The appearance of a print is affected by Where did the ink...

VIPP VALUES CREATED IN FIBRE-BASED PROCESSES AND PRODUCTS


VIPPThe industrial graduate school. Industry and academia in synergy for tomorrow’s solutions.

Print & layoutUniversitetstryckeriet, Karlstad 2018

Cover photo:Hans Christiansson

The appearance of a print is affected by the individual ink layers, and when the

ink is unevenly distributed it will bring negative consequences. As an extension,

print quality can affect how packaging is perceived and may even create ideas

about the product inside the packaging. Therefore, it is important to understand

the reasons behind poor and uneven print and to ensure that the board substrate

provides a stable quality.

This thesis puts focus on how the absorbency of a coated substrate impacts the

ink distribution in flexographic printing. Both monochrome and multicoloured

areas are dealt with. To generate an understanding of these interactions, two new

measurement techniques were developed where the non-uniformities in liquid

absorption and ink trapping can be characterised. Surface roughness often has a

strong influence on the print and this may hide interactions from ink absorbency.

Therefore, the materials used here were carefully selected (or designed) to highlight

the way liquid uptake affects the ink layers. Interactions between substrate and

printing ink were studied in several ways, from monochrome printing at 0.5 ms-1 in

laboratory-scale to multicolour printing at 10 ms-1 in production-scale.

Sofia Thorman is working at RISE Bioeconomy, Stockholm, where she started in 2007. Her main focus is on print quality related issues and ways to characterise and predict printability of paper and board substrates. Interest in the science of graphical arts started when obtaining a Bachelor of Science in Graphic Technology from Dalarna University, Sweden, in 2003 and after that getting acquainted with research and development as a trainee at Stora Enso.

Sofia Thorman

Where did the ink go?The effect of liquid absorption on ink distribution in flexography

So

fia T

horm

anW

here did the ink go?

KAU.SE/EN/VIPPDOCTORAL THESIS | Karlstad University Studies | 2018:52

ISSN 1403-8099

ISBN 978-91-7063-987-6 (pdf)

ISBN 978-91-7063-892-3 (print)


Sofia Thorman

Sofia Thorm

an | Where did the ink go? | 2018:52

Where did the ink go?

The appearance of a print is affected by the individual ink layers, and when the ink is unevenly distributed it will bring negative consequences. As an extension, print quality can affect how packaging is perceived and may even create ideas about the product inside the packaging. Therefore, it is important to understand the reasons behind poor and uneven print and to ensure that the board substrate provides a stable quality.

This thesis puts focus on how the absorbency of a coated substrate impacts the ink distribution in flexographic printing. Both monochrome and multicoloured areas are dealt with. To generate an understanding of these interactions, two new measurement techniques were developed where the non-uniformities in liquid absorption and ink trapping can be characterised. Surface roughness often has a strong influence on the print and this may hide interactions from ink absorbency. Therefore, the materials used here were carefully selected (or designed) to highlight the way liquid uptake affects the ink layers. Interactions between substrate and printing ink were studied in several ways, from monochrome printing at 0.5 ms-1

in laboratory-scale to multicolour printing at 10 ms-1 in production-scale.

DOCTORAL THESIS | Karlstad University Studies | 2018:52

Faculty of Health, Science and Technology

Chemical Engineering


ISSN 1403-8099

ISBN 978-91-7063-987-6 (pdf)

ISBN 978-91-7063-892-3 (print)



Sofia Thorman

LICENTIATE THESIS | Karlstad University Studies | 2014:64

Åsa Nyflött

Structural Studies and Modelling of Oxygen Transport in Barrier Materials for Food Packaging

Print: Universitetstryckeriet, Karlstad 2018

Distribution:Karlstad University Faculty of Health, Science and TechnologyDepartment of Engineering and Chemical SciencesSE-651 88 Karlstad, Sweden+46 54 700 10 00

© The author

ISSN 1403-8099

urn:nbn:se:kau:diva-69984

Karlstad University Studies | 2018:52

DOCTORAL THESIS

Sofia Thorman

Where did the ink go? - The effect of liquid absorption on ink distribution in flexography

WWW.KAU.SE

ISBN 978-91-7063-987-6 (pdf)

ISBN 978-91-7063-892-3 (print)

Knowledge rests not upon truth alone, but upon error also

- Carl Jung1

1 source: www.calgaryjungsociety.org

i

Abstract

The appearance of a print is affected by the individual ink layers. If the ink is

unevenly distributed in inked areas on the substrate, it lowers the quality. This

thesis puts focus on how the liquid absorbency of a coated substrate impacts on

the ink distribution in flexographic printing. Both monochrome and

multicoloured areas are dealt with. It is well known that a smooth surface

increases the chances of a uniform print, whereas the influence from an uneven

absorption is not established and has even been difficult to measure. If the ink is

applied directly onto the substrate, or as an overprint onto already present ink

layers, the outcome is even more complex. Ink trapping behaviour affects the

uniformity of overprint layers. As of yet, this has been largely overlooked in

flexographic printing.

This work includes several situations, from monochrome printing at 0.5 ms-1 in

laboratory-scale to multicolour printing at 10 ms-1 in production-scale. These

studies showed that the ink absorption interacted directly with monochrome ink

layers and that pore-structures with larger pores and greater liquid uptake

generated more homogeneous prints. The tolerance varied between samples for

how uneven pore-structure, and thereby absorption, could be allowed without

severe impacts on print mottle.

In multicolour printing, the overprint layer interacted directly with the preceding

ink and indirectly with the absorbency (rate and uniformity) of the substrate.

Overprint layers became thicker when the first ink layer was thinner and,

consequently, they became uneven when the first layer was uneven. Additionally,

the time between the applications of the two layers was important. When

immobilisation of the first ink was too slow or uneven, it disturbed the ink

trapping so that the overprint layer became irregular.

Output from this thesis project offers a palette of new tools to use when studying

liquid absorbency and its impact on print quality: a) an experimental approach to

generate a “property matrix” in which the influence of uneven absorption can be

separated from surface roughness, b) an aqueous staining technique to

characterise absorption non-uniformity, and c) a scanner-based technique to

characterise ink trapping non-uniformity.

Additionally, a limited study within this project deals with water/moisture

interference in offset printing. Reasons for print mottle problems, related to a

non-uniform absorption of fountain solution, was investigate with the new

staining technique.

ii

List of publications

I. Thorman, S., Ström, G., Hagberg, A. & Johansson, P. A. (2012).

Uniformity of liquid absorption by coatings-Technique and impact of coating

composition. Nordic Pulp & Paper Research Journal, vol. 27 no. 2, pp. 456-

465.

II. Thorman, S., Yang, L. & Hagberg, A. (2013). Simultaneous determination

of absorption mottle and white-top mottle in the same area on coated boards. In:

Enlund, N. & Lovreček, M. (eds.) Advances in Printing and Media

Technology, vol. XL. Proceedings of: 40th International Research

Conference of iarigai, 8-11 Sep. 2013, Chemnitz, Germany. Darmstadt:

iarigai. pp 225-232.

III. Thorman, S., Yang, L., Hagberg, A. & Ström, G. (2018). The impact of

non-uniform ink absorption on flexographic print mottle. Journal of Print Media

Technology Research, vol. 7, no. 1, pp. 7-18.

IV. Thorman, S., Ström, G. & Gane, P. (2018). Impact of non-uniform water

absorption on water-interference print mottle in offset printing. Nordic Pulp & Paper

Research Journal, vol. 33, no. 1, pp. 150-163.

V. Thorman, S., Yang, L. & Lestelius, M. (2018). Mechanism and

characterisation of ink trapping defects in conventional multicolour printing.

(Manuscript)

VI. Thorman, S., Yang, L. & Lestelius, M. (2018). Influence of board properties

on the flexographic ink distribution in single and multicoloured prints. (Manuscript)

Related publications:

Thorman, S. & Yang, L. (2016), Impact of individual Board Properties on Flexographic

Print Mottle. In: 14th TAPPI Advanced Coating Symposium, 4-6 Oct 2016,

Stockholm, Sweden. Norcross: TAPPI.

Thorman, S. & Yang, L. (2015) Critical evaluations of liquid absorption testing methods

for package printing. In: Gane P. & Ridgway C. (eds.) Advances in Printing and

Media Technology. Proceedings of: 42nd International Research Conference of

iarigai, 6-9 Sep. 2015, Helsinki, Finland. Darmstadt: iarigai. pp. 139-145.

iii

The thesis author’s contribution to the papers

Paper I: Experimental design and set-up, entire data analyse, and main part of the

writing was made by Sofia Thorman. Laboratory work was carried out by Anni

Hagberg.

Paper II: Planning of trials, data analysis and main part of the writing was made

by Sofia Thorman. Laboratory work was handled by Anni Hagberg and method

development was made jointly by Li Yang, Hans Christiansson and Sofia

Thorman.

Paper III: Experimental design and set-up, entire data analyse, and main part of

the writing was made by Sofia Thorman. Laboratory work was mainly handled

by Anni Hagberg.

Paper IV: Sofia Thorman contributed with absorption non-uniformity and burn-

out characterisations of samples, which had already been produced, printed and

analysed regarding print quality within a separate PhD project (Hajer Kamal

Alm). Analysis of print quality with respect to the new characterisations was made

by Sofia Thorman and writing was done jointly.

Paper V: Formulation of research questions, method development, experimental

work, entire data analyse, and main part of the writing was made by Sofia

Thorman. Cupper contents were determined by Chatleen Karlsson. Samples

were produced and printed within a separate project, where Sofia Thorman

participated in planning and data analysis.

Paper VI: Formulation of research questions and, in addition to, experimental

work from Paper V, Sofia Thorman contributed with characterisations of

absorption non-uniformity, entire data analyse and main part of the writing.

Characterisations of the substrates were made within a separate project, where

Sofia Thorman participated in planning and analysis, but did not carry out the

experimental work.

Parts of the background chapter of this thesis have previously been published in

a licentiate thesis by the same author.

iv

Acknowledgements

I am truly thankful for the support I have had on the way and would like to

express my deep gratitude to my supervisors, who have been with me along this

journey: Associate Professor Li Yang, you inspire me to do better and I truly

appreciate our discussions and your positive spirit. Associate Professor Göran

Ström, you helped me getting off to a good start and then offered your guidance

and support, even after retirement, thank you. Professor Magnus Lestelius, I’m

thankful for your help and I believe that your input has sharpened my arguments.

My grateful thanks also go to Associate Professor Anita Teleman who initiated

this project and ensured that it could be finished. Earlier, I have had the pleasure

of working together with Dr Per-Åke Johansson and Dr Petter Kolseth, thank

you for getting me hooked on science.

I wish to extend my thanks to my colleagues, especially Anni Hagberg and Hans

Christiansson. Without our teamwork, this thesis would be short of many results

and ideas, and the journey would not have been as enjoyable. To my fellow PhD

candidates within VIPP and at Innventia/RISE, it was a pleasure sharing this

experience with you. In particular, I found support in having Dr Hajer Kamal

Alm and Dr Aron Tysén by my side.

This thesis project was performed within the Industrial Graduate School VIPP

(Values Created in Fiber Based Processes and Products) with financial support

from the Knowledge Foundation, Sweden. I would also like to acknowledge the

financial support and cooperation with BillerudKorsnäs, Tetra Pak, Stora Enso,

Klabin, Mondi Kraft Paper, Miller Graphics and RISE. Special thanks are also

given to Dr Hajer Kamal Alm and to Omya International and Professor Patrick

Gane, who willingly shared results and samples from their joint project and

offered guidance in my study.

I have received assistance in carrying out various measurements. I appreciate the

help from Mikael Bouveng and Chatleen Karlsson at RISE, Martin Taylor and

Eli Gaskin at Imerys and Bernt Boström at former FIBRO system.

To my parents, Kerstin and Jan-Peder, thank you for the care you have offered,

my appreciation goes beyond words. The thanks are also extended to Margareta

and Björn, who have always been willing to lend a hand. Olle, my life companion.

I have been leaning on you for many years and from now on I am back to share

the load and joy with you. Aston and Eddie, you make my day, every day.

~ Thank you ~

v

Table of contents

Page

1 Introduction .................................................................................................... 1

1.1 Introduction to the research area ......................................................................... 1

1.2 Research questions ............................................................................................. 2

1.3 Scope ................................................................................................................... 2

1.4 A broader context: adequate and consistent print quality .................................... 2

2 Background .................................................................................................... 4

2.1 Coated liquid packaging boards ........................................................................... 4

2.2 Flexographic and offset printing and common print defects .............................. 10

2.3 Ink transfer, trapping and setting ....................................................................... 18

2.4 Liquid absorption and wetting ............................................................................ 22

2.5 Characterisation of liquid absorption.................................................................. 34

3 Experimental ................................................................................................ 40

3.1 Board and paper substrates ............................................................................... 40

3.2 Printing ............................................................................................................... 43

3.3 Print quality analysis .......................................................................................... 44

3.4 Absorption non-uniformity by staining technique ............................................... 48

3.5 Mean absorbency and surface energy............................................................... 51

3.6 Pore-structure of coating layers ......................................................................... 52

3.7 Coating layer uniformity by burn-out treatment .................................................. 53

3.8 Topography: surface roughness ........................................................................ 53

4 Results and discussion................................................................................. 54

4.1 Print uniformity of monochrome flexographic prints ........................................... 54

4.2 Print uniformity of multicolour flexographic prints .............................................. 64

4.3 Validation of the absorption non-uniformity technique ....................................... 74

5 Conclusions ................................................................................................. 78

6 Future work .................................................................................................. 79

7 References ................................................................................................... 80

Appendix A: Laboratory flexographic printing of pilot coated papers .................... 94

Appendix B: List of nomenclature ........................................................................ 95

1

1 Introduction

1.1 Introduction to the research area

A printed colour image consists of areas covered by single or multiple ink layers.

The uniformity of these layers affects the appearance of the print and is,

therefore, an important factor in flexographic printing. In printing, inks are

transferred either directly to the substrate or as overprints to ink layers already

present on the substrate. Therefore, adequate interactions of the printing ink with

both the substrate and the preceding inks are critical.

The uniformity of a monochrome area (monolayer) depends on the direct

interactions with the substrate. When dealing with paper board as a substrate,

sources of print quality issues include topographic variation of the substrate’s

surface, often known as surface roughness (Jensen, 1989; Barros et al., 2005;

Barros and Johansson, 2007), and inadequate ink absorbency (Jensen, 1989).

However, the individual influences from surface roughness and absorbency are

difficult to separate from each other as they co-exist and simultaneously affect

the print (Aspler et al., 1998; Barros and Johansson, 2006; Bohlin et al., 2016).

Unlike roughness, which has often been the centre of research and development

of testing methodology, accumulated knowledge concerning the characterisation

of aqueous absorption is limited. Except for methods that measure average liquid

uptake, for example, Cobb and Bristow’s wheel, there have only been a few

proposed methods in the literature for the characterisation of absorption

uniformity: the Microtack tester with silicon oil (Shen et al., 2005) and the ultra-

sound technique equipped with a multi-sensor (Skowronski et al., 2005).

However, only the latter uses aqueous liquids and has a resolution no higher than

25 dpi. Consequently, there is no quantitative knowledge on how or to what an

extent absorption non-uniformity contributes to flexographic print mottle.

The quality of multicoloured prints with polychrome areas (bilayers) relies on ink

trapping, that is, the acceptance of printing ink by already present inks on the

substrate. Recently, Preston et al. (2017a) presented one of few studies related to

the ink trapping of flexographic inks. It was shown that trapping in laboratory

printing improved because of quicker absorption and more hydrophilic

substrates. Previously, research regarding ink trapping has mainly dealt with

offset printing with the average behaviour of ink-trapping being characterised

over large print areas (Chung et al., 2009; Ragab and El Kader, 2012; Hauck and

Gooran, 2013). Local variations of ink-trapping and its behaviour in flexographic

printing have thus been overlooked and need to be addressed.

2

1.2 Research questions

This thesis project deals with the following research questions: In what way does

the liquid absorbency of a coated paperboard substrate impact the ink

distribution in flexographic printing

a) of a monochrome area, where ink interacts directly with the substrate?

b) of a polychrome area, where ink is transferred onto a previously printed

ink layer, rather than the substrate surface?

To tackle these questions, measurement techniques to characterise absorption

non-uniformity and ink trapping non-uniformity have been developed, together

with a systematic approach to separate the influence of absorbency from that of

surface roughness on print non-uniformity (print mottle).

1.3 Scope

The title, “Where did the ink go?”, refers to the distribution of ink on paper and

board substrates, and specifically to the reasons why ink distributes unevenly.

The liquid absorbency of coating layers was the outset that unified the six

separate papers. The primary focus was flexographic printing with water-based

inks on coated liquid packaging boards, which are a type of carton board used to

package liquid foodstuffs.

A new staining technique was developed to characterise absorption non-

uniformities. The details of the method were introduced in Papers I and II. The

influence of liquid absorbency on monochrome prints was studied in Paper III,

by means of an experimental strategy to modify the uniformity of the absorption

of substrates. The applicability of the absorption non-uniformity technique was

also tested for water/moisture interference issues in offset printing in Paper IV.

A new technique to quantify non-uniformities in the overprint layer was

introduced in Paper V and used in Paper VI. Reasons for non-uniform ink

distribution in monochrome and polychrome areas were studied in Paper VI, with

a specific focus on the influence of coating pore-structure and liquid absorbency.

1.4 A broader context: adequate and consistent print quality

This doctoral project was part of a multidisciplinary research school, where the

technical issue of print uniformity has been discussed from several perspectives.

This broader picture also deserves to be mentioned here. The target of these

studies was not to enable higher print quality, but to assist board manufacturers

and converters in producing printed packaging materials with consistent and

adequate quality, with as little waste as possible. To limit the environmental

3

impact, printability issues need to be identified early in the value chain, before

unnecessary transportation and when the material is easily recycled. A worst-case

scenario is that both packaging and (food) product end up as waste.

One way to limit waste during production is to ensure that the substrate offers adequate

and consistent quality. Preferably, the printability of a board material can already

be predicted and controlled at the board mill, to ensure that it yields the expected

results in the print shop. Thus, a deep understanding of how the substrate

interacts with the ink and printing press is needed. When packaging boards

behave consistently, printers can achieve comparable quality each time and have

a faster start-up for each printing job. This may increase value for the printer and

at the same time reduce the environmental impact.

Print quality is important because the appearance of the packaging influences how

food products are perceived (Magnier et al., 2016; van Ooijen et al., 2017).

Sustainability aspects of packaging are important and if packaging is perceived as

being harmful to the environment, as much as 60 % of Indian and 20 % of

Swedish consumers have stated that they refrain from buying the product

(Innventia, 2013). Separate items of the same product need to display a consistent

quality, to avoid value and safety being questioned by consumers. For customer

satisfaction, the quality needs to be sufficiently high so that it does not deteriorate

the communicative attributes of the packaging, such as information, aesthetics,

branding and appearance (Löfgren and Witell, 2005; Löfgren et al., 2011). Food

safety and freshness information is considered to influence food waste

(Wikström et al., 2014). In fact, a survey showed that 70 %, 31 % and 13 % of

the respondents in India, the USA and Sweden, respectively, threw away milk

without first tasting it when the “best before” date had expired (Innventia, 2013).

Therefore, print quality needs to clearly convey information.

The total impact of food products and their packaging need to account for food losses,

besides the impacts of agriculture, food processing, packaging, transport, retail

and (other aspects of) consumer phases (Wikström and Williams, 2010; Conte et

al., 2015). Energy consumption is the category where the packaging has the

highest impact relative to the food (Wallman and Nilsson, 2011; Williams and

Wikström, 2011). This includes impacts from the production of packaging

materials and treatment of packaging waste. Sometimes it may even be justified

to increase the environmental impact of the packaging, when it helps reduce food

waste and the total impact (Williams and Wikström, 2011; Innventia, 2013).

Assessing the sustainability of a food product and its packaging is complex and

goes beyond the scope of this thesis, but a holistic view is certainly needed.

4

2 Background

2.1 Coated liquid packaging boards

The term liquid packaging board refers to the material being used in packaging

for liquid foodstuffs. Typical, for the liquid packaging board, is the barriers that

are added on both sides to hinder water, oxygen and sometimes also light

transport through the packaging. The carton boards are commonly constructed

in multiple layers, Figure 1, with a structure resembling the folding boxboard or

solid bleached board. Today, this is a solution used for many types of boards,

originating from corrugated board and taking advantage of the engineering

capacity of the I-beam construction for strength and lightweight. The number

and types of layers can differ with, for example, a bleached or unbleached middle

layer, or layers, and an uncoated or coated top. Carton boards, for liquid

packaging applications, contains only virgin fibres and no optical brightening

agents. This study focussed on what occurs when printing directly onto a coated

board before applying any barrier. The focus was on how the final coating

structure impacts ink penetration and print quality. However, a general discussion

on the formation of the final structure has some relevance.

Figure 1. Illustration (not to scale) of a typical liquid packaging board construction. The

fibre layers can be bleached or unbleached and some boards are uncoated.

The fibre network of an uncoated base consists of pores of many different sizes

and a coating layer creates a more uniform pore structure with finer pores that is

more suitable for high quality printing. The coating also serves to increase the

brightness of the board, which requires low opacity and high brightness of the

coating layer itself. Opacity is controlled by light scattering and light absorption.

However, the latter can be achieved through coloration, which affects brightness

in a negative way (Pauler, 1999).

In brief, a coating layer consists of air voids and particles of mineral pigments

bound together with latex or a soluble binder, such as starch. The fluid

absorbency and optical properties will largely be determined by the air filled voids

(pores) and the interfaces between the air and the binder/pigment (Lepoutre,

1989). Pigments and binders typically have a refractive (RI) index around 1.5-1.7

Chemical and/or mechanical pulp layer(-s)

(often unbleached softwood)

Chemical pulp layer

(often unbleached softwood)Barriers on inside of packaging

Chemical pulp layer

(often bleached hardwood)

Top-coating layer(-s)Barriers on outside of packaging

5

(Pauler, 1999). There are exceptions, for example, titanium dioxide has higher RI

than clay and calcium carbonate and will, therefore, give higher light scattering.

Having practically the same RI, means that light will propagate through the

materials in a similar way and, thus, light scattering depends on the interfaces

towards air within the coating structure. Opacity increases with the amount of

interfaces and, thus, a thicker layer of a specific coating colour will yield higher

opacity, as will a greater number of pores. Additionally, finer pores will result in

higher opacity, when compared with a layer of the same pore volume but having

coarser pores (Alince and Lepoutre, 1980; Pauler, 1999).

2.1.1 Coating techniques

In a coater, the colour is either applied to the substrate and then metered, as in

blade coating for example, or pre-metered before being transferred to the

substrate, as in film and curtain coating for example. The final step is drying,

which takes place through liquid penetrating into the base substrate or

evaporating, with the additional help from dryers (Engström and Rigdahl, 1992).

In blade- and film-coating, there is a direct contact between substrate and the

blade, or roll, whereas the curtain coater applies the coating colour without

contact between applicator and substrate. Therefore, a curtain-coated layer

follows the structure of the base and has a uniform thickness, but rough surface

profile (Endres and Tietz, 2007; Lee et al., 2012). It is said to be a contour coating.

The opposite is seen for blade- or roll-coated layers, which follows the structure

of the compressed base and, therefore, yield an uneven thickness, but smooth

surface profile. A high blade pressure, compressible and plasticised (from water-

uptake) base substrate, can create a more uniform mass distribution in blade-

coated layers (Engström, 2016). However, runnability issues, in form of web

breaks on low grammage papers, can arise when running with high blade pressure

and having quick dewatering of the coating colour.

In blade-coating, the colour is applied with an applicator roll or through a jet

applicator. Liquid starts to migrate into the base stock, causing some of the

coating layer to immobilise and viscosity to increase (Dean, 1997). To limit the

viscosity increase, a short dwell time before the metering blade removes excess

colour, is preferred (Engström and Rigdahl, 1992). As the blade meters the

coating, voids in the base are filled and peaks are left with a thinner coating

(Dahlström and Uesaka, 2009). The coat weight is controlled by the blade

pressure and the angle between web and blade. A bent blade, as compared to a

stiff blade, gives a bigger contact area which lowers the pressure. Another

6

metering device is air-knife, which forms more of a contour coating (Engström

and Rigdahl, 1992).

In a film-coater, the coating colour is pre-metered on distributing rollers and then

transferred to the substrate via a roller nip (Endres and Tietz, 2007). Generally,

limited coat weights are applied with this technique.

In a curtain coater, the coating colour is pre-metered and distributed through a

slot die, from above the web. A continuous coating film bridges between the slot

and the web, and this “curtain” must maintain without disruption. Schmid (2011)

reported that the drying energy and water usage was reduced when shifting from

an air-knife to curtain coater, because of the higher solids content used in the

curtain coater.

2.1.2 Coating immobilisation and consolidation

Immediately after applying the coating colour, the binder and pigment particles

can move around freely in the wet coating. When the fluid phase is removed

(dewatered) by means of drainage into the base stock or through evaporation,

the mobility of the particles is greatly restricted (Watanabe and Lepoutre, 1982).

When it comes to rapid dewatering and/or evaporation, there is no time for the

particles to pack tightly together and a bulky and open structure becomes

consolidated, whereas slower dewatering allows more time for pigment

orientation to take place, resulting in there being more efficient packing and

thereby finer pores (Lepoutre, 1978b; Al-Turaif et al., 2002). On the contrary,

Ragnarsson et al. (2013) did not find that there was any effect on the mean

porosity or variations in porosity when the drying strategy was altered, that is, the

dewatering rate was changed.

As the structure is consolidated, large pigments are immobilised first while the

smaller particles, that is, pigment fines and binders, can still move around. They

follow the fluid phase, causing what is referred to as migration. Over the years, it

has been debated whether binders migrate towards the surface (Fujiwara et al.,

1991) or the base (Lepoutre, 1978b; Al-Turaif et al., 2002; Preston et al., 2017b).

Regardless of an enrichment or a dilution of binders on the surface, any migration

may cause a non-uniform distribution in the z-direction. Dahlström and Uesaka

(2012) indicated that conflicting results may originate in the measuring depth of

the analysis techniques used. In their study, latex enrichment was shown in the

very thin (500 nm thick) outermost layers of the coating, both on the coating

surface and in the interface between the coating and the base.

7

Dewatering can be slowed down with the aid of water-retention additives, less

absorbent base stock, increased coating weight and/or changes in the solids

content of the coating colour (Lepoutre, 1978b; Larrondo and St-Amour, 1994;

Laudone et al., 2006). A coating layer reaches a higher void fraction when being

applied to a more absorbent base stock (Larrondo and St-Amour, 1994). The

coating porosity has been linked to coat-weight, with thinner coatings being more

porous and thereby having more water-uptake. In thicker coatings, the pigments

have more freedom to become rearranged and to pack tightly together, which is

possible due to a slower dewatering as a result of high resistance by the filter-

cake (Lepoutre, 1978b; Fujiwara et al., 1991; Laudone et al., 2003; Endres and

Tietz, 2007; Azimi et al., 2011).

Apart from absorbance, the paperboard’s formation (fibre density variations) and

its roughness will impact on the properties of the coating layer. Dahlström and

Uesaka (2009) and Tomimasu et al. (1990) both showed that small-scale

variations in coating thickness depended on surface roughness, whereas larger

scale variations were more dependent on base-sheet formation. The coating layer

became thicker in positions where the base sheet had less mass.

It cannot be expected that the pore-structure, the pigment or the binder

distribution will be completely uniform laterally (Arai et al., 1988). An uneven

coating coverage and coating mass distribution may give rise to lateral variations

in: (i) surface latex profiles (Fujiwara et al., 1990; Fujiwara and Kaga, 1992), (ii)

porosity (Lepoutre, 1978b; Fujiwara et al., 1991; Laudone et al., 2003; Endres and

Tietz, 2007; Azimi et al., 2011), and (iii) optical properties, such as opacity,

reflectance and brightness, which can be seen as white-top mottle on unbleached

base stocks (Pauler, 1999).

2.1.3 Coating colour

2.1.3.1 Pigments

A coating colour is normally composed of pigments, binders, thickeners, water

and various additives. The particle size, shape and particle size distribution (PSD)

will have an impact on the packing and, therefore, the porous structure. With a

wide PSD, the fine particles can fill the pores between the larger pigments, which

means that the packing is tighter and the pores become smaller, Figure 2 (Dean,

1997; Preston et al., 2008a; Preston et al., 2017b).

Often in combination, clay (kaolin) and calcium carbonates are pigments that are

commonly used in coatings for liquid packaging board. The calcium carbonates

are normally divided into natural or ground (GCC) and precipitated calcium

8

carbonates (PCC), where the latter is synthetically prepared and is of a more

uniform size. The crystal forms of the pigments differ. Kaolin has a platy shape,

whereas GCC is sometimes compared to a cauliflower shape. The calcium

carbonate packs more tightly than the clay particles. The clay platelets have

positively charged edges and negatively charged surfaces, which cause them to

pack with an open and porous structure (Larsson et al., 2007). This structure

resembles a house of cards and can easily be deformed, by for example

calendering. Kaolin increases the stiffness of the coating layer in the xy-direction,

but decreases z-directional strength (Preston et al., 2017b).

Figure 2. Illustration (not to scale) of particle packing with mineral pigment of a) narrow

and b) broad particle size distribution.

It is possible for finer pigments to create a smoother surface that is also glossier

(Tyagi et al., 2010). However, the pigments must be large enough to fill the

roughness of the base board and not disappear into the surface pores.

Alternatively, multiple coating layers are used where larger pigments are typically

used in the pre-coating, Figure 3.

Figure 3. Illustration (not to scale) of a coarser pre-coating layer, which fills the roughness

of the base sheet, and a finer top-coating layer that generates a smooth surface.

The mineral pigments are supplied to the paper or board mill pre-dispersed as

high-solids slurries. A common dispersant in calcium carbonate coatings is

sodium polyacrylate. Polyacrylate adsorbs to the calcium carbonate and increases

its surface charge, so that repulsive interactions hinder the particles from

flocculating (Rohmann, 1974; Shaw, 1980). To achieve high levels of rheological

a) Broad particle size distribution b) Narrow particle size distribution

Mineral pigment Latex binder

Base

Pre coating

Top coating

9

stability of the coating colour, an additional dose of dispersant is commonly

added during the makedown of the coating colour (Kane, 1987). This additional

dose may, however, be critical for the printability of the paper (Kamal Alm,

2016).

The interaction of sodium polyacrylate, calcium ions and solid calcium carbonate

is complex and at high polyacrylate concentrations, excess polyacrylate that is not

adsorbed into the carbonate surface forms soluble complexes with calcium ions

(Järnström, 1993; Rogan et al., 1994; Geffroy et al., 2000; Loiseau et al., 2005).

When the calcium ion to polyacrylate ratio increases, colloids are formed, which

may flocculate and form agglomerates (Sinn et al., 2004; Molnar and Rieger, 2005;

Bulo et al., 2007).

Polyacrylate is hygroscopic (Omidian et al., 1998; Sohn and Kim, 2003) and

hydrophilic (Koivula et al., 2011) and has shown an ability to block fine pores

without major influence on the total pore volume of the coating layer (Kamal et

al., 2010; Ridgway et al., 2011)

2.1.3.2 Binder

The properties of a coating layer are affected by the type of binder, its content

and distribution. The higher the binder content, the more closed the structure

will be, resulting in lower porosity and negative effects on the paper optics

(Lepoutre et al., 1979; Aspler and Lepoutre, 1991; Preston et al., 2008a; Bohlin,

2013). Higher binder (latex) content is needed to achieve surface strength of a

coating layer with fine mineral particles, which have a greater surface area

compared to coarser particles (Preston et al., 2017b). Latex has a tendency to

accumulate in the small pores, which means that the pore-size distribution will

change when a binder is added (Ström et al., 2010; Dahlström and Uesaka, 2012).

The polymer stiffness (glass transition temperature, Tg) impacts on the shrinkage

of the coating layer during drying (Watanabe and Lepoutre, 1982; Laudone et al.,

2006). A severe shrinkage causes low porosity and greater surface roughness.

High Tg polymers tend to distribute more towards the coating surface, leaving

the bulk with higher porosity (permeability) and the surface structure somewhat

smoother (Olsson, 2007; Järnström et al., 2010). A less cross-linked (styrene-

butadiene) latex, is more deformable and, thus, the coating structure can stay

more stretched during coalescence and becomes more permeable (Lamminmäki

et al., 2005). The size of latex particles affects immobilisation rate of the coating

colour, and there is more time for pigment packing when using small particles.

10

2.2 Flexographic and offset printing and common print defects

2.2.1 Water-based flexographic printing

2.2.1.1 Flexographic process

In flexographic printing, the ink is transferred from an anilox onto a relief

patterned flexible printing plate and further onto the substrate. It is the relief

pattern that separates printing from non-printing areas. Several different

substrates can be printed on, from plastic to paper and board. This thesis deals

with printing on coated boards and, to a limited extent, coated papers, with a

focus on liquid packaging boards.

The anilox is filled with ink and the excess ink is removed with a doctor blade.

There are different designs of inking unit, and enclosed or chambered systems

with double doctor blades are commonly used (Leach and Pierce, 1993). The

anilox can be engraved to carry different ink volumes through the size and

frequency of cells. This is an important factor for the ink transfer to the plate.

However, other variables will also affect the transfer, such as the cell geometry,

ink properties, surface chemistry and topography of the plate, as well as the

pressure between the anilox and the plate.

Plate materials vary in hardness, thickness, surface chemistry and surface structure;

in addition, the plate making process can give rise to characteristic relief profiles,

such as the depth and shape of the dots. The plates are mounted on a roller or

sleeve with adhesive tape, or mounted directly onto a self-adhesive sleeve. When

printing on corrugated boards, for example, a thick compressible cushioning tape

is often used.

It is important to achieve contact between the ink and substrate and this depends on a

large set of variables where the printing pressure and compressibility of substrate,

plate, mounting tape and/or sleeve are important. The flexibility of the image

carrier and its backing means that they can conform to a rather macro-rough

surface of the substrate and, therefore, print density, ink transfer and ink

coverage increase when the plate is softer and/or the nip pressure is higher

(Lagerstedt and Kolseth, 1995; Johnson et al., 2004; Barros and Johansson, 2007).

At the same time, this deformation means that the image can become distorted

and the pressure can become unevenly distributed, which may lead to increases

in tone-value and print mottle (McGratten, 1990; Aspler et al., 1998; Hallberg

Hofstrand et al., 2006; Olsson et al., 2006b).

As described by, for instance, Taylor and Zettlemoyer (1958), Myers et al. (1959),

Coyle (1988) and Yang (2013), the pressure-pulse in a nip can be divided into

11

three regimes: (i) nip pressure increases when approaching the centre of the nip

(compression), (ii) after the narrowest gap the pressure decreases

(decompression) and then (iii) there may even be a negative force acting on the

substrate during ink film splitting.

The nip pressure, as such, is not defined in full-scale printing presses, but instead

the nip gap relative to the substrate thickness and compression stiffness

determines the pressure. The cylinders are aligned and brought together so that

the image begins to transfer. From this position the gap is made narrower (that

is, impression is increased) until a satisfactory print is achieved, where ink

coverage in solid tones is balanced against a tone-value increase in halftones. The

pressure involved in flexographic printing is normally considered to be low, and

pressures as low as 0.1 MPa to 0.5 MPa have been suggested (Lim et al., 1996;

Kipphan, 2001; Bohan et al., 2002); however, significantly higher pressures

between 1 MPa and 2 MPa have also been found (Mirle and Zettlemoyer, 1988;

Johnson et al., 2004).

2.2.1.2 Flexographic presses

For roll-to-roll flexographic printing, there are at least three classes of presses:

central impression cylinder (CI), in-line and stack type presses. The classification

refers to the location of the printing decks. In a CI press, all the printing decks

are geared around a large common impression cylinder, whereas each deck has a

separate impression cylinder in the other two types. An in-line press refers to the

horizontal arrangement of the decks and a stack type press to a vertical

arrangement of decks, in which the web often passes half of the decks on the

way down and the other half on the way up. The register shifts, from colour to

colour, are minimised in a CI press, where the substrate is supported by the

central impression cylinder (Leach and Pierce, 1993). Intermediate drying stations

usually dry the ink between the printing units (Kipphan, 2001).

There is also laboratory equipment available for flexographic printing, for

example, the F1 Printability tester (IGT Testing Systems), Flexiproof 100/UV

(RK PrintCoat Instruments Ltd), Little Joe (Little Joe Industries), Perfect

Proofer™ (Print Proof Solutions B.V.) and Printing Proofer (Biuged Laboratory

Instruments (Guangzhou) Co, Ltd.). However, there are some obvious

differences when compared to full-scale printing, such as geometry, inking unit

and press speed. Yet, Aspler (2004) saw a high correlation and visual similarities

when comparing unbleached liner boards printed on a commercial web press and

the IGT F1.

12

2.2.1.3 Flexographic inks

A water-based flexographic ink consists of dispersed pigments and a vehicle with

binder resins, a solvent (water) and various additives, such as defoamers,

surfactants and waxes. The resins act as a carrier for the pigments during printing

and bind them to the substrate upon drying. There are commonly two types of

binders, which are either used separately or combined: dispersed resin and

alkaline solutions (Luu et al., 2010a). In the dispersed vehicle, there is an emulsion

of fine particle resins, suspended in water. As the emulsion dries, the particles

coagulate or flow together to form a film. A water-insoluble carboxylated resin

(commonly acrylic) can also be solubilised by addition of volatile alkaline

material, such as ammonia or organic amines (FFTA, 1999). The carboxylic acid

group in the resin reacts with the base and this way forms a water-soluble salt.

Flexographic inks are normally diluted with water in the print shop to reach the

required viscosity (FTA, 1992). On the press, viscosity is commonly measured

with an efflux flow cup, which is fast and easy to use, but it can give an inadequate

response if the ink is tixotropic (Leach and Pierce, 1993). When diluted to a lower

viscosity (that is, lower pigment concentration), the optical density decreases as

observed on non-absorbent polyethylene (Lavelle et al., 1996). However, when

heated to a lower viscosity, the optical density can increase as observed on coated

substrates (Olsson et al., 2007b). In addition, a high degree of dynamic surface

tension and polarity will have a negative impact on the transfer and spreading

ability of the ink (Aspler et al., 2004). Furthermore, Olsson et al. (2006a) showed

that fluids with greater dipole moments penetrate deeper into the coating layer.

More about the liquid rheology can be found in Section 2.4.6.

It is suggested that the binder type affects how a filter-cake builds up in the ink

when the fluid phase separates from the ink solid phase. This filter-cake affects

how easily the ink is dewatered and solution polymer-based inks have shown a

slower dewatering than emulsion polymer-based inks (Olsson et al., 2007a; Luu

et al., 2010a; Luu et al., 2010b).

2.2.1.4 Common print defects in flexographic printing

Print mottle, non-uniform optical density, in a solid tone or halftone area, has a

major influence on print quality in flexographic printing on both coated and

uncoated substrates (Wågberg and Wennerblom, 1992; Aspler et al., 1998;

Aspler, 2004). Print mottle is commonly divided into classes of spatial size, and

wavelengths above 1 mm and up to 8 mm, or possibly 9 mm, are considered the

most disturbing for the perceived quality (Jensen, 1989; Johansson, 1993;

Johansson and Norman, 1996; Fahlcrantz and Johansson, 2004). Small-scale

13

variations are sometimes referred to as speckle (MacGregor and Conners, 1987),

ink-splitting pattern (Barros and Johansson, 2007), filament pattern (Bohlin et al.,

2013), marbling (Quesne et al., 2013) or viscous fingering (Sauer et al., 2018) and

often appear in the sub-millimetre wavelengths.

Uncovered areas (UCAs), that is, white spots in the print, are characteristic defects

in solid tone flexographic prints. Many small UCAs affect print mottle

measurement and, hence, the two can be closely connected.

Optical print density characterises the darkness of the ink layer relative to the

unprinted substrate. It increases exponentially up to a certain limit, from which

there is no further increase in density when applying more ink (Tollenaar and

Ernst, 1961; Aspler et al., 1992). Optical density depends on the ink amount, but

also on the ability to retain ink pigments close to the substrate’s surface (ink

holdout). Similar to uncoated papers (Pauler, 1987), print density on coated

substrates can be assumed to decrease when the ink pigments penetrate into the

pore-structure, compared to staying at the surface. In addition, the uniformity of

the ink layer affects the measured optical density. When unevenly distributed over

the surface, the average density can be far lower than the saturated density of the

spots with full ink coverage (Tollenaar and Ernst, 1961; Lavelle et al., 1996).

Ink trapping occurs in multi-colour printing when an ink is transferred to a

previously applied ink film on the substrate. Ink trapping describes the

acceptance of the overprint ink on the first ink layer, relative to the acceptance

directly onto the substrate. The mean ink trapping is commonly assessed by

gravimetric, densitometric or spectral measurements (TECHKON, 2000;

Yuanhong and Zhihong, 2008; Chung et al., 2009; Hauck and Gooran, 2013).

This area has mainly been studied with respect to offset printing, where trapping

has shown to be a function of ink tack (Tollenaar and Ernst, 1965; Chung et al.,

2009). The tack of the first ink layer increases as it sets and, thereby, the ink

trapping rises.

Preston et al. (2017a) recently presented one of few studies on ink trapping in

flexographic printing. From laboratory trials, it was found that surface roughness

had a dominant and negative effect on ink trapping. Furthermore, ink trapping

behaviour became more effective and complete with a quick setting of the first

ink. Pre-dampening (not flooding with water) of the surface before printing

increased the absorption and ink setting rate of the first ink, which had a positive

effect on ink trapping. Calendering, on the other hand, had a negative effect,

which was explained by the decreased porosity of the coating layer and thereby

14

slower ink setting. Ink trapping was determined by the show-through of the first

ink through the overprint layer.

Experiments related to flexographic multi-colour printing were conducted by

Johnson et al. (2005); Johnson et al. (2008) on uncoated substrates. Aqueous pre-

treatments were made prior to ink application to simulate ink-on-ink printing.

On low absorbing boards, a pre-treatment with water or surfactant solution

reduced print mottle, whereas a surfactant/wax solution increased print mottle

(Johnson et al., 2005). Similar pre-treatments had no impact on papers that were

already highly absorbent (Johnson et al., 2008).

2.2.1.5 Interactions with the substrates

Surface roughness makes ink transfer uneven and, as a result, print density and the

perceived print quality of uncoated and coated substrates is impaired (Jensen,

1989; Wågberg and Wennerblom, 1992; Lagerstedt and Kolseth, 1995; Aspler,

2004; Barros and Johansson, 2007; Luu et al., 2010b). On rough substrates the

pressure will be unevenly distributed, which will push ink away from surface

peaks and result in print mottle (Barros and Johansson, 2007). Surface craters

hinder ink transfer and, therefore, generate uncovered areas (UCAs) when the

craters are too wide for the ink to bridge (> 0.15 mm wavelength) but still too

narrow to be flattened in the nip (< 0.75 mm wavelength) (Barros and Johansson,

2007).

Not all structures of the surface give rise to printability issues, and it has been

pointed out that the commonly used air-leak techniques for roughness

measurement may neither be adequate for coated (Aspler, 2004; Barros et al.,

2005) nor uncoated substrates (Steadman et al., 1993). The various air-leak

techniques are sensitive when it comes to different ranges; Bendtsen and

Sheffield instruments capture long wave variations (2-5 mm), whereas Parker

Print Surf (PPS) can detect structures across the whole range but without

significantly high correlation in any given band (Wågberg and Wennerblom,

1992; Steadman et al., 1993). Instead, more sophisticated techniques are

recommended, where the topography of the surface is captured and can be

analysed with focus on various characteristics, such as variations in different

wavelength ranges or crater areas (Barros and Johansson, 2007).

Print quality may also be influenced by the surface energy and pore-structure of

the substrate, sometimes seen as ink absorbency. In a review by Aspler (1993), it

is suggested that a high degree of absorbency may lead not only to lower ink hold-out,

but also to higher degree of ink transfer. For uncoated boards, several reports

state that print density increases with a lower level of absorbency (McGratten,

15

1990; Wågberg and Wennerblom, 1992; Steadman et al., 1993). At the same time,

the slow absorbency may increase print mottle (Wågberg and Wennerblom,

1992). Jensen (1989) explained that a higher print mottle occurs when the ink

setting is slowed down and gives more time for ink spreading.

McGratten (1990) pointed out that ink can be adjusted separately for high- and

low-absorbent substrates, but problems mainly arise when a range of boards with

different rates of absorbency are used. During printing, it may be more difficult

to overcome or compensate for laterally non-uniform absorption and print. For

example, Preston et al. (2008a) suggested that uneven ink-spreading is an

important parameter, which causes lateral variations in dot gain.

Increasing the surface energy and making paper or board easier to wet with the ink,

increase ink transfer and increase print density (Lagerstedt and Kolseth, 1995;

Aspler et al., 1998; Sheng et al., 2000). More polar coating layers have shown

higher print density (Olsson et al., 2006b) and more uniform prints (Bassemir

and Krishnan, 1991) in water-based flexography. Hou and Hui (1990) pointed

out that both polar and dispersive components of the surface energy are of

importance for water-based ink absorption.

The impact from surface energy on ink spreading may have a greater impact on

halftones, and their tone-value increase, than on solid tones (Lagerstedt and

Kolseth, 1995; Aspler, 2004). Ink-paper boundaries are formed around each dot

(at least in lower tone values) and the spreading potential will impact the dot size.

In the solid tones, ink spreading is more likely to have a levelling effect and yield

a smoother ink layer. Ink-paper boundaries are also formed where there are

UCAs and, consequently, spreading may reduce their size.

A more open coating structure with large pores allows for deeper ink penetration and

a lesser ink holdout (Aspler and Lepoutre, 1991; Zang and Aspler, 1995b; Mesic

et al., 2004; Preston et al., 2008a; Bohlin et al., 2013). On a coating layer, the

possibility of ink pigments penetrating into the pores is limited because of size

exclusion (Xiang and Bousfield, 2003). Measurements on coated substrates have

indicated that offset pigments retain on the surface with no, or only slight,

penetration into the pore structure of the substrate (Preston et al., 2000; Ström

et al., 2001; Koivula et al., 2008). Zang and Aspler (1995a) concluded that an

uneven porosity in the lateral direction may lead to print mottle, due to an uneven

ink holdout and uneven ink setting.

The individual effects of surface roughness and ink absorbency (surface energy or pore-

structure) are often difficult to see and separate from each other because they co-

exist and may simultaneously affect the print. This was shown in studies by

16

Barros and Johansson (2006) and Bohlin et al. (2013), where 30 % to 50 % of the

UCAs on coated boards were linked to the surface structure, and the rest were

suggested to originate from wetting. Also, micro-mottle has been linked to both

topography and capillary absorption (Wågberg and Wennerblom, 1992; Barros

and Johansson, 2007).

Additionally, the relative sensitivity to surface roughness and surface energy can

be affected by the press settings; for example, ink binder type (Olsson, 2007; Luu

et al., 2010b), press speed (Lagerstedt and Kolseth, 1995), plate hardness and nip

pressure (Barros and Johansson, 2007). Possibly, the interaction between

absorbency and surface roughness determine their relative impact on the print.

Zang and Aspler (1995a) suggested that a smooth surface may not be enough to

generate high print quality, since parameters such as surface energy can start

influencing the print when the surface is smooth. Lagerstedt and Kolseth (1995),

on the contrary, proposed that the combination of two unfavourable properties

(for example, roughness and surface chemistry) might be needed to cause poor

print quality. Aspler (2004) did not see any influence of surface chemistry on

print quality for white liner boards, but advised that “water absorbency beyond

the commercial norm – either greater or less – may produce problems.”

2.2.2 Offset printing

Offset printing is not the main target of this thesis, but one of the papers touches

upon this area: In Paper VI the influence of non-uniform absorption on

water/moisture interference mottle in offset printing was studied. Hence, a short

description of the printing technique and printing quality issues is given.

2.2.2.1 Lithographic process

In offset printing, an oil-based ink is transferred to the paper via a printing plate

and rubber blanket. Image areas on the plate are oleophilic (with low surface

energy), whereas non-image areas are hydrophilic and will be wetted by the

aqueous fountain solution which is applied prior to the ink. The fountain solution

consists of water, a wetting agent (often isopropyl alcohol and/or surfactant), a

buffering system to control pH and other additives (for example, plate

preservative agents) (Kipphan, 2001). This is an important part of the process

and will act to keep non-image areas free of ink (MacPhee, 1979). This means

that the fountain solution will be transferred to the paper substrate from the non-

image areas and, for each printing unit, the amount in the paper will increase. The

fountain solution is also emulsified in the ink and transferred to the substrate in

this way.

17

2.2.2.2 Water/moisture related print defects

Ink and fountain solution are applied in each printing unit and thus the ink can

meet a moist paper surface. The feed of fountain solution is a delicate balance

between the amount needed to keep image areas free of ink and the ability of the

substrate to remove this excess water. When moisture is retained and distributed

unevenly at the paper’s surface, it may give rise to ink refusal and water-

interference mottle (Kolseth, 2004; McConkey et al., 2004; Ström and Preston,

2013) or a bond to the ink that is so weak that the ink is detached at the contact

with the rubber blanket in the following unit (Kamal Alm et al., 2015a), see

Figure 4. The latter is referred to as ink lift-off. Water/moisture interference

problems normally occur in some of the last units in the press.

Figure 4. Illustration of water/moisture interference problems, which arise in the fifth unit,

and how the inked area may suffer from additional ink-lift-off failure in the subsequent

(sixth) unit.

2.2.2.3 Back-trap mottle

An ink layer on the substrate will split in each of the subsequent printing units.

This means that part of the non-immobilised ink transfers back to the next

blanket and later contacts the next signature (sheet). Optimally, an equilibrium is

reached after a number of signatures. Under those circumstances back-trapping

will make the ink film more uniform, as irregularities are levelled out (Purfeerst

and van Gilder, 1991; Kolseth, 2004).

In many cases, the situation is not optimal and uneven oil absorption results in

uneven ink setting (Xiang et al., 1999; Xiang et al., 2000). This will affect the

uniformity of the ink that is transferred in the back-trap nip and, consequently,

the next signature meets a non-uniform ink layer on the blanket. This causes

back-trap mottle. The problem is foremost seen in ink layers that are applied in

some of the first printing units, where the ink will start to set before the last units

have passed. A non-uniform surface pore-structure is often the cause of the

uneven ink setting and back-trap mottle (Kim et al., 1998; Chinga and Helle,

2003; Shen et al., 2005; Preston et al., 2008b).

Moist paper from

previous press unitsUneven laydown of ink:

Water-interference

mottle (WIM)

Ink may be removed

due to insufficient

adhesion (Ink-lift-off)

and cause UCA

C5

Ink

C6

No ink

18

2.3 Ink transfer, trapping and setting

In a printing nip, the ink is transferred to a substrate where it can redistribute and

spread laterally over the surface and form a uniform ink layer before the ink film

sets. This process is often referred to as ink levelling. The mechanisms of ink

transfer, directly onto the substrate and trapping onto a previously applied ink

layer, and ink setting will be described in the coming section and are affected by

conditions and materials in the printer, the ink and the substrate properties. The

transfer is, for example, affected by the substrate’s roughness, compressibility

and absorbency, and the ink’s rheology under shear and tension as well as its

adhesion to the substrate. Ink setting rate is affected by the rate of liquid uptake,

rate of evaporation and formation of a filter-cake in the ink.

2.3.1 Mechanisms of ink transfer

Ink transfer is often divided into three general steps: (1) contact and wetting

between the ink and substrate when entering a nip, (2) immobilisation of part of

the ink layer and (3) still fluid ink splits, some of it follows the substrate and the

rest remains on the plate. One of the commonly used models for ink transfer

mechanisms was derived empirically from studies of offset printing by Walker

and Fetsko (Walker and Fetsko, 1955):

𝑦 = (1 − 𝑒−𝑘𝑥)[𝑏(1 − 𝑒−𝑥 𝑏⁄ ) + 𝑓{𝑥 − 𝑏(1 − 𝑒−𝑥 𝑏⁄ )}] [Eq. 1]

Where y = the amount of ink transferred per unit area to the substrate, k = a

constant related to the smoothness of the substrate that regulates contact

between the ink and substrate, x = the initial amount of ink on the plate, f = the

split factor (the fraction of ink remaining on the paper) and b = the amount of

ink that is immobilised during impression. When the ink film is sufficiently thick

and achieves full contact with the substrate, Eq. 1 can be shortened into:

𝑦 = 𝑏 + 𝑓(𝑥 − 𝑏) [Eq. 2]

The Walker-Fetsko model is often used and can give valuable input, but one

criticism that has been raised is that there is interdependence between the

immobilisation (b) and the splitting (f) parameters (Aspler, 1993). As described by

Zang (1993), the above equation has in many cases given ink transfer curves that

do not agree with the published experimental data and/or unrealistic ink transfer

parameters, such as negative splitting coefficients.

2.3.1.1 Contact and wetting

First, ink transfer depends on physical contact between the ink and substrate. A

narrower gap or thicker ink film increase the contact, whereas the surface

19

roughness of the substrate is a hindering factor. To some extent, this can be

overcome by the compressibility of the substrate, the nip pressure and the

softness of the plate material, which enables it to conform to the substrate’s

surface. It is sometimes stated that the ink needs to wet the substrate in order to

achieve contact and transfer (Aspler and Lepoutre, 1991). However, transfer has

been shown to be possible despite the surface energy of the substrate being lower

than that of the ink (De Grace and Mangin, 1983; Lagerstedt and Kolseth, 1995;

Shen et al., 2004). Liu and Shen (2008) explained that the transfer of high energy

oil-based ink to low energy Polytetrafluoroethylene (Teflon) is enabled under

forced wetting, when the adhesion strength between the ink and substrate is

greater than the force required to split the ink film. Still, when the nip pressure is

removed, the low energy substrate may de-wet so that the ink film beads up after

the nip.

2.3.1.2 Ink immobilisation

Upon contact with the substrate, ink will start to immobilise by absorption

and/or filling the micro-structure of the surface (De Grace and Mangin, 1983).

Absorption is driven by the external nip pressure, but it can also be governed by

both wetting and capillary action. When it comes to offset printing, it has been

claimed that the immobilisation is more dependent on the filling of micro-

structures than on absorbency (Lepoutre et al., 1979; Isoard, 1983; Zang and

Aspler, 1995b).

Filtration resistance of the ink may influence the immobilisation and, therefore,

the ink splitting. A lower resistance allows for a higher pressure-driven

absorption (Gros et al., 2002). Ink tack will rise when a filter-cake is formed

during the setting of a flexographic or offset ink (Zang and Aspler, 1995b; Ma et

al., 2009). Increased tack close to the coating layer shifts the split position away

from the substrate.

2.3.1.3 Ink splitting

As the ink film leaves the nip, the free (non-immobilised) ink-film will split

between the plate and substrate. Myers et al. (1959) showed that cavitation forms

at high rotational speeds and expands when approaching the role nip exit. This

leaves ink filaments attached to both the plate and the substrate, which will

elongate until they break up. The split of offset inks is considered to be

asymmetric, with a ratio that is in favour of the plate (Taylor and Zettlemoyer,

1958; De Grace and Mangin, 1983; De Grace and Mangin, 1987). Different

reasons behind the asymmetry were proposed: (a) cavitation starts from the

centre of the nip and expands more rapidly towards the substrate where the ink

20

viscosity is lower (Taylor and Zettlemoyer, 1958), and (b) cavitation initiates from

the micro-roughness of the substrate’s surface where air gets entrained and do

not have sufficient time to reach the centre of the nip (De Grace and Mangin,

1983; De Grace and Mangin, 1987).

Ink splitting patterns are often seen in prints and these non-uniformities may be

linked to Saffman-Taylor instabilities in the ink-air interface at the exit of a

printing nip and are, therefore, referred to as viscous fingering. Ink forms a

meniscus between the printing plate and the substrate and, when low viscous air

displaces the more viscous ink, finger-shaped air intrusions are created, Figure 5

(Sauer et al., 2018). These turn into liquid bridges and further into thin sheets

that finally collapse, either smoothly or with a release of small droplets, when the

nip gap expands (Sauer et al., 2018). Sauer et al. (2018) describes that the size of

the fingers increases with print cylinder radius, surface tension of the ink and

thickness of the liquid film, but decreases with velocity and ink viscosity. Also

Wang et al. (2014) saw that the size of the fingers grew larger with ink volume

and Morgan et al. (2018) concluded that more elastic inks decreased their size.

Figure 5. Illustration of viscous fingering at the exit of a printing nip (a), which leads to ink

splitting patterns in the print (b).

2.3.2 Mechanisms of ink trapping

In multi-colour printing, ink does not only transfer directly onto the substrate

but may also transfer onto an already present wet/semi-wet ink layer. In offset

printing, a third alternative is transfer onto a thin layer of fountain solution. To

acquire effective ink trapping, the convention in offset printing has been to

Air intrusion

Centre of nip: Top view

Web

direction

Onset of finger

formation at

meniscus front

Air intruding behind

fingers/ink bridges

at meniscus front

Filaments begin to

detach from

meniscus front

Air intrusion

a)

Centre

of nip

100 µm

b)

21

ensure that subsequent inks have a lower tack than the ink film present on the

substrate (De Grace and Mangin, 1987; Aspler and Lepoutre, 1991). When the

substrate leaves the printing nip where the overprint layer is applied, the ink split

is suggested to occur in the ink layer where the least force is needed (Tollenaar

and Ernst, 1965). In a situation where the first layer is not strong enough to resist

the separating forces, it can split and thereby prevent transfer of the overprint

ink. In successful ink trapping, cohesive forces in the first ink layer withstands

splitting, due to the greater tack and immobilisation that has reduced the effective

thickness (of mobile ink) in the first ink layer (Tollenaar and Ernst, 1965). De

Grace and Mangin (1987) suggested that the splitting mechanisms were similar

in both multi-colour and monochrome printing, with a tendency for the split

ratio to be asymmetric.

MacPhee (1979) explained that the splitting of the weakest film enables a thin

layer of fountain solution to keep non-image areas on an offset plate free of ink.

Shen et al. (2004) pointed out that the adhesion strength between ink and

fountain solution is substantial when brought into intimate molecular contact

and, therefore, it is unlikely with adhesion failure and more likely that the ink or

fountain solution layer splits. The fountain solution layer must be thick enough

for this to happen, otherwise the required split force will be too high. Both

MacPhee (1979); Shen et al. (2004) linked the split force (F) to Stephen’s law;

𝐹 = 𝜂𝑉𝐴𝐶 𝑡3⁄ [Eq. 3]

Where η is liquid viscosity, V is the separation speed, A the contact area, C a

constant and t the film thickness.

2.3.3 Mechanisms of ink setting and drying

In the following description, ink is considered to have set when no further

redistribution (levelling) can occur on the micrometre-scale, although there may

still be redistribution on a molecular level. The ink binder will film-form and

polymerise during drying, which makes the dried ink water-insoluble.

When the fluid phase is removed, the ink pigments are consolidated and

immobilised. The immobilisation process happens very quickly and before

becoming immobilised the ink film can level and form a uniform ink layer on the

substrate, as described for offset printing by, for example, Donigian (2006). The

levelling process is governed by the film thickness, viscosity and surface tension

of the ink. As described by Kheshgi (1997), a capillary pressure and gravity will

act to level and smooth irregularities in thin films. The surface tension of the

liquid will strive to minimise the surface area and create a boundary between the

22

ink and air which is as flat as possible. Hence, liquid will be transported from

convex to concave regions of the film. This action can create a uniform film

thickness when the solid surface is smooth, or an uneven film thickness when

the solid surface is rough, see Figure 6. Gravity will promote the liquid motion

and speed up the levelling of a horizontal surface, whereas viscoelastic effects

may slow down the levelling. Levelling may be counteracted by centrifugal forces

when rotating at high speeds (Bousfield, 1991).

Figure 6. Illustration of the levelling effect of surface tension acting on an (ink) liquid

film. The direction of fluid transport (top) and the final levelled film (bottom) are

shown for a film on a solid that is smooth (left) and one that is rough (right).

As the coating reaches a high solids content, the levelling will stop. This means

that thin liquid films, a quick removal of fluid phase through absorption or

evaporation, a high initial solids content and high viscosity will shorten the time

for ink immobilisation and thereby levelling (Aspler and Lepoutre, 1991;

Bousfield, 1991; Leach and Pierce, 1993; Kheshgi, 1997). When there is not

sufficient time for levelling, ink splitting patterns may remain visible in the print

(Barros and Johansson, 2007). The longer the spatial wavelength of the

irregularity, the longer it will take to level it (Kheshgi, 1997). Therefore, large-

scale disturbances may not have time to heal before the film has dried, and only

short wavelength disturbances can be smoothened when drying quickly.

2.4 Liquid absorption and wetting

It is generally accepted that liquid absorption into porous networks can depend

on wetting, liquid properties and the pore structure. The mechanism of

absorption may be approximated by penetration into a single pore, a bundle of

Flow FlowFlow Flow

Healing of irregularities

Ink film that has levelled

23

pores or simulations of penetration into entire pore networks. As described in

the previous section (2.3 Ink transfer, trapping and setting), absorption can be part of

the ink transfer as well as the ink setting processes.

In a review by Aspler (1993), it was suggested that the sorption of water into

newsprint is determined by surface chemistry, whereas the sorption of oil is

mainly affected by the pore structure. Thus, the liquid properties that influence

the behaviour and results drawn from absorption testing of oil-based liquids may

not translate to the absorption of water-based liquids.

There is a limitation to the modelling and measuring of the absorption of “pure”

liquids in order to adequately describe the penetration of the fluid phase from a

thin layer of pigmented ink (Aspler and Lepoutre, 1991; Aspler, 1993). A high

degree of absorption could result not only in more ink being transferred to a

substrate, but also in a deeper level of penetration of either the whole ink or just

the fluid phase of the ink. To add to the difficulty of modelling and understanding

ink absorption mechanisms, the ink changes as the fluid phase dewaters, which

means that viscosities and the contact angles are not constant (Aspler, 1993). The

following sections focus on how the coating layer may control aqueous

absorption, without specific consideration being given to ink pigments.

2.4.1 Driving forces for fluid penetration

Liquid can be dragged into a capillary or pore by the capillary pressure (PC)

and/or pushed into the capillary by an external pressure (PE), such as nip

pressure, during the printing process. When it comes to a single capillary tube,

the Young-Laplace2 pressure difference shows that the capillary pressure (PC) is

proportional to the product of the liquid surface tension (γ) and the cosine of the

contact angle between the liquid and the solid (θ); and is inversely proportional

to the radius (r) of the capillary:

𝑃𝐶 = 2𝛾𝑐𝑜𝑠 𝜃

𝑟 [Eq. 4]

Eklund and Salminen (1987) stressed that the capillary pressure is more correctly

described as a dynamic than a static pressure. Liquid vapour will precede the

liquid phase and, as the liquid molecules adsorb into the pore wall, the contact

angle will change. The dynamic contact angle will depend on the penetration

velocity and possibly also on liquid viscosity. Salminen (1988) suggested that the

2 This equation is often used, for example, Washburn (1921), Schoelkopf et al. (2000b) and Ridgway et al. (2001), but seldom cited. Finn (1999) described that Young (1805) and Laplace independently arrived at the same conclusions.

24

influence of the dynamic behaviour of contact angles diminishes under external

pressure.

Ma et al. (2007) described liquid imbibition thermodynamically, in terms of

lowering the Gibbs free energy. The change in Gibbs free energy will be greater

when capillary pressure is greater, and they suggested that, under equilibrium

imbibition, only the driving force should be accounted for and that the

retardational forces should be disregarded. The change in Gibbs free energy

(∆𝐺𝑖) for a capillary of any given pore size (i) is defined as the volume (Vi) times

the Young-Laplace pressure difference (Eq. 4):

∆𝐺𝑖 = −𝑉𝑖4𝛾𝑐𝑜𝑠𝜃

𝐷𝑖= −4𝐴𝑖ℎ𝑖

𝛾𝑐𝑜𝑠𝜃

𝐷𝑖= −𝜋𝐷𝑖ℎ𝑖𝛾𝑐𝑜𝑠𝜃 [Eq. 5]

This means that, at given contact angle (θ) and surface tension (γ), the change in

Gibbs free energy in the final stage (filled coating) is determined by the diameter

of the pore (Di) and its length (hi). The negative sign accounts for the penetration

of wetting liquids, with cosθ>0 being spontaneous.

2.4.2 Resistive forces for fluid penetration

Several factors can slow down absorption into a capillary. There is a resistive

viscous drag, accounted for by the Hagen-Poiseuille flow. The pressure loss (PF)

is proportional to the liquid viscosity (η), capillary length (h) and volumetric flow

rate (Q = dV/dt), and inversely proportional to the capillary radius (r) to the

power of four:

𝑃𝐹 =8𝜂ℎ𝑄

𝜋𝑟4, 𝑄 =

𝜋𝐷4𝑃𝐹

128𝜂ℎ 𝑜𝑟 𝑄 =

𝜋𝑟4𝑃𝐹

8𝜂ℎ [Eq. 6]

As described by Sutera and Skalak (1993), Eq. 6 is commonly referred to as

Poiseuille’s law, but this form was not derived by Poiseuille. In his equation,

π/128η was denoted with K’’, which represents a function of the temperature

and the type of liquid.

The liquid accelerates at the very beginning of absorption into a capillary. At this

point, inertia and the density of the liquid are important. There is also the

possibility of a counter pressure caused by entrapped (or at least not easily

flowing) air, as described by Salminen (1988).

Formation of a filter-cake of flexographic ink pigments and binders can also

affect the removal of fluid from inks (Gane and Ridgway, 2009; Ma et al., 2009).

When the permeability of a filter-cake is low, the absorption rate is mainly

controlled by the filter-cake itself, and not by the substrate (Ma et al., 2009). A

high filter-cake resistance leaves a longer period for ink levelling and spreading,

25

so that, for example, print gloss can increase (Olsson et al., 2007a). However, a

low resistance lets the substrate´s surface energy become more influential on

both the absorption rate and the print (Ma et al., 2009; Luu et al., 2010a; Luu et

al., 2010b). Another aspect is that the filter-cake may act to retain and hinder

polymers from entering and blocking fine pores in the coating layer and,

consequently, it prevents absorption retardation (Gane and Ridgway, 2009).

Similarly for offset printing, the filter-cake reduces permeability and, thereby,

impedes fluid drainage (Zang and Aspler, 1995b; Xiang and Bousfield, 2000;

Xiang et al., 2004). However, Ström et al. (2001) and Ström (2005) suggested that

the concentration of pigment in offset inks is too low to form a filter-cake and,

instead, binder solution and dispersed pigments form a gel phase.

2.4.3 Models of absorption

Liquid absorption can be described using the Navier-Stokes model. However,

due to its complexity, various simplified forms are commonly used in the paper

industry. According to a summary by Ridgway et al. (2002), the best known

solution is the Lucas (Lucas, 1918) and Washburn (Washburn, 1921) equation

(Eq. 7). This does not account for inertia dependency during short time

absorption, which the Bosanquet equation does (Eq. 10) (Bosanquet, 1923).

At the liquid front there is a balance between the driving pressures, for example,

capillary and external pressure, and the resistive pressure, that is, the viscous flow:

PE+PC=PF (Salminen, 1988). This assumes the capillaries to be open at both ends

and, therefore, the atmospheric pressure can be neglected. In the Lucas-

Washburn model, fully laminar flow is assumed and the penetration rate (dh/dt)

is driven by the total effective pressure acting on the liquid (ΣP) and slowed down

by the viscous flow (PF). When setting the slip coefficient to zero, which is true

when the liquid wets the pore walls (Washburn, 1921), the penetration rate is

described by the following expression:

𝑑ℎ

𝑑𝑡=

𝑟2∑𝑃

8𝜂ℎ [Eq. 7]

When penetration is driven by external and capillary pressure, the expression

takes the form 𝑑ℎ

𝑑𝑡=

𝑟2(𝑃𝐸+𝑃𝐶)

8𝜂ℎ and when driven by capillary pressure alone it

takes the form 𝑑ℎ

𝑑𝑡=

𝑟𝛾 cos𝜃

4𝜂ℎ. In the latter case, the depth of liquid penetration can

be explicitly expressed as:

ℎ = √𝑡𝑟𝛾 cos𝜃

2𝜂 [Eq. 8]

26

The combined effect of external and capillary pressure gives (Salminen, 1988):

ℎ = √𝑡2𝑟𝛾 cos𝜃+𝑟2𝑃𝐸

4𝜂 [Eq. 9]

The force balance actually includes both spontaneous absorption (inertia of the

accelerating fluid, viscosity drag and capillary forces) and forced absorption

(external pressure) (Bosanquet, 1923):

𝑑

𝑑𝑡(𝜋𝑟2𝜌ℎ

𝑑ℎ

𝑑𝑡) + 8πηh

𝑑ℎ

𝑑𝑡= 𝑃𝐸𝜋𝑟

2 + 2𝜋𝛾𝑟 cos 𝜃 [Eq. 10]

The first term on the left-hand side is related to the inertia of an accelerating

liquid with a mass density (𝜌), at the depth of a liquid column (h) in a capillary

tube with a specific radius (r). The second term is the viscous drag of the ink. On

the right-hand side, the first term originates from the external pressure (PE), such

as the nip pressure in printing, which is usually a short time pulse. The second

term also corresponds to the capillary force. In the absence of external pressure,

a solution to the equation for short time-scales was presented by Schoelkopf et

al. (2000b):

ℎ = 𝑡√2𝛾 cos𝜃

𝑟𝜌 (

8𝜂

𝑟2𝜌𝑡 ≪ 1, 𝑃𝐸 = 0) [Eq. 11]

The Bosanquet model assumes that there is constant pressure, although the

pressure exerted from two rotating cylinders is a time-dependent pressure pulse.

To account for this, Yang (2013) changed the constant pressure in Eq. 10, 𝑃𝐸 , to

a time dependant function, PE(t). It was shown that amplitude and duration of

the pressure pulse were of importance in relation to the liquid penetration.

There are some important differences between the Lucas-Washburn model

(Eq. 8) and the Bosanquet model (Eq. 11). According to the latter, the depth of

liquid penetration is inversely proportional to the square root of the capillary

radius, while the Lucas-Washburn model states the opposite. In other words, the

Bosanquet model indicates that ink initially moves faster in smaller pores, since

it is slowed down more by the liquid density in larger pores, after which liquid

moves faster into the larger pores. Calculations based on the Bosanquet model,

made by Ridgway et al. (2002), predict that a 100 µm capillary is filled faster than

a 10 µm capillary after t=0.26 ms. According to the Lucas-Washburn model,

liquid always penetrates faster in larger pores. The depth of ink penetration is

linearly proportional to contact time in the Bosanquet model, while the Lucas-

Washburn model states that it is proportional to the square root of the contact

time. Eklund and Salminen (1987) suggested capillary driven penetration to be

27

linearly proportional to time and, when adding an external pressure, it instead

became linearly proportional to the square root of contact time.

2.4.3.1 Comments on mechanically forced penetration

Salminen (1988) stated that pressure penetration into uncoated papers is a key

factor in many industrial processes. Applying an external pressure will change

how the substrate and liquid interact. The impact from pore structure increases

under external pressure, while at the same time, the impact from surface tension

and the substrate’s surface chemistry decrease (Salminen, 1988; Rennes and

Eklund, 1989). The external pressure only contributes during the nip-pressure

pulse, whereas the capillary movements are active for a longer period.

Yang (2013) illustrated that the maximum penetration depth is reached at the

central point of a nip. Pressure penetration pushes the ink into the substrate when

approaching the nip centre and then pulls it out again, in the regime between the

nip centre and the exit. However, the capillary force holds a portion of the ink

inside the pores until leaving the nip.

A higher nip pressure can increase ink penetration and, thereby, increase the

optical density of solid tones (Holmvall and Uesaka, 2008). It has also been

indicated that the greatest influence of high nip pressure on print quality is

increased squeezing of the ink, which results in ink spreading and dot gain

(Megat Ahmed et al., 1997; Bohan et al., 2000; Dubé et al., 2008).

2.4.3.2 Scaling from a single capillary to a porous network

The absorption models, Lucas-Washburn and Bosanquet, often deal with

absorption into a single capillary tube. This needs to be scaled up to a porous

coating layer consisting of many capillaries. There are different approaches to this

and, in many cases, the coating layer is approximated by a bundle of capillaries

of a single size and without any interconnection among them. This implies the

multiplication of the volume absorbed in one capillary by the total number of

pores. The volume absorbed per unit area (V) will depend on the number of

capillaries (N), their radius (r) and the length of penetration as given by Eq. 8:

𝑉 = 𝑁𝜋𝑟2√𝑡𝑟𝛾 𝑐𝑜𝑠 𝜃

2𝜂 [Eq. 12]

To approximate the actual distance the liquid has travelled, the porosity and

tortuosity of the pore structure have often been used, as described by Gate and

Windle (1973) and Lepoutre (1978a). The porosity (ε) is normally defined as the

ratio of the pore volume of the coating (or paper) to the total volume of the

coating (or paper), and is expressed as a fraction or percentage (Murakami and

28

Imamura, 1983). Lepoutre (1978a) described the porosity of the coating layer by

the number of pores, their radii and lengths (L), divided by the coating thickness

(E): 𝜀 =𝑁𝜋𝑟2𝐿

𝐸. Furthermore, the pores are usually not straight capillaries, but

tortuous, with the tortuosity factor (τ) being defined as 𝜏 = 𝐿/𝐸. Hence, the

porosity can be written as 𝜀 = 𝑁𝜋𝑟2𝜏. This gives the absorbed volume per unit

area as a function of porosity and tortuosity:

𝑉 =𝜀

𝜏√𝑡

𝑟𝛾 𝑐𝑜𝑠 𝜃

2𝜂 [Eq. 13]

Preston et al. (2002) showed that offset ink setting is mainly determined by the

pore density (the number of pores per area of coating) and the pore size. It is

often stated that small pores set ink faster than larger ones, when the coatings are

compared at an equal pore volume. This suggestion comes from the idea that

structures with finer pores also consist of a greater number of pores. However,

when compared at an equal pore density, the larger pores display a faster ink

setting. Ma et al. (2009) showed an inverse dependence between the flexographic

ink setting time (time to reach maximum tack) and the pore-wall surface area

(∑𝜋𝐷𝑖ℎ). This, in turn, was proportional to the combined effect of pore volume

and pore size (∑𝑉𝑖 𝐷𝑖⁄ ).

Schoelkopf et al. (2002) showed that the equivalent hydraulic radius calculated

from the liquid uptake determined experimentally (using the Lucas-Washburn

model) can deviate from the pore radius measured by mercury intrusion.

Furthermore, Ma et al. (2005) concluded that there are bigger differences in

capillary forces among small pores than among larger ones. Thus, the pore-size

distribution may need to be considered, instead of making approximations from

a single pore size. It has been proposed to calculate the absorption as the sum of

each pore size (Xiang et al., 2004; Ma et al., 2007).

Adding a pore-size distribution to the bundle of capillaries may give a more

accurate approximation, but this still ignores the fact that the pores in the coating

layer are interconnected. Imbibition into a network of connected pores with a

porosity, connectivity and defined sizes of pores and throats can be modelled

from, for example, mercury intrusion data, using the PoreXpert software

(PoreXpert Ltd., Plymouth, United Kingdom). Schoelkopf et al. (2000a; 2000b)

concluded that liquid is selective in filling a pore network and that inertial

preference controls the selectivity for both short and long periods, which creates

a “preferred pathway”. Initially, liquid prefers fine pores and moves faster in

them, whereas the larger pores are excluded. As penetration proceeds, the inertial

29

effects decline and, instead, the absorption follows the Lucas-Washburn

dynamics. When new pores are reached, inertia will still be effective and exclude

the very large pores. Consequently, only the preferred paths in the network will

be filled, leaving other paths unfilled. Ridgway et al. (2002) have shown that not

only is the pore radius important in the filling of individual elements, but also the

aspect ratio of the capillaries (length divided by the radius).

Similar to the preferred pathway, Xiang and Bousfield (2003) showed

experimentally that an offset ink mainly filled the small pores in a coating.

However, the possibility of an ink vehicle initially filling the large pores was not

excluded. The ink vehicle in these pores would then be drained and pulled by

capillary pressure into the small pores.

2.4.4 Wetting and surface energy

When a solid, liquid and vapour interact, their interfacial energies will determine

how the liquid spreads over the solid. There will be a balance of adhesive forces

(between the liquid and the solid) and cohesive forces within the liquid. A drop

will spread and wet the solid when the cohesive force is weaker, and the adhesive

force is stronger. The interaction can be characterised in terms of the contact

angle that the liquid makes with the solid. Using Young’s equation, the contact

angle (θ) can be determined from the interfacial energies between the solid-

vapour (γsv), the solid-liquid (γsl) and the liquid-vapour (γlv) (see Eq. 14 and

Figure 7). The density of air is much lower than that of the solid and the liquid.

Thus, the impact from air on the interfacial tensions is negligible. The

thermodynamic work required to separate the interface of two phases, such as

liquid and solid, defines the work of adhesion (WA), Eq. 15. Interfacial attraction

between phases acts to increase the work of adhesion. The work required to

separate the interface when the two phases are identical, for example, one liquid,

defines the work of cohesion (WC), Eq. 16. Finally, the spreading coefficient (S),

Eq. 17, is defined as the difference between work of adhesion and cohesion.

𝛾𝑠𝑣 = 𝛾𝑠𝑙 + 𝛾𝑙𝑣 cos 𝜃 [Eq. 14]

𝑊𝐴 = 𝛾𝑠𝑣 + 𝛾𝑙𝑣 − 𝛾𝑠𝑙 [Eq. 15]

𝑊𝐶 = 2𝛾𝑙𝑣 [Eq. 16]

𝑆 = 𝑊𝐴 −𝑊𝐶 = 𝛾𝑠𝑣 − 𝛾𝑙𝑣 − 𝛾𝑠𝑙 = 𝛾𝑙𝑣(cos 𝜃 − 1) [Eq. 17]

30

Figure 7. Diagram showing the interfacial energies and contact angle

formed by a liquid drop, a solid and vapour. Young’s equation,

Eq. 14, describes the interactions.

Ideally, an equilibrium contact is formed between the solid/liquid interface and

the liquid/vapour interface. This requires the solid to be non-porous, flat,

chemically homogeneous, insoluble and non-reactive by the liquid. The initial

penetration into porous materials will also impact on the contact angle

measurement (Elftonson and Ström, 1995). Experimental measurements capture

the apparent contact angle. This can be made at a low metastable energy state,

but the equilibrium energy state cannot be captured. To ensure that the drop is

relaxed, it can beforehand be subjected to a stimuli, such as vibrations (Ruiz-

Cabello et al., 2014). The contact angle around a drop, on a heterogenous surface,

is not equal on the microscopic level. Chemical heterogeneities and surface

roughness cause local protrusions or pinning of the liquid front, which give a

contact angle that diverge from the rest of the drop and the macroscopic average

(Oliver and Mason, 1973; Duursma et al., 2010). The area fraction and contact

angle of individual patches, of a chemically heterogenous surface, will determine

its contact angle. This behaviour is described, for surfaces with large-scale

patches, by the Cassie equation and when the patches approach molecular or

atomic dimensions, a modified equation by Israelachvili and Gee (1989) can be

used. The roughness of a substrate can either enhance or reduce the wetting

behaviour and the outcome depends on the surface chemistry. Conditioned that

the liquid fills the structure, a rough surface increases the wetting if contact angle

is below 90⁰, and decreases the wetting if it is above 90⁰ (Wenzel, 1936).

There is often a hysteresis between the advancing (maximum) and receding

(minimum) contact angles. This means that the angle will not be the same when

the liquid spreads out and recedes over the solid. A heterogeneous surface can

give the moving liquid front a stick and slip behaviour, since liquid pins to certain

areas until the energy barrier is overcome, and then the liquid progresses to the

next metastable equilibrium (Brandon and Marmur, 1996).

γsv

γlv

γslθ

Solid

VapourLiquid

31

The total surface energy (𝛾𝑇𝑂𝑇) of a solid or liquid can be divided into apolar and

polar components. According to van Oss et al. (1987), the apolar component is

assigned to electrodynamic interactions, termed Liftshitz-van der Waals (𝛾𝐿𝑊)

and are comprised of London dispersion, Keesom dipole-dipole and Debey

dipole-induced dipole forces. The polar component relates to acid-base

interactions (𝛾𝐴𝐵) and comprises hydrogen bond interactions (Brønsted acids

and bases), Lewis acid and base interactions and dipole moment interactions. The

basic behaviour (𝛾−) is due to proton acceptor or electron donor functionality,

whereas the acid behaviour (𝛾+) is due to proton donor or electron acceptor

functionality. Surfaces can have acidic and/or basic groups and functionalities.

When both are present, it is termed bipolar, as opposed to a monopolar surface,

where one of the functionalities is dominant (van Oss et al., 1987).

Furthermore, Owens and Wendt (1969) divided the surface energy into apolar

(dispersive) and polar components. Here, the dispersive component (𝛾𝑑)

constitutes London dispersion forces only and the polarity (𝛾𝑝) comprises

hydrogen bonds and dipole-dipole interactions. van Oss et al. (1987) suggested

that the Keesom and Debey forces only make a small contribution to the surface

and interfacial tensions, hence, 𝛾𝐿𝑊 and 𝛾𝑑 may be regarded as comparable.

The surface energy of a solid can be calculated and separated into its components

based on contact angle measurements with liquids of known surface tension and

polarity. It is common that one of the liquids is only able to make apolar

interactions, for example, diiodomethane, which gives a straight forward estimate

of the apolar component. In the case of the Owens and Wendt approach, two

liquids are required and calculations are made according to (Owens and Wendt,

1969):

𝛾𝑙(𝑐𝑜𝑠 𝜃 + 1) = 2√𝛾𝑠𝑑𝛾𝑙

𝑑 + 2√𝛾𝑠𝑝𝛾𝑙𝑝 [Eq. 18]

𝛾𝑇𝑂𝑇 = 𝛾𝑑 + 𝛾𝑝 [Eq. 19]

The surface energy (γ) of the solid and liquid is separated by subscripts (s and l,

respectively). Superscript d denotes the dispersive, p the polar component and

TOT the total surface energy.

32

Determination of the bipolar nature of a solid in the van Oss approach requires

contact angles from three liquids (van Oss et al., 1987):

𝛾𝑙(𝑐𝑜𝑠 𝜃 + 1) = 2√𝛾𝑠𝐿𝑊𝛾𝑙

𝐿𝑊 + 2√𝛾𝑠+𝛾𝑙

− + 2√𝛾𝑠−𝛾𝑙

+ [Eq. 20]

𝛾𝑥𝑎𝑏 = 2√𝛾𝑥

−𝛾𝑥+ (𝑥 = 𝑠, 𝑙) [Eq. 21]

𝛾𝑥𝑇𝑂𝑇 = 𝛾𝑥

𝐿𝑊 + 𝛾𝑥𝑎𝑏 (𝑥 = 𝑠, 𝑙) [Eq. 22]

Superscript (LW) denotes the apolar (Liftshitz van der Waals), (ab) the polar

component, (+) the acidic character and (–) the basic character.

2.4.5 Dewetting

Dewetting describes the rupture of liquid films on non-wettable solids, which the

liquid was first forced to cover. The research area is highly relevant, for instance,

to the design of super-hydrophobic self-cleaning surfaces (Yilgör et al., 2018) and

the formation of 3D patterns by controlled dewetting (Wu et al., 2015). This area

is briefly touched upon here, with no attempt to fully cover it.

The dewetting of thin liquid films depends on their thickness and surface energy

(that is, on the molecular interactions). A thick film stays stable due to gravity,

whereas a film that is below a critical thickness may rupture and contract to seek

a new low energy equilibrium. Xue and Han (2012) explained that the thickness

becomes critical when the surface energy forces dominate the gravitational

forces. Thin unstable films rupture through “spinodal dewetting” when capillary

waves on the films spontaneously amplify. This forms a characteristic pattern.

Metastable films dewet through nucleation, where, for example, dust particles or

defects induce the rupture. Chemical heterogeneities will induce liquid to flow

from areas with low to high wettability and, eventually, the film ruptures in areas

where it has become too thin (Kargupta et al., 2001). In coating applications,

higher speeds are known to cause dynamic wetting problems when air is

entrapped underneath the coating layer and this air serves as a nucleation site for

dewetting (Blake and Ruschak, 1997).

When holes have formed, a negative spreading coefficient (Eq. 17) will drive the

hole growth. This may be inhibited by, for example, lowering the contact angle

between liquid and solid or by reducing the mobility of the liquid (for example,

increasing viscosity) (Redon et al., 1991; Xue and Han, 2012). A hole in a water

film will close if the film is thick, whereas it grows if the film is thin and the

contact angle is in the mid-range (Mulji and Chandra, 2010). Furthermore, Mulji

and Chandra (2010) found that surface roughness promotes wetting and inhibits

the dewetting of thick films, because liquid becomes attached to protrusions.

33

Similarly, Luo and Gersappe (2004) identified that dewetting is retarded by

pinning to immobile nanofillers, although their size must be below the film

thickness. The surface roughness of the solid can initiate dewetting in unstable

thin films (spinodally), when the structure has both a high frequency relative to

the spinodal wavelength and a large enough amplitude (Volodin and Kondyurin,

2008). Kim and Kim (2018) showed that a liquid film follows geometrical features

in the solid’s surface when it retracts and may leave residual liquid in groves

depending on the geometry and contact angle.

2.4.6 Liquid rheology

The rheological aspects of an ink will affect its flow properties and, thereby, its

behaviour during the printing process, such as its transport through the press,

but also its transport onto or into the paper. As already mentioned, the ink

viscosity will slow down capillary absorption and dewetting.

The viscosity is defined as the resistance to deform (flow) when a force acts on

the liquid (that is, when applying a shear stress). For Newtonian liquids that are

sheared between two parallel plates, see Figure 8, the viscosity (η) is calculated as

the shear stress (τ) divided by shear rate (�̇�) (Leach and Pierce, 1993):

𝜂 = 𝜏 �̇�⁄ [Eq. 23]

The shear stress is defined as the force (F) per unit area (A), 𝜏 = 𝐹 𝐴⁄ , and the

shear rate is the velocity gradient, �̇� = 𝑑𝑣 𝑑𝑦⁄ (Leach and Pierce, 1993). Eq. 23

gives the apparent viscosity for non-Newtonian fluids.

Figure 8. Diagram showing the velocity gradient within a fluid that is

sheared between one moving and one stationary plate. When a force

(F) is applied to the moving plate, the velocity (v(y)) in the fluid will

depend on the y-position between the plates.

Water is Newtonian, but a flexographic ink deviates from this behaviour and the

flow response to shear stress is no longer linear. The particle association,

chemical bonds and physical interactions during flow will change the linear

FMoving plate

Stationary platex

yv(y)

34

relationship between shear rate and shear stress (Leach and Pierce, 1993).

Viscosity responses that are of interest when considering inks:

• Shear thinning (pseudo-plastic): Viscosity decreases with increasing shear rate.

• Shear thickening (dilatant): Viscosity increases with increasing shear rate.

This can occur when particles pack and form rigid structures.

• Tixotropic: Viscosity decreases during shearing at a constant rate and stress.

A gel structure is formed when the ink is at rest and this progressively

breaks down when being sheared. The ink recovers to the original gel

structure when the shearing stops, but the recovery takes a longer time

than it takes to break down the structure.

• Yield value: Minimum shear stress that is needed for the ink to start to flow.

• Visco-elastic implies that the ink possesses both viscous and elastic

properties. Ink deforms when applying a stress and recovers when the

stress is removed. Leach and Pierce (1993) suggested that the visco-elastic

property is important in printing processes. This determines not only how

the ink will flow in nips, but also how it recovers after splitting, which

affects its ability to level.

Flexographic inks usually have low viscosities and are shear thinning (Aspler and

Lepoutre, 1991; Havlínová et al., 1999; Olsson et al., 2007a; Olsson et al., 2007b;

Gane and Ridgway, 2009). Olsson et al. (2007a) showed that the ink binder has a

dominant impact on the viscosity curve, whereas it was not as affected by the

pigment particle type. Furthermore, they showed that the viscosity was

temperature-dependent (in the range investigated) and decreased at higher

temperatures. A high yield value, high viscosity or excessive thixotropic

behaviour will prevent the ink flow. A viscosity which is too low may, on the

other hand, cause dot spreading (FTA, 1992).

2.5 Characterisation of liquid absorption

Commonly used techniques for absorption characterisations are described briefly

below together with a selection of the more advanced and/or non-standardised

methods.

2.5.1 Absorbed volume and absorption rate

The total absorbed volume can be determined gravimetrically or volumetrically,

either by the increased weight of a sample or the liquid from the reservoir. The

manual Cobb test is frequently used, and the liquid uptake is determined by

weighing the sample before and after absorption (Zang and Aspler, 1995a; Lyne,

35

2002; Skowronski et al., 2005). In an automated version, the Automatic Cobb

Tester, the sample is clamped against a water-filled porous membrane. Liquid

that is withdrawn from the liquid chamber is monitored continuously and gives

the absorption as a function of time (TMI, 2015).

In the Bristow wheel test, a stain is made from a defined amount of dyed liquid

(Bristow, 1967). The substrate passes beneath a liquid holder with a slit, where

the liquid can transfer. The liquid uptake is calculated from the length of the

track, while being dependent on the surface roughness and absorptivity of the

substrate. Several stains applied at different speeds are needed to obtain the

absorbed volume as a function of time. The curves reveal both the roughness

index (which is where the curve intersects with the y-axis) and the absorption

coefficient (absorption rate). Bristow (1967) presented experimental data

indicating that absorption did not start immediately on some samples and it was

suggested that this was related to a delayed wetting of the solid. The concept of

a wetting delay has been debated and Salminen (1988) claimed the term to be

misleading and that a paper surface is wetted more quickly than the shortest

contact that is experimentally measurable. As pointed out by Aspler (1993), it

might be more accurate to refer to this as absorption delay.

In a recent review, Ovaska and Backfolk (2018) mentions that several absorption

testers has a testing principle that resembles the Bristow wheel. The tester has,

for example, been equipped with corona charger and infra-red dryer and the

tester has been automated by KRK Kummagai. In the Automatic Scanning

Absorptometer (ASA), testing of different speeds has been incorporated into a

single measurement (KRK, 2015b). It is much like a gramophone, where the

liquid holder has replaced the stylus and the sample is mounted on the turntable.

As the liquid holder moves outwards at an accelerating turntable speed, it leaves

a spiral track. The liquid transfer is detected continuously, resulting in an

absorption curve as a function of time in one measurement.

A Micro Drop Absorption Tester (MicroDAT) detects the rate of absorption of

water droplets into the paper substrate (Ström et al., 2008). Water droplets that

are roughly 300 pl are formed using a piezoelectric dispenser and hit the surface

at a velocity of about 1.5 ms-1. This forces a certain amount of the volume of the

drop into the substrate and flattens the drop into a half sphere. The absorption

and evaporation are captured dynamically at a frequency of 8 ms. The MicroDAT

test is mainly used in research work and is suitable for coated grades, but not for

rough and uncoated grades.

36

There are also various other gravimetric approaches for measuring absorption.

Schoelkopf et al. (2000b) measured imbibition into compacted pigment tablets.

Similar tablet blocks were also used to measure pressurised liquid permeability

(Schoelkopf et al., 2004).

2.5.2 Contact angle and surface energy

There are various techniques for characterising the contact angle, surface energy

and surface tension. There are instruments that measure the static apparent

contact angle, whereas others capture the dynamic contact angle. It has been

suggested that the dynamics are important for understanding what takes place in

a printing nip (Bassemir and Krishnan, 1990). The surface energy of a solid can

be calculated by using contact angle measurements involving two or three liquids,

as discussed in Section 2.4.4.

The contact angle between liquid and solid, for example a board sample, can be

detected with static techniques such as goniometers. These instruments capture

images of sessile droplet profiles. These measurements can be dynamic, in the

sense that contact angles are detected as a function of time. This can reveal how

quickly a metastable state is reached. To capture the effects of hysteresis, the plate

holding the sample can be tilted to identify the maximal (downhill side) and

minimal (uphill side) angle achievable without a drop starting to roll. Another

option, to capture the hysteresis, is to add and remove liquid from the droplet,

without changing the contact line (ramé-hart, 2014).

Figure 9. Example of testing principles (not to scale), for testing of liquids’ dynamic surface

tension (1), forced wetting/dewetting (2), contact angle hysteresis (3, 4) and capillary force

(4). Numbers are connected to descriptions in text.

Static surface tension of a liquid can be measured with, for example, a Wilhelmy

plate, which detects the force acting on a solid that is lowered into and withdrawn

from the liquid. The dynamic surface tension, of liquids, can be measured with

techniques that vary the rate of wetting and contact. Some examples are

illustrated in Figure 9 and work according to the following principles: 1) In a

1) Pressure to

form gas bubbles

2) Rotating

cylinder

3) Drops sliding

on inclined plane

BALANCE

4) Compression and

stretching of liquid bridge

37

bubble tensiometer, gas bubbles are formed in the liquid at different rates and

the surface tension is calculated from the maximum pressure (Krüss, n.d.). 2)

Forced wetting and dewetting can be studied dynamically, by rotating a cylinder

at different speeds, while being half-way immersed in the testing liquid (Fell et

al., 2011). 3) The advancing and receding contact angles can be studied from the

shape of drops that slide on an inclined plane (Grande). The inclination of the

plane controls the velocity. 4) Cyclic compression and stretching of a liquid

bridge, formed between two plates, reveals the dynamic hysteresis of contact

angle (Shi et al., 2018). A balance, placed under the bottom plate, detects the

capillary force. The dynamic measurements have in common that the testing

liquid can be changed, but the solids are limited to a few and well controlled

materials. This means that the dynamic interaction between liquid and a paper

substrate cannot be tested this way.

2.5.3 Oil-based ink stain test

There are various oil-based stain tests, such as K&N, Croda and Noiré

Porometric. Although the testing oils differ, the working principle is the same for

each test. Dyed viscous oil is absorbed for a given period, before any excess is

removed, and the reflectance of a stain is used as an indicator of absorbency

(Bristow and Bergenblad, 1981). The reflectance of the stain depends on the oil

pick-up, as it changes with the oil distribution in the z-direction (Hattula and

Oittinen, 1982). Over the years, some modifications to the tests have been

explored, such as weighing the samples after different contact times have elapsed

in order to obtain the absorption rate (Ranger, 1981) and calculating the void

fraction of a coating (Larrondo and St-Amour, 1992).

2.5.4 Ultra sound attenuation

There are also ultrasound techniques (adopted from use in metallic materials

testing and in medical analysis), where various liquids can be used as testing

liquids. The transmission of ultrasound signals through a sample will change as

the liquid penetrates the sample, due to the reflectance of the ultrasonic wave at

material interfaces. The ultrasound technique allows for relative comparisons but,

as pointed out by Lyne (2002), the exact relationship between the ultrasonic

signal response and liquid absorption has not been well defined. Yang et al. (2013)

proposed a physical model to convert the transmittance values to penetration

depth into uncoated papers. It was stressed that it would be more challenging to

create a model for a multi-layered coated board. In his master’s thesis, Bergvall

(2006) found that it was possible to determine the time it takes to penetrate

through a coating layer.

38

2.5.5 Flexographic ink-tack test

Ma et al. (2009) introduced a “micro-tack” device to measure ink-tack curves for

water-based flexographic inks. This device produces curves similar to those for

offset inks, by using, for instance, the ISIT device. The ink setting is defined as

the time needed to reach maximum tack. A 2 mm x 2 mm rubber blanket probe

is covered with ink and then brought into contact with paper. The force needed

to pull the probe away from the paper is measured. As liquid volumes in the test

are larger than in real printing, the time-scale is also longer and takes couple of

hundred seconds to reach maximum force.

2.5.6 Pressurised absorption

There are not many commercially available techniques that characterise forced

absorption. Various attempts have been made to pressurise the Bristow

technique. One example of this is from Kummagi KRK (KRK, 2015a), where

the traditional Bristow wheel is equipped with a pressurised head-box (a contact

pressure of 0.1 MPa to 0.5 MPa and liquid pressurisation of 0.003 MPa to

0.01 MPa).

The Clara Device is a measuring apparatus for research purposes (Lamminmäki

et al., 2010; Lamminmäki et al., 2011), which is based on the capacitance changes

during liquid absorption by paper. By adjusting the pressure inside a liquid

chamber, it is possible to study absorption under different external pressures,

ranging from -0.5 bar to 5 bars.

In the IGT print penetration test (W24), a lab printer is equipped with a syringe

containing red oil (Diethylhexyl adipate) (IGT, 2006). An oil droplet falls onto a

print disc and is then squeezed in a nip between the disc and sample, at an

accelerating speed. The length of the stain is related to the absorbency of the

substrate, so that the faster the absorption occurs, the shorter the stain is. It has

also been used with diluted flexographic inks (Steadman et al., 1993) and with

dyed water (Wågberg and Wennerblom, 1992).

2.5.7 Absorption uniformity

Uniformity in absorption is seldom characterised, particularly not in relation to

water-based ink absorption. Variations may be roughly estimated using multiple

measurements with some of the already mentioned techniques, such as

MicroDAT, the contact angle of sessile drops or the flexographic micro-tack

device. Skowronski et al. (2005) introduced a multi-sensor to the ultrasound

technique. The evenness of the signal is detected with a row of 32 sensors, each

1 mm². It is not clear if this is commercially available anymore but was previously

39

a module of the Dynamic Penetration Analyser (emtec Electronic GmbH).

Another option is to examine the evenness of the Bristow Test stains. However,

this includes an influence from the roughness, due to a filling of surface pores,

and does not only reflect variations in penetration and wettability (Bristow, 1967).

In the absorption curves, roughness index and absorption rate can be separated,

but in the stains, it is not possible to separate reflectance variations that originate

form roughness filling from those originating from capillary absorption.

When it comes to offset printing, the oil-based staining techniques mentioned

above can give an indication about uniformity in oil absorption. Xiang et al.

(1999) introduced the Micro-Tackmeter, which measures the force required to

pull an inked probe (millimetre-scale contact area) off the paper surface after

contact has been made. This splitting force is measured repeatedly at the same

location to indicate the tack build-up. Variations are captured by means of

multiple measurements distributed across the surface.

40

3 Experimental In this section, the methods and materials are accounted for and their advantages

and disadvantages are discussed. However, full descriptions are given in the

respective papers. The techniques developed in the doctoral project are

addressed with some additional details.

3.1 Board and paper substrates

Coated substrates of paper and board types were used in these studies. Papers I,

V and VI included substrates from a set of pilot-coated boards, Paper IV included

pilot-coated fine papers and Paper II included commercially produced boards. In

Paper III, surface modifications were made to coated boards.

3.1.1 Coated boards

Fourteen pilot-coated substrates were produced from the same pre-coated

commercial 200 gm-2 duplex base board (Klabin, S.A., Brazil) by different coating

formulations in the top-coating layer. The formulations were divided into four

groups (Table 1): 1) increasing the binder content of latex type A, 2) increasing

the binder content of type B, 3) decreasing the clay content in favour of ground

calcium carbonate (GCC) pigment and 4) GCC pigments of different kinds.

Changing the coating colour compositions by using different mineral pigments

and latexes, and their dosage, influenced pore structure and surface energy. At

the same time, the roughness range was limited since the coating layers were

applied to the same pre-coated base board. As a result, the influence of properties

other than surface roughness was made easier to distinguish.

Pre- and top-coating layers of roughly 12 gm-2 each were applied with a blade

coater in the pilot facilities at BillerudKorsnäs (Frövi, Sweden). The coating

colours were formulated with the following constituents in various portions:

• Binders: Latex type A (vinyl acetate acrylate with Tg 15⁰C, Acronal S722

from BASF, Ludwigshafen, Germany) or type B (styrene butyl acrylate

with Tg 23⁰C, Acronal S722 from BASF, Ludwigshafen, Germany).

• GCC mineral pigments: Covercarb® 75 (narrow particle size distribution,

PSD, and 95 % of its particles < 2 μm), Hydrocarb® 90 (broad PSD and

90 % < 2 μm) and Setacarb® HG (broad PSD, 98 % < 2 μm). These are

referred to as CC, HC90 and SC, respectively. These were delivered from

Omya International AG (Oftringen, Switzerland).

• Clay pigment: Brazilian delaminated clay Capim NPTM (Imerys Minerals

AB, Sundsvall, Sweden).

41

Table 1. Coating formulations for the top-coating and pre-coating layers. Sample codes for

the separate studies are indicated. Each constituent is given as parts per hundred parts

pigment (pph). *Note that sample P01 occurs twice.

Sample codes

Group Paper V, VI Paper I Paper III Coating formulation

1. Increase latex A

↓

P00 12.5 latex A Pigments: HC90:SC (60:40 pph)

Binder: Latex A (12.6, 15, 17.5, 20.0 pph)

Thickener 0.5 pph and soda 0.08 pph. P01* 15 latex A P1

P02 17.5 latex A

P03 20 latex A P2

2. Increase latex B

↓

P04 Pigments: HC90:SC (60:40 pph)

Binder: Latex B (12.6, 15, 19.9 pph)

Thickener 0.5 pph and soda 0.08 pph. P05

P06 P3

3. Decrease Clay

↓

P07 40 clay Pigments: Clay HC90 (40:60, 20:80, 10:90 pph)

Binder: Latex A (15 pph)

Thickener 0.5 pph and soda 0.08 pph. P08 20 clay

P09 10 clay

4. Different GCCs

↓

P10 B90 GCC Pigments:

P10: 100 pph of HC90,

P01: 60:40 pph of HC90:SC,

P11: 100 pph of SC, P12: 20:80 pph of SC:CC,

P13: 100 pph of CC

Binder: Latex A (15 pph)

Thickener 0.5 pph and soda 0.08 pph.

P01*

P11 B98 GCC

P12

P13 N75 GCC P4

Pre-coating 100 pph Hydrocarb® 60 and latex A 13 pph.

Thickener 0.5 pph and soda 0.08 pph.

Commercial coated liquid packaging boards for flexographic printing (provided

by Stora Enso, Sweden; BillerudKorsnäs, Sweden and Tetra Pak, Sweden) was

included in Paper II (named A-F) and in Paper III (named CA and CB). One cast-

coated offset grade was included in Paper III (Zanders GmbH, Germany).

3.1.2 Pilot-coated papers

The pilot-coated papers were produced and printed within a separate study

(Kamal Alm et al., 2015b). An industrially pre-coated wood-free 80 gm-2 paper

(Sappi Europe, a division of Sappi Ltd, Johannesburg, South Africa) was pilot-

coated using a blade coater (Modular Combi Blade manufactured by Voith Paper

GmbH & Co. KG, Heidenheim, Germany) and then supercalendered (SK 14/12

manufactured by Bruderhaus Maschinen GmbH, Reutlingen, Germany).

42

For each sample, calcium carbonate-based coating formulations were modified

only in the additional dosages of dispersants. A commercial dispersant, Dispex

N40 (BASF SE, Ludwigshafen, Germany), of sodium polyacrylate (referred to as

NaPA) was used. It was also modified with the addition of calcium nitrate

tetrahydrate (Merck KGaA, Darmstadt, Germany), which increased the calcium

ion concentration and partially calcium-neutralised the sodium polyacrylate

(referred to as NaCaPA). One reference coating was included, in which the pre-

dispersed pigment slurry (with sodium polyacrylate) was used without

modification. In the other coating colours, the additional doses of NaPA or

NaCaPA were 0.2 pph, 0.4 pph and 0.8 pph, respectively.

3.1.3 Absorption-modified surfaces

The absorption non-uniformity of coated board samples was modified by adding

a barrier pattern that locally created compact areas in the coating layer. The

patterns consisted of a coarse halftone screen (31 lpi). And different nominal

tone-values (ranging from 2 % to 14 %) generated various levels of absorption

non-uniformity, Figure 10 (characterised by the absorption staining that detects

variations in liquid uptake over the surface, see Section 3.4). The absorbency of

seven coated boards (Section 3.1.1) was modified by so-called barrier patterns to

generate a “property matrix”, Figure 11. The influence on print mottle of

absorption non-uniformity was studied on surfaces where the level of surface

roughness the same. Vice versa, the influence of surface roughness was studied

on different substrates where absorption non-uniformity was interpolated to the

same level.

The patterns were applied with a laboratory flexographic press (F1 Printability

Tester from IGT Testing Systems, Amsterdam, Netherlands) and a flexographic

ink vehicle (colorant excluded). Details are given in Paper III.

To acquire relevant information from the matrix, the barrier modified areas only

varied in pore-structure and surface chemistry, but not surface roughness. This

was verified by (a) determination of the surface energy of the untreated board

and solid tones with the barrier material, (b) SEM imaging and the surface pore

structure of a modified surface on two boards and (c) the topography of a

modified surface on each board, measured from replicas of the surfaces.

43

Figure 10. Example of barrier patterns of

different tone-values that cause absorption

non-uniformity on sample P1 [reproduced

with permission, Thorman et al. (2018b)].

Figure 11. Illustration of “property matrix”.

Print quality may be compared when the

impact of roughness or absorption non-

uniformity are isolated [redrawn from

Thorman et al. (2018b)].

3.2 Printing

Substrates in three of the studies were printed in a flexographic laboratory press

(Paper III) or a flexographic production-scale press (Paper V and VI). In both

cases, water-based inks were used. In Paper IV, water/moisture interference in

offset printing was studied in a production-scale sheet-fed offset press.

3.2.1 Laboratory-scale flexographic printing

Boards with absorption-modified surfaces were best handled at the laboratory-

scale, where parameters can be controlled with greater accuracy and samples

treated individually. A flexographic laboratory press was used (F1 Printability

Tester, IGT Testing Systems, Amsterdam, Netherlands). Additional details are

given in Paper I. The pilot-coated papers were also laboratory printed. These

results have not been published and details of the testing are given in Appendix A.

3.2.2 Full-scale flexographic printing

The pilot-coated boards were printed in an inline production flexographic press

(VT-Flex 175ES) at Tetra Pak (Lund, Sweden), where the speed was high and

multicolour printing possible. This emphasised the influence of ink absorbency.

Two separate printing speeds, 7 ms-1 and 10 ms-1, were run and cyan ink was

applied in two units and magenta in one. The units were equipped with aniloxes

of separate volumes (4.0 cm³·m-2 as standard and 6.2 cm³·m-2 as high volume),

and identical polymer plate materials. Monochrome areas of cyan (referred to as

monolayers) and polychrome areas where magenta overprinted cyan (referred to

as bilayers) were evaluated, Figure 12. Details are given in Papers V and VI.

Ab

so

rpti

on

no

n-u

nif

orm

ity

,S

.Dev

[%]

Wavelength [mm]

0.0

0.4

0.8

1.2

1.6

0.06

-0.1

2

0.12

-0.2

5

0.25

-0.5

0.5-

1.0

1-2

2-4

4-8

2%

5%

8%

14%

Absorp

tion n

on-u

niform

ity

Surface roughness

Impact of

roughness

Impact of

absorbency

44

Figure 12. Illustration (not to scale) of the press set-up and the inked areas that were

analysed. [Paper V]

3.2.3 Full-scale offset printing

Pilot-coated papers were printed in a production sheet-fed offset press

(Manroland R 706 LTTLV, Manroland sheet-fed GmbH, Offenbach am Main,

Germany) within the study presented by Kamal Alm et al. (2015b).

Water/moisture interference issues were in focus. Therefore, an elevated feed of

fountain solution was used, and cyan ink was printed in the second, fifth and

sixth printing units. Additional details are given in Paper IV.

3.3 Print quality analysis

In the print quality analysis, ink layer non-uniformities (print mottle) and

uncovered areas (UCA) in solid tones were prioritised. Mean optical density was

also evaluated. Halftones were not included in the work presented here. Light

microscopy was utilised to study the prints qualitatively and was often used for a

limited selection of samples to gain an understanding of the print quality data.

3.3.1 Print mottle and uncovered areas

Print mottle was analysed from reflectance variations in printed monolayers and

bilayers. A flatbed scanner (Epson Perfection V750 Pro, or alternatively V850

Pro, from Epson Europe B.V., Amsterdam, Netherlands) captured images of the

prints, together with a calibration frame. Reflectance levels of the areas in the

calibration frame were well-defined and enabled the calibration of each pixel

value to a reflectance value. Calibration and analysis of mottle were made with

the software STFI Mottling Expert (RISE Bioeconomy, Stockholm, Sweden),

where outputs were: mean reflectance, �̅�, coefficient of variation of reflectance,

𝐶𝑉𝑅, and uncovered area (UCA) fraction. UCAs were identified with a threshold

that was set relative to the median reflectance of the image.

In Papers III, V and VI, the mottle data was processed further. In Paper III,

standard deviation of reflectance, 𝜎𝑅, was obtained from 𝜎𝑅 = 𝐶𝑉𝑅 �̅�⁄ .

Consequently, absorption-modified surfaces were compared accurately even

Paper web

MC CM

C.Vol C.Vol

M

Printed areas(ink 1)

C.Vol

Cyan

Anilox: High

volume

First unit

(ink 1)

C

Cyan

Anilox:

Standard

Second unit

Special case,

not evaluated

now

Third unit

Ink 2

(M)

Magenta

Anilox:

Standard

Fourth unit

45

when the barrier pattern covered different surface fractions. Ink trapping

analyses from Papers V and VI are described in Section 3.3.2.

Fast Fourier Analysis (FFT) was utilised to divide reflectance variations into

lateral wavelength bands (ranges). The visual appearance of non-uniform prints

were expected to correspond to the measured print mottle in wavelengths

ranging from 1 mm up to 8 mm (Johansson, 1993; Johansson and Norman, 1996;

Fahlcrantz and Johansson, 2004). However, disturbances of shorter wavelengths

also occurred and may be relevant for the understanding of ink distribution and

UCAs. Therefore, the studies encompassed both fine-scale (0.25-1 mm) and

large-scale (1-8 mm) non-uniformities, which were treated separately in Papers IV

to VI, and combined (0.5-8 mm) in Paper III. Measurement areas from laboratory

printing were too small to acquire information from the wavelengths above

8 mm.

Images were analysed in grey-scale (Papers III and IV) or in the individual channels

of RGB images (Papers V and VI). Grey-scale images were standard in STFI

Mottling analyses and commonly showed correlation with the visual appearance

(Johansson, 1993; Johansson and Norman, 1996; Fahlcrantz and Johansson,

2004). From RGB images, the channel of the complimentary colour of the ink

can be used. This offered greater contrast and sensitivity to ink thickness

variations. RGB images were used in the ink trapping analyses and when

analysing absorption stains. Further details are given in Papers III to VI.

Details regarding the print mottle analysis of flexographic prints on the pilot-

coated papers have not been published and details are given in Appendix A.

3.3.2 Ink trapping non-uniformity

The multicolour prints (Papers V and VI) contained monolayers where cyan

(ink 1) and magenta (ink 2) were each printed directly onto the board and bilayers

where magenta overprinted cyan. Figure 13 illustrates the terminology.

Figure 13. Terminology used for the polychrome and monochrome areas. [Paper V]

Bilayer Monolayer

Overprint layer

Bottom layer Ink 1 (cyan)

Ink 2 (magenta)

Ink 1 (cyan) Ink 2 (magenta)

Monolayer

46

3.3.2.1 Mean optical density, trapping ratio and overprint density

The relative print density, �̅�𝑟𝑒𝑙, related to the mean reflectance of the printed,

�̅�𝑝𝑟𝑖𝑛𝑡 , and unprinted, �̅�𝑜. Both was measured with the complimentary filter of

the ink in focus:

�̅�𝑟𝑒𝑙 = log �̅�𝑜 − log �̅�𝑝𝑟𝑖𝑛𝑡 [Eq. 24]

Ink trapping was calculated by Preucil trap ratio, 𝑇𝑅̅̅ ̅̅ , and optical densities were

given by the complimentary filter of the second ink, that is, green filter, when

magenta was overprinting cyan (Preucil, 1953 )3:

𝑇𝑅̅̅ ̅̅ =�̅�12𝑟𝑒𝑙−�̅�1

𝑟𝑒𝑙

�̅�2𝑟𝑒𝑙 × 100 [Eq. 25]

Subscripts 1, 2 and 12 refer to the first ink (cyan), the second ink (magenta) and

the bilayer of both inks, respectively. The numerator in Eq. 25, defines mean

overprint density, �̅�𝑜𝑣𝑒𝑟:

�̅�𝑜𝑣𝑒𝑟 = �̅�12𝑟𝑒𝑙 − �̅�1

𝑟𝑒𝑙 = �̅�12 − �̅�1 [Eq. 26]

3.3.2.2 Non-uniformities in monolayer, overprint layer and ink trapping

Mottle values from STFI Mottling defined how reflectance values varied around

their mean. Thus, mottle was an error value. In Papers V and VI, standard

deviation of relative optical density, 𝜎𝐷𝑟𝑒𝑙, was used for the characterisation of

print mottle of the monolayers. The variation in relative optical density depended

on variations within inked areas and un-inked board surface. It was estimated by

applying conventional rules of error propagation to Eq. 24 (Beers, 1957). When

errors (reflectance variations) were assumed to be uncorrelated and independent:

𝜎𝐷𝑟𝑒𝑙 = √(

𝜎𝑅(𝑝𝑟𝑖𝑛𝑡)

�̅�𝑝𝑟𝑖𝑛𝑡× log10 𝑒)

2

+ (𝜎𝑅0

�̅�0× log10 𝑒)

2

= log10 𝑒 √(𝐶𝑉𝑅0)2+ (𝐶𝑉𝑅𝑝𝑟𝑖𝑛𝑡)

2

[Eq. 27]

Where CV is the coefficient of variation, σ the standard deviation and subscripts

𝑅(𝑝𝑟𝑖𝑛𝑡) denote the reflectance of the printed area and 𝑅0 the reflectance of

unprinted board. For clarification, reflectance variations of the board described

white-top mottle.

Similarly, standard deviation of optical density in a bilayer, 𝜎𝐷12𝑟𝑒𝑙 , was the

combined effect of the variations in the individual layers: ink layer one, 𝜎𝐷1𝑟𝑒𝑙, and

overlayer, 𝜎𝑜𝑣𝑒𝑟 . The first layer and bilayer were determined experimentally and,

3 The original source has not been read. Information about the Preucil formula is found, for example, in Hauck & Gooran (2013) and TECHKON (2000).

47

from this, the contribution of the overlayer was estimated. Applying the rules of

error propagation to Eq. 26 yielded:

𝜎𝑜𝑣𝑒𝑟 = √(𝜎𝐷1𝑟𝑒𝑙)2 + (𝜎𝐷12

𝑟𝑒𝑙 )2 − 2𝜌(𝐷1𝑟𝑒𝑙,𝐷12𝑟𝑒𝑙)𝜎𝐷1𝑟𝑒𝑙𝜎𝐷12

𝑟𝑒𝑙 [Eq. 28]

Absolute, instead of relative, optical densities can be used when estimating the

overprint density. Each layer was characterised in the G-channel, which was

complimentary to the second (magenta) ink. The third term, under the square

root sign, related to the dependence (or independence) between local values in

the bilayer and corresponding values in the first ink layer. The correlation

coefficient, 𝜌(𝐷1𝑟𝑒𝑙,𝐷12𝑟𝑒𝑙), between bilayer and ink layer one, can take a value

between -1 and +1. A negative value indicated that they were inversely

proportional, a positive value that they were proportional, and zero that they were

uncorrelated and independent. Values between -1 and +1 in steps of 0.5 were

evaluated.

Coefficient of variations of density mottle, 𝐶𝑉𝐷𝑟𝑒𝑙, and overprint mottle, 𝐶𝑉𝑜𝑣𝑒𝑟 :

𝐶𝑉𝐷𝑟𝑒𝑙 = 100 × (𝜎𝐷

𝑟𝑒𝑙 �̅�𝑝𝑟𝑖𝑛𝑡𝑟𝑒𝑙⁄ ) and 𝐶𝑉𝑜𝑣𝑒𝑟 = 100 × (𝜎𝑜𝑣𝑒𝑟 �̅�𝑜𝑣𝑒𝑟⁄ ) [Eq. 29]

The ink trapping ratio (Eq. 25) was the acceptance of magenta ink on another ink

layer (overprint density) relative to the acceptance on the substrate (monolayer

density). Similarly, ink trapping mottle (𝜎𝑇𝑅) was calculated as variations in

overprint density (that is, overprint mottle) relative to mean optical density of a

magenta monolayer (�̅�2𝑟𝑒𝑙):

𝜎𝑇𝑅 =100

�̅�2𝑟𝑒𝑙 𝜎𝑜𝑣𝑒𝑟 [Eq. 30]

3.3.3 Validation of scanner-based estimates of optical densities

Optical densities of the prints from the scanner (described in Section 3.3.1) and

the spectrophotometer (SpectroDens from TECHKON GmbH, Königstein,

Germany) were compared. Both readings were taken from the same areas. Details

were given in Paper V.

Cupper contents of a few full-scale printed cyan areas were determined by

Inductively Coupled Plasma-Optical Emission Spectroscopy (Optima 8300 ICP-

OES, Perkin Elmer Inc., Waltham, USA). It was confirmed that the ink contained

cupper, but the exact relation with ink volume was not determined. Instead, the

content was compared between areas printed with different anilox volumes and

different printing speeds. The outcome was compared to the corresponding

optical densities. Details were given in Paper V.

48

3.3.4 Light microscopy

The structure of printed areas was examined by light microscopy (Zeiss Axioplan

2 from Carl Zeiss Microscopy GmbH, Jena, Germany) and stereomicroscopy

(SMZ-U zoom 1:10 from Nikon Crop.). In paper V, RGB images of bilayers and

monolayers were separated into their individual image channels, which indicated

the distribution between ink layers. Sequential images of the same area were

captured with transmitted and then reflected light. Beforehand, the board was

soaked with water and the unbleached layer was removed manually.

3.4 Absorption non-uniformity by staining technique

3.4.1 Staining technique

Absorption non-uniformities were characterised by a staining technique using a

coloured aqueous liquid. Samples absorbed the liquid during a brief and well-

defined period, Figure 14. This generated a stain, where darker areas corresponded

to higher liquid uptake, and the uniformity of the stain was related to the

absorption uniformity. The technique was introduced in Papers I and II and later

utilised in the studies related to flexographic print mottle (Paper III and VI) and

offset water/moisture interference mottle (Paper IV).

Figure 14. Illustration of the measurement set-up in F1 Printability Tester for the testing

of absorption non-uniformity. Dimensions of applicator: 30 mm x 15 mm x 20 mm (WxHxL)

with an opening of 26 mm x 16 mm.

A laboratory print tester, F1 Printability tester (IGT Testing Systems,

Amsterdam, Netherlands), was equipped with a specially made liquid applicator

that was open at the top and bottom. The polymer printing plate was replaced by

a blotting paper (BF3 from Munktell Filter ab, Falun, Sweden) backed with a soft

cushioning material (SF950 from Rogers Corporation, Evergem, Belgium) and

the nip force was set to 350 N.

When a sample passed underneath the applicator, a thin liquid film was dispensed

from the applicator and covered the surface of the sample until the blotting nip

Board sample

Liquid applicator

Stain

Cylinder with

Blotting paper

49

was reached. Thus, the time for absorption was defined by both the speed and

distance between the liquid applicator and blotting nip. Samples were tested at

0.5 ms-1 (0.2 ms-1 in Paper IV) running speed, which gave 0.15 s (0.37 s in

Paper IV) from first contact with the liquid until blotting was started and 0.21 s

(0.52 s in Paper IV) until reaching exit of the blotting nip.

3.4.2 Testing liquid

Methylene blue dye was used as the colorant and this was mixed with deionised

water (treated by reverse osmosis and deionisation at RISE, Stockholm, Sweden)

and a surfactant. The dye concentration and additives changed during the project

and arrived at: 0.075 wt.-% of methylene blue (C.I. 52015, Merck KGaA,

Darmstadt, Germany) and 0.007 wt.-% of Surfynol 2502 (acquired from Air

Products Chemicals Europe BV, now marketed by Evonik Industries AG, Essen,

Germany) mixed with deionised water. The dye concentration was increased

from 0.025 wt.-% to 0.075 wt.-% (Papers IV and VI), which generated darker

stains with higher contrast.

3.4.3 Analysis of stains

Stains were analysed in a similar way to mottle of printed samples: scanned

together with a calibration frame and analysed in STFI Mottling Expert.

Reflectance variations were divided into wavelength bands and absorption non-

uniformity was given as coefficient of variation, and in Paper III as standard

deviation.

Stains were semi-transparent, and reflectance of a stain depended on light

transmission through the stain and the reflectance of the board substrate. This

meant that a substrate with non-uniform reflectance, for example, white-top

mottle, affected the stain reflectance, especially at the lower dye concentration

(Papers I to III) and when analysis was made with grey-scale images (Paper I).

Contribution of non-uniform absorption (that is, non-uniform transparency) was

separated from that of white-top mottle in two different ways (0.025 wt-% dye):

(I) Separate measurements were made of both the board and stain in grey-scale.

Influence of absorbency was estimated based on error propagation (described in

Paper I). (II) Stains were scanned in RGB. The blue stain was hardly visible in the

B-channel and clearly visible in its complementary R-channel, Figure 15. Thus,

white-top mottle was indicated by the B-channel. From the two channels, an

absorption image was reconstructed (described in Paper II).

50

Figure 15. Separation of a stained surface into the individual channels of an RGB image.

White-top mottle dominated the B-channel and absorption dominated the R-channel

[reproduced with permission, Thorman et al. (2013)].

When using the higher dye concentration (0.075 wt-% dye), the light absorption

in the R-channel was deemed strong enough to suppress the influence of white-

top mottle and this channel was used without further modifications.

3.4.4 Validation of the absorption non-uniformity technique

3.4.4.1 Stain reflectance versus liquid uptake

To ensure that reflectance variations in the stain originated from uneven

absorption, it was:

I) verified that a greater liquid uptake created darker stains (see

description below);

II) ruled out that variations in the stain came from the roughness of the

substrate (Paper II);

III) confirmed qualitatively that there was excess liquid on the samples

and, thus, liquid access was not a limiting factor (Paper II);

IV) verified that the fading of stains was counteracted when storing them

in the dark (Paper II).

The stain reflectance does not necessarily quantify the absolute liquid uptake in

the test. Instead, variations in reflectance were used to determine how unevenly

liquid was absorbed within a specific sample. This was verified by applying

controlled liquid volumes to three of the pilot-coated papers from Paper IV.

These results have not been published. The coloured liquid was applied with

Automatic Scanning Absorptometer (KRK Kummagai Riki Kogyo Co., Ltd.,

Tokyo, Japan), which dynamically determined the transferred volume at different

contact times. Three tests were made per sample, and each left a stain that

contained a range of defined liquid volumes. The stains were scanned and

analysed in the same way as the absorption non-uniformity stains and reflectance

Red channel Green channel Blue channelRGB image

51

values were compared with the transferred volumes (for contact times of 131 ms

to 1751 ms).

The mean reflectance of the stains can differ because of the penetration and

adsorption behaviour of the cationic dye. The dye showed chromatographic

separation on the substrates studied, which was probably linked to adsorption

(Gane et al., 1999) rather than size exclusion. Hence, the coating colour

composition may alter the absorption and adsorption behaviour. This supported

that the reflectance variations should be compared and not the absolute levels.

3.4.4.2 Reproducibility

The reproducibility of stains was analysed by comparing stains made on six

different coated boards at three different laboratories (Paper II). The

reproducibility standard deviation, 𝑠𝑅, was calculated (unpublished) for each

board according to (Huang et al., 2008):

𝑠𝑅 = 𝑚𝑎𝑥{𝑠𝑟 , √𝑠�̅�2 + 𝑠𝑟

2(1 − 1 𝑛⁄ )} [Eq. 31]

𝑠𝑟 = √∑ (𝑠𝑖2 𝐼⁄ )𝑖 [Eq. 32]

Where 𝑠�̅� is the standard deviation between the laboratories, n, the number of

sub-areas per laboratory, 𝑠𝑟 , the repeatability standard deviation, I, the number

of laboratories and 𝑠𝑖 , the standard deviation of measurements at the i-th

laboratory. The reproducibility standard deviation was given relative to the mean

absorption non-uniformity of each sample, that is, as the coefficient of variation,

CV.

3.5 Mean absorbency and surface energy

In Paper VI, the mean water absorption was determined with an automatic Cobb test

where the liquid uptake was monitored dynamically by a level sensor (ACT 2500,

FIBRO system AB, now marketed by Testing Machines Inc., New Castle, USA).

No operator dependent steps were involved in the automatic test, which is

expected to significantly reduce sources of errors. Therefore, fewer repeated tests

than in manual testing were deemed necessary.

Wettability was indicated by contact angle measurements using DAT (FIBRO system

AB, now marketed by Testing Machines Inc., New Castle, USA) and deionised

water (treated by reverse osmosis and deionisation at RISE, Stockholm, Sweden).

To mimic the surface tension of a flexographic ink, 8.6 wt-% n-propanol was

added. This lowered the surface tension of water from approximately 73 mNm-1

to 38 mNm-1 (Papers I and VI). In Paper III, the surface energy, with its polar and

52

apolar components, was determined from the contact angles of deionised water

and Diiodomethane (for synthesis, Merck KGaA, Schuchardt, Germany),

according to Owens and Wendt (1969).

3.6 Pore-structure of coating layers

The pilot-coated boards were characterised by mercury intrusion porosimetry (Papers I

and VI), in a Micrometric AutoPore IV (Micromeritics Instrument Corp.,

Norcross, USA). The pore volume was reported as cm³ voids per m² board

surface, in the pore radius interval of 25 nm to 100 nm, and pore size (R50) where

50 % of the void volume consisted of finer pores was reported. Pore-structures

of the double-coated layer and the top coat were separately obtained by

subtracting the void volume of the uncoated (or pre-coated) board from that of

the fully-coated board. The void volume of the bulk was large relative to the

coating layers and, if bulk differed between areas where the uncoated and fully-

coated board were measured, the outcome was affected.

Local measurements of the refractive indices (RI) characterised the local surface

pore-structures of two selected barrier patterns (Paper III), using Surfoptic Imaging

Reflectometry System (Data Systems Ltd, Bristol, United Kingdom). The

technique detected RI at less than 1 µm depth (Hiorns et al., 2005) and was

reported to work well for substrates with a real RI of approximately 1.5, which

coating layers normally have (Elton and Preston, 2006b). The effective RI

decreased with air content in the surface, that is, pore diameter and number of

pores (Elton and Preston, 2006a). When the ink vehicle closed the surface pores,

the effective RI was expected to increase.

Scanning electron microscopy (SEM) technique was used to qualitatively assess

surface structures of barrier patterns and pilot-coated papers. A limited selection of

the barrier patterns (Paper III) was analysed after the sputtering of a conductive

monolayer of gold/palladium (Au/Pd) alloy. The secondary electron emission

was detected in a Jeol 6700 field emission SEM (JEOL Ltd., Tokyo, Japan), using

a primary electron beam of 5.0 kV.

In Paper IV, the pilot-coated papers were analysed without prior sputtering, under

a low vacuum (40 Pa), and the back-scatter emission of electrons were detected

using SU3500 (Hitachi High-Technologies Europe GmbH, Krefeld, Germany).

Electron acceleration voltage was 5.0 kV. The back-scatter may come from

emissions deeper in the structure than the secondary emission, but has the

advantage of being affected by the atomic mass of the nuclei that give rise to

scatter.

53

3.7 Coating layer uniformity by burn-out treatment

In Papers I and IV, the boards and papers were burn-out treated and then analysed

as print mottle (Section 3.3.1). Burn-out treatment was made according to Dobson

(1975), with some exceptions; samples were emerged into the testing liquid until

saturated and then air-dried before being heat treated at 225 ⁰C. The non-

uniformities were defined as standard deviation of reflectance. An uneven

coating layer will lead to reflectance variations in burn-out treated substrates

(Kolseth, 2004), as will an uneven pore-structure (Ragnarsson et al., 2013).

3.8 Topography: surface roughness

Topographies of the substrates were captured with either MicroProf® (MPR

1106 with sensor CWL, Fries Research & Technology GmbH, Bergisch

Gladbach, Germany) or Innventia OptiTopo Expert (RISE Bioeconomy, now

marketed by ABB, Zürich, Schweiz as L&W OptiTopo together with OptiTpoo

Expert from RISE Bioeconomy, Stockholm, Sweden). The MicroProf used

chromatic aberration and the OptiTopo used a photometric stereo technique to

analyse the surfaces. The surface roughness was calculated as a standard deviation

of height from the mean plane and divided into lateral wavelength bands (ranges)

using one-dimensional FFT. In Paper VI, surface craters were identified by a

depth threshold.

To acquire accurate detection of the transparent barrier patters in Paper III, an

intermediate replica step was necessary. Topography was measured from a cast

of the surface, made with a silicone rubber compound (RepliSet-GF1 from

Struers A/S, Ballerup, Denmark). The height needed to be inverted to represent

the topography of the board surface after measuring the replica.

Details are found the individual papers (I, III, IV and VI).

54

4 Results and discussion

4.1 Print uniformity of monochrome flexographic prints

The uniformity of monochrome ink layers (monolayers) was first studied by

laboratory printing on coated boards (Section 4.1.1) and pilot-coated papers

(Section 4.1.2). These results did not conclusively separate the influence of

absorbency from surface roughness and, therefore, coated boards were modified

so that each possessed a range of absorption non-uniformity levels. This enabled

the study of individual influence of absorbency and surface roughness

(Section 4.1.3). To resolve the influence of absorbency on monochrome ink layers,

a set of boards with varying coating colour compositions of the top-coating layer,

resulting in varying structure and surface chemistry, was printed in a full-scale

press (Section 4.1.4).

4.1.1 Laboratory-printing of coated boards (Paper III)

Seven coated board samples were studied, which differed in surface roughness

and porous structure. These substrates were later used for absorption

modifications. Print mottle of the untreated boards was higher for the rougher

substrates and a linear dependence was observed within the studied experimental

property space, Figure 16a. The roughness range was extended by the commercial

board samples and, without those, the correlation was not high. Instead, print

mottle ranked the pilot-coated boards in a similar way to absorption non-

uniformity, Figure 16b. This illustrates that surface roughness was the major factor

that determined print quality. However, when boards of similar roughness were

compared, the influence of non-uniform absorption was evident. Thickness and

compressibility may also be important factors.

Figure 16. Print mottle (0.5 mm to 8 mm) plotted versus a) surface roughness (0.25 mm

to 1 mm) and b) absorption non-uniformity (0.5 mm to 8 mm). Error bars indicate a 95 %

confidence interval [reproduced with permission, Thorman et al. (2018b)].

0.0

0.5

1.0

Pri

nt

mo

ttle

,S

.Dev

[%]

0.0 0.5 1.0 1.5

Height, S.Dev [µm]

0.0

0.5

1.0

Pri

nt

mo

ttle

,S

.Dev

[%]

0.0 0.5 1.0 1.5

Absorption non-uniformity, S.Dev [%]

b)a)

Pilot-coated

Commercial

Linear fit (pilot-coated): R² = 0.98Linear fit (all samples): R² = 0.69

55

4.1.2 Laboratory-printing of pilot-coated papers (unpublished)

The pilot-coated papers, with varying dosage and type of dispersant, were of

interest for flexographic studies, since they had already shown water/moisture

interference issues in offset printing. However, in laboratory tests, the effect of

uneven water absorption on flexographic print quality was modest. There was no

significant difference between the samples, with respect to solid tone mottle in

the spatial wavelength range of 1 mm to 8 mm, see Table 2. The laboratory

printing caused an intra-sample variation which was too high, relative to the

limited inter-sample variations. Fine-scale print mottle, of 0.25 mm to 1 mm,

differed between the samples on a 5 % significance level.

Table 2. Analysis of variance between sample means by one-way

ANOVA. [Unpublished]

Property (df1, df2) F-ratio p-value

Mottle C100 (1-8 mm) F(6,39) 1.3 0.30

Mottle C100 (0.25-1 mm) F(6,39) 3.3 0.01

UCA fraction F(6,39) 1.6 0.19

The samples with more uneven absorption also had higher fine-scale print

mottle, see Figure 17a. However, the linear fit (R²=0.69) was largely dependent

on the most uneven sample. Fine-scale mottle was showing a similar connection

to surface roughness, with R²=0.65, when given as height variations in the

wavelength range of 0.25 mm to 1 mm (Figure 17b). From this test, it was not

possible to conclude that either non-uniform absorption or surface roughness

was controlling fine-scale print mottle.

Figure 17. Fine-scale (0.25 mm to 1 mm) print mottle versus a) absorption non-uniformity

(0.25 mm to 1 mm) and b) surface roughness as height variations (0.25 mm to 2 mm).

Error bars indicate a 95 % confidence interval. [Unpublished]

a)NaPA

Ref.

NaCaPA

Absorption non-uniformity, CV [%]

Fle

xo

gra

phic

Prin

tM

ott

le,C

V[%

] b)NaPA

Ref.

NaCaPA

Height, S.Dev [µm]

Fle

xo

gra

ph

ic P

rin

t M

ottle

, C

V [%

]

56

4.1.3 Laboratory-printing of absorption-modified surfaces (Paper III)

In order to systematically study the impact of non-uniform absorption on print

mottle, the absorbency of coated boards of different surface roughness was

modified to form a “property matrix”. Various levels of absorption non-

uniformity were generated by means of printing barrier patterns of different area

coverage onto the substrates. Several methods were used to ensure that the

barrier patterns solely modified local pore structure and/or surface chemistry but

not the surface roughness of the coating layers. In the following sections, the

influence of the surface modifications on print mottle (4.1.3.1), surface roughness

(4.1.3.2), pore structure (4.1.3.3), and absorbency and wettability (4.1.3.4) are

presented.

4.1.3.1 Printability of absorption-modified surfaces

Absorption characteristics of the boards became non-uniform when their pore-

structures were modified with the addition of barrier patterns. This had a negative

influence on the print, on both rough and smooth board surfaces, Figure 18ab.

The barrier patterns made the print brighter in positions where the pores were

covered by the barrier coating, Figure 19. A correlation between print mottle and

absorption non-uniformity was observed. The correlation was linear in several

cases, but with different slope. A possible consequence of this may be that it is

more important to keep absorption non-uniformity on a low level in certain

samples. For products such as CB, on the other hand, a moderate increase in

absorption non-uniformity caused only minor effects on print mottle and it may,

therefore, be less critical to maintain a highly uniform absorption on this product.

Figure 18. Print mottle versus absorption non-uniformity of the paper boards with barrier

tones ranging from 0 % to 14 %. Error bars indicate a 95 % confidence interval [reproduced

with permission, Thorman et al. (2018b)].

a) b)

57

Figure 19. Microscopy images illustrating how the barrier pattern on sample N75 GCC was

affecting a) absorption non-uniformity and b) printed solid tone. There are six barrier dots

in each of the images. The brightness and contrast have been increased by 20 % for

visibility.

Based on linear regression statistics from Figure 18, print mottle values were

extrapolated/interpolated to different absorption non-uniformity levels. The

influence of surface roughness on print mottle was higher when comparing the

samples at the same absorption non-uniformity level (0.5 %). R2 increased from

0.69 to 0.84 when minimising the influence of different absorption non-

uniformities. This indicates that print mottle of untreated boards was largely

affected by surface roughness and, at the same time, by absorption non-

uniformity, although to a lesser extent.

4.1.3.2 Surface roughness of absorption-modified surfaces

A surface roughness pattern that followed the barrier pattern was only seen on

sample CC. Because of this, the sample was excluded from the printability

analysis. On this sample, 93 % of the barrier dots were raised above the rest of

the surface. Barrier dots were about 0.7 µm above the mean plane (zero level)

and board surface in-between the dots was often as high as 0.6 µm. The height

of the barrier dots on two other samples (P3 and CB) also needed to be examined

in detail. Here it was indicated that only 9 % and 10 % of the dots were raised

above their untreated surface, while 83 % and 72 % of surface peaks were not

connected to the barrier dots.

The surface roughness was evaluated in spatial wavelengths from 0.04 mm to

6 mm. When looking at the surface roughness in much greater detail, down to

individual mineral pigments or pores, all surface structures were affected by the

barrier material, Figure 20.

1 mm

b) Solid cyan print on top of 10% barrier tonea) Absorption on top of 10% barrier tone

58

4.1.3.3 Pore structure of absorption-modified surfaces

From SEM images, it was evident that pores covered by the barrier dots became

partly blocked, Figure 20. The pore-volume of the coating layer reduced in

positions where the barrier dots were located. This was seen in surface porosity

measurements based on RI, Figure 21. The effect was greater on a coating layer

that was more open to start with (P4, compared to CC), where the contrast

became higher.

4.1.3.4 Absorbency and wettability of absorption-modified surfaces

With one exception, the solid tones of the barrier material were more easily

wetted than the untreated board surfaces, Figure 22, due to a higher surface energy

(shown in Paper III). The staining technique revealed that the barrier patterns

indeed made the absorption uneven, Figure 23. The impact of the barriers’

nominal tone-values differed between the separate coating layers, because pores

were blocked more effectively on some sample surfaces and this generated a

greater contrast between highly absorbent surfaces and low-absorbent barrier

dots. Additionally, tone-value increase and transfer of barrier material may differ

between the substrates. Therefore, the absorption non-uniformity needed to be

characterised.

Absorption images (in Paper III) showed that the barrier patterns on samples P3

and CC became darker than the board surface surrounding the stains. This may

be a consequence of the low wettability (Figure 22) and compact pore structures

(Table 3 and Paper III) of the untreated surfaces, which limited the absorption into

these samples. The methylene blue dye stained the barrier dots slightly on all

samples, probably through adsorption of dye. This was, in most cases, difficult

to see because of the high contrast between the pale stain of dots and relatively

dark stain of untreated surfaces. Two samples (P3 and CC) were omitted from

the printability analysis because of the problems with the absorption testing and

the topographical structure from the barrier dots on sample CC.

59

Figure 20. SEM images of a barrier dot on the open structure of the pilot-coated sample P4.

Figure 21. Surface porosity maps of a 10 %

barrier tone on the open coating of P4 and

closed coating of CC; darker pixels

correspond to lower RI, that is pores are

darker.

Figure 22. Contact angle of water droplets on

the coating layers and their counter parts

covered by a solid barrier tone.

Figure 23. Absorption non-uniformity of samples with applied barrier patterns. Plotted as a

function of the nominal tone value. Error bars indicate a 95 % confidence interval.

[Figures 20 to 23: Reproduced with permission, Thorman et al. (2018b)]

10 μm

b) Edge inside 10% barrier dota) 10% barrier tone

1 μm100 μm

65

75

85

95

Co

nta

ct

an

gle

,w

ate

r0.1

s[°

]

b)

P1 P2 P3 P4 CA CB CC30

40

50

To

tal

Su

rface e

nerg

y [

mJ/m

²]

0

10

Po

lar

a)

P1 P2 P3 P4 CA CB CC

0.3

1.3

2.3

Ab

so

rpti

on

no

n-u

nif

orm

ity

S.D

ev [

%]

0 % 5 % 10 % 15 %Nominal tone value, barrier

CB

P2

P1

P3

CA

CC

P4

60

4.1.4 Full-scale printing of pilot-coated boards (Paper VI)

The cyan monolayers were more uneven when printed on the substrates whose

coating layers had small pores, a high contact angle and low liquid uptake in

Cobb60 testing, Figure 24. Similar trends were observed from prints at both

printing speeds, 7 ms-1 and 10 ms-1, respectively.

The cyan ink layer with high ink volume was most uneven on samples P06, which

had high latex content in the coating formulation and can be seen as an outlier in

Figure 24ab (pore sizes from analyses of the whole boards are plotted). When

using the standard ink volume, this sample no longer had the highest print mottle.

Possible explanations were the small pores in the top-coating layer and/or the

rather low pore volume of this sample (Table 3). If 50 % of the ink volume

transferred in each of the nips (anilox-plate and plate-substrate) and the anilox

volumes were 4.0 cm³·m-2 and 6.2 cm³·m-2, respectively, ink volumes on the

substrates would approximately be 1.0 cm³·m-2 and 1.6 cm³·m-2. This suggests

that the pore volume of the top-coating layer (1.6 cm³·m-2) may have been a

limiting factor for ink transfer when the ink volume was high.

Table 3. Characteristics of the fully-coated board samples (Full) and estimates of the double-

coating layer (Coat) and the top-coating layer (Top). Wavelength range for absorption non-

uniformity (Abs. Var.) was 0.25 mm to 0.5 mm and for height variations (Height Var.) was

0.06 mm to 0.25 mm. ± Indicates standard deviation of eight to twelve repeated tests, except

for Cobb60, where the min/max of two repeated tests are given. [Paper VI].

Pore Radius R50 [µm]

Pore Volume

[cm³·m-2] Contact Angle 0.5 s [⁰]

Cobb60

[ml·m-2] Abs.Var. [%]

Height Var. [µm]

Crater [%]

Full Coat Top Full Coat Top Water Alco Full Full Full Full

1. la

tex

A

P00 0.06 0.06 0.06 5.8 2.9 2.2 64 ±1 34 ±1 58 ±2 1.4 ±0.1 0.95 ±0.07 1.3 ±0.1

P01 0.06 0.06 0.05 6.4 3.4 2.7 69 ±1 37 ±1 55 ±0 1.1 ±0.0 1.04 ±0.09 1.3 ±0.2

P02 0.06 0.06 0.06 5.9 3.0 2.3 76 ±1 38 ±1 55 ±2 1.7 ±0.1 1.02 ±0.07 0.8 ±0.1

P03 0.06 0.06 0.06 5.2 2.2 1.5 79 ±2 40 ±1 51 ±1 1.5 ±0.1 0.88 ±0.03 0.6 ±0.1

2. la

tex

B P04 0.06 0.05 0.05 6.2 3.3 2.6 78 ±0 51 ±1 50 ±1 1.0 ±0.0 1.18 ±0.11 1.1 ±0.1

P05 0.06 0.05 0.05 5.9 3.0 2.3 84 ±2 54 ±1 43 ±0 1.2 ±0.1 1.23 ±0.11 1.1 ±0.2

P06 0.06 0.05 0.04 5.2 2.3 1.6 88 ±2 57 ±1 26 ±2 1.2 ±0.1 1.08 ±0.09 0.5 ±0.1

3. C

lay

P07 0.07 0.07 0.07 6.3 3.4 2.7 63 ±1 35 ±1 62 ±1 2.0 ±0.2 0.90 ±0.04 0.9 ±0.2

P08 0.07 0.07 0.06 7.3 4.4 3.7 67 ±1 36 ±0 59 ±1 2.1 ±0.1 0.88 ±0.03 0.9 ±0.1

P09 0.07 0.07 0.07 5.9 2.9 2.3 71 ±1 37 ±0 56 ±0 1.5 ±0.2 0.91 ±0.06 1.0 ±0.2

4. G

CC

s

P10 0.06 0.06 0.06 6.6 3.6 2.9 74 ±1 37 ±1 56 ±1 1.6 ±0.1 0.89 ±0.07 1.0 ±0.2

P01 0.06 0.06 0.05 6.4 3.4 2.7 69 ±1 37 ±1 55 ±0 1.1 ±0.0 1.04 ±0.09 1.3 ±0.2

P11 0.05 0.05 0.05 6.1 3.1 2.4 66 ±1 33 ±0 54 ±2 1.2 ±0.1 1.07 ±0.05 1.8 ±0.2

P12 0.07 0.07 0.07 6.5 3.5 2.8 60 ±1 33 ±1 73 ±2 1.4 ±0.1 1.18 ±0.06 1.4 ±0.1

P13 0.08 0.08 0.08 6.1 3.2 2.5 62 ±1 33 ±0 71 ±3 1.2 ±0.1 1.08 ±0.10 1.1 ±0.1

61

Figure 24. Density mottle of cyan monolayers versus board characteristics: pore radius (a,

b), contact angle of water (c, d) and absorbed volume in cobb testing (e, f). Print mottle

results from the 7 ms-1 printing speed (a, c, e) are plotted in separate graphs from 10 ms-1

(b, d, f). Error bars indicate standard deviation. [Paper VI]

R² = 0.68

R² = 0.65

0.00

0.04

0.08

0.12

0.03 0.05 0.07 0.09

De

nsity M

ottle

D1

S.D

ev

Pore size, R50Pore radius, R50 [µm]

Density M

ott

le, 𝜎𝐷10 2 −1

𝑟𝑒𝑙

C7 ms-1: C(Vol.)a)

R² = 0.73

R² = 0.55

0.00

0.04

0.08

0.12

0.03 0.05 0.07 0.09

De

nsity M

ottle

D1

S.D

ev

Pore size, R50

10 ms-1: C C(Vol.)

Pore radius, R50 [µm]

Density M

ott

le, 𝜎𝐷10 2 −1

𝑟𝑒𝑙

b)

R² = 0.55

R² = 0.34

R² = 0.59

0.00

0.04

0.08

0.12

50 60 70 80 90 100

De

nsity M

ottle

D1

S.D

ev

Contact angle H2O 0.5 sContact angle, water 0.5 s [⁰]

Density M

ott

le, 𝜎𝐷10 2 −1

𝑟𝑒𝑙

C7 ms-1: C(Vol.)c)

R² = 0.30

R² = 0.54

0.00

0.04

0.08

0.12

50 60 70 80 90 100

De

nsity M

ottle

D1

S.D

ev

Contact angle H2O 0.5 s

10 ms-1: C C(Vol.)

Contact angle, water 0.5 s [⁰]

Density M

ott

le, 𝜎𝐷10 2 −1

𝑟𝑒𝑙

d)

R² = 0.54

R² = 0.33

R² = 0.79

0.00

0.04

0.08

0.12

0.0 20.0 40.0 60.0 80.0

De

nsity M

ottle

D1

S.D

ev

ACT 60 sVolume, Cobb 60 [ml·m-2]

Density M

ott

le, 𝜎𝐷10 2 −1

𝑟𝑒𝑙

C7 ms-1: C(Vol.)e)

R² = 0.31

R² = 0.74

0.00

0.04

0.08

0.12

0.0 20.0 40.0 60.0 80.0

De

nsity M

ottle

D1

S.D

ev

ACT 60 s

10 ms-1: C C(Vol.)

Volume, Cobb 60 [ml·m-2]

Density M

ott

le, 𝜎𝐷10 2 −1

𝑟𝑒𝑙

f)

R² = 0.74

62

Coating layers with fine pores can be assumed to produce higher capillary

pressure (Eq. 4) that drives the liquid flow. On the contrary, the liquid uptake

(Cobb60) was observed to be low on these substrates. There were two possible

reasons. First, the substrates with finer pores tended to have a lower pore volume

per square meter of board, which affects the liquid uptake and ink penetration.

Second, the viscous drag, a resistance to liquid flow, is inversely proportional to

the pore radius, raised to the power of four (Eq. 6). This means that the flow

resistance (viscous drag per volumetric flow rate) was 16 times greater in a pore

with a radius of 0.04 µm compared to a radius of 0.08 µm, given the prerequisite

that penetration depth was the same in both pores. If instead compared at equal

volumes, the difference in flow resistance was even greater. A consequence of

this was that the substrates with fine pores hindered ink penetration driven by

nip pressure.

From this followed, that when reaching the exit of the printing nip, there was a

larger amount of free (mobile) ink available between the printing plate and a

substrate with finer, as compared to larger, pores, Figure 25. Ink that had

penetrated the pore-structure, or had been drained from its fluid phase, was

considered to be immobilised, meaning that it could not redistribute anymore.

Thereby, a larger volume of non-immobilised ink could be transferred via ink

splitting (Walker and Fetsko, 1955). Possibly, a filtercake was created at the

coating surface when the fluid phase of the ink penetrated into the coating layer

(Olsson et al., 2007a; Gane and Ridgway, 2009; Luu et al., 2010b). A greater

immobilisation and ink transfer may yield a more stable optical density (Tollenaar

and Ernst, 1961), and therefore lower density mottle. Additionally, the non-

immobilised ink may have generated an heterogenous ink splitting pattern and/or

may have become unevenly thick in the subsequent ink levelling process. Ink

splitting patterns, from viscous fingering, were anticipated to create coarser

structures in thicker ink films (Wang et al., 2014; Sauer et al., 2018). On a rough

substrate, ink levelling may result in an uneven ink thickness (Kheshgi, 1997).

Thus, at the point of ink splitting, a high pressurised ink penetration may have

immobilised a greater ink volume and left a minor volume mobile, which could

yield a more homogenous split and/or less risk that ink levelling created and

uneven ink thickness.

63

Figure 25. Illustration (not to scale) of possible ink distribution in a coating layer with a) large

pores and high pore volume and b) narrow pores and low pore volume. [Paper V]

The cyan ink layers became more uneven when using the anilox that carried more

ink and, at the same time, the mean optical density increased. For some samples,

density mottle increased substantially more than the mean optical density when

using the high ink volume. The high-volume ink layers on those samples had a

high fraction of UCA. These were the substrates with a high latex content in the

coating colour. UCA levels were not explained by the substrates being rougher

or having a larger fraction of craters than the other samples. Instead water made

a high contact angle on these substrates, possibly caused by local accumulation

of latex at the surface (Bohlin and Johansson, 2018) and thus local areas of poor

wettability (Kargupta et al., 2001) may have induced UCAs.

Print mottle and optical density in the monolayer remained on the same level

when increasing the printing speed from 7 ms-1 to 10 ms-1 and, on some

substrates, those properties even improved, see Figure 26. This indicated that the

time interval for capillary driven absorption, ink levelling and dewetting was of

little importance for how uniform the ink layer became. Hence, the effects seen

in Figure 24, where finer pores, higher contact angle and lower liquid uptake

increased print mottle, happened very quickly, either in the nip (pressurised

penetration) or immediately after it.

Ink split in

free ink

a) Coating layer: large pores b) Coating layer: fine pores

Immobilised ink

Voids (pores)Whe

n re

achin

g

printing

nip

exit

Aft

er p

rinting

nip

64

Figure 26. Properties of the cyan monolayer when printing with speeds

7 ms-1 and 10 ms-1: a) mean optical density, b) density mottle (0.25 mm to

1 mm). Error bars indicate standard deviation between a minimum of five

printed signatures. [Paper VI]

4.2 Print uniformity of multicolour flexographic prints

This section deals with flexographic multicolour printing where ink transfers

onto ink layers already present on the substrate, that is, ink-on-ink printing. Focus

is on the uniformity of the overprint layer (top layer) in areas with two layers of

ink on top of each other (bilayers). First, the ink distribution of the bilayer

(Section 4.2.1), impact of cyan ink volume (Section 4.2.2) and mechanisms of ink

trapping (Section 4.2.3) are discussed. Then, a direct influence of the preceding ink

layer (Section 4.2.4) and an indirect influence of the substrates’ ink absorbency on

ink trapping (Section 4.2.5) are proposed.

4.2.1 Ink distribution within bilayers (Paper V)

4.2.1.1 Density mottle

Density mottle of the bilayers (with magenta printed over cyan) showed similar

tendencies to density mottle of the cyan monolayer, Figure 27. This indicated that

there was a correlation between the two. The bottom of the bilayer was assumed

to be identical to the monolayer of cyan.

1.0

1.1

1.2

1.3

P0

0P

01

P0

2P

03

P0

4P

05

P0

6

P0

7P

08

P0

9

P1

0P

01

P1

1P

12

P1

3

Me

an D

_cyan

0.04

0.06

0.08

P0

0P

01

P0

2P

03

P0

4P

05

P0

6

P0

7P

08

P0

9

P1

0P

01

P1

1P

12

P1

3

Mo

ttle

1𝐷1

𝜎1

a)

b)

0.01

0.02

0.03

0.04

P0

0P

01

P0

2P

03

P0

4P

05

P0

6

P0

7P

08

P0

9

P1

0P

01

P1

1P

12

P1

3

Sd

ev_

ove

r (0.25-1 mm)

C: 7 ms-1 C: 10 ms-1

Latex A Latex B Clay cont. GCC’s types

65

Figure 27. Density mottle of the bilayer (G-channel) plotted versus the cyan

monolayer (R-channel). Wavelength range: 0.25 mm to 1 mm. Error bars

indicate standard deviation between a minimum of five printed signatures.

[Paper V]

4.2.1.2 Light reflection and transmission microscopy

Light reflection (Figure 28) and transmission (Figure 29) microscopy images of

printed bilayers indicated how the layer thickness of both the bottom and

overprint were related to each other. This was seen after separating the RGB-

images into the red (R) and green (G) image channels. Cyan ink layers absorbed

light in the G and foremost in the R-channel, whereas magenta absorbed nearly

no light in the R, but very much in the G-channel. This is illustrated by the

monolayers seen on the right-hand side in, Figures 28 and 29.

Inspection of reflection and transmission images from the same area of a printed

bilayer indicated that the amount of magenta was high (corresponding to dark

areas in the G-channels) when the amount of cyan was low (bright in the R-

channels), which can be seen in the outlined areas in Figures 28 and 29. To ensure

that a bright area in the R-channel corresponded to the (relatively) thin cyan layer,

the images of both reflection and transmission needed to be bright. If (red) light

scattered inside the magenta layer, it was partly prevented from reaching the cyan

layer and could therefore not be absorbed, see Figure 30. Thus, a thick magenta

layer may increase the brightness of the R-channel reflection image, but the

opposite effect was then expected to occur in the transmission image.

R² = 0.40

R² = 0.83

0.04

0.08

0.12

0.04 0.08 0.12

Ove

rprint M

ottle

S.D

ev

Monochrome mottle D1

Bila

yer

mott

le,𝜎𝐷12 −𝑐ℎ

𝑎𝑛𝑛𝑒𝑙

𝑟𝑒𝑙

C7 ms-1: C(Vol.)

Monochrome Mottle, 𝜎𝐷1 𝑅−𝑐ℎ𝑎𝑛𝑛𝑒𝑙𝑟𝑒𝑙

66

Figure 28. Images from the reflective light microscopy of a bilayer (cyan + magenta)

and monolayers of magenta and cyan, respectively: a) RGB images and corresponding

b) red channel and c) green channel images of sample P10 (standard ink volume,

10 ms-1 printing speed). The scales mark 100 μm in each image. Dashed cyan outlines a

few areas that were bright in the green channel and dashed red outlines dark areas in the

green channel. [Paper V]

Bilayer (cyan + magenta)



Magenta

Cyan

Magenta

Cyan

Magenta

Cyan

100 µm 100 µm

a)

RG

Bb

) R

ed

ima

ge

ch

an

ne

lc

) G

ree

n im

ag

e c

ha

nn

el

67

Figure 29. Images from transmitting light microscopy of a bilayer (cyan + magenta)

and monolayers of magenta and cyan, respectively: a) RGB images and corresponding

b) red channel and c) green channel images of sample P10 (standard ink volume,

10 ms-1 printing speed). The scales mark 100 μm in each image. Same outlines as in

Figure 28. [Paper V]

a)

RG

Bb

) R

ed

ima

ge

ch

an

ne

lc

) G

ree

n im

ag

e c

ha

nn

el




Magenta

Cyan

Magenta

Cyan

Magenta

Cyan

100 µm 100 µm

68

Figure 30. Illustration of how red light may scatter, absorb, transmit and reflect

in the magenta and cyan ink of a bilayer in reflection and transmission light

mode. [Paper VI]

4.2.2 Impact of cyan ink volume (Paper V)

The volume of cyan ink was increased in one of the trial points. This confirmed

the inverse dependence between bottom and overprint layers and was seen by

spectrophotometric- and scanner-based measurements, the latter is shown in

Figure 31. Using an anilox that carried a larger cyan volume caused the overprint

density to decrease for each tested sample. On average, the cyan optical density

increased by 16 % ± 2 when shifting to the high volume anilox (Figure 31a), and

this resulted in an average reduction of the overprint densities by 7 % ± 1

(Figure 31b).

Figure 31. Mean optical density measured with the scanner of a) the cyan monolayer

(from the R-channel) and b) the overprint layer (from the G-channel). Densities when

using the anilox with standard and high ink volume are shown in separate series.

Error bars indicate standard deviation between a minimum of five printed signatures.

[Paper V]

b) Red light transmissiona) Red light reflection

Absorbed

Scattered

Transmitted

Incoming light

Incoming light

Absorbed

Reflected

Scattered

Board Cyan Magenta

0.6

0.7

0.8

P0

0

P0

1

P0

2

P0

3

P0

4

P0

5

P0

6

P0

7

P0

8

P0

9

P1

0

P0

1

P1

1

P1

2

P1

3

D_

ove

r𝐷𝑜𝑣𝑒𝑟

b)

1.0

1.1

1.2

1.3

1.4

P0

0

P0

1

P0

2

P0

3

P0

4

P0

5

P0

6

P0

7

P0

8

P0

9

P1

0

P0

1

P1

1

P1

2

P1

3

D_

1𝐷1

a)More ink 1

More ink 1

C(stand.) C(Vol.)


69

4.2.3 Mechanisms of ink trapping (Paper V)

4.2.3.1 Overprint ink transfer

The plot of density mottle in Figure 27 indicated that there was a correlation

between the uniformity of the cyan layer and uniformity of the bilayer.

Furthermore, the microscopy images (Figures 28 and 29) and the overprint

densities resulting from the two cyan volumes (Figure 31), showed that the

overprint and bottom layers were inversely proportional. This was the case for

the average optical densities, as well as on the local level.

On the local level, the connection between the bottom and overprint layers,

suggested that the two layers interacted and affected how the bilayer was split

and the overprint ink transferred. When using a high cyan volume, the thickness

of the bottom layer was assumed to increase. Evidently, this moved the split

position (of the bilayer) away from the printing plate so that less overprint ink

transferred, illustrated in Figure 32ab. Consequently, an uneven thickness of the

first ink may yield an uneven transfer of the subsequent ink, Figure 32c.

Possibly, the split position moved from the magenta into the cyan layer in

localised areas where the cyan ink was thick. Increased film thickness may lower

the resistance to split forces (Tollenaar and Ernst, 1965; MacPhee, 1979; Shen et

al., 2004). However, this was not supported by the microscopy images, which

indicated that the cyan was thickest in positions where magenta was low.

Figure 32. Illustrations (not to scale) of how the thickness of the first layer (cyan

ink) may affect the transfer of an overprinting ink in cases of; a) thinner cyan layer,

b) thicker cyan layer, c) uneven cyan layer and d) optical densities corresponding

to uneven layers. The mobile ink is assumed to split at the same fraction. [Paper V]

c) Uneven layer of cyan

Ink

splitMobile ink

Immobile

Mobile ink

a) Thinner layer of cyan

Immobile

Ink

splitMobile ink

Immobile

b) Thicker layer of cyan

Bilayer

Cyan

d) Optical density of uneven layer

Ink

split

70

In situations where the roughness of the substrate controls the first ink layer, it

may also control the overprint layer. For example, both the first ink and overprint

ink may be pushed away from surface peaks and accumulate in surface craters.

However, the monolayer printed on these substrates was, as already discussed,

affected by the pore structure (Figure 24) of the coating layer rather than by

surface roughness.

4.2.3.2 Correlation term in overprint mottle formula

Overprint mottle was estimated from Eq. 28, which includes a correlation term.

Thus, the correlation coefficient between the optical densities of cyan and the

bilayer needs to be approximated. The correlation coefficient was likely ≤ 0

(Figure 32d) because of the bottom and overprint layers being inversely

proportional and optical density in the bilayers being more influenced by magenta

than cyan ink, when measured in the G-channel. The overprint mottle values

gave very similar rankings of the samples when estimated with correlation

coefficient -1, -0.5 or 0, see Figure 33. Therefore, the choice was not considered

critical and -1 has been used in Sections 4.2.4 and 4.2.5.

Figure 33. Values of overprint mottle when estimated using different correlation

coefficients (ρ): a) -1 vs. 0; b) -0.5 vs. 0. Coefficient of determinations (R2) were between

0.99 and 1.00. Error bars indicate standard deviation between a minimum of five printed

signatures. [Paper V]

0.05

0.09

0.13

0.17

0.05 0.09 0.13 0.17

ρ(D

1, D

12

) -1

ρ(D1, D12) 0

C7 ms-1: C(Vol.)

𝜎𝑜𝑣𝑒𝑟,

ℎ𝑒𝑛

1

, 12

=

−1

𝜎𝑜𝑣𝑒𝑟 , ℎ𝑒𝑛 1 , 12

= 0

a)

0.05

0.09

0.13

0.05 0.09 0.13

ρ(D

1, D

12

) -0

.5

ρ(D1, D12) 0

C7 ms-1: C(Vol.)

𝜎 𝑜𝑣𝑒𝑟,

ℎ𝑒𝑛

1

, 12

=

−0

𝜎𝑜𝑣𝑒𝑟 , ℎ𝑒𝑛 1 , 12

= 0

b)

71

4.2.4 Direct influence of the preceding ink layer on trapping (Paper VI)

As expected from the described mechanisms, the unevenness of an overprint

layer (magenta) was proportional to that of the monolayer (cyan), Figure 34ab.

The overprint layers may also be affected by properties of the magenta ink, such

as ink pigmentation or anilox volume, which may affect the magenta layer

regardless of where it was printed. Trapping mottle estimated how unevenly the

ink was distributed in overprint layers, relative to the average acceptance of

magenta on the substrate (Eq. 30). Samples were ranked in a similar way for

trapping and overprint mottle, and both characteristics correlated with the non-

uniformity of the cyan ink layer, Figure 35. If comparing substrates that are

printed with, for example, different magenta inks or magenta aniloxes, it is

probably better to compare trapping mottle instead of overprint mottle. Under

those circumstances, ink trapping mottle may separate the influence of substrates

from the influence of different inks.

Figure 34. Overprint mottle at printing speed 7 ms-1, plotted as a function of density mottle

of the first ink layer: a) Wavelength range 0.25 mm to 1 mm and b) Wavelength range 1 mm

to 8 mm. Error bars indicate standard deviation between a minimum of five printed

signatures. [Paper V, VI]

R² = 0.56

R² = 0.93

0.04

0.08

0.12

0.16

0.00 0.04 0.08 0.12

Ove

rprint M

ottle

S.D

ev


Ove

rprint

Mott

le, 𝜎𝑜𝑣𝑒𝑟0 2 −1

Monochrome Mottle, 𝜎𝐷1 0 2 −1 𝑟𝑒𝑙

a) C7 ms-1: C(Vol.)

R² = 0.80

R² = 0.93

0.01

0.02

0.03

0.04

0.05

0.01 0.02 0.03 0.04 0.05

Ove

rprint M

ottle

S.D

ev

Monochrome mottle

Ove

rprint

Mott

le, 𝜎𝑜𝑣𝑒𝑟1−8

Monochrome Mottle, 𝜎𝐷1 1−8 𝑟𝑒𝑙

b) C7 ms-1: C(Vol.)

72

Figure 35. Trapping mottle at printing speed 7 ms-1, plotted as a function of density mottle

of the first ink layer: a) Wavelength range 0.25 mm to 1 mm and b) Wavelength range 1 mm

to 8 mm. Error bars indicate standard deviation between a minimum of five printed

signatures. [Paper V, VI]

4.2.5 Indirect influence of substrates’ ink absorbency on trapping (Paper VI)

The overprint layers on certain samples (P02-P03, P05-P08 and P11) were highly

sensitive to increases in speed, see Figure 36. The effect was evident in overprint

density (Figure 36a) and overprint mottle in the wavelength range of 0.25 mm to

1 mm (Figure 36b), but not as clear in the 1 mm to 8 mm range (Figure 36c). The

samples that stood out had either high latex content (type A or B), the finest

GCC pigment or high clay content. Substrates with high latex content had low

absorbency, high contact angles, compact coating layers and a tendency towards

fine pores, see Table 3. Likewise, the coating layers with the finest GCC pigment

had small pores and low liquid uptake. The substrates with high clay content

were, on the other hand, easily wetted, had large pores and high pore volume

(Table 3), which ought to favour a quick ink immobilisation. Fine-scale absorption

non-uniformity may instead be the cause for ink trapping issues here. Table 3

indicated that this substrate had the most uneven absorption when investigated

in the wavelength range of 0.25 mm to 0.5 mm.

A large fraction of uncovered areas emerged in the overprint layer at the higher

printing speed, especially on the samples whose absorption was slow or uneven.

Commonly, print mottle reflects heterogeneities in ink film thickness. To a

greater or lesser extent, areas with no film thickness at all (that is, UCAs) will

contribute to both the print mottle appearance and value. The results here

indicated that fine-scale overprint mottle was closely connected to the UCA

R² = 0.42

R² = 0.90

4.0

8.0

12.0

16.0

0.00 0.04 0.08 0.12

Tra

pp

ing

mo

ttle


Tra

ppin

g m

ott

le, 𝜎𝑜𝑣𝑒𝑟

0 2 −1

a) C7 ms-1: C(Vol.)

Monochrome Mottle, 𝜎𝐷1 0 2 −1 𝑟𝑒𝑙

R² = 0.78

R² = 0.98

0.0

2.0

4.0

6.0

0.01 0.02 0.03 0.04 0.05

Tra

pp

ing

mo

ttle

Monochrome mottle

Tra

ppin

g m

ott

le, 𝜎𝑜𝑣𝑒𝑟

1−8

b) C7 ms-1: C(Vol.)

Monochrome Mottle, 𝜎𝐷1 1−8 𝑟𝑒𝑙

73

fraction at the high printing speed, Figure 37. It is suggested that a slow or uneven

absorption in combination with a short time between applications of the first and

second ink layers gave insufficient time for ink immobilisation. This appeared to

be critical. Magenta ink in the overprint layer either dewetted or was prevented

from transferring onto the bottom layer in positions where this ink was thick.

Figure 36. Properties of the overprint layer when printing with speeds of

7 ms-1 and 10 ms-1: a) mean optical density, b) overprint mottle in wavelength

range 0.25 mm to 1 mm and c) overprint mottle in wavelength range 1 mm to

8 mm. The overprint was strongly affected by changed printing speed on

samples marked by dashed circles. Error bars indicate standard deviation

between a minimum of five printed signatures. [Paper VI]

Figure 37. Overprint mottle (fine-scale) plotted as a function of UCA. Error bars

indicate standard deviation between a minimum of five printed signatures.

[Paper VI]

0.5

0.6

0.7

0.8

P0

0P

01

P0

2P

03

P0

4P

05

P0

6

P0

7P

08

P0

9

P1

0P

01

P1

1P

12

P1

3

Me

an D

_o

ve

r

0.04

0.06

0.08

0.10

0.12

P0

0P

01

P0

2P

03

P0

4P

05

P0

6

P0

7P

08

P0

9

P1

0P

01

P1

1P

12

P1

3

Mo

ttle

ove

r

0.02

0.04

0.06

P0

0P

01

P0

2P

03

P0

4P

05

P0

6

P0

7P

08

P0

9

P1

0P

01

P1

1P

12

P1

3

Sd

ev_

ove

r𝜎𝑜𝑣𝑒𝑟

𝜎𝑜𝑣𝑒𝑟

b)

c)

(0.25-1 mm)

0.01

0.02

0.03

0.04

P0

0P

01

P0

2P

03

P0

4P

05

P0

6

P0

7P

08

P0

9

P1

0P

01

P1

1P

12

P1

3

Sd

ev_

ove

r (1-8 mm)

C: 7 ms-1 C: 10 ms-1𝐷𝑜𝑣𝑒𝑟

a)


R² = 0.87

R² = 0.91

0.04

0.08

0.12

0.16

0.0 3.0 6.0 9.0

Ove

rpri

nt

Mo

ttle

S.D

ev

UCA multilayer G-channel [%]

Ove

rprint

Mott

le, 𝜎𝑜𝑣𝑒𝑟0 2 −1

UCA in overprint, G-channel [%]

10 ms-1: C C(Vol.)a)

R² = 0.91

74

4.3 Validation of the absorption non-uniformity technique

The absorption non-uniformity technique was developed as part of the doctoral

project. The intention was to demonstrate absorption behaviours that were

related to water-based flexographic printing (Sections 4.1.3.1 and 4.2.5). The

technique was tested in several ways, both to verify its accuracy and to create an

understanding of its limitations. A part of this is presented in the following

sections, 4.3.1 to 4.3.3. Additionally, the applicability of the technique was

evaluated in a nearby field. Namely, the interaction between absorption of

aqueous fountain solution and water/moisture interference mottle in offset

printing (Section 4.3.4).

4.3.1 Stain reflectance, a function of liquid volume (unpublished)

Methylene blue dye was used in the staining technique. This dye was cationic and

separated chromatographically when the liquid was absorbed by the paper

substrate. The stain became darker when a larger liquid volume was absorbed,

but the stain reflectance was not calibrated to a liquid volume. Thus, the stain

technique gave a relative measurement rather than an absolute measurement. It

is likely that each individual substrate requires its own calibration. Therefore, the

test should be used for comparisons of variations in reflectance, and not to assess

the total volume of absorption. Material components and their distribution in a

coating layer may influence the dye adsorption, thus an uneven stain implies non-

uniformity either in material distribution or coating structure. For example, Gane

et al. (1999) pointed out that cationic dyes were easily attracted by anionically

dispersed calcium carbonate pigments.

Figure 38. Average reflectance of stains made with the Automatic Scanning

Absorptometer instrument at contact times from 131 ms to 1751 ms. The same

testing liquid as in the absorption non-uniformity technique. The error bars

represent a 95 % confidence interval. [Unpublished]

0.2 NaPA

0.8 NaPA

0.8 NaCaPA

75

Figure 38 depicts the reflectance levels of stains made by applying different

volumes of the coloured liquid (applied with an Automatic Scanning

Absorptometer) on three paper substrates. As shown, the stains became brighter

(that is, had a higher reflectance) when the liquid volume got smaller. This

confirms that lateral non-uniformities of the stain were a result of uneven

absorption into the pore structure.

4.3.2 Reproducibility (Paper II)

The coefficients of variation associated with reproducibility of the absorption

non-uniformity testing were between 6 % and 7 % in the 0.25 mm to 1 mm

wavelength range, and higher, up to 14 %, in the 1 mm to 8 mm range, Table 4.

Table 4. Reproducibility of the (absorption non-uniformity) staining technique between

three laboratories. Non-uniformities in shorter (0.25 mm to 1 mm) and longer (1 mm to

8 mm) wavelength ranges. Results for seven coated boards (A-F). [Unpublished]

1 mm to 8 mm 0.25 mm to 1 mm

Mean [%] sR [%] CV [%] Mean [%] sR [%] CV [%]

A 0.46 0.05 12 0.46 0.04 6

B 0.56 0.05 10 0.56 0.05 7

C 0.50 0.07 14 0.50 0.03 6

D 0.44 0.04 9 0.44 0.05 7

E 0.52 0.07 14 0.52 0.05 7

F 0.42 0.04 10 0.42 0.05 7

4.3.3 Additional verification (Paper II)

There was high correlation between white-top mottle measured before making a

stain and from the same area after applying a stain. The latter was evaluated from

the blue channel of RGB images of the stained areas, when methylene blue dye

was used as the colour agent. This procedure was successful when the stains were

bright (0.025 wt-%), see Figure 39, but not when using a higher dye concentration

in the testing liquid (0.075 wt.-%) (not depicted in the graph). The darker stains

were still visible in the blue image channel.

76

Figure 39. White-top mottle measured through the stain compared to when

measured without any stain. Error bars indicate a 95 % confidence interval

[reproduced with permission, Thorman et al. (2012)].

4.3.4 Relevance for water/moisture interference in offset printing (Paper IV)

In offset printing, oil-based inks may be transferred to the dry substrate, or to an

ink layer applied in one of the previous units or to a dampened surface. Paper IV

deals with the latter case, which was similar to an ink trapping situation but with

a layer of water or moisture instead of ink in the bottom layer.

In a separate study, water-interference mottle and ink lift-off issues were

identified on a set of pilot-coated papers where the dosage and type of dispersant

(NaPA and NACaPA) had been varied (Kamal Alm et al., 2015b).

Characterisation of the absorption non-uniformity of those papers was made in

Paper IV. This revealed that the substrates with uneven absorption also had

higher UCA fraction, Figure 40a, and print mottle, Figure 40b. Coating structure

and absorbency were disturbed by a moderate dose of excess dispersant in the

coating colour. In turn, ink that was applied downstream in the press presumably

met a paper surface with uneven moisture content. This is suggested to be the

cause of the moisture/water interference problems. Lingering fountain solution

(moisture), may form a thin aqueous layer on the surface, be capillary-absorbed

within the structure or be bonded within the hygroscopic polyacrylate (Kamal

Alm et al., 2010). All situations may affect the ink transfer and adhesion between

the coating and the ink. Shen et al. (2004), and earlier (MacPhee, 1979), suggest

that fountain solution prevents ink from being transferred to non-image areas on

the plate through a film splitting of the weakest layer, which is the fountain

solution. Possibly, similar effects occurred when liquid accumulated in the paper.

R² = 0.99

0.0

0.5

1.0

0.0 0.5 1.0

Th

rou

gh

Sta

in, C

V [

%]

Before Stain, CV [%]

White-top mottle, 1-8 mm

77

There was one substrate, 0.8 NaPA, which did not follow the general trend in

Figure 40. Absorption stain was highly uneven on this paper, but offset printing

exhibited only modest print mottle and low UCA fraction. A closer look revealed

that absorption tests on this paper had a particularly high measurement

uncertainty. Qualitative assessment of the coating layers by SEM, indicated that

there were small “closed” areas (less permeable) in the coating layers of all

samples (Figure 41a), except for 0.8 NaPA, where no such areas were observed

(Figure 41b). Kamal Alm et al. (2010) indicate that the addition of NaPA,

compared to NaCaPA, slightly increases the fraction of large pores and increases

the instant water uptake. Possibly, the pore structure was more open and rapidly

removed the fountain solution on sample 0.8 NaPA. This may have compensated

for the uneven absorption.

Figure 40. UCA (a) and print mottle (b) as functions of absorption non-uniformity. Error bars

indicate a 95 % confidence interval [reproduced with permission, Thorman et al. (2018a)].

Figure 41. SEM image showing: a) sample 0.8 NaCaPA, where two closed areas are

outlined, and b) sample 0.8 NaPA, where no closed areas were found [reproduced with

permission, Thorman et al. (2018a)].

b) 0.8 pph NaPAa) 0.8 pph NaCaPA

78

5 Conclusions These studies give new insight into the interaction among coated substrates and

flexographic water-based inks, specifically regarding the influence from liquid

absorbency. This is useful for producers of liquid packaging boards and

packaging converters since it offers a palette of new tools to use when

investigating liquid absorbency and its impact on print quality:

a) Aqueous staining technique to characterise absorption non-uniformity: This

technique was applied not only for the prediction of print mottle in

flexographic printing, but also for water/moisture interference mottle in

offset printing.

b) Experimental approach to create a board “property-matrix” by adding barrier

patterns and thereby modifying the substrate’s absorption: This way, the

influence on print mottle of non-uniform absorption can be separated

from that of other board properties, for example, surface roughness. It

was found that non-uniform absorption affected the flexographic print

mottle on both rougher and smoother boards.

c) Technique to reveal ink trapping non-uniformity: Combined scanner and

microscopy analysis gave new insights into ink trapping and ink

transferring mechanisms in multicolour printing. It was found that

overprint layers became thicker when the first ink layer was thinner, and

vice versa. Consequently, an uneven thickness distribution of the first ink

made the overprint layer uneven.

It is concluded that the liquid absorbency of the substrates interacted directly with the ink in

monochrome areas. Low liquid absorption was seen to increase the risk of uneven

print. Coating layers with a pore-structure consisting of small pores and low

liquid uptake generated higher print mottle because of high flow resistance in

pressure driven liquid absorption. The absorption-modified surfaces clearly

showed that locally compact pore-structures hindered ink penetration and caused

brighter spots in the print.

It is concluded that overprint layers in multicolour prints interacted directly with the preceding

ink layer and only indirectly with the absorbency of the substrate. The ink amount in the

bottom layer inversely controlled the transfer of overprint ink. In addition, the

substrates controlled the rate and uniformity of ink immobilisation taking place

between applications of inks one and two. An ink absorption which was either

too slow or uneven made ink trapping more sensitive to printing speed.

79

6 Future work The work presented here has left a few loose ends and led to new questions being

raised. Following areas would be valuable to investigate further:

1. The results indicated that pressurised ink penetration in the nip was of

high importance, for print quality in monochrome areas. Testing of

capillary driven absorption can give misleading results, since surface

chemistry has high influence on spontaneous but not on forced

absorption testing. Two of few commercially available methods is the

IGT Ink Penetration tester (IGT Testing Systems, Almere, Netherlands)

and pressurised Bristow Wheel (KRK Kummagai, Tokyo, Japan). These

techniques cannot separate pressurised ink penetration, from liquid filling

the surface roughness. Therefore, a new, or modified, absorption testing

technique for pressurised absorption is needed.

2. There is still much to learn regarding flexographic ink trapping. For

example, the influence of surface roughness and board compressibility

need to be studied. In the present work, it was possible to gain an

understanding for the influence of ink volume in the first ink layer. The

second ink was kept constant in the trial, so there are still open questions

on the influence of, for example, magenta ink volume.

3. The full-scale print trial indicated that print mottle was lower on samples

with coating structures having large pores and high pore-volume. Others

have found that highly porous coating layer may cause deeper ink

penetration and thus lower optical density (Preston et al., 2008a).

However, in the study here, no negative effects were seen form porous

coating layers. Is this because of the tested range of samples? Will the

effects be different under other circumstances?

4. In the absorption stain test, a dye is used as colorant. Depending on its

ionic charge, it will adsorb on different coating constituents. When chosen

wisely, the dye may act as a marker for components in the coating layer

where the affinity is greater. With a deeper knowledge on dye adsorption,

testing liquids may be designed for different purposes, which expands the

versatility of the technique. For example, the colorant should probably be

chosen differently if predicting ink jet or flexographic printability.

(Young, 1805; Finn, 1999; Ridgway et al., 2001; n.d.) (TECHKON, 2000; Hauck

and Gooran, 2013) (Schoelkopf et al., 2000a){Schoelkopf

80

7 References Al-Turaif, H., Bousfield, D. W. & Lepoutre, P. (2002). The influence of substrate absorbency on coating surface chemistry. Prog. Org. Coat., vol. 44 no. 4, pp 307-315.

Alince, B. & Lepoutre, P. (1980). Porosity and optical properties of clay coatings. J. Colloid Interface Sci., vol. 76 no. 2, pp 439-444.

Arai, T., Yamasaki, T., Suzuki, K., Ogura, T. & Sakai, Y. (1988). The relationship between print mottle and coating structure. Tappi J., vol. 71 no. 5, pp 47-52.

Aspler, J. (1993). Interactions of ink and water with the paper surface in printing. Nord. Pulp Paper Res. J., vol. 8 no. 1, pp 68-74, 112.

Aspler, J. (2004). Print quality test methods for linerboard. Proceedings of: 56th TAGA Annual Technical Conference, 20 April 2004, San Antonio, USA. TAGA. pp 428-452.

Aspler, J., Bèland, M.-C., Boluk, M. Y. & Dalphond, J. E. (1992). Transfer and setting of water-based ink, part I: A new flexographic proof press. vol. 2. Proceedings of: 1992 TAGA Annual Technical Conference, April 1992, Vancouver, Canada. TAGA office.

Aspler, J., Cormier, L. & Manfred, T. (2004). Linerboard surface chemistry and structure affect flexographic print quality. Proceedings of: International Printing & Graphic Arts Conference, 4-6 Oct. 2004, Montreal, Canada. Montreal, Canada: PAPTAC. pp 167-177.

Aspler, J., Jordan, B., Zang, H. & Nguyen, N. (1998). Print Quality of Linerboard in Commercial Water-based Flexography. Proceedings of: 1998 TAGA Annual Technical Conference, 26-29 April 1998, Chicago, USA. TAGA office. pp 749-781.

Aspler, J. S. & Lepoutre, P. (1991). The transfer and setting of ink on coated paper. Prog. Org. Coat., vol. 19 no. 4, pp 333-357.

Azimi, Y., Kortschot, M. T. & Farnood, R. (2011). Surface roughening by wetting of coated papers at meso- and microscale. Appita J., vol. 64 no. 5, pp 428-432.

Barros, G. G., Fahlcrantz, C. M. & Johansson, P. A. (2005). Topographic Distribution of UnCovered Areas (UCA) in Full Tone Flexographic Prints. TAGA Journal, vol. 2 no. 1, pp 43-57.

Barros, G. G. & Johansson, P. A. (2006). Prediction of uncovered area occurrence in flexography based on topography: a feasibility study. Nord. Pulp Paper Res. J., vol. 21 no. 2, pp 172-179.

Barros, G. G. & Johansson, P. A. (2007). Mottle and ink distribution in full-tone flexography prints on coated linerboard. J. Pulp Pap. Sci., vol. 33 no. 4, pp 217-226.

Bassemir, R. W. & Krishnan, R. (1990). Practical Applications of Surface Energy Measurements in Flexography. Flexo, vol. 15 no. 7, pp 31-40.

Bassemir, R. W. & Krishnan, R. (1991). Surface and Colloidal Aspects of Dynamic Ink/Paper Interactions during Printing. Proceedings of: 1991 Symposium on Papercoating Fundamentals, 17-18 May 1991, Montreal, Canada. Montreal, Canada: Canadian Pulp and Paper Association. pp 69-75.

Beers, Y. (1957). Introduction to the Theory of Error, 2nd edition, Addison-Wesley Publishing Company Inc.: Reading, Massachusetts, USA.

Bergvall, R. (2006). Penetration och absorption av vätska i kartong. M.Sc. Thesis, Luleå University, http://epubl.ltu.se/1402-1617/2006/186/index.html.

Blake, T. D. & Ruschak, K. J. (1997). Wetting: Static and Dynamic Contact Lines. In: S.F., K. & P.M., S. (eds.) Liquid Film Coating. Chapman & Hall: London, UK. ISBN 0 412 06481 2.

Bohan, M. F. J., Claypole, T. C. & Gethin, D. T. (2002). The application of roughness model to a soft EHL contact. Int. J. Numer. Methods Fluids, vol. 12 no. 4, pp 356-374.

81

Bohan, M. F. J., Claypole, T. C., Gethin, D. T. & Megat Ahmed, M. M. H. (2000). A model for ink impression in printing contacts. J. Pulp Pap. Sci., vol. 26 no. 11, pp 414-419.

Bohlin, E. (2013). Surface and porous structure of pigment coatings: Interactions with flexographic ink and effects on print quality. PhD Thesis, Karlstad University, Karlstad, Sweden. ISBN 978-91-7063-528-1.

Bohlin, E. & Johansson, C. (2018). Print quality of flexographic printed paperboard related to coating composition and structure. TAPPI J., vol. 17 no. 1, pp 43-51.

Bohlin, E., Johansson, C. & Lestelius, M. (2016). Flexographic ink-coating interactions: effects of latex variations in coating layers. TAPPI J., vol. 15 no. 4, pp 254-262.

Bohlin, E., Lestelius, M. & Johansson, C. (2013). Flexographic ink-coating interactions - Effects of porous structure variations of coated paperboard. Nord. Pulp Paper Res. J., vol. 28 no. 4, pp 573-581.

Bosanquet, C. H. (1923). On the flow of liquids into capillary tubes. Philos. Mag., Series 6, vol. 45 no. 267, pp 525-531.

Bousfield, D. W. (1991). The leveling of coating defects. Proceedings of: 1991 Coating Conference, 19-22 May 1991, Montreal, Canada. Atlanta, USA: TAPPI Press. pp 101-111.

Brandon, S. & Marmur, A. (1996). Simulation of Contact Angle Hysteresis on Chemically Heterogeneous Surfaces. J. Colloid Interface Sci., vol. 183 no. 2, pp 351-355.

Bristow, A. J. (1967). Liquid Absorption into Paper During Short Time Intervals. Svensk Papperstid., vol. 70 no. 19, pp 623-629.

Bristow, A. J. & Bergenblad, H. (1981). Interpretations of ink-stain tests on coated papers. In: Banks, W. H. (ed.) Advances in Printing Science and Technology, vol. 16. Proceedings of: 16th International Conference of Printing Research Institutes, June 1981, Key Biscayne, USA. London, Plymouth, UK: Pentech Press. pp 373-391.

Bulo, R. E., Donadio, D., Laio, A., Molnar, F., Rieger, J. & Parrinello, M. (2007). “Site Binding” of Ca2+ Ions to Polyacrylates in Water: A Molecular Dynamics Study of Coiling and Aggregation. Macromolecules vol. 40 no. 9, pp 3437-3442.

Chinga, G. & Helle, T. (2003). Relationships between the coating surface structural variation and print quality. J. Pulp Pap. Sci., vol. 29 no. 6, pp 179-184.

Chung, R., Hsu, F., Clark, D. & Husain, K. (2009). Weight-based ink trapping assessment. Proceedings of: 61st TAGA Annual Technical Conference, 15-17 March 2009, New Orleans, USA. Sewickley, USA: TAGA.

Conte, A., Cappelletti, G. M., Nicoletti, G. M., Russo, C. & Del Nobile, M. A. (2015). Environmental implications of food loss probability in packaging design. Food Res. Int., vol. 78, pp 11-17.

Coyle, D. J. (1988). Forward roll coating with deformable rolls: A simple one-dimensional elastohydrodynamic model. Chem. Eng. Sci. , vol. 43 no. 10, pp 2673-2684.

Dahlström, C. & Uesaka, T. (2009). New insights into coating uniformity and base sheet structures. Ind. Eng. Chem. Res., vol. 48 no. 23, pp 10472-10478.

Dahlström, C. & Uesaka, T. (2012). Microstructure variations in paper coating: Direct observations. Ind. Eng. Chem. Res., vol. 51 no. 24, pp 8246-8252.

De Grace, J. H. & Mangin, P. (1984). A mechanistic approach to ink transfer part I: Effect of substrate properties and press conditions. In: Banks, W. H. (ed.) Advances in printing science and technology, vol. 17. Proceedings of: 17th International Conference of Printing Research Institutes, June 1983, Saltsjöbaden, Sweden. Plymouth, UK: Pentech Press. pp 312-332.

82

De Grace, J. H. & Mangin, P. J. (1988). A mechanistic approach to ink transfer. Part II - The splitting behavior of inks in printing nips. In: Banks, W. H. (ed.) Advances in Printing Science and Technology, vol. 19. Proceedings of: 19th International Conference of Printing Research Institutes, June 1987, Eisenstadt, Austria. London, UK: Pentech Press. pp 146-161.

Dean, T. W. R. (1997). The essential guide to aqueous coating of paper and board, PITA: Bury. ISBN 0-9530227-0-6.

Dobson, R. L. (1975). Burnout, a coat weight determination test re-examined. Proceedings of: 1975 Coating Conference, 21-23 April 1975, Chicago, USA. Atlanta, USA: TAPPI. pp 123-131.

Donigian, D. W. (2006). A new mechanism for the setting and gloss development of offset ink. J. Pulp Pap. Sci., vol. 32 no. 3, pp 163-177.

Dubé, M., Drolet, F., Daneault, C. & Mangin, P. J. (2008). Hydrodynamics of fluid transfer. J. Pulp Pap. Sci., vol. 34 no. 3, pp 174-181.

Duursma, G. R., Sefiane, K. & David, S. (2010). Advancing and receding contact lines on patterned structured surfaces. Chem. Eng. Res. Des., vol. 88 no. 5, pp 737-743.

Eklund, D. & Salminen, P. (1987). Water sorption in paper during short time. Appita J., vol. 40 no. 5, pp 340-346.

Elftonson, J. E. & Ström, G. (1995). Penetration of Aqueous Solutions into Models for Coating Layers. Proceedings of: 1995 Coating Fundamentals Symposium, 19-20 May 1995, Atlanta, USA. Tappi Press. pp 17-25.

Elton, N. J. & Preston, J. S. (2006a). Polarized light reflectometry for studies of paper coating structure Part II. Application to coating structure, gloss and porosity. Tappi J., vol. 5 no. 8, pp 10-16.

Elton, N. J. & Preston, J. S. (2006b). Polarized light reflectometry for studies of paper coating structure. Part I. Method and instrumentation. Tappi J., vol. 5 no. 7, pp 8-16.

Endres, I. & Tietz, M. (2007). Blade, film, and curtain coating techniques and their influence on paper surface characteristics. TAPPI J., vol. 6 no. 11, pp 24-32.

Engström, G. (2016). Causes of back-trap mottle in lithographic offset prints on coated papers. TAPPI J., vol. 15 no. 2, pp 91-101.

Engström, G. & Rigdahl, M. (1992). Kompendium i bestrykningsteknik, STFI: Stockholm, Sweden.

Fahlcrantz, C. M. & Johansson, P. A. (2004). Print mottle evaluation of flexogrphic prints: using a scanner-based measurement system. Flexo Magazine. FTA.

Fell, D., Auernhammer, G., Bonaccurso, E., Liu, C., Sokuler, R. & Butt, H.-J. (2011). Influence of Surfactant Concentration and Background Salt on Forced Dynamic Wetting and Dewetting. Langmuir, vol. 27 no. 6, pp 2112-2117.

FFTA (1999). Flexography: Principles & Practices. Ronkonkoma, USA: Foundation of Flexographic Technical Association, Inc.

Finn, R. (1999). Capillary surface interfaces. Notices of the AMS, vol. 46 no. 7, pp 770-781.

FTA (1992). Flexography Principles and Practices, 4th edition, Foundation of Flexographic Technical Association, Inc.: New York.

Fujiwara, H., Fujisaki, N. & Kaga, C. (1991). The Effect of Water Penetration on Offset Mottling (2). The Effect of Dwell Time on Offset Mottling with a Vdta Coater. Proceedings of: 1991 Coating Conference 19-22 May 1991, Montreal, Canada. Atlanta, USA: TAPPI Press. pp 179-188.

Fujiwara, H. & Kaga, C. (1992). Single and Double Blade Coating: The Variations in Sub-Millimeter Scale and Their Effects on Sheet and Print Qualities. Proceedings of: 1992 Coating Conference, 17-20 May 1992, Orlando, USA. Atlanta, USA: TAPPI Press. pp 147-159

83

Fujiwara, H., Kano, I. & Kaga, C. (1990). Measurement of Gloss Profile. Proceedings of: 1990 Coating Conference, 13-16 May 1990, Boston, USA. Atlanta, USA: TAPPI Press. pp 209-218

Gane, P. a. C. & Ridgway, C. J. (2009). Interactions between flexographic ink and porous coating structures. Proceedings of: PaperCon 2009 : solutions for a changing world, 31 May-3 June 2009, St. Louis, USA. Norcross (CD-rom), USA: TAPPI Press.

Gane, P. a. C., Schoelkopf, J., Spielmann, D. C., Matthews, G. P. & Ridgway, C. J. (1999). Observing fluid transport into porous coating structures: some novel findings. Tappi J., vol. 83 no. 5, pp 77.

Gate, L. F. & Windle, W. (2003). Absorption of oils into porous coatings. In: Bolam, F. (ed.) The fundamental properties of paper related to its use: Trans. 5th Fund. Res. Symp. 1973, Cambridge. Manchester, UK: FRC. pp 438-451.

Geffroy, C., Persello, J., Foissy, A., Lixon, P., Tournilhac, F. & Cabane, B. (2000). Molar mass selectivity in the adsorption of polyacrylates on calcite. Colloids Surf., A, vol. 162 no. 1–3, pp 107-121.

Grande, B. T. (2012). Ink trapping in offset printing [Online]. Accessed 15 June 2018 from https://www.dod.uni-wuppertal.de/fileadmin/_migrated/content_uploads/Ink_Trapping _in_Offset_Printing.pdf

Gros, M., Gandini, A., Blayo, A., Xiang, Y. & Bousfield, D. W. (2002). The influence of ink pigment concentration and filter cake resistance coefficient on ink setting rates. Proceedings of: 11th International Printing and Graphic Arts Conference, Oct. 2002, Bordeaux, France. Paris, France: Association technique de l'industrie papetière.

Hallberg Hofstrand, E., Lestelius, M. & Rättö, P. (2006). Dynamic measurements of nip force variations during post-printing of corrugated board. Nord. Pulp Paper Res. J., vol. 21 no. 1, pp 111-114.

Hattula, T. & Oittinen, P. (1982). An analysis of ink penetration into pigment coatings. Pap. Puu, vol. 6-7, pp 407-412.

Hauck, S. & Gooran, S. (2013). Investigation of the effect of ink penetration and gloss on a proposed spectral trapping model for high quality gloss coated paper. J. Print Media Technol. Res., vol. 2 no. 4, pp 235-244.

Havlínová, B., Cicák, V., Brezová, V. & Horňáková, L. (1999). Water-reducible flexographic printing inks-rheological behaviour and interaction with paper substrates. J. Mater. Sci., vol. 34 no. 9, pp 2081-2088.

Hiorns, A. G., Kent, D. & Parsons, J. (2005). Enhanced performance through multilayer coating. Proceedings of: 2005 Coating conference and exhibit, 17-20 Apr. 2005, Toronto, Canada. Atlanta (CD-rom), USA: TAPPI Press.

Holmvall, M. & Uesaka, T. (2008). Print uniformity of corrugated board in flexo printing: Effects of corrugated board and halftone dot deformations. Packaging Technology and Science, vol. 21 no. 7, pp 385-394.

Hou, M.-J. & Hui, S. H. (1990). Interfacial phenomena in surface sizing. Proceedings of: 1990 Papermakers Conference, 23-25 April 1990, Atlanta, USA. Atlanta, USA: TAPPI Press. pp 209-214.

Huang, W.-X., Lee, S.-H., Sung, H. J., Lee, T.-M. & Kim, D.-S. (2008). Simulation of liquid transfer between separating walls for modeling micro-gravure-offset printing. International Journal of Heat and Fluid Flow, vol. 29 no. 5, pp 1436-1446.

IGT (2006). W24 IGT print penetration [Online]. IGT Testing Systems. Almere, Netherlands. Accessed 16 Nov. 2018 from www.igt.nl/EN_bestanden/W_information_leaflets/W24.pdf

84

Innventia (2013). Packaging 2020: Innventia global outlook. Stockholm, Sweden.

Isoard, J. C. (1983). Ink Transfer and Retransfer - Mottling and Offset Picking of Coated Papers. Proceedings of: 1983 Coating Conference, 15-18 May 1983, San Fransisco, USA. Atlanta, USA. pp 143-152

Israelachvili, J. N. & Gee, M. L. (1989). Contact angles on chemically heterogeneous surfaces. Langmuir, vol. 5 no. 1, pp 288-289.

Jensen, K. W. (1989). Flexo Printability of Coated White-Top Liner with Consideration of Convertability. Graphic Arts in Finland, vol. 18 no. 1, pp 14-20.

Johansson, P. A. (1993). Print mottle evaluation by band-pass image analysis. Advances in Printing and Media Technology, vol. 22. Proceedings of: International Research Conference, 5-8 Sept. 1993, Munich, Germany. Munich, Germany: Pentech Press. pp 403-413.

Johansson, P. A. & Norman, B. (1996). Methods for evaluating formation, print unevenness and gloss variations developed at STFI. Proceedings of: 1996 Process and product quality conference, 14-17 Oct 1996, Cincinnati, USA. Atlanta, USA: TAPPI Press. pp 139-145.

Johnson, J., Lestelius, M., Blohm, E., Rättö, P. & Järnström, L. (2008). The interaction between Water and Liner and Newsprint in Flexographic CI-Printing Press. Proceedings of: 60th TAGA Annual Technical Conference, 16-19 March 2008, Sant Francisco, USA. TAGA. pp 267-284.

Johnson, J., Rättö, P., Järnström, L., Lestelius, M. & Blohm, R. (2005). Interaction Between Water And Paperboard And Liner In A Flexographic Printing Press. TAGA Journal, vol. 1, pp 110-128.

Johnson, J., Rättö, P., Lestelius, M. & Blohm, E. (2004). Measuring the dynamic pressure in a flexographic central impression printing press. Nord. Pulp Paper Res. J. , vol. 19 no. 1, pp 84-88.

Järnström, J., Väisänen, M., Lehto, R., Jäsberg, A., Timonen, J. & Peltonen, J. (2010). Effect of latex on surface structure and wetting of pigment coatings. Colloids Surf., A, vol. 353 no. 2, pp 104-116.

Järnström, L. (1993). The polyacrylate demand in suspensions containing ground calcium carbonate. Nord. Pulp Paper Res. J., vol. 8 no. 1, pp 27-33.

Kamal Alm, H. (2016). Interfacial Adhesion Failure; Impact on print-coating surface defects. PhD Thesis, KTH Royal Institute of Technology, Stockholm, Sweden. ISBN 978-91-7729-136-7.

Kamal Alm, H., Ström, G., Karlström, K., Schoelkopf, J. & Gane, P. a. C. (2010). Effect of excess dispersant on surface properties and liquid interactions on calcium carbonate containing coatings. Nord. Pulp Paper Res. J., vol. 25 no. 1, pp 82-92.

Kamal Alm, H., Ström, G., Schoelkopf, J. & Gane, P. a. C. (2015a). Ink-lift-off during offset printing: a novel mechanism behind ink–paper coating adhesion failure. J. Adhes. Sci. Technol., vol. 29 no. 5, pp 370-391.

Kamal Alm, H., Ström, G., Schoelkopf, J., Ridgway, C. & Gane, P. a. C. (2015b). Ink adhesion failure during full scale offset printing: causes and impact on print quality. J. of Print and Media Tech. Res., vol. 4 no. 4, pp 257-278.

Kamal, H., Ström, G., Schoelkopf, J. & Gane, P. a. C. (2010). Characterization of Ink Adhesion Failure; Effect of Pre-dampening of Carbonate Containing Coatings. J. Adhes. Sci. Technol., vol. 24 no. 3, pp 449-469.

Kane, R. J. (1987). The use of auxiliary dispersant to improve high speed coated performance. Proceedings of: TAPPI Coating conference, 17-21 May 1987, Houston, USA. Atlanta, USA: Tappi Press. pp 117.

85

Kargupta, K., Konnur, R. & Sharma, A. (2001). Spontaneous dewetting and ordered patterns in evaporating thin liquid films on homogeneous and heterogeneous substrates. Langmuir, vol. 17 no. 4, pp 1294-1305.

Kheshgi, H. S. (1997). 6. The fate of thin liquid films after coating. In: Kistler, S. F. & Schweizer, P. M. (eds.) Liquid film coating. Chapman & Hall: London. ISBN 0 412 06481 2.

Kim, L. H., Pollock, M. J., Wittbrodt, E. L., Roper, J. A., Smith, D. A., Stolarz, J. W., Rolf, M. J., Green, T. J. & Langolf, B. J. (1998). Reduction of back-trap mottle through optimization of the drying process for paper coatings, part I. TAPPI J., vol. 81 no. 8, pp 153-164.

Kim, T. & Kim, W. (2018). Viscous dewetting of metastable liquid films on substrates with microgrooves. J. Colloid Interface Sci., vol. 520, pp 11-18.

Kipphan, H. (2001). Handbook of print media: technologies and production methods (Illustrated), Springer: Berlin, Heidelberg, New York. ISBN 3-540-67326-1.

Koivula, H., Kamal Alm, H. & Toivakka, M. (2011). Temperature and moisture effects on wetting of clacite surfaces by offset ink constituents. Colloids Surf., A, vol. 390 no. 1, pp 105-111.

Koivula, H., Preston, J. S., Heard, P. J. & Toivakka, M. (2008). Visualisation of the distribution of offset ink components printed onto coated paper. Colloids Surf., A, vol. 317 no. 1, pp 557-567.

Kolseth, P. (2004). Coating structure and print quality (slides only). Proceedings of: New concepts in paper surface treatment, 19-20 Oct. 2004, Stockholm, Sweden. Stockholm, Sweden: Institute of Surface Chemistry.

KRK (2015a). Bristow tester (pressurized type) [Online]. Kumagai Riki Kogyo Co. Ltd. Tokyo, Japan. Accessed 2018-11-15 from www.krk-kumagai.co.jp/dcms_media/other/ KRK%202072-1%2C2%20Bristow%20tester.pdf

KRK (2015b). No.2071 Automatic scanning liquid absorptmeter KM500win [Online]. Kumagai Riki Kogyo Co. Ltd. Tokyo, Japan. Accessed 16 Nov. 2018 from http://www.krk-kumagai.co.jp/dcms_media/other/KRK%202071%20absorption.pdf

Krüss (n.d.). Dynamic surface tension [Online]. Krüss GmbH. Hamburg, Germany. Accessed 5 Nov. 2018 from https://www.kruss-scientific.com/services/education-theory/glossary/dynamic-surface-tension/

Lagerstedt, P. & Kolseth, P. (1995). Influence of surface energetics on ink transfer in flexo printing. Advances in printing science and technology, vol. 23. Proceedings of: International Research Conference, 17-20 Sept. 1995, Paris, France. Chichester, UK: John Wiley & Sons. pp 269-299.

Lamminmäki, T., Forsström, U., Tienari, M., Saharinen, E., Mikkilä, J. & Koskelainen, J. (2005). Calendering of coated paper at high speeds and temperatures. Nord. Pulp Paper Res. J., vol. 20 no. 3, pp 354-360.

Lamminmäki, T., Kettle, J., Puukko, P., Ketoja, J. & Gane, P. (2010). The role of binder type in determining inkjet print quality. Nord. Pulp Paper Res. J. , vol. 25 no. 3, pp 380-390.

Lamminmäki, T., Kettle, J., Rautkoski, H., Kokko, A. & Gane, P. (2011). Limitations of current formulations when decreasing the coating layer thickness of papers for inkjet printing. Ind. Eng. Chem. Res., vol. 50 no. 12, pp 7251-7263.

Larrondo, L. & St-Amour, S. (1992). A Method for Measuring the Void Fraction of Coatings on Porous Substrates. Proceedings of: 1992 Coating Conference, 17-20 May 1992, Orlando, USA. Atlanta, USA: TAPPI Press. pp 37-51.

Larrondo, L. & St-Amour, S. (1994). Method for measuring the void fraction of coatings on porous substrates. J. Pulp Paper Sci., vol. 20 no. 8, pp 215-219.

86

Larsson, M., Engström, G., Vidal, D. & Zou, X. (2007). Impact of calendering on coating structures. Nord. Pulp Paper Res. J., vol. 22 no. 2, pp 267-274.

Laudone, G. M., Matthews, G. P. & Gane, P. a. C. (2003). Coating shrinkage during evaporation: observation, measurement and modelling within a network structure. Proceedings of: 8th Advanced coating fundamentals symposium, 8-10 May 2003, Chicago, USA. Atlanta, USA: Tappi Press. pp 18-.

Laudone, G. M., Matthews, G. P. & Gane, P. a. C. (2006). Effect of latex volumetric concentration on void structure, particle packing, and effective particle size distribution in a pigmented paper coating layer. Ind. Eng. Chem. Res., vol. 45 no. 6, pp 1918-1923.

Lavelle, J. S., Quinn, J. R., Gallagher, J. E. & Micale, F. J. (1996). Measurement of flexographic ink transfer on a modified Prufbau. Proceedings of: 1996 International Printing & Graphic Arts Conference, 16-19 Sept 1996, Minneapolis, USA. Atlanta, USA: TAPPI Press. pp 199-222.

Leach, R. H. & Pierce, R. J. (eds.) (1993). The printing ink manual, 5th edition, Blueprint: London. ISBN 0 948905 81 6.

Lee, H. L., Kim, J. D., Lee, K. H., Kim, C. H. & Youn, H. J. (2012). Effect of coating formulations and drying methods on the coverage and smoothness of brown grade base papers. Nord. Pulp Paper Res. J., vol. 27 no. 1, pp 79-85.

Lepoutre, P. (1978a). Liquid Absorption and Coating Porosity. Paper Technology and Industry, vol. 19 no. 9, pp 298-304.

Lepoutre, P. (1978b). Paper Coatings: Substrate Absorbency and Coating Structure. Tappi, vol. 61 no. 5, pp 51-55.

Lepoutre, P. (1989). The structure of paper coatings: an update. Prog. Org. Coat., vol. 17 no. 2, pp 89-106.

Lepoutre, P., Degrace, J. H. & Mangin, P. J. (1979). Printability of coated papers: Influence of coating absorbency. Tappi, vol. 62 no. 5, pp 33-36.

Lim, C. H., Bohan, M. F. J., Claypole, T. C., Gethin, D. T. & Roylance, B. J. (1996). A finite element investigation into a soft rolling contact supplied by a non-Newtonian ink. J. Phys. D: Appl. Phys., vol. 29 no. 7, pp 1894-1903.

Liu, F. & Shen, W. (2008). Forced wetting and dewetting of liquids on solid surfaces and their roles in offset printing. Colloids Surf., A, vol. 316 no. 1, pp 62-69.

Loiseau, J., Ladavière, C., Suau, J. M. & Claverie, J. (2005). Dispersion of calcite by poly(sodium acrylate) prepared by reversible addition–fragmentation chain transfer (RAFT) polymerization. Polymer, vol. 46 no. 19, pp 8565-8572.

Lucas, R. (1918). Ueber das Zeitgesetz des kapillaren Aufstiegs von Flüssigkeiten. Kolloid Z., vol. 23 no. 1, pp 15-22.

Luo, H. & Gersappe, D. (2004). Dewetting Dynamics of Nanofilled Polymer Thin Films. Macromolecules, vol. 37 no. 15, pp 5792-5799.

Luu, W. T., Bousfield, D. W. & Kettle, J. (2010a). Influence of ink chemistry and uncoated paper surface energy on setting and transfer of water-based flexographic inks. vol. 4. Proceedings of: PaperCon 2010, 2-5 May 2010, Atlanta, USA. Norcross (CD-rom), USA: TAPPI.

Luu, W. T., Bousfield, D. W., Kettle, J. & Aspler, J. (2010b). 8.4 - Influence of ink chemistry and surface energy on flexographic print quality. Proceedings of: 11th Advanced coating fundamentals symposium, 11-13 Oct. 2010, Munich, Germany. Atlanta, USA: TAPPI Press.

87

Lyne, M. B. (2002). 7. Wetting and the Penetration of Liquids into Paper. In: Borch, J., Lynce, M. B., Mark, R. E. & Habeger, C. C. J. (eds.) Handbook of Physical Testing of Paper. 2nd edition. Marcel Dekker: New York.

Löfgren, M. & Witell, L. (2005). Kano’s Theory of Attractive Quality and Packaging. Quality Management Journal vol. 12 no. 3, pp 7-20.

Löfgren, M., Witell, L. & Gustafsson, A. (2011). Theory of attractive quality and life cycles of quality attributes. The TQM Journal, vol. 23 no. 2, pp 235-246.

Ma, D., Carter, R. D. & Bousfield, D. W. (2009). Coating structure influence on flexographic printing. Solutions for a changing world. Proceedings of: PaperCon 2009, 31 May-3 June 2009, St. Louis, USA. Norcross (CD-rom), USA: Tappi Press.

Ma, D., Carter, R. D., Chen, C. Y. & Hardy, R. E. (2005). Print mottle reduction through clay engineering and pore structuring in paper coating. Proceedings of: 2005 Coating conference and exhibit, 17-20 April 2005, Toronto, Canada. Atlanta (CD-rom), USA: TAPPI Press. pp Session 27.

Ma, D., Carter, R. D., Laine, J. & Stenius, P. (2007). Gibbs energy analysis of ink oil imbibition during ink setting. Nord. Pulp Paper Res. J., vol. 22 no. 4, pp 523-528.

MacGregor, M. A. & Conners, T. E. (1987). Image analysis of an LWC paper reveals wire mark in the print density variations. Tappi J., vol. 70 no. 9, pp 95-99.

MacPhee, J. (1979). An engineer´s analysis of the lithographic printing process. Proceedings of: 1979 TAGA Annual Meeting, 7-10 May 1979, Kansas City, USA. Rochester, USA: TAGA. pp 237-277.

Magnier, L., Schoormans, J. & Mugge, R. (2016). Judging a product by its cover: Packaging sustainability and perceptions of quality in food products. Food Qual. Preference, vol. 53, pp 132-142.

McConkey, F., Preston, J., Parsons, C. G. & Husband, J. C. (2004). Optimization of latex level to control water interference mottle of coated paperboard: a mill study. Proceedings of: 2004 Coating and graphic arts conference and exhibit, 16-19 May 2004, Baltimore, USA. Atlanta, USA: TAPPI Press. pp 16pp.

McGratten, W. (1990). Key characteristics of linerboard, corrugating medium, and roll stock mechanical condition and their influence on the manufacture of corrugated products, part 1. Tappi J., vol. 73 no. 11, pp 99-108.

Megat Ahmed, M. M. H., Gethin, D. T., Claypole, T. C. & Roylance, B. J. (1997). A model for ink impression into a porous substrate. J. Phys. D: Appl. Phys., vol. 30 no. 16, pp 2276-2284.

Mesic, B., Jarnstrom, L., Hjarthag, C. & Lestelius, M. (2004). Effects of application temperature in paper surface sizing with temperature-responsive starch on water repellency and flexographic printability. Appita J., vol. 57 no. 4, pp 281-298.

Mirle, S. K. & Zettlemoyer, A. C. (1988). Viscoelasticity of photopolymer plates and ink hydrodynamics in modelling flexographic printing. Mathl Comput. Modelling, vol. 11 no. C, pp 1162-1165.

Molnar, F. & Rieger, J. (2005). “Like-charge attraction” between anionic polyelectrolytes: molecular dynamics simulations. Langmuir, vol. 21 no. 2, pp 786-789.

Morgan, M. L., Holder, A., Curtis, D. J. & Deganello, D. (2018). Formulation, characterisation and flexographic printing of novel Boger fluids to assess the effects of ink elasticity on print uniformity. Rheol. Acta, vol. 57, pp 105-112.

Mulji, N. & Chandra, S. (2010). Rupture and dewetting of water films on solid surfaces. J. Colloid Interface Sci., vol. 352 no. 1, pp 194-201.

88

Murakami, K. & Imamura, R. (1983). 17. Porosity and Gas Permeability. In: Mark, R. E. (ed.) Handbook of Physical and Mechanical Testing of Paper and Paperboard. 1st edition. Marcel Dekker: New York, Basel. ISBN 0-8247-7052-8.

Myers, R. R., Miller, J. C. & Zettlemoyer, A. C. (1959). The splitting of thin liquid films. Kinematics. J. Colloid Sci., vol. 14 no. 3, pp 287-299.

Oliver, J. F. & Mason, S. G. (2003). Scanning electron microscope studies of spreading of liquids on paper. In: Bolam, F. (ed.) The Fundamental Properties of Paper Related to its Uses: Trans. 5th Fund. Res. Symp. 1973, Cambridge. Manchester, UK: FRC. pp 428–437.

Olsson, R. (2007). Some aspects on flexographic ink-paper and paperboard coating interaction (manuscript no. 5). PhD Thesis, Karlstad University, Karlstad, Sweden. ISBN 978-91-7063-126-9.

Olsson, R., Van Stam, J. & Lestelius, M. (2006a). Imbibition into model calcium carbonate coatings of fluids with different dipole moments. Appita J., vol. 59 no. 6, pp 471-485.

Olsson, R., Yang, L. & Lestelius, M. (2007a). Water retention of flexographic inks and its influence on final print gloss. Nord. Pulp Paper Res. J., vol. 22 no. 3, pp 287-292.

Olsson, R., Yang, L., Van Stam, J. & Lestelius, M. (2006b). Effects on ink setting in flexographic printing: Coating polarity and dot gain. Nord. Pulp Paper Res. J., vol. 21 no. 5, pp 569-574.

Olsson, R., Yang, L., Van Stam, J. & Lestelius, M. (2007b). Effects of elevated temperature on flexographic printing. In: Enlund, N. & Lovreček, M. (eds.) Advances in Printing and Media Technology, vol. 34. Proceedings of: International Research Conference, Sept 9-12 2007, Grenoble, France. Zagreb, Croatia: Acta Graphica Publishers.

Omidian, H., Hashemi, S. A., Sammes, P. G. & Meldrum, I. (1998). A model for the swelling of superabsorbent polymers. Polymer, vol. 39 no. 26, pp 6697-6704.

Ovaska, S.-S. & Backfolk, K. (2018). The versatility of the Bristow absorption tester - a review. Nord. Pulp Paper Res. J., vol. 33 no. 2, pp 279-296.

Owens, D. K. & Wendt, R. C. (1969). Estimation of surface free energy of polymers. J. Appl. Polym. Sci., vol. 13 no. 8, pp 1741-1747.

Pauler, N. (1987). A model for the interaction between ink and paper. In: Banks, W. H. (ed.) Advances in printing science and technology. Proceedings of: 19th International conference of Printing Research Institutes, June 1987, Eisenstadt, Austria. London, UK: Pentech Press. pp 116-136.

Pauler, N. (1999). Paper optics, AB Lorentzen och Wettre: Sweden. ISBN 91-971 765-6-7.

Preston, J., Gaskin, E., Toivakka, M., Koppolu, R. & Salminen, P. (2017). Impact of pre-wetting on the subsequent absorption of water and its applicability to drying of multi coloured flexo prints. Proceedings of: 28th PTS Coating Symposium, 5-6 Sept. 2017a, Munich, Germanypres. Munich, Germany: Papiertechnische Stiftung PTS. pp 74-87.

Preston, J., Hiorns, A., Parsons, D. J. & Heard, P. (2008a). Design of coating structure for flexographic printing. Paper Technology, vol. 49 no. 3, pp 27-36.

Preston, J., Hiorns, A. G., Elton, N. & Strom, G. (2008b). Application of imaging reflectometry to studies of print mottle on commercially printed coated papers. Tappi J., vol. 7 no. 1, pp 11-18.

Preston, J., Toivakka, M., Findlay, A., Tsukuda, R. S., Oliveira, E. & Costa, E. (2017b). Pigment impacts on strength in a paper coating & movement of binder during coating colour consolidation. O Papel, vol. 78 no. 8, pp 94-104.

Preston, J. S., Elton, N. J., Husband, J. C., Heard, P. J., Allen, G. C. & Legrix, A. (2000). SIMS analysis of printed paper surfaces to determine distribution of ink components after printing.

89

Proceedings of: 2000 TAPPI International Printing and Graphic Arts Conference, 1-4 Oct. 2000, Savannah, USA. Atlanta, USA: TAPPI Press. pp 101-120.

Preston, J. S., Elton, N. J., Legrix, A., Nutbeem, C. & Husband, J. C. (2002). The role of pore density in the setting of offset printing ink on coated paper. TAPPI J., pp 3-5.

Preucil, F. Color hue and ink transfer—their relation to perfect reproduction. Proceedings of: TAGA1953. pp 102-110.

Purfeerst, R. D. & Van Gilder, R. L. (1991). Tail-edge picking, back-trap mottle and fountain solution interference of model latex coatings on a six-colour press. Proceedings of: 1991 Coating Conference, May 19-22 1991, Montreal. Atlanta, USA: TAPPI Press. pp 461-472.

Quesne, B., Reverdy-Bruas, N., Beneventi, D. & Chaussy, D. (2013). Investigation into the causes of marbling on PVC printed with water-based inks. In: Enlund, N. & Lovreček, M. (eds.) Advances in Printing and Media Technology, vol. 40. Proceedings of: International Research Conference, 8-11 Sep. 2013, Chemnitz, Germany. Darmstadt, Germany: iarigai. pp 219-224.

Ragab, A. E. R. & El Kader, M. E. A. (2012). Analysis of trapping of color sequences of multicolor offset printing. International Design Journal, vol. 1 no. 1, pp 55-61.

Ragnarsson, M., Engström, G. & Järnström, L. (2013). Porosity variations in coating layers - Impact on back-trap mottle. Nord. Pulp Paper Res. J., vol. 28 no. 2, pp 257-263.

Ramé-Hart (n.d.). ramé-hart: Information on Contact Angle [Online]. ramé-hart instrument co. Succasunna, USA. Accessed 9 Sept. 2014 from http://www.ramehart.com/contactangle.htm

Ranger, A. E. (2003). Coating Pore Structure Analysis by Fluid Penetration and Permeation. In: Brander, J. (ed.) The Role of Fundamental Research in Papermaking: Trans. 7th Fund. Res. Symp. , vol. 2, Sept. 1981, Cambridge. Manchester, UK: FRC. pp 685-705.

Redon, C., Brochard-Wyart, F. & Rondelez, F. (1991). Dynamics of dewetting. Phys. Rev. Lett., vol. 66 no. 6, pp 715-718.

Rennes, S. & Eklund, D. (1989). The influence of binders on the structure and water sorption of coated paper. Pap. Puu, vol. 71, pp 698-703.

Ridgway, C. J., Gane, P. a. C. & Schoelkopf, J. (2002). Effect of Capillary Element Aspect Ratio on the Dynamic Imbibition within Porous Networks. J. Colloid Interface Sci., vol. 252 no. 2, pp 373-382.

Ridgway, C. J., Kukkamo, V. & Gane, P. a. C. (2011). Effects of binders and additives on inkjet coating pigment pore structures. Proceedings of: PaperCon ‘11 1-4 May 2011, Covington, USA. Norcross, USA: TAPPI. pp 1305-1319.

Ridgway, C. J., Schoelkopf, J., Gane, P. a. C., Matthews, G. P. & James, P. W. (2001). The effects of void geometry and contact angle on the absorption of liquids into porous calcium carbonate structures. J. Colloid Interface Sci., vol. 239 no. 2, pp 417-431.

Rogan, K. R., Bentham, A. C., George, I. A. & Skuse, D. R. (1994). Colloidal stability of calcite dispersion treated with sodium polyacrylate. Colloid. Polym. Sci., vol. 272 no. 10, pp 1175-1189.

Rohmann, M. E. (1974). Influence of sodium polyacrylate on properties of coating colors. Tappi, vol. 57 no. 5, pp 96-100.

Ruiz-Cabello, F. J. M., Rodríguez-Valverde, M. A. & Cabrerizo-Vílchez, M. A. (2014). Equilibrium contact angle or the most-stable contact angle? Adv. Colloid Interface Sci., vol. 206, pp 320-327.

Salminen, P. (1988). Studies of water transport in paper during short contact times. PhD Thesis, Åbo Akademi, Åbo, Finland. ISBN 951-649-459-9.

Sauer, H. M., Roisman, I. V., Dörsam, E. & Tropea, C. (2018). Fast liquid sheet and filament dynamics in the fluid splitting process. Colloids Surf., A.

90

Schmid, M. (2011). Curtain coater as air knife replacement. Proceedings of: PaperCon 2011, 1-4 May 2011, Covington, USA. Norcross, USA: TAPPI. pp 2514-2521.

Schoelkopf, J., Gane, P. a. C. & Ridgway, C. J. (2004). Observed non-linearity of Darcy-permeability in compacted fine pigment structures. Colloids Surf., A, vol. 236 no. 1–3, pp 111-120.

Schoelkopf, J., Gane, P. a. C., Ridgway, C. J. & Matthews, G. P. (2000a). Influence of inertia on liquid absorption into paper coating structures. Nord. Pulp Paper Res. J., vol. 15 no. 5, pp 422-430.

Schoelkopf, J., Gane, P. a. C., Ridgway, C. J. & Matthews, G. P. (2002). Practical observation of deviation from Lucas–Washburn scaling in porous media. Colloids Surf., A, vol. 206 no. 1–3, pp 445-454.

Schoelkopf, J., Ridgway, C. J., Gane, P. a. C., Matthews, G. P. & Spielmann, D. C. (2000b). Measurement and Network Modeling of Liquid Permeation into Compacted Mineral Blocks. J. Colloid Interface Sci., vol. 227 no. 1, pp 119-131.

Shaw, D. J. (1980). Introduction to colloid and surface chemistry, 3rd ed. edition, Butterworths. ISBN 10: 0408710497.

Shen, W., Hutton, B. & Liu, F. (2004). A new understanding on the mechanism of fountain solution in the prevention of ink transfer to the non-image area in conventional offset lithography. J. Adhes. Sci. Technol., vol. 18 no. 15-16, pp 1861-1887.

Shen, Y., Bousfield, D. W., Van Heiningen, A. & Donigian, D. (2005). Linkage between coating absorption uniformity and print mottle. J. Pulp Paper Sci. , vol. 31 no. 3, pp 105-108.

Sheng, Y. J., Shen, W. & Parker, I. H. (2000). Importance of the substrate surface energetics in water-based flexographic printing. Appita J., vol. 53 no. 5, pp 367-370.

Shi, Z., Zhang, Y., Liu, M., Hanaor, D. a. H. & Gan, Y. (2018). Dynamic contact angle hysteresis in liquid bridges. Colloids Surf., A, vol. 555, pp 365-371.

Sinn, C. G., Dimova, R. & Antonietti, M. (2004). Isothermal titration calorimetry of the polyelectrolyte/water interaction and binding of Ca2+: effects determining the quality of polymeric scale inhibitors. Macromolecules, vol. 37 no. 9, pp 3444-3450.

Skowronski, J., Gruener, G. & De Groot, W. (2005). Ultrasonic Systems for Liquid Absorption Measurement: Part. 2. Validation of Ultrasonic Testing for Liquid Absorption Uniformity. Proceedings of: Practical Papermaking Conference, 22-26 May 2005, Milwaukee, USA. Atlanta, USA: TAPPI Press. pp 1-22.

Sohn, O. & Kim, D. (2003). Theoretical and experimental investigation of the swelling behavior of sodium polyacrylate superabsorbent particles. J. Appl. Polym. Sci., vol. 87 no. 2, pp 252-257.

Steadman, R., Woodall, M. & Lesniak, M. (1993). The Flexographic Printability of Linerboard. vol. 2. Proceedings of: 47th Appita Annual General Conference, 19-23 April 1993, Rotorua, New Zealand. Parkville, Australia: Appita. pp 629-637.

Ström, G. (2005). Interaction between offset ink and coated paper: a review of the present understanding. In: I'anson, S. J. (ed.) Advances in Paper Science and Technology: 13th Fund. Res. Symp. , vol. 3, Sept. 2005, Cambridge, UK. Bury, UK: Pulp and Paper Fundamental Research Society. pp 1101-1137.

Ström, G., Borg, J. & Svanholm, E. (2008). Short-time water absorption by model coatings. Proceedings of: 10th Advanced Coating Fundamental Symposium, 11-13 June 2008, Montreal, Canada. Atlanta, USA: TAPPI Press. pp 204-216.

Ström, G., Gustafsson, J. & Sjolin, D. (2001). Printing papers separation of ink constituents during ink setting on coated substrates. Proceedings of: 2001 Coating and graphic arts conference and trade fair, 6-9 May 2001, San Diego, USA. Atlanta, USA: TAPPI Press. pp 303-313.

91

Ström, G., Hornatowska, J. & Terasaki, O. (2010). A novel SEM cross-section analysis of paper coating for separation of latex from void volume. Nord. Pulp Paper Res. J., vol. 25 no. 1, pp 107-113.

Ström, G. & Preston, J. (2013). Impact of local variation in coating structure on uniformity in print gloss. TAPPI J., vol. 12 no. 5, pp 43-51.

Sutera, S. P. & Skalak, R. (1993). The history of Poiseuille's law. Annu. Rev. Fluid Mech., vol. 25 no. 1, pp 1-19.

Taylor, J. H. & Zettlemoyer, A. C. (1958). Hypothesis on the mechanism of ink splitting during printing. Tappi, vol. 41 no. 12, pp 749-757.

TECHKON (2000). Measurement of Trapping - Application Note 4 [Online]. TECHKON GmbH. Königstein, Germany. Accessed 15 June 2008 from http://www.beijersbergen.com/ downloads/techkon%20trapping%20of%20inktaanname.pdf

Thorman, S., Ström, G. & Gane, P. a. C. (2018a). Impact of non-uniform water absorption on water-interference print mottle in offset printing. Nord. Pulp Paper Res. J., vol. 33 no. 1, pp 150-163.

Thorman, S., Ström, G., Hagberg, A. & Johansson, P. A. (2012). Uniformity of liquid absorption by coatings-Technique and impact of coating composition. Nord. Pulp Paper Res. J., vol. 27 no. 2, pp 456-465.

Thorman, S., Yang, L. & Hagberg, A. (2013). Simultaneous determination of absorption mottle and white-top mottle in the same area on coated boards. In: Enlund, N. & Lovreček, M. (eds.) Advances in Printing and Media Technology, vol. 40. Proceedings of: International Research Conference, 8-11 Sep. 2013, Chemnitz, Germany. Darmstadt, Germany: iarigai. pp 225-232.

Thorman, S., Yang, L., Hagberg, A. & Ström, G. (2018b). The impact of non-uniform ink absorption on flexographic print mottle. J. Print Media Technol. Res., vol. 7 no. 1, pp 7-18.

TMI (2015). Automatic Cobb Tester ACT [Online]. Testing Machines Inc. Accessed 25 Oct. 2018 from www.testingmachines.com/pdf/61-76-automatic-cobb-tester.pdf

Tollenaar, D. & Ernst, P. a. H. (1961). Optical density and ink layer thickness. In: Banks, W. H. (ed.) Advances in printing science and technology, vol. 2. Proceedings of: 6th International Conference of Printing Research Institiutes1961, Elsinore, Denmark. Oxford, UK: Pergamon Press. pp 214-234.

Tollenaar, D. & Ernst, P. a. H. (1965). Tack experiments for wet-on-wet printing. In: Banks, W. H. (ed.) Advances in Printing Science and Technolgy, vol. 4. Proceedings of: 8th International Conference of Printing Research Institutes1965, Aulanko, Finland. Oxford, UK: Pergamon Press. pp 445-454.

Tomimasu, H., Suzuki, K. & Ogura, T. (1990). The effect of basestock structure on coating weight distribution. Tappi J., vol. 73 no. 5, pp 179-187.

Tyagi, S., Ray, A. K. & Sood, Y. V. (2010). Surface characteristics of coating layers formed by coating pigments with different particle sizes and size distribution. Journal of Coatings Technology Research, vol. 7 no. 6, pp 747-756.

Walker, W. C. & Fetsko, J. M. (1955). A concept of ink transfer in printing. American Ink Maker, vol. 33 no. 12, pp 38-44, 69-71.

Wallman, M. & Nilsson, K. (2011). Klimatpåverkan och energianvändning från livsmedelsförpackningar. Livsmedelsverkets rapportserie. https://www.livsmedelsverket.se/globalassets/publikationsdatabas/rapporter/2011/2011_livsmedelsverket_18_klimatpaverkan_energianvandning_livsmedelsforpackningar.pdf: SIK.

Van Ooijen, I., Fransen, M. L., Verlegh, P. W. J. & Smit, E. G. (2017). Packaging design as an implicit communicator: Effects on product quality inferences in the presence of explicit quality cues. Food Qual. Preference, vol. 62, pp 71-79.

92

Van Oss, C. J., Chaudhury, M. K. & Good, R. J. (1987). Monopolar surfaces. Adv. Colloid Interface Sci., vol. 28 no. C, pp 35-64.

Wang, Z., Winslow, R., Madan, D., Wright, P. K., Evans, J. W., Keif, M. & Rong, X. (2014). Development of MnO2 cathode inks for flexographically printed rechargeable zinc-based battery. J. Power Sources, vol. 268, pp 246-254.

Washburn, E. W. (1921). The Dynamics of Capillary Flow. Phys. Rev., vol. 17 no. 3, pp 273-283.

Watanabe, J. & Lepoutre, P. (1982). A mechanism for the consolidation of the structure of clay-latex coatings. J. Appl. Polym. Sci., vol. 27 no. 11, pp 4207-4219.

Wenzel, R. N. (1936). Resistance of solid surfaces to wetting by water. Ind. Eng. Chem. , vol. 28 no. 8, pp 988-994.

Wikström, F. & Williams, H. (2010). Potential environmental gains from reducing food losses through

development of new packaging – a life-cycle model. Packaging Technology and Science, vol. 23 no. 7, pp 403-411.

Wikström, F., Williams, H., Verghese, K. & Clune, S. (2014). The influence of packaging attributes on consumer behaviour in food-packaging life cycle assessment studies - a neglected topic. Journal of Cleaner Production, vol. 73, pp 100-108.

Williams, H. & Wikström, F. (2011). Environmental impact of packaging and food losses in a life cycle perspective: a comparative analysis of five food items. Journal of Cleaner Production, vol. 19 no. 1, pp 43-48.

Volodin, P. & Kondyurin, A. (2008). Dewetting of thin polymer film on rough substrate: II. Experiment. J. Phys. D: Appl. Phys., vol. 41 no. 6, pp 065307.

Wu, L., Dong, Z., Kuang, M., Li, Y., Li, F., Jiang, L. & Song, Y. (2015). Printing Patterned Fine 3D Structures by Manipulating the Three Phase Contact Line. Adv. Funct. Mater., vol. 25 no. 15, pp 2237-2242.

Wågberg, P. & Wennerblom, A. (1992). A Correlation between Results Achieved with an Optical Profile Tester, Conventional Paper Evaluation and Printability. Proceedings of: 1992 International Printing & Graphic Arts Conference, 18-21 Oct. 1992, Pittsburgh, USA. Atlanta, USA: TAPPI Press. pp 187-196.

Xiang, Y. & Bousfield, D. W. (2000). Influence of coating structure on ink tack dynamics. J. Pulp Paper Sci., vol. 26 no. 6, pp 221-227.

Xiang, Y. & Bousfield, D. W. (2003). Coating pore structure change after printing. Proceedings of: 8th Advanced coating fundamentals symposium, 8-10 May 2003, Chicago, USA. Atlanta, USA: TAPPI Press. pp 107-112.

Xiang, Y., Bousfield, D. W., Coleman, P. & Osgood, A. (2000). The cause of backtrap mottle: chemical or physical? Proceedings of: 2000 Coating Conference, 1-4 May 2000, Washington, USA. Atlanta, USA: TAPPI Press. pp 45-58.

Xiang, Y., Bousfield, D. W., Hassler, J., Coleman, P. & Osgood, A. (1999). Measurement of local variation of ink tack dynamics. J. Pulp Pap. Sci., vol. 25 no. 9, pp 326-330.

Xiang, Y., Bousfield, D. W., Hayes, P. C. & Kettle, J. (2004). A model to predict ink-setting rates based on pore-size distributions. J. Pulp Paper Sci., vol. 30 no. 5, pp 117-120.

Xue, L. & Han, Y. (2012). Inhibition of dewetting of thin polymer films. Prog. Mater Sci., vol. 57 no. 6, pp 947-979.

Yang, L. (2013). A physical model for liquid movement into a porous substrate under the action of a pressure pulse. Nord. Pulp Paper Res. J., vol. 28 no. 1, pp 94-100.

93

Yang, L., Liu, J. & Gu, L. (2013). Detailed insights to liquid absorption and liquid-paper interaction. In: I'anson, S. J. (ed.), vol. 1. Proceedings of: Advances in pulp and paper research: 15th Fund. Res. Symp. , 8-13 Sept. 2013, Cambridge, UK. Bury, UK: The Pulp and Paper Fundamental Research Society. pp 585-598.

Yilgör, E., Söz, C. K. & Yilgör, I. (2018). Wetting behavior of superhydrophobic poly(methyl methacrylate). Prog. Org. Coat.

Young, T. (1805). An essay on the cohesion of fluids. Philosophical Transactions of the Royal Society of London, vol. 95, pp 65-87.

Yuanhong, Z. & Zhihong, L. (2008). The Evaluation Methods for Ink Trapping Based on Ideal Overprint and Psychological Chroma. vol. 6. Proceedings of: International Conference on Computer Science and Software Engineering, CSSE 2008, 12-14 Dec. 2008, Wuhan, China. IEEE Computer Society. pp 389-392.

Zang, H. (1993). A new approach for modelling ink transfer. Tappi J., vol. 76 no. 7, pp 97-103.

Zang, Y. H. & Aspler, J. S. (1995a). Factors that affect the flexographic printability of linerboards. Tappi J., vol. 78 no. 10, pp 240-223.

Zang, Y. H. & Aspler, J. S. (1995b). The influence of coating structure on the ink receptivity and print gloss of model clay coatings. Tappi J., vol. 78 no. 1, pp 147-154.

94

Appendix A: Laboratory flexographic printing of pilot

coated papers Details regarding print settings and print quality analysis of the unpublished trial

(flexographic printing of pilot-coated papers) are presented below:

• Press: F1 Printability tester (IGT Testing Systems, Amsterdam, Netherlands).

• Printing plate and tape: UVR, 1.14 mm thick (MacDermid Inc., Waterbury,

USA) mounted with 1020 (3M Europe N.V./S.A, Diegem, Germany).

• Ink: PremoNova Pack NF cyan (Flint Group Sweden AB, Trelleborg,

Sweden).

• Ceramic Anilox: volume 2.7 cm³m-2 (IGT Testing Systems, Amsterdam,

Netherlands).

• Nip force: 30 N in anilox-plate and plate-sample nip.

• Press speed: 0.8 ms-1.

• Scanning: RGB, 600 dpi. Flatbed scanner Perfection V850 Pro (Epson

Europe B.V., Amsterdam, Netherlands)

• Print mottle analysis: STFI Mottling Expert (RISE Bioeconomy, Stockholm,

Sweden). RGB images from the scanner were converted to grey-scale before

calibration to reflectance. Wavelength ranges: 0.25 mm to 1 mm and 1 mm

to 8 mm. Averages calculated for a minimum of six (up to nine) areas, each

was 22 mm x 22 mm.

One-way independent sample analysis of variance (ANOVA) was performed to

identify if mean values of samples’ print mottle differed from each other. The

null hypothesis was that there is no significant difference between the sample

means. F-ratios were calculated according to:

𝐹 =𝑏𝑒𝑡𝑤𝑒𝑒𝑛 𝑔𝑟𝑜𝑢𝑝 𝑣𝑎𝑟𝑖𝑎𝑏𝑖𝑙𝑖𝑡𝑦

𝑤𝑖𝑡ℎ𝑖𝑛 𝑔𝑟𝑜𝑢𝑝 𝑣𝑎𝑟𝑖𝑎𝑏𝑖𝑙𝑖𝑡𝑦=

∑ 𝑛𝑖(�̅�𝑖 −�̅�)2 (𝐾−1)⁄𝐾

𝑖=1

∑ ∑ (𝑌𝑖𝑗−�̅�𝑖 )2

(𝑁−𝐾)⁄𝑛𝑖𝑗=1

𝐾𝑖=1

[Eq. 33]

where �̅�𝑖 is the sample mean in i-th group, 𝑛𝑖 is the number of observations

(subareas) in the i-th group, �̅� is the overall mean (all groups together), K is the

number of groups, 𝑌𝑖𝑗 is the j-th observation in the i-th group and N is the overall

sample size (total number of subareas). The calculated F-ratios were compared

with the probability density function with df1 (𝑑𝑓1 = 𝐾 − 1) and df2 (𝑑𝑓2 =

𝑁 − 𝐾) degrees of freedom to determine the likelihood of each F-ratio (that is,

the p-value). When the means were significantly separated, the F-ratio was larger

and the p-value smaller. P-values < 0.05 corresponds to 5 % significance level.

95

Appendix B: List of nomenclature 𝐴 Unit area (eq. 3, 23)

𝐴𝑖 Cross-sectional area of capillary

with diameter di (Eq. 5)

𝑏 Amount of immobilised ink

𝐶 A constant (Eq. 3)

𝐶𝑉𝐷𝑟𝑒𝑙 Coefficient of variation of

relative print density

𝐶𝑉𝑜𝑣𝑒𝑟 Coefficient of variation of

overprint density

𝐶𝑉𝑅0 Coefficient of variation of

reflectance of unprinted area

𝐶𝑉𝑅𝑝𝑟𝑖𝑛𝑡 Coefficient of variation of

reflectance of printed area

𝑑𝑓1, 𝑑𝑓2 Degrees of freedom 1 and 2

𝑑ℎ 𝑑𝑡⁄ Penetration rate

𝐷𝑖 Capillary diameter (i-th)

�̅�𝑜𝑣𝑒𝑟 Mean overprint density

�̅�𝑟𝑒𝑙 Mean relative print density,

subscript 1, 12, 2 to define

monolayer of ink 1, bilayer of

ink 1+2, monolayer of ink 2

𝑑𝑉 𝑑𝑡⁄ Volumetric flow rate

𝐸 Coating thickness

𝑓 Split factor

𝐹 Force (Eq. 3 and 23)

𝐹 F-ratio (Eq. 33)

∆𝐺𝑖 Gibbs free energy, in capillary

with diameter 𝑑𝑖

ℎ Liquid column (height)

ℎ𝑖 Length filled with liquid, in

capillary with diameter 𝑑𝑖

𝐼 Number of laboratories

𝑘 Constant (Eq. 1-2)

𝐾 Number of groups

𝐿 Length of pores in coating layer

𝑁 Number of capillaries (Eq. 12)

𝑁 Overall sample size (Eq. 33)

𝑛 Number of sub-areas per

laboratory (Eq. 31-32)

𝑛𝑖 Number of observations in i-th

group (Eq. 33)

𝑃𝐶 Capillary pressure

𝑃𝐸 External pressure

𝑃𝐹(𝑡) Time dependent pressure

function

𝑃𝐹 Pressure loss

𝑄 Volumetric flow rate

𝑟 Radius

�̅�𝑜 Mean reflectance of unprinted

area

�̅�𝑝𝑟𝑖𝑛𝑡 Mean reflectance of printed area

𝑆 Spreading coefficient

𝑠𝑖 Standard deviation of

measurements at i-th laboratory

𝑠𝑅 Reproducibility standard

deviation

𝑠𝑟 Repeatability standard deviation

𝑠�̅� Standard deviation between the

laboratories

96

𝑡 Film thickness (Eq. 3)

𝑡 Time (Eq. 7-13)

𝑇𝑅̅̅ ̅̅ Mean trapping ratio

𝑉 Separation speed (Eq. 3)

𝑉 Volume in capillary (Eq. 6)

𝑉 Volume per unit area (Eq. 12-13)

𝑉𝑖 Liquid volume, in capillary with

diameter di (Eq. 5)

𝑊𝐴 Work of adhesion

𝑊𝐶 Work of cohesion

𝑥 Initial ink amount on plate

𝑦 Ink transfer per unit area

�̅� Overall mean

�̅�𝑖 Sample mean in i-th group

𝑌𝑖𝑗 j-th observation in the i-th group

𝜀 Porosity

𝛾 Liquid surface tension or solid

surface energy

𝛾𝑙𝑣 Interfacial energy between liquid

and vapour

𝛾𝑠𝑙 Interfacial energy between solid

and liquid

𝛾𝑠𝑣 Interfacial energy between solid

and vapour

𝛾𝑠𝑇𝑂𝑇, 𝛾𝑙

𝑇𝑂𝑇 Total surface energy/tension of

solid (s)/liquid (l)

𝛾𝑠𝑎𝑏, 𝛾𝑙

𝑎𝑏 Acid-base component of surface

energy/tension of


𝛾𝑠𝑑 , 𝛾𝑙

𝑑 Dispersive component of surface

energy/tension of


𝛾𝑠𝐿𝑊, 𝛾𝑙

𝐿𝑊 Lifshitz-van der waals

component of surface

energy/tension of


𝛾𝑠𝑝

, 𝛾𝑙𝑝

Polar component of surface

energy/tension of


𝛾𝑠−, 𝛾𝑙

− Base component of surface

energy/tension of


𝛾𝑠+, 𝛾𝑙

+ Acid component of surface

energy/tension of


�̇� Shear rate

𝜂 Liquid viscosity

𝜃 Contact angle

𝜌 Mass density

𝜌(𝐷1𝑟𝑒𝑙,𝐷12

𝑟𝑒𝑙) Correlation coefficient between

bilayer and ink layer one

𝜎𝐷𝑟𝑒𝑙 Standard deviation of relative

optical density. Subscript 1 and

12 to define monolayer of ink 1,

bilayer of ink 1+2

𝜎𝑜𝑣𝑒𝑟 Standard deviation of overprint

density

𝜎𝑅(𝑝𝑟𝑖𝑛𝑡) Standard deviation of reflectance

of printed area

𝜎𝑅0 Standard deviation of reflectance

of unprinted area

𝜎𝑇𝑅 Standard deviation of ink

trapping

𝜏 Tortuosity factor (Eq. 13)

𝜏 Shear stress (Eq. 23)


Sofia Thorman

Sofia Thorm

an | Where did the ink go? | 2018:52

Where did the ink go?

The appearance of a print is affected by the individual ink layers, and when the ink is unevenly distributed it will bring negative consequences. As an extension, print quality can affect how packaging is perceived and may even create ideas about the product inside the packaging. Therefore, it is important to understand the reasons behind poor and uneven print and to ensure that the board substrate provides a stable quality.

This thesis puts focus on how the absorbency of a coated substrate impacts the ink distribution in flexographic printing. Both monochrome and multicoloured areas are dealt with. To generate an understanding of these interactions, two new measurement techniques were developed where the non-uniformities in liquid absorption and ink trapping can be characterised. Surface roughness often has a strong influence on the print and this may hide interactions from ink absorbency. Therefore, the materials used here were carefully selected (or designed) to highlight the way liquid uptake affects the ink layers. Interactions between substrate and printing ink were studied in several ways, from monochrome printing at 0.5 ms-1

in laboratory-scale to multicolour printing at 10 ms-1 in production-scale.


Faculty of Health, Science and Technology

Chemical Engineering


ISSN 1403-8099

ISBN 978-91-7063-987-6 (pdf)

ISBN 978-91-7063-892-3 (print)



VIPP The industrial graduate school. Industry and academia in synergy for tomorrow’s solutions.

Print & layoutUniversitetstryckeriet, Karlstad 2018

Cover photo:Hans Christiansson

The appearance of a print is affected by the individual ink layers, and when the

ink is unevenly distributed it will bring negative consequences. As an extension,

print quality can affect how packaging is perceived and may even create ideas

about the product inside the packaging. Therefore, it is important to understand

the reasons behind poor and uneven print and to ensure that the board substrate

provides a stable quality.

This thesis puts focus on how the absorbency of a coated substrate impacts the

ink distribution in flexographic printing. Both monochrome and multicoloured

areas are dealt with. To generate an understanding of these interactions, two new

measurement techniques were developed where the non-uniformities in liquid

absorption and ink trapping can be characterised. Surface roughness often has a

strong influence on the print and this may hide interactions from ink absorbency.

Therefore, the materials used here were carefully selected (or designed) to highlight

the way liquid uptake affects the ink layers. Interactions between substrate and

printing ink were studied in several ways, from monochrome printing at 0.5 ms-1 in

laboratory-scale to multicolour printing at 10 ms-1 in production-scale.

Sofia Thorman is working at RISE Bioeconomy, Stockholm, where she started in 2007. Her main focus is on print quality related issues and ways to characterise and predict printability of paper and board substrates. Interest in the science of graphical arts started when obtaining a Bachelor of Science in Graphic Technology from Dalarna University, Sweden, in 2003 and after that getting acquainted with research and development as a trainee at Stora Enso.

Sofia Thorman


So

fia T

horm

an W

here did the ink go?

KAU.SE/EN/VIPPDOCTORAL THESIS | Karlstad University Studies | 2018:52

ISSN 1403-8099

ISBN 978-91-7063-987-6 (pdf)

ISBN 978-91-7063-892-3 (print)

The appearance of a print is affected by Where did the ink...

Documents

Transcript of The appearance of a print is affected by Where did the ink...