Mechanical Behaviour of Adhesive Joints in Cartonboard for ...241381/FULLTEXT01.pdf · hot -melt...

Faculty of Technology and ScienceChemical Engineering

Christer Korin

Mechanical Behaviour of Adhesive Joints in

Cartonboard for Packaging

DISSERTATIONKarlstad University Studies

2009:48

Karlstad University Studies2009:48

Christer Korin

Mechanical Behaviour of Adhesive Joints in

Cartonboard for Packaging

Christer Korin. Mechanical Behaviour of Adhesive Joints in Cartonboard for Packaging

DISSERTATION

Karlstad University Studies 2009:48ISSN 1403-8099 ISBN 978-91-7063-271-6

© The Author

Distribution:Faculty of Technology and ScienceChemical EngineeringSE-651 88 Karlstad+46 54 700 10 00

www.kau.se

The cover is printed on Korsnäs WHITE 265.

Printed at: Universitetstryckeriet, Karlstad 2009

i

Abstract

A cartonboard package is often sealed and closed with an adhesive – either a hot-melt adhesive (adhesives that are applied in a molten state on the cartonboard) or a dispersion adhesive (adhesives that are applied as water-based dispersions). This thesis focuses on the process of hot-melt gluing, and how material properties and process conditions affect the performance of the adhesive joint.

Requirements vary depending on how the package is to be used. A package that is only supposed to protect the product during transport differs from one that is supposed to attract consumers and facilitate their use of the product. If a package has been opened, due to external or internal forces that cause a fracture in the adhesive joint, the consumer may choose another package instead.

A fracture of the adhesive joint may occur in several different ways; for example, a cohesive fracture in the adhesive, an interfacial fracture between the adhesive and one of the cartonboard surfaces, and a cohesive fracture in the cartonboard. The traditional way of testing the adhesive joint is to subjectively evaluate the fibre tear after manually tearing the joint apart.

The primary interest of this study has been to find an objective method that can characterise the adhesive joint – that is, its strength and joint characteristics. The work has principally concentrated on physical experiments where the Y-peel method has been evaluated and further developed, including the construction of a laboratory adhesive applicator.

Adhesive joint failure is analysed and correlated to the force-elongation curve during Y-peel testing in order to explore various mechanisms of the failure. The force versus elongation curves are transformed into a force versus inelastic deformation curve for the adhesive joint. The inelastic deformation of the adhesive joint is defined as the inelastic opening of the adhesive joint perpendicular to the cartonboard surface. The dissipative descending energy has been used to characterise the adhesive joint. High descending dissipative energy showed high resistance against final failure of the joint. This correlates very well with the manual fibre-tear test. Characteristic force-elongation curves in Y-peel testing – that is, the shape of the curve – have been analysed, and four main failure modes have been identified. The finite element method has been used to predict mechanical behaviour in the ascending part of the force-elongation curve. When it comes to local behaviour, a high stiffness adhesive results in more shearing than low adhesive stiffness, but on a global scale, no big difference was detected on the ascending part of the force-elongation curve.

ii

The new laboratory adhesive applicator and finite element method can be used to objectively design the interaction between the adhesive and the cartonboard for a specific application. This can be achieved by modifying the cartonboard, the adhesive or the process parameters.

iii

Papers included in the thesis

Paper I

J. Tryding, M. Lestelius, C. Korin and J. Lewandowski, Interpretation of Y-peel testing of adhesive sealed carton board, Verarbeitungsmaschinen und Verpackungstechnik, Dresden: 277–294 (2003)

Paper II

C. Korin, J. Tryding, M. Lestelius and N. Hallbäck, Y-peel characterization of adhesively-bonded carton board joints: An objective method, Journal of Adhesion Science and Technology 21 (2): 197–210 (2007)

Paper III

C. Korin, N. Hallbäck and R. Junghans, Failure modes in adhesively bonded carton boards, Journal of Adhesion Science and Technology 22 (16): 2079–2104 (2008)

Paper IV

C. Korin, R. Seppänen, M. Vähä-Nissi and N. Hallbäck, Influence of surface treatment on the mechanical strength of hot melt adhesive joints in carton-board, Submitted for publication (2009)

Paper V

C. Korin, N. Hallbäck, C. Barbier and M. Nygårds, Finite element analysis of the influence of adhesive stiffness variation on the mechanical strength of hot melt adhesive joints in carton board, Manuscript in preparation (2009)

Related report by the same author

M. Vähä-Nissi, C. Korin, R. Seppänen, C. Laine and N. Hallbäck, Influence of paperboard on bond formation and strength of adhesive joint, TAPPI 12th European PLACE Conference, Budapest (2009)

1

Table of contents

1. INTRODUCTION ...................................................................................................................2

1.1 CARTONBOARD .............................................................................................................................................. 3 1.1.1 Transport and consumer packages .............................................................................................................. 4

1.2 SEALING OF CARTONBOARD ........................................................................................................................ 5 1.2.1 Pre-fill sealing ............................................................................................................................................ 5 1.2.2 Closure sealing ........................................................................................................................................... 5 1.2.3 Adhesives ................................................................................................................................................... 6 1.2.4 Adhesive-application process ....................................................................................................................... 6

1.3 DEMANDS ON THE ADHESIVE JOINT .......................................................................................................... 7 1.3.1 Problem areas ............................................................................................................................................ 8

2. METHODS AND MATERIALS ..............................................................................................9

2.1 ADHESIVE APPLICATOR .............................................................................................................................. 10 2.2 Y-PEEL............................................................................................................................................................ 11

2.2.1 Force elongation ....................................................................................................................................... 13 2.2.2 Interpretation and failure characterisation ................................................................................................. 14

2.3 SURFACE TREATMENT AND CHARACTERISATION .................................................................................. 15 2.3.1 Roughness characterisation ....................................................................................................................... 15 2.3.2 X-ray photoelectron spectrometer ............................................................................................................... 15 2.3.3 Dynamical mechanical thermal analysis .................................................................................................... 16

2.4 FINITE ELEMENT ANALYSIS........................................................................................................................ 16 2.5 MATERIALS .................................................................................................................................................... 16

2.5.1 Cartonboard ............................................................................................................................................ 16 2.5.2 Hot-melt adhesives ................................................................................................................................... 17

3. THEORY ................................................................................................................................ 18

3.1 FRACTURE, WETTING AND ADHESION ..................................................................................................... 19 3.2 STRUCTURE MECHANICAL ANALYSIS ........................................................................................................ 22

3.2.1 Cartonboard model................................................................................................................................... 22 3.2.2 Continuum behaviour of plies ................................................................................................................... 23 3.2.3 Cohesive behaviour of interfaces ................................................................................................................ 24 3.2.4 Adhesive .................................................................................................................................................. 27

4. RESULTS ............................................................................................................................... 27

4.1 TYPICAL FAILURE MODES ........................................................................................................................... 27 4.2 CHARACTERISATION OF FAILURE MODES................................................................................................ 29 4.3 Y-PEEL CHARACTERISATION AND FIBRE TEAR ....................................................................................... 31 4.4 SURFACE TREATMENTS ............................................................................................................................... 32 4.5 COMPARISON BETWEEN EXPERIMENTS AND MODELLING .................................................................. 34

5. DISCUSSION ......................................................................................................................... 36

5.1 Y-PEEL TEST METHOD ................................................................................................................................. 36 5.2 Y-PEEL CHARACTERISATION ...................................................................................................................... 37 5.3 Y-PEEL FAILURE ........................................................................................................................................... 37 5.4 FINITE ELEMENT STUDY ............................................................................................................................. 39 5.5 DEMANDS ON THE ADHESIVE JOINT ........................................................................................................ 39

6. CONCLUSIONS .................................................................................................................... 40

7. FUTURE RESEARCH .......................................................................................................... 41

8. ACKNOWLEDGEMENTS .................................................................................................... 41

9. REFERENCES ...................................................................................................................... 43

10. NOTATIONS ....................................................................................................................... 47

2

1. Introduction

People have long made packages and receptacles in which to protect their goods. Later, when people started to transport and store materials, demands on the packages increased. Initially the packages were made of available natural materials best suited for their content [1]. Examples of the first packages are earthenware jars, wooden cases, leather cases or even leaves. As technology and marketing concepts developed, a demand emerged on the market for more complex and sophisticated packages to meet transport requirements, storage conditions and display provisions. At the dawn of the industrial era, mechanical converting equipment expanded the possibilities for optimising package functionalities and creating new designs of modern packages. Today nearly every product is packed in some sort of package during its entire product life or part of it. The products being packaged are very diversified; they could be anything from ball bearings, batteries, toys, food, liquids, and medicine, to luxury products such as cosmetics, cigarettes and liquor. All these products require different things of their packaging. The packaging material is also much more varied; for example, paper, wood, glass, metal and plastic.

The traditional role of the package was to protect the product during storage and transport. An additional important and increasing demand today from consumers and the industry is to integrate information and marketing into the package. Thus, the initial function of the package evolved from not only protecting the product during transport and storage but also to informing and displaying essential data on it. The package should give the consumer a positive overall experience of the contained product [2] throughout the product’s life, from production, transport, storage, display and up until the product is finally used and disposed of. On a shop shelf, the package acts as a “silent salesman”, urging the consumer to buy the item [3]. The package/product that attracts attention and looks good is chosen. This is called “the first moment of truth”. “The second moment of truth” [2] is when the consumer decides to buy the product a second time, and this depends on the relation between the consumer’s expectations and experiences when using or consuming the product. The combination of the package and its product thus creates the total experience of the product.

The process of producing a packaged product from cartonboard reels includes several converting steps such as printing, embossing, creasing, cutting, punching, folding, gluing, erecting, filling and sealing. This can be done in one single in-line converting process or in several discrete stand-alone machines that cover one or more converting steps. Each converting step is a scientific and technical challenge in itself.

Packages can be sealed in several different ways, such as by gluing, heat sealing, stapling or mechanical interlocking. This thesis focuses on the process of hot-

3

melt gluing, and how material properties and process conditions affect the performance of the adhesive joint. Understanding and characterising the mechanical strength of adhesive joints in cartonboard packages is important to ensure that the sealed package meets the demands it will face throughout its complete life cycle, from converting, distribution and recycling perspectives. A reduced operating window caused by increasing production speed and automation leads to an increased demand for an objective method for characterising the mechanical behaviour of adhesive joints. How to optimise package sealing for a specific package application becomes critical, because a failure in the converting process causes more and more losses in productivity as the production speed is increased. Using the wrong adhesive can turn out to be a costly mistake, with respect to either customer problems or reduced profit for the manufacturer. Therefore, it is important to have an objective and quantitative joint-testing method as a tool for gaining more knowledge about the quality of adhesive joints in cartonboard packages and for optimising the cost-performance relations in the entire value chain.

The purpose of this work is to find an objective test method and a way to predict the pre-peak behaviour of the Y-peel test in order to interpret how good or bad an adhesive joint is when exposed to mechanical loading. An attempt is made to predict the fracture behaviour based on the pre-peak stress and strain state. The local stress results may make it possible to predict how the process will develop; for example, whether the behaviour will be brittle or tough (that is, a pop-up failure or ply-tear) [Paper III].

In addition, the test method should be repeatable and correlate with practical experience in the converting industry. This study of adhesive joints is mainly done through physical experiments where the Y-peel test method is evaluated and further developed for the characterisation of the mechanical strength of adhesive joints in cartonboard packages. The adhesive-joint failure is analysed and correlated to the force-elongation curve to explore various mechanisms of adhesive-joint failure. The impact of the surface properties on the fracture mechanics, due to coating and/or calendering of the cartonboard, is also investigated. The force-elongation curves are analysed to gain knowledge about the behaviour of the adhesive joint during crack propagation and to make it possible to predict the behaviour of the adhesive joint during stress. The effect of the stiffness of the hot-melt adhesive (HMA) on the joint failure is investigated by a finite element analysis (FEA) using the commercial program Abaqus/Standard.

1.1 Cartonboard

Cartonboard is usually defined as a paper material ranging from about 200 g/m2 up to 600 g/m2. If the grammage is less than 200 g/m2, the material is usually

4

called paper; if it is higher than 600 g/m2, millboard. Cartonboard is often made as a multi-ply board, a sandwich construction that gives the producer the possibility to design the properties of each ply to meet various demands for different applications. Cartonboard is an anisotropic material with different properties in the in-plane directions (the machine direction, or MD, which is the direction in which the board is produced on the board machine, and the cross-machine direction, or CD) and the out-of-plane direction (the thickness direction, or ZD). The in-plane stiffnesses are 100 times or more higher than the out-of-plane stiffness; see Table 1 [4]. Furthermore, the in-plane tensile yield stresses are at least 10 times higher than the failure stress in the out-of-plane direction.

Table 1. Experimental results of uniaxial tensile tests in the machine direction

(MD) and cross-machine direction (CD) and of through-thickness tensile tests in the thickness direction (ZD) of a multi-ply board [4].

Elastic modulus [MPa]

Tensile yield stress [MPa]

Tensile failure stress [MPa]

MD 5600 12 44 CD 2000 6.5 18 ZD 18 - 0.40

1.1.1 Transport and consumer packages

Package requirements vary depending on how the package is used. A package that is only supposed to protect the product during transport differs from one that is supposed to attract customers and facilitate the use of the product by consumers.

The board and the transport package must withstand the mechanical loads they are exposed to during the complete life cycle of the package; this life cycle spans the time from when the blanks are converted up until when the package’s materials are recycled. These loads include embossing, creasing, folding, stacking, water/moisture (creep) and vibrations. Smaller consumer packages (primary packages) can be packed in a transport package (secondary packages or tertiary packages) for distribution from the manufacturer to the supermarket or shop. The size of the transport package is then adapted to contain the number of consumer packages that is convenient for the supermarket or shop to handle. The different sizes and applications define the specific demands on the sealing until the package is opened. In addition, the transport package may also be used as a display package – that is, a package that, from the shelf of a shop, urges the consumer to buy the product – which adds further demands on the sealing. In this case, it is important that the sealing opens as desired.

Along with the production and transportation demands placed on the adhesive joint, the package must also live up to the expectations of the consumer to

5

reinforce the total product experience [2]. From the consumer’s point of view, the transport package is useless immediately after opening (for example, a transport box for a refrigerator), whereas the consumer package may be used for a longer period of time (for example, packages for breakfast cereals, butter, fruit juices or cosmetics).

The demands on adhesive joints differ depending on whether the package is a transport package, consumer package or combined transport/consumer package. It is not desirable that the first time a package is opened, its appearance or function be destroyed if the package is elaborately decorated, if it is used by consumers over a long period, or if it is used as a display package on a store shelf. In such cases, fracture involving tearing or delaminating is not desired. On the other hand, maximum toughness may be preferred irrespective of the fracture process for packages where the entire content will be used at once. In some cases, it is desirable that the adhesive joint opens with a distinct snap-like manner. This requires low toughness but high strength for the package to resist loads during converting and to withstand transportation to the consumer.

1.2 Sealing of cartonboard

Fracture of the adhesive joint may occur in several different ways; for example, there may be a cohesive fracture in the adhesive, an interfacial fracture between the adhesive and one of the cartonboard surfaces, or a cohesive fracture in the cartonboard. Literature reports that most bonding is due to chemical interactions across the interface, such as Lifshitz-van der Waals and acid-base interactions (cf. [5]). In addition, diffusion and entanglement of polymer molecules between the substrate and the adhesive may increase the adhesion; mechanical interlocking between the adhesive and the substrate on different size levels may also have the same effect.

1.2.1 Pre-fill sealing

Here pre-fill sealing is defined as all kinds of sealing that occur before filling. These seals are generally not opened when the content is being used. Therefore it is important that this type of joint is intact and closed during the complete package chain. The demand on this type of joint is to resist the stresses and not to fail. If the joint eventually fails, then the character of the failure is normally of no interest to the consumer.

1.2.2 Closure sealing

The closure sealing, here defined as the closure of a filled cartonboard package, is sometimes easy to open, with a short click sound, and in other cases, it is

6

tamper-proof. When a package is nicely decorated and is to be used over an extended period of time, then it is desirable that the first time its seal is opened that the package’s appearance not be damaged through fibre tear. On the other hand, when a tamper-proof seal is required, then the opening of the seal must result in significant fibre tear as an obvious sign or evidence that someone has opened or tried to open the package.

1.2.3 Adhesives

The adhesives used in cartonboard packages are either HMAs (adhesives that are applied at high temperatures in a molten state on the cartonboard) or dispersion adhesives (adhesives that are applied as water-based dispersions). The present investigation and its results refer to HMAs.

A typical traditional HMA has these main chemical components:

- 30–40% polymer

- 20–30% wax

- 30–40% resin

The polymer gives the adhesive its cohesive mechanical strength. The resin is a wetting substance, whereas the wax gives the viscosity and regulates the adhesion (that is, the tack) when the adhesive is applied on the cartonboard. Usually a small amount of antioxidants is also added to protect the adhesive from ageing [6].

HMAs often have a chemical basis of ethyl vinyl acetate or polyolefin, and are designed to suit the cartonboard packaging industry.

Dispersion adhesives are applied as water-based dispersions, and they set when water is removed. The polymer is typically polyvinyl acetate [7]. These adhesives also contain for example surfactants, plasticisers and tackifiers. The tackifiers affect wet tack, while the surfactants improve wetting. The plasticisers increase the wet tack and improve film formation [8].

1.2.4 Adhesive-application process

In the machine converting process of producing cartonboard packages (Figure 1), there are different types of seals; for example, stapling, adhesive sealing, adhesive tape and mechanical interlocking. As the machine speed increased so did the use of adhesive seals. Adhesives also make it easier to produce nicely designed packages with high-quality printed decorations and different, untraditional shapes. At the same time, it is important for the industry that the converting process runs with very few interruptions, and that it has a high and constant quality due to tough competition (from other materials and brands).

7

Figure 1. A schematic illustration of a folder-gluer machine redrawn from [9].

It is essential to be able to control the complete adhesive-application process in order to control the quality of the adhesive joints (in this context, the adhesive joint is defined as two sheets of cartonboard with an HMA in between); this process includes the application of the adhesive on the substrate, the wetting of the board by the adhesive, the consolidation of the adhesive, and the final joint formation. This is possible by controlling certain process parameters. The most important parameters to get the full and intended strength of the joint are:

- open time: the time from application of the adhesive on one board surface (Part 1) until the other surface (Part 2) is pressed against the adhesive string

- pressure time: the time during which the adhesive joint is set under pressure

- pressure on the joint: the pressure that is applied on the adhesive joint in the adhesive applicator

- amount of applied adhesive: the amount of adhesive that is applied on the surface

- temperature of applied adhesive: the temperatures in the adhesive tank, hose and nozzle

To withstand certain package conditions, the adhesive must be applied in such a manner as to avoid indications of fracture.

1.3 Demands on the adhesive joint

In the beginning of packaging history, the demands on a package were that it should be safely closed and that the content be protected from the environment. The appearance of the seal was practically of no importance. In some cases, however, the seal needed to be tamper-proof. Later during industrialization, it was crucial for profitability that the machines ran fast, smoothly and with few stops. The optimisation of package sealing for specific package applications became more and more important, because a jam in the converting machine would cause a large drop in productivity. In addition, poor sealing can be very expensive with respect to damaged goods, customer

Carton

blanks

Pre-folding Gluing Compression

8

complaints, and loss of market image and reputation. The joint has to resist the demands throughout the complete life cycle of the package.

To optimise the adhesive joint, it is not only necessary to know whether the adhesive joint is good or bad but also how good or bad. The contribution from mechanical bonding can be particularly important for cartonboard where the mechanical bonding is achieved through mechanical interlocking of the adhesive into irregularities and pores of the cartonboard surface and the embedding of fibres sticking out from the surface (see [10], [11]) creating fibre tear (see Figure 2, top). In other cases, pop-up (see Figure 2, bottom) behaviour is required where a low degree of fibre tear is needed. A tensile test machine can be used to objectively measure the force-elongation curve of adhesive joints.

1.3.1 Problem areas

The demands on the adhesive joints originate from mechanical loading and environmental conditions (for example, forces from the surroundings, such as the weight of other packages, dynamical forces from the machines, or climate changes such as temperature and relative humidity) during converting, transport, storage, exposure on shelf, end use of the product and finally recycling. Fracture of the adhesive joint may occur in mainly three different ways:

- cohesive fracture in the adhesive

- interfacial fracture between the adhesive and the cartonboard

- cohesive fracture in the cartonboard

The failure may also be a combination of the above-listed failure modes; for example, starting as cohesive fracture in the cartonboard and, after some distance, ending as an interfacial failure. Several different test methods exist ([Paper II], [12], [13], [14]), such as the T-peel, angle-peel and Y-peel methods; see Figure 3.

9

Figure 2. Two Y-peel specimens torn apart, with and without fibre tear.

The traditional way to test a joint is to conduct a manual peel test [Paper I], [15], [16] and subjectively evaluate the degree of fibre tear. To characterise a joint and control the converting process, an objective method has been developed. This method includes a laboratory adhesive applicator and a test procedure that tears apart the adhesive joint in a well-defined manner.

2. Methods and materials

The traditional way to analyse the mechanical strength of an HMA joint in the converting industry is to visually inspect the fracture surfaces after manual peeling. The joint is acceptable if 50% of the fractured adhesive surface is covered by fibres. However, this method is very subjective and based on personal skills and experience. There are other semi-manual peeling methods in which the joint is subjected to a load, either increasing the load until the joint

Fibre tear

No fibre tear

10

fails or determining the time to failure at a constant load [12]. Still other methods are based on using tensile testing equipment in which the force-elongation curve is recorded during peeling until the joint fails [17].

Figure 3. Testing geometries for the T-peel, angle-peel and Y-peel methods

[Paper II].

The literature suggests these peel tests for packaging materials based on cartonboard: the T-peel tests [12], Figure 3 (left); the angle-peel tests [13], Figure 3 (middle); and the Y-peel tests [Paper II], [18], Figure 3 (right). These methods all have their advantages and disadvantages.

In the T-peel test method, the adhesive joint is not fixed relative to the tail, and the tail is free to move during the test. Ways to overcome this movement are to stabilise the tail by holding the tail in a fixed position or by adding a stabilising tab [12], [19]. In the angle-peel and the Y-peel test methods, no arbitrary movement is allowed. It is further observed that in the angle-peel test, the sample needs to be bonded to a sliding support block. This makes the preparation work for the angle-peel test more time consuming compared to the other two methods. The Y-peel method is a redesign of the constrained T-peel test. The Y-peel method is discussed in detail below.

2.1 Adhesive applicator

It was necessary to have a laboratory adhesive applicator to be able to simulate the different process parameters in the adhesive line of a converting and filling machine. The intention was that this laboratory adhesive applicator could be used for production control and development work, by the cartonboard industry, and in academic research. It was designed and developed for HMAs and dispersion adhesives; cf. Paper I. In the remaining studies, however, dispersion adhesives were not used.

F

FR

R

FF

F

Leg Stabilizing

Tab

Tail

Crease

Line

Leg

Base leg

Support block

Moving trolleyFixed platform

Leg a

Leg b

Crease

line

B

11

The following process parameters can be set in the developed adhesive applicator:

- open time

- pressure time

- pressure on the joint

- amount of adhesive

- temperature of the applied adhesive (especially relevant for HMAs)

- application speed

In co-operation with IM-Teknik, a prototype laboratory adhesive applicator was built; see Figure 4. Today this equipment is fully developed and commercially available from IM-Teknik, Gothenburg, Sweden. To glue two board surfaces together with an HMA or dispersion adhesive involves several process steps [Paper I], all of which were incorporated into the laboratory applicator process.

Figure 4. Photo of the adhesive applicator, with its various components labelled.

2.2 Y-peel

To produce repeatable test results, the Y-peel equipment was redesigned and adapted to the Y-peel samples. Sample preparation for the Y-peel test method is fully described by Tryding et al. [Paper I]. The Y-peel fixture is set up into the uniaxial tensile tester (Figure 5). During a test, the uniaxial tester pulls the upper clamp with a constant speed and continuously records the resulting forces with corresponding prescribed deformations (maximum 3.0 mm). As the upper clamp moves, the Y-peel specimen stretches, which ultimately leads to a

Transport plate

Hot-melt adhesive nozzle

Press plate

Weights

Dispersion adhesive nozzle

12

fracture in the adhesive joint of the Y-peel specimen. The Y-peel specimen is mounted in hinged upper and lower clamps so that movement is free of friction. The free ends cannot move uncontrolled, because they are fixed in the clamps.

Figure 5. Photo of the fixture with the mounted Y-peel specimen [Paper IV].

Figure 6. The Y-peel set-up together with a sketch of the geometry and boundary conditions. Δ is the displacement of the point of fracture opening during testing.

Creased

line

Adhesive

joint

Rotationally

unrestrained

clamp

Y-peel

specimen

Rotationally

unrestrained

clamp

Adhesive

string Creased

line

Adhesive

joint

Rotationally

unrestrained

clamp

Y-peel

specimen

Rotationally

unrestrained

clamp

Adhesive

string

P

R R

P/2

R

α

Δ

Mb

F

Ft

n

13

The set-up of the Y-peel method is shown in Figure 6 together with sketches of the geometry and boundary conditions. In the middle of Figure 6, the load P is acting on the upper simply supported clamp of the set-up. The resulting reaction forces R at the two simply supported clamps at the lower part of the set-up are of the same magnitude. The forces are all aligned parallel to the specimen legs. Force equilibrium gives the relation

0)2cos(2RP (1)

where is the peel angle; see the middle of Figure 6. Using the symmetry line in the Y-peel set-up gives the reaction forces and moment at the adhesion joint zone; see the right of Figure 6. The reaction forces in the vertical and horizontal directions are denoted tF and nF , respectively. The moment is denoted as bM .

In this set-up, Δ is considered small enough to justify the approximation of

bM ≈0. Force equilibrium in the vertical and horizontal planes then gives, using

eq. (1), the relation

)2/tan(2

and 0P

FF nt (2)

respectively. Note that the reaction force, nF , on the right of Figure 6, is

perpendicular to the board surface (for detailed discussion, [Paper II]).

2.2.1 Force elongation

The force-elongation curves have to be separated into their elastic and dissipative (inelastic opening) regions before they can be analysed in a relevant and objective way; see Figure 7 [Paper II].

14

Figure 7. Typical total force-elongation curves from Y-peel tests and identification

of the elastic component and extraction of the dissipative part [Paper II].

Typical force-elongation curves for a test sample in Y-peel are shown in

Figure 7. The area for extracting the remaining elastic energy is marked remain

elW ;

the dissipative ascending and descending energy are marked asc

DW and desc

DW

respectively [Paper II]). The total elongation is denoted u , whereas the elastic part of u is denoted eu , and the dissipative part of u is derived as the

difference euu . The initial energy caused by initial tightening of the specimen

is called in itial

DW .

The crack propagation begins and ends in the dissipative part of the force-elongation regime, and the fracture development depends on how the crack propagates after its initiation.

2.2.2 Interpretation and failure characterisation

The inelastic behaviour of the adhesive joint can be derived by separating the recorded force-elongation curve into elastic and dissipative parts (Figure 7). The testing method is described in [Paper I] and the mathematical derivations in [Paper II]. This is based on the fact that the proportional limiting yield stress in the in-plane direction is at least 10 times higher than the out-of-plane failure stress [4], [20]. Inelastic deformation is hence confined to out-of-plane deformation of the adhesive joint. Using the equations

F

u

Elastic stiffness

asc

DW

desc

DW

remain

elWinitial

DW

(u-ue)

MaxF

15

2

FFn

and the inelastic opening )(2 e

ie

n uu (3)

[Paper II], [Paper III], the force versus elongation curve can be divided into inelastic force and deformation acting on the HMA joint, where F is the total

measured force and nF is the force acting on the HMA joint. The inelastic

deformation of the adhesive joint ( ie

n) is defined as the inelastic opening of the

adhesive joint in the direction of nF .

2.3 Surface treatment and characterisation

The surface treatment experiment was made to gain knowledge about the effect of changes in surface chemistry or surface roughness on the mechanical strength of the HMA joint [16], [21], [22].

The surface roughness was modified through coating or calendering. Surface coating was either single-/double-pigment coating or surface sizing with starch. In a pilot trial, five different board qualities were produced using the same cartonboard as a coating substrate [Paper IV]. The five board qualities generated six different types of surface qualities, which were combined in an experiment to generate 31 cases, including the important factor of which surface the HMA was applied on and which surface was subsequently pressed onto the glue string [Paper IV].

2.3.1 Roughness characterisation

The surface roughness was characterised by the root-mean-square roughness,

qR unfiltered, since the different surfaces showed a span too wide for the

commonly used Parker Print-Surface method (PPS). The qR value was

measured with white-light interferometry [Paper IV]. The instrument (ZYGO NewView 5010 optical profilometer from Zygo LOT GmbH, Darmstadt, Germany) scans the topography of the surfaces in a non-contact mode [23]. The surface roughnesses in this thesis are given as the root-mean-square value

( qR ). The maximum Z-range of this scanner is 100 μm.

2.3.2 X-ray photoelectron spectrometer

The X-ray photoelectron spectroscopy (XPS) is a surface-sensitive analysis technique. The principle technique is to use X-rays to cause the emission of electrons from the surface substrate and then measure the kinetic energy of the emitted photoelectrons. With this information, the energy of the surface-bound electron can be calculated and related to the chemical state of bonding [24],

16

[25]. The surface chemical composition of the samples was determined with a Kratos AXIS HS X-ray photoelectron spectrometer (Kratos Analytical, UK) using a monochromatic Al Kα X-ray source (1486.6 eV). The XPS spectra were run using a take-off angle of 90°, resulting in information being collected from a roughly 10 nm thick layer. It should be noted that the signal decays exponentially with increasing depth in the material.

Analyses were made on an area approximately 1 mm2 at two different locations on each sample. Surface elemental concentrations and oxygen-to-carbon (O/C) ratios were calculated from the survey spectra. The relative amounts of carbons with different bonds to oxygen were determined from high-resolution carbon C1s spectra using a curve-fitting program from the spectrometer manufacturer [Paper IV].

2.3.3 Dynamical mechanical thermal analysis

Dynamical mechanical thermal analysis (DMTA) can simply be described as applying a periodical prescribed displacement and measuring the force applied to deform the sample. With these dynamical measurements, a storage modulus (storage of energy) and loss modulus (damping and heat generation) can be calculated. The DMTA allows the study of the influence of temperature and frequency on the viscoelastic properties of materials [26].

A dynamical mechanical thermal analyser [27] (DMA/SDTA681, Mettler Toledo, Switzerland) was used to estimate the relevant stiffness variation, which should be used in the finite element (FE) calculations. The shear mode with frequencies from 0.0001 Hz to 1000 Hz was used in these tests in the temperature range –50°C up to 200°C.

2.4 Finite element analysis

The finite element method (FEM) with an implicit solution was used to investigate how the mechanical properties of the adhesive joint affect the Y-peel test. The commercial FE program Abaqus/Standard [28] was used to analyse the Y-peel test. A more detailed description of the modelling can be found in section 3.2.

2.5 Materials

2.5.1 Cartonboard

The cartonboard used in this thesis (in Papers I–III and V) is mostly four-ply cartonboard from Korsnäs, Sweden; see Figure 8. The cartonboard is built up

17

from the bottom ply with unbleached softwood sulphate pulp, and the two middle plies are a mixture of unbleached softwood sulphate pulp, unbleached softwood chemi-thermo-mechanical pulp (CTMP), and internally produced broke from the board machine. The top ply is a mixture of softwood and hardwood bleached sulphate pulp, and on top of that is a pigment-coated layer.

Figure 8. The Korsnäs cartonboard in cross section, both schematically and as a scanning electron microscope micrograph.

In one of the investigations (Paper IV), cartonboard made by a pilot machine was used. The top and back plies of all the boards used in this study were made of bleached chemical pulp, while the middle ply was made of bleached chemi-thermo-mechanical pulp (BCTMP).

2.5.2 Hot-melt adhesives

The HMAs used in the study were mainly Swift B569/38 and Henkel Technomelt Supra 100 [Papers I–III]. In Paper III, other HMAs were also used. In Paper IV, a custom-made HMA from National Adhesives was used. This HMA was composed of high molecular weight ethylene-vinyl acetate, tall oil ester and synthetic wax. In Paper V, the Henkel Technomelt Supra 100 was compared to the FE analysis.

Cartonboard

Coating

Bleached cellulose pulp

Unbleached cellulose pulp and

CTMP

Unbleached cellulose pulp and

CTMP

Unbleached cellulose pulp

18

Figure 9. The storage modulus (G´) vs. temperature. The shear modulus was

measured with dynamic mechanical thermal analysis (DMTA) for a hot-melt adhesive (HMA) (Henkel Technomelt Supra 100) [29].

The variation of the real part of the complex storage modulus with temperature is shown in Figure 9. The curve is obtained by the DMTA of an HMA (Henkel Technomelt Supra 100), at the frequency of 1 Hz, heating rate 9 °C per minute. Figure 9 shows that the storage modulus (G´) for this HMA varies between 300 MPa to about 10 MPa in the temperature range of –45°C to 75°C. Other tested HMAs have shown even bigger variations in the storage modulus, from 1000 MPa down to 10 MPa in the same temperature range [29]. This order of magnitudes has also been reported earlier in other studies [30][31].

3. Theory

The dissipative energy can be computed from the force versus inelastic deformation curve of the Y-peel test. The dissipative energy is caused by the formation of cracks, which in turn is due to the formation of microcracks and the subsequent coalescence of these cracks in the thickness direction of the cartonboard, and by inelastic deformation and damage of the adhesive. Hence, the dissipative energy is a measure of the amount of energy that is consumed during opening of the joint. As can be seen in Figure 7, the dissipative energy can be divided into an ascending and a descending part.

The force equilibrium means that the cartonboard and the tensile test machine show a completely elastic behaviour, while the adhesive joint fails out-of-plane (Table 1). The relative proportions and the magnitudes of the elastic, ascending

G’(M

Pa

)

Temperature ( C)

19

dissipative work, and descending dissipative work [Paper II] make it possible to objectively characterise how suitable a certain adhesive joint is for a particular package application. The ascending part is the first part of the curve that ends at the maximum force ( maxF ); cf. Figure 7. The energy that is consumed during

the ascending part of force versus inelastic deformation is denoted asc

DW . The

descending part of the curve starts at the maximum force and ends where final failure occurs. During the descending part of force versus inelastic deformation, the coalescence of microcracks and fibres pulled out from the cartonboard results in a softening behaviour [4]. Part of the softening behaviour may also be due to the deformation and fracture in the adhesive. In some cases, however, a rapid failure occurs at the maximum force so that the descending dissipative energy is essentially zero. The energy that is consumed during the descending

part of force versus inelastic deformation is denoted desc

DW .

3.1 Fracture, wetting and adhesion

Fracture of the adhesive joint may occur in several different ways, such as:

- cohesive fracture in the adhesive

- interfacial fracture between the adhesive and one of the cartonboard surfaces

- cohesive fracture in the cartonboard

The failure may also be a combination of the above-listed failure modes.

The work of cohesion a

Cw – that is, the surface energy per unit area associated

with cohesive fracture in the adhesive – is equal to a2 , [5], where a is the

surface tension of the adhesive. The work of cohesion in the cartonboard s

Cw

could similarly be expressed as s2 , where s denotes the surface tension of

the cartonboard.

The work of adhesion in the interface between the adhesive and the cartonboard is expressed by the Young-Dupré equation according to

assaAw (4)

Assuming that the surface tensions of the cartonboard and the liquid adhesive in vacuum are approximately the same as the corresponding surface tensions in adhesive vapour [32], then the contact angle could be derived from Young’s equation as

20

a

asscos (5)

The contact angle is the angle between the upper surface of the liquid and the interface between the liquid and the substrate, when a small liquid drop is placed on a horizontal plane substrate. Hence, the contact angle is a measure of wetting of the substrate by the liquid. Complete wetting occurs when

aass )( , leaving 0 . This corresponds to the non-equilibrium situation

when the liquid spreads completely on the substrate. The expression )( ass

corresponds to the empirical critical surface tension c for the substrate [33].

Hence the substrate is completely wetted by the adhesive if c for the substrate

is greater than the surface tension of the liquid adhesive. A combination of eqs. (4) and (5) results in the work of adhesion.

)cos1(aAw (6)

From eq. (6), it is evident that enhanced wetting gives higher work of adhesion, a fact that has also been verified experimentally ([34]). From eq. (6), it may seem as if the work of adhesion is always less than or equal to the work of cohesion of the liquid adhesive. Note, however, that for complete wetting

aass . As pointed out in [5], this implies that a

CaA ww 2 for

complete wetting conditions. This indicates that the fracture in such cases would occur as a cohesive fracture in the adhesive, rather than an interfacial fracture between the substrate and the adhesive. This may be true for ideally brittle conditions, but in most cases, the bulk of the material (the adhesive and the cartonboard) experiences energy dissipation not only due to crack growth but also due to inelastic deformation. To take such effects into account, more complex fracture mechanics must be considered.

Consider two solid sheets of cartonboard joined together with an adhesive. During the progressive creation of free surfaces through crack growth in the board [35], in the adhesive or along the board-adhesive interface, the energy balance for the body requires that

kf WWW (7)

Here denotes the energy put into the system by external forces, W is the

change in strain energy, fW is the total surface energy consumption during

crack growth, and kW is the kinetic energy of the body during crack growth.

The strain energy can be divided into its elastic and inelastic parts, elW and

ieW . The elastic energy is the part of the total strain energy that is recovered

during unloading of the body. Eq. (7) can now be rewritten as

21

kfieel WWWW (8)

The dissipative energy during crack growth is defined as fieD WWW .

Note that inelastic energy dissipation does not result in a change of elastic stiffness of the body as opposed to the energy dissipation associated with crack growth. In most cases, the strain energies vary as a function of position in the

body, which means that elW and ieW are the integrated elastic and inelastic

strain energy densities over the entire volume of the body. For non-homogeneous conditions, crack growth may occur simultaneously at different locations in a body meaning that

asAs

s

Ca

a

Cf AwAwAwW (9)

Here assa AAA and, respectively are the incremental crack surface areas

created in the adhesive, the solids, and at the interface between the solids and the adhesive. The onset and progression of inelastic deformation and crack growth in a material is governed by the condition of forces at the point of the crack and the material’s ability to sustain those forces.

The work of adhesion according to eq. (4) represents the thermodynamic bonding between the adhesive and the cartonboard. High work of adhesion

results if as is small; that is, when there is a favourable interaction of the

materials in the interface. This requires strong bonding across the interface. Most of this bonding energy is due to interactions across the interface, such as Lifshitz-van der Waals and acid-base interactions (cf. [5]). Diffusion and entanglement of polymer molecules into the substrate may increase the adhesion as well as pure mechanical bonding between the adhesive and the substrate. The latter contribution is particularly important for fibrous materials such as cartonboard where mechanical bonding is accomplished by the mechanical interlocking of the adhesive into irregularities in the cartonboard surface and by the embedding of fibres sticking out from the cartonboard surface (see [10], [11]).

If the total bonding across the interface is large enough, the adhesive joint may fail through cohesive fracture in the cartonboard or in the adhesive.

In polymer-based adhesives, such as HMAs, the cohesive strength depends on the molecular structure of the polymer (see [36]). Micromechanically, cohesive fracture occurs by cavitation, fibrillation and crazing (see [37]). These processes involve a large portion of inelastic deformation. Cartonboard is known to have inelastic deformation when subjected to loading in the thickness direction

22

(cf. [38]). This means that cohesive fracture involves energy dissipation due not only to crack growth but also to inelastic deformation.

Cohesive fracture in multi-layer cartonboard either takes place deep in the board structure between the plies of the cartonboard (see [4]) or in the outermost ply adjacent to the adhesive and close to the surface of the cartonboard. The former type of fracture is often referred to as delamination, while the latter is referred to as tearing of the cartonboard. Note that the difference between interfacial fracture and cohesive fracture in the cartonboard (by tearing) is not very distinct.

Interfacial fracture with partial tearing is revealed by the observation that parts of the glue-string area after fracture are covered by fibres. Complete tearing is when 100% of the glue string is covered by fibres, and pure interfacial fracture is when no fibres at all stick to the glue string. The expression “tearing” is normally used when more than 50% of the glue-string area is covered by fibres after a fracture.

The interlaminar strength between the plies of a multi-layer cartonboard depends on how the board is made; for example, how the paper webs are dewatered and dried, as well as whether starch or fines are added to improve the bonding between the board plies. The tearing strength and the bonding strength of fibres of the surface of the cartonboard depend on the pulp composition, pulp quality, refining, added chemicals and a number of other process parameters. Many methods to measure the behaviour and strength of cartonboard in the thickness direction exist, including Z-toughness [39] and Arcan test[38].

All in all, it appears that brittle failure is possible only if the joint fails by pure interfacial fracture. The fact that no fibres are loosened from the cartonboard would then indicate that the mechanical contribution to the interfacial strength is small in such cases. Maximum toughness (that is, dissipated energy) requires cohesive fracture, either in the cartonboard (by delamination or by tearing) or in the adhesive.

3.2 Structure mechanical analysis

3.2.1 Cartonboard model

The cartonboard is modelled as a layered structure, where the layers are described with a non-linear orthotropic continuum mechanical model, and the adhesion between the layers is modelled as a cohesive damage model with softening behaviour ([4], [40]). Each board ply is represented by at least one

23

layer, but in some cases, it may be necessary to divide an individual ply into several layers, especially in cases where the stress state is inhomogeneous [41].

In this study, the material was modelled according to Figure 10: the ply in contact with the adhesive string is divided into two layers, where the adhesive contact side layer, Layer 1, has a thickness corresponding to one to two times the cellulose-fibre thickness in the cartonboard.

Figure 10. Definition of finite element analysis (FEA) layers for a cartonboard model.

3.2.2 Continuum behaviour of plies

The continuum model used was orthotropic elastic-plastic, with a Hill yield surface and isotropic hardening. The Hill yield criterion [42] defines the flow surface as

2

12

2

13

2

23

2

2211

2

1133

2

3322 222)()()()( NMLHGFf (10)

where ij are rotated components of the Cauchy stress tensor, while F, G, H,

L, M and N are constants obtained by material tests in different orientations, where the orientation is indexed 1 in MD, 2 in ZD and 3 in CD.

The cartonboard has been characterised with respect to elastic-plastic properties by Nygårds [43], [44]; the elastic material constants used in this study have been fitted to the experimental curves by using a least-square fitting

15

Ply 4

Ply 3

Ply 2

Ply 1

Layer 5

Layer 4

Layer 3

Layer 1

Layer 2

Perfectlybonded

Cohesivemodel

60

100

250

60

Hot-melt Adhesive

24

model. The elastic modulus in the ZD, E2, was however extrapolated from the stress-strain curves of the compression tests in the regime 5–10 MPa. The

isotropic hardening modulus ( ) was fitted to the MD tension test.

In Paper V, it was assumed, using the data obtained in Paper I, that mainly the MD and ZD properties affect the results. Moreover, most of the ZD behaviour is determined by the cohesive model, since delamination is enabled in the model. Therefore, plasticity mainly develops due to high MD stresses. Used constants for the FE calculations are listed in Table 2.

Table 2. Continuum properties used for the different plies in the cartonboard.

Bottom ply Middle ply Top ply

Elastic constants

1E [MPa] 8800 3200 5900

2E [MPa] 130 160 230

3E [MPa] 3000 1200 2700

12G [MPa] 68 30 60

13G [MPa] 1600 640 1400

23G [MPa] 68 30 60

12 0 0 0

13 0.51 0.47 0.45

23 0 0 0

Yield criterion and hardening

11y [MPa] 66 22 40

22y [MPa] 33 11 20

33y [MPa] 24 10 14

12y [MPa] 4.0 0.66 4.8

13y [MPa] 33 13 28

23y [MPa] 3.0 0.28 3.5

[MPa] 4300 800 2300

3.2.3 Cohesive behaviour of interfaces

The cohesive model was a damage traction-displacement model with exponential softening, which is available in Abaqus/Standard [28]. The initial

elastic stiffness in ZD ( 0

2K ) and the initial elastic shear stiffness in MD ( 0

1K )

and CD ( 0

3K ) were set according to Table 3. This implies that the interface was

significantly stiffer than the continuum model. Therefore the behaviour of the undamaged cartonboard is mostly governed by the continuum model.

25

The condition for damage initiation was expressed with an equivalent yield

surface using the traction components in the normal direction, 2t , and the two

shear directions, 1t and 3t , as

1

2

0

3

3

2

0

1

1

2

0

2

2

t

t

t

t

t

t (11)

provided that 02t (otherwise 2t is set to zero). In the above equation 0

2t , 0

1t

and 0

3t are the tractions needed to initiate damage in their respective direction.

The traction components evolve with the damage (D) as

33

11

2

22

2

)1(

)1(

otherwise,

0f,)1(

tDt

tDt

t

titDt

(12)

where 2t , 1t and 3t are the traction components predicted by initial stiffness

components at the current separations. This implies that the initial stiffness components of the interface are reduced by the factor (1 – D). The effective

separation ( m ) of the interface is defined as

2

3

2

1

2

2m (13)

where 2 is the separation in ZD ( 02 ), and 1 and 3

are the shear

separations in MD and CD, respectively.

Assuming an exponential damage evolution law, the damage could be expressed as

1

0

0max

1

1

111

max

0

e

eD

mf

m

mm

m

m (14)

where max

m refers to the maximum value of the effective separation attained

during the loading history (see Figure 11), and 0

m is the effective separation at

26

damage initiation. f

m is the effective separation at failure and 1 is the damage

evolution parameter.

Figure 11. Interpretation of cohesive law in the case of pure tension in the

thickness direction (ZD).

The necessary material properties for the cohesive model were determined by the experimental data [44]; see Table 3. The effective separation, generated by these material constants and this model, is of the order of some tenths of the fibre length, and this seems reasonable from a physical point of view.

traction

0

m

f

m

0

2)1( KD

effective separation

0

2K

0

2t

2t

2tD

max

m

27

Table 3. Cohesive properties for different interfaces in the cartonboard.

Ply interface Layer interface within the ply

Elastic stiffness components 0

2K / MPa/mm 780 1000

0

1K / MPa/mm 2200 2900

0

3K / MPa/mm 1300 1700

Damage initiation and evolution 0

2t / MPa 1.0 1.3

0

1t / MPa 0.28 0.36

0

3t / MPa 1.0 1.3

0

m

f

m/mm 1.0 1.0

1 12 15

3.2.4 Adhesive

The HMA used for the experiments was Henkel Technomelt Supra 100. For this type of polyolefin-based HMA, the E-modulus can vary from below 100 up to some 1000 MPa in the normal temperature range for packages [30].

The HMA was assumed to behave as an isotropic elastic material in the Y-peel set-up. In the FE calculations, the E-modulus of the HMA was varied from 100 to 1000 MPa [Paper V].

4. Results

The newly designed adhesive applicator and the Y-peel test method have been used in combination to test adhesively bonded cartonboard. The FEM has also been used to investigate and analyse the behaviour of different joints during the Y-peel test.

4.1 Typical failure modes

In Paper III, 310 different samples were tested. It was observed that the force-elongation curves could be grouped into four main types of failure modes: M1–M4. See Figure 12.

28

Figure 12. Characteristic force-elongation curves in Y-peel testing with four main

failure modes.

The modes in Figure 12 are categorised by the shape of the force-elongation curves. Mode M1 specifically has one single peak with an abrupt failure after a very short displacement with no or very little descending dissipative energy, a pop-up failure. Mode M2 has a quick failure at a relatively short displacement, sometimes with several peaks, but it shows a small but measurable amount of dissipative descending energy. Mode M3, on the other hand, does not break completely within the specified maximum 3.0 mm displacement during testing. Several peaks are not uncommon, and mode M3 has a significant proportion of dissipative descending energy. The force is more or less constant during the final part of the force-elongation curve. Finally, mode M4 is similar to mode

Mode Characteristics Example

M1

- single peak

- sudden breakage at peak load

- short displacement before breakage

- no or very little recorded descending dissipative

energy

M2

- single or multiple peaks

- sudden breakage

- relatively short displacement before breakage

- small part of descending dissipative energy

M3


- no breakage within target max displacement

(3.0 mm)

- last part of force-elongation curve shows

approximately constant force

M4


- significant displacement before breakage

- moderate slope after peak

0 1 2 3

0.0

0.5

1.0

1.5

2.0

Path in mm

Force in N

/mm

0,0 0,2 0,4 0,6 0,8

0,0

0,2

0,4

0,6

0,8

1,0

1,2

Path in mm

Forc

e in

N/m

m

0 1 2 3

0,0

0,5

1,0

1,5

2,0

Path in mm

Fo

rce

in N

/mm

0 1 2 3

0,0

0,5

1,0

1,5

2,0

Path in mm

Fo

rce

in

N/m

m

29

M3, with a significant proportion of dissipative descending energy, but after one or several peaks, the force gradually decreases during elongation until complete failure.

The majority of the 310 Korsnäs cartonboard specimens had a mode M3 failure (Figure 13a). Mode M2 was the least frequent failure, and modes M1 and M4 were more or less equally frequent.

Maximal force in the force-elongation curves showed no correlation with the failure modes except for failure mode M1 (Figure 13b), which appeared to have a generally lower maximum peak force with a higher standard deviation. A cause for the higher standard deviation may be that the failure sometimes is completely interfacial but sometimes starts with a small cohesive fracture in the cartonboard. The small difference in maximum peak force between modes M2 to M4 is due to the fact that at this point, in all three failure modes, fibres start to be torn out of the surface of the board. The resistance to this process is the same for all failure modes, since the surface strength of the board is constant in the present investigation.

Only failure modes M3 and M4 had a significant proportion of dissipative

descending energy, desc

DW (Figure 13c). Failure mode M1 had in practice no desc

DW at all.

Figure 13a.

Number of tested Y-peel specimens with different modes of failure. A Henkel adhesive was used.

Figure 13b.

Average Fmax (maximal force) vs. failure mode of 310 tested specimens with the Henkel adhesive.

Figure 13c. desc

DW (descending dissipative

energy) vs. failure mode of 310 tested specimens with the Henkel adhesive.

4.2 Characterisation of failure modes

The force and elongation curves were recorded during all the tests. In some tests, a digital microscope equipped with a video camera (TIMM 400 C, SPI GmbH, Oppenheim, Germany) was used to take pictures of the joint during testing. The picture sequences were automatically synchronised with the force-elongation curve (Figure 14). It was observed that the major crack was initiated through small cracks during the ascending part of the force-elongation curve.

Amount of different failure

0

50

100

150

200

250

M1 M2 M3 M4

Failure mode

Nu

mb

er

of

sp

ec

imen

s

Amount of different failure

0

50

100

150

200

250

M1 M2 M3 M4

Failure mode

Nu

mb

er

of

sp

ec

imen

s

Fmax

0

20

40

M1 M2 M3 M4

Failure mode

(N)

Dissipative descending energy

0

10

20

30

40

50

60

M1 M2 M3 M4

Failure mode

(mJ

)

Dissipative descending energy

0

10

20

30

40

50

60

M1 M2 M3 M4

Failure mode

(mJ

)

30

The small cracks then merged during further elongation into a propagating major crack in the descending dissipative part, which finally led to failure.

Figure 14. Numbered photographs from the formation and propagation of a delamination fracture with corresponding positions numbered on the Y-peel force-

elongation curve. This sample is characterised as a mode M3 fracture with

Fmax =36.8 N, desc

DW =45.4 mJ.

In failure mode M3 (Figure 14), the failure starts with small microcracks in the ascending part of the curve (Pictures 2–4). The microcracks appear between the outer ply and the nearest middle ply of the upper cartonboard in Pictures 2–4. In the descending part of the force-elongation curve, a major crack has developed (Picture 5) in the lower cartonboard. This crack leads to the final failure, after more than 3.0 mm, with fibres and the complete outer ply being torn off from the lower cartonboard (Picture 6). The lower cartonboard surface is the one on which the HMA was applied in the adhesive applicator.

In modes M1 and M2, the microcracks start in the same way as in modes M3 and M4. The difference is that once the microcracks are initiated, they grow

0

15

30

45

0,00 0,50 1,00 1,50 2,00 2,50 3,00

Elongation (mm)

Fo

rce

(N

)

1

4

3

2

5

6

1 2 3

4 5 6

Cartonboard

Adhesive

FibreMajor crack

Microcrack

Final failure Adhesive

1 2 3

4 5 6

Cartonboard

Adhesive

FibreMajor crack

Microcrack

Final failure Adhesive

31

rapidly to form a major crack and the final joint failure in modes M1 and M2. After failure, no fibre tear is visible in mode M1 and only sparsely in mode M2.

Mode M4 has a smooth curve, and the failure starts with the formation of microcracks in the ascending part. The microcracks then merge in the descending part of the curve to form the major crack leading to a final failure, where fibres abundantly have been torn off from either of the cartonboard surfaces.

In this study, the major crack always appeared in the adhesive, in the board ply closest to the adhesive, or at the interface between the outer and middle board ply. The failure never occurred in the middle board plies or at the interface between middle plies.

4.3 Y-peel characterisation and fibre tear

In a trial, the open time parameter in the adhesive applicator was changed over an interval to investigate the effect on the fracture process and how the crack propagated [Paper II]. The idea was to test whether there was a correlation between the traditional method with manual peeling and, on the one hand, subjective evaluation of fibre tear and, on the other, force-elongation curves or failure modes. The test showed that fibre tear increases with increased descending dissipative energy until the strength of the adhesive joint exceeds the internal strength of the fibre structure in the outer ply of the board, leading to 100% fibre tear (the descending dissipative energy decreased with increasing open time [Paper II]); see Figure 15.

32

Figure 15. Average descending dissipative energy vs. per cent samples showing fibre tear. Dashed lines indicate maximum internal strength of the outer plies of

LGT250 (Korsnäs cartonboard with a grammage of 250 g/m2) and CRY425 (Korsnäs cartonboard with a grammage of 425 g/m2).

4.4 Surface treatments

Surface treatment affected the surface roughness and the surface chemistry [Paper IV]. It was difficult to separate the influence from the surface roughness from the surface chemistry, since the surface roughness was affected not only by calendering but also to a large extent by the surface treatment through pigment coating or surface sizing.

The descending dissipative energy’s, dW , dependence on the surface roughness

of Part 1 (on which the HMA is applied) and Part 2 (the cartonboard pressed onto the HMA) is shown in Figures 16a–d.

0

10

20

30

40

50

60

70

80

0 25 50 75 100

LGT250

CRY425

0

10

20

30

40

50

60

70

80

0 25 50 75 100

LGT250

CRY425

Per cent of fibre tear samples (%)

Desc

en

din

gd

issip

ati

ve

en

erg

y(m

J)

33

Figure 16a: The descending dissipative energy

( dW ) in mJ/m vs. surface roughness in μm.

Part 1 double-coated and calendered.

Figure 16b: The descending dissipative energy


Part 2 double-coated and calendered.

In Figure 16a, dW vs. qR is plotted when Part 1 is double-pigment–coated and calendered, and

Part 2 is surface-sized (0sC); single-pigment–coated and calendered (1cC); or double-pigment–coated and calendered (2cC). In Figure 16b, Part 2 is instead double-pigment–coated and calendered, whereas Part1 is single-pigment–coated and calendered (1cC); double-pigment–coated and calendered (2cC); surface-sized, calendered (0sC); or uncalendered (0s).

Figure 16c: The descending dissipative energy


Part 1 double-coated and uncalendered.

Figure 16d: The descending dissipative energy


Part 2 double-coated and uncalendered.

In Figure 16c, Part 1 is double-pigment–coated and uncalendered, and Part 2 is double-pigment–coated, uncalendered (2c); surface-sized, calendered (0sC); or uncalendered (0s). In Figure 16d, Part 2 is double-pigment–coated and uncalendered, and Part 1 is double-coated, calendered (2cC); uncalendered (2c); or surface-sized, uncalendered (0s).

0.00

0.05

0.10

0.15

0.20

0.25

0.30

Wd

calendered

Part 1, double coated

0sC

1cC

2cC

0.0 1.0 2.0 3.0 4.0 5.0

Rq, Part 2

0.00

0.05

0.10

0.15

0.20

0.25

0.30

Wd

calendered


0sC

1cC

2cC

0.0 1.0 2.0 3.0 4.0 5.0

Rq, Part 2

0.0 1.0 2.0 3.0 4.0 5.0

0.00

0.05

0.10

0.15

0.20

0.25

0.30

Rq, Part 1

Wd

calendered


0s

0sC

1cC

2cC

0.0 1.0 2.0 3.0 4.0 5.0

0.00

0.05

0.10

0.15

0.20

0.25

0.30

Rq, Part 1

Wd

calendered


0s

0sC

1cC

2cC

0.0 1.0 2.0 3.0 4.0 5.0

0.00

0.05

0.10

0.15

0.20

0.25

0.30

Rq, Part 2

Wd

not calendered


0s

0sC

2c

0.0 1.0 2.0 3.0 4.0 5.0

0.00

0.05

0.10

0.15

0.20

0.25

0.30

Rq, Part 2

Wd

not calendered


0s

0sC

2c

0.0 1.0 2.0 3.0 4.0 5.0

0.05

0.10

0.15

0.20

0.25

0.30

Rq, Part 1

Wd

not calendered


0s

2c

2cC

0.0 1.0 2.0 3.0 4.0 5.0

0.05

0.10

0.15

0.20

0.25

0.30

Rq, Part 1

Wd

not calendered


0s

2c

2cC

34

From these figures, the conclusion can be drawn that dW was lower in cases

when Part 1 was smooth and Part 2 was rough than for the opposite situation; see, for example, 0sC in Figure 16a and Figure 16b. It therefore indicates which part the adhesive string is first applied on is important.

4.5 Comparison between experiments and modelling

The FEA was performed with a symmetric model of the Y-peel test up to peak load. In the analysis, truss elements, 8-node linear brick incompatible mode elements, and cohesive elements were used, as shown in Figure 17 [Paper V].

Figure 17. a) The symmetric finite element (FE) model of the Y-peel, b) closed up in the hot-melt adhesive (HMA) region and c) showing the mesh close to the edge of the HMA. The cohesive elements are placed in the cohesive regions as labelled

in Figure 10.

Cartonboard

Truss elements

AdhesiveSymmetry contact plane

Continuum elements C3D8I

Cohesive elements COH3D8

a)

b)

c)

Cartonboard

Truss elements

AdhesiveSymmetry contact plane

Continuum elements C3D8I

Cohesive elements COH3D8

a)

b)

c)

35

The stiffness – that is elastic moduli of the HMA – was varied one order of magnitude.

A comparison of the FEA and the measured experimental results shows that the global behaviour (that is, the pre-peak force-elongation curve; cf. Figure 19), and also the more local behaviour (that is, the delamination; cf. Figure 18), is similar to the calculation and the experiments; cf. Figure 19.

Figure 18. The upper image is a finite element analysis (FEA) simulation of delamination with a hot-melt adhesive (HMA) E-modulus 500 MPa. The bottom

image is a picture taken from a film at the same deformation from an experiment. The adhesive parts of each image are facing one another.

The measured Y-peel tests did not show any big differences in the ascending pre-peak part of the force-deformation curve. The global behaviour only shows small differences in the ascending part, in the same manner as the simulations show; cf. Figure 19. When the E-modulus increases, it causes only small differences in the global ascending pre-peak force-deformation curve in the FEA simulations. In Figure 18, it is observed that the delamination is more dominating in the interface between Layer 2 and Layer 3 than the interface between Layer 1 and Layer 2. Locally, variations in the cartonboard can be observed with differences in the stiffness of the HMA. The analysis with a higher E-modulus shows (even if the mesh is rough) higher shear stresses in the board close to the interface. This suggests that interface failure may occur more easily when the E-modulus is high than if it is low. Finally, independently of the HMA stiffness, the σ11-stress field (depicted in Figure 18 for 500 MPa) is almost the same both in location and in value. In combination with the result that the

(MPa)

36

delamination behaves almost in the same way with high and low HMA stiffness, these observations may explain why the global force-elongation curve in the ascending part looks very similar.

Figure 19. Y-peel testing results compensated for the machine compliance. The samples show different final failure modes. Smooth curves shown for calculated

finite element (FE) simulations with different E-moduli [Paper V].

The analytic stiffness calculation results in too high stiffness, but the FE simulation results are closer to the measurement in the ascending part. The initial stiffness for the FE simulation is a bit too high.

5. Discussion

5.1 Y-peel test method

The Y-peel method is an attempt to create a method that objectively measures the mechanical properties in adhesively bonded cartonboard. An adhesive applicator was developed to prepare the samples (the adhesive joints between two cartonboard sheets); this was done to ensure good repeatability. This laboratory adhesive applicator makes it possible to simulate a production machine and independently set the hot-melt sealing parameters: open time, press time, pressure on the joint, amount of applied adhesive, temperature of applied adhesive, application speed and different adhesives. After the sheets

Elongation

Fo

rce

37

have been glued together, the laminate is cut in pieces suitable for the Y-peel test equipment, mounted and then torn apart in the uniaxial tensile tester, in which the force-elongation curve is recorded. Three parameters of the force-elongation curve are identified: an elastic part, an ascending dissipative part and a descending dissipative part. These three parameters are the basis for evaluation and characterisation of the force-elongation curve and the joint.

5.2 Y-peel characterisation

From the force-elongation graphs, it is possible to draw conclusions about when and how the crack propagates in the specimen.

- The maximum measured force gives information about when the major crack and the softening behaviour start.

- The dissipative descending part gives information about how the major crack propagates and how the failure behaves.

- The ascending dissipative part has so far not shown any correlation with fracture characteristics or crack propagation, probably because the microcracks start in the same way in all tested samples in this study.

In this study, a correlation between dissipative descending energy and the concept of fibre tear, commonly used in the converting industry, has been found. Therefore, the dissipative descending energy can be used as a practical assessment of the mechanical strength of an adhesive joint, a value that can be used by the converter to set the appropriate level. The maximal force Fmax, on the other hand, gives information about when a crack in the joint is initiated, but it does not show any obvious correlation with fibre tear.

The Y-peel method in combination with the adhesive applicator is an objective method to measure and characterise when fracture is initiated and how the joint behaves in specific applications in a way that correlates well with practical experience (for example, fibre tear).

5.3 Y-peel failure

It appears that brittle failure (that is, fracture without inelastic energy dissipation) would be possible only if the joint fails by pure interfacial fracture. The fact that no fibres are loosened from the cartonboard would indicate that the mechanical contribution to the interfacial strength is small in such cases. To explore this matter, consider the brittle case of mode M2 with multiple peaks as in Figure 20. The fracture initiates as an interfacial pop-up failure reflected by the first two peaks of the force-elongation curve. Consider the second peak of

38

the curve. From the video capture, the pop-up crack growth was estimated to be 0.3 mm, and the width of the specimen is 25 mm. The grey areas in Figure 20 represent the elastic energy released during the pop-up crack growth.

This is estimated to elW = -3.2 mJ. Tests made with controlled displacement,

as in the Y-peel test, led to 0 during fracture. The kinetic energy can be neglected because quasi-static conditions are restored after a small amount of crack growth. Combining eqs. (8) and (9) then gives

0asAc

c

Ca

a

Cieel AwAwAwWW (15)

Figure 20. The release of elastic energy during the pop-up crack growth.

The inelastic part of the dissipated energy corresponds approximately to the

light grey area in Figure 20. This estimate gives 65.0ieW mJ. No crack

growth is visible in the adhesive, that is, 0aA , and there is no visible crack

growth taking place in the cartonboard during the pop-up fracture, implying

0cA . The surface created at the interface due to the pop-up crack growth is 2mm5.7253.0asA . Inserting this into the above equation gives

0asAieel AwWW (16)

Solving for Aw , based on experimental data, gives Aw = 340 J/m2.

Introducing Zisman’s [33] empirical surface energy c into eq. (4) gives

caAw . Assuming that the value of the critical surface energy of cellulose

Forc

e (

N)

Elongation (mm)

50

25

0

0 0.5 1.0 1.5 2.0 2.5 3.0

Forc

e (

N)

Elongation (mm)

50

25

0

0 0.5 1.0 1.5 2.0 2.5 3.0

39

is c =40 mJ/m2 [45] and the value of the surface energy of cold consolidated

HMA is a =30 mJ/m2 [46] gives wA = 0.070 J/m2. There is apparently a huge

discrepancy between the surface energy derived from mechanical testing compared to the surface energy derived from thermodynamic calculations, even in cases behaving macroscopically “brittle”. This is in line with the conclusion by Gent and Schultz [34], who stated that enhanced wetting gives higher work of adhesion. Gent and Schultz [34] recognised that the experimentally measured work of adhesion was much higher than the calculated thermodynamic work, but there was still a correlation. A cause for the correlation between the thermodynamically calculated work of adhesion and the experimentally measured work of adhesion may be that a better wetting of the cartonboard may give larger interdiffusion and more interlocking with fibres and fibrils [22].

5.4 Finite element study

This FE study of the behaviour of the joint failure caused by differences in the elastic modulus shows that the maximal force and the stiffness of the complete test does not change on a global point of view. To see differences, it is necessary to consider local effects. A comparison of differences in stiffnesses shows more shearing effect when the elastic modulus is high than if it is low. This effect may be one reason why there is often fibre tear when the open time is short and more brittle behaviour when it is long [Paper II]. If the joint is cold, it behaves more brittlely than if it is warm [21].

Comparison between the FEA and the experimental measurements shows that the calculation result in the ascending part of the force-elongation curve is close to the experiment result. The initial high stiffness could be assumed to be caused by differences in the HMA stiffness, but the simulation shows very small differences, depending on the HMA stiffness, which indicates that other factors – such as microcracks in the initial part – cause this small deviation between the global stiffness in the measurement compared to the calculations. Finally the FEM can easily take into account the viscosity change of the HMA, since it can be seen as a lower E-modulus. With a temperature change, the E-modulus varies. This variation may cause a new stress situation with more or less shearing; for example, cold conditions may more easily cause interfacial failure for the same load [47].

5.5 Demands on the adhesive joint

The Y-peel test, in combination with the developed adhesive applicator, makes it possible to design the performance of an adhesive joint in cartonboard packages or develop the cartonboard surface to match the adhesive for specific applications.

40

If the demand on an adhesive joint is a small amount of dissipative energy and if the joint should open easily with little damage to the cartonboard surface, then a seal with a force-elongation curve, as in fracture mode M2, should be the aim. This requires a combination of a relatively strong cartonboard surface and a matching surface energy of the adhesive that permits the fracture to propagate at the interface between the board and the adhesive.

On the other hand, if it is important that the adhesive joint is tamper-proof with a relatively low force needed to open it, then a seal with a force-elongation curve such as the one for fracture mode M3 should be developed. This in turn requires a relatively weak and porous board surface structure with matching viscosity and surface energy of the adhesive during application to favour good wetting and mechanical interlocking of the board surfaces. This forces the crack to propagate into the board structure and damage the surface. Moderation of the force to open the joint is achieved by, for example, changing the amount of adhesive applied or the temperature of the adhesive during application, shortening the opening time, chemically modifying the adhesive itself, or lowering the internal strength of the paper board.

6. Conclusions

The Y-peel method combined with the adhesive applicator opens up new possibilities for converters to define their demands on adhesive joints quantitatively and in relevant theoretical terms. It also gives the converters possibilities to optimise the converting parameters in the adhesive-application process in cases of a specific combination of adhesives and cartonboard, and for a specific package application. Furthermore, the technique gives possibilities to develop the adhesive and/or cartonboard to match a specific application, with the intention to obtain joints that work well throughout the entire life of the package.

A higher surface smoothness on the cartonboard that is pressed against (Part 2) the HMA results in higher toughness of the HMA joint. It is possible for the converter to decide, in the adhesive-application step, how tough the joint should be by where the adhesive is applied on the cartonboard (on the rough or the smooth part).

The fact that the FEA results are close to the experimental results shows that the method (with the used model for the cartonboard and the HMA) can be used to satisfactorily predict the behaviour of the adhesive joint. The difference between different E-moduli shows only small deviations in the ascending part of the force-elongation curve, but the local stress situation can be different,

41

which probably causes interfacial failure more easily for higher E-modulus HMA than for lower E-modulus HMA.

7. Future research

There are a number of areas that could be interesting for future research to examine. First, it could be tested whether a refined FE model could be used to calculate the dissipative descending energy. Another interesting possibility would be to find out whether the Y-peel method could be used to gain a better understanding of which physical properties and process parameters have the biggest influence on the mechanical behaviour of an adhesive joint, and how a joint can be optimised for a specific package application by deciding its material and converting properties. Yet another area to explore is the analysis of different converting properties – for example, pressure and pressure time (on the HMA joint when the joint is made) – for the HMA joint in cartonboard packages in correlation with the Y-peel results. Also, it would be fascinating to analyse the mechanical demands on the HMA joint in different steps in the complete package chain. Further it should also be of interest to use the Y-peel equipment to investigate dispersion adhesion joints in cartonboard packages and to analyse the behaviour with the FEM.

The Y-peel could perhaps also be used to measure barrier coating adhesion – for example, polyethylene adhesion – on cartonboard. The force-elongation curve for the barrier coating adhesion measurement in Y-peel tests could be examined to understand what happens in the interface between the barrier coating and the cartonboard when the failure initiates. This information could help us understand and characterise the barrier coating adhesion on cartonboard packages. Lastly, the different converting operations and each of their effects on the coating adhesion could be analysed; for example, flame and ozone treatment for PE adhesion.

8. Acknowledgements

This work of research has been financially supported by Sveaskog and Korsnäs. The research was initiated and supported by my previous manager Dr Johan Tryding at former AssiDomän Frövi with continued support from my current manager at Korsnäs, Dr Ola Karlsson. Dr Johan Tryding gave me inspiration, and I truly appreciate the interest he showed my work as a coach, co-writer and project member all the way. Dr Ola Karlsson supported me and encouraged me, particularly at the end of this work. I especially want to thank Prof Lars Ödberg for all his encouragement, great advice, theoretical support and patience while answering all of my “silly” questions.

42

Thanks to my supervisors, examiners and co-writers Prof Gunnar Engström, Assoc Prof Nils Hallbäck, Assoc Prof Magnus Lestelius and Dr Christophe Barbier who introduced me to the world of research. Their joy and enthusiasm to help me out, no matter the time of day or day of the week, was highly appreciated. Thank you for all the encouragement, discussions, support and laughter.

I also want to thank the people who were involved in the design and technical development of the adhesive applicator and the Y-peel method: Mr Thomas Neumann, Mr Johan Granat, Mr Åke Eriksson and Mr Stefan Andreasson. Mr Roogher Johansson and Mr Jonas Jarhäll supplied the HMA for development and experiments. Ms Inga-Lill Jansson did a great portion of the Y-peel testing, and Ms Carita Diedrichs showed me in scanning electron microscope studies what happens when an HMA joint fails.

All my colleagues at Karlstad University really made me feel like part of the research team. You supported me on both a philosophical and a scientific level; our discussions were both serious and light-hearted. You gave me confidence and support to finish my thesis even when unpredictable things happened.

Dr Torbjörn Widmark helped me as I wrote the thesis and made me see different nuances in the text, for which I am truly grateful. Thank you for your constant interest. Also my grateful thanks to Ms Jennifer Palley who helped me polish the English.

Thanks to all my colleagues at Korsnäs who supported and encouraged me in my work.

My special gratitude goes to my family. My wonderful wife, Sofia, who was always there for me and encouraged me when it was tough. Without her great understanding and support, I would not have been able to finish this thesis. My beloved sons, Evald and Edvin, who never understood why their dad had to spend so much time working in front of the computer. Evald started to find adhesive strings almost as fun as his dad does. Edvin quickly made me think of other matters. I especially appreciate the support and understanding I have had from my mum, my dad and my mother-in-law. They were always very interested in my work and how it proceeded. My brother always supported me, believed in me and encouraged me in my work.

43

9. References

1 Å. Andersson, Nästan hela förpackningshistorien, Nord Emballage 10: 25–55 (2004)

2 M. Löfgren, The leader of the pack: A service perspective on packaging and

customer satisfaction, PhD thesis, Karlstad University (2006) 3 D. Judd, B. Aalders and T. Melis, The silent salesman: Primer on design,

production and marketing of finished package goods, Continental Press, Singapore (1989)

4 Q. S. Xia, Mechanics of inelastic deformation in paperboard, PhD thesis,

Massachusetts Institute of Technology (2002) 5 J. C. Berg, The importance of acid-base interactions in wetting,

coating, adhesion and related phenomena, Nordic Pulp and Paper Research Journal 8 (1): 75–85 (1993)

6 E. F. Eastman and L. Fullhart Jr., Polyolefin and ethylene

copolymer-based hot melt, Handbook of adhesives (3rd ed.), I. Skeist (ed.), Van Nostrand Reinhold, New York (1990)

7 J. J. Latimer, A study of adhesive penetration and cylinder board

surface characteristics as related to gluability, Tappi 61 (5): 37–41 (1978)

8 H. Edquist, Lim, Emballage Förpackningshandbok V, Thorsten Fahlskog (ed.), Vällingby, 83-87 (1997)

9 R. Joukio and S. Mansikkamäki, Papermaking science and technology,

vol. 12, Fapet Oy, Helsinki: 239 (1998) 10 J. J. Bikerman, Adhesion to fibrous materials, Tappi 44 (8): 568–

570 (1961) 11 R. Pelton, W. Chen, H. Li and M. R. Engel, The peeling behaviour

of pressure sensitive adhesives from uncoated papers, Journal of Adhesion 77 (4): 285–308 (2001)

12 M. Edin, L. Ödberg and J. Sterte, Hot melt adhesion of liner sized

with alkylketene dimer, Nordic Pulp and Paper Research Journal 17 (4): 395–400 (2002)

44

13 A. J. Kinloch, C. C. Lau and J. G. Williams, The peeling of flexible laminates, International Journal of Fracture 66: 45–70 (1994)

14 R. J. Dickson and P. Lepoutre, Mechanical interlocking in coating

adhesion to paper, Tappi Journal 50 (11): 149–157 (1997) 15 P. Lepoutre, K. H. Huynh and A. A. Robertson, The setting of

water soluble adhesives on paper – effect of adhesive thickness and of pressure, Svensk papperstidning 83 (10): 520–524 (1980)

16 S. W. Albrecht, Study of corrugated liner paper and bondability by

hot-melt adhesives, Tappi Journal 80 (9): 190–192 (1997) 17 J. R. Huntsberger, Interpreting peel tests, Treatise on adhesion and

adhesives 6: R. L. Patrick (ed.): 9–32 (1989) 18 B. Persson, Finite Element Modelling and Testing of Crack Propagation in

Package Seals. MSc Thesis, Lund University (1997). 19 R. P. Wool, Polymer interface: structure and strength, Hanser/Gardner

Publication, New York (1995) 20 Q. S. Xia, M. C. Boyce and D. M. Parks, A constitutive model for

the anisotropic elastic-plastic deformation of paper and paperboard, International Journal of Solids and Structures, 39: 4053–4071 (2002)

21 M. Viljanmaa, A. Södergård and P. Törmälä, Adhesion properties

of lactic acid based hot melt adhesives and storage stability in different packaging application, International Journal of Adhesion & Adhesives 22: 447–457 (2002)

22 D. E. Packham, Surface topography and adhesion, International

Journal of Adhesion and Adhesives 23, 437–448 (2003) 23 R. Seppänen and M. Sjöberg, Optimizing the use of hot melt adhesives for

packaging, YKI, Institute for Surface Chemistry, Report B1018, Stockholm (2008)

24 G. M. Dorris and D. G. Gray, The surface analysis of paper and

wood fibres by ESCA (electron spectroscopy for chemical analysis), Cellulose Chemistry and Technology., 12: 9–23, 721–734, 735– 743 (1978)

25 T. E. Conners and S. Banerjee (eds.), Surface analysis of paper, CRC

Press, Boca Raton (1995)

45

26 K. P. Menard, Dynamical mechanical analysis: A practical introduction,

CRC Press, Boca Raton (1999) 27 Dynamic mechanical analysis sets new standards DMA/SDTA861,

Mettler Toledo GMBH, Switzerland (2002) 28 Hibbitt, Karlsson and Sorensen, ABAQUS user’s manual, version 6.7,

Hibbit, Karlsson and Sorensen Inc., Providence, (2006) 29 C. Korin and F. Thuvander, DMTA analysis of hot melt adhesives,

Manuscript in preparation (2009) 30 Y. J. Park, H. S. Joo, H. S. Do and H. J. Kim, Viscoelastic and

adhesion properties of EVA/tackifier/wax ternary blend systems as hot-melt adhesives, Journal of Adhesion Science and Technology 20 (14): 1561–1571 (2006)

31 B. W. Foster and A. E. Barret, Characterizing hot-tack differences

in hot-melt adhesives using DMTA, Tappi Journal 74 (9): 135–138 (1991)

32 M. E. Schrader, Effect of adsorbed vapour on liquid-solid

adhesion, Contact Angle, Wettability and Adhesion (A. A. Donskoi, M. A. Shashkina, and G. E. Zaikov, eds.) 3: 67–91 (2003)

33 M. K. Bernett and W. A. Zisman, Relation of wettability by

aqueous solutions to the surface constitution of low energy solids, Journal of Physical Chemistry 63 (8): 1241–1246 (1959)

34 A. N. Gent and J. Schultz, Effect of wetting liquids on the

strength of adhesion of viscoelastic materials, Journal of Adhesion 3 (4): 281–294 (1972)

35 A. A. Griffith, The phenomena of rupture and flow in solids,

Philosophical Transactions of Royal Society, A221: 163–173 (1921) 36 I. Skeist and J. Miron, Introduction to adhesives, Handbook of

adhesives (3rd ed.), I. Skeist (ed.), Van Nostrand Reinhold, New York (1990)

37 H. J. Sue, Craze-like damage in a core-shell rubber-modified epoxy

system, Journal of Materials Science 27 (11): 3098–3107 (1992)

46

38 N. Stenberg, C. Fellers and S. Östlund, Measuring the stress-strain properties of paperboard in the thickness direction, Journal of Pulp and Paper Science 27 (6): 213–221 (2001)

39 A. Lundh and C. Fellers, The Z-toughness method for measuring

the delamination resistance of paper, Nordic Pulp and Paper Research Journal 16 (4): 298–305 (2001)

40 M. Nygårds, N. Hallbäck, M. Just and J. Tryding, A finite element

model for simulations of creasing and folding of paperboard, Proceeding at the 2005 ABAQUS Users’ Conference (2005)

41 N. Hallbäck, O. Girlanda and J. Tryding, Finite element analysis of

ink-tack delamination of paperboard, International Journal of Solids and Structures 43 (5): 899–912 (2006)

42 R. Hill, The mathematical theory of plasticity, Oxford University Press,

Oxford (1950) 43 M. Nygårds, Experimental techniques for characterization of

elastic-plastic material properties in paperboard, Nordic Pulp and Paper Research Journal 23 (4): 432–437 (2008)

44 H. Huang and M. Nygårds, Two dimensional finite element analysis of

paperboard creasing, KTH Royal institute of technology, Report 480 (2009)

45 L. Wågberg and G. Annergren, Physico-chemical characterisation

of papermaking fibers, Transactions of the 11th Fundamental Research Symposium at Cambridge, Pira International, Leatherhead, Volume 1: 1 (1997)

46 M. F. Tse, Hot melt adhesive model: interfacial adhesion and

polymer/tackifier/wax interactions, Journal of Adhesion 66 (1–4): 61–88 (1998)

47 A. Jansson and T. Björklind, Klistring av kartong, MSc Thesis,

Örebro University (2003)

47

10. Notations

HMA Hot-melt adhesive FEA Finite element analysis MD The machine direction; the direction in which the

board is produced on the board machine CD The cross-machine direction; the direction in which

the board is produced on the board machine ZD The out-of-plane direction; that is, the thickness

direction P The load acting on the upper simply supported clamp R The reaction force at the two simply supported clamps

at the lower part of the set-up

The peel angle

tF The reaction forces in the vertical direction

nF The reaction forces in the horizontal direction, normal

to the board surface

bM The bending moment of the Y-peel specimen

Δ The distance from the HMA to the crease u The total elongation

eu The elastic part of u

ie

n The inelastic deformation of the adhesive joint; the

inelastic opening of the adhesive joint in the direction

of nF

maxF The maximum force in the Y-peel test

remain

elW The remaining elastic energy

48

asc

DW The energy that is consumed during the ascending part

of force versus inelastic deformation

desc

DW The energy that is consumed during the descending

part of force versus inelastic deformation

initial

DW The initial energy caused by the initial force

PPS Parker print surface roughness

qR Root mean square surface roughness

XPS X-ray photo electron spectroscopy O/C Oxygen to carbon ratio, determined from XPS C1s XPS signal from 1s electron in carbon DMTA Dynamical mechanical thermal analysis FE Finite element FEM Finite element method (B)CTMP (Bleached) Chemi-thermo-mechanical pulp

a

Cw The work of cohesion in the adhesive

a The surface tension of the adhesive

s

Cw The work of cohesion in the cartonboard

s The surface tension of the cartonboard

as The surface tension of the interface

The contact angle

c The empirical critical surface tension

The energy put into the system by external forces

W The change in strain energy

49

fW The total surface energy consumption during crack

growth

elW The elastic part of the strain energy

ieW The inelastic part of the strain energy

aA The incremental crack surface areas created in the

adhesive

sA The incremental crack surface areas created in the

solids

asA The incremental crack surface areas created in the

interface between the solids and the adhesive

ij The rotated components of the Cauchy stress

tensor

F, G, H, L, M and N constants obtained by material tests

iE The elastic modulus in the principal directions

ijG The shear modulus in the principal directions

ij The contraction in the principal directions

The hardening modulus

0

2K The initial elastic stiffness in ZD

0

1K The initial elastic shear stiffness in MD

0

3K The initial elastic shear stiffness in CD

2t The traction component in the normal direction

1t and 3t The traction component in the shear direction

0

1t , 0

2t and 0

3t The damage initiation traction in their direction

50

1t , 2t and 3t The traction components predicted by initial stiffness components at current separation

D The damage

m The effective separation

2 The separation in ZD

1 and 3 The shear separation in MD and CD

max

m The maximum value of the effective separation

0

m The effective separation damage initiation

f

m The effective separation at failure

1 The damage evolution parameter used in the

calculation of D M1, M2, M3 and M4 The failure modes

Karlstad University StudiesISSN 1403-8099

ISBN 978-91-7063-271-6

Mechanical Behaviour of Adhesive

Joints in Cartonboard for Packaging

A cartonboard package is often sealed and closed with an adhesive – either a hot-melt adhesive (adhesives that are applied in a molten state on the cartonboard) or a dispersion adhesive (adhesives that are applied as water-based dispersions). This thesis focuses on the process of hot-melt gluing, and how material properties and process conditions affect the performance of the adhesive joint.

The primary interest of this study has been to find an objective method that can characterise the adhesive joint – that is, its strength and joint characteristics. The work has principally concentrated on physical experiments where the Y-peel method has been evaluated and further developed, including the construction of a laboratory adhesive applicator. To understand the details of the mechanical behaviour, the finite element method has been used to simulate the ascending part of the force-elongation curve.

The new laboratory adhesive applicator and finite element method can be used to objectively design the interaction between the adhesive and the cartonboard for a specific application. This can be achieved by modifying the cartonboard, the adhesive or the gluing process parameters.

Mechanical Behaviour of Adhesive Joints in Cartonboard for ...241381/FULLTEXT01.pdf · hot -melt...

Documents

Transcript of Mechanical Behaviour of Adhesive Joints in Cartonboard for ...241381/FULLTEXT01.pdf · hot -melt...