view.
6.2.1.2 Use of D.I.C during the process
As presented earlier in this document, a digital stereo-vision system have been installed on
the coating pilot for process monitoring. This adapted system using stereo-vision technology
have been designed to enable wrinkles detection and characterisation during processing on the
coating pilot. The APS-C (Advanced Photo System type-C) sensors of the two DSLR with the
24 mm focal length of the two lenses give a 49,84 ° horizontal field of view that allows placing
the cameras at approximately 35 cm from the characterised paper to visualise the whole width.
Considering that configuration and the definition of the DSLR sensors (5184 x 3456 px2 ) the
resolution of the digital images is about 63 µm · px−1 . Such a system is compact and might
detect and analyse submillimetric wrinkles. The distance and the tilt between the cameras are
approximately, and respectively, 30 cm and 30 °.
The stereo-system is installed on a rigid metallic truss-structure that enables adjustments
Figure 6.20: Stereo-system mounted on the coating pilot (left) and printed paper with wrinkle
taken from both cameras (top right). 3D Reconstructions (bottom right) have been performed
in order to analyse the wrinkle geometry.
for clear and sharp images acquisition. During the first trials done, it seems that this truss
strucutre does not suffer form vibration created by the coating pilot. DSLR cameras are remotely
controlled and energetically autonomous for eliminating every noise after calibration. After
installation of the image acquisition system, the first step consists on calibrating the cameras
using a chess-board that is located and rotated at different positions and angles every time a
picture has been taken. The calibration phase is done once, since the system has not been moved
or changed, and give the intrinsic parameters of the DSLRs and the positions and rotations of
one camera relative to the other one. Knowing these parameters, it is possible to compute the
position of a point in the 3D space if that point is observable on both pictures taken from the
two DSLRs thanks to triangulation computation.
A hand-crafted code has been developed in order to reconstruct the surface of the charac-
terized paper thanks to stereo-vision. Difficulties in that project consist on correlating both
images since the surfaces to characterised are uniformly diffused whether the wrinkles are not
big enough, and computing quickly to correct the pilot behaviour fast enough. At the time
of writing, the first difficulty has been resolved by projecting a precise and random speckled
pattern on the surface of the paper. Considering the other difficulty, the code is being developed
in order to increase the computation speed.
A first experiment of the stereo-system coupling to the computation code is shown in Figure
6.20. The characterised wrinkle is large, however that test proves the feasibility of the surface
reconstruction. The 3D reconstruction (at the centre bottom) was chosen to be fast to compute
with a large stereo-grid and small stereo research areas. Because of that last point, it was
impossible to find the stereo-correspondent points at the top of the wrinkle (white areas on
Figure 6.20). Another computation was performed in order to test the resolution of the system:
the linear reconstruction of a cross-section of the wrinkle (at the right bottom). The linear
profile was chosen to be dense (very thin stereo-grid) with large stereo research areas. The
linear reconstruction gave a complete profile of the wrinkle and proves that 3D reconstruction
with dense stereo-grids and large research areas might characterise wrinkles that appear on the
coating pilot. Tests on smaller sizes of wrinkles are currently performed and multi-threading
(a) Left picture of the DIC system. (b) Right picture of the DIC system.
Figure 6.21: CPRAW paper with wrinkles passing in front of the DIC system and a speckles
spotlight (red laser). A contrast issue is seen.
techniques are currently developed on the stereo-correlation part in order to speed up the process.
The DIC system have also been tested on the CPRAW paper as presented Figure 6.21. As
seen, the image quality is not good and not enough contrast could be recovered and the exposure
is not good. In fact, the photo set-up used were the ones for the GHBS 35. It means that if the
paper is changed during the process, the camera set-up should also be changed. This could lead
to movement and leads to another calibration.
Figure 6.22 show the GHDS 40 paper used on the pilot. One wrinkled s easily visible on
the right photo but much less on the left one. The light used could be the reason of this, while
creating a shadow on one side of the wrinkle. The pictures exposure are better even if they are
still blur. The shutter speed of the cameras is difficult to adjust because a compromise must be
found between a large aperture (f /18), a not too high ISO (=3200 here) value to avoid digital
noise linked to the cameras and the shutter speed (1/250 s). It was therefore necessary to invest
in industrial cameras more adapted to the phenomena (Figure 6.23).
The set-up with the new cameras is presented Figure 6.23. This new set-up used another
code to take the picture. The same code than before is used to reconstruct the surface of the
characterized paper thanks to stereo-vision
6.2.2 Characteristics of finished products

6.2.2.1 A wrinkle free coated paper
Thanks to the coating process, it has been possible to produce some coated, wrinkle-free GHDS
35 papers. The characteristics of these papers are presented in Table 6.2. Firstly, there is a
clear increase in grammage for the coated papers. This increase is significantly greater for the
"GHDS 35 coated 1" reference because a Mayer .012 bar was used, unlike the "GHDS 35 coated
2" reference, which was coated with a Mayer .002 bar (less coating color). This impact of the
coating color is also visible on the thickness, which increased following the coatings, but to a
greater extent for the reference "GHDS 35 coated 1".
Regarding the mechanical properties, the Young’s moduli according to MD increased with
(b) Right picture of the DIC system. A

(a) Left picture of the DIC system. wrinkle is present.
Figure 6.22: GHDS 40 paper with wrinkles passing in front of the DIC system and a speckles
spotlight (red laser). The wrinkle is easily visible on the right photo but much less on the left
one.
Figure 6.23: The new industrial cameras (in red) with the lens, installed on the D.I.C. system
mounted on the coating pilot (model cameras: AV ALVIUM 1800U-1620M-CH-C - High
resolution monochrome matrix camera).
Characteristics GHDS 35 GHDS 35 coated 1 GHDS 35 coated 2

G (g · m−2 ) 36.2 ±0.2 37.9 ±0.3 37.1 ±0.2
tf (µm) 32 ±1 37 ±1 35 ±1
ECD (GPa) 4.15 ±1.17 4.94 ±0.44 5.01 ±0.28
EMD (GPa) 8.92 ±0.88 10.24 ±0.17 10.62 ±0.58
Strain at break (MD) (%) 1.53 ±0.38 1.49 ±0.21 1.63 ±0.28
Strain at break (CD)(%) 5.52 ±0.22 4.18 ±1.28 4.51 ±0.97
Stress at break (MD) (MPa) 76.23 ±14.85 85.71 ±7.13 85.82 ±6.36
Stress at break (CD) (MPa) 49.89 ±1.79 44.07 ±3.27 44.38 ±1.50
Table 6.2: Characteristics of GHDS 35 uncoated, coated with .006 Mayer bar and 15%wt
coating color (GHDS coated 1) and coated with .012 Mayer bar and 15%wt coating color by
weight (GHDS coated 2) Both colors contained 1% particles.
the coating. The Young’s moduli following CD also seem to have increased but the standard
deviation of ECD of GHDS 35 does not allow for a statistical difference. The strains at break
following CD or MD cannot be considered identical due to the overlapping standard deviations.
Finally, the stresses at break can be considered identical according to MD but it seems that the
coating reduces the stresses at break according to CD after coating.
The coated papers presented were made with a starch gel containing uncooked starch grains.
As part of the Roll and Fun project preceding this thesis, a specification was defined with the
following constraints:
→ The height of the starch grains outside the layer must be between 5 and 10 µm.
→ The density of starch grains on the surface of the deposited layer must be close to 3.2 107
particles · m−2 .
→ The surface condition of the paper must be homogeneous.
Point number 1 of the specification is one of the goals of this thesis. Point number 2 has been
studied briefly here, as shown in Figure 6.24, representing a coated paper with starch grains on
its surface.
The number of grains was determined using thresholding and particle detection with the
ImageJ software. Thus, the surface studied here is 1.70 mm2 and 637 particles were detected.
We therefore obtain a particle density of 375 particles · mm−2 , or 3.75 · 108 · particles · m−2 .
The result obtained is above the value required by the specifications. We are therefore able to
meet this criterion with our coating technique.
6.2.2.2 Extension to very lightweight paper grades
As mentioned earlier in this document, an opportunity arose during this project to obtain a
roll of CPRAW and a roll of CPBS. These papers could only be roughly characterised, and the
studies carried out for the most part on GHDS 35 could not be completed. However, it was
possible to use these two papers during a test on the coating pilot. Phenomena already observed
on GHDS 35 and GHDS 40 were observed but in a more pronounced way.
Figure 6.24: Starch particles on a GHDS 40 paper after coating with a .002 Mayer rod (picture
obtained using SEM).
Figure 6.25a shows the CBPS paper between the two hot air dryers. There are large out-of-
plane displacements: the paper undergoes what is called "flutting" in the printing industry. This
term has been defined in the literature review. For fluting to occur there must be a combination
of web tension, heat and hygroexpansion or water drying. In our case, it can be due to drying
heterogeneity on the web as shown in the temperature profiles Figure 6.6. Figure 6.25b shows
wrinkles forming on the roll before the first hot air dryer in the pilot. The wrinkles formed are
parallel to the direction of the paper flow (the roll not being misaligned). The buckling tension
phenomena could therefore be the cause of these folds. It should also be noted that the roller
in the photo has a strong tendency to heat up. Consequently, this can cause thermal shock and
promote shrinkage when the paper dries. Since wrinkles are a mechanical instability, this kind
of thermal disturbance can promote the wrinkling phenomenon or even accentuate it.
In Figure 6.26, it shows the CPRAW paper in use on the pilot. In Figure 6.26a, the paper
is fluttering and 8 bands of water are visible, making the paper more translucent. No reason for
these bands was found. Figure 6.25b shows numerous wrinkles forming on a slightly misaligned
roll (the amount of misalignment is not known).
In view of the large number of wrinkles, thresholding with ImageJ software was carried out
in order to be able to discern them as well as possible. The results are presented in Figure 6.27.
Each yellow band is linked to a peak where the gray value locally tends towards black. In this
way, 18 wrinkles could be identified, without being evenly spaced. These wrinkles appear to be
shear wrinkling, but in view of the number of folds, tension buckling phenomena may also be
present and thus amplify the wrinkling phenomenon.
(a) Fluting of CBPS paper between the two hot

air dryers. (b) Wrinkling of CPBS paper after the IR dryer.
Figure 6.25: Fluting and wrinkling on CPBS paper during coating with water.
(a) Fluting of CPRAW paper after the IR dryer. (b) Shear wrinkling of CPRAW paper after the
Bands of water are visible and seem regular IR dryer.
Figure 6.26: Fluting and wrinkling on CPRAW paper during coating with water.
Figure 6.27: Counting the number of wrinkles using image analysis and tresholding.
6.3 Conclusion
In this chapter, different points concerning the process have been discussed. Firstly, the temper-
ature of the sheet and its measurement through the use of temperature sensors. It was shown
that the temperature sensors should be as perpendicular as possible to the surface of the paper,
with an emissivity εs = 0.95. The distance between the sheet and the sensor depends on what is
being measured. The distance between the sheet and the sensor depends on whether one wishes
to measure a global temperature (distant sensor) or a local temperature (very close sensor). It
was also shown that the temperature profil of the paper at the exit of the IR dryer shows that the
sheet edges are cooler than the centre of the web. This heterogeneity during the drying process
can be the cause of a heterogeneity of the water content on the web and consequently favour
the wrinkling phenomena. It has been shown that, in the context of this study, temperature
monitoring provides access to production events and better monitoring of the latter.
In a second step, the optimisation of the process was addressed. Through the creation of a
control chart using the models developed in this project and those found in the literature. This
map allows to link web speed, web tension and misalignment angle. It can be further improved
by taking into account buckling tension phenomena and the limit stress at break of the papers,
especially when they are wet. These phenomena have also been observed on CPRAW and CBPS
papers, which seem to be much more sensitive to moisture. Finally, the GHDS 35 and GHDS 40
papers were characterised. It was even found that the specifications of the Roll and Fun project
could be met to some extent. This process chapter also opens many doors and possibilities. For
the time being, only one web tension measurement is on the pilot, but others could be installed.
They could also be installed to a control system allowing to vary it and thus avoid wrinkles.
The brake should also be changed : a torque-controlled unwind motor should be installed for
better web tension control and connected to the web tension measuring roller. Finally, a study
on the link between the Mayer bar and its paper coating color supply could be carried out.
Empirically, it has been observed that a .002 or .006 mayer bar would limit the appearance of
wrinkles because the water supply is lower. The sensor for this system could be the D.I.C system
developed during this project. Finally, in order to best meet the specifications of the Roll and
Fun project, a recirculation of the coating color could be implemented, with a system as shown
in Figure 6.28.
6.3. CONCLUSION 173
Figure 6.28: Schematic diagram for the recirculation of the coating color on the coating pilot
using an Arduino controller and two volume pumps.
Chapter 7
General conclusion
In the context of this thesis, various points have been put forward. We will go through them
in the chronological order of this document. First of all, we were interested in the penetration of
water into the paper. Two main aspects were studied: the properties of the coating sauce and the
wettability and absorption of water by the paper. In the study of the coating sauce it was shown
that the cooking of the starch is an important point in the process, as this impacts on the rheology
of the sauce. It was shown that there starch sauces has a rheofluidic behaviour (OdW law) and
that its behaviour also depends on the temperature at which it is used (Eyring law), which is in
line with the literature. The sedimentation of starch particles was studied in water and quickly
observed in starch gel. The particles have a significant speed when considering the dimensions
of the sauce storage tank on the pilot: it was estimated that it took less than 12 minutes for
the heaviest particles to settle in the storage tank. For the coating sauce, the models calculated
a sedimentation rate 10,000 times lower than in water. However, experimentally, sedimentation
and aggregation of the starch particles in the starch gel was observed. A recirculation of the
coating sauce should be installed to avoid this kind of sedimentation phenomenon.
The wettability and water absorption of GHDS 35 was also discussed. We started from the
Lucas-Washburn law, for which we had to retrieve the different parameters allowing us to use
it. An interesting point is that the SFEs of the paper showed larger dispersive components
than the polar components on both sides. This possibly means that the fibres were treated with
wet strength, thus explaining the high contact angles between water and paper on all sides.
The study of water absorption by GHDS 35 paper was carried out using three methods: the
COBB method, the DPM method, and the Paper discs stacking method. Over long times, the
Cobb method showed a first order water absorption behaviour of the paper. The paper disc
stacking method is a new method proposed to follow water absorption in paper at short times.
Three models were tested on the experimental values: LW, first order and power law. The
first order model seems to be the one that best models the observed behaviour. However, it is
necessary to take into account the very large standard deviation of the experimental data. Then
comes the need to optimise the protocol and the measurement apparatus. Finally, the DPA
technique has shown a particular behaviour of the coated papers looking at an AKD treated
paper. This confirms the hypothesis put forward following the determination of the SFE (Surface
175
176 CHAPTER 7. GENERAL CONCLUSION
Free Energy).
The second experimental part of this work was concerned with the properties and mechanical
behaviour of wrinkled papers. Based on the literature on the occurrence of wrinkles in web
materials, we investigated the impact of a roll misalignment angle and a critical tension at which
wrinkles can occur. During in-situ tests, the tension buckling was observed in the following CD
and MD. Wrinkles were also observed on the coating pilot by varying the web tension and the
misalignment angle of a roll. These observations were made on 4 different light weight papers
and the conclusions were the same: wrinkles appeared. The ex-situ observations were made
on samples of wrinkled papers when using the coating pilot. Using the Alicona device, it was
possible to confirm the asymmetric shape of the wrinkles, as well as a pinching phenomenon
observed along the wrinkles. The shape of the wrinkles could be modelled by a 4-parameter
Weibull law. Furthermore, it was found that the wrinkles do not follow the MD: they are
oriented due to the misalignment of the rollers. Microscopic observations using SEM and X-Ray
tomography have allowed us to highlight the pinching phenomenon. The latter seems to be
responsible for a decohesion of the fibrous network as well as a delamination of the industrially
deposited layer on the GHDS 35.
Following these observations, modelling was carried out using models from the literature.
The mechanical properties of the papers had to be determined. Two challenges were faced:
the characterisation of wet paper (saturated in water) and the determination of Poisson’s ratio.
An elegant method was used to access the Poisson’s ratio. Regarding the behaviour of the
water-saturated GHDS 35 paper, we were able to quantify its drying in a conditioned room by
a decreasing first order law. Following this, it was possible to model the evolution of Young’s
moduli as a function of the quantity of water absorbed per square meter of paper. The law used
is inspired by Cross’ law, and is a key element of this project. The measurement method still
needs some improvement as a hysteresis in the water absorption of the paper may exist and may
impact the evolution of the Young’s moduli. Nevertheless, the observed behaviour is close to
that of the literature: this is reassuring for the possible continuation of this work. The WSD
allowed to highlight the 2nd order behaviour according to the CD and a first order behaviour
according to the MD during this test. Thanks to the results obtained, a new parameter was
identified: the tangent modulus obtained with WSD measurement. This parameter makes it
possible to quantify the evolution of the mechanical properties of the papers as a function of the
WSD test time, and consequently as a function of the quantity of water during an absorption
phase. To improve the experiment, a measurement of local water content of paper, on the width
should be installed.
All this information was used to feed the Hashimoto model for the appearance of wrinkles on
the coating pilot. Our experimental values were compared with the model predictions. It was
surprising to find that our values, especially the misalignment angle, are very large compared
to the literature. Although not inconsistent, the slow speed used during the experiment could
explain this difference. To a lesser extent, the discretization of the number of wrinkle in the
Hashimoto model has been studied. We were able to show that making the variable representing
the number of wrinkles integer led to not respecting the bounding equation allowing to define i
177
Figure 7.1: Evolution of paper quality. N°0 is a sample prepared during the Roll and Fun
project, before this PhD thesis; N°1 is one of the first sample coated during this project; N°2
were made in March 2022 (the one below the arrow is nearly "perfect" but has no uncooked
starch particles. Finally N°3 is the last sample coated in May 2022.
and to over or underestimate the misalignment angle from which wrinkles could appear. Nev-
ertheless, the very notion of a non-natural number of wrinkles is difficult to grasp. However,
this could be related to equilibrium transition zones. Finally, the analysis of the tensile tests
and their comparison with the Zhu model validated our values. This model could also be used
to predict the tension buckling that can occur on the pilot when the rolls are well aligned, the
paper is wet and wrinkles are formed.
The last experimental chapter deals with the study of the coating process on the pilot itself.
Two points were addressed: temperature measurement during the process and process optimisa-
tion. Temperature measurement during the process was studied using IR sensors. It was shown
that the emissivity set on the sensors should be close to 0.95 and that the sensors should be as
perpendicular as possible to the web. The distance between the sensor and the web is dependent
on the value sought: local or global. This study also highlighted the non-homogeneity of the
temperature isotherms at the outlet of the IR dryer on the paper. This phenomenon could also
justify the fluting of the paper, due to heterogeneous drying, although no roll misalignment was
observed. The sensors made it possible to follow the coating process during different runs and to
illustrate the different events taking place during the process such as web breakage, equipment
change or coating at different speeds. The last aspect of this chapter was the optimisation of the
process by linking web tension, misalignment angle and web speed. For this, all the tools and
models developed earlier have been used. Thus, we linked the web speed to the water absorption,
which in turn was linked to the Young moduli by the Cross type law. This model was itself
integrated with the laws proposed by Hashimoto, by adding a variation of Poisson coefficients
thanks to the empirical model of Baum. This allowed the generation of process control and
driving maps that allowed the determination of the misalignment angle not to be exceeded as a
function of the pilot speed and the web tension. However, these charts do not take into account
the phenomena of buckling tension. For this, it is necessary to use the kc coefficient proposed
by Zhu as a first approach. This thesis allowed to understand the phenomena related to the
178 CHAPTER 7. GENERAL CONCLUSION
appearance of wrinkles. This work offers a new perspective to the study of this kind of phe-
nomenon. It would be interesting to acquire more data concerning the typical Cross like model
of water absorption and Young’s modulus of the paper. Being able to acquire more information
with the DIC system developed during this thesis when the process is running would also be a
plus for understanding the in-situ phenomena and regulating the process through measurement
and control of web tension. A measurement of the moisture profile of the paper on the web
would be a plus as well as a measure on the web tension along the web width. To conclude,
we will answer this last question: was it possible to coat a paper, without wrinkles, with a
starch gel containing starch granules? The answer is yes (Figure 7.1, because the process is now
controlled through the creation of control maps and the process is waiting to be improved for
this particular application.
Bibliography
[1] Timothy J Walker. The taxonomy of wrinkles. In Tenth International Conference on Web
Handling (IWEB, page 15, Stillwater, OK., June 2009. Oklahoma State University.
[2] FCBA. Mémento FCBA - 2020, https://www.fcba.fr/ressources/memento-2020/.
[3] Erik Orsenna. Sur la route du papier. Petit précis de mondialisation. Stock, Paris, 2012.
Publication Title: Sur la route du papier.
[4] Joël Boulier and Laurent Simon. Atlas des forêts dans le monde. Atlas/Monde. Autement,
Paris, 2022.
[5] Tatjana Stevanovic and Perrin Dominique. Chimie du bois. Presses polytechniques et
universitaires romandes, Lausanne, 2009.
[6] E. Renault. Délignification du bois de châtaignier par une approche de chimie verte.
Mise en œuvre et impacts sur la structure et le potentiel anti-radicalaire des phyto-
polysaccharides extraits. Phd dissertation, Université de Limoges, December 2014.
[7] Marie-Christine Trouy and Pascal Triboulot. Matériau bois structure et caractéris-
tiques. Techniques de l’ingénieur Les superstructures du bâtiment, base documentaire :
TIP253WEB.(ref. article : c925), 2019. fre.
[8] Eero Sjöström. Wood Chemistry: fundamentals and applications. Academic Pr, San Diego,
2. ed. edition, 1997.
[9] Harry Phillips Brown, A. J. Panshin, and C. C. Forsaith. Textbook of wood technology.
American forestry series. McGraw-Hill, New York, 1st edition edition, 1949. Publication
Title: Textbook of wood technology.
[10] D. H. Page. A method for determining the fibrillar angle in wood tracheids. Journal of
Microscopy, 90(2):137–143, October 1969.
[11] Mohamed Naceur Belgacem and Antonio Pizzi. Lignocellulosic Fibers and Wood Handbook:
Renewable Materials for Today’s Environment. Wiley, 2016.
179
180 BIBLIOGRAPHY
[12] Dietrich Fengel and Gerd Wegener, editors. Wood: Chemistry, ultrastructure, reactions.
De Gruyter, 2011.
[13] C. Monot. Contribution à l’étude des complexes lignine-hydrates de carbone (LCC) dans
le bois: étude de l’impact des différentes étapes d’un procédé de bioraffinerie sans soufre
sur les LCC. Phd dissertation, Université Grenoble Alpes, December 2015.
[14] Roger M. Rowell, editor. Handbook of wood chemistry and wood composites. Taylor &
Francis, Boca Raton, Florida, 2005.
[15] Monica Ek. Pulping chemistry and technology. Number 2 in Pulp and paper chemistry
and technology. De Gruyter, Berlin, 2009.
[16] Herbert Holik. Handbook of paper and board. Wiley-VCH, Weinheim, 2006.
[17] Valentin I. Popa, editor. Pulp production and processing: from papermaking to high-tech
products. Smithers Rapra, Shawbury, 2013.
[18] Herbert Sixta. Handbook of pulp, volume 1. Wiley-VCH, Weinheim, 2006.
[19] C.J. Biermann. Handbook of Pulping and Papermaking. Elsevier Science, 1996.
[20] Jean-Claude Roux. Désintégration et raffinage de la pâte à papier. Techniques de

l’ingénieur Bois et papiers, base documentaire : TIP590WEB.(ref. article : j6903), 2008.
fre.
[21] André Lemaître. Procédés papetiers préparation des pâtes. Techniques de l’ingénieur Bois
et papiers, base documentaire : TIP590WEB.(ref. article : j6901), 2003. fre.
[22] E. Tourette. Étude mécanique et physico-chimique du contact rouleau - papier lors du

calandrage. Phd dissertation, École Nationale Supérieure des Mines de Paris, December
2007.
[23] Ruben Miranda, Ana Balea, Emilia Sanchez de la Blanca, Isabel Carrillo, and Angeles
Blanco. Identification of Recalcitrant Stickies and Their Sources in Newsprint Production.
Industrial & Engineering Chemistry Research, 47(16):6239–6250, August 2008.
[24] COPACEL. COPACEL, https://www.copacel.fr/, 2022-03-06.
[25] Martin A. Hubbe, Douglas J. Gardner, and Wei Shen. Contact Angles and Wettability of
Cellulosic Surfaces: A Review of Proposed Mechanisms and Test Strategies. BioResources,
10(4):8657–8749, September 2015.
[26] Niskanen Kaarlo. Papermaking science and technology / book ed. Kaarlo Niskanen. 16.
Paper physics. Papermaking science and technology. Fapet Oy Published in cooperation
with the Finnish Paper Engineers’ Association and TAPPI, Helsinki (Finlande), cop. 1998.
[27] M. Bouchon. Etude de l’influence des phénomènes de mouillage sur la pénétration de l’eau
dans le papier aux temps courts. PhD dissertation, Grenoble INP, May 2004.
BIBLIOGRAPHY 181
[28] T. Karppinen. Characterizing physical properties of wetting paper by air coupled ultrasound
and light. Phd dissertation, University of Helsinki, September 2008.
[29] M.J. Moura, P.J. Ferreira, and M.M. Figueiredo. Mercury intrusion porosimetry in pulp
and paper technology. Powder Technology, 160(2):61–66, December 2005.
[30] S. Rolland du Roscoat. Contribution à la quantification 3D de réseaux fibreux par microto-

mographie au rayonnement synchrotron : applications aux papiers. PhD Thesis, Université
de Grenoble, 2007.
[31] E. A. Barclay, J. E. Donahue, D. A. Hughes, R. H. Rogers, J. R. Simpson, and R. N.

Thompson. Preparation of paper coating colors. TAPPI Press, New York, 1954.
[32] Robert H. Mosher. The technology of coated and processed papers. Chemical Pub. Co,
New York, 1952.
[33] Lehtinen Esa. Papermaking science and technology / book ed. Esa Lehtinen ; book reviewer
Charles P. Klass. 11. Pigment coating and surface sizing of paper. Papermaking science
and technology. Fapet Oy Published in cooperation with the Finnish Paper Engineers’
Association and TAPPI, Helsinki, cop. 2000.
[34] Jan C. Walter. The coating processes. TAPPI Press, Atlanta, 1993.
[35] Pierre Valette and Christian De Choudens. Le bois, la pâte, le papier. CTP, Grenoble,
second edition, 1989.
[36] Naceur Belgacem. De la fibre à l’imprimé / Grenoble INP - Pagora. Tome III. Procédés de
finition et de transformation des papiers et cartons. Agefpi, Saint-Martin-d’Hères, France,
2016. Backup Publisher: École internationale du papier, de la communication imprimée
et des biomatériaux Saint-Martin-d’Hères, Isère Publication Title: De la fibre à l’imprimé
. Tome III . Procédés de finition et de transformation des papiers et cartons.
[37] S. Cena. Etude d’un système d’enduction par presse à transfert de film. PhD dissertation,
Grenoble INP, December 1997.
[38] S. Moreau. Structure des couches pigmentaires : étude de leur séchage. PhD dissertation,
Grenoble INP, 1994.
[39] Alain Buléon, Gabrielle Véronèse, and Jean-Luc Putaux. Self-Association and Crystalliza-
tion of Amylose. Australian Journal of Chemistry, 60(10):706, 2007.
[40] D. Le Corre. Starch nanocrystals: preparation and application to bio-based flexible pack-
aging. Phd dissertation, Université de Grenoble, October 2011.
[41] Costas G. Biliaderis. Structural Transitions and Related Physical Properties of Starch. In
Starch, pages 293–372. Elsevier, 2009.
182 BIBLIOGRAPHY
[42] Bernard Boursier. Amidons natifs et amidons modifiés alimentaires. Techniques de

l’ingénieur Agroalimentaire, base documentaire : TIP700WEB.(ref. article : f4690), 2005.
fre.
[43] E. Schwach. Etude de systèmes multiphases biodégradables à base d’amidon de blé plas-
tifié : relations structure-propriétés, approches de la compatibilisation. PhD dissertation,
Université de Reims - Champagne Ardenne, 2004.
[44] James N. BeMiller and Kerry C. Huber. Starch. In Wiley-VCH Verlag GmbH & Co. KGaA,
editor, Ullmann’s Encyclopedia of Industrial Chemistry, page a25_001.pub4. Wiley-VCH
Verlag GmbH & Co. KGaA, Weinheim, Germany, January 2011.
[45] G Bacquet. La physico-chimie des bains de couchage. Revue A.T.I.P, 50(4):8, July 1996.
[46] J.F. Richardson and W.N. Zaki. The sedimentation of a suspension of uniform spheres
under conditions of viscous flow. Chemical Engineering Science, 3(2):65–73, April 1954.
[47] B. Gueslin. Etude de la sédimentation de particules non-browniennes dans un fluide

thixotrope à seuil: la laponite. Phd dissertation, Université de Paris 6, November 2006.
[48] James Brander and Ian Thorn, editors. Surface Application of Paper Chemicals. Springer
Netherlands, Dordrecht, 1997.
[49] J. Jäder. Consolidation and rheology at high solids content. PhD dissertation, Karlstad
University, 2004.
[50] Reinhard Sangl, Werner Auhorn, Werner Kogler, and Martin Tietz. Surface Sizing and
Coating. In Herbert Holik, editor, Handbook of Paper and Board, pages 745–784. Wiley,
1 edition, April 2013.
[51] Stephan F. Kistler, editor. Liquid film coating: scientific principles and their technological
implications. Springer, Dordrecht, 1. ed edition, 1997. OCLC: 935042650.
[52] Savolainen Antti. Papermaking science and technology. Fapet Oy Published in cooperation
with the Finnish Paper Engineers’ Association and TAPPI, Helsinki (Finlande), cop. 1998.
[53] Jan C. Walter. Chapter 4: Drying. In The coating processes, page 260. Tappi, Atlanta,
1993.
[54] David R. Roisum. The mechanics of web handling. TAPPI PRESS, Atlanta, 1998.
[55] David R. Roisum. The mechanics of rollers. TAPPI Press, Atlanta, 1996.
[56] Roll-2-Roll Technologies LLC. Roll-2-Roll, https://r2r.tech/, 2022-08-09.
[57] William B. Gilbert. Optimizing web tension control for high productivity: Part 1. Official
Publication of the Association of International Metallizers, Coaters and Laminators, (2019
- 2nd quarter):66–71, 2019.
BIBLIOGRAPHY 183
[58] PFF-online. Web tension control fundamentals, https://www.pffc-

online.com/ pffcon5/news/15515-web-tension-control-fundamentals, 2022-08-09.
[59] J J Shelton. Simplified model for lateral behavior of short web spans. In Sixth International
Conference on Web Handling (IWEB), page 16, Stillwater, OK., June 2001. Oklahoma
State University.
[60] David R. Roisum. The mechanics of web spreading Part 1. TAPPI Journal, 76(10):8,
October 1993.
[61] D. K. Owens and R. C. Wendt. Estimation of the surface free energy of polymers. Journal
of Applied Polymer Science, 13(8):1741–1747, August 1969.
[62] D. H. Kaelble. Dispersion-Polar Surface Tension Properties of Organic Solids. The Journal
of Adhesion, 2(2):66–81, April 1970.
[63] Naceur Belgacem and Alain Dufresne. Phénomène d’adhésion et le collage. In De la fibre à
l’imprimé - tome 1 : Sciences de l’ingénieur pour les industries des fibres, des biomatéraux
et de la communication imprimée, volume 1, page 28. Agefpi, Saint Martin d’Hères, 2016.
[64] Robert N. Wenzel. Resistance of Solid Surfaces to Wetting by Water. Industrial & Engi-
neering Chemistry, 28(8):988–994, August 1936.
[65] David Quéré. Wetting and Roughness. Annual Review of Materials Research, 38(1):71–99,
August 2008.
[66] Nuria Gómez, Ester Quintana, and Juan Carlos Villar. Effect of Paper Surface Properties
on Coated Paper Wettability with Different Fountain Solutions. BioResources, 9(3):4226–
4241, May 2014.
[67] Hedieh Modaressi and Gil Garnier. Mechanism of Wetting and Absorption of Water
Droplets on Sized Paper: Effects of Chemical and Physical Heterogeneity. Langmuir,
18(3):642–649, February 2002.
[68] A. B. D. Cassie and S. Baxter. Wettability of porous surfaces. Transactions of the Faraday
Society, 40:546, 1944.
[69] A. Pons. Simulation numérique de la montée capillaire en espace confiné, en vue de

l’application à des procédés d’élaboration de matériaux composites par imprégnation non-
réactive ou réactive. PhD dissertation, Univesité de Bordeaux, October 2017.
[70] D. Dallel. Etude expérimentale modèle de l’imbibition capillaire de substrats poreux à

volume et structure de pores ajustables. Phd dissertation, Université de Haute-Alsace,
December 2012.
[71] P. Salminen. Studies of Water Transport in Paper during Short Contact Times. Phd
dissertation, Abo Akademi, October 1988.
184 BIBLIOGRAPHY
[72] N. Fries. Capillary transport processes in porous materials - experiment and model. PhD
dissertation, Universität Bremen, 2010.
[73] C. H. Bosanquet. On the flow of liquids into capillary tubes. The London, Edinburgh, and
Dublin Philosophical Magazine and Journal of Science, 45(267):525–531, 1923. Publisher:
Taylor & Francis _eprint: https://doi.org/10.1080/14786442308634144.
[74] D Quéré. Inertial capillarity. Europhysics Letters (EPL), 39(5):533–538, September 1997.
[75] Carina Waldner and Ulrich Hirn. Ultrasonic Liquid Penetration Measurement in Thin
Sheets—Physical Mechanisms and Interpretation. Materials, 13(12):2754, June 2020.
[76] J. Hyväluoma, P. Raiskinmäki, A. Jäsberg, A. Koponen, M. Kataja, and J. Timonen.

Simulation of liquid penetration in paper. Physical Review E, 73(3):036705, March 2006.
[77] R. J. Roberts. Liquid penetration into paper. PhD dissertation, Australian National
University, September 2004.
[78] Jianchao Cai, Tingxu Jin, Jisheng Kou, Shuangmei Zou, Junfeng Xiao, and Qingbang
Meng. Lucas–Washburn Equation-Based Modeling of Capillary-Driven Flow in Porous
Systems. Langmuir, 37(5):1623–1636, February 2021.
[79] Sarah Krainer and Ulrich Hirn. Short timescale wetting and penetration on porous
sheets measured with ultrasound, direct absorption and contact angle. RSC Advances,
8(23):12861–12869, 2018.
[80] T. Karppinen, I. Kassamakov, J. Aaltonen, H. Pajari, and E. Hæggström. Measuring

liquid penetration in the thickness direction of paper. The European Physical Journal
Applied Physics, 32(1):65–71, October 2005.
[81] Narpinder Singh, Jaspreet Singh, Lovedeep Kaur, Navdeep Singh Sodhi, and Balmeet
Singh Gill. Morphological, thermal and rheological properties of starches from different
botanical sources. Food Chemistry, 81(2):219–231, May 2003.
[82] Narpinder Singh, Naoto Isono, Sathaporn Srichuwong, Takahiro Noda, and Katsuyoshi
Nishinari. Structural, thermal and viscoelastic properties of potato starches. Food Hydro-
colloids, 22(6):979–988, August 2008.
[83] Dominique Dupuis and Alain Ponton. Mesure de la viscosité principes généraux. Tech-
niques de l’ingénieur Mesures physiques, base documentaire : TIP672WEB.(ref. article :
r2350), 2021. fre.
[84] Henry Eyring. Viscosity, Plasticity, and Diffusion as Examples of Absolute Reaction Rates.
The Journal of Chemical Physics, 4(4):283–291, April 1936.
[85] Malcolm M. Cross. Rheology of non-Newtonian fluids: A new flow equation for pseudo-
plastic systems. Journal of Colloid Science, 20(5):417–437, June 1965.
BIBLIOGRAPHY 185
[86] Luiz Claudio F. Andrade, Jamilson A. Petronílio, Calos Edilson de A. Maneschy, and
Daniel Onofre de A. Cruz. The carreau-yasuda fluids: a skin friction equation for turbu-
lent flow in pipes and kolmogorov dissipative scales. Journal of the Brazilian Society of
Mechanical Sciences and Engineering, 29(2):162–167, 2007.
[87] Mohd I Nurul, B. M. N. Mohd Azemi, and D. M. A. Manan. Rheological behaviour of

sago (Metroxylon sagu) starch paste. Food Chemistry, 64(4):501–505, 1999.
[88] Yongfeng Ai and Jay-lin Jane. Gelatinization and rheological properties of starch. Starch
- Stärke, 67(3-4):213–224, March 2015.
[89] T. Emilia Abraham. Stabilization of Paste Viscosity of Cassava Starch by Heat Moisture
Treatment. Starch - Stärke, 45(4):131–135, 1993.
[90] J.-L. Doublier. Rheological Studies on Starch — Flow Behaviour of Wheat Starch Pastes.
Starch - Stärke, 33(12):415–420, 1981.
[91] R.G.M. van der Sman, Job Ubbink, Marina Dupas-Langlet, Magdalena Kristiawan, and
Isabel Siemons. Scaling relations in rheology of concentrated starches and maltodextrins.
Food Hydrocolloids, 124:107306, March 2022.
[92] J.-Y Thebaudin, A.-C Lefebvre, and J.-L Doublier. Rheology of Starch Pastes from
Starches of Different Origins: Applications to Starch-based Sauces. LWT - Food Science
and Technology, 31(4):354–360, April 1998.
[93] Douglas W Bousfield. Rheology Issues in the paper Industry. The British Society of
Rheology, pages 47–70, 2008.
[94] Dominique Dupuis and Alain Ponton. Mesure de la viscosité viscosimètres et rhéomètres.
Techniques de l’ingénieur Mesures physiques, base documentaire : TIP672WEB.(ref. arti-
cle : r2351), 2021. fre.
[95] El Kissi, S Nigen, and Frédéric Pignon. Glissement en rhéologie. Rhéologie, 10:13–39,
November 2006.
[96] Ting Wang, Yifan Yang, and Fan Xu. Mechanics of tension-induced film wrinkling and
restabilization: a review. Proceedings of the Royal Society A: Mathematical, Physical and
Engineering Sciences, 478(2263):20220149, July 2022.
[97] Juner Zhu, Xiaowei Zhang, and Tomasz Wierzbicki. Stretch-induced wrinkling of highly
orthotropic thin films. International Journal of Solids and Structures, 139-140:238–249,
May 2018.
[98] N Friedl, F.G Rammerstorfer, and F.D Fischer. Buckling of stretched strips. Computers
& Structures, 78(1-3):185–190, November 2000.
[99] F. G. Rammerstorfer. Buckling of elastic structures under tensile loads. Acta Mechanica,
229(2):881–900, February 2018.
186 BIBLIOGRAPHY
[100] E. Cerda and L. Mahadevan. Geometry and Physics of Wrinkling. Physical Review Letters,
90(7):4, February 2003.
[101] N. Jacques, A. Elias, M. Potier-Ferry, and H. Zahrouni. Buckling and wrinkling during
strip conveying in processing lines. Journal of Materials Processing Technology, 190(1-
3):33–40, July 2007.
[102] Nicolas Jacques and Michel Potier-Ferry. On mode localisation in tensile plate buckling.
Comptes Rendus Mécanique, 333(11):804–809, November 2005.
[103] N. G. M. Jacques. Modélisation et étude du plissement des tôles lors de leur transport en
continu dans les usines sidérurgiques. Phd dissertation, Université Paul Verlaine, Metz,
October 2004.
[104] Eric Puntel, Luca Deseri, and Eliot Fried. Wrinkling of a Stretched Thin Sheet. Journal
of Elasticity, 105(1-2):137–170, November 2011.
[105] Philippe G. Ciarlet. A justification of the von Kármán equations. Archive for Rational
Mechanics and Analysis, 73(4):349–389, December 1980.
[106] Wei Tang, Jiankui Chen, and Zhouping Yin. Elastic buckling analysis of webs transported
through rollers with misalignment. Thin-Walled Structures, 121:1–7, December 2017.
[107] Chung Hwan Kim. Stress distribution and wrinkling of webs under non-uniform and
asymmetric edge loading. International Journal of Solids and Structures, 46(3-4):851–859,
February 2009.
[108] C. C. Lin and C. D. Mote. The Wrinkling of Rectangular Webs Under Nonlinearly Dis-
tributed Edge Loading. Journal of Applied Mechanics, 63(3):655–659, September 1996.
[109] A. Kulachenko. Mechanics of Paper Webs in Printing Press Applications. PhD disserta-
tion, Mid Sweden University, 2006.
[110] Artem Kulachenko, Per Gradin, and Tetsu Uesaka. Tension wrinkling and fluting in
heatset web offset printing process. Post-buckling analyses. In Advances in Paper Science
and Technology, page 29, Cambridge, UK, 2005.
[111] Artem Kulachenko, Per Gradin, and Tetsu Uesaka. Basic mechanisms of fluting formation
and retention in paper. Mechanics of Materials, 39(7):643–663, July 2007.
[112] Cecilia Land. Laboratory method for the study of moisture-induced waviness in paper.
Licentiate dissertation, Karlstads universitet, 2004.
[113] Tetsu Uesaka. Variability, non-uniformity, and complexity: from product to process. In
Advances in Pulp and Paper Research,, page 38, Cambridge, UK, 2013.
[114] Artem Kulachenko. Chapter 9: Moisture-induced deformations. In Mechanics of Paper

Products, pages 163–180. De Gruyter, Berlin, Boston, 2011.
BIBLIOGRAPHY 187
[115] Douglas Coffin. A Buckling Analysis Corresponding to the Fluting of Lightweight Coated
Webs. In International Paper Physics Conference, page 5, Vancouver, 2003.
[116] T. Uesaka. General formula for hygroexpansion of paper. Journal of Materials Science,
29(9):2373–2377, May 1994.
[117] Artem Kulachenko, Per Gradin, and Bo Westerlind. Analysis of Paper Web Tension
Profiles. TAGA JOURNAL, 2:72–81, 2005.
[118] Frédéric Parent, Jean Hamel, and J D McDonald. Web lateral instability caused by
nonuniform paper properties. TAPPI Journal, 21(1):12, January 2022.
[119] Tetsu Uesaka. Chapter 6: Web dynamics in paper transport systems. In Mechanics of
Paper Products, pages 93–110. De Gruyter, Berlin, Boston, 2011.
[120] Yannick Martz and Dominique Knittel. Dynamics of an elastic web in roll-to-roll sys-
tems using finite element method. In 11th World Congress on Computational Mechanics
(WCCM XI), page 12, Barcelon, July 2014.
[121] D. Knittel. Web dynamics analysis using finite element simulations. In AIMCAL Web
Coating and Handling Conference, June 2014.
[122] L. S. Gehlbach, D M Kedl, and James K. Good. Predicting shear wrinkles in web spans.
TAPPI Journal, 72(8):6, August 1989.
[123] James K Good, Joseph A Beisel, and H Yurtcu. Instability of webs: the prediction of
troughs and wrinkles. In Advances in Pulp and Paper Research, pages 517–556, Oxford,
2009.
[124] H. Hashimoto. Prediction model of paper-web wrinkling and some numerical calculation
examples with experimental verifications. Microsystem Technologies, 13(8-10):933–941,
April 2007.
[125] J. A. Beisel. Single span web buckling due to roller imperfections in web process machinery.
Phd dissertation, Oklahoma State University, December 2006.
[126] Y. Martz. Modélisation et commande de systèmes d’entraînement de bandes flexibles :

nouvelles approches à l’aide des éléments finis. Phd dissertation, Université de Strasbourg,
June 2017.
[127] J K Good, D M Kedl’, J J Shelton, and M Company. Shear wrinkling in isolated spans.
In Fourth International Conference on Web Handling (IWEB), page 19, Stillwater, Okla-
homa, June 1997. Oklahoma State University.
[128] Stephen P. Timoshenko and James M. Gere. Theory of elastic stability. McGraw-Hill book
company Inc, 2nd edition, 1963.
[129] Pierre Lecomte du Noüy. An interfacial tensiometer for universal use. The Journal of
General Physiology, 7(5):8, May 1925.
188 BIBLIOGRAPHY
[130] M.E Biancolini, C Brutti, Ernesto Mottola, and S Porziani. Numerical evaluation of buck-
ling and post-buckling behaviour of corrugated board containers. In Associazioneitaliana
per l’analisi delle sollecitazioni xxxiv convegno nazionale, page 14, Milano, September
2005. Politecnico Di Milano.
[131] G. A Baum, D. C. Brennan, and C.C Habeger. Orthotropic elastic constants of paper.
TAPPI Journal, 64(8):5, August 1981.
[132] G. A Baum and L.R Borhoeft. Estimating Poisson ratios in paper using ultrasonic tech-
nique. TAPPI Journal, 62(5):5, May 1979.
[133] J. Viguié. Comportements mécanique et hygroexpansif des matériaux ligno-cellulosiques

pour l’emballage. Phd dissertation, Université de Grenoble, May 2010.
[134] Christer Fellers and Leif A. Carlsson. Bending stiffness, with special reference to paper-
board. In Handbook of physical testing of paper, volume 1, pages 233–256. Taylor & Francis
Group„ New-York, 2nd edition edition, 2002.
[135] Operating Instructions thermoMETER CT, micro-epsilon, 2019.
[136] M. Dolz, M. J. Hernández, C. Cabeza, A. Casanovas, and J. Delegido. Analysis of stability

of food emulsions by Eyring’s theory: Influence of different biopolymers: Stability of Food
Emulsions. Journal of Applied Polymer Science, 92(4):2653–2657, May 2004.
[137] GuangZe Han, ZengKe Fang, and MingDong Chen. Modified Eyring viscosity equation
and calculation of activation energy based on the liquid quasi-lattice model. Science China
Physics, Mechanics and Astronomy, 53(10):1853–1860, October 2010.
[138] Micael Stehr, Douglas J. Gardner, and Magnus E. P. Wålinder. Dynamic Wettability of
Different Machined Wood Surfaces. The Journal of Adhesion, 76(3):185–200, August 2001.
[139] D. H. R. Sarma and R. W. Vest. Kinetics of Liquid Spreading and Penetration with Ap-
plication to RuO2-Glass Thick-Film Resistors. Journal of the American Ceramic Society,
68(5):249–253, May 1985.
[140] Belgacem Naceur. Agefpi, Saint-Martin-d’Hères, France, 2016.
[141] Ying Li, Jie Zhang, Haishan Li, Wen juan Gu, and Banggui He. Analysis of the Coating
surface Properties of Coated Paper:. In 2015 International Conference on Materials,
Environmental and Biological Engineering, Guilin, China, 2015.
[142] Joel Songok, Pekka Salminen, and Martti Toivakka. Temperature effects on dynamic water
absorption into paper. Journal of Colloid and Interface Science, 418:373–377, March 2014.
[143] Craig R. Friedrich. Geometry and characterization of wrinkles in paper. Tappi Journal,
73(8):159–163, August 1990.
BIBLIOGRAPHY 189
[144] Jean-Francis Bloch, Pierre J.J. Dumont, Sabine Rolland du Roscoat, Laurent Orgéas,
Christian Geindreau, and Nicolas Bouquet. Quelques aspects du comportement hygromé-
canique du papier. Mécanique & Industries, 10(1):43–48, January 2009.
[145] S. T. Lehti, J. A. Ketoja, and Kaarlo Niskanen. Measurement of Paper Rheology at Varied
Moisture Contents. In International Paper Physics Conference, page 4, Vancouver, 2003.
[146] Lennart Salmen and Myat Htun. A Treatise on the Elastic and Hygroexpansional Prop-
erties of Paper by a Composite Laminate Approach. Fibre science and technology, 20:14,
1984.
[147] Micro-Epsilon. Mesure thermique sans contact: fondements, 2019.
[148] Patrice Nortier. Chapitre 2: Transferts de chaleur. In De la fibre à l’imprimé, volume 1,

page 398. Agefpi, Saint Martin d’Hères, agefpi edition, 2016.
[149] Jean-Pierre Dory, Françoise Evin, and Michel Piro. Chauffage par rayonnement infrarouge,
1999. fre.
190 BIBLIOGRAPHY
List of Figures
1.1 Conformed (left) and non-conformed (right) paper. . . . . . . . . . . . . . . . . . 1

1.2 From troughs to creases on the RK coating pilot of the LGP2 laboratory. . . . . 2
1.3 Visual abstract presenting the manuscript organisation. . . . . . . . . . . . . . . 3
2.1 From living cell to wood fibre [8]. . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.2 Distribution of principal chemical constituents in softwood cell wall layers [9]. . . 8
2.3 Cellobiose (between brackets) and 2 β-D-glucopyranose units (which composed
cellobiose unit) [11]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.4 The 3 lignin precursors for plants (from left to right): p-hydroxyphenyl unit,
guaiacyl unit and syringyl unit [12]. . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.5 An overview of chemical and papermaking process: from logs to customers [19]. . 12
2.6 Example of double disks refiner. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.7 Scheme of a paper machine with size press section and calenders [23] . . . . . . . 14
2.8 Possible pore structures in paper, from left to right : surface pore, connected pore
and cavity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.9 Amylose (top) and amylopectin (bottom) structure [40]. . . . . . . . . . . . . . . 19
2.10 Maltese cross observed within potato starch granules viewed under plane-polarized
light (x120) [42]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.11 Pasting profile of a wheat starch granule suspension over the course of a heating
and cooling cycle [44]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.12 Three different way to deposit coating color on paper. . . . . . . . . . . . . . . . 25
2.13 A size press system [50]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.14 Three different way to meter coating color on paper. . . . . . . . . . . . . . . . . 26
2.15 Three different way to spread web on roll to roll system [60]. . . . . . . . . . . . 28
2.16 The four guiding technologies [56] [56]. . . . . . . . . . . . . . . . . . . . . . . . . 29
2.17 Tension sensing system [58]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.18 Surface tension appearing at the interface of gas (G) and liquid (L). . . . . . . . 32
2.19 Surface tension measuring systems. . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.20 Scanning absorptometer apparatus [79] . . . . . . . . . . . . . . . . . . . . . . . . 40
2.21 Salminen apparatus [71]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
191
192 LIST OF FIGURES
2.22 Parameters for the evaluation of DPM measurement proposed by Waldner and
Hirn [75]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
2.23 Picture of the measurement device proposed by Karppinen [80] and its schematic
presentation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.24 Viscosity profile of native and modified starch as a function of shear rate [89]. . . 45
2.25 Starch viscosity as a function of temperature [87]. . . . . . . . . . . . . . . . . . . 46
2.26 Three different rheometer techniques. . . . . . . . . . . . . . . . . . . . . . . . . . 47
2.27 Deformation at different tensile strains during tensile test of highly orthotropic
films in low stiffness direction. Each line represents a different sample width
(name W here). Each column is a different strain rate. The red rectangles are
configurations where out of plane displacement appear [97] . . . . . . . . . . . . 48
2.28 Critical buckling coefficient according to Friedl et al. versus the length/width
ratio [98]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
2.29 Example of wrinkles formation, during an uni-axial test, on a polyethylene sample
of length a = 250 mm and width b = 100 mm. The thickness tf is around 0.1
mm. On the picture, the strain is about 10 % [100] . . . . . . . . . . . . . . . . . 49
2.30 Appearance of wrinkle during sheet conveying. Here TB is the critial buckling
load and TW is the critical wrinkling load [101]. . . . . . . . . . . . . . . . . . . . 50
2.31 Highlighting of cockling and fluting appearence on paper. . . . . . . . . . . . . . 51
2.32 Buckling generation by shearing and impact of friction force . . . . . . . . . . . . 54
3.1 Blade used to cook and blend starch gel in cooking robot. . . . . . . . . . . . . . 60
3.2 Sedimentation test evolution at 2 different times : t and t + ∆t. . . . . . . . . . 61
3.3 Highlighting of the cooking evolution of starch particle by optical microscope
analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.4 Rotationnal rheological measurement with a cone-plan system. . . . . . . . . . . 63
3.5 Density measurement of a 20 %wt starch gel with a glass sphere. . . . . . . . . . 63
3.6 Use of a platinum Wilhelmy plate (19.5 mm x 0,12 mm) to measure starch gel
surface tension with 5 round-trip cycles. . . . . . . . . . . . . . . . . . . . . . . . 64
3.7 Highlighting of the phenomena that occurs during a surface tension measure: (1)
the plate is at its "zero" position; (2) the plate is in contact with the liquid, the
advancing contact angle can be recorded; (3) the plate continues to go deeper in
the liquid; (4) the plate is going back to its "zero" position, the surface tension is
measured.The receding contact angle can be recorded [25]. . . . . . . . . . . . . . 64
3.8 The resonance method. Resonance corresponds to maximum deflection δ and w
to the out of plane displacement [134]. . . . . . . . . . . . . . . . . . . . . . . . . 67
3.9 WSD (Water Stretch Dynamics analyser) test principle. . . . . . . . . . . . . . . 67
3.10 Finch’s test principle. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.11 Highlighting of the different scales that could be measured while focusing on
surface roughness. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
3.12 Evolution of a contact angle over time during a sessile drop method test. . . . . . 70
3.13 Size and organisation of a typical paper sample for DCA measurements. . . . . . 70
LIST OF FIGURES 193
3.14 Cobb test principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

3.15 Setting up the water absorption measurement by stacking paper discs on top of
each other and immersed it in water. . . . . . . . . . . . . . . . . . . . . . . . . . 72
3.16 DPM measurement principle [75]. . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.17 RK laboratory coating pilot. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.18 The coating station with the different coating tools. . . . . . . . . . . . . . . . . 74
3.19 Non-contact IR temperature measurement system. . . . . . . . . . . . . . . . . . 75
3.20 R pictures taking with Flir-i7 (Web speed: 5 m · min−1 . . . . . . . . . . . . . . . 76
3.21 DIC system installed on the tensile test (left) and on the coating pilot (right). . . 77
3.22 Example of result that could be obtain with Alicona InfinteFocus device. . . . . . 78
4.1 Evolution of starch suspension at different temperature during cooking (Optical

microscopy). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
4.2 Measured temperature by the IR sensor (°C) versus the heating temperature (°C)
and measuring height hs (mm) (standard deviation: ±1 °C). 3 trials were carried
out. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
4.3 Evolution of starch gel viscosity versus shear rate, for different starch mass frac-
tion, at different temperature. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
4.4 Evolution of viscosity of a starch gel at 15%wt,at 50 °C and γ̇ = 0.1 s−1 . The
starch dried and now stabilization of the viscosity appears. . . . . . . . . . . . . 85
4.5 Evolution of starch gel viscosity versus temperature at different shear rate for
20%wt starch gel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
4.6 Population of starch particles, expressed in volume or number, as a function of
their diameter. The cross symbols are the experimental values. The dotted lines
are used to facilitate reading. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
4.7 Starch particles of different sizes. Photo obtain by SEM. . . . . . . . . . . . . . . 89
4.8 Evolution of the height of the sedimentation front and of the sedimentation rate
over time. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
4.9 Dimension of the coating tank used on the coating pilot. . . . . . . . . . . . . . . 91
4.10 Evolution of the contact angle between different liquids and GHDS 35 paper
coated side. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
4.11 Evolution of the contact angle between different liquids and GHDS 35 paper
uncoated side. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
4.12 View of the GHDS 35 paper section at different magnifications (SEM). . . . . . . 95
4.13 Ω versus Π for the coated and uncoated side of GHDS 35, at different time. . . . 97
4.14 FTIR spectrum of the GHDS 35 paper sides and bleached cellulose (for reference). 98
4.15 Evolution of the starch gel (20%wt) surface tension over time at 40°C. A, B, C
and D are different tests made in same conditions. . . . . . . . . . . . . . . . . . 99
4.16 Differential intrusion of mercury as a of the GHDS 35 paper pores diameter. . . . 100
4.17 Absorptiveness of paper over time obtained trough the Cobb method.Water is in
contact with the coated side of GHDS 35 paper. . . . . . . . . . . . . . . . . . . 101
194 LIST OF FIGURES
4.18 Parity plot comparing the modelled and experimental values obtained with the
Cobb method. Water is in contact with the coated side of GHDS 35 paper. . . . 102
4.19 Signal intensity overs time curves obtained from DPM measurement.Coated and
uncoated refereed to the GHDS 35 paper side facing water. Grooved and flat
refereed to the sample holder surface. . . . . . . . . . . . . . . . . . . . . . . . . . 103
4.20 GHDS 35 paper water absorption versus time. 50 stacks of paper are used, the
coated side facing the water. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
√
4.21 GHDS 35 paper water absorption versus t. 50 stacks of paper are used, the
paper discs stacking method for the GHDS 35 coated side down. . . . . . . . . . 107
4.23 GHDS 35 paper water absorption versus time. 50 stacks of paper are used, the
uncoated side facing the water. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
√
4.24 GHDS 35 paper water absorption versus t. 50 stacks of paper are used, the
paper discs stacking method for the uncoated face. . . . . . . . . . . . . . . . . . 108
4.26 Evolution of the mass of water absorbed by the paper over time; and the coefficient
of determination for a linear model applied to the mass, as a function of time.
The GHDS 35 paper is used with the coated and uncoated sides facing the water. 109
5.1 3D topographic view of speckled GHDS 35 paper obtained with the AIF set up. . 112
5.2 Evolution of the stress versus time during a tensile test in MD. The associated
pictures correspond to the out of plane displacement of the sample at a given
time (speed: 5 mm · min−1 ). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
5.3 Evolution of the stress versus time during a tensile test in CD. The associated
pictures correspond to the out of plane displacement of the sample at a given
time (speed: 2 mm · min−1 ). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
5.4 Tensile stress linked to out of plane displacement obtained during tensile tests of
speckle GHDS 35 paper. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
5.5 The evolution of the wrinkles formation with GHDS 35 paper, on the coating
pilot in function of the paper state, the misalignment angle and the web tension. 115
5.6 Experimental highlighting the appearance of wrinkles as a function of the pair:
web tension - angle of misalignment (GHDS 35 paper, uncoated side touching the
rollers). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
5.7 Web tension versus misalignment angle proposed by David Roisum [54]. 3 Regions
are highlighted: planar web, slack edge web and wrinkled web. . . . . . . . . . . 117
5.8 Experimental highlighting the appearance of wrinkles as a function of the pair:
web tension - angle of misalignment (GHDS 35 paper, coated side touching the
rollers). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
5.9 Troughs and wrinkles appearing on CPRAW paper during "coating" with water,
while using the coating pilot. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
LIST OF FIGURES 195
5.10 CPRAW, CPBS, GHDS 35 papers appearence after "coating" with water, using
the coating pilot. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
5.11 3D and 2D reconstructions of a wrinkled GHDS 35 paper. . . . . . . . . . . . . . 120
5.12 Wrinkle height map for a GHDS 35 paper (1.39 mm deep) . . . . . . . . . . . . . 120
5.13 Wrinkle height as a function of its width. Comparison between experimental and
modelled values obtained with the Weibull law on a wrinkled GHDS 35 paper.
The experimental values are the average of a wrinkle with a depth of 1.39 mm. . 121
5.14 Elevation profiles of a 25 mm deep wrinkle (GHDS 35 paper) and its height map. 122
5.15 Wrinkle height as a function of its width. Comparison between experimental and
modelled values obtained with the Weibull law on a wrinkled GHDS 35 paper.
The experimental values are the average of a wrinkle with a depth of 25 mm. . . 122
5.16 The 2 sides of GHDS 35 paper observed using SEM (Low vaccum mode, secondary
electrons). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
5.17 Delamination phenomenon observed by SEM on a wrinkled and pinched area of
GHDS 35 paper. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
5.18 X-Ray tomography of a GHDS 35 paper sample. . . . . . . . . . . . . . . . . . . 124
5.19 Cross-section of a wrinkled GHDS 35 sample observed by X-ray tomography. . . 125
5.20 Damaged cross-section of a wrinkled GHDS 35 sample observed by X-ray tomog-
raphy. The observed area is delaminated. . . . . . . . . . . . . . . . . . . . . . . 125
5.21 Modelisation of X-ray tomography wrinkle using 4 -parameters Weibull law. . . . 126
5.22 Typical stress - strain curves of the 4 different papers in CD and MD. . . . . . . 127
5.23 Evolution of tangent elastic modulus Etan versus the angle bewtween MD and
test direction Φ and parity plot comparing experimental and modelled for GHDS
35 and GHDS 40 papers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
5.24 Drying dynamic of tensile test sample (GHDS 35 paper) in conditionned room. . 130
5.25 Parity plot comparing modelled and experimental values from the drying of a
tensile test sample (GHDS 35 paper) in a conditioned room. . . . . . . . . . . . . 131
5.26 Evolution of Young’s modulus in MD and CD versus absorptiveness As for GHDS
35 paper. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
5.27 Impact of the absorptiveness at half height λs value on the evolution of the CD
Young’s modulus ECD as a function of the GHDS 35 paper absorptiveness As ,
for ne = 38.43. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
5.28 Impact of the factor of severity ne value on the evolution of the CD Young’s
modulus ECD as a function of the GHDS 35 paper absorptiveness As , for λs =
0.14. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
5.29 The relative modulus, where E0 is the modulus at 0 % moisture content, as a
function of moisture content. Solid line is the laminate model and the black dots
the measured value. The paper sheet used was isotropic [146]. . . . . . . . . . . . 134
196 LIST OF FIGURES
5.30 Example of experimental values and average value obtained during WSD test.
The sample here are in CD, tested 5 times with a force of 2 N. The average curve
is also represented. The GHDS 35 paper is used, with the coated face in contact
with water. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
5.31 Model curves obtained from WSD tests on GHDS 35 paper. The tests were
performed according to CD and MD (row), with the coated or uncoated side
of the paper in contact with water (column). The tests were carried out for 4
different strength: 1, 1.5, 2 and 2.5 N. . . . . . . . . . . . . . . . . . . . . . . . . 136
5.32 Parameters of the second order models obtained from the WSD curves, at different
forces, according to CD or MD, with the coated or uncoated side of the GHDS
35 paper in contact with water. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
5.33 Stress versus strain curves, at different time, from WSD. The sample are in CD,
with the GHDS 35 paper uncoated side facing water. Each markers represent a
given time in second. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
5.34 Evolution of WSD tangent modulus Ew,tan over time. The values are for GHDS
35 paper in CD and MD direction, with the coated side or uncoated side facing
water. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
5.35 Evolution of WSD vertical intercept bw,tan over time.The values are for GHDS 35
paper in CD and MD direction, with the coated side or uncoated side facing water.139
5.36 Wet tensile strength, obtained with the Finch’s test, for GHDS 35 and GHDS 40
in CD and MD. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
5.37 Buckling coefficient kc versus aspect ratio β for GHDS 35 for tensile test in CD
direction (left) and MD (right). i is the number of strip on the sample. . . . . . . 141
5.38 Comparison between experimental data and model values of critical tensile buck-
ling limits for GHDS 35 paper of "dry" GHDS 35 paper. . . . . . . . . . . . . . . 142
5.39 Comparison of experimental and model values for the evolution of misalignment
angle versus web tension for dry paper (µ= 0.17). The paper studied is the GHDS
35, uncoated side touching the rollers. . . . . . . . . . . . . . . . . . . . . . . . . 143
5.40 Highlighting the web width (L here) (row) and web speed (columns) on the areas
of wrinkle appearance by Hashimoto [124]. . . . . . . . . . . . . . . . . . . . . . . 143
5.41 Demonstration of the impact of the friction coefficient on the critical band voltage
value by Martz [126]. The critical band tension is inversely proportional to the
friction coefficient (all other things being equal). . . . . . . . . . . . . . . . . . . 144
5.42 Evolution of misalignment angle and number of wrinkles versus web tension for
a considered dry or wet GHDS 35 paper, with discrete and continuous i. The
uncoated side is touching the rollers. . . . . . . . . . . . . . . . . . . . . . . . . . 145
and measuring height hs (mm) (standard deviation: ±1 °C). . . . . . . . . . . . 148
6.2 Measured temperature by the IR (°C) sensor versus the heating temperature (°C)
and measuring angle κs (°) (standard deviation: ±1 °C). . . . . . . . . . . . . . . 149
LIST OF FIGURES 197
6.3 Schematic views of the impact of the measurement angle on measuring area shape
(h = 150 mm). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
6.4 Highlighting of the evolution of the surface of (a) the measurement area and (b)
the evolution of the ellipse-centre distance of the sample (OM distance). . . . . . 150
and sensor emissivity εs (standard deviation: ± 1 °C). . . . . . . . . . . . . . . . 151
6.6 Highlighting of the evolution of temperature isotherms on the paper web according
to different web speeds (standard deviation: ±2 °C). 300 mm correspond to the
driver side of the coating pilot. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
6.7 Displacement of the 85°C isotherm at different web speeds (standard deviation:
±2 °C). 300 mm correspond to the driver side of the coating pilot. . . . . . . . . 154
6.8 Area illuminated by an IR lamp and overlapping of the illumination areas). . . . 154
6.9 Different placement for IR sensors on a web after an IR dryer (web speed: 5
m · min−1 , hs and εs constant, κs = 60° for the ellipses). . . . . . . . . . . . . . . 154
6.10 Temperature monitoring screen installed next to the driver. . . . . . . . . . . . . 155
6.11 Trial conducted on 22-03-2022 on the GHDS 35 paper with a starch gel coating. 156
6.12 Trial conducted on 23-05-2022 on CPRAW paper and GHDS 35 paper with a
"water coating". . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
6.13 Trial conducted on 24-05-2022 on CPRAW paper and CPBS paper with a "water
coating". . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
6.14 Construction diagram of the control maps as a function of the paper absorptive-
ness or of the web speed. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
6.15 Control map for GHDS 35, at different paper friction coefficient µ and for a web
aspect ratio β = 0.2.The green lines at the bottom of each subfigures corresponds
to the value of the critical web tension that should not be exceeded in order to
avoid shear wrinkling. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
6.16 Control map for GHDS 35, at different paper friction coefficient µ and for a web
aspect ratio β = 0.5. The green lines at the bottom of each subfigures corresponds
to the value of the critical web tension that should not be exceeded in order to
avoid shear wrinkling. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
6.17 Control map for GHDS 35 as a function of web speed, at different paper friction
coefficient µ and for a web aspect ratio β = 0.2. The purple lines at the bottom
of each subfigures corresponds to the value of the critical web tension that should
not be exceeded in order to avoid shear wrinkling. . . . . . . . . . . . . . . . . . 163
6.18 Control map for GHDS 35 as a function of web speed, at different paper friction
coefficient µ and for a web aspect ratio β = 0.5. The purple lines at the bottom
of each subfigures corresponds to the value of the critical web tension that should
not be exceeded in order to avoid shear wrinkling. . . . . . . . . . . . . . . . . . 164
6.19 Buckling coefficient kc versus aspect ratio β for GHDS 35 in MD, in dry (left) or
fully wet (right) state. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
198 LIST OF FIGURES
6.20 Stereo-system mounted on the coating pilot (left) and printed paper with wrinkle
taken from both cameras (top right). 3D Reconstructions (bottom right) have
been performed in order to analyse the wrinkle geometry. . . . . . . . . . . . . . 166
6.21 CPRAW paper with wrinkles passing in front of the DIC system and a speckles
spotlight (red laser). A contrast issue is seen. . . . . . . . . . . . . . . . . . . . . 167
6.22 GHDS 40 paper with wrinkles passing in front of the DIC system and a speckles
spotlight (red laser). The wrinkle is easily visible on the right photo but much
less on the left one. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
6.23 The new industrial cameras (in red) with the lens, installed on the D.I.C. system
mounted on the coating pilot (model cameras: AV ALVIUM 1800U-1620M-CH-C
- High resolution monochrome matrix camera). . . . . . . . . . . . . . . . . . . . 168
6.24 Starch particles on a GHDS 40 paper after coating with a .002 Mayer rod (picture
obtained using SEM). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
6.25 Fluting and wrinkling on CPBS paper during coating with water. . . . . . . . . . 171
6.26 Fluting and wrinkling on CPRAW paper during coating with water. . . . . . . . 171
6.27 Counting the number of wrinkles using image analysis and tresholding. . . . . . . 172
6.28 Schematic diagram for the recirculation of the coating color on the coating pilot
using an Arduino controller and two volume pumps. . . . . . . . . . . . . . . . . 173
7.1 Evolution of paper quality. N°0 is a sample prepared during the Roll and Fun
project, before this PhD thesis; N°1 is one of the first sample coated during this
project; N°2 were made in March 2022 (the one below the arrow is nearly "perfect"
but has no uncooked starch particles. Finally N°3 is the last sample coated in
May 2022. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
A.1 The full standard space frame assembly used for the more complex coating, print-
ing and laminating arrangements. Four hot air impingement drying modules (op-
tional equipment) are shown fitted in typical positions. (1) Pilot structure; (2)
Lamination station; (3) Winding station; (4) Unwinding station; (5) Free roller;
(6) Speed measurement roller; (7) Coating station; (8) Drive engine; (9) Com-
mand board; (10) Fixing table; (12) Drying system (hot air drying, infra-red
drying or ultraviolet drying). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
B.1 Evolution of the water mass in the paper over time using the "Paper Disc Stacking"
method. Concerning the notation "Xs_Y ": X is the number of stacked discs and
Y is the sample number. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
C.1 Generated speckle with the parameters used in the following code. Parameters :
400 x 400 dpi with a noised step equal to 25, for a size of 210 x 297 mm. . . . . . 209
D.1 Evolution of the stress versus time during a tensile test in MD (sample A). The
associated pictures correspond to the out of plane displacement of the sample at
a given time (speed: 5 mm · min−1 ). . . . . . . . . . . . . . . . . . . . . . . . . . 211
LIST OF FIGURES 199
D.2 Evolution of the stress versus time during a tensile test in MD (sample B). The
D.3 Evolution of the stress versus time during a tensile test in CD (sample A). The
D.4 Evolution of the stress versus time during a tensile test in CD (sample B). The
D.5 Evolution of the stress versus time during a tensile test in CD (sample C). The
E.1 Highlighting the Young’s modulus of papers used in this project according to MD
and CD, versus their bulk density. . . . . . . . . . . . . . . . . . . . . . . . . . . 215
F.1 SEM picture of CPRAW paper. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217

F.2 SEM picture of CPBS paper. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
G.1 Evolution of strain over time during WSD test. The first line correspond to test
in CD, the second line test following MD. . . . . . . . . . . . . . . . . . . . . . . 219
H.1 3D definition of the cone-plane intersection problem. . . . . . . . . . . . . . . . . 222

H.2 2D definition of the cone-plane intersection problem. . . . . . . . . . . . . . . . . 222
I.1 Papier conforme (gauche) et non conforme (droite). . . . . . . . . . . . . . . . . . 225

I.2 Du papier tuilé au pli définitif sur le pilote de couchage RK du laboratoire LGP2. 226
I.3 Résumé visuel présentant l’organisation du manuscrit. . . . . . . . . . . . . . . . 227
I.4 Courbes d’intensité du signal en fonction du temps obtenues à partir de la mesure
DPM. Couché et non-couché renvoient à la face du papier GHDS 35 tournée vers
l’eau. Rainuré et plat renvoient à la surface du porte-échantillon. . . . . . . . . . 230
I.5 Hauteur du pli en fonction de sa largeur. Comparaison entre les valeurs expéri-
mentales et modélisées obtenues avec la loi de Weibull sur un papier GHDS 35
plié. Les valeurs expérimentales sont la moyenne d’un pli d’une profondeur de
1,39 mm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
I.6 Phénomène de délamination observé par MEB sur une zone pliée et pincée du
papier GHDS 35. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
I.7 Schéma de construction des cartes de contrôle en fonction de l’absorption d’eau
par le papier ou de la vitesse du papier. . . . . . . . . . . . . . . . . . . . . . . . 234
200 LIST OF FIGURES
List of Tables
2.1 Prediction of web wrinkles proposed by Martz [126](high arrow: increase, low
arrow: decrease, *****: very important impact, *: low weak) . . . . . . . . . . . 53
3.1 Surface tension, polar and dispersive components of liquids used during contact
angle measurements [63]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.2 Tests carried out in conditioned or unconditioned rooms . . . . . . . . . . . . . . 65
3.3 Cut-off values as a function of periodic profile RSM . . . . . . . . . . . . . . . . . 69
3.4 Test duration for Cobb tests according to ISO 535 standard. . . . . . . . . . . . . 71
3.5 Link between Mayer rod characteristics and the thickness of the wet film deposited. 74
4.1 Parameters obtained for the OdW model.Φbinder is the cooked starch mass fraction. 84
4.2 Parameters obtained for the Eyring model. . . . . . . . . . . . . . . . . . . . . . 85
4.3 Particle familly. "Heavy" or "light" are linked to the phase study, "number" or
"volume" are linked to the analyse of population. . . . . . . . . . . . . . . . . . . 89
4.4 Particle speed results from the 2 different models. . . . . . . . . . . . . . . . . . . 90
4.5 Sdr (Developed interfacial area ratio) obtained for different cut-off valuesλc , in
the case of roughness or waviness. The GHDS 35 coated face is studied here. . . 94
4.6 Parameters obtained from the modelling using the Stehr model. The first column
informs about the GHDS 35 paper side tested (coated or uncoated). The second
column indicates if the Wenzel correction is used or not. θc,i is expressed in ° and
Kc in s−1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
4.7 Parameters obtained from the modelling using the Sarma model.The first column
informs about the GHDS 35 paper side tested (coated or uncoated). The second
column indicates if the Wenzel correction is used or not. θc,i is expressed n ° and
ks in m−1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
4.8 Surface Free Energy (SFE) results of GHDS 35 paper for the 2 faces, at two
different time. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
4.9 Advancing and receding contact angle between starch gel and GHDS 35 paper,
obtained with DCA. ACA is Advancing Contact Angle, RCA is Receding Contact
Angle and SD is the Standard Deviation. . . . . . . . . . . . . . . . . . . . . . . 100
201
202 LIST OF TABLES
4.10 Parameters obtained after the modelling of the paper absorptiveness over time,
during Cobb test. Water is in contact with the coated side of GHDS 35 paper. . 102
4.11 Characteristics parameters measured with DPM method. Coated and uncoated
refereed to the GHDS 35 paper side facing water. Grooved and flat refereed to
the sample holder surface. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
4.12 The 3 different models used to modelled experiemntal data obtained with the
paper disc stacking method. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
4.13 The modelled parameters obtained from the 3 different models used. ρ, γlg , θc , η
are input values. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
4.14 The modelled parameters obtained from the 3 different models used.ρ, γlg , θc , η
are input values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
5.1 Parameters obtained from the 3 and 4 parameters Weibull models (Σr2 is the sum
of the squares of the deviations used is the LSM methode). The experimental
values are the average of a wrinkle with a depth of 1.39 mm. The paper is the
GHDS 35. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
5.2 Parameters obtained from the 3 and 4 parameters Weibull models (Σr2 is the
sum of the squares of the deviations used is the LSM methode).The experimental
values are the average of a wrinkle with a depth of 25 mm. The paper is the
GHDS 35. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
5.3 Parameters obtained from the 4 parameters Weibull models (Σr2 is the sum of
the squares of the deviations used is the LSM methode).The experimental values
(height as a function of width) are obtained from Figure 5.19. . . . . . . . . . . . 125
5.4 Mechanical characteristics of tested papers. . . . . . . . . . . . . . . . . . . . . . 127
5.5 Bending stiffness in CD and MD of GHDS 35 and GHDS 40 papers with and
without speckle. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
5.6 Poisson’s coefficient νxz,Baum , νzx,Baum and shear modulus Gxz,Baum of tested
papers calculated following Baum’s method. . . . . . . . . . . . . . . . . . . . . . 128
5.7 Poisson’s coefficient νxz and shear modulus Gxz of tested papers calculated fol-
lowing mechanical constitutive equation and error compare to Baum’s method. . 128
5.8 Results of modelling Young’s modulus versus absorptiveness using a Cross like
model for GHDS 35 paper. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
5.9 Buckling coefficient and critical tension stress to have buckling determined with
Zhu model, for the GHDS 35 paper following CD and MD. . . . . . . . . . . . . 141
6.1 Relative error (%) on the measured temperature compared to a reference (εs = 0.95).151
6.2 Characteristics of GHDS 35 uncoated, coated with .006 Mayer bar and 15%wt
coating color (GHDS coated 1) and coated with .012 Mayer bar and 15%wt coating
color by weight (GHDS coated 2) Both colors contained 1% particles. . . . . . . 169
C.1 Python code to generate speckle made by Maxime Teil. . . . . . . . . . . . . . . 210

LIST OF TABLES 203
I.1 Les paramètres modélisés obtenus à partir des 3 différents modèles utilisés. ρ, γlg ,
θc , η sont des valeurs d’entrée. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
204 LIST OF TABLES
Appendix A
Coating pilot diagram
In this Appendix, a more complete diagram of the coating pilot, provided by the manufacturer,
is presented.
Figure A.1: The full standard space frame assembly used for the more complex coating,
printing and laminating arrangements. Four hot air impingement drying modules (optional
equipment) are shown fitted in typical positions. (1) Pilot structure; (2) Lamination station;
(3) Winding station; (4) Unwinding station; (5) Free roller; (6) Speed measurement roller; (7)
Coating station; (8) Drive engine; (9) Command board; (10) Fixing table; (12) Drying system
(hot air drying, infra-red drying or ultraviolet drying).
205
206 APPENDIX A. COATING PILOT DIAGRAM
Appendix B
Paper disc stacking : complementary results
In this Appendix, a complementary Figure is presented concerning the Paper Disc stacking.
Figure B.1: Evolution of the water mass in the paper over time using the "Paper Disc
Stacking" method. Concerning the notation "Xs_Y ": X is the number of stacked discs and Y
is the sample number.
207
208 APPENDIX B. PAPER DISC STACKING : COMPLEMENTARY RESULTS
Appendix C
Generation of speckle for tensile test used in
D.I.C
In this Appendix, the Python code used to generate speckle and the associated speckle picture.
The code have been created by Maxime Teil.
Speckle picture generated
Here a example of the speckle that has been printed on D.I.C sample.
Figure C.1: Generated speckle with the parameters used in the following code. Parameters :
400 x 400 dpi with a noised step equal to 25, for a size of 210 x 297 mm.
209
210 APPENDIX C. GENERATION OF SPECKLE FOR TENSILE TEST USED IN D.I.C
Python code to generate speckle

Here the code used to generate the speckle printed.
# −∗− c o d i n g : u t f −8 −∗−
# Import modules
from PIL import Image
from random import r a n d r a n g e
from o s . path import j o i n
# −−−−− D e f i n e p r o p e r t i e s −−−−− #
p r i n t _ r e s o l u t i o n = (400 , 400) # i n d p i ( l e f t t o r i g h t , top t o
bottom )
image_size = (210 , 297) # i n mm ( l e f t t o r i g h t , top t o bottom )
n o i s e d _ s t e p = 25
saving_path = ’ . / ’
INCH_MM = 2 5 . 4 # 1 i n c h = 2 . 5 4 cm
# −−−−− Generate image −−−−− #
image_definition = [ ]
f o r m, n i n z i p ( image_size , p r i n t _ r e s o l u t i o n ) :
i m a g e _ d e f i n i t i o n . append ( i n t (m / INCH_MM ∗ n ) )
image_definition = tuple ( image_definition )
img = Image . new ( mode=’L ’ , s i z e=i m a g e _ d e f i n i t i o n , c o l o r =255)
p i x e l s = img . l o a d ( )
# −−−−−Generate n o i s e −−−−− #
f o r j in range ( image_definition [ 1 ] ) :
f o r i in range ( image_definition [ 0 ] ) :
gray_value = r a n d r a n g e ( s t a r t =0,
s t o p =256 ,
s t e p=i n t ( n o i s e d _ s t e p ) )
p i x e l s [ i , j ] = gray_value
# −−−−− Saving t h e image −−−−− #
image_name = ’ n o i s e d _ s t e p { : 0 3 d}_{ : 0 4 d}x { : 0 4 d} dpi_ { : 0 3 d}x { : 0 3 d}mm.
jpg ’ . format (
in t ( noised_step ) , i nt ( print_resolution [ 0 ] ) , int (
print_resolution [ 1 ] ) ,
i n t ( image_size [ 0 ] ) , i n t ( image_size [ 1 ] ) )
img . s a v e ( j o i n ( saving_path , image_name ) , ’ jpe g ’ )
Table C.1: Python code to generate speckle made by Maxime Teil.

Appendix D
Out of plane displacement of sample during
tensile test
Tensile test in MD
Here are presented other results for tensile linked to buckling. The sample are in MD.
Figure D.1: Evolution of the stress versus time during a tensile test in MD (sample A). The
associated pictures correspond to the out of plane displacement of the sample at a given time
(speed: 5 mm · min−1 ).
Tensile test in CD
Here are presented other results for tensile linked to buckling. The sample are in CD.
211
212APPENDIX D. OUT OF PLANE DISPLACEMENT OF SAMPLE DURING TENSILE TEST
Figure D.2: Evolution of the stress versus time during a tensile test in MD (sample B). The
Figure D.3: Evolution of the stress versus time during a tensile test in CD (sample A). The
213
Figure D.4: Evolution of the stress versus time during a tensile test in CD (sample B). The
Figure D.5: Evolution of the stress versus time during a tensile test in CD (sample C). The
214APPENDIX D. OUT OF PLANE DISPLACEMENT OF SAMPLE DURING TENSILE TEST
Appendix E
Mechanical properties of papers
Young’s modulus versus apparent density

Here are presented the Young’s modulus of papers used in this project according to MD and
CD, versus their apparent density.
Figure E.1: Highlighting the Young’s modulus of papers used in this project according to MD
and CD, versus their bulk density.
215
216 APPENDIX E. MECHANICAL PROPERTIES OF PAPERS
Appendix F
S.E.M picture of CPRAW and CPBS paper
This Appendix presents complementary SEM pictures of CPRAW and CPBS papers, for both
side.
CPRAW paper
(a) Marked side - x200 (b) Marked side - x400
(c) Unmarked side - x200 (d) Unmarked side - x400
Figure F.1: SEM picture of CPRAW paper.
217
218 APPENDIX F. S.E.M PICTURE OF CPRAW AND CPBS PAPER
CPBS paper
Here are presented CPBS paper pciture at different magnifications.
(a) Marked side - x200 (b) Unmarked side - x200 (c) Marked side - x400
(d) Unmarked side - x400 (e) Marked side - x800 (f) Unmarked side - x800
Figure F.2: SEM picture of CPBS paper.

Appendix G
WSD: complementary data
In this Appendix, experimental data not presented in the main body of this document are
available.
(a) CD - coated side in contact with water (b) CD - uncoated side in contact with water
(c) MD - coated side in contact with water (d) MD - uncoated side in contact with water
Figure G.1: Evolution of strain over time during WSD test. The first line correspond to test in
CD, the second line test following MD.
219
220 APPENDIX G. WSD: COMPLEMENTARY DATA
Appendix H
Determination of the elliptical shape of the
intersection between a plane and a cone
Definition of the problem in 2D and 3D
This demonstration is linked to the subsection 6.1.1 focusing of the factors influencing a non
contact IR temperature measurement. The evolution of the measuring area is discussed and
it has been shown that a parallel could be done between this practical question and the more
theoretical problem of the intersection of a cone and a plane. The goal of this demonstration
is to find the parameters of the shape created with the intersection of this two objects while
knowing the cone shape and the angle between a the two object.
Consider a reference orthonormal frame (xyz), of centre O(0, 0, 0). A cone of revolution
Π is defined. Its axis of revolution is (Oz) and its apex is the point C(0, 0, c) where c ∈ R∗+ .
The shape of the cone is known and its aperture angle is Γa . A cutting plane P is defined. It
intersects the cone Π. The plane makes an angle Ψ with the plane (Oxy). The angle between
the cone axis and the plane P is then 90 − Ψ as illustrated in Figure H.1.
In the present case, the cutting plane P does not pass through the apex C of the cone.
Moreover, the inclination angle of the cutting plane to the cone axis 90 − Ψ is greater than
Γa (the angle formed by the cone axis and a cone generatrix). It is then well known that the
intersection will be an ellipse.
To make the solution of this problem easier, it will be studied in 2D as shown in Figure H.2.
The plane P is shown as a line here (in cross section). This line passes through O(0, 0). Ψ is
the angle formed between P and the vector → −
x . C(0, c) is the apex of the cone Π, shown in
cross section here. The points A1 (xA1 , zA1 ) and B1 (xB1 , zB1 ) are known. The unknowns of the
problem are the points A2 (xA2 , zA2 ) and B2 (xB2 , zB2 ) and M (xM , zM ) which is the centre of
the segment [A2 B2 ]. The vector →
−n is the normal of P .
221
222APPENDIX H. DETERMINATION OF THE ELLIPTICAL SHAPE OF THE INTERSECTION BETWEE
Figure H.1: 3D definition of the cone-plane intersection problem.
Figure H.2: 2D definition of the cone-plane intersection problem.
Intersection between a plane and a cone
First, let’s define the equation H.1 of the line (P) which is z(x):
z(x) = tan(Ψ) · x (H.1)
We must also define the equations of the line (A1 C) defined as follows by the equation H.2
with z+ (x):
z+ (x) = p+ · x + c (H.2)
Where:
223
zc − zA1
p+ = (H.3)
xc − xA1
The same way, (B1 C) is defined by equation H.4 with z− (x):
z− (x) = p− · x + c (H.4)
Where:
zc − zB1
p− = (H.5)
xc − xB1
The intersection between P and (A1 C) is now of interest. Therefore, the equation H.6 is
obtained. Moreover, using FigureH.2 it is easily seen that the lines intersect at A2 . Then:
z(x) = z+ (x) (H.6a)

⇐⇒tan(Ψ) · xA2 = p+ · xA2 + c (H.6b)
c
⇐⇒xA2 = (H.6c)
tan(Ψ) − p+
The intersection between P and (B1 C) is now studied. The equation H.7 is obtained. More-
over, using FigureH.2 it is easily seen that the lines intersect at B2 . Then:
z(x) = z− (x) (H.7a)

⇐⇒tan(Ψ) · xB2 = p− · xB2 + c (H.7b)
c
⇐⇒xB2 = (H.7c)
tan(Ψ) − p−
Once the abscissae of A2 and B2 are known, the corresponding ordinates can be determined
c
as follow with equations H.8 and H.9. Knowing that xA2 = , it leads to:
tan(Ψ) − p+
tan(Ψ) · c
yA2 = (H.8)
tan(Ψ) − p+
c
The same way, knowing that xB2 = , it leads to:
tan(Ψ) − p−
tan(Ψ) · c
yB2 = (H.9)
tan(Ψ) − p−
The coordinates of the intersections between the cone Π and the cutting plane P are now known
in a plane problem.
Determination of the parameters of the ellipse

Knowing the coordinates of points A2 and B2 , the parameters of the ellipse formed by the
intersection of the cone Π and the cutting plane P will be determined. For this, the previously
224APPENDIX H. DETERMINATION OF THE ELLIPTICAL SHAPE OF THE INTERSECTION BETWEE
calculated data will be used. Indeed, the segment [A2 B2 ] corresponds to the major axis of the
ellipse, the point M being the middle of this segment. Thus:
s 2 2
c c tan(Ψ) · c tan(Ψ) · c

∥A2 B2 ∥ = − + − (H.10)
tan(Ψ) − p+ tan(Ψ) − p− tan(Ψ) − p+ tan(Ψ) − p−
Finally, the coordinates of M are obtained with equation H.11 and equation H.12:
1 c c

xM = · + (H.11)
2 tan(Ψ) − p+ tan(Ψ) − p−
1 tan(Ψ) · c tan(Ψ) · c

zM = · + (H.12)
2 tan(Ψ) − p+ tan(Ψ) − p−
Appendix I
French extended abstract
Contexte
Depuis sa création, le papier en tant que matériau n’a cessé d’évoluer pour répondre aux be-
soins de la société. Actuellement, la recherche se concentre sur la réduction des grammages à
caractéristiques mécaniques constantes et la fonctionnalisation des papiers afin d’obtenir des pro-
priétés de surface spécifiques (rugosité, barrière à l’huile et à l’humidité, perméabilité à l’oxygène
et à l’eau, etc.) pour la réduction de l’empreinte environnementale de l’industrie papetière. Ces
caractéristiques peuvent être obtenues en enduisant la surface du papier avec une sauce de
couchage aqueuse appropriée. La qualité du couchage obtenu dépend du triplet : papier/sauce
de couchage/procédé de couchage. Une fois le papier couché, des propriétés importantes doivent
être contrôlées : absorption d’eau, grammage déposé, surface et hydro-expansion. Pour la sauce
de couchage, les propriétés rhéologiques, la teneur en matière sèche et la composition chimique
doivent être prises en compte. Quant au procédé, les principaux paramètres sont la quantité de
sauce de couchage transférée, le temps de parcours entre le couchage et le séchage, les conditions
de séchage, la tension de la bande et l’alignement des rouleaux.
Lors du couchage de papiers de très faibles grammages, des plis peuvent apparaître à la
sortie de la zone de couchage (Figure I.1) qui, après séchage et selon la nature de la sauce de
couchage, peuvent être irréversibles et entraîner des non-conformités et des déclassements des
produits.
Figure I.1: Papier conforme (gauche) et non conforme (droite).
225
226 APPENDIX I. FRENCH EXTENDED ABSTRACT
Il s’agit ici d’étudier un couplage entre l’absorption d’eau, qui réduit les propriétés mé-
caniques du papier, et les phénomènes de flambage, éventuellement dus au cisaillement, au
désalignement et à la tension de bande. Premièrement, l’eau provenant de la sauce de couchage
affaiblit les propriétés mécaniques du réseau de fibres.
Deuxièmement, la charge imposée par la tension de bande (utilisée pour tendre le papier
pendant le processus) peut être non uniforme sur la laize du papier. Cela peut être dû à un
mauvais alignement des rouleaux. Cette répartition non uniforme de la charge pourrait générer
des contraintes de compression secondaires perpendiculaires à la direction de transport du papier
et conduire à une déformation hors plan : c’est le flambement. En passant d’un brin libre à un
autre, il se tuile (sur le rouleau). Si le papier tuilé passe sur le rouleau, il créera des plis. Le
flambement pourrait aussi n’être qu’un "flambement par tension" si le critère de flambement est
faible.
Une question vient alors immédiatement à l’esprit : "qu’est-ce qu’un pli ?" Le pli est un
défaut du papier le plus courant et le plus coûteux que l’on rencontre dans la fabrication et la
transformation des papiers de faible grammage, de film et de textile. La Figure I.2 représente
l’évolution d’un pli : du papier tuilé au pli définitif.
Figure I.2: Du papier tuilé au pli définitif sur le pilote de couchage RK du laboratoire LGP2.
Ce projet se déroule au sein des équipes MatBio et BioChip,du laboratoire LGP2, à Grenoble
(France). Il s’inscrit également dans la continuité d’une collaboration avec l’INSA Lyon, soutenue
par le LGP2 (projet Roll and Fun, 2018) dans le cadre du développement d’un isolant thermique
à base de papier. L’objectif de cet ancien projet était de produire un isolant thermique à base
de papier en process continu. Cependant, la problématique du plissement du papier après
couchage et séchage demeurait. Il faut également noter que cette thématique des variations
dimensionnelles liées à l’hydro-expansion des supports fibreux et à la déformation hors plan est
un problème récurrent pour fabriquer des produits fibreux.
227
Objectifs de ce doctorat
Ainsi, l’objectif principal de cette thèse de doctorat est de bien comprendre les mécanismes
sous-jacents à l’apparition des plis lors du processus de couchage bobine-bobine. L’objectif final
est de limiter ou d’éviter totalement l’apparition de plis lors du dépôt d’une couche fonctionnelle
à la surface de papiers de très faibles grammages. Ce projet est également basé sur une étude
multi-échelle des phénomènes physiques impliqués et de leur compréhension/approche fonda-
mentale. La compréhension et la résolution technologique de ce défi peuvent ensuite conduire
au développement de nouvelles fonctionnalités pour la technologie bobine-bobine. Pour attein-
dre le but du projet, trois objectifs majeurs devront être atteints, chacun constituant un chapitre
de résultats, listés ci-dessous et illustrés par la figure I.3.
Figure I.3: Résumé visuel présentant l’organisation du manuscrit.
→ Chapitre 4 - connaissance du comportement du papier lorsqu’il est couché :

Ce premier chapitre de résultats s’attachera à comprendre les phénomènes d’absorption
d’eau du papier GHDS 35. La mouillabilité du papier sera étudiée en tenant compte de
sa rugosité. Ensuite, l’absorption d’eau sera analysée à l’aide de 3 techniques : Cobb,
DPM et "empilement de disques de papier". Ce chapitre traitera également de la sauce de
couchage, en se concentrant sur la cuisson de l’amidon, la rhéologie et les phénomènes de
sédimentation qui peuvent se produire avec les particules d’amidon.
→ Chapitre 5 - compréhension de la mécanique de la formation des plis à travers

une approche multi-échelle: Le deuxième chapitre de résultats s’intéressera fortement
de la mécanique des plis. A travers l’observation statique et dynamique du phénomène
de pli, des observations seront faites lors d’essais de traction par corrélation d’images
numériques (D.I.C) et d’essais sur un pilote de couchage. Les plis produits sur ce dernier
seront caractérisés a posteriori aux échelles macroscopique et microscopique. Des modèles
issus de la littérature seront utilisés pour mieux comprendre les phénomènes.
→ Chapitre 6 - instrumentation du pilote de couchage et développement d’un

outil d’aide à la conduite : Le dernier chapitre de résultats présentera en premier lieu

l’instrumentation du pilote de couchage. L’étude d’un capteur de température et de son
suivi au cours du processus fera l’objet de la première partie de ce chapitre. La seconde
partie sera consacrée à la réalisation d’une carte de contrôle, afin de mieux maîtriser le
procédé. Les papiers couchés dans le laboratoire seront également présentés, ainsi que
leurs caractéristiques.
Ces 3 objectifs (et chapitres) seront précédés d’une revue de la littérature (Chapitre 2) et d’une
présentation des matériaux et méthodes utilisés dans ce projet (Chapitre 3).
Résultats principaux
Absorption de la sauce de couchage
Dans le cadre de cette thèse, différents points ont été mis en avant. Nous allons les parcourir
dans l’ordre chronologique de ce document. Tout d’abord, nous nous sommes intéressés à la
pénétration de l’eau dans le papier. Deux aspects principaux ont été étudiés : les propriétés
de la sauce de couchage et la mouillabilité et l’absorption de l’eau par le papier. Dans l’étude
de la sauce de couchage, il a été démontré que la cuisson de l’amidon est un point important
du processus, car elle a un impact sur la rhéologie de la sauce. Il a été montré que la sauce
d’amidon a un comportement rhéofluidique (loi d’Ostwald-de-Waele) et que son comportement
dépend également de la température à laquelle elle est utilisée (loi d’Eyring), ce qui est en accord
avec la littérature.
La sédimentation des particules d’amidon a été étudiée dans l’eau et rapidement observée
dans le gel d’amidon. La vitesse de sédimentation des particules est importante compte tenu des
dimensions de la cuve de stockage de la sauce sur le pilote : il a été estimé qu’il fallait moins de
12 minutes pour que les particules les plus lourdes se déposent dans la cuve de stockage. Pour
la sauce de couchage, une vitesse de sédimentation 10 000 fois plus faible que dans l’eau a été
calculé grâce aux modèles. Cependant, expérimentalement, une sédimentation et une agrégation
des particules d’amidon dans le gel d’amidon ont été observées. Une recirculation de la sauce
de couchage devrait être installée pour éviter ce type de phénomène de sédimentation.
La mouillabilité et l’absorption d’eau du GHDS 35 ont également été discutées. Nous
sommes partis de la loi de Lucas-Washburn, pour laquelle nous avons dû récupérer les dif-
férents paramètres nous permettant de l’utiliser. Un point intéressant est que les énergies de
surface du papier ont montré des composantes dispersives plus importantes que les composantes
polaires des deux côtés. Cela signifie probablement que les fibres ont été traitées pour la résis-
tance à l’état humide, ce qui explique les angles de contact élevés entre l’eau et le papier sur
toutes les faces. L’étude de l’absorption d’eau par le papier GHDS 35 a été réalisée à l’aide de
trois méthodes : la méthode COBB, la méthode DPM et la méthode d’empilement des disques
de papier. Sur de longues périodes, la méthode Cobb a montré un comportement d’absorption
d’eau de premier ordre du papier. La méthode d’empilement des disques de papier est une nou-
velle méthode proposée pour suivre l’absorption d’eau dans le papier sur de courtes périodes.
Trois modèles ont été testés sur les valeurs expérimentales : Lucas-Washburn, premier ordre et
229
loi de puissance (Table I.1). Le modèle du premier ordre semble être celui qui modélise le mieux
le comportement observé.
L-W Premier ordre Loi puissance

c (m5 ) = 7.40 10−12 mwa,eq = 0.029 λa = 0.0063
ρ (kg · m3 ) = 1000 τa = 4.900 na = 0.6376
γlg [0.001N · m−1 ] = 0.072 Σr2 = 0.00000438 Σr2 = 0.00007
θc = 75° R2 = 0.999 R2 = 0.988
η P a · s−1 = 0.001 - -
Σr2 = 1.64 10−4 - -
R2 = 0.993 - -
Table I.1: Les paramètres modélisés obtenus à partir des 3 différents modèles utilisés. ρ, γlg ,
θc , η sont des valeurs d’entrée.
Cependant, il est nécessaire de tenir compte de l’écart-type très important des données
expérimentales. Il est donc nécessaire d’optimiser le protocole et l’appareil de mesure. Enfin,
la technique DPM a montré un comportement particulier des papiers couchés pouvant rappeler
un papier traité AKD (Figure I.4). Ceci confirme l’hypothèse émise suite à la détermination de
l’énergie de surface.
Phénomènes de plis sur le papier

La deuxième partie expérimentale de ce travail portait sur les propriétés et le comportement
mécanique des papiers pliés. En nous basant sur la littérature relative à l’apparition de plis dans
les matériaux sous forme de bande, nous avons étudié l’impact d’un angle de désalignement d’un
rouleau et d’une tension critique à laquelle les plis peuvent apparaître. Lors des essais in-situ,
le plissement sous tension a été observé dans les CD et MD suivants.
Des plis ont également été observés sur le pilote de couchage en faisant varier la tension de
bande et l’angle de désalignement d’un rouleau. Ces observations ont été faites sur 4 papiers
différents et les conclusions ont été les mêmes : des plis sont apparus. Les observations ex-situ
ont été faites sur des échantillons de papiers pliés lors de l’utilisation du pilote de couchage. En
utilisant le dispositif Alicona, il a été possible de confirmer la forme asymétrique des plis (Figure
I.5), ainsi qu’un phénomène de pincement observé le long des plis. La forme des plis a pu être
modélisée par une loi de Weibull à 4 paramètres I.1.
kW r −1 kW r
kW r lW r − θW r lW r − θW r

−
HW r (lW r , kW r , λW r , θW r , AW r ) = AW r · · ·e (I.1)
λW r λW r λW r
De plus, il a été constaté que les plis ne suivent pas le MD : ils sont orientées en raison du
désalignement des rouleaux. Les observations microscopiques par MEB et tomographie à rayons
X nous ont permis de mettre en évidence le phénomène de pincement (Figure I.6). Ce dernier
semble être responsable d’une décohésion du réseau fibreux ainsi que d’une délamination de la
couche déposée industriellement sur le GHDS 35.
Suite à ces observations, une modélisation a été réalisée à l’aide de modèles issus de la
(b) coated and uncoated - flat - zoom in

(a) coated and uncoated - flat 0-2s
(d) coated - flat and grooved - zoom

(c) coated - flat and grooved between 0-2 s
(f) coated and uncoated - grooved - zoom

(e) coated and uncoated - grooved between 0-2 s
(h) Uncoated - flat and grooved - zoom

(g) Uncoated - flat and grooved between 0-2 s
Figure I.4: Courbes d’intensité du signal en fonction du temps obtenues à partir de la mesure
DPM. Couché et non-couché renvoient à la face du papier GHDS 35 tournée vers l’eau.
Rainuré et plat renvoient à la surface du porte-échantillon.
231
Figure I.5: Hauteur du pli en fonction de sa largeur. Comparaison entre les valeurs
expérimentales et modélisées obtenues avec la loi de Weibull sur un papier GHDS 35 plié. Les
valeurs expérimentales sont la moyenne d’un pli d’une profondeur de 1,39 mm.
Figure I.6: Phénomène de délamination observé par MEB sur une zone pliée et pincée du
papier GHDS 35.
littérature. Les propriétés mécaniques des papiers ont dû être déterminées. Deux défis ont été
relevés : la caractérisation du papier humide (saturé en eau) et la détermination du coefficient
de Poisson. Une méthode élégante a été utilisée pour accéder au coefficient de Poisson, présentée
équation I.2).
1 1 −νxz 1 1

= · cos4 ϕf + 2 · + · sin2 ϕf · cos2 ϕf + · sin4 ϕf (I.2)
Etan Ex Ex Gxz Ez
En ce qui concerne le comportement du papier GHDS 35 saturé en eau, nous avons pu
quantifier son séchage dans une pièce conditionnée par une loi du premier ordre décroissante.
En suivant cette loi, il a été possible de modéliser l’évolution des modules d’Young en fonction
de la quantité d’eau absorbée par mètre carré de papier. La loi utilisée est inspirée de la loi de
Cross, et constitue un élément clé de ce projet (équation I.3).
EM D,min · (λs · As )ne + EM D,max

EM D = (I.3)
1 + (λs · As )ne
La méthode de mesure doit encore être améliorée car une hystérésis dans l’absorption d’eau
par le papier peut exister et avoir un impact sur l’évolution des modules d’Young. Néanmoins, le
comportement observé est proche de celui de la littérature : ceci est rassurant pour la poursuite
éventuelle de ce travail. Le test WSD a permis de mettre en évidence un comportement d’ordre
2nd selon la CD et un comportement de premier ordre selon la MD lors de cet essai. Grâce
aux résultats obtenus, un nouveau paramètre a été identifié : le module tangent obtenu avec
la mesure WSD. Ce paramètre permet de quantifier l’évolution des propriétés mécaniques des
papiers en fonction du temps de test WSD, et par conséquent en fonction de la quantité d’eau
pendant une phase d’absorption. Pour améliorer l’expérience, une mesure de la teneur en eau
locale du papier, sur la largeur, devrait être installée.
Toutes ces informations ont été utilisées pour alimenter le modèle de Hashimoto (équation I.4
et équation I.5) pour l’apparition des plis sur le pilote de couchage. Nos valeurs expérimentales
ont été comparées aux prédictions du modèle.
6τcr a2
θcr = (I.4)
Ex b2
2t2f
s
Ex · Ez
Tcr = (I.5)
Il a été surprenant de constater que nos valeurs, en particulier l’angle de désalignement, sont
très grandes par rapport à la littérature. Sans être incohérente, la vitesse lente utilisée lors de
l’expérience pourrait expliquer cette différence. Dans une moindre mesure, la discrétisation du
nombre de plis dans le modèle de Hashimoto a été étudiée. Nous avons pu montrer que rendre
la variable représentant le nombre de plis entière conduisait à ne pas respecter l’équation limite
permettant de définir i et à surestimer ou sous-estimer l’angle de désalignement à partir duquel
les plis pouvaient apparaître. Néanmoins, la notion même d’un nombre de plis non entier est
difficile à appréhender. Cependant, cela pourrait être lié à des zones de transition d’équilibre.
Enfin, l’analyse des essais de traction et leur comparaison avec le modèle de Zhu (équation I.6)
ont validé nos valeurs. Ce modèle pourrait également être utilisé pour prédire le flambage en
traction qui peut se produire sur le pilote lorsque les rouleaux sont bien alignés, que le papier
est humide et que des rides se forment.
3ξ1 − 1 + 2ξ12 νxz 2

( )
2
h i
kc = β + ξ1 β −2 n2 (n + 2) + n2 + (n + 2)2 (I.6)
ξ1 + 1 + 2ξ1 νxz
Étude du processus de couchage sur le pilote de laboratoire
Le dernier chapitre expérimental porte sur l’étude du procédé de couchage sur le pilote lui-même.
Deux points ont été abordés : la mesure de la température pendant le procédé et l’optimisation
du procédé. La mesure de la température au cours du procédé a été étudiée à l’aide de capteurs
IR. Il a été montré que l’émissivité réglée sur les capteurs devait être proche de 0,95 et que
les capteurs devaient être aussi perpendiculaires que possible à la bande. La distance entre le
capteur et la bande dépend de la valeur recherchée : locale ou globale. Cette étude a également
mis en évidence la non-homogénéité des isothermes de température à la sortie du sécheur IR
sur le papier. Ce phénomène pourrait également justifier le tuilage du papier, due à un séchage
233
hétérogène, bien qu’aucun désalignement des rouleaux n’ait été observé. Les capteurs ont permis
de suivre le processus de couchage au cours de différents passages et d’illustrer les différents
événements se produisant au cours du processus tels que la rupture de la feuille de papier, le
changement d’équipement ou le couchage à différentes vitesses.
Le dernier aspect de ce chapitre était l’optimisation du processus en reliant la tension de la
bande, l’angle de désalignement et la vitesse de la bande. Pour cela, tous les outils et modèles
développés précédemment ont été utilisés. Ainsi, nous avons lié la vitesse de la bande de papier
à l’absorption d’eau, qui à son tour était liée aux modules d’Young par la loi de type Cross. Ce
modèle a lui-même été intégré aux lois proposées par Hashimoto, en ajoutant une variation des
coefficients de Poisson grâce au modèle empirique de Baum. La Figure I.7 résume la philosophie
de création des cartes de contrôle utilisant les différents modèles
Ceci a permis de générer des cartes de contrôle et de conduite du processus permettant
de déterminer l’angle de désalignement à ne pas dépasser en fonction de la vitesse du papier
sur le pilote et de la tension de bande. Cependant, ces cartes ne prennent pas en compte les
phénomènes de tension de flambage. Pour cela, il est nécessaire d’utiliser en première approche
le coefficient kc proposé par Zhu.
Conclusion finale
Cette thèse a permis de comprendre les phénomènes liés à l’apparition des plis. Ce travail offre
une nouvelle perspective à l’étude de ce type de phénomène. Il serait intéressant d’acquérir plus
de données concernant pour nourrir le modèle de Cross liant absorption d’eau et module d’Young
du papier. Pouvoir acquérir plus d’informations avec le système DIC, lorsque le processus est
en cours, serait un plus pour comprendre les phénomènes in-situ et réguler le processus par la
mesure et le contrôle de la tension de bande. Une mesure du profil d’humidité du papier sur
la laize du papier serait un plus. Pour conclure, nous répondrons à cette dernière question :
a-t-il été possible de coucher un papier, sans plis, avec un gel d’amidon contenant des granulés
d’amidon ? La réponse est oui car le procédé est maintenant maîtrisé grâce à la création des
cartes de contrôle.
Figure I.7: Schéma de construction des cartes de contrôle en fonction de l’absorption d’eau par
le papier ou de la vitesse du papier.
RESUME EN FRANÇAIS :
Depuis sa création, le papier en tant que matériau n'a cessé d'évoluer pour répondre aux
besoins de la société. Actuellement, la recherche se concentre sur la réduction des
grammages avec des caractéristiques mécaniques constantes et la fonctionnalisation des
papiers afin d'obtenir des propriétés de surface spécifiques et pour réduire l'empreinte
environnementale des papeteries. Ces caractéristiques peuvent être obtenues en enduisant
la surface du papier avec une couche de revêtement aqueuse appropriée. La qualité du
couchage obtenu dépend du triplet : papier/couche déposée/procédé d’enduction. Lors du
couchage de papiers très légers, des plis peuvent apparaître à la sortie de la zone de
couchage. Après séchage et selon la nature de la couche, les plis peuvent être irréversibles
et entraîner des non-conformités. Le but de ce projet est donc de comprendre comment et
pourquoi ses plis peuvent apparaitre. Dans un premier lieu, l’étude de l’absorption de l’eau
par le papier a été menée. Cette dernière a permis de déterminer une loi du premier ordre
reliant la quantité d’eau absorbé par le papier en fonction du temps. Dans un second temps,
le phénomène de plis est étudié, à différentes échelles, du point de la mécanique. Cela à
permis de mettre en avant l’impact des modules d’Young sur l’apparition des plis, notamment
en présence d’eau. De plus, à une échelle microscopique, une décohésion du réseau fibreux
est constatée. Les plis ont donc un impact macroscopique et microscopique. Le dernier point
abordé lors de ce projet est l’optimisation du procédé de couchage. Trois paramètres de
process ont été mis en relation : la vitesse de défilement du papier, l’angle de désalignement
des rouleaux les uns par au rapport aux autres et la tension de bande. Cela a permis
l’élaboration d’une carte de contrôle permettant de conduire le couchage dans les meilleures
conditions. Enfin un papier de faible grammage a pu être couché sans pli.
ABSTRACT IN ENGLISH :
Since its creation, paper as a material has constantly evolved to meet society's needs.
Currently, research focus in the reduction of grammages with constant mechanical
characteristics and the functionalization of papers in order to obtain specific surface
properties for the reduction of the environmental foot printing of the paper mills. These
characteristics can be obtained by coating paper surface with suitable aqueous coating color.
The quality of the coating obtained depends on the triplet: paper/coating color/coating
process. When coating very light papers, wrinkles may appear at the exit of the coating area.
After drying and depending on the nature of the coating, wrinkles can be irreversible and lead
to non-conformities. The aim of this project is therefore to understand how and why these
wrinkles can appear. Firstly, the study of water absorption by paper was carried out. This
allowed to determine a first order law relating the quantity of water absorbed by the paper as
a function of time. In a second step, the wrinkle phenomenon is studied, at different scales,
from the point of view of mechanics. This allowed to highlight the impact of Young's modulus
on the appearance of wrinkles, particularly in the presence of water. Moreover, on a
microscopic scale, a decohesion of the fibrous network was observed. Therefore, wrinkles
have a macroscopic and microscopic impact. The last point addressed in this project is the
optimisation of the coating process. Three process parameters were related to each other:
the web speed, the angle of misalignment of the rolls with respect to each other and the web
tension. This has enabled the development of a control map that allows the coating to be
carried out under the best possible conditions. Finally, a low-weight paper could be coated
without wrinkle.

LE-GALLIC 2022 Archivage

Transféré par

Droits d'auteur :

Formats disponibles

LE-GALLIC 2022 Archivage

Transféré par

Informations du document

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

LE-GALLIC 2022 Archivage

Transféré par

Droits d'auteur :

Formats disponibles

Understanding of the wrinkles formation during the

coating functionalization of lightweight fibrous material :

To cite this version:

HAL Id: tel-04048638

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

DOCTEUR DE L’UNIVERSITÉ GRENOBLE ALPES

Compréhension des phénomènes physiques relatifs à l'apparition de

The necessaries of life for man in this climate may, accurately

lc,jp Characteristic length (m)

As,eq Plateau value of the aborptiveness (gwater · m−2

Tx Tension in x-direction (N · m−1 )

νzx Poisson’s ratio for transverse strain in the x-direction

∆U Strain energy of bending (N · m)

3 Materials and methods 57

3.3.3 Mechanical properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

4 Absorption of the coating color by papers 79

5 The wrinkling phenomena of paper 111

6 Study of the coating process on the laboratory pilot 147

7 General conclusion 175

A Coating pilot diagram 205

B Paper disc stacking : complementary results 207

C Generation of speckle for tensile test used in D.I.C 209

D Out of plane displacement of sample during tensile test 211

E Mechanical properties of papers 215

F S.E.M picture of CPRAW and CPBS paper 217

G WSD: complementary data 219

H Determination of the elliptical shape of the intersection between a plane and

I French extended abstract 225

Figure 1.1: Conformed (left) and non-conformed (right) paper.

→ A web creases is permanent deformation from wrinkling

Figure 1.2 represented the evolution from troughs to creases.

Objectives of this PhD thesis

Figure 1.3: Visual abstract presenting the manuscript organisation.

→ Chapter 5 - understanding of the wrinkle formation mechanics throughout a

→ Chapter 6 - instrumentation of the coating pilot and development of a driving

Scientific valorisation and presentation

2.2.3.4 Tension control and sensing . . . . . . . . . . . . . . . . . . . . 30

Figure 2.1: From living cell to wood fibre [8].

2.1.1.1 Cellulose and hemicellulose

Lignin is the second largest constituent of wood by weight: 20 to 30%. It is a non-linear,

2.1.1.3 Interaction between lignin and cellulose

2.1.1.4 The others components of wood

2.1.2 Paper making process

2.1.2.1 A short history of paper

2.1.2.2 Pulp manufacturing

Figure 2.6: Example of double disks refiner.

2.1.3 Structure and surface of paper

Paper products could be divided in 4 categories [24]:

→ Wrapping and packaging paper-cardboard: they protect and preserve large-consumer

→ Macro-roughness: at 100 µm-1 mm,is the result of paper formation;

→ Pores connected to one side of the paper only;

2.2 Sizing and coating process

2.2.1 Coating and sizing color

2.2.1.2 Binders and co-binders

Figure 2.9: Amylose (top) and amylopectin (bottom) structure [40].

2.2.1.4 Formulation and preparation of coating color

VRichardson−Zaki = VStokes (1 − ϕ)S (2.4)

2.2.2 Sizing and coating process

2.2.2.1 Coating color application on papers

Figure 2.12: Three different way to deposit coating color on paper.